COMPOSITION AND EVOLUTION OF THE PROTEROZOIC VIOOLSDRIF BATHOLITH (INCLUDING THE ORANGE RIVER GROUP), NORTHERN CAPE PROVINCE, SOUTH AFRICA Hendrik Minnaar Submitted for the degree of philosophiae doctor in the Department of Geology University of the Free State December 2011 Supervisor: Prof. A.E. Schoch CONTENTS Abstract i Samevatting iv 1. INTRODUCTION 1 1.1 MOTIVATION, AIMS AND LIMITATIONS 1 1.2 LOCATION 3 1.3 PREVIOUS WORK 5 1.4 THE RICHTERSVELD SUBPROVINCE IN GLOBAL CONTEXT 10 2. FIELD DESCRIPTIONS 13 2.1 EXTENT, SHAPE AND FORM 13 2.2 DISTINCTION BETWEEN THE HAIB AND RICHTERSVELD AREAS 15 2.3 ORANGE RIVER GROUP 16 2.3.1 Haib Subgroup 19 2.3.1.1 Tsams Formation 20 2.3.1.2 Nous Formation 21 2.3.2 De Hoop Subgroup 24 2.3.2.1 Windvlakte Formation 26 2.3.2.2 Paradys River Formation 29 2.3.2.3 Abiekwa River Formation 30 2.3.2.4 Klipneus Formation 31 2.3.2.5 Kook River Formation 33 2.3.2.6 Kuams River Formation 33 2.3.2.7 Rosyntjieberg Formation 34 2.4 VIOOLSDRIF SUITE 37 2.4.1 Vuurdood Subsuite 38 2.4.2 Goodhouse Subsuite 41 2.4.2.1 Khoromus Tonalite 43 2.4.2.2 Xaminxaip River Granodiorite 47 2.4.2.3 Blockwerf Migmatite 50 2.4.2.4 Gaarseep Granodiorite 53 2.4.2.5 Hoogoor Granite 55 2.4.3 Ramansdrif Subsuite 57 2.4.3.1 Ghaams Granite 58 2.4.3.2 Sout Granite 60 3. PETROGRAPHY 61 3.1 ORANGE RIVER GROUP 61 3.2 VIOOLSDRIF SUITE 64 3.2.1 Vuurdood Subsuite 64 3.2.2 Goodhouse Subsuite 66 3.2.3 Ramansdrif Subsuite 71 4. GEOCHEMISTRY 77 4.1 GEOCHEMICAL VARIATIONS 77 4.1.1 Classification diagrams 77 4.1.2 Harker diagrams 82 4.1.3 Eigenvectors and eigenvalues 87 4.1.4 Multi-element diagrams 89 4.1.5 REE patterns 98 4.2 MAGMATIC PROCESSES 105 4.2.1 Vuurdood Subsuite in relation to the rest of the batholith 105 4.2.2 Vuurdood Subsuite as remnants of the primary magmas 111 4.2.3 Radiometric evidence reviewed 114 4.2.4 Unique signatures of the subduction environment 115 4.2.5 Origin of the Blockwerf Migmatite 119 4.2.6 Ramansdrif Subsuite 124 4.3 THE ORIGINAL TECTONIC ENVIRONMENT 125 4.4 TTG AND ADAKITE IN THE RICHTERSVELD SUBPROVINCE 134 5. CONCLUSIONS 146 5.1 SUBDIVISIONS 146 5.2 MAGMATIC PROCESSES AND TECTONIC ENVIRONMENT 149 5.3 THE RICHTERSVELD SUBPROVINCE IN ITS REGIONAL SETTING 155 6. ACKNOWLEDGEMENTS 158 REFERENCES APPENDIX 1: ANALYSES, SAMPLE LOCALITIES, DATA TREATMENT AND PROCESSING APPENDIX 2: NEW GEOCHRON DATA MAP (IN FOLDER) i Abstract The Vioolsdrif Suite and Orange River Group repnret sgeenetically related calcalkaline plutonic and volcano-sedimentary assemblages oaf ePoaplroterozoic age formed during the Orange River orogeny. Together they occup yla trhge st part of the Richtersveld Subprovince – a unique tectono-stratigraphic terer.a nRadiometric data indicate the period of formation roughly between 2.0-1.73 Gah.e Tsubprovince has been vastly eroded and isolated from its original tectonic ernovniment by subsequent tectonic processes, leaving a relatively small portion os fo irtiginal extent for investigation. Previous studies have dealt with limited partsh oef stubprovince and although informal subdivisions of the Orange River Group and Viooifls Sdur ite are generally in use, some correlations and further subdivisions remained crovnet rsial. This study has two main aims, viz., to propose formal subdivisions of twhoe ut nits and to investigate the magmatic processes and original tectonic environtm oef ntheir formation. Geochemical evidence is presented here to support the propsousbedi visions, which were previously based entirely on field evidence. The subdivis liaorngsely follow that of previous studies. The Orange River Group is subdivided into the Haanibd De Hoop Subgroups. Geochemical evidence show that these two subgrdoiuffpesr in the magmatic processes that led to their formation. The Haib Subgroup wshs oa genetic gradational relationship with the Vuurdood Subsuite, which is regarded amsn raents of the primary magmas. The De Hoop Subgroup does not display this relation swhiitph the Vuurdood Subsuite. The Vioolsdrif Suite is subdivided into the Vuurd,o Gooodhouse and Ramansdrif Subsuites. Mafic-ultramafic bodies of the Vuurd oSoudbsuite are regarded as remnants of the primary magmas. This is based on multmi-elnet variation diagram patterns and comparison to source magmas in modern island aMrcOsR (B). Previous studies have also shown that initial isotope ratios for the Vduourod Subsuite are similar to those of the Goodhouse Subsuite and Orange River Group, rel athtinegm to a similar source. Dark mineral cumulate material are contained in the mc-auflitramafic bodies. ii The Goodhouse Subsuite is subdivided into the Kmhuosro Tonalite, Blockwerf Migmatite, Xaminxaip River Granodiorite, Gaarseep Granodio raitned Hoogoor Granite. The Khoromus Tonalite is identified as the oldest uwnitiht in the subsuite based on contact relationships as well as radiometric data, whichn tcinouously render older ages for the Khoromus Tonalite compared to the other units wni thie subsuite. Certain field and petrological observations in this unit may be inptrerted as products of magma mixing processes. The Blockwerf Migmatite is distingudis fhroem the other units by its migmatitic character and anomalous La/Yb ratiosh.e Tunit is identified as a possible remnant volcanic centre. The Xaminxaip River Grdaionroite is interpreted as a subvolcanic unit and is characterised by the depvmeleont of migmatite in places. This migmatite is attributed to metamorphic conditionhsi cwh locally reached high grade in an orogenic geothermal regime. The Gaarseep Griaonrioted represents the main phase of the Vioolsdrif Suite. Its compositional rangnec lui des all those represented by the other units individually from gabbro through dioer,i ttonalite and granodiorite to granite. Its development spans almost the entire evolutiyo nhaisrtory of the Richtersveld Subprovince as a whole. The Hoogoor Granite islu idnecd in the Goodhouse Subsuite based on the geochemical variation patterns andil abvlae radiometric evidence. The Ramansdrif Subsuite is subdivided into the Gmhsa and Sout Granites based on grain size variation and petrological evidence. e T shubsuite could have been formed by partial melting of the older plutonic phases of Vthieoolsdrif Suite. The deformation associated with this partial melting event has bneoet n identified. Previous studies have related the evolution ofR tihceh tersveld Subprovince to modern subduction zone magmatism similar to that of thdee Aan volcanic arc. This is largely supported by the current study, however, a chan gmea igmatic processes and the tectonic environment can be observed in the geoichael mvariation patterns. Multi- element diagrams show patterns typical of subdnu cztoione magmatism for both the Orange River Group and Vioolsdrif Suite. Duringe tihnitial stages, primary magmas, now represented by the Vuurdood Subsuite, werev edde rfriom a depleted mantle reservoir. The first volcanic eruptions – thos eth oef Haib Subgroup – represent iii fractional crystallization products off the prima mryagmas. With progressive development of the arc, newly formed crust wasi ncuoonut sly recycled back into the mantle and crustal partial melting led to a chaning em agmatic processes with magma mixing and contamination becoming increasingly imrtapnot. Tectonic discrimination diagrams suggest that the initial stages of theh oblaitth development may be compared to a primitive continental arc, while the later stag mesay be be compared to a typical continental arc. iv Samevatting Die Vioolsdrif Suite en Oranjerivier Groep verteeonowrdig Palaeoproterosoïese kalkalkaliese plutoniese en vulkanies-sedimentêperee novolgings wat gevorm is tydens die Oranjerivier orogenese. Tesame beslaan huiell eg rdootste deel van die Richtersveld Subprovinsie – ‘n unieke tektonies-stratigrafieesrere tin. Volgens radiometriese data was die tydperk van ontwikkeling tussen ongeveer 27.03- G1.a. Die subprovinsie is grootliks verweer en latere tektoniese prosesse het dit ivea ono drspronklike tektoniese terrein geïsoleer sodat ‘n relatiewe klein gedeelte virt ubdeesring oorgebly het. Vorige studies het slegs oor beperkte gedeeltes van die subprioev gineshandel. Alhoewel informele onderverdelings van die Oranjerivier Groep en Vsiodorilf Suite algemeen in gebruik is, is sekere korrelasies en verdere onderverdelings noongtr okversieël. Die doel van die huidige studie is veral tweërlei nl., om algemeean-vaaarde onderverdelings van die twee eenhede voor te stel en om die tektonies-mtaiegsme aprosesse waardeur hulle ontwikkel het te ondersoek. Geochemiese data hwioerd a angebied om die onderverdeling, wat voorheen geheel en al op verwldavnetskappe berus het, te steun. Die voorgestelde onderverdeling wyk nie noemensdwiga aarf van dié in vorige studies nie. Die Oranjerivier Groep word onderverdeel in die Hb aein De Hoop Subgroepe. Geochemiese data dui daarop dat dié twee subgrdoeupre v erskillende magmatiese prosesse ontstaan het en dat die tektoniese ter rweainarin hulle ontstaan het nie heeltemal dieselfde was nie. Die Vioolsdrif Suite word onderverdeel in die Vuouord, Goodhouse en Ramansdrif Subsuites. Die mafies-ultramafiese liggame va nV duiuerdood Subsuite word beskou as oorblyfsels van die primêre magmas en dit beva tk ouomkulate van donkerminerale. Die Vuurdood Subsuite toon ‘n genetiese verwantska pd ime eHtaib Subgroep maar nie met die De Hoop Subgroep nie. Die verwantskap tussiee nV udurdood Subsuite en die Haib Subgroep kan moontlik die gevolg van fraksioneilset aklrlisasie wees. v Die Goodhouse Subsuite word onderverdeel in dier oKmhous Tonaliet, Blockwerf Migmatiet, Xaminxaiprivier Granodioriet, Gaarseepra Gnodioriet en Hoogoor Graniet. Die Khoromus Tonaliet word beskou as die oudsteh eiedn in die subsuite op grond van kontakverwantskappe sowel as radiometriese dataaa. r Dis veld- en petrografiese getuienisse wat daarop dui dat magma vermengingssepsrsoe ‘n rol gespeel het in die vorming van hierdie eenheid. Die Blockwerf Migmeta wti ord van die ander eenhede onderskei deur sy migmatitiese karakter en anomLa/lYe b verhoudings. Die eenheid word geïdentifiseer as ‘n moontlike vulkaankeëlb-olyofsf el. Die Xaminxaiprivier Granodioriet word geïnterpreteer as ‘n subvulkanei esenheid en word gekenmerk deur die ontwikkeling van migmatiet op plekke. Hiermdiieg matiet word geïnterpreteer as die produk van lokale hoëgraadse metamorfe toestandt eg ewhaeers het in die orogene subduksieregime. Die Gaarseep Granodioriet ish doioef fase van ontwikkeling van die Vioolsdrif Suite. Dit oorheers in volume en oourvelel met feitlik die hele vormingsgeskiedenis van die Richtersveld Subprioev. i nDsit bevat die volledige samestellingsreeks wat in die ander plutoniese edeen hindividueel verteenwoordig word van gabbro deur dioriet, tonaliet en granodiorioet gt raniet. Die Hoogoor Graniet word in die Goodhouse Subsuite ingesluit op grond vaonc hgemiese variasiepatrone en beskikbare radiometriese data. Die Ramansdrif Subsuite word onderverdeel in diea aGmhs en Sout Graniete op grond van korrelgroottevariasies en petrologiese getusi.e nDi ie subsuite is skynbaar gevorm deur gedeeltelike insmelting van die ouer Goodh oSuusbesuite plutoniese eenhede. Die vervormingsepsiode wat met hierdie gedeeltelikme einltsing gepaard sou gaan, is nie geïdentifiseer nie. Vorige studies vergelyk die ontwikkeling van diceh Rteirsveld Subprovinsie met tektonies- magmatiese prosesse wat werksaam is in modernaea vnublkoë soos die Andesboog. Hierdie vergelyking word grootliks deur die huid isgteudie ondersteun maar ‘n verandering in die magmatiese prosesse en dien tieeksteo omgewing gedurende die ontwikkelingsgeskiedenis is waarneembaar in diec hgemo iese data. Multi-element diagramme vertoon tipiese subduksiesone variasrioenpea tvir beide die vi Oranjeriviergroep en die Vioolsdrif Suite. Gedudren die aanvangstadium is primêre magmas gegenereer in die mantel, wat ‘n verarmdmee satelling gehad het. Die Vuurdood Subsuite verteenwoordig oorblyfsels vaen p driimêre magmas. Die eerste vulkaniese lawas – dié van die Haib Subgroep – pwroadsu kte van fraksionele kristallisasie vanaf die primêre magmas. Met deiele gidelike ontwikkeling van die vulkaanboog is nuwe kors in groter-wordende volu gmeepsroduseer. Hierdie kors is voortdurend in die mantel herwin en gedeelteliksem inelting het daartoe gelei dat prosesse soos magma vermenging en kontaminasineg brieklear geword het as fraksionele kristallisasie in die bepaling van die magmasamlleinsgtes. Tektoniese diskriminasiediagramme dui daarop dat die aanvaandgisutms van die batoliet se ontwikkeling moontlik plaasgevind het in ‘n primewitie kontinentale boog, terwyl die latere stadiums in ‘n gewone kontinentale boog spglaeavind het. 1 1. INTRODUCTION The Richtersveld Subprovince is largely occupie dth bey Vioolsdrif Suite and Orange River Group. Together these two units record Poaplraoeterozoic juvenile crust formation processes. At the current level of erosion, thoeo Vlsidrif Suite (in particular the Goodhouse Subsuite) volumetrically predominatesh wthiet Orange River Group occurring as rafts within the plutonic suite. Ttehrem Vioolsdrif Batholith in the title is meant to include the Orange River Group (with w hiti cish genetically related), which is not mentioned repeatedly. 1.1 MOTIVATION, AIMS AND LIMITATIONS During 2002, the statutory mapping program of thoeu nCcil for Geoscience came to include the extreme northwestern parts of the croyu, nintcluding the current study area. At the time, there were still some unresolved coovnetrsies concerning the Orange River Group and Vioolsdrif Suite and it was decided thaet field mapping should be concentrated on some of these issues. The fatc tth tehsae two units are genetically related and occupy a distinct tectonic subprovince has lboenegn recognized. The recognition of individual units, their stratigraphic positions a tnhde formalization of nomenclature are important aims of statutory mapping programs an dre ncoent compilation in which the two units are studied as a whole, are availabnled.i v iIdual studies were previously concentrated on parts of it. The statutory pro pjerectsented the opportunity to accumulate a large number of samples, field data and new ovabtsioenr s, so that a comprehensive academic study became possible. Field mapping ceonmcemd from 2002 to 2005 and 1:50 000 scale field maps were produced in contsiounlt awith maps produced during previous studies. Aerial photographs of 1:60 0c0a0le s were used in the field. The statutory mapping of 1:50 000 scale is used to ciloem 1p:250 000 scale maps for publishing. Although Jouberet t al. (1986) included the Vioolsdrif Suite and Orangivee Rr Group in a single compilation, some important subdivisions ew aedrded in later studies. The 2 distinction between the Khoromus Tonalite and Geaeapr sGranodiorite was recognized by Prinsloo (1987) and first published by Mareati sa l. (2001). Although Ritter (1980) recognized the Xaminxaip River Granodiorite, covnetrrosies remained concerning the unit’s genesis and stratigraphic position (De Verilsli and Söhnge, 1959; Ritter, 1980; Minnaar et al., 2011). New radiometric age data (U-Pb zirco nL Aby-ICPMS) is presented here as additional evidence as to aittsig srtarphic position (see 2.4.2.2 Xaminxaip River Granodiorite). Controversy alsoig nres over the inclusion of the Hoogoor Granite in the Vioolsdrif Suite (Strydoemt a l., 1987; Moen and Toogood, 2007). This originates from the fact that the Hoorg Granite is pervasively overprinted by Namaqua foliation and metamorphism, which haliste orabted evidence and contact relationships in the field. Colliston and Scho2ch0 0(6; Fig. 5) addressed the effects of high strain deformation on the rocks of the Vioorilfs Sduite. Field observations are the foundation of any geiocalol gstudy, as will also become apparent for this investigation. However, in igunse oterranes, geochemical and isotopic studies have become indispensable additions fiovri nagr rat meaningful conclusions as to the evolutionary history and original tectonic isnegt tof the rock units. The recognition of the results of processes such as fractional crliyzsattaiol n, magma mixing, wall rock assimilation, homogenization in a magma chambertri,a pl amelting, mantle wedge metasomatism, and basaltic underplating, is vasidtleyd by these studies. Indeed, the modeling of modern plate tectonic processes rehleieasv ily, in some cases entirely, on geochemical and isotopic evidence. The evolutionary history of the Richtersveld Subvpinroce also has consequences for the understanding of the surrounding Namaqua Metamco rPprhoivince. Current theories suggest that this tectono-metamorphic province fworamse d by accretion of a number of individual, previously separated, terranes of w hthiceh Richtersveld Subprovince is one. The terrane accretion model has gained wide acncceep tian recent literature but controversies concerning the exact configuratiodn taenrrane boundaries persist. This study aims to make a positive contribution to tnhseig iht into the geological evolution of the entire Namaqua Metamorphic Province. 3 The aims of the current study are: a) To present acceptable subdivisions of the Or aRnivger Group and Vioolsdrif Suite in South Africa. Geochemical data availafrbolem the Namibian side for the two units from previous studies were also includedt hfoer geochemical investigation (see Appendix 1: Analyses, sample localities, data mtreeantt and processing). b) To recognize and study the igneous and tectpornoice sses involved during the formation of the Orange River Group and VioolsdSruifi te. The study suffers from two limitations, viz., incpolmete sampling coverage and limited isotopic analyses, which were the result of budcgoents traints. The Namibian portion of the Richtersveld Subprovince (representing aboluf to hf ait) was not included in this study. For the geochemical study, the analyseRse oidf (1977) on the Namibian side in the Haib area are included. However, due to tchke olaf field data the subdivisions and resulting map does not include the Namibian siRdaed. i ometric age determinations were obtained for only four samples (U-Pb by LA-ICPMS zoirncons). Two of these determinations are owed to collaboration with caoglluees at the University of Gothenburg, Sweden (see 6. Acknowledgements). 1.2 LOCATION The Richtersveld Subprovince is located in northtewrens South Africa and southwestern Namibia, straddling the Orange River from the Reicrshvteld eastwards into the Bushmanland of South Africa (Figure 1.1). For the most part it is a rugged, mountainous itne rsreat in an arid climate with a very low population density and sparse vegetation and a nlifme.a lThe Orange River is the only perennial river in the area. The natural scense rdyo iminated by outcrops, especially in the Haib area (east of the Neint Nababeep plat e Aacuc).ess is relatively good with a well developed network of field tracks, in places onclcy eassible by four-wheel drive vehicle. 4 The tarred road leading south from Vioolsdrif inthtoe Namaqualand interior is the main access route and runs through the batholith tyepae saoruth of Vioolsdrif. There are two important geographic features ins tuhdey area, which also represent geological boundaries. In the Richtersveld, thes yRnotjieberg mountain range includes the highest mountain peaks in the area (Vandebrsetregr ris the highest point at 1 400 m). This mountain range forms an almost inaccessibuleth seorn boundary to the Richtersveld National Park (Figure 1.2). Figure 1.2: Part of the Rosyntjieberg mountaing rea wnhich represents the southern boundary of tchhet eRrisveld National Park. Photo taken in a southerly direnc tfirom Tatasberg in the eastern part of the pa7rk.2 (6150°E; 28.3251°S). Distance to horizon approximately m15. kThe highest mountain peak in this photo is erseepnr ted by Rosyntjieberg (1 329 mamsl). Stretching in an east-west direction, it was alseod u by Ritter (1980) as a geological division, including all the volcanic rocks of ther aOnge River Group to the south thereof in the Windvlakte Formation. Between the settletms eonf Vioolsdrif and Eksteenfontein, the Neint Nababeep Plateau represents the sedirmy esnutcacession of the Nama Group 5 (Germs, 1972; Gresse and Germs, 1993) in the satrueday. It divides the Vioolsdrif Batholith into two distinct parts namely the Hairbe a to the east, and the Richtersveld area to the west. This represents a historicadl ivsuisbion, the basis of which will be discussed in detail later (2.2 Current subdivis.i on) 1.3 PREVIOUS WORK The first comprehensive description of rocks of Vthioeolsdrif Batholith is to be found in the survey mapping of Rogers (1916) in the Richvteeldrs area. In those early years, the volcanic rocks of the current Orange River Groupre w ceorrelated with similar rocks of the “Kheis System” (Rogers, 1911) in the Upingtornie-sPka area (Geverest al., 1937; Gevers, 1941; Von Backström, 1953; De Villiers aSnödh nge,1959; Von Backström and De Villiers, 1972). This correlation was largelays bed on the similar stratigraphic succession and the assumption that the “grey gicn egirsasnite” (current Goodhouse Subsuite) was part of the Archean magmatic cyDcle .V illiers and Burger (1967) obtained an U-Pb zircon age of 1 850 Ma for thersaen igtes, showing that they could not belong to the Archean magmatic cycle. Early in the nineteen-seventies, the Precambriasne aRrech Unit of the University of Cape Town, started studies on the rocks in the currteundty s area. The radiometric age of De Villiers and Burger (1967), along with the lackr oigfo rous evidence for the correlation of the rocks over the large distance between the eRriscvhetld and the Upington-Prieska area, inspired Blignault (1974) to propose a new subdioivni sfor the volcanic pile in the Richtersveld. He included them in the Orange R Givreorup and distinguished between the Haib Subgroup in the Haib area (to the eatsht eo fN eint Nababeep Plateau), which is largely similar to the “Wilgenhout Drift Series” othfe “Kheis System”, and the De Hoop Subgroup in the Richtersveld area (to the wesht eo fN teint Nababeep Plateau), which is largely similar to the “Marydale Series” of the “eKihs System”. The predominantly quartzitic unit of the current Rosyntjieberg Forimona twas termed the Rosynebos Formation. This was changed to the Rosyntjiebeorrgm Fation by Kröner and Blignault (1976) after the name of the mountain range. Baluigltn (1977) subdivided the Haib 6 Subgroup into the predominantly felsic Tsams, arned opminantly mafic Nous, Formations. Ritter (1980), on the basis of texl tduirfaferences and the recognition of protoliths for the metavolcanic rocks, subdividheed Dt e Hoop Subgroup into the Paradys River, Abiekwa River, Klipneus, Kook River and Kusa mRiver Formations. He included all the volcanic rocks of the Orange River Grou pth teo south of the Rosyntjieberg mountain range in the Windvlakte Formation ande sdt atht at the original textures in this formation have been obliterated by deformation manedta morphism. Geverse t al. (1937) and Gevers (1941) regarded the rockse o cf uthrrent Vioolsdrif Suite as a genetically related group; the latter autdheonr tiified the more basic phases as the oldest. Coetzee (1941) recognized the older a gthee o Vf ioolsdrif Suite relative to the Namaqualand gneisses and referred to it as the m“beanst granite”. De Villiers and Söhnge (1959) mapped large areas of the “greys sgince gi ranite” (current Goodhouse Subsuite) and regarded the type area to be so uVthio olfsdrif. Von Backström and De Villiers (1972) referred to the leucogranites oef tVhioolsdrif Suite (current Ramansdrif Subsuite) in the unfoliated Richtersveld domain“r eads granites”, and correlated them with the “pink gneisses” in the foliated Namaquam daoin to the east. De Villiers and Burger (1967) proposed that the “grey gneissic igtera” nbe renamed the “Vioolsdrif Granite”. Blignault (1977, p. 13) defined the current Vioorlisf dSuite on the basis of spatial association and evidence that the various compoonsailt iintrusive types constitute an igneous rock series. The subdivisions of Blign a(1u9lt77) and Ritter (1980) were accepted and formalized by SACS (1980). Strydeot mal. (1987) assigned this rock association suite status and subdivided it into V tuhuerdood, Goodhouse and Ramansdrif Subsuites. Within the Goodhouse Subsuite, Maerta aisl. (2001) were the first to recognise the distinction between the KhoromusG aanadr seep “Gneisses”. Minnaeatr al. (2011) suggested a distinction between the Gh aanmd sSout Granites within the Ramansdrif Subsuite. 7 Reid (1974, 1977, 1979, 1982, 1997) made a grenatrt ibcoution to the understanding of the magmatic evolution of the Vioolsdrif Batholiathn d provided excellent radiometric data (Rb-Sr and U-Pb) for the Haib area. Reid 4(1) 9e7stablished the chronological order within the batholith, which is in accordance witvha ailable field evidence. The eruption of the Orange River volcanics (1 996±15 Ma) waslo fwoel d by the intrusion of the Vuurdood Subsuite (<1 996 Ma) and then the intrnu soifo the Goodhouse Subsuite (1 900±30 Ma). After a 170 m.y. period of igneouse qsucience, the Ramansdrif Subsuite was intruded around 1 730 Ma. Reid (1977) geochemically identified the environmt oefn formation of the batholith as that of an island arc. The study furthermore shdo twheat a model of fractional crystallization can account for the magmatic eviolnu tof most of the various igneous rock types, although not for the batholith as a lwe.h oThe isotopic evidence presented in the study indicated that the members of the Goosdeh oSuubsuite as well as that of the Orange River Group are linked to the same paremnataglm a. A model of fractional crystallization from a high-Mg basaltic magma un tdheer influence of high water pressure, was proposed for the magmatic evolutfi othne o Vioolsdrif Suite. A similar model but under low water pressure, is also appiartoep fror the Orange River Group. Reid (1982) showed that the mafic-ultramafic bo doife tshe Vuurdood Subsuite are also linked to the same parental magma as the GoodhSouubsseu ite and Orange River Group and that the bodies may represent cumulate ma wteirtihailn the early magmas. The study furthermore suggested that the granites of the Rnasdmraif Subsuite were probably derived by partial melting from the earlier toneasli tand granodiorites. Reid (1997) put forward strong evidence from Smi-sNodto pe ratios, that the Vioolsdrif Batholith represents juvenile crust formation froa mbu lk earth-like mantle. He stated that the isotope signatures excluded the possyi boifl itrecycling of older crust into the Vioolsdrif Batholith source material. Nordin (20)0, 9on the other hand, considered the source material to be that of a depleted mantle db aosn Lu-Hf isotope evidence from a 8 single sample from the Haib Subgroup. A singlee irnithed zircon age of 2.7 Ga also suggested the presence of pre-existing Archeant. crus Various previous studies investigated the relathioipn sbetween the largely undeformed, low metamorphic terrane which is the Vioolsdrif hBoaltith type area, and the surrounding Namaqua Metamorphic Province. On regional schaele ,c ot ntrast between these two terranes is rather obvious. Gevetr sa l. (1937) established, on field evidence, that the “Namaqualand Granite and Gneiss” were younger wthhaant is currently termed the Vioolsdrif Suite. Beukes (1973) and Blignaeutl ta l. (1974) mapped the position of isograds in southern Namibia, indicating the metrapmhioc transition from the low-grade Vioolsdrif Batholith type area, to the surroundimnge dium-grade Namaqua Province. Based on this, Beukes (1973) indicated the “Nam aFqrounat” as the boundary between his Richtersveld and Namaqualand “Provinces” (sigeuer Fe 2.1). Kröner and Blignault (1976) distinguished the two areas as two sepatercatoen o-stratigraphic terranes, each with its own unique stratigraphic, structural anedt ammorphic imprints. During the first half of the nineteen-eighties, Bthueshmanland Project (University of the Orange Free State) led to extensive field mappihnigc hw included part of the current study area. By this time, a tectonic model of ep lactcretion was suggested by all the available evidence. The wedge-shaped terrane oiecdc ubpy the Vioolsdrif Suite and Orange River Group (Richtersveld Subprovince) wnatesr pi reted as a terrane which was thrusted upon the Namaqua Province, being pres efrovmed the Namaqua deformation and metamorphism (Blignauelt al., 1983; Joubert, 1986; Van der Merwe, 1986). e Rr itt (1980) did not agree with this model, claiming tthhaet differences in deformation, metamorphism and age could be explained by diffceerse nin crustal level. Blignaueltt al. (1983), by field mapping, extended the Vioolsdruifi tSe eastward from the type area across Beukes’s (1973) Namaqua Front, into the Nqauma aProvince as far as the vicinity of Onseepkans (Colliston and Schoch, 2006; seer eF i2g.u1). They identified the Lower Fish River thrust as the northern boundary whiele G throothoek thrust (Van der Merwe, 1986) was identified as the southern boundary F(sigeuer e 2.1). The eastward extension of the Vioolsdrif Suite was not accepted by Moend aTnoogood (2007), claiming that 9 their “Hoogoor Suite” could be distinguished frohme tVioolsdrif Suite on the basis of field evidence (this will be discussed in more dil eutnader 3.2.3.2.4 Hoogoor Granite). This controversy was largely the result of the paesirve overprint of the Namaqua deformation and metamorphism on the proposed Horo Sguoiote. In a global context, following various studies s uacsh those by Stoweet al. (1984), Hartnadye t al. (1985), Jouberet t al. (1986) and Piper (2000), the Richtersveld Subprovince is currently interpreted by most recsheearrs to represent a Kheisian (2.0-1.73 Ga) tectono-stratigraphic terrane which is surroeudn adlong tectonic boundaries by the 1.3-1.0 Ga Namaqua Metamorphic Province. Igneocukss r of the Vioolsdrif Suite and Orange River Group occupy most of the RichtersvSeulbdp rovince and record a major juvenile crust-formation event, while the Namaquroav Pince correlates with the global Grenville and Kibaran orogenies of North Americad aCnentral Africa respectively. The latter orogenies record the assembly of the supnetirnceont Rodinia. On the evolutionary history of the Vioolsdrif Su aitend Orange River Group, it has always been generally agreed that the environmfe fnotr moation was that of an island arc, as evidenced by the calcalkaline bulk compositinodn tahe stratigraphy. Ritter (1983) compared it specifically to a Cordilleran-type isnegt,t such as typified by the modern Andes. However, some controversy still prevaRilse.i d (1977, 1997) propose the igneous rocks of the Richtersveld Subprovince ptore rseent a period of juvenile crust formation from a bulk earth-model mantle in an a wrehaere no crust existed before. Nordin (2009), however (as noted before; p.8), voind eence of a single relict Archean zircon, as well as Lu-Hf model ages, claims thaetv piorus crust must have been present. Noteworthy is the fact that no basement to the Ogera Rniver Group has been identified as yet. Miller (2008) suggests the volcanics mighvt eh abeen deposited on an ocean floor. Isotopic and radiometric age data are availablme fvroarious previous studies including Reid (1975, 1977, 1982, 1997), Allsoeptp a l. (1979), Barton (1983) and Krönetr al. (1983). 10 The Vioolsdrif Batholith is host to numerous deptso soif economic important minerals. A thorough investigation of the tungsten depositsi c(hw hwere exploited in earlier days) has been made by Bowles (1988). Some of the more psrionmg iporphyry-type copper deposits have been investigated (e.g., Minnitt 1).9 79 1.4 THE RICHTERSVELD SUBPROVINCE IN GLOBAL CONTEXT Various previous studies (e.g., Blignault, 1977i;d R, e1977; Ritter, 1980; Barton, 1983) have identified the Vioolsdrif Batholith as a calklcaaline batholith formed in a volcanic arc environment. On the current globe, two typfe ssu obduction-related arcs are recognized, viz., the intra-ocean island arc ( eK.egr.,madec, Tonga, Marianas, Izu-Bonin and Kurile arcs) and the continental arc, wherea noicce crust is subducted beneath continental crust (e.g., Lesser Antilles, Cascaadneds Andes arcs). A third orogenic arc – the continent-continent collision arc (e.g., them Haliayas) – represents an area where the ocean separating the two continents has been dyedst,r soubduction has ended and tectonism is dominated by metamorphic processedsin lge ato the production of S-type dominated magmatism. When the Richtersveld Subprovince is considereitds irne gional setting within the Namaqua Metamorphic Province, formation of the Vlsiodorif Batholith itself coincides with a global era of Late Archean-early Protero zcoruicst formation events around 2.0-1.9 Ga, while the Namaqua Metamorphic Province coinsc iwdieth the Grenville orogeny (1.3-1.0 Ga) during which the supercontinent Road winai s assembled (Condie, 2005b). According to Condieo (p cit.), it is recognized that continental crust was aadlrye present on the globe by 4.0 Ga, probably even 4.4 Ga. eEnvcide indicate that crust formation processes during the Archean were much more aicnt itvhee northern hemisphere than in the south (where the Vioolsdrif Batholith was fordm).e In North America, about five provinces are recognized which were assembled gd uthrien same time in which only a part of the Vioolsdrif Batholith was developing 9(15-.1.75 Ga). They include the Rae and Hearne provinces (both Archean), the Trans-oHnu dpsrovince (up to 500 km wide) and the Penokean and Yavapai provinces (Conodpi ec,it .). Suturing of the latter two 11 provinces occurred at 1.90 and 1.75 Ga respec t(ivCeolyndie, 2005b). It seems therefore that much less older crust, if any, was presentht ein s tudy area compared to the northern hemisphere, during the time of formation of the oVlsiodrif Batholith. Reconstructions of Rodinia also support this as the continental cirnu tsht e northern hemisphere already occupied a much larger area than in the south;e Lnatiuar already dominated the crustal configuration at that time. Hamilton (1998) put forward strong evidence thatd meron plate tectonic processes (rifting and subduction) could not have been ac itniv tehe Archean, but that they were operating by about 2.0 Ga, which is the time ofe ot nosf the evolution of the Vioolsdrif Batholith. Therefore, subduction-related tectoannicd magmatic theories would be realistic to account for the development of theo Vlsiodrif Batholith. According to Miyashiro (1974) the main rocks in island arc snegtsti are usually basalts and basaltic andesites of the tholeiite series while those int icnoental arc settings are andesites and dacites of the tholeiite and calcalkaline seriTehs.e main volcanic rocks in continental margins are andesites, dacites and rhyolites o cf athlcealkaline series and the proportion of calcalkaline rocks among all the volcanic rotceknsd s to increase with advancing development of continental crust. Therefore, b aosne tdhe absence of basaltic volcanic rocks and the proportion of calcalkaline rocks eotfw been 60-95% (Miyashiro, 1974), the bulk of the Vioolsdrif Batholith can be compared m todern continental arcs such as typified by the Andean arc, as suggested by presv siotudies. This is as opposed to the island arc settings where basaltic composition sim apreortant and the proportion of calcalkaline rocks does not exceed 50% (common1l0y %0-). Reid (1997), however, concluded that Sm-Nd isot ospigicnatures preclude the possibility of incorporation of previously existing continen ctarul st. Initial8 7Sr/86Sr values from Reid (1977) support this with values for all of tVhueurdood Subsuite, the Orange River Group and the Goodhouse Subsuite being low (~0.. 7 V0a3l)ues for the Ramansdrif Subsuite does suggest crustal reworking (~0.7R07e)i.d (1997) concluded that the best interpretation of the isotopic signatures involvseligsh t enrichment in radiogenic Pb and Sr by subduction zone metasomatism of a mantle ew ewditgh an undifferentiated bulk 12 earth signature. He proposed that the evolutio tnh eo fVioolsdrif Batholith involved juvenile crust formation from the mantle with suqbusent reworking of newly-formed crust to produce felsic magmatism. The Vioolsdrif Batholith provides evidence on twmop iortant issues relating to Proterozoic tectonic and magmatic processes. l yF, irthstere is the question of whether pre-existing crust was present in the study arieoar ptor the development of the batholith. No evidence of such pre-existing crust and base mtoe tnhte Orange River Group has as yet been identified in the field. A single inherdit zircon age of 2.7 Ga in a lava sample from the Haib Subgroup (Nordin, 2009) currently rbse tahe only testimony to the possibility of Archean crust in the study area. - NSdm isotope signatures, however, strongly argue against the presence of pre-exi sctriunsgt in the area (Reid, 1997). Secondly, the Vioolsdrif Batholith provides eviden acs to the tectonic and magmatic processes which were active during a time whene athrteh ’s crust-formation processes apparently changed dramatically. Various linees voidf ence show a distinction between the Archean processes which produced mainly TTGak (iatidc) magmas, and modern processes which produce mainly calcalkaline mag m Tahse.se changes, in global context, approximately spans the evolutionary history of Vthioeolsdrif Batholith. 13 2. FIELD DESCRIPTIONS A map of the Vioolsdrif Suite and Orange River Gpr oinu South Africa, with the subdivisions as motivated for in this study, iss perneted in a folder at the back of this thesis. Comprehensive field descriptions are g inve pnrevious studies and recent reviews of such descriptions can be found in Moen and Tood g(o2007), Miller (2008) and Minnaar et al. (2011). Only observations which are consideereledv rant to the current study (including those from previous work and frtohme current study), will be discussed here. 2.1 EXTENT, SHAPE AND FORM Consensus has not been reached concerning the abroieusn odf the Vioolsdrif Batholith. Although most current researchers agree that tthheo lbitah is tectonically bounded within the Richtersveld Subprovince between the Lower RFiisvher thrust in the north (Namibia) and Groothoek thrust in the south (RSA), observnast iforom various studies suggest that it might in fact extend into the Namaqua Provin cTeh.i s is not, however, a major concern of the present study. The extent of the batho(ilnitchl uding the Orange River Group) according to most current literature, is preseninte Fdi gure 2.1. The type area of the batholith is defined as thlaet ivre ly undeformed and low- metamorphic-grade part of the batholith. Along bthoeundaries of the batholith, it has been subjected to deformation and metamorphismhe o Nf tamaqua orogeny, the foliation varying from weak to intense. The batholith isr osunr ded to the north, east and south by the Namaqua Metamorphic Province in which the r oacrkes pervasively overprinted by intense, predominantly east-west trending folia taionnd medium-grade metamorphism. The batholith type area can be referred to in tneic ttoerms as the Vioolsdrif Terrane of the Richtersveld Subprovince (after Joubeet rat l.,1986). It is delineated to the east by the Namaqua Front of Beukes (1973) who defined thisn bdoaury as a tectonic transition and described it as “well defined north of Goodhouste b beucoming vague eastward”. At 14 Ramansdrif 135 it is said to be tentative by Varn M derwe and Botha (1989). North of the Orange River it corresponds to the sillimaninit eis-ograd, as such representing a metamorphic transition. Blignault (1977) define sdi ma ilar line further towards the northwest in the Ais-ais area, marking the pos iwtiohnere the generally non-penetrative fabric of his “Richtersveld Domain” becomes regilo annad penetrative, and called it the Southern Front. This represents the western eixotne nosf Beukes’s Namaqua Front. Both these authors stated that these lines merely mhae rpko tsition where the intensity of foliation changes markedly, and do not represernrat nte boundaries. In Namibia, Blignault included the “grey gneisses” to the no orft hthe Southern Front in the Vioolsdrif Suite, correlating it with the rocks of the Goodsheo uSubsuite. Moen and Toogood (2007) omitted all indications of a sharp boundbaerytw een the unfoliated Vioolsdrif and foliated Namaqua rocks in the Goodhouse arean, sgt athtiat the transition is entirely gradational. The Lower Fish River thrust (Blignauelt al., 1983) and Groothoek thrust (Van der Merwe, 1986) were identified during the Bushman laPnrodject (1981-1986; see 1.3 Previous work). They represent the northern anudth seorn boundary of the Vioolsdrif Batholith respectively. To the west, the batho ilsit hoverlain by the Gariep Supergroup (780-520 Ma; Von Veh, 1993). To the east, the Gthroeok thrust combines with the Onseepkans thrust (Moen and Toogood, 2007), whsi ct he i extension of the Lower Fish River thrust (Namibia) in South Africa. As suche, tRichtersveld Subprovince (after Joubert 1986) is enclosed within these bounda rIine sit.s eastern parts, the Richtersveld Subprovince includes rock groups other than theo lVsidorif Batholith, however, this study is only concerned with the batholith and tmheaninly the type area. The only unit which is investigated beyond the type area, isH tohoeg oor Granite (Hoogoor Suite of Moen and Toogood, 2007). Contributing to solvihneg ct ontroversy regarding its inclusion or non-inclusion in the Vioolsdrif Sui twe,as one of the initial aims of this study. The Vioolsdrif Batholith (in the current tectonirca mf ework) covers an area of approximately 220 km in length and with an averwagidet h of 110 km. Evidence show 15 that most of the original batholith has been ero. d Beldignault (1977) proposed that the current level of erosion (in the Haib area) repnretss ethe root zone of the Orange River Group at a depth of 6-10 km. This interpretatiso nb aised on the recognition of features indicating a position in the transitional epizone-smozone of the crust (after Buddington, 1959). 2.2 DISTINCTION BETWEEN THE HAIB AND RICHTERSVELD AREAS In the batholith type area, the Orange River Griosu spe parated into two parts by an area of no outcrop (the Nein Nababeep plateau) withD teh eH oop Subgroup occurring to the west thereof, and the Haib Subgroup to the eahsits. dTistinction is a historical one and is primarily based on the fact that the Haib Subgriosu opv erwhelmingly volcanic in composition, while the De Hoop Subgroup is relalyti vreich in interlayered metasedimentary rock types, the Rosyntjieberg Ftoiormn arepresenting an unique, rather thick and mature (arenitic), metasedimentary unitiht iwn the subgroup. During early regional mapping programs, the current Orange R Givreorup was correlated with the “Kheis System” of Rogers (1911) in the Upingtone-Psrkia area (e.g., Rogers, 1916; De Villiers and Söhnge, 1959; Von Backström and Deli eVrisl , 1972). In this three-fold subdivision, most of the current Haib Subgroup mrebsles the “Wilgenhout Drift Series”, most of the De Hoop Subgroup the “Marydale Seraiensd” most of the Rosyntjieberg Formation the “Kaaien Series”, of the Kheis Syst e Bmli.gnault (1974) did not support this correlation over such a large distance anrdo dinutced the Orange River Group as a separate unit from the Kheis System. However,i sin s uhbdivision he retained the distinction between the Haib and De Hoop Subgro ups. The possible correlation of the Haib and De Hoobpg Srouups has been considered by a number of previous investigators (e.g., Minnitt7, 91;9 Ritter, 1980). Reid (1979) stated that this possible correlation cannot be ruledo onu pt etrological grounds. Minnaeatr al. (2011) considered the available evidence at the atimnd came to the conclusion that there is enough evidence to retain the subdivision. T hhiegyhlighted the most important 16 differences between the two areas, not only inO trhaen ge River Group but also in the Vioolsdrif Suite, as follow: a) The Pan African deformation is absent in theb H aarei a, while it is pervasively developed in the Richtersveld. This suggestst hthea bt oundary between the two areas (currently concealed by the Nama Group) may repnrte asne important tectonic feature. b) A large proportion of sedimentary rocks (greykwea, cconglomerate and quartzite) are interbedded in the volcanics in the Richterds vaerel a but are relatively scarce in the Haib area. c) The absence of a thick, relatively mature, qzuitaicr tunit (Rosyntjieberg Formation) overlying the volcanic pile in the Hairbe a. d) The Haib Subgroup is on average much more mthaafinc the De Hoop Subgroup. e) The Goodhouse Subsuite has an overall morec fnealstui re in the Richtersveld than in the Haib area. Both the Vuurdood Subsaunitde Khoromus Tonalite (the two most mafic units in the Vioolsdrif Suite) are absent hine Richtersveld. 2.3 ORANGE RIVER GROUP The stratigraphic sequence within the Orange RGivreoru p has been a matter of controversy but this comes as no surprise if thger edee of deformation and dismemberment of the unit, largely due to the isnitornu of the plutonic suite, is taken into account. The uncertain relationship between thieb Hanad De Hoop Subgroups, the deformed contact relationships within the De Hooupb gSroup, and the isolated occurrences of individual fragments are the maimn pcloicating factors in deciphering the stratigraphic sequence. 17 Within the Haib Subgroup, Blignault (1977) founds inag le location where he considered that the relationship between the Tsams (under)l yaingd Nous Formations could be established with certainty. This stratigraphicu seenqce within the Haib Subgroup is substantiated by the Rb-Sr radiometric age resouf lRtse id (1977), which show the Tsams Formation to be 2 020±70 Ma (dacites-rhyolites) tahned Nous Formation 1 970±70 Ma (basaltic andesites-andesites). Within the De Hoop Subgroup, Ritter (1980) baseed s tthratigraphic sequence on a model of deposition from a common volcanic centre sitdu ainte the northeast of the Richtersveld, and the recognition of proximal anisdta dl facies. He considered the Rosyntjieberg Formation to be interlayered betwtehen u nderlying rocks to the north thereof and the overlying Windvlakte Formation. is T chontradicts De Villiers and Söhnge (1959), who considered it to occur betwhee nu nt derlying Haib and overlying De Hoop Subgroups. Blignault (1977), again, thot uthgaht the Rosyntjieberg Formation overlies all the volcanic rocks, forming the top t hoef succession along a discordant contact. Minnaaer t al. (2011) is in agreement with Blignauoltp ( cit.) but consider the contact to be gradational (over a few meters). Radiometric data for the De Hoop Subgroup is viemryit el d with available results including only one sample from the volcanic Para Rdiyvser Formation and one from the quartzitic Rosyntjieberg Formation. Both these pslaems are from the current study and single zircon U-Pb ages were determined by LA-ICP (MthSe analyses are included in Appendix 2). The volcanic (dacite) sample rend earne dage of 1 883±7.4 Ma, which is notably younger than the ages of Reid (1977) feo rH thaib Subgroup. The quartzite sample rendered detrital zircon ages which suptphoer itn ferred stratigraphic position of the formation, overlying the volcanic pile. Theo bpar bility distribution curve of 42 analyses shows a major peak between 1 850-1 80w0i tMh as ingle grains showing ages of 1 760 Ma, 1 920 Ma and 2 050 Ma. This also consf iram Vioolsdrif Batholith provenance to the formation. 18 Although deformation has rendered primary textudrieffsic ult to recognize and even obliterated in some areas, it is possible to idfye nat iprotolith for the rocks in almost the whole of the group. Aphanitic lav asre very fine- to fine-grained and often the mainlse r cannot be discerned under the microscope. How eover,all the grain size of the volcanic rocks in the Orange River Group is sucaht the individual minerals can be discerned under the microscope (if not pervasivaeltleyr ed, which is often the case). The feldspar porphyroblasts of porphyritic lav caosmmonly exhibit a milky white or greenish colour owing to alteration. The quartz porphyrosbtsla are mostly rounded but may be subhedral. Some andesitic la vdaisplay quartz-containing vesicles, such as thino sthee Nous and Klipneus Formations. Figure 2.2: Foliated leucocratic tuff (Pan Afric daenformation) of the Paradys River Formation. Ltioocna: 16.9883°E; 28.2912°S. Tuffs are recognized by the presence of rock fragmemnatsll esr than 2 cm in the matrix (e.g., Figure 2.2). Rock fragments, glass shanrds p aumice fragments occur along with phenocrysts and are elongated and flattened pla troa ltlhee foliation. Agglomerate asre recognized by the presence of fragments larger 2th camn in the matrix (e.g., Figure 2.3). These fragments vary greatly in size, commonly beetnw 1-20 cm but they grade into 19 millimeter sizes at the one end and rare sizesp otof uhalf a meter have been encountered. They comprise mostly lava and chert and are inbvalyr ilaocally derived, correlating with the rocks of the same or at least associated, ftoiornmsa. Agglomerates are rather abundant throughout the Orange River Group andc eiaslplye in certain parts, e.g., in the Klipneus Formation in the Richtersveld National kP anrd the Windvlakte Formation in its northeastern parts. Figure 2.3: Foliated leucocratic agglomerate (PAafrnic-an deformation) from the Windvlakte Formati oena,st of Eksteenfontein (17.2817°E; 28.7383°S). Large freangtms comprise lava and chert. 2.3.1 Haib Subgroup The stratigraphy of the Haib Subgroup may be sumizmeda ras in Table 1. 20 Table 1: Subdivision and descriptions of the HSauibb group . Formation Description Nous Melanocratic volcanics (aphanitic lava, velsairc ulava, tuff, volcanic breccia). Tsams Leucocratic to mesocratic aphanitic and pyoryrptihv lava. 2.3.1.1 Tsams Formatio n In volume, the Tsams Formation is subordinate eto N thous Formation and predominantly leucocratic. Thyolite is the predominating comptiosni. Minor melanocratic lenses and discontinuous layers occur sporadically. Apha nlaitvica, quartz-porphyry and tuff predominates while volcanic breccia has been dbeesdc rfirom a number of localities. Rare interlayered quartzites have been reportemd tfhroe formation (Minnaaer t al., 2011). At one locality in the course of the Oerpn oReiver, where the track from Nous enters the river, leucocratic volcanics of the Ts aFmormation are overlain by a thin (approximately 1 m thick) layer of quartzite wh iicsh persistent over a distance of approximately 1 km. The contact relationship iasd gartional, with interlayered quartzite lenses making their appearance in the felsic laevaar tnhe contact and becoming increasingly abundant upward towards the contEacats. t ward, the Tsams Formation undergoes facies changes with metasedimentary rboecckosming more abundant untill they predominate over the volcanic rocks in them faotrion (Colliston and Schoch, 1996; Moen and Toogood, 2007). At a point approximately 2,5 km northeast of Gha,a tmo sthe west of the track leading from Ghaams to Kamgab, the leucocratic volcanicnsta cino spherical leucocratic nodules (Figure 2.4). These nodules comprise mainly qu, aKr-tfzeldspar and muscovite with accessory plagioclase. Biotite and ore are thke cdoanrstituents and are concentrated in the centre of the nodules. The nodules are ona agvee 2r x 3 x 7 cm in diameter. In the 21 same vicinity these nodules may also occur in othrme fof veins approximately 5 cm wide and 50 cm long. Figure 2.4: Leucocratic nodules in leucocraticc vaonlics of the Tsams Formation east of Ghaams. t iLoonc: a 17.933083°E, 28.978832°S. 2.3.1.2 Nous Formatio n The Nous Formation dominates the Haib Subgrou pis. o Ivt erwhelmingly melanocratic with interlayered leucocratic lavas which are coicnusopus in outcrop from a distance. Aphanitic lava and tuff are the most common rocpke tsy while vesicular lava and volcanic breccia also occur. Agglomerate and vnoiclc aconglomerate (lava containing pebbles of dacite) have been described and aott ereswt torking in the formation (Figure 2.5). Xenoliths of older volcanics are quite comnm ino the Nous lavas. Xenoliths of granitoids (tonalite, granodiorite and granite) ew aelrso found in melanocratic Nous lavas at one locality (Figure 2.6). No radiometric agaes hbeen obtained for these xenoliths but they present rare evidence of pre-existing crusict hw ihs older than the lava in which they are contained. They may also represent reworketedr miala of earlier phases of the Vioolsdrif Suite (not Archean crust). 22 Figure 2.5: Conglomerate from the Nous Formatiopnr icsing a matrix of melanocratic metalava (amplhitieb-obiotite schist) containing rounded fragments of dacitea (tliocn: 17.9822°E; 28.8344°S). Figure 2.6: Well rounded xenoliths of granitoidns m i elanocratic Nous lava at a hill south of Vioorilfs (d17.6859°E; 28.823°S). The occurrence of locally developed linear, crouststi-ncg veins (Figure 2.7) is a common phenomenon in the melanocratic aphanitic lavash eo fN tous Formation. These veins 23 vary in thickness but are usually not thicker th5a mn m. They can be densely concentrated and cut each other in all directio Tnhse. y could have been caused by contraction during cooling in a particularly homongeeous aphanitic lava. Another characteristic feature of the mafic lavas of theu sN oFormation (although not occurring frequently), is the occurrence of quartz vesictluebse, s and veinlets. Figure 2.7: Linear cross-cutting veins in melanaoticr lava of the Nous Formation east of the Noubg aRsiever. Location: 17.7476°E; -28.8043°S. A speckled appearance is often induced in the moecrlantic volcanics, especially the tuffs, by the clustering of dark minerals in aggartegs which are on average 5 mm in diameter. Where such aggregates are weathere tdh eo urto,ck is left with a pitted surface. Growth of secondary minerals also often leads tion carnease in the grain size to medium- grained. The Nous Formation is often highly foliated, thses l ecompetent lavas often being altered to schist. This foliation is near-vertical and wtheearing patterns such as flaggy, bladed, 24 columnar, rod-shaped or needle-shaped, are chraisratict eof the formation in highly deformed areas such as in the vicinity of Nousr.a lPleal layering, on average 5 mm thick, can be discerned in places and represents a pr ifmloawr ytexture. Interlayering and interfingering between lava and tuff can be obsde ravte certain localities (Figure 2.8). Figure 2.8: Interfingering relationship betweevna l a(fine-grained) and tuff in the Nous Formationa rn Gehaams, Nous area. Location: 18.000179°E; 28.954834°S. 2.3.2 De Hoop Subgroup The Richtersveld area (between the Neint Nababelaetepa Pu and the Gariep Supergroup) is pervasively overprinted by roughly north-sourtehn tding, low metamorphic-grade Pan African (720-550 Ma) foliation (see Figures 2.1 a2n.9d). 25 Figure 2.9: North-south trending Pan-African foioliant imprinted on rocks of the Vioolsdrif Batholiwthi thin the Devil’s Castle shear zone (Ritter 1980), east of Eksteetenifno.n Looking south from 17.3111°E; 28.7923°S.l i aFtioon is indicated with red trend lines. In the Richtersveld National Park, north of the yRnotjsieberg mountain range, the De Hoop Subgroup consists of two distinct success iio.en.s, ,in the northeast along the Orange River, and in the southwest underlying tohsey Rntjieberg Formation. These two entities are nowhere in contact with each other t haenrdefore age relationships are difficult to establish. Ritter (1980) considerehde mt to form the limbs of a regional anticline of which the core is occupied by grandisto oi f the Vioolsdrif Suite. According to Ritter (op cit.), the northeastern succession represents a coef nvtorelc anism which was active throughout its development, while the sorunt hseuccession can be seen as a recipient of sediments derived from this volcaneircr atne. Based on this interpretation, the Abiekwa River, Kook River and Kuams River Fotrimonas are interpreted as distal facies in which volcanic rocks predominate, whhile Pt aradys River and Klipneus Formations are interpreted as proximal facies inic hw ha relatively large proportion of reworked material occur. 26 Due to the fact that the Windvlakte Formation isw nhoere in contact with the other volcanic units of the De Hoop Subgroup, being saetpeadr from them by the Rosyntjieberg Formation, it is not possible to determine its tsigtraphic position within the subgroup. The stratigraphy of the De Hoop Subgroup may bem saurmized as in Table 2. Table 2: Subdivision and descriptions of the Deo pH oSubgroup . Formation Description Rosyntjieberg Massive and feldspathic quartziteh winiterlayered metapelite and black, iron-rich quartzite and overlying conglomter. a Discontinuous conglomerate bed at the base. Kuams River Melanocratic and mesocratic aphannitdic paorphyrytic lava with interlayered leucocratic lava and metasedimentasr (tzq,u conglomerate). Kook River Mesocratic and melanocratic quartz-fpealdrs porphyry. Klipneus Mesocratic and melanocratic tuff, lavad angglomerate with interlayered metasediments (conglomerate, qua,r tczhiteert) Abiekwa River Leucocratic, mesocratic and melantoicc rwaelded tuff and lava with interlayered metasediments in places. Paradys River Leucocratic, mesocratic and minora mnoeclratic tuff. Windvlakte Undifferentiated leucocratic, mesocr atnicd melanocratic volcanics; partly recrystallized with primary textures oftebnl ioterated. 2.3.2.1 Windvlakte Formatio n This formation is extensively intruded by the grtoaindis of the Vioolsdrif Suite which led to its dismemberment into isolated occurrencese. rTohcks are highly deformed and recrystallized and primary textures are difficuol tr et cognize or have been obliterated. 27 The Windvlakte Formation is a heterogeneous unmit pcroising undifferentiated leucocratic, mesocratic and melanocratic volcaoncick sr with interlayered metasedimentary rocks. There is a compositionaadl agtrion in the formation from southwest to northeast with leucocratic volcanircesd opminating in the areas to the south of Eksteenfontein, mesocratic volcanics predomninga itni the central parts of the outcrop area, and melanocratic volcanics predominatingh ein a treas bordering the Orange River. Porphyrytic lava, agglomerate and agglomeratic ptureffdominate in localities where original textures can be identified. Contact relationships among the volcanic rocksh ein f ot rmation are sharp in places and gradational in others. Lenses and discontinuoyuesr sla of the one occur in the other. Contacts between the volcanics and the Xaminxavipe rR Gi ranite are mostly gradational, locally over narrow (> 0.5 m) zones. Sharp const abcettween these two rock types have been observed but are rare and not convincingrluy sinivte (not cross-cutting). Contacts between the volcanics and the Gaarseep graniteoaidvse lno doubt that the latter are intrusive (sharp, cross-cutting contacts and abnutn xdeanoliths of volcanics in the granites). [The Xaminxaip River Granite is alstor uinded by Gaarseep granitoids.] The presence of metasedimentary rock types asesdo cwiaith the felsic volcanic unit in the area between Jenkinskop and Eksteenfontesinu,g igse sted by several occurrences of sillimanite in the rocks. In this highly shearendv ieronment the rocks are schistose and the occurrence of quartz-sericite-sillimanite stc phoisints to an originally sedimantary assemblage. It is however, also possible thast ctheis t represents an original acid volcanic. In areas immediately to the west ands ec lto the contact with the Gariep Supergroup, the schist commonly contains chlohriotern, blende and haematite, pointing to a volcanic origin. Sillimanite is especiallye pnltiful at Jenkinskop itself. Porphyritic felsic lava was found in a number oaf cpels in the Windvlakte Formation. Approximately 8 km south of Eksteenfontein, thea l aovccurs within a strongly foliated zone (Figure 2.10). The foliation trends N-S asn dre ilated to Pan-African deformation. The quartz porphyroblasts are milky white, elondg aptaerallel to the foliation and are on 28 average 5 mm in diameter (long axes). The folnia atinoastomises around the porphyroblasts. Figure 2.10: Foliated felsic porphyritic lava aopxpirmately 8 km south of Eksteenfontein. Locati1o7n.:2 5894°E, 28.901451°S. The geographically northeastern part of the foromna, tci losest to the Orange River, is especially rich in agglomerate (see figure 2.3h).i s Tagglomerate zone can be traced along strike up to an area just south of the Rojiseybnetrg Formation, where extensive development of the Xaminxaip River Granodioritem tienrates its extension along strike. To the north of the Rosyntjieberg mountain ranhge ,a tgglomerates of the Klipneus Formation are situated along strike of these, ssutgingge that the Klipneus Formation may represent the continuation of at least part ofW thined vlakte Formation, north of the Rosyntjieberg mountain range. 29 2.3.2.2 Paradys River Format ion The basement of the Paradys River Formation ha bs eneont identified. It is intruded by the Vioolsdrif Suite and overlain by the Rosyntejiergb Formation. The contact with the overlying Rosyntjieberg Formation is mostly obscdu brey deformation but can, in some places, be observed to be gradational (see 2.R3.o2s.7yn tjieberg Formation). Along strike to the northwest, the Rosyntjieberg Formation peinsc ohut and the Paradys River Formation is overlain by the Abiekwa River Formant,i othe contact being gradational. Figure 2.11: Intensely foliated leucocratic porrpyhtiyc dacite of the Paradys River Formation at ftohoet of the Rosyntjieberg mountain range, Richtersveld Nati oPnaarlk. Lenses of the overlying Rosyntjieberg Fotiromna quartzite occur in the dacite. Location: 17.0617°E; 28.3°4S7. 5 The largest part of the Paradys River Formatiomn pcroises leucocratic volcanic rocks, mostly tuff (see Figure 2.2), while lava and aggeloramte occur interlayered. Along strike to the southeast, the rocks become mesocratic wath tihlee foot of Rosyntjieberg, minor melanocratic volcanics occur in the formation. e nInste northwest-trending foliation is developed throughout the Paradys River Formatioignu (rFe 2.11). This is thought to be the result of sinistral shear associated with Pfarinc aAn deformation, during which the 30 overlying quartzite unit acted as a more compeutennit tr elative to the underlying volcanic unit. A dacite sample which was dated during this sturodmy f the Paradys River Formation rendered an U-Pb zircon age (LA-ICPMS) of 1 883 ±M7.a4 (Appendix 2). 2.3.2.3 Abiekwa River Formatio n The Abiekwa River Formation was the main locusn otrfu ision by the Vioolsdrif granitoids in the Richtersveld National Park. Nonolty was most of the formation engulfed by the granites, but the intrusion alsuos cead thinning and uplifting of the supracrustal succession on both sides of the invetr ubsody, to compensate for the lack of space. In the northern parts of the Richtersveld NatioPnarlk , the Abiekwa River Formation is overlain by both the Kook River Formation and Kua Rmivser Formation, the boundary represented by a thrust with the Abiekwa River Faotriomn overriding the latter two formations. In the eastern parts of the park,A tbhie kwa River Formation grades upwards into the Kuams River Formation. The Abiekwa River Formation comprises approxima eteqluyal volumes of mesocratic and leucocratic volcanics, mostly represented bffy a tnud interlayered lava. Agglomerates are also fairly common and compris leo coaflly derived material (Figure 2.12). 31 Figure 2.12: Agglomeratic leucocratic porphyrylaticv a of the Abiekwa River Formation, Richtersveladt iNonal Park. Location: 17.1081°E; 28.2239°S. 2.3.2.4 Klipneus Formatio n The Klipneus Formation occurs along strike with Pthaeradys River Formation but is separated from it on lithological grounds, a proemnitn shear zone between the two being interpreted as an important tectonic boundary K(twhea ggarug shear, Ritter 1980). It is nowhere in contact with the other volcanic unitst ibt uis thought to be associated with the Kook and Kuams River Formations because of its opmreindantly melanocratic character. It contains interlayered meso- and leucocratic sro acnkd is the most heterogeneous unit of the Orange River Group comprising mainly tuff, l a(ivnacluding vesicular lava) and agglomerate but also sedimentary rock types su cho nagslomerate and chert. Well developed agglomerate can be observed along thde i nr othae eastern part of the Richtersveld National Park north of the Oudinnis Rieivper course. Here the contact between the steeply foliated Orange River Group t haen dweakly deformed, near- horizontal Karoo Supergroup, can also be obserFviegdu r(e 2.13). 32 Figure 2.13: Contact (largely concealed by ta lbluest)ween foliated (Pan-African foliation) schis tt hoef Klipneus Formation and undeformed strata of the Karoo Surpouerpg, Richtersveld National Park. Location: 17.03°5E9; 28.3467°S. Figure 2.14: Contact between mesocratic (bottonmd) m aelanocratic quartz-feldspar porphyry of the kK oRoiver Formation in the Richtersveld National Park. Thone tcrats is not well defined in the photo but follso wthe horizontal joint. Location: 17.2599°E; 28.2420°S. 33 2.3.2.5 Kook River Formatio n The upper contact of the Kook River Formation wthiteh overlying Kuams River Formation is gradational. The Kook River Forma tcion sists entirely of quartz-feldspar porphyry, the lower part being mesocratic and tphpee ur part melanocratic. The contrast between these two compositional varieties is cocnusopuis and the contact is sharp (Figure 2.14). The contrast can be observed from the lreoaaddin g down the course of the Kook River immediately to the west of De Hoop in the hRteicrsveld National Park. In hand specimen, the phenocrysts are represented by blaogtiho cplase and quartz. 2.3.2.6 Kuams River Formati on The Kuams River Formation consists predominant lpyo orfphyritic lava, but contains intercalated tuff layers. Mafic enclaves are sdpiocrally contained in the lava (Figure 2.15). Figure 2.15: Mafic enclave in porphyrytic melanaotcicr lava of the Kuams River Formation north of HDoeo p, Richtersveld National Park. Location: 17.1848°8E.;1 2631°S. 34 Two types of enclaves were recognized, viz. xehnso loitf other lava, and restites composed of dark minerals. The lavas are predonmtlyin ma elanocratic, with subordinate interlayered leucocratic units. Epiclastic metaimsednts are present in places in the form of quartzite, cross-bedded biotite-muscovite gn aenisds conglomerate. To the northeast, the unit is intruded by granitoids of the Vioolsf dSriuite. Xenoliths (of various sizes) of the Kuams River Formation occur within the pluto unnicits of the Vioolsdrif Suite. 2.3.2.7 Rosyntjieberg Formati on The Rosyntjieberg Formation is a quartzite-domidn autneit that gives rise to the Rosyntjieberg mountain range. The formation grallyd uthains out along strike in a northwesterly direction. Duplication in the formioant due to isoclinal folding and thrusting is apparent where the formation is a tth iitcskest. The contact relationship with the underlying volica pnile is, for the most part, sharp, giving the impression that the metasediments oev ethrlei volcanics unconformably. However, this sharp contact relationship resuoltsm f rPan African deformation, which led to displacement between the volcanic and quar tuznititcs. In places, the original contact relationship is preserved and can be observed gtora bdeational with lenses and discontinuous layers of quartzite becoming moren tpifluel in the underlying volcanics until the quartzite predominate, which in turn caoinst lenses of the underlying volcanics (Figure 2.16). These lenses of volcanic rocks teuvaellny disappear completely higher up in the succession. In Paradyskloof in the Richtersveld National Painrktr,u sive contact relationships can be observed between the Quartzites of the Rosyntjgie Fboermation and granites of the Gaarseep Granodiorite unit (Figure 2.17). 35 Figure 2.16: Interlayered quartzite in intenseollyia fted, leucocratic porphyrytic lava of the Parsa dRyiver Formation near the base of the Rosyntjieberg Formation (ttrhaeti gsraphic bottom-to-top in the photo is fromh rti gto left, location: 17.0618°E; 28.3471°S, Rosyntjiewater, RichtersvNealdti onal Park). Figure 2.17: Xenolith of Rosyntjieberg quartzinte g iranite of the Goodhouse Subsuite (Gaarseep Gdiroarnitoe) in Paradyskloof, Richtersveld National Park (17.024; 72°8E.3141°S). Length of measuring-stick in the tpoh iso 50 cm. 36 The bulk of the formation comprises quartzite, vinagry in composition between brown feldspathic and white, fine- to very fine-grainerde naitic quartzite, with the latter predominating. A variety of sedimentary structu areres preserved in the unit including cross- and parallel bedding, ball-and-pillow stururecst (Figure 2.18), and ripple marks. Figure 2.18: Parallel- and cross-bedding as wse bll aall-and-pillow structures in the Rosyntjieberogr mFation, Richtersveld National Park (17.0052°E; 28.3284 °HSe).a d of hammer on cross-bedding. At the base of the formation, a conglomeratic la isy esrporadically developed. A persistent argillaceous unit approximately 50 mc kt hoiccurs interlayered in the quartzitic sequence and can be traced along the entire sotfr itkhe formation (Figure 2.19). Seen from the north, the unit forms a conspicuous daorlko-ucred band within the white quartzites. On close inspection it is seen to ebrey vheterogeneous, being dominated by black, iron-rich quartzite and further comprisinge tmapelite, mudstone, siltstone and arkose. At the summit of Oemsberg and the surrionugn adrea, a quartzitic conglomerate constitutes the top of the Rosyntjieberg Forma t iTonh.e matrix as well as the pebbles are quartzitic. Both the bottom and top conglomeraetdes b comprise only locally derived pebbles. 37 Figure 2.19: Dark-weathering argillaceous unitr n tehae base of the Rosyntjieberg Formation, Richvteeldrs National Park. Location: 17.0647°E; 28.3936°S. Detrital zircon ages (U-Pb by LA-ICPMS) from onem spale of quartzite from the Rosyntjieberg Formation rendered a probabilityr dibisution curve with a major peak between 1 850-1 800 Ma and single grains showinegs aogf 1 760 Ma, 1 920 Ma and 2 050 Ma (n=42; see Appendix 2) 2.4 VIOOLSDRIF SUITE Contrary to the Orange River Group, the VioolsdSruifi te shows little deformation in the batholith type area. Clear contact relationshmipso nag the various units of the Vioolsdrif Suite, as well as between the suite and the vocl cgarnoiup, are common within the type area, with the exception of the Vuurdood Subsuite. Compositional references in the names of the u(Knihtso romus “Tonalite”, Gaarseep “Granodiorite”, Xaminxaip River “Granodiorite” anHdo ogoor “Granite”), are based on that unit’s average composition as indicated on c thheemical classification diagram of De le Rochee t al. (1980; see Figure 4.2). 38 The stratigraphy of the Vioolsdrif Suite may be smuamrized as in Table 3. Table 3: Subdivision and description of the Viodorilfs Suite. Subsuite Unit Description Sout Granite Medium- to coarse-grained leucocratic Ramansdrif granite. Ghaams Granite Fine- to medium-grained leucocgraraticn ite. Hoogoor Granite Light-brown weathering floiated ngirtoaids (granite, granodiorite, tonalite). Gaarseep Granodiorite Homogeneous brown-weath ecroinagrs, e- grained granitic and mafic intrusives (granodiorite, granite, tonalite, diorite, gabbro). Blockwerf Migmatite Heterogeneous migmatitic maafincd granitic Goodhouse plutonic rocks (diorite, tonalite, granodiorite). Xaminxaip River Orange-brown weathering, even-grained Granodiorite (medium-grained) granodiorite. Khoromus Tonalite Dark-brown to black weatherinrge,y g, medium-grained granitic and mafic intrusives (tonalite, granodiorite, diorite, gabbro). Vuurdood Mafic-ultramafic bodies of limited exte n t. Composites of gabbro, pyroxenite, peridotite, minor troctolite. 2.4.1 Vuurdood Subsuite The Vuurdood Subsuite is mainly developed in theib Haarea to the south of Vioolsdrif and is conspicuously absent in both the Richtedrs avreela and within the Hoogoor 39 Granite. Mafic-ultramafic bodies which are pres ienn tthe Richtersveld area to the east and southeast of Eksteenfontein (De Villiers anhdn Sgöe, 1959; Middlemost, 1965; Minnaar et al., 2011) is not included in the Vuurdood Subs bueitecause field evidence suggest that they were emplaced during post-Vioriof ltsimd es (reliable radiometric dating is needed to confirm this). But even if they dorr ecloate with the Vuurdood Subsuite, then they are still volumetrically vastly inferitoor those in the Haib area. In the Hoogoor Granite, sporadic bodies of amphibolite are enceoruendt which may represent metamorphosed bodies of the Vuurdood Subsuite gbauint ,a even if that is the case, they cannot volumetrically be compared to those in thaeib H area. The Vuurdood Subsuite comprises composite bodoienst,a cining various proportions of gabbro, amphibolite, peridotite and pyroxenite.i d R (e1977) distinguished troctolite in the Swartkop body. The bodies vary in size andp es,h raanging from about 100 m to about 2 km across and are invariably enclosed nw itthhei Goodhouse Subsuite. Contact relationships between the bodies and the rockhse o Gf toodhouse Subsuite are either obscured due to shearing, or covered by talus fTanhsis. has previously led to some controversy as to the nature of these bodies. S reogmaerded them as earlier phases of the Vioolsdrif Suite (e.g., Blignault, 1977), some ausm culate material within the Goodhouse Subsuite (e.g., Reid, 1982), while others regartdhemd as intrusive bodies in the Goodhouse Subsuite (e.g., Beukes, 1973). Radioicm aegter determinations and isotopic studies by Reid (1977, 1982) have established bde ryeoansonable doubt that they are among the oldest in the batholith (overlappingg ine awith the early volcanics) and that they originate from similar parent magmas as bhoeth O trange River Group and Goodhouse Subsuite. During the current study, lxitehnso of gabbro, closely resembling those of the Vuurdood Subsuite, were found in Gooudshe Subsuite granites (Figure 2.20). 40 Figure 2.20: Xenoliths of gabbro, probably frome tVhuurdood Subsuite, in granitoids of the KhoromTuosn alite at a locality southeast of Vioolsdrif (17.9411°E; 28.804°S2). Figure 2.21: Gabbro of the Vuurdood Subsuite laotc al ity southeast of Vioolsdrif (17.9431°E; 28.88°S7). The bodies of the Vuurdood Subsuite are melanocc arantdi the gabbros often show a speckled surface appearance with white (averagme )5 f meldspar porphyroblasts in the melanocratic matrix (Figure 2.21). Original rocekx ttures are invariably destroyed by the 41 growth of large secondary minerals, leading to as smivae texture and the production of boulder-sized rubble. Zoning was recognized du trhineg current mapping in some of the bodies where the ultramafic rocks (pyroxenite aenrdid potite) form a marginal zone around a central core of gabbro. The contact beentw riem and core is gradational. Ultramafic compositions also occur within the ceanl gtrabbroic part, probably as cumulates. 2.4.2 Goodhouse Subsuite The Goodhouse Subsuite is volumetrically dominna ntht ei Vioolsdrif Suite. The Gaarseep Granodiorite unit can be considered tiofy ttyhpe suite and its type area is along the national Namibian-South African highway souft hV iooolsdrif (see map). The subsuite as a whole comprises a range of compnos iftriom gabbro, diorite, tonalite and granodiorite to monzo- and syenogranite. Howefvoeurr, units can be distinguished within the subsuite within which mineralogical agnedo chemical variation is limited. The exact proportions of tonalite, granodiorite agnradnite within the subsuite is difficult to determine as they resemble each other clos ethlye i nfield in physical character. Diorite and gabbro show a gradation, making thefmfic duilt to distinguish from each other but can, however, be distinguished from trhaen igtoids fairly easily by an overall darker weathering colour. They represent about o1f0 t%he total volume within the Goodhouse Subsuite. In the Haib area, contacts between the Goodhoubsseu Siteu and Orange River Group are sharp and cross-cutting with numerous xenolithtsh eo fl atter contained in the former (Figure 2.22). In the Richtersveld area, the ccotnst abetween these two units are typically concordant, with xenoliths of the volcanics stbillu andant in the intrusives. 42 Figure 2.22: Typical contact relationship betwetheen volcanics of the Orange River Group and graidnsit of the Vioolsdrif Suite in the Haib area as seen at al iltoyc saoutheast of Vioolsdrif (17.9617°E; 28.8850.° S) Figure 2.23: Heterogeneous enclave swarm in thoer oKmhus Tonalite at a locality in the Oernoep Ricvoeur rse, southeast of Vioolsdrif (17.9457°E; 28.9339°S).e yT hconsit of xenoliths of the Orange River Groupd aonlder phases of the Vioolsdrif Suite as well as dark mineral cuulamtes. 43 A characteristic feature of the Goodhouse Subsisu itthee widespread occurrence of dark mafic enclaves in the granitoids. These enclavr e so fa various types and include xenoliths of the Orange River Group and restite csu omrulates of dark minerals which formed within the unconsolidated magmas. Ther ea lasroe enclaves that were deformed, occurring as flattened and elongated dark mineurmalu clates, the origin of which is uncertain. In places, enclave swarms, some homeoguesn and some heterogeneous (e.g., Figure 2.23), occur. The xenoliths of Orange R iGverorup volcanics are the most abundant. They vary in shape from angular to roeudn adnd in size from a few millimeters to tens of meters or even hundreds eotfe mrs in diameter. A detailed investigation of the enclaves was not done for sthtuisdy. 2.4.2.1 Khoromus Tonali te The Khoromus Tonalite is the oldest unit within Gthoeodhouse Subsuite, as can be deduced from contact relationships, and is extenlys idveveloped only in the Haib area. In the field, the Khoromus Tonalite can be distiinsghued from the Gaarseep Granodiorite from a distance by an overall darker weatheringo ucro tlhat can be recognised even on aerial photographs. A characteristic feature loafn ad scape built by the Khoromus Tonalite, is the presence of conspicuous hills ecadp bpy black-weathering bodies with dark debris along the slopes (Figure 2.24). This is in contrast with the areas underlain by rGseaeap Granodiorite, which display homogeneous brown colours (Figure 2.35). Contaecttws eben the Khoromus Tonalite and Gaarseep Granodiorite are sharp and cross-cuttiitnhg x wenoliths of the former occurring in the latter (Figure 2.25). 44 Figure 2.24: Typical view over an area underlayin th be Khoromus Tonalite unit. Locality near theu Ngoaseb River course (17.9619°E; 28.8649°S). Direction of viesw to i the east. Figure 2.25: Xenoliths of Khoromus Tonalite (da irnk )Gaarseep Granodiorite (17.9253°E; 28.9191°S). 45 Figure 2.26: Typical Khoromus Tonalite granitonid o iutcrop. Course of the Oernoep River, southoefa Vsito olsdrif (17.9417°E; 28.9286°S). The well-rounded xeno lpitrhosbably represent mafic volcanic rocks of then Ogrea River Group. In outcrop, granitoids of the Khoromus Tonalite tayrpeically medium-grained with a speckled grey colour (Figure 2.26). The speckrle sd aue to aggregates (average 5 mm) of dark minerals (biotite and hornblende). Sucghr eagates may be up to 2 cm across and are sometimes weathered out, leaving a pitted cseu.r f aFeldspar phenocrysts, on average 5 mm in diameter, are often discernible in outc rop. The black-weathering rocks which are so conspic ufrooums a distance (see Figure 2.24), occur as numerous isolated patches of a few mteot etersn ths of meters across. The black weathering is due to a coat of desert varnish eo ns uthrface of the rocks. They typically produce a metallic clang when hit with a hammehr.e Tfresh rock is blue-grey and medium-grained, while grey quartz and green plalagsioec crystals can be discerned. Dark minerals (biotite and hornblende) sometimes alsuos ec a speckled appearance to this melanocratic rock. Contacts between these blacakt-hweering rocks and the brown- weathering varieties are gradational over a fewe mrse (tFigure 2.27). Leucocratic varieties of the Khoromus Tonalite rocks also o cbcuutr are scarce and limited in extent. The unit is predominant lmy edium-grained but coarse-grained variations o c cTuhre. 46 presence of mafic enclaves is a characteristicu rfe aotf both the Khoromus Tonalite and the Gaarseep Granodiorite. Figure 2.27: Gradational contact relationship beetnw melanocratic and mesocratic varieties withein K thoromus Tonalite at a locality north of Ghaams. Locati1o8n.:0 008°E; 28.8936°S. The rocks of the Khoromus Tonalite unit commonlsyp dlai y textures in outcrop which may be interpreted as products of magma mixinge r( asfttudies like, e.g., Perugienti al. 2002; Perugini and Poli 2004). These include bnagn daind “streaky” textures (Figure 2.28), which conform with descriptions of “filame” natnd “globular” regions in rocks which have been shown to represent products of ma amgimxing. 47 Figure 2.28: Banding and “streaky” textures comlmy odnisplayed in the Khoromus Tonalite unit which ym bae interpreted as “filament” (leucocratic areas) angldo b“ ular” (melanocratic areas) areas as definePde irnu ginie t al. (2002) and interpreted as the result of chaotici nmgi xdynamics during magma mixing processes. Soausth oef Vioolsdrif (17.9253°E; 28.9191°S). 2.4.2.2 Xaminxaip River Granodior ite The Xaminxaip River Granodiorite is best developine tdh e area to the east of Eksteenfontein and is named after a dry river bne tdh ai t area (see map). It was first mapped as an independent unit by Ritter (1980), rwehfeorred to it as “even-grained granodiorite”. Minnaaer t al. (2011) named it the Xaminxaip River Granodio r ite. Unfortunately this name is uncommon in the Engllaisnhg uage but was chosen for reference purposes on published maps. Being art edde saerea, such references are few and far between on the relevant maps. The namme i“nXxaaip” should be pronounced “kaminkaip” (refer to the international phoneticp haal bet) as it originates from a Nama word in which a clicking sound (with the tongue )r eisplaced by “x” (after De Villiers and Söhnge, 1959). As there is no such clicking soinu ntdhe English alphabet, “k” is the closest to it. 48 U-Pb zircon (LA-ICPMS) ages were obtained for a pslaem from the unit (see Appendix 2). A crystallization age of 1 892.5±4.8 Ma is yv esrimilar to that of the Gaarseep granitoids as determined by Reid (1977) in the H aareiba. It is chronologically placed older than the Gaarseep Granodiorite since conretalactti onships attest to it being intruded by the Gaarseep Granodiorite (Figure 2.29). Ine s ocamses, the relationship is unclear with the contact being gradational, however, thter ipnretation is supported by previous studies of De Villiers and Söhnge (1959) and R i(t1te9r80), who came to the same conclusion. Figure 2.29: Xenolith of Xaminxaip River Granoditieo rin granite of the Gaarseep Granodiorite (176.4°E40; 28.8035°S). The Xaminxaip River Granodiorite is spatially asisaotecd with the volcanic rocks of the Windvlakte Formation. Contact relationships betnw tehee two units are mostly gradational and it is often possible to identifylc vaonic textures, such as tuffaceous fragments, in the granitoids (Figure 2.30). R i(tt1e9r80) considered the Xaminxaip River Granodiorite to represent a subvolcanic unit bentw tehe Vioolsdrif Suite and the overlying volcanics. 49 Figure 2.30: One of the textures often found ien Xthaminxaip River Granodiorite unit which is intererpted to represent an original volcanic character (mesocratic tuftfh inis case) from the Windvlakte Formation after yresctarllization (location: 17.3203°E; 28.7466°S, east of Eksteetnefionn). The Xaminxaip River Granodiorite as a whole is v heorymogeneous in composition and grain size (medium-grained, average 1 mm) altho turgahn,sitions to fine- and coarse- grained varieties do occur. It displays a charaiscttice grey colour and weathers light- brown. Xenoliths of Orange River Group volcanicss w aell as other types of enclaves occur in places, although much less frequent tnh athne i Khoromus Tonalite and Gaarseep Granodiorite. The development of metamorphic banding, plastico rdmeaf tion and, locally, fully developed migmatite (Figure 2.31), may be consdid echrearacteristic of the Xaminxaip River Granodiorite. De Villiers and Söhnge (195co9n) sidered the migmatite to be a direct result of the intrusion of the Gaarseep Godraionrite into the Xaminxaip River Granodiorite (thus, by contact metamorphism). e Rr i(t1t 980) regarded the migmatite in Helskloof to the east of Eksteenfontein, as a pcrot douf metamorphism at a deeper crustal level. Contacts between the migmatite and the hgoemneoous granodiorite within the unit 50 are entirely gradational. Areas of migmatite doevpemlent are commonly a few metres in extent. Figure 2.31: Migmatite developed in the XaminxRaiipv er Granodiorite at Helskloof, east of Eksteentefoinn (17.4398°E; 28.8037°S). 2.4.2.3 Blockwerf Migmatit e The Blockwerf Migmatite was originally termed thelo Bckwerf Migmatite Complex by Ritter (1980). It is situated on the bank of threa nOge Riverin the northeastern part of the Richtersveld National Park in South Africa. Itr iundtes the Kuams River Formation and is itself intruded by the Gaarseep Granodioriteh.e sTe relationships are indicated by cross-cutting contacts between the migmatites haen dG taarseep Granodiorite (Figure 2.32), xenoliths of the Kuams River Formation vonlicas found within the granitoids of the Blockwerf Migmatite, and xenoliths of the Blowcekrf Migmatite in the granitoids of the Gaarseep Granodiorite. 51 Figure 2.32: Contact between granodiorite of thaea rGseep Granodiorite unit (top) and migmatite oef B thlockwerf Migmatite at Blockwerf, Richtersveld National Pa rLko. cation: 17.1221°E; 28.0798°S. Figure 2.33: Migmatite of the Blockwerf Migmati tteh,e migmatitic character is owed to a large num abnedr high concentration of mafic enclaves in the intrudinagn gitroids. Location: 17.1525°E; 28.1074°S. The migmatitic character of the Blockwerf Migma tiiste a direct result of the large number and high concentration of mafic enclavetsh ein i ntruding granitoids. In other words the mafic enclaves (which vary in size aned d aernsely and randomly distributed) 52 cause a deformation of the granitoid’s matrix inch s ua way that the impression of plastic deformation is created, which is being termed mitgitme a(Figure 2.33). The majority of the enclaves closely resemble the volcanic roc kths eo fKuams River Formation, while others are represented by cumulates of mafic mlisn.e ra Two phases of migmatite can be distinguished w ihthei nBlockwerf Migmatite, based on how well the migmatitic character is developed (cwhh iis directly related to the density of enclaves in the granitic matrix). The first phaiss ree presented by those parts where the density of mafic enclaves is high and the migmca ctihtiaracter is well developed. In this phase, the dark mineral contents is also high haen dg rtanitoids are melanocratic. The second phase is represented by those parts in wthheic dhensity of enclaves are low and the granotoids are not migmatitic. This phasea h laosw er dark mineral contents and is mesocratic. This relative difference in composni tbioetween the two phases can also be observed on a TAS classification diagram (Figu1re9 ;4 4. .2.7 Blockwerf Migmatite). The two phases may be referred to as phase 1 and 2p h(as ein Figure 4.19). There is a complete gradation between the two phases. There is a large variation in the sizes of the aevnecsl from a few centimetres to bodies of about 3 km in length and 1 km wide. Patches onf igtera occurring within such large bodies, are melanocratic. For the most part, tahter icmes of the Blockwerf Migmatite are coarse-grained (average 4 mm) and the dark min cearaulse a speckled appearance in the fresh rock. In phase 1, the matrix is often finraei-nged. Gradational contact relationships between the enclaves and the granitoid matrix ains ep h1, suggest that the granitoid may be a hybrid rock (i.e., formed by the complete masilasition of the country rock which it intruded; Figure 2.34). The intruding granites of the Gaarseep Granodi oariete characterised by homogeneity in all of composition, grain size and weathering cor.l o Iun these characteristics they closely resemble phase 2 of the Blockwerf Migmatite whehree e tnclave densities are low. The Gaarseep granitoids have, compared to the grasn iotof idthe Blockwerf Migmatite, much fewer enclaves and their concentration is very ilno wth e unit. They are evenly 53 distributed, well-rounded and their sizes are e veanry,ing only between 2 – 5 cm. In addition to the enclave types found in the Blockfw Meirgmatite, the Gaarseep granitoids also contain xenoliths of the Blockwerf Migmatitsee ilf. The composition of the Gaarseep granitoids do not seem to be influenc ethde b eyngulfment of the migmatites, nor by the engulfment of the Kuams River volcan(ficrosm observations elsewhere). Figure 2.34: Hybrid rock of the Blockwerf Migmaet.i t Location: 17.1373°E; -28.1002°S. Blockwerfc, hReirsveld National Park. 2.4.2.4 Gaarseep Granodio rite The Gaarseep Granodiorite has by far the most seixvte ndistribution in the Vioolsdrif Suite and is very homogeneous. It is coarse-gdra ained invariably exhibits a brown weathering colour and a woolsack weathering pa t(tFeirgnure 2.35). The type area for the unit has historically been considered to be thea asoreuth of Vioolsdrif (see map). The occurrence of mafic enclaves is a characterisatictu fre , as for the Khoromus unit. These enclaves are relatively homogeneous in size anpde s, hbaeing mostly very well rounded and on average 5 cm in diameter. In the fringea sa oref the batholith, where the Gaarseep Granodiorite is overprinted by Namaqua foliatioanr,g le (up to 3 cm) idiomorphic 54 secondary hornblende crystals are commonly devde loinp tehe granotoids (see Figure 3.9). Figure 2.35: Typical view over an area underlayin th be Gaarseep Granodiorite. Dark-weathering oupt corf the Khoromus Tonalite are visible in the backgroundo.c aLlity near the Oudanisiep River course (17.961; 92°8E.8649°S). Direction of view is to the south. Based on field evidence and contact relationshiti pcsa,n be said that the timespan of development of the Gaarseep Granodiorite unit oavpesr tlhat of the entire Vioolsdrif Batholith,from the earliest phases overlapping wthiteh Orange River Group, through to the final stages, even as young as Ramansdrif Siteu.b sXuenoliths of it are found in some of the earliest volcanics (see 2.3.1.2 Nous Foromna),t iand it can be observed to intrude all of the Orange River Group, Khoromus Tonalitea,m Xinxaip River Granodiorite, Blockwerf Migmatite and Rosyntjieberg Formatione, tlhatter of which render some of the youngest radiometric ages in the batholith 2(s.3e.e2.7 Rosyntjieberg Formation). Contact relationships with the Hoogoor Granite dco nuolt be established in the field. The Gaarseep Granodiorite is on average more felstihce i nR ichtersveld than in the Haib area. In the southern parts of the Richtersveld NatioPnaarlk , the granites commonly display green and pink alteration of the feldspars in ooupt.c r 55 A sample from the Gaarseep Granodiorite in the tRericshveld, which is found to intrude the Rosyntjieberg Formation, rendered U-Pb zircLoAn- I(CPMS) ages of 1896±12 Ma (see Appendix 2). This age is problematic becaitu csoen tradicts the field evidence. According to contact relationships it should be nygoeur than all of the Paradys River Formation (1 883±7 Ma), Rosyntjieberg Formation 8(<010-1 850 Ma), and Xaminxaip River Granodiorite (1 892±5 Ma). Given the relaetlyiv wide time span of development of the Gaarseep Granodiorite, it is possible thatw trhoen g outcrop was sampled for the age determination. According to field mapping, the spalem is from the same pluton that intrudes the Rosyntjieberg Formation (see map lidne fro). However, the sample is not from the actual locality where the two units aruen fdo in contact (Paradyskloof, Richtersveld National Park). 2.4.2.5 Hoogoor Grani te The Hoogoor Granite is located within the areasro suunrding the batholith type area where it is pervasively overprinted by foliationd a mnetamorphism of the 1.3-1.0 Ga Namaqua orogeny (Figure 2.36), which has led t oo bthlieteration of original textures and contact relationships. This has been the cause of controversies regatrhdein cgo rrelation of the unit with the Vioolsdrif Suite (Gaarseep Granodiorite). Blignta eutl al. (1983) extended the Vioolsdrif Suite through field mapping out of its relativelnyd ueformed type area eastward up to the vicinity of Onseepkans, thus correlating what rism ted the Hoogoor Granite in this study, with the Gaarseep Granodiorite. Followihnigs ,t Strydome t al. (1987) included the unit in the Goodhouse Subsuite. However,c tohrisre lation was not followed by Moen and Toogood (2007) who, in agreement with Von Btaröckms (1953), distinguished the unit separately as the Hoogoor Suite. They bahsiesd s ut bdivision on the following field evidence: 56 Figure 2.36: Intense foliation associated with Nthaemaqua (1.3-1.0 Ga) orogeny, which is attaine pdl ainces within the Hoogoor Granite unit. Location: 18.5039°E; 28.9°2S3.6 a) the overall more felsic nature of the Hoogooar nGitre compared to the Gaarseep Granodiorite in the Haib area, b) the general absence of the Vuurdood Subsuiteic -mulatrfamafic bodies, which is typically associated with the Goodhouse Subsu itiets itnype area, c) the common association of the Hoogoor Graniteh wmietasedimentary rocks, as opposed to the Goodhouse Subsuite's general atsiosno cwiaith metavolcanic rocks, d) the virtual absence of mafic enclaves in the gHoooor Granite, which is considered characteristic of the Goodhouse Subsuite. Moen and Toogood (2007) furthermore report fieldid envce which suggest that parts of the Hoogoor Granite actually belong with the medtaimsentary supracrustal units. Such evidence include gradational contacts with metatepse laind quartzites, horizontal facies changes and the occurrence of abundant sillimainn itthee gneiss in places. Colliston and Schoch (2006), on the other hand, demonstrate dm tahnayt of the textural differences between the Gaarseep Granodiorite and Hoogoor tGe,r amnai y be accounted for by differences in strain intensity between the OraRngivee r orogeny and Namaqua orogeny, 57 thereby maintaining the correlation between the utwniots. This correlation is supported by a U-Pb zircon (SHRIMP) age of 1 890 Ma, whic hre ipsorted by Moen and Toogood (2007; pers. comm. R. Armstrong). This age agwreeells w ith the intermediate members of the Gaarseep Granodiorite unit, as well as wthieth X aminxaip River Granodiorite. The fact that the Gaarseep Granodiorite and Hoo Ggoraonr ite are time equivalents does not necessarily imply that they are geneticallya tredl . This study investigates the geochemical relationship between them. Some o gf ethoechemical characteristics, especially those involving trace and rare earthm elnets, are regarded to reflect primary igneous processes. Moen and Toogood (2007) suegdg feusrther investigations into the age relationship between the Hoogoor Granite aen dm tehtasedimentary units which are typically associated with them, especially the Gouma dFormation of Strydome t al. (1987). Indeed, in the eastern extension of thite a unnd in Namibia, it becomes part of much more complicated correlation issues involvain ngu mber of supracrustal units. Such investigations are not within the scope o fc tuhrerent study. In outcrop, the Hoogoor Granite displays light-bnro weathering colours. The Namaqua foliation is pervasively developed, varying in insteity, and the rocks display a granoblastic texture with a relatively wide varoiant iin grain size. Moen and Toogood (2007) recognized an association between coarsinee-gdr a(>4 mm) varieties and the development of an augen texture. The augen macyo mbep osed of alkali feldspar or intergrowths of quartz and feldspar. 2.4.3 Ramansdrif Subsuite The Ramansdrif Subsuite includes the leucocratdicm eenmbers of the suite. It is found intrusive into the Gaarseep Granodiorite (Figur3e7 )2 .and represents the final stage of the batholith evolution (Reid 1977). The granites of the Ramansdrif Subsuite weathleigr htot- brown colours, similar to other units in the Namaqua Metamorphic Province which c aormemonly referred to as "pink 58 gneisses". Two distinct grain size variations bcea nd istinguished in outcrop and Minnaar et al. (2011) proposed the subdivision of the uniot itnhte Ghaams (fine- to medium- grained) and Sout (medium- to coarse-grained) Gtersa,n ni amed after wells in the study area (see map). Contacts with the rocks of thed Gh ouose Subsuite are sharp and discordant. Figure 2.37: Intrusive relationship between theu tS Goranite and Gaarseep Granodiorite at a locsaoliutyt heast of Vioolsdrif (17.9419°E; 28.8878°S). 2.4.3.1 Ghaams Gran ite The Ghaams Granite is fine- to medium-grained aonmdm conly displays a gneissic texture. In the southern part of the Haib areah iwn ithe foliated fringe areas of the batholith, the Ghaams Granite’s characteristic olecuractic nature contrasts sharply with the granodioritic rocks of the Goodhouse SubsuFiitgeu (re 2.38). Also in these foliated 59 areas, the gneiss is exceptionally rich in musceo avnitd commonly associated with the development of pegmatite and greissen. Figure 2.38: Conatact between the leucocratic Gmhsa Garanite (right) and mesocratic granitoids of Gthoeodhouse Subsuite in the southern, foliated parts of theh oblaitth south of Ghaams (17.8850°E; 28.9659°S).e cDtior n of view is to the east. Figure 2.39: Leucocratic nodules and lenses i nG thaeams granite at a localtion south of Ghaamsc.a tLi on: 17.9227°E; -28.9647°S. 60 Occurrences of leucocratic nodules, lenses ando ndtiisncuous veins is a characteristic feature in certain parts of the Ghaams Graniteu (rFei g2.39). They comprise quartz, K- feldspar, subordinate plagioclase and muscovithe bwioittite and opaque minerals being concentrated along the centres. They are simoi lathro tse described for the Tsams Formation and are thought to originate during dmefaotrion and metamorphism through migration of felsic melts which originate by palr tmiaelting of the granites. 2.4.3.2 Sout Grani te The Sout Granite is medium- to coarse-grained,o lceruactic and typically void of dark minerals. The K-feldspar typically displays a riesdhd alteration (Figure 2.40). Very coarse-grained varieties are also found. It comlym woneathers to a woolsack weathering pattern. Figure 2.40: Reddish alteration of K-feldspar oina rcse-grained Sout Granite at a locality south aomf Kgab. Location: 18.0154°E; -28.8657°S. 61 3. PETROGRAPHY 3.1 ORANGE RIVER GROUP Within the Orange River Group, the matrix of aphtiacn lai vas are sometimes too fine- grained to recognize individual minerals under mthicer oscope. Commonly however, even though fine-grained, the matrices of the Oera Rnigver Group volcanics lend itself to microscopic investigation. The porphyrytic lavaoss mtly contain both quartz and feldspar phenocrysts. The latter phenocrysts are typiceaullhye dral to subhedral and the degree of alteration varies from entirely altered to unaldte (re.g., Figure 3.1). Alteration of the feldspars may be related to metamorphism durin gO trhaenge River Orogeny or to the presence of fluids during the late stages of clrlyizsatation of the volcanics. Sericite, soussurite and epidote are common alteration prtso dleuacding to the formation of a cloudy mass within the crystals. Figure 3.1: Thin section of leucocratic porphycr ylativa of the Tsams Formation showing a fine-grda imneatrix and euhedral, relatively unaltered plagioclase phensotcsr. y location of sample: 17.6632°E; 28.8044°S. 62 Figure 3.2: Thin section of quartz-feldspar porrpyh oyf the Kook River Formation showing porphyrobtsla csomprising quartz aggregate (flattened along the foliationd) uanaltered, subhedral feldspar. Location of saem: 1p7l .1771°E; 28.1833°S. The quartz porphyroblasts are rounded to subhe dTrhael.y may also be comprised of aggregates and may be flattened along the foli,a atiso nis commonly seen in the Paradys River Formation, which was greatly affected by PAafrnic an deformation (Figure 3.2). Tuffs throughout the Orange River Group containg mfraents of very fine-grained lava and chert. The dark mineral constituents in the Orange Rivreoru Gp include hornblende, biotite, chlorite, epidote and opaque minerals and musc ovcicteurs as an additional phase. Sphene occurs in accessory amounts. The prop oorft idoanrk minerals in the rock varies depending on the composition. The basaltic anedse soift the Nous Formation are dominated by plagioclase in which the alteratiornie vsa between completely altered to unaltered. Hornblende, biotite and epidote are p trheedominating dark minerals. Epidote occur mostly as alteration products but may alspore rseent primary crystals. Andesites throughout the Orange River Group also contain bhloernnde but to a lesser extent than in the Nous basaltic andesites. Muscovite becomeigsn aifi csant phase in the andesites. The 63 dark mineral contents is low in the dacites. Holerndbe is still present as a minor phase. Rhyolites do not contain hornblende while musco ivsi tae common phase. Hornblende, biotite and muscovite may occur either as primahrays pes or alteration products, while chlorite and epidote are alteration products thhrougt. Regarding the metamorphic grade, the greenschciisets-f aconditions inferred for the Orange River Group and Vuurdood Subsuite in var pioruesvious studies (e.g., Beukes, 1973; Blignault, 1977; Reid, 1977) is confirmedt hbey mineral assemblages studied. The green amphibole actinolite is a common constitueesnpte cially in the Vuurdood Subsuite and both hornblende and biotite display well-dedfi npeleochroism. The rocks of the Haib Subgroup are commonly altered to schist in are ahsig ohf strain deformation. In thin section, the foliation in these schists is definbey dd ark minerals orientated in a parallel array and varying in proportion in accordance wthiteh composition. Hornblende often occurs as large, secondary crystals overgrowin gp rtihmeary phases. A speckled appearance in hand-specimen may be caused eit hdearr kb ymineral aggregates or large, poikilitic hornblende or biotite flakes. In the DHeoop Subgroup, foliation related to the Pan African orogeny is pervasively present and irmtsp aa schistose texture in the rocks. Mainly the Paradys River Formation has been suebdje tcot intense strain. Two generations of feldspar are present. One gaetinoen rcomprises unaltered crystals while the other comprises cloudy, altered crys t aTlhse. degree of alteration within the latter generation varies among individual crys(tea.lsg ., Figure 3.3). This is mainly true for the Windvlakte and Abiekwa River Formationst,h b of which were intensely subjected to recrystallization, as seen in thed .f i eAls such, alteration of the first feldspar generation in these rocks could possibly be re ltaot emdetamorphism during which fluids were liberated. 64 Figure 3.3: Thin section of dacite from the Winadkvtle Formation. Two generations of feldspar ca nd isbteinguished namely an altered and unaltered phase. Locati osna mofple: 17.3549°E; 28.9432°S. 3.2 VIOOLSDRIF SUITE 3.2.1 Vuurdood Subsuite In the Vuurdood Subsuite, only remnants of the inoarilg minerals can be recognized in the altered phases. In some cases, the rock mill abye s utnaltered (e.g., Figure 3.4). The gabbros comprise plagioclase, orthopyroxene, cylirnooxpene, olivine and hornblende. Biotite, muscovite, chlorite, epidote, zoisite, aonpdaque minerals occur in varying amounts. Actinolite is a common alteration prod wuhcitle K-feldspar, quartz, apatite and allanite represent accessory phases. The peerisd oatnitd pyroxenites are commonly highly altered to serpentinite. Orthopyroxene (ehryspthene), olivine, phlogopite and opaque minerals comprise the other constituenatsd dinit ion to serpentine. The amphibolites consist of actinolite, hornblende asunbdo rdinate epidote. Limited occurrences of troctolite at Swartberg comprisvei noeli and plagioclase as major components and clinopyroxene and hornblende asr dsiunbaote constituents. 65 Figure 3.4: Thin section of olivine gabbro frome tVhuurdood Subsuite showing relatively unalteredn emrai logy comprising plagioclase, hornblende, clinopyroxende oalivine. Location of sample: 18.1197°E; 28.9°0S8.6 In the Vuurdood Subsuite as a whole, green ampeh icbonl stitutes the major phase – an estimated 41%. It is difficult to determine how cmhu of this amphibole is primary. REE patterns with positive Eu anomalies in some sam opfl ethse Vuurdood Subsuite (Figure 4.12), suggest that a substantial amount may bmea pryri, since positive Eu anomalies are commonly thought to indicate large-scale amphibfroalcet ionation or amphibole retention in the source (Rollinson, 1993). However, ther alttieon process from pyroxene to amphibole is very commonly observed (almost througt)h, suggesting that a large proportion of the amphibole represents alteratiroond upcts from pyroxene. The alteration is thought to be related to greenschist facies moertpahic conditions. This is based on the fact that the Vuurdood Subsuite and Oranger R Givroeup are of similar age and as such, probably underwent the same deformationa lm aentdamorphic processes. Greenschist facies metamorphism is common in thaen gOer River Group. 66 3.2.2 Goodhouse Subsuite In the Goodhouse Subsuite, alteration of the fealrdss pvaries between completely altered and unaltered. In partly altered crystals, altioenra mt ay be concentrated along the rim or in the centre. Epidote and sericite are commoenr aatliton products of the feldspars. Epidote alteration may be associated with a clomudays s or may occur as single flakes which are distributed randomly throughout the fpealdrs crystal. In situations where the plagioclase composition could be estimated by thiceh éMl-Levy method (using the extinction angle), an andesine composition predva. i lOene occurrence revealed an oligoclase composition (Hoogoor Granite). The dark mineral contents in the Goodhouse Sub rseusiteembles that of the Orange River Group. Biotite is found both as alteration prod oufc at mphibole and as a primary phase. Amphibole often occurs as a primary and secondhaarys ep. Most commonly, the amphibole crystals will display a poikilitic chartearc (Figure 3.5). Such phases can be identified as secondary in some samples where f oitu insd in contact with the primary phase. The primary phase is altered to biotite d aoneds not show a poikilitic character. Figure 3.5: Typical poikilitic character of primya hrornblende in granitoids of the Khoromus Tona. lLitoecation of sample: 17.9259°E; 28.9200°S. 67 Epidote most commonly represents an alterationu pcrto odf feldspar but also occurs as a primary phase in the form of idiomorphic matrix sctrayls. It is also often seen to develop in the centre of biotite flakes. Muscovite too m, isost commonly associated with alteration but also occur as primary crystals. o Crithel is an alteration product throughout, either of hornblende or biotite. Ortho- and clinyroopxene are limited to the mafic endmembers in which they occur in accessory am.o unts Estimated modal persentages in the Goodhouse Steu basreu ias follow (number of samples studied in brackets): Unit Qtz Fsp Bt Hbl Ep Chl Ttn Op Ms Cpx Opx Act Khoromus Tonalite (61) 33 38 13 3 7 2 2 2 acc acc acc Xaminxaip River Granodiorite (21) 34 49 5 5 3 acc 1 2 Gaarseep Granodiorite (106) 29 39 12 5 8 1 2 1 2 acc Hoogoor Granite (28) 37 38 11 4 3 1 2 2 acc 2 In the Khoromus Tonalite, contacts between soe-dca “lflilament” and “globular” areas as identified in the field (see Figure 2.28), are w deellfined in thin section (Figure 3.6). The “filament area” is represented by a relatively csoea frelsic matrix, the feldspar being generally unaltered and the dark mineral contewn.t loThe “globular area” is represented by a mafic matrix, rich in dark minerals and hig halytered feldspar. These textures may be interpreted as possible products of magma m ipxrinocgesses (Perugienti al., 2002). They may however, also represent products of pl amrteialting during migmatite-forming processes. 68 Figure 3.6: Thin section of granite from the Khmoruos Tonalite showing a well-defined contact betw ae emnafic (top of photo) and a felsic matrix (the contact trans ethcet scale bar at approximately 1.35 mm). Thweose m tatrices may be interpreted as “globular” and “filament” areaesp rectively in terms of magma mixing models (Penriu egti al., 2002). Location of sample: 17.9260°E; 28.9187°S. Figure 3.7: Thin section showing a normally zonpeladg ioclase crystal in granite from the Khoromusn aTloite. Location of sample: 17.9260°E; 28.9187°S. 69 Another observation speaking in favour of magmai nmgi xprocesses in the Khoromus Tonalite, is the frequent occurrence of zoned fpealdr scrystals (Figure 3.7). This zonation is found to be both normal and reevde.r s Although less common, zoned hornblende crystals can also be found in the Khourso Tmonalite. Zoned crystals may also be explained in terms of normal fractionals ctaryllization processes during which they were in disequilibrium with the melt when th cerystallized. However, zoned crystals were not noted for any of the other u n Tiths.e Gaarseep Granodiorite, with which the Khoromus Tonalite is closely associaltaecdk, s any evidence of magma mixing processes, both in the field and in thin sectiMono.r e detailed studies are needed to confirm a magma mixing model for the Khoromus Toitnea. l In the coarse-grained granitoids of the Gaarseeapn oGdriorite, coarse, unaltered microcline crystals often contain inclusions of eorl dphases such as altered feldspar, quartz and dark minerals (Figure 3.8). Figure 3.8: Thin section of Gaarseep Granodiosrhitoew ing inclusions of well rounded, pervasivelye raeltd feldspar and quartz within an euhedral and relatively unreadlte microcline phenocryst. Location of sample:0 175.2°E; 29.0294°S. 70 Figure 3.9: Thin section showing euhedral, uneadlt esrecondary hornblende containing inclusions oafq oupe minerals and altered primary dark minerals, overgrowingp trhime ary, altered mineralogy in the Gaarseep Graonroitdei. Location of sample: 17.9664°E; 28.9919°S. The altered feldspar generation may occur inteiarlslyti tin the form of pods, veins and stringers, or disseminated within a matrix of uenraeldt feldspar and quartz. In areas affected by the Namaqua deformation, hornblende b aiontdite often also occur as two generations, the second being due to metamorpfheict se.f These second generation crystals are occasionally seen in contact witho tlhde r altered phases and may contain inclusions of them, attesting to their secondartyu rnea (Figure 3.9). Both the hornblende and biotite are orientated in a preferred orieonnta, timparting a foliation on the rock. The Hoogoor Granite is entirely overprinted by Naqmuaa foliation. The typical Hoogoor Granite has a common granitic composition with sceo aarlkali feldspar and a low dark mineral content. Microcline dominates the alkealdi fspar compositions but orthoclase also occur. Plagioclase also forms part of thelt eurnead matrix. Hornblende and biotite represent the dark mineral constituents and aretl ym iodsiomorphic, unaltered and strongly pleochroic. 71 The degree of metamorphism of the minerals in tohoeg Hoor Granite increases from west to east. In the western parts around Goodhouisgein, aolr amphibole-rich rocks may be altered to actinolite-epidote gneiss, after greheinsst-cfacies metamorphism. The feldspars are extremely altered and the original characticesri sctan be recognized only in rare cases. Hornblende is associated with alteration and fnlaetdte along the foliation (Figure 3.10). In places, the rock is porphyroblastic with thep phoyrroblasts consisting of quartz aggregates which also encloses certain alteredsp faerld masses. Further to the east, altered feldspars are rarely observed and the isro dcokm inated by coarse, unaltered and undeformed alkali feldspar, Here, altered feld spphaarses occur only as small, rounded inclusions in the coarse and porphyrytic, unalt earlekdali feldspars, or as interstitial stringers. Figure 3.10: Thin section of foliated Hoogoor Gitrea ndisplaying pervasively altered feldspar, qu,a grtrzeen amphibole altered to epidote in places and biotite, fromw thees tern part of the unit. Location of sample: 1881.73°E; 28.9235°S. 3.3.3 RAMANSDRIF SUBSUITE In the Ramansdrif Subsuite, estimated modal comtipoonsi are as follow: 72 a) Ghaams Granite: (24 samples studied) quartz% -, 4fe3ldspar - 38%, muscovite - 9%, biotite - 5%, chlorite - 1%, epidote - 1%, oupea qminerals - 1%. The muscovite contents is exceptionally high in the southerng feri nareas of the batholith where the rocks are affected by the Namaqua foliation. Also ins teh esouthern fringe areas, the gneiss occasionally contain leucosome nodules with gadrneevet loped in their cores. b) Sout Granite: (12 samples studied) quartz - 5 a0l%ka,li feldspar - 39%, plagioclase - 5%, muscovite - 2%, biotite - 1%,o crihtel - 1%, opaque minerals - 2%. Figure 3.11: Thin section of Ghaams Granite wai tfho liation caused by the linear distribution oaft tfel ned pervasively altered feldspar and biotite occurring interstliyti almong undeformed and unaltered feldspar andt zq.u aLrocation of sample: 17.9573°E; 28.9947°S. Muscovite is a common constituent in the Raman sSdurbifsuite, mostly associated with sericite as alteration products of the feldspaHros.w ever, secondary muscovite also occur as unaltered, idiomorphic flakes. The Ghaams Gter aisn iconfined to the areas affected by the Namaqua foliation and here, idiomorphic snedc goeneration biotite flakes typically display strong pleochroism varying between lighot-wbnr and dark-green. The foliation in the Ghaams Granite in these areas is impartedi ebnyt aotred biotite and/or muscovite. It 73 may also be the result of pervasively altered fpealdr sand biotite being aligned and flattened in parallel layers, occurring interstlliyti aamong the unaltered second phase (Figure 3.11). Based on petrographic correlati oanll st,he occurrences of the Ramansdrif Subsuite which are located within the Hoogoor Gtrea, nciorrelate with the Ghaams Granite. In the Sout Granite, the feldspars are predominy acnotal rse- to very coarse-grained, unaltered and void of inclusions. Where limitetde raaltion do occur, sericite is almost exclusively the alteration product. A finer-gradin meatrix is typically developed interstitially among the coarse phenocrysts. Tinhtiesr stitial matrix occur in the form of pods and veins. As in the case of the phenoc rtyhset sin, terstitial feldspars do not show any appreciable alteration or deformation (e.gg.,u rFei 3.12). It is interpreted to represent the late stages in a normal crystallization his.t ory Figure 3.12: Thin section of Sout Granite showtiwnog alkali feldspar phenocrysts with interstitiianle f-grained matrix occurring in the form of a vein. location of same:p 1l7.8901°E; 28.8774° S. 74 Figure 3.13: Thin section showing first phaser aeldte feldspar and dark minerals being deformed a tlhoen gmargins of a second phase feldspar in granodiorite of the Xamaipin xRiver unit. Location of sample: 17.4437°E;7 2784.7°S. Two generations of coarse feldspar may therefo rre cboegnized in the Vioolsdrif Batholith namely phenocrysts (magmatic in origind) aporphyroblasts (metamorphic in origin). In porphyrytic samples of the Sout Graen, itnterstitial feldspars are nowhere found to be deformed along the rims of the phensotcsr, yattesting to a magmatic origin (Figure 3.12). Evidence attesting to feldspar tbhleasis is found where the interstitial matrix is seen to be deformed along the rims odfs fpeal r porphyroblasts, e.g. as in Figure 3.13. In such cases, minerals in the deformedi xm matary sometimes be seen to be broken or cracked. Exsolution textures are very rarely observed anlyd on small scale, in feldspars of the Goodhouse Subsuite as well as the Ghaams Graen.igte.,, F( igure 3.14). 75 Figure 3.14: Thin section showing rare myrmekietev edloped along the edge of a quartz crystal in Hooor gGranite. Location of sample: 18.5467°E; 28.8793°S. Figure 3.15: Thin section showing perthite whisc hc oi mmonly developed in the Sout Granite in itse tayprea south of Nous. Location of sample: 17.8785°E; 28.8703°S. 76 However, these textures are very common in thes pfealrd of the Sout Granite. In the Sout Granite type area around Nous, as well as in thceh tRerisveld National Park to the east of Sendelingsdrif, perthite represents the prevaielixnsgo lution type (Figure 3.15). In the occurrences in the Haib area to the east of Kam egxasbo,lution is less well developed, appearing only in the larger phenocrysts. The opchreynsts at this locality are notably unaltered. 77 4. GEOCHEMISTRY 4.1 GEOCHEMICAL VARIATIONS A number of classification and variation diagramilsl bwe presented here for the Vioolsdrif Batholith with two aims in mind: a) To geochemically compare the various units ein b tahtholith with each other. b) To identify the tectonic environment and asstoecdi amagmatic processes. Some of the evidence presented here have alreaedny p bresented in previous studies. However, a repeat of such evidence is justifietdh ein light of significant additional data as well as some advances in classification syssteinmcse the last published previous study. Furthermore, no previous study presentgeedo ac hemical comparison of the various units based on the current subdivision. 4.1.1 Classification diagrams A diversity of classification systems for igneouosc krs are in use, reflecting the fact that they may be produced by a variety of processe sv ainr iaety of tectonic settings. The total alkalis-silica (TAS) diagram is one oef tmh ost useful variation diagrams and has been shown by Coext al. (1979) to present a sound theoretical bas isth feo r classification of volcanic rocks. The current vioenrs of the diagram was constructed by Le Maitre et al. (1989) from a large database of volcanic ro cFkisg.ure 4.1 shows the Orange River Group plotted on this diagram. 78 Figure 4.1: The two subgroups of the Orange RGivreoru p on the total alkali-silica diagram of Le Mrea iet t al (1989). Alkaline-subalkaline boundary is that of of Irvinane d Baragar (1971). The boundary dividing the alkaline from the subalinlkea (tholeiite) series after Irvine and Baragar (1971) is also shown. The Orange Riveru pG roepresents a well defined subalkaline series. The average composition insi tgicr awith SiO2>63%. Deviations into high alkali contents may be caused either by pryim fealrdspar phenocrysts or later feldspar blasthesis (as suggested in thin sec twiohni)c,h lead to a high feldspar:quartz ratio in the rock. Deviations into low-alkali arse ma ay be caused by alkali depletion during metamorphism. A distinction can be obse rbvetdween the Haib and De Hoop Subgroups in that the latter subgroup has no elqeunitv acompositions to the basaltic andesites of the former. 79 Figure 4.2: The Vioolsdrif Suite on the classitfiocan diagram of De la Rocheet al. (1980). R1 = [4Si – 11(Na + K) – 2(Fe + Ti)]; R2 = Al + 2Mg + 6Ca. The R1-R2 classification diagram of De la Roceht ea l. (1980) is used here (Figure 4.2) to distinguish between the individual units within tVhieoolsdrif Suite. The names of these units are based on their average compositionsh. o Auglth the diagram has the disadvantage that the data range plots in a smreaall oaf the diagram, it incorporates the more commonly used granite nomenclature. Furthreer,m ito takes into account all the more abundant major elements and depicts the ivoanr ioaft SiO2 as well as the changes in Fe/(Fe+Mg) ratio and plagioclase composition. The Khoromus Tonalite and Gaarseep Granodiorite heanvdmembers sharing the gabbro field with the Vuurdood Subsuite, thus coincidinigth w the basaltic andesites of the Haib 80 Subgroup. In the Gaarseep Granodiorite, thesec m enafdimembers are all from the Haib area. The sample from the Blockwerf Migmatite which pl oints the syenodiorite field is porphyrytic and seen in thin section to be highltley raed. This suggests that its anomalous plot is the result of alkali loss during metamorspmh.i The Xaminxaip River Granodiorite may be distinguished on the diagram based onm itiste ldi compositional variation within the granodiorite field. Although the Khoromus Tloitnea and Gaarseep Granodiorite largely overlap, the two units can be distinguis ohne dtheir average compositions. The Gaarseep Granodiorite also extends across the ecnotmirpositional range defined by the two units, while the Khoromus Tonalite does note enxdt into the granite field. The Gaarseep Granodiorite in the Richtersveld area bmea dyi stinguished from that in the Haib area based on the fact that it shows a dtislyt incarrower and more felsic range, being almost limited to the granodiorite field. e T Hhoogoor Granite can be distinguished from the Gaarseep Granodiorite being more liminte cdo imposition, as well as more felsic. Although the Ramansdrif Subsuite overlaps withf ethlsei c endmembers of both the Gaarseep and Hoogoor units, it does define a dctis ativnerage composition. The Ghaams and Sout Granites both occupy the range monzoger-asnyeitnogranite-alkali granite. However, the Ghaams Granite average compositioshni fitse d to the right along the R1 axis, compared to the Sout Granite. This can tbrieb uatted to the higher dark mineral contents of the Ghaams Granite (see 3. Petrogr,a lpehayd)ing to elevated Fe and Ti contents. Norm calculations have been executed on all theila abvlea samples from the Vioolsdrif Batholith. In Figure 4.3, the Vioolsdrif Suite a Ondrange River Group are plotted on the classification diagrams of Streckeisen (1976) flourt opnic and volcanic rocks. This classification scheme is currently the most widueslye d for igneous rocks. The calcalkaline trend is clearly defined for both amsbselages. 81 Figure 4.3(a): The Vioolsdrif Suite (after normlc cualations) plotted on the classification diagrafm S otreckeisen (1976) for plutonic rocks. Figure 4.3(b): The Orange River Group (after n ocramlculations) plotted on the classification diamg roaf Streckeisen (1976) for volcanic rocks. 82 4.1.2 Harker diagrams Figure 4.4 shows Harker diagrams for selected enletsm ine the Vioolsdrif Batholith. For the La/Yb ratio, chondrite normalized values aBfteory nton (1984) are used. On Harker diagrams, the trend of an igneous rocitke swuill typically follow what Bowen (1928) defined as the “liquid line of descent”, wchh ican be directly related to the fractional crystallization process. However, o tmheargmatic processes, specifically partial melting, will produce the same trend. Nuromues studies show that only rarely will a suite of volcanic rocks, showing a progressiveem cihcal change, erupt as a time sequence (Rollinson, 1993). This is also truet hfoer O range River Group in which the oldest volcanics have felsic compositions (the Tss aFmormation; Reid, 1977). Therefore, Harker diagrams do not go a long way in distingiunigs hmagmatic processes. They are however, useful in comparing geochemical variat iaomnsong the various units within the same igneous suite. They are also useful in ifdyeinngti the influence of metamorphism and alteration. Figure 4.4: Harker diagrams for selected elemienn tthse Vioolsdrif Batholith. 83 Figure 4.4: (Continue). 84 Figure 4.4: (Continue). Metamorphic alteration in the Hoogoor Granite i sb eto expected given the pervasive overprint of the Namaqua orogeny. The elemen2Ots, KNa2O, Pb and Rb all show a linear increase from mafic to felsic and relativ heilgyh values are to be expected in the felsic Hoogoor granites. However, some sampletsh iso fu nit show significant depletion in these elements. This can be ascribed to ele moebnitlity during metamorphism since these elements are classified as incompatible me oebleilments (e.g., Rollinson, 1993). The depletion in Rb and Pb specifically, has imaptliiocns for age determinations by the Rb-Sr and U-Th-Pb systems. However, for the puer poof sthe current study, the deviation of these samples from the average trse ndo ti considered a major factor. When correlation among units is considered, multi-eletm aennd REE patterns carry more weight as evidence since an array of elements are usmedu ltini- element diagrams and not single ones, and the REE are notoriously resistant tou einfcle from metamorphism. Alteration is also evident in the Windvlakte andi eAkbwa River Formations of the De Hoop Subgroup, which is in accordance with field apnetrological evidence suggesting large-scale recrystallization in these formatio nAs n. umber of samples from these two formations are depleted in2 OK and Rb, both of which are classified as mobielem elnts (e.g., Rollinson, 1993). The rest of the De Hooupb gSroup coincide closely with the average trend. 85 The Vuurdood Subsuite is unique in its mafic-ultarafimc composition and represents a clearly defined group at the mafic end of the tr eond all the diagrams. Despite this, it displays a definite association with the rest oef bthatholith, always falling on the extension of the trend defined by the rest of tahteh oblith. For some elements, the trend defined by the rest of the batholith continuesa lirnlye into the Vuurdood Subsuite, e.g., MgO, K2O, Rb and Pb. For others, (e.g.2, OA3l, Cr and Sr), the trend defined by the rest of the batholith widens and curves at the mafic seon das to coincide with the field defined by the Vuurdood Subsuite. The Sr diagraismpl adys this especially well. Within the Goodhouse Subsuite (in all of the diamgsr ain Figure 4.4), there is a broad overlap between the Khoromus Tonalite, Gaarseepn oGdrioarite and Hoogoor Granite. The Gaarseep Granodiorite extends across the efinetldir eoccupied by these three units while the Khoromus Tonalite is concentrated atm thaefi c end and the Hoogoor Granite at the felsic end. The latter two units do not ovpe.r l aThe Xaminxaip River Granodiorite is distinguished by its limited variation, which coiidnecs with equivalent compositions in the Gaarseep Granodiorite. (Although the Xamin xRaiivper Granodiorite samples are few, they do represent the entire compositionagl era wnithin the unit, based on field evidence.) The Blockwerf Migmatite is distinguisdh oen the Harker diagram (Figure 4.4) for the La/Yb ratio, on which it shows values ab othve average trend. This elevated La/Yb ratios is the single most characteristic gheomcical feature of this unit. The Ramansdrif Subsuite coincides with the felseicm mbers of the Goodhouse Subsuite in Figure 4.4 (in all the elements) and trends iwn ithe two subsuites are similar for individual elements. Within the Ramansdrif Subes,u tiht e Ghaams Granite is on average slightly less felsic than the Sout granite. The Harker diagram (Figure 4.4) for the La/Yb r ahtigohlights three distinguishing features in the Vioolsdrif Batholith when compar ignegochemical variations among the units: 86 a) It distinguishes the three subsuites of the lVsidoroif Suite. The Vuurdood Subsuite does not show a linear trend, while tehned t rfor the Ramansdrif Subsuite is steeply positive. In the Goodhouse Subsuite, rtehned t is flat-lying linear (if the anomalous samples of the Blockwerf Migmatite iso irgend) with a moderate positive slope. b) It shows a good correlation between the Goodeh oSuusbsuite and the Orange River Group, suggesting that they share similar maatgic processes during their evolution, which are not shared by the Vuurdood Ranadmansdrif Subsuites. Note that only the De Hoop Subgroup is represented on thaigsr daim for the Orange River Group, as no REE analyses are available for the Haib Soubpg. r c) It distinguishes the Blockwerf Migmatite in eitsle vated La/Yb ratios compared to the rest of the batholith. There is a notable association between the Vuur dSouobdsuite and the Haib Subgroup in the elements A2Ol 3 and Cr (also for TiO2, P2O5 and Ni, which are not shown). At the mafic end, the trend for the Haib Subgroup dev iafrtoems the main trend defined by the other units, curving into lower values of the Vuouord Subsuite. This association is not displayed by the De Hoop Subgroup, nor by the Gouosdeh Subsuite, both of which continue linearly into the higher values of the Vrduouod Subsuite. On all the other diagrams, the Haib Subgroup also represents, efo mr tohst part, the range of values intermediate between the Vuurdood Subsuite andre tshte o f the batholith, between which there is typically a notable gap. The De Hoop Srouubpg displays a striking association with the Goodhouse Subsuite throughout. The association between the Vuurdood Subsuite anibd SHubgroup as described above, can be shown to be a genetic one, and not influde bnyc elater metamorphic or alteration processes. The fact that both Cr and Ni follows threind supports a primary origin since both these elements are highly compatible and imilme o(eb.g., Rollinson, 1993). Furthermore, deviations from the average trendig thoe hr and lower values for the Haib 87 Subgroup in a wide range of elements (2T, iAOl2O3, CaO, K2O, MgO, P2O5, Cr, Ni, Rb, Sr, V and Zn), are distributed evenly and not conntrcaeted below the average trend. If these deviations were caused by alteration, theo rmitya wj ould fall below the average trend, as is the case for the Hoogoor Granite haen dW tindvlakte and Abiekwa River Formations. 4.1.3 Eigenvectors and eigenvalues Le Maitre (1976) used eigenvectors and eigenvatlou esstu dy in more detail the geochemical variability within diferent igneous kro tcypes. Eigenvectors represent the comparison of straight lines in a multidimensiosnpaal ce in the direction of maximum variation. Most commonly, the first two or threieg eenvalues represent between 80-90% of the chemical variation within a rock and as s, uitc ihs possible to construct a two dimensional projection representative of the maxmim vuariation. In Figure 4.5(a), the Vioolsdrif Batholith is plotted along eigenvectowrsh ich were calculated for all igneous rocks (Le Maitre, 1976). On this diagram, the eth vrectors represent 92.1% of the total chemical variation in the rocks. No clear distiionnc tcan be made between the units. Since the average composition of the Vioolsdrifh Boalith is granodioritic, eigenvectors calculated for this composition may be able to rdimisicnate among the various units. This is shown in Figure 4.5(b) with the accumuliavtee teigenvalue representing 90.3% of the chemical variation in the rocks. Again, not idnicstion can be made among any of the units. 88 Figure 4.5(a): The Vioolsdrif Batholith plottedo anlg eigenvectors calculated for all igneous rocLkes M ( aitre, 1976). Figure 4.5(b): The Vioolsdrif Batholith plottedo anlg eigenvectors calculated for granodiorites (Laeit rMe, 1976). 89 The fact that the eigen diagrams do not disting uamishong the various units of the batholith means that field evidence must be regda ardse the most important indicator for subdividing the batholith into its composite un itTsh.e reason why the geochemical composition of similar rock types among the var iouunsits of the batholith coincide so closely, must be related to primary igneous proecse,s ssince variation diagrams (in agreement with field evidence) show that metamocr pmhoibilization of the elements did not play a very sigificant role. The gradationealal rtionship between the Vuurdood Subsuite and Haib Subgroup as displayed on Hariakegr adms is also evident on the eigen diagrams. This indicates that this relationsh iipn diseed the result of chemical properties and is not due to projections. 4.1.4 Multi-element diagrams Multi-element diagrams compare trace element viaornias tin rocks relative to a standard. Since trace element fractionation is more sens tithivaen major elements, and since it is primarily controlled by magmatic processes, mulletim-eent diagrams are useful to provide evidence on the nature of these proceasse wse, ll as in the correlation among various rock units. In this study, a chondrite model with normalizinagl uves of Thompson (1982) is used. A selection of incompatible trace elements are arerda nfrgom left to right in order of decreasing mobility (LILE to the left, HFSE to trhigeh t). For evidence on the magmatic processes involvethde i nfo rmation of the Vioolsdrif Batholith, as well as for initial correlation amo tnhge main units, multi-element diagrams are constructed for average concentrations of enletsm ine the Orange River Group and the three subsuites of the Vioolsdrif Suite (Figure) .4 .6 90 Figure 4.6: Comparison of the multi-element vaiorina tpatterns for the three subsuites of the Vioroifl sSduite and the Orange River Group, including REE. The Vioolsdrif Batholith as a whole displays chateraricstics which are considered distinct of the subduction zone environment. They incluhdee f ot llowing: a) High LILE:HFSE ratios (decoupling of the LILEo fmr the HFSE). This is considered to be the result of metasomatism inm tahnet le wedge by hydrothermal fluids originating from the dehydration of the subductbinags altic slab which underlies it (e.g., Pearce and Parkinson, 1993; Pearce and Peate ,D 1h9u9i5m,e et al., 2007). These fluids also act as a flux which promotes melting in then tmlea wedge, leading to the extraction of the LILE to be carried upward and concentrante tdh ei evolving overlying crust. Dry peridotite solidus is too high for melting to tapklea ce in the mantle wedge. The high LILE:HFSE ratios of arc magmas show that water sp la ysignificant role in arc magmatism. This subduction of volatile componeton tpsr ovide fluxing materials for the melting of the overlying material, is what distinisghues convergent plate boundaries from all other tectonic settings (Wyllie, 1983). 91 b) Prominent troughs of especially Nb and Ta bust tm ooften also Ti, Zr and Hf. This is a characteristic feature of modern arc avnoilcs and is often referred to as “the subduction component” (e.g., Condie, 2005b; Caos, t2il0l 06). The mechanism of how it is acquired is still unclear as it implies deconugp liof these elements from the LILE and REE. Depletion of Nb and Ta is thought to be cda ubsye the retention of rutile in the source. The patterns in Figure 4.6 for average values ein G thoodhouse Subsuite and Orange River Group correlate very well. Patterns for Vthueu rdood and Ramansdrif Subsuites are each unique. Troughs at Zr in the Vuurdood R aanmdansdrif Subsuites are not developed in the Goodhouse Subsuite, nor the O rRanivgeer Group. The Vuurdood Subsuite is furthermore distinguished by small nSdr Pa troughs. For more detailed comparison of trace element tvioanrisa among the various units of the batholith, individual rock compositions have toc boen sidered. Due to limited ICPMS analyses (specifically insufficient REE data foer tOhrange River Group), the REE, Hf and Th are omitted from the diagrams that folloFwig. u re 4.7 shows multi-element diagrams for average trace element concentratnio tnhse i various rock types within the various units of the Vioolsdrif Suite. Also showarne the variations in the series from mafic to felsic (gabbro-diorite-tonalite-granoditoer-igranite) in the Goodhouse Subsuite. For the Vuurdood Subsuite, patterns for the ultrfaicm raocks and gabbronorites are also shown, however, gabbro is the only rock type wqithu ivealent compositions in the rest of the batholith. The patterns of the individual r otycpkes in the Vuurdood Subsuite are variable both in total concentrations and in sh a Hpoew. ever, it is important to note that this subsuite, like the other units, displays haell ct haracteristics of the subduction zone environment. This implies that it was formed bey tshame processes as the other units. The lowest concentrations of both Nb and Zr aren dfo iun the Vuurdood Subsuite. The pattern of gabbro in the Vuurdood Subsuite dnoet scorrelate with the gabbros in the Khoromus Tonalite and the Gaarseep Granodioritoet.h Bthe LILE and Nb 92 concentrations increase systematically from ther Vdououd Subsuite to the Khoromus Tonalite to the Gaarseep Granodiorite. The HFS tEh eo fKhoromus Tonalite and Gaarseep Granodiorite are very similar. Sinceg tahbeb roic composition is shared by the three units, the increase in LILE concentrationnsn coat be attributed to fractional crystallization. It may be a function of progrensgs imetasomatism in the mantle wedge. This is in agreement with their relative ages, edaescirng from the Vuurdood Subsuite to the Khoromus Tonalite to the Gaarseep Granodio rBitoet.h Nb and Zr show a sudden large increase between the Vuurdood Subsuite aen tdw toh units of the Goodhouse Subsuite. While Nb still displays a trough in tghaeb bros of the latter subsuite, Zr does not. This implies the fusion of zircon and par ftuiaslion of rutile during the formation of the Goodhouse Subsuite, which was not the casneg d tuhrei formation of the Vuurdood Subsuite. This must be the result of an increna sfues iion temperatures and a possible explanation would be that the zone where meltincgu rosc above the subducting slab, increased in depth possibly due to an increashee i nth tickness of the overlying continental crust. Figure 4.7: Multi-element diagrams for averagec ceonntrations in the various rock types and unittsh eo fV ioolsdrif Suite. 93 Figure 4.7: (Continue). Within the Goodhouse Subsuite, a good degree oref lcaotiron is displayed when the various rock types are compared. Diorites occ uthr ein Khoromus Tonalite, Blockwerf Migmatite and Gaarseep Granodiorite and their rpnast tfeit very closely. Tonalites occur in the Khoromus Tonalite, Blockwerf Migmatite, Gsaeaerp Granodiorite and Hoogoor Granite and again their patterns also correlatel. w Tehle fact that the Blockwerf Migmatite is not distinguished from the other u noints these multi-element diagrams where the REE are not considered, is noteworthrya.n oGdiorites occur in the Khoromus Tonalite, Xaminxaip River Granodiorite, Blockwerfi gMmatite, Gaarseep Granodiorite and Hoogoor Granite, their patterns being almodsits tininguishable. Granites occur in the Gaarseep Granodiorite, Hoogoor Granite and Ramiaf nSsudbrsuite. Their patterns show 94 some variation in Sr, P and Ti, however, therea rigs el variation among the patterns in each unit individually (not shown) and thereforaer,i avtion in their average patterns is to be expected. The variations do not distinguish oafn thye units as unique among the others. When the Gaarseep Granodiorite is conesdid ienrdependently, no distinction can be made in any of the rock types between the Hnadib R aichtersveld areas. The variation pattern in the series from mafic to felsic (gabdbiror-ite-tonalite-granodiorite-granite) in the Goodhouse Subsuite shows systematic incre aLsILeE in and Nb concentrations and systematic decrease in Sr, P and Ti concentra t iAonll st.hese are inaccordance with variations on Harker diagrams and in accordanche aw iftractional crystallization mod el. Figure 4.8 shows multi-element diagrams for ave rtaragcee element concentrations in the various rock types within the various units of Othrea nge River Group. Also shown is the variation in the series from mafic to felsic (batisca al ndesite-andesite-dacite-rhyolite). Anomalous samples were omitted from the calcula otifo tnhese average trends. Figure 4.8: Multi-element diagrams for averagec ceonntrations in the various rock types and unittsh eo fO range River Group. 95 Figure 4.8: (Continue). The basaltic andesites in the Tsams Formation snhootawb ly lower LILE concentrations than those in the Nous Formation. Andesites o fA tbhiekwa River Formation show on average lower LILE values than those of the othoermr fations, which in turn show good correlation. Based on field and petrographic envcidee as well as evidence from Harker diagrams, lower LILE contents of the Abiekwa RivFeorr mation can be attributed to element mobility during metamorphism. The dacaitett eprns of all the formations coincide very well. The Rb and K concentration sth ien pattern for the Windvlakte Formation andesite are anomalous compared to sr irmoiclaks in the rest of the Orange River Group. However, this pattern representsv aenra age which was calculated off only one sample which may not be representative. A tsh feo rAbiekwa River Formation, all 96 the evidence points to element mobility during mmeotraphism as the cause of the low Rb and K values. The average patterns for daciter haynodli te in the Windvlakte Formation are not anomalous in relation to the other unTithse. r efore there is no reason to believe that the anomalous pattern in the single andeasmitep lse is related to magmatic processes. As in the Goodhouses Subsuite, there is a systce minactriease in LILE concentrations and an associated decrease in Sr, P and Ti, in thaet vioanr ifrom mafic to felsic. Again this is in agreement with variations on Harker diagrams manady be related to a fractional crystallization process. In Figure 4.9, the Vioolsdrif Suite and Orange Rr iGveroup are compared through average multi-element diagrams. Figure 4.9: A comparison of the multi-element avatiorin patterns between rock types of the Viools Sduriifte and Orange River Group. 97 Figure 4.9: (Continue) For the Vuurdood Subsuite, only the pattern forb groa bis shown as it is the only rock type in the subsuite which is comparable to thet orfe tshe batholith. It displays the largest Nb trough in the batholith while it is a ltshoe only one which displays a Zr trough. The pattern for the basaltic andesites of the HSauibg roup is distinct in its LILE and Nb concentrations and falls between that of the Vuoudrd Soubsuite and the gabbros of the Goodhouse Subsuite. In most of the HFSE (exce, pit cPo)incides very closely with the gabbros of the Goodhouse Subsuite. Since thet ibca asnadl esite composition is comparable to that of gabbro, the increasing LILnEd Nab trend from the Vuurdood Subsuite to the De Hoop Subgroup to the Goodhoubsesu Site, cannot be attributed to fractional crystallization. Again (as previousloys ptulated for a similar pattern between the Vuurdood Subsuite, Khoromus tonalite and Gaeapr sGeranodiorite) it may be the function of progressing metasomatism in the mawnetled ge with progressing subduction. However, according to radiometric data, the Haib gSrouup is either contemporaneous with the Vuurdood Subsuite, or predates it. Anro tphoessible explanation of the elevated LILE and Nb concentrations in the Haib Subgroup pcaormed to the Vuurdood Subsuite, is contamination of the Haib magmas with these enletms as they rise to surface. But if this was the only mechanism leading to the highaelur evs, then the Haib Subgroup’s values should also be higher than that of the Gouosdeh Subsuite, which is not the case. As such, the two processes probably both playeodle a. r 98 There is no equivalent composition to the Haib Srouubpg’s basaltic andesites in the De Hoop Subgroup. For the other rock types, the rpnast tfeor diorites in the plutonic suite and andesites in the volcanic group coincide cylo, stheel granodiorites and dacites correspond as well as the granites and rhyolites. 4.1.5 REE patterns A limited number of REE analyses were done forO thraen ge River Group during this study and none are available from previous stu d Oiefs t.hose that were done during this study, all are for the De Hoop Subgroup. There, faonre REE investigation of the volcanic group extended only as far as the constructionn oafv aerage pattern (Figure 4.10), which, as such, is representative only of the De Hoop rSouubpg. Like multi-element diagrams, REE patterns are ul sine ftuhe correlation among various rock units of the same suite, since their fracttiioonna is even more sensitive than the other trace elements, and also primarily controlled byg matic processes. They are especially useful in testing the process of fractional crylisztatlion because theoretically, their behaviour during this process can be accurateldyi cptreed, based on their distribution coefficients (e.g., Rollinson, 1993). According s tuoch predictions, both the size of the Eu-anomaly and the La/Yb ratio, should increasteh ein r ange from mafic to felsic in a rock unit which is the product of this process e(gni vthe mineral composition of the Vioolsdrif Batholith). Figure 4.10 shows REE patterns of average conctieonntsra in the Orange River Group and the three subsuites of the Vioolsdrif Suiteh.e TBlockwerf Migmatite is shown here independantly as its slope is different in rela ttio nthe rest of the batholith to such an extent that it distorts the pattern for the Goodsheo Suubsuite if it is included in it. 99 Figure 4.10: Patterns for average REE abundanc tehse ithree subsuites of the Vioolsdrif Suite ahned Ot range Rive r There is very good correlation between the pat toefr nthse De Hoop Subgroup and Goodhouse Subsuite, while the Vuurdood and Ramiaf nSsudbrsuites each defines a distinct pattern. Although the Vuurdood Subsuiatett eprn has lower overall concentrations than those of the Goodhouse Sub asnudit eOrange River Group, its shape is nearly parallel to these two units. This spe ina kfsavour of a similar source for the three units. The pattern for the average Ramafn Ssudbrisuite shows lower overall concentrations and a flatter slope than that o fG tohoedhouse Subsuite. This would not have been the case if the Ramansdrif Subsuites renptreed the final stages of crystallization in a continuous fractional crysitzaalltion process for the suite as a whole. The pattern for the Vuurdood Subsuite in Figure0 4s.h1ows a negative Tb anomaly. In mafic melts, the MREE are controlled by amphibonled aclinopyroxene with their distribution coefficients the highest in hornble n(dee.g., Rollinson, 1993). Therefore, the development of negative Tb anomalies in the Vuudrd Souobsuite may be regarded to indicate an amphibolitic source composition. WhReEnE patterns for the different rock 100 types in the Vuurdood Subsuite are considered (rFesig 4u.11 and 4.12), it can be seen that the Tb anomalies are characteristic of the ultraicm raofcks and not of the gabbro. This may indicate that some of the ultramafic samplepsre rseent cumulates of amphibolite and pyroxene. For a more detailed investigation of the Viools dSruifite, REE patterns for average concentrations in the various rock types within vthaerious units of the suite are shown in Figure 4.11. Due to insufficient REE data for Othrea nge River Group, the unit is excluded from these investigations. Figure 4.11: Patterns for average REE concenntrsa tiino the various rock types within the varioust su noif the Vioolsdrif Batholith. Chondrite normalized valuoefs B oynton (1984). 101 Figure 4.11: (Continue). Within the Vuurdood Subsuite, the ultramafic rochkasv e much lower overall concentrations than the gabbro, but similar sloepxec e(pt the Tb anomaly). The gabbro in the Vuurdood Subsuite shows good correlation whiatht itn the Gaarseep Granodiorite, both in slope and overall concentrations. The Blockwerf Migmatite is distinguished in its espteer slope (higher La/Yb ratios) compared to similar compositions in other unitsio. r Dite patterns in the Khoromus Tonalite and Gaarseep Granodiorite show good caotirorenl. Patterns for tonalite among the various units show variation mainly in the HR cEoEncentrations, while the LREE concentrations show better correlation. Patteornr sg rfanodiorite among the various units in which it occurs, show very good correlation epxtc tehe pattern for the Khoromus Tonalite, which shows higher overall concentrat.io Gnsranite patterns for the Gaarseep Granodiorite and Hoogoor Granite show very goodre claotrion, while the granites of the Ramansdrif Subsuite can be distinguished by ae frl astltope. The Gaarseep Granodiorite occurs in both the Haib and Richtersveld areasa.n oGdriorites of this unit from the two areas cannot be distinguished in the REE patteurtn isn bthe patterns for granite, there is a distinction with those in the Richtersveld showlionwg er overall concentrations than 102 those in the Haib area. However, this patternf iosn oly one sample and may not be representative. In Figure 4.12, the variation in REE patterns fvoer rage concentrations is shown for the range from mafic to felsic compositions, in unithsi cwh comprise a compositional series. These units include the Vuurdood Subsuite (ultraicm-gaafbbronorite-gabbro), Khoromus Tonalite (diorite-tonalite-granodiorite), Blockwe Mrfigmatite (diorite-tonalite- granodiorite), Gaarseep Granodiorite in the Haeiba a(rgabbro-diorite-tonalite- granodiorite-granite), Gaarseep Granodiorite in R thicehtersveld (tonalite-granodiorite- granite) and the Hoogoor Granite (tonalite-granroitdeio-granite). Figure 4.12: Variation of REE patterns within tuhneit s of the Vioolsdrif Batholith. Chondrite normlizaing values after Boynton (1984). 103 Figure 4.12: (Continue). For the Vuurdood Subsuite, the patterns of indiavil dsuamples are shown because of the large degree of variation displayed. Patternsv oefr aage concentrations would not be representative, especially of the ultramafic ro c Ikns .the ultramafic samples, patterns show variation in overall concentration, slope aesll was the extent of the Tb anomaly. The patterns for gabbronorite are more consisnte snlto ipe. There is a gradual decrease in the size of the Tb anomaly with increasing ovecroanll centrations. The single gabbro sample shows the highest overall concentrationhs nwoit Tb anomaly. As such, the influence of fractional crystallization can be rgencoized in the range from gabbronorite to gabbro. However, among the ultramafic samplesv athriea bility in the patterns suggest the influence of alternative or additional process. s Tehe suggestion made earlier that some of these samples may represent cumulate mal,a itse rsiupported by this observation, as there is no systematic process involved in othrme aftion of cumulates. Within the Khoromus Tonalite there is a graduarl einacse in overall concentrations. The Eu anomaly in the granodiorite is larger than ien dthiorite and tonalite, however, no clear systematic increase in the size of the anomalyis icse drnable from diorite to tonalite to granodiorite. These observations do not suppmorot dae l of fractional crystallization as a single process in the development of the Khoromounsa Tlite. 104 Patterns for the Blockwerf Migmatite (individualm spales) are characterized by variation in overall concentration, slope and the Eu-anom aOlyn.e granodiorite samples shows a large positive Eu anomaly, which is commonly autttreibd to hornblende fractionation (e.g., Rollinson, 1993). The large degree of vtiaornia in all the aspects of the patterns rules out fractional crystallization as the domitn parnocess in the formation of this unit. A distinction is made between the Gaarseep Granrioted ion the Haib and Richtersveld areas. Within the Gaarseep Granodiorite in theb Haareia, no systematic trend can be observed. The gabbros and diorites show flattienrg- lyslopes than the more felsic rocks, with the gabbros having lower overall concentrast iothnan the diorites. There is a sudden and large increase in slope between the dioritde st oannalites with the latter showing the steepest slope of all the rocks in this unit. Teh ies ra decrease in the slope from tonalite to granodiorite and granite, the latter two patteronisn cciding almost exactly. The Eu- anomaly systematically increases in size from m taof ifcelsic compositions. Most of these observations also speaks against a singlel mofo fdractional crystallization for the development of the Gaarseep Granodiorite in theb Haraeia. The Gaarseep Granodiorite in the Richtersveld dairsepala ys an increase in slope from mafic to felsic compositions, along with a gradiunaclr ease in the size of the Eu anomaly. The increase in slope, however, is not the resf ualt moore rapid increase in LREE concentrations compared to HREE concentrations s. u Acsh, it does not support a model of fractional crystallization for its developmen Tt.h e increase in slope from mafic to felsic compositions is contrary to the trends ien Gthaarseep Granodiorite in the Haib area. But again, the single granite sample froem R tihchtersveld may not be representative. Patterns in the Hoogoor Granite show a systemnactirce ai se in slope as well as the size of the Eu anomaly from mafic to felsic compositionTsh.i s speaks in favour of a fractional crystallization model for the development of theit .u n 105 As such, within the Goodhouse Subsuite, when indduiavli units are considered, the evidence presented suggest that various magmaotcice spsres may have been involved during its evolution. Except for the Hoogoor Grtaen, ithe influence of contamination processes in addition to fractional crystalliza,t iiosn indicated. 4.2 MAGMATIC PROCESSES 4.2.1 Vuurdood Subsuite in relation to the rest of the batholith Due to unclear contact relationships in the fiethlde, exact nature of the Vuurdood Subsuite and its relationship with the rest of bthaeth olith, has been controversial among previous investigators. Possibilities as to ittsu rnea included remnants of the mantle sources, xenoliths of earlier plutonic phases, clautme umaterial in the primary magmas, and later intrusions in the Goodhouse Subsuite p. r eAstent, enough evidence have amassed to rule out the possibility that the Vuoudrd Soubsuite (i.e., the bodies occurring in the Haib area) may represent post-Goodhousues iinotnrs. Radiometric data (Reid, 1977, 1982) show that tuhuer Vdood Subsuite is the oldest unit in the Vioolsdrif Suite and overlaps in age withe tOhrange River Group. Initial isotope data (Reid, 1982) show that the Vuurdood Subsuhiatere s similar parental sources with the Goodhouse Subsuite and Orange River Group.t i -Meluelment diagrams indicate that the Vuurdood Subsuite experienced similar magmparotic esses as the Goodhouse Subsuite and Orange River Group in having simialattre prns (Figures 4.6, 4.7, 4.9). REE patterns support the initial isotope data in shogw thinat the Vuurdood Subsuite share similar sources with the Goodhouse Subsuite, stihnecier slopes are nearly parallel. Overall REE concentrations in gabbros (the onlyk rtoycpe which the two units have in common) also closely coincide (Figure 4.12). Ahlisl tevidence taken into account, it would be realistic to postulate that the Vuurdooudb sSuite may represent remnant bodies of the primary source magmas. 106 The REE patterns also suggest that the procesras cotifo fnal crystallization can be recognized as one of the magmatic processes whaicsh a wctive during the evolution of the Vuurdood Subsuite, but that other factors mhuasvte played a role in producing the observed variations among the trends. Based o lna rtghe negative Tb anomalies and large variation in slopes of some of the ultram asaficmples, these samples may represent cumulates of amphibole and pyroxene. Evidence from Harker diagrams (4.1.3 Harker diagsr)a smuggest a genetic link between the Vuurdood Subsuite and Haib Subgroup which ti so nboserved between the Vuurdood Subsuite and the De Hoop Subgroup, nor the Goodeh Soubssuite. A progressive increase in LILE concentrations from the Vuurdooudb sSuite, through the Haib Subgroup, to the Goodhouse Subsuite in rocks oifla sri mcomposition, is observed on multi-element diagrams (4.1.4 Multi-element diagsra, mFigure 4.9). Increases in LILE concentrations in subduction zone magmatism is conmlym attributed to metasomatic processes in the mantle wedge (e.g., Pearce ankidn sPoanr, 1993). This suggests that the observed systematic increases in LILE concentrsa tmioany be the result of progressing metasomatic processes from the formation of theb HSauibgroup to that of the De Hoop Subgroup and Goodhouse Subsuite, from a commond Vououdr-like source. The Jensen cation plot (Jensen, 1976) is a clcaastsioifni scheme for subalkaline rocks. It has the advantage that it distinguishes tholeaiintidc komatiitic rocks from basalts and calcalkaline rocks. In Figure 4.13, the Viools dBriaftholith is plotted on this diagram. 107 Figure 4.13: The Vioolsdrif Batholith on the Jen sceation plot (Jensen 1976). Most of the Vuurdood Subsuite samples show resenmceb lato high-Mg tholeiite basalt. These are the primary magmas which are producmedo dine rn volcanic arcs from a MORB-like source (e.g., Condie, 2005b). Some oef uthltramafic samples plot in the komatiite field with a space separating them frohme ot ther samples of the subsuite. This plot in the komatiite field is related to elevatMedg contents and may be explained in terms of cumulate material which is rich in pyroex eannd olivine. While the other units of the batholith are clearly parted from the Vuuord oSubsuite on this diagram, the Haib Subgroup represents a continuous range towardVsu tuhred ood Subsuite. Evidence can furthermore be provided that the matigcm parocesses which were active during the formation of the Haib Subgroup, diffe rfreodm those which were active during the formation of the De Hoop Subgroup and Goodh oSuusbesuite. Kostitsyne t al. (2007) used diagrams of compatible vs. incompatible trealecme ents to study the crystallization history of an igneous rock unit. If two igneousc kr ounits originate from the same source, the distribution of any two elements within themll wdei pend on the distribution 108 coefficient of the elements and the magmatic prsoecse sthat were active during their formation. Figure 4.14 shows a plot of Rb vs. Cr (compatibsl.e i nvcompatible element) for rocks of similar composition in the Vioolsdrif Batholith. h Tese include gabbros, diorites and basaltic andesites, which can be shown on bothi -meluelmt ent and REE diagrams to overlap closely in composition. Figures 4.14: Variations of Rb vs. Cr in mafic krso cof the Vioolsdrif Batholith (gabbros, dioritensd a basaltic andesites). The Vuurdood Subsuite and Haib Subgroup are astseodc iina a single linear trend, while the Goodhouse Subsuite defines a linear trend w ish iochriented at an angle to the former. (There are no equivalent rock compositions in thee H Doop Subgroup.) Since these units in Figure 4.14 originated from a similar source i(dR, e1977; 1982), it can be assumed that the magmatic processes which led to the formatfio thne o two trends, differed. As such, 109 the Vuurdood Subsuite and Haib Subgroup are geanlleyt ilcinked, while the Goodhouse Subsuite is not genetically linked to the VuurdoSoudb suite. All classification and variation diagrams show a close association bet wtheee nGoodhouse Subsuite and the De Hoop Subgroup. Therefore, it can be assumed hthea Dt et Hoop Subgroup is also not genetically linked to the Vuurdood Subsuite. The use of compatible vs. incompatible element rdaimags can be extended to include all the rock compositions in the batholith. Kostitseytn a l. (2007) showed that rocks produced by fractional crystallization from a comnm poarent, must display linear trends on logarithmic variation diagrams of compatible ivnsc.ompatible elements. The variation of Rb vs. Cr among the various rock ty apneds within various units of the Vioolsdrif Batholith, is displayed on logarithmicia dgrams in Figure 4.15. Within the Haib Subgroup, there is a transitione orbvesd from the basaltic andesites to the rest of the compositions. The trend of the bacs alntidesites is steep and positive and coincides with that of the Vuurdood Subsuite. r aItd ges into the trend for andesites, which is flat-lying and negative, and continuese alinrly into those for the dacites and rhyolites. Since Rb is a highly mobile elemenet, sthingle values on the diagram which plot below the average trends, are considered toh eb eresult of later metamorphism. None of the Goodhouse Subsuite, nor the De Hoopg rSoubp trends show a linear association with the Vuurdood Subsuite, such ats o tfh tahe Haib basaltic andesites. Compositions in the Goodhouse Subsuite which opv ewrliath the Vuurdood Subsuite coincide with the negative trend defined by thet oref sthe rock types. The trends of individual rock types within the Goodhouse Subs auritee curved, as is the combined trend from mafic to felsic. This curved nature distinsghueis it from the linear trend defined by the Haib Subgroup. The De Hoop Subgroup also edse fain linear trend as does the Haib Subgroup, however, the trend for the De Hoop Suubpg riso slightly positive, which distinguishes it from the slightly negative Haibb Sguroup trend. 110 Figure 4.15: Variations in Cr vs. Rb for the vaursio units of the Vioolsdrif Batholith. The linear trends in the Haib and De Hoop Subgr osupgsgest that fractional crystallization played the dominant role in the edleovpment of the volcanic rocks. The curved trends in the Goodhouse Subsuite, howeuvgegr,e sst major influence from alternative processes in the development of thteo npilcu suite. Through major and trace element modeling, Reid (1977) found that a simrpalec tfional crystallization model could not account for the production of the whole conotiunsu range from diorite through tonalite and granodiorite to granite in the Goodsheo Suubsuite. Reido p( cit., p. 278) 111 proposed that models of contamination and magmain mg imxay explain some of these discrepancies. Von Backström and De Villiers (1,9 p7.2 40), as well as Beukes (1973, p. 68) noticed an influence on the composition of gthraen itoids through processes of contamination. During field work for the currentut dsy, it was often noticed that the granotoids have a more mafic composition in thein vitiyc of mafic and ultramafic bodies. This is attributed to processes of magma mixingce s itnhe contacts were found to be gradational and since the granitoids and the maanfdic ultramafic bodies are so closely related in time. 4.2.2 Vuurdood Subsuite as remnants of the primary magmas The features displayed by the Vuurdood Subsuitme uolnti -element diagrams (Figure 4.6) are typical of subduction zone magmatism (e.g.r, cPee and Parkinson, 1993). In Figure 4.16, the Vuurdood Subsuite is compared to simroilcakr s in modern volcanic arcs. Also shown is the pattern for MORB and the modeled dtedp lemantle. Modern island arc volcanics originate from MORBa. s Bed on the fact that the pattern of the Vuurdood Subsuite closely parallels that of emrond island arc volcanics, it can be assumed that the processes which led to their ftoiornm, awere similar. The concentration of Ba, K, La, Ce and Sr in the Vuurdood Subsuitreth feurmore closely match their concentrations in modern arc volcanics. Accordtoin Pg earce and Parkinson (1993), these are elements which originate in the mantle souurcee t od metasomatic processes. Therefore, it can also be assumed that the souf rtche oVuurdood Subsuite was similar to that of modern arc volcanics. Pearce and Park in(ospo ncit.) furthermore considered the concentrations of Nb, Zr, Ti, Y and the HREE, tpor reesent that in the source before subduction started. Based on these assumptiopnastt,e ar n can be projected for the mantle source from which the Vuurdood Subsuite origina(tinedi cated on Figure 4.16). 112 Figure 4.16: Multi-element diagrams of averageu evsa lfor the Vuurdood Subsuite compared to that oodf emrn volcanic arcs (including the Kermadec, Tonga, Mnaarsia, Izu-Bonin, Kurile, Lesser Antilles, Cascadneds Caentral Volcanic Zone of the Andes arcs) and MORB (valufe Ps eoarce, 1983). Values for the volcanic arcs woebrteained from the internet georoc database (http://georoch.m-mpainz.gwdg.de/georo)c. / Chondrite normalizing values are from Thompson (1982). The modeled depleted mantle in Figure 4.16 reprtse stheant of the mantle residue after continental crust extraction around 2 Ga (Patc h1e9t8t,9). This time coincides with the onset of formation of the Vioolsdrif Batholith. eT hdifference in overall concentrations between the modeled depleted mantle and the perodj escoturce of the Vioolsdrif Batholith, indicates that the Vioolsdrif Batholidthid not originate from the model depleted mantle, but from a mantle which was hi ginh eorverall concentrations. Various previous studies point out the fact that the ma inst nleot continuously being depleted through time (e.g., Polaett al., 2011). This is again shown in Figure 4.16 fotrh b the Vioolsdrif Batholith source as well as MORB. Onae yw in which such elevated concentrations can be achieved, is through thec lriencgy of continental crust into the mantle (e.g., Griffine t al., 2009). This speaks in favour of pre-existinugs ct rin the study 113 area before the evolution of the Vioolsdrif Baththo.l i If this is indeed the case, this pre- existing crust must have been completely recycsliendc,e no trace of it is currently found. However, recycling of pre-existing crust contrasd itchte Sm-Nd isotope evidence presented by Reid (1997). Another way in which e tlheement concentrations in the mantle can be elevated, is through continuous traelifzeartion from the asthenosphere (Griffin et al., 2009). However, since the time of onset ofe tvhoel ution of the Vioolsdrif Batholith coincides with the time the depleted mlea nwtas formed (2 Ga), this mechanism cannot be invoked for the Vioolsdrif Batholith. Nordin’s (2009) model DTM ages suggest that crust formation in the area daylr estarted at 2 300 Ma, which favours the presence of pret-ienxgi scrust. This was supported by the presence of a single inherited zircon of about 02 M70a, indicating the presence of Archean material. CTHUR ages from the same study do not allow time fo rf othrme ation of pre-existing crust, however, some of these age ns eagraetive, which suggest that this model is not applicable. Reid (1997), however,e aegs rwith the latter model, especially based on the initial isotope signatures, which lpurdeec the possibility of incorporation of pre-existing crust. He suggested a juvenile cprurosdt-ucing episode at around 2 000 Ma from a bulk earth mantle, followed by the rapidy rcelicng of young crust that generated granitic and rhyolitic components. The deeper troughs at Nb, P and Zr and the lowleure sv aof Ti, Y and Yb in the Vuurdood pattern compared to that of modern arucgsg, esst either that the processes which create these troughs were more efficienht ein P troterozoic than they are today, or that they were different processes all togethne rm. oI dern arcs, they are associated with metasomatism of the mantle wedge by dehydratiothne o sf ubducting slab and there are currently no alternative models. The mechanism w hbiych the pre-Vioolsdrif mantle became enriched relative to the model depleted lme aisn nt ot resolved here if the presence and complete recycling of pre-existing continenctrauls t is denied. Most of the evidence presented here support this theory. 114 4.2.3 Radiometric evidence reviewed Reid (1977) determined initia87lS r/86Sr ratios for the Vuurdood Subsuite (0.7031), Nous Formation (0.7035), Tsams Formation (0.7028), Khmourso Tonalite (0.7029) and Gaarseep Granodiorite (0.7030). This means thea Vt uthurdood Subsuite, Orange River Group and Goodhouse Subsuite originate from a asri msoilurce. The available radiometric data for the Richters vaerleda is not sufficient to indicate the chronological order for all the units in the Viodorlisf Batholith. In the Haib area, the whole rock Rb-Sr results of Reid (1977) are in ardcaconce with field evidence. It shows the dacite-rhyolite compositions of the Tsams Fotiromna to be 2 020±70 Ma and the basaltic andesite-andesite compositions of the NFourms ation to be 1 970±70 Ma. This means that the felsic volcanics in the Haib Subpg raorue not only older than their mafic counterparts, but that they are indeed the oldfe asltl othe rocks in the batholith, including the Vuurdood Subsuite. This is another piece oidfe envce arguing for the presence of pre-existing crust in the area (which was compyle rtecl ycled) as felsic magma simply cannot be produced as a primary melt from a masnotulerc e (e.g., Leake, 1983). Although the study of Reid (1977) did not recogn tihze current subdivision of the Goodhouse Subsuite, his sample localities and sea dmepslcriptions offer reliable indications as to which of the current units theylo bng to. Gabbros and diorites (chemical classification) from the Khoromus Tonea lriet ndered an age of 1 975 Ma±55 Ma. Of the remaining thirty samples representhineg ttonalites, granodiorites and granites in the Goodhouse Subsuite, 28 are from the Gaa rGseraenpodiorite. The tonalites dated at 1 965±80 Ma, the granodiorites at 1 925±45 Mda tahne adamellites at 1 800±40 Ma. As such, there is a clear relationship betweenc hthroen ological sequence and the compositional range within the Goodhouse Subsun iteh ei Haib area, the units becoming progressively younger from mafic to felsic. Howre, vaes stated before, the study concluded through major and trace element mode tlhlinagt fractional crystallization cannot account for the formation of the entiree sse frriom diorite to adamellite. 115 A single age determination on the Xaminxaip Riverar nGodiorite (U-Pb zircon by LA- ICPMS; Appendix 2) of 1 892±4.8 Ma coincides witimh islar compositions in the Gaarseep Granodiorite. Field evidence indicatte t hthea Gaarseep Granodiorite intrude the Xaminxaip River Granodiorite. An age of ab1o u8t9 0 Ma (SHRIMP U-Pb zircon) is reported by Moen and Toogood (2007) for the Hoo gGoroarnite which is also more or less similar in age to the granodiorites in thes suuitbe. In the Richtersveld, the Paradys River Formatiosn bheaen dated at 1 883±7.4 Ma (U-Pb zircon by LA-ICPMS; Appendix 2). This age is sifgicnai ntly younger than any of the ages obtained by Reid (1977) for the Haib Subgr o Tuhpe. Rosyntjieberg Formation rendered detrital zircon ages which indicate a Vlsidoroif Batholith provenance with one as young as 1 760 Ma (Appendix 2). The latter aaggre es well with the crystallization age of the Ramansdrif Subsuite and no deformaetiloante rd to this age has been identified. Therefore, metamorphic overprint ist an olikely explanation to the 1 760 Ma detrital zircon age. Rather it is thought to rcetf ltehe presence of Ramansdrif-age rocks in the Rosyntjieberg Formation provenance. The fhaactt tthe Rosyntjieberg Formation is intruded by granites of the Gaarseep Granodiosriete (2.3.2.7 Rosyntjieberg Formation), suggests that the Gaarseep Granodiorte in the eRrsicvhetld may be significantly younger than that in the Haib area. However, a U-Pb zi racgoen (LA-ICPMS) of 1896±12 Ma for this granite (Appendix 2) coincides with the agbetsa oined by Reid (1977) for those in the Haib area and also predates most of the detriteasl aing the Rosyntjieberg Formation. As stated before, it is possible that the sample w whiacsh taken for dating was taken from the wrong outcrop. Even though the map suggests tt hisa ftr iom the same pluton that intrudes the quartzites, it is not from the actloucaal tion where the two units are found in contact (Paradyskloof, Richtersveld National Paserke; map). 4.2.4 Unique signatures of the subduction environment In the subduction zone environment, the element,s T Na,b Zr, Hf, Ti, Y and the REE in calcalkaline basalts, are derived almost entirreolym f the mantle wedge while Rb, Ba, Th and K are among those which include a significaontt rcibution from the subducting slab 116 (Pearce and Parkinson, 1993). Therefore, a ruactiho ass Ba/Nb would indicate the involvement of hydrothermal fluid from the subduncgt islab, while a ratio such as La/Nb would indicate the involvement of arc-like process (see.g., Kay, 2005; Kay and Mpodozis, 2002). Furthermore, the influence of stuhbeducting slab can be divided into two processes, i.e., the influence of hydrotherfmluaidl and the influence of melt generated through the partial melting of the stlhaibs (represents the adakitic component). These two processes can be distinguished from oetahcehr on a plot of Ba/Nb (fluid component) vs. Sr/Y (melt component) (e.g., Keatu ar l., 2009). Figure 4.17 shows the Vioolsdrif Suite plotted o dni aagram of Ba/Nb vs. La/Nb (the Orange River Group is not included due to lack ao fa Lnalyses), as well as both the Vioolsdrif Suite and Orange River Group on a diamg roaf Ba/Nb vs. Sr/Y. The positive linear trend on both diagrams suggest that thee tphroecesses envisaged to be operating in subduction zones (fluid-induced metasomatismlt,i nmge of a subducting slab and processes leading to the production of a “subdnu cctoiomponent” on multi-element diagrams), are applicable to the Vioolsdrif Batthho. li The fact that values for the various units overlap suggests that these processes wteivre daucring the entire evolutionary history of the batholith and that they remaineds ctaont throughout its evolutionary history. 117 (a) (b) Figure 4.17: Plots of (a) Ba/Nb vs. La/Nb (basne dK oay, 2005 and Kay and Mpdozis, 2002) for the Vlsiodorif Suite, and (b) Ba/Nb vs. Sr/Y (based on Kaeut ra l., 2009) for both the Vioolsdrif Suite and Orangivee Rr Group. 118 The temperature in the mantle wedge in a subdu cztoionne tectonic setting is most likely controlled by heat exchange with the cooler subindgu cstlab (Pearce and Parkinson, 1993). Temperature control in the source regiolnl d weitermine which mineral assemblages will fuse, as such controlling ther ibdiusttion of specific elements between the source and the magma. The elements Nb anrde Zarm aong those which Pearce and Parkinson o(p cit.) considered to originate almost entirely from mthaen tle wedge in the subduction zone environment. They are also bortyh hvieghly incompatible elements and their fractionation from each other is dependan lto own degrees of melting since, with higher degrees of melting, the distinction betwteheenm is progressively smoothed out (Pearce and Parkinsono p, cit.). In Figure 4.18, these two elements are ploattgeadi nst each other for various units in the Vioolsdrif Boaltithh. Figure 4.18: Zr-Nb variations in the various u noift sthe Vioolsdrif Batholith. 119 Variation within the Haib Subgroup follows a povseit ilinear trend while, in the rest of the batholith, there is no systematic relations hAips. such, these two elements behaved as a unity during the development of the Haib Subgr, obuupt became decoupled from each other during the rest of the batholith’s developmt. e Cnoncentrations for the rest of the batholith cluster around the maximum values attda ine the Haib Subgroup. 4.2.5 Origin of the Blockwerf Migmatite The two phases within the Blockwerf Migmatite, daesn itified in the field (2.4.2.3 Blockwerf Migmatite), is also clearly distinguish eond the total alkali-silica classification diagram (Figure 4.19). Phase 1 (high concentr aotfio mnafic enclaves and well developed migmatitic character) is on average mmoarfeic than phase 2 (low concentration of mafic enclaves and granitic teex)t.u rPhase 2 is also compositionally similar to the granitoids of the Gaarseep Granoitdei owrhich intrudes it. Figure 4.19: Plots of the Blockwerf Migmatite, Grsaeaep Granodiorite which intrudes it and Kuams Rr Fivoermation volcanics which are intruded by it, on a TAS clfaicsastiion diagram. 120 REE patterns for the Blockwerf Migmatite have bedeisnc ussed earlier (4.1.5 REE patterns; Figure 4.12). The large variations einir t hoverall concentrations, slopes and Eu- anomalies rule out fractional crystallization aes dthominant magmatic process in the development of the unit. Figure 4.20 shows the RpEatEterns for the Gaarseep Granodiorite and Kuams River Formation volcanicsic wh hare found in contact with the Blockwerf Migmatite. (a) (b) Figure 4.20: Chondrite-normalized REE patterns t hfoer Gaarseep Granodiorite (a) and Kuams River aFtoiormn (b) which are associated with the Blockwerf Migmat i(teG.r een = granodiorite/dacite, blue = andesite) The Gaarseep Granodiorite rocks which intrude tlhoec kBwerf Migmatite have the same composition (granodiorite), however, show rathergr ela variations in concentrations, slope and the size of the Eu anomaly. One of tshaemseples also shows an elevated La/Yb ratio (Figure 4.20a), the single most impnotr tfaeature which distinguishes the Blockwerf Migmatite geochemically. The volcanicfs t hoe Kuams River Formation show a systematic increase in slope in the range frodme sainte to dacite, suggesting a fractional crystallization model for their development. Multi-element diagrams in which the REE are notr erespented, do not distinguish the Blockwerf Migmatite from the rest of the bathol(it4h. 1.4 Multi-element diagrams, Figure 121 4.7). The fact that immobile trace elements susc Bh aa, Nb, Ti, Zr and Hf do not distinguish the Blockwerf Migmatite migmatitic rosc kfrom the rest of the batholith, illustrates that the processes leading to theire ldoepvment were similar, i.e., subduction- related. The steep slopes (high La/Yb ratiosg),e l avrariations in overall concentrations and the presence of positive Eu anomalies in thoec kBwlerf Migmatite REE patterns, distinguish them from the other batholith rocksh. e Tse features suggest that the Blockwerf Migmatite owes its characteristic natutor ep rimary differences in the source (which is hornblende-dominated), and not to latearg matic processes which acted during its evolution. It can therefore be concldu dtheat the mafic enclaves to which the Blockwerf Migmatite owes its characteristic natu rreep,resents xenoliths of its source, and not of the volcanics which it intrudes. Thailso,n g with field evidence indicating the presence of hybrid rocks, suggests a magma mixiondge ml for the origin of the Blockwerf Migmatite. Note that many of the rock types in the Blockwerigf mMatite are adakitic in nature, bearing evidence as to the origin of adakite anGd TinT the Vioolsdrif Batholith (see 4.3.4 TTG and adakite in the Richtersveld Subpnrcoev)i. The single most distinguishing feature of the Blockwerf Migmatite is the elevatLead/Y b ratios. According to Haschke et al. (2006), REE fractionation is an indicator ash teo dt epth of magma generation. The La/Yb, La/Sm as well as Sm/Yb ratios relate to mthaen tle domain in which melting occurs with the following serving as a rough gu ide: a) depths of 30-35 km; source dominated by gab bcroomicpositions (clinopyroxene- bearing); La/Yb<20, La/Sm<4, Sm/Yb<3, b) depths of approximately 40 km; amphibolitic scoeu; rLa/Yb>20-30, La/Sm>4, Sm/Yb=3-5 (indicate MREE retention by amphibole), c) depths>45-50 km; eclogite (garnet-bearing) seo;u Lrca/Yb>30, Sm/Yb>5 (indicate HREE retention by garnet). In Figure 4.21, values of the La/Yb ratio are counretod across the batholith and a cross section is constructed to show the anomalous cl rtuhsictakness underlying the Blockwerf 122 Migmatite during its development, relative to thees tr of the batholith. The crustal thickness during the development of the VioolsBdraift holith is also compared with that of higher La/Yb ratios. The typical modern Andeaarcn has crustal thicknesses within the eclogite zone (up to 70 km crustal thickness)i.s tIht erefore clear that the Vioolsdrif Batholith developed on and in a relatively thins ct raubove the subduction zone. Figure 4.21(a): Distribution of La/Yb values ino driites, tonalites, granodiorites, andesites andit edsa accross the Vioolsdrif Batholith. To minimize the effects orfa fctional crystallization, samples of similar comsiption were used (tonalite-granodiorite for the Vioolsdrif Suite a nddacites for the Orange River Group). An interptiolna method of kriging was used as executed by the ArcGIS soft wcoamreputer program. (b): A cross section to show inferred crustal kthniecsses. The Blockwerf Migmatite represents a more or leirscsu clar REE fractionation anomaly in the batholith. This suggests the Blockwerf Maigtmite as being representative of a 123 volcanic centre and supports the theory of Rit1te9r8 (0), envisaging a volcanic centre in the northeast of the Richtersveld area. The thniicnkge of the crust underneath the unit may be explained by a process of mafic underpl.a ting If the Blockwerf Migmatite does represent a volca cneintre, any volcanics which originated from it will also bear the high La/Ybt iroa signature. Figure 4.22 shows a plot of the La/Yb ratios vs. La for migmatites from tuhnei t, the Gaarseep granodiorites associated with it, and all the Orange River Grsoaump ples for which REE analyses are available. Figure 4.22: La vs. La/Yb for the Blockwerf Migmitaet, the Gaarseep Granodiorite which intrudes dit alnl the samples from the De Hoop Subgroup for which REEly asneas are available. None of the samples from the Orange River Grourp w, fhoich REE analyses are available, match the high La/Yb values displaye dth bey Blockwerf Migmatite. This means that, if the Blockwerf Migmatite does inder epdr esent a volcanic centre, it was 124 not the source for any of the volcanic formatiounrsr ecntly preserved within the batholith. However, the current REE analyses which are avlaei lfaobr the Orange River Group are not sufficient to be considered representativeh eo fg troup as a whole. Also apparent from Figure 4.22, is the fact that the granodiosr iftreom the Gaarseep Granodiorite, although not included in the Blockwerf Migmatited a cnlearly intrusive into it, display some of the highest La/Yb ratios. This is to bpee ecxted if the theory of mafic underplating is accepted and they were generatoevde a tbhe same area. 4.2.6 Ramansdrif Subsuite Geochemically, the granites of the Ramansdrif Siuteb scuorrelate closely with the granites of the Goodhouse Subsuite, as well arsh tyhoeli tes of the Orange River Group, on Harker diagrams for all elements analysed ( eF.igu.,re 4.4) as well as multi-element diagrams (Figure 4.7). However, REE patterns sthhoew R amansdrif Subsuite granites to be lower in overall concentrations than the grasn oitfe the Goodhouse Subsuite, as well as its slope being flatter (Figure 4.11). This predcelsu the possibility that the Ramansdrif Subsuite represents the final stage of the Viooif lsSdurite in a continuous fractional crystallization process. The Ramansdrif Subsuite postdates the Goodhousseu iSteu by approximately 170 m.y. and initial 87Sr/86Sr ratios for the Ramansdrif Subsuite (Reid, 19s7u7g)g ests an origin through melting of pre-existing crust. The closoerr eclation of the Ramansdrif Subsuite granites with the granites of the Goodhouse Sueb souni tvariation diagrams as mentioned above, suggests that the Ramansdrif granites choauvled been derived through partial melting of the granodiorites and tonalites of thoeo Gdhouse Subsuite. This was also suggested by Reid (1977) after major and trace eenlet modelling. Cornelel t al. (2007) mentioned that the Ramansdrif Subsuite might alcyt unaolt be genetically linked to the Goodhouse Subsuite but rather represents a mag emvaetnict of its own, limited in extent and producing only felsic magmatism. 125 4.3 THE ORIGINAL TECTONIC ENVIRONMENT The original extent of the Vioolsdrif Batholith niso t known and subsequent deformation processes further obliterated much of the physeivciadle nce related to the original tectonic environment. In modern convergent plate margihnes ,m tost distinctive feature is the magmatic arc – a linearly orientated feature, plealr taol the subduction trench and associated with eruptive volcanic centers. Othear accteristic features include turbiditic trench deposits, accretionary wedge deposits aangdn doistic fore- and backarc assemblages. In the Vioolsdrif Batholith, mosth oefs e features, which are located at surface, have been eroded. The only significadnitm seentary unit is the Rosyntjieberg Formation, which was interpreted by Ritter (198s0 )a a marine deposit, indicating a marine ingression towards the end of the volcayncicle c. He also inferred the presence of a volcanic centre in the northwest of the Richtelrds varea, based on the distribution of the volcanic deposits. This is supported by threre cnut study in the recognition of the Blockwerf Migmatite as a possible volcanic cent4r.e2 .(7 Blockwerf Migmatite). However, the current configuration of these feast uinre relation to their original settings is difficult, if not impossible, to assess, due to sseuqbuent deformation. In Figure 4.23, the Vioolsdrif Batholith is geomiceatrlly compared to Mesozoic-Tertiary batholiths in the Andean and North American arcs. Based on a comparison with the Mesozoic-Tertiarthy oblaiths, the preserved part of the Vioolsdrif Batholith represents but a fraction hoef toriginal one. Individual plutons in the Mesozoic-Tertiary batholiths have various otraietinons but collectively and on a regional scale, the batholiths’ axes are paraoll ethl et subduction trench. It is not possible to determine the arc axis of the Vioolsdrif Baththo loi n physical evidence. 126 Figure 4.23: Comparison of scale between the Vsidoroifl Batholith and its Mesozoic-Tertiary counterrtpsa. 127 Radiometric data indicate the Mesozoic-Tertiaryh boali ths to have developed in a time span of up to 200 m.y. (e.g. the Coast Range bitaht hoof lChile; Paradae t al., 1999). Some developed in as little as about 50 m.y. (the.eg .Id aho batholith; Kinge t al., 2007). The preserved Vioolsdrif Batholith retains evide nocf eabout 270 Ma if the Ramansdrif Subsuite is included. Note however, that the tsimpaen with which the Ramansdrif Subsuite postdates the main intrusion (Goodhoubsesu Situe; 170 m.y.), is sufficient for an entire batholith to form. Still however, the prrevseed part of the Vioolsdrif Batholith retains evidence on a major span of its evolutiyo nhaisrtory. There are currently no alternative theories to othf amt odern-day subduction processes that can explain the tectono-magmatic evolutionth oef Vioolsdrif Batholith. Referring to the modern configuration Poleat al. (2011) state that, fundamentally, continentasl tc ru cannot grow, supercontinents cannot form, and caarncsn ot generate melts over millions of years without subduction, collision, subductoiof nw ater, and slab dehydration or melting. Currently, convergent plate margins ahree otnly tectonic settings where magmatic processes produce granitic bodies of blitahtihco extent and widespread felsic and andesitic magmas can only be produced in ea tpelactonic regime (Condie, 2005b). Based on various studies (e.g., Patchett, 1989;i lHtoanm, 1998; Condie, 2005a; Castillo, 2006), it seems likely that the tectonic crust-fionrgm processes which operated on the earth during geological time, underwent dramatiacn cghes during the Late Archean-Early Proterozoic. The formation of the Vioolsdrif Baltihtho coincides with the late stages of this period. Studies show that about half of tuhrer ecnt crust on the earth was produced during the Archean and that it had a TTG compons i(tPioatchett, 1989; Condie, 2005a; Castillo, 2006). Modern crust-forming processeosd purce precious little crust of that composition and also much less volumes of crusctc. o Arding to Hamilton (1998), evidence of modern plate tectonic processes, ssu cohb vaious continental truncations, rifted margin assemblages and sutures, are onolyg nreizced from about 2.0 Ga. Frost et al. (2001) introduced the modified alkali-lime ind(eMxA LI) as Na2O+K2O-CaO and delineated the fields occupied by the most conm gmranite types of the world. In 128 Figure 4.24, the Vioolsdrif Batholith units are tptelod on this diagram showing the fields for Mesozoic Cordilleran batholiths, Archean totnica lgi neisses and peraluminous leucogranites Figure 4.24: Units of the Vioolsdrif Batholith pttleod on the SiO2 vs. MALI diagram of Froset t al. (2001) with the fields for Mesozoic Cordilleran batholiths (brokleinne ), Archean tonalitic gneisses (dotted surfaacned) peraluminous leucogranites (grated surface) indicated. The Vioolsdrif Batholith coincides well with thee flid defined for Mesozoic Cordilleran batholiths. The Ramansdrif Subsuite plots in bthoeth f ield defined for Mesozoic Cordilleran batholiths and peraluminous leucogreasn. it The amount of iron enrichment relative to magnes diuemfines two distinct rock suites and Miyashiro (1974) showed that these two suites c obuel d istinguished on a plot of FeO/(FeO+MgO) vs. Si2O. The former ratio is commonly referred to as Fthee-n umber and numerous workers have shown that this plote o-nf uFmber vs. SiO2 distinguishes between granites from two different tectonic segtsti,n i.e., anorogenic and arc-related 129 (Petroe t al., 1979; Anderson, 1983; Maniar and Piccoli, 19F8r9o;s t and Lindsley, 1991; Frost and Frost 1997). Froestt al. (2001) defined these fields and added those for Archean TTG and peraluminous gneisses, on a FeFOe(OT)(/T)+MgO vs. SiO2 diagram. In Figure 4.25, the Vioolsdrif Batholith units aprleo tted on this diagram. Figure 4.25: The Vioolsdrif Batholith plotted ohne t SiO2 vs. Fe-number diagram of Froestt al. (2001). Again, the Vioolsdrif Batholith shows a close aiftfyin to the field for Cordilleran batholiths. The Ramansdrif Subsuite is not distltyin acssociated with any of the fields. There is a rather wide variety of discriminationa gdriams available on which data is compared with the fields defined by modern tect oenivcironments. These diagrams seldom provide unequivocal confirmation of a for mteecrtonic environment and at best they can be used to suggest an affiliation fori tae saus a whole, and not for single samples (Rollinson, 1993). For the current stuthdey ,V ioolsdrif Batholith has been plotted on a variety of discrimination diagramsl uindcing Ti-Zr-Y, Ti vs. Zr (normal 130 scale), Ti vs. Zr (log scale) and Ti-Zr-Sr of Peea arcnd Cann (1973), Zr/Y vs. Zr of Pearce and Norry (1979), Cr vs. Y of Pearce (19 Z8r2/Y), vs. Zr of Pearce (1983), T2iO- MnO-P2O5 of Mullen (1983) for basaltic compositions, and vNsb. Y and Rb vs. Y+Nb of Pearcee t al. (1984) for granitic compositions. All these driamg s identify a volcanic arc setting for the whole of the batholith. Figure 64 .s2hows only two of these discrimination diagrams as examples for basaltic and granitic coosmitipons in the batholith. Figure 4.26: The units of the Vioolsdrif Batho lipthlotted on the Ti-Zr-Sr discrimination diagramP oef arce and Cann (1973) for basaltic rocks and the Rb-Y+Nb discriamtionn diagram of Pearceet al. (1984) for granitic rocks. 131 The Vuurdood Subsuite is identified as originatiin ga n island arc environment, confirming its genetic relationship with the resf t hoe suite (e.g., Figure 4.26). This island arc origin for the Vuurdood Subsuite is ciromnef d and better illustrated by the log- transformed immobile trace element diagram of Agarla ewt al. (2008), which distinguishes it more clearly from MORB and bas oarltisginating in other tectonic settings (Figure 4.27). Figure 4.27: The Vuurdood Subsuite on a DF1-DFg2-t rloansformed immobile trace element tectonic dimisicnration diagram of Agrawael t al. (2008). [DF1 = 0.3518 * Log(La/Th) + 0.6013 * gL(oSm/Th) - 1.3450 * Log(Yb/Th) + 2.1056 * Log(Nb/Th) - 5.4763; and DF2 = -0.3050 o*g L(La/Th) - 1.1801 * Log(Sm/Th) + 1.6189 * Log(Ybh/)T + 1.2260 * Log(Nb/Th) - 0.9944]. Ross and Bédard (2009) also defined fields fore tihitoicl and calcalkaline basalts with the transitional field between them, on a diagram o vf sZ.r Y. In Figure 4.28, the Vuurdood Subsuite is plotted, coinciding with the transitaiol nfield. 132 Figure 4.28: The Vuurdood Subsuite on a Y-Zr dimisicnration diagram (after Ross and Bédard, 2009). Figure 4.29: Plots of the various units of the oVlsiodrif Batholith on a Nb vs. Rb/Zr diagram indicnagt arc maturity after Browne t al. (1984). 133 Miyashiro (1974) showed that most arc systems haanv oel der outer, and younger inner arc which is coupled with an increase in arc matyt uwrith the gradual development of overlying continental crust. According to Browetn a l. (1984), the maturity of a volcanic arc may be exhibited in the Nb contents and Rba/Ztior .r Figure 4.29 shows a plot of the various units of the Vioolsdrif Batholith on a Nbs. vRb/Zr diagram. The Vuurdood Subsuite plots mostly in the field pforirmitive island and continental arcs. The Haib Subgroup shows a transition from thisd f ienlto the field for normal continental arcs while the De Hoop Subgroup, along with mo stht eo fGoodhouse Subsuite, plots entirely in the latter field. This is in agreem ewnith earlier evidence showing a gradational genetic relationship between the Vuoudrd Soubsuite and Haib Subgroup. It suggests a gradual development of the batholitmh far oprimitive island arc to a normal continental arc starting with the Vuurdood Subs, uthiterough the Haib Subgroup to the rest of the batholith. The largest part of theh boali th formed in the latter environment. Certain lines of evidence strongly argue againes tc tohmparison of the Vioolsdrif Batholith with modern tectonic environments. Mond erquivalents (e.g., the Andean arc), demand the presence of a thick overlying incoental crust. This is the only way in which the average felsic composition and calcanlkea cliharacter can be accounted for in terms of modern theories. A survey of modern suctbiodnu arcs (Miyashiro, 1974) shows that the ratio of tholeiite:calcalkaline rocks i nb atholith relates to crustal thickness with 0% calcalkaline rocks indicating crustal thickness osfe between 12-15 km, 10% of 15 km, 50% of 18-30 km and 60-95% of 30-70 km. The abes eonf cbasaltic volcanics and a basement to the Orange River Group in the Vioofl sBdartiholith may be explained in terms of deep erosion. However, the prevailings tcarlu thickness during its development may be inferred from the La/Yb ratios (Haschekt ea l., 2006). This was considered during the investigation of the Blockwerf Migma t(it4e.2.5 Blockwerf Migmatite) and Figure 4.21 shows that only a small part of thnist umiatches such crustal thicknesses as is observed for modern arcs. In the Andean arc, aclr uthsitcknesses vary between 30-70 km, the thickest being in the Central Volcanic Z.o nTehe volcanic rocks also show abundant evidence of crustal contamination. Nod eenvcie of crustal contamination can 134 be found in the Orange River Group with all elemse cnotrrelating well with the concentrations in their intrusive counterparts.e rTehfore it can be inferred that the major part of the batholith developed in a tectonic tneerr aof which the crustal thickness cannot be compared to modern equivalents. If the La/Yb ratios are considered in relation gtoe ,a no indication can be found of a gradually thickening overlying crust. La/Yb rat iions the Orange River Group (the oldest among the units) are between 15-20, while thaht ein y toungest granites (from the Gaarseep in the Richtersveld area, intruding thsey nRtojieberg Formation) resembles that of the oldest units. La/Yb ratios in the KhoromTuosn alite (the oldest unit in the Goodhouse Subsuite) are more or less similar tto o tfh tahe youngest granites in the Gaarseep (Richtersveld area). 4.4 TTG AND ADAKITE IN THE RICHTERSVELD SUBPROVINCE The recognition (or non-recognition) of a tonaltirtoen- dhjemite-granodiorite (TTG) rock association and adakites in granitic batholiths b heacsome an important topic especially since the work of Barker (1979). TTG suites domtein tahe earth’s Archean crust and their geochemical similarities to modern adakitaes bheen taken by many as an indication that similar tectonomagmatic processoe tsh et modern ones have been active throughout earth’s history (Drummond and Defant9, 01;9 Martin, 1999; Martine t al., 2005; Van Tonder and Mouri, 2010). According ton Cdioe (2005a), the fact that similar types of granitoids are recorded from the eartliemset s, suggest that similar processes of their origin to the present ones were already otipoenral on the early earth (the oldest accretionary orogens recorded date back to 4.03 .a9n Gd a namely the Acasta and Amitsoq TTG gneisses in northwest Canada and Garenedn).l As such, Archean processes have been studied at the hand of thtoinsge oanc the earth today. Modern adakite was first identified by Kay (197r8o)m f the Adak Island in the western section of the Aleutian arc volcanic chain and ttehrem was defined by Defant and Drummond (1990) who showed that their geochemihcarl accter suggested an origin by 135 partial melting of hydrated mafic source rocks caoinnitng garnet. Many observations favour the derivation of adakitic magmas from paal rmti elting of the subducting oceanic slab in volcanic arc tectonic settings. Howevelthr,o augh the original definition of adakite referred specifically to rock generatedm f rtohe melting of slab basalt, later studies have shown that adakite magma may be perodd inu ctectonic settings which are unrelated to subduction in both arc and non-artcin sgest (Castillo, 2006). Furthermore, while Archean TTG is compositionally similar to meordn adakite, it is not so to modern TTG and according to Condie (2005a), studies hahvoew ns that TTG and adakite are not the same and have different origins. Nowhere on earth today are adakitic magmas beiondgu pcred in volumes comparable to those in the Archean. For the generation of saurcghe l volumes of TTG as found in the Archean crust, slab melting is not the most efvfeec mti echanism, unless a higher geothermal gradient acted in the Archean (Cas t2il0lo0,6). In the Archean, higher mantle geotherms may have resulted in subducting slabcsh irnega partial melting temperatures at shallower dephts before dehydration rendered tahbe inslfusible (Richard and Kerrich, 2007). Alternative models for the generation oaf kaitdic magma include fractional crystallization of basaltic melts (Arth and Hans 1o9n7, 5), partial melting of mantle rocks (Moorbath, 1975) and partial melting of pre-exigst itnonalites (Johnson and Wyllie, 1988). Kaye t al. (1999) proposed a model of crustal thickenintgh ein Central Volcanic Zone of the Andes and melting of the lower cru set xtpolain the origin of TTG magmas in that area. Richard and Kerrich (2007) state theay ta kdakitic geochemical signatures, such as low Y and Yb concentrations and high Snr/dY Laa/Yb ratios, can be generated in normal asthenosphere-derived tholeiitic to calclainlkea arc magmas by common upper plate crustal interaction and crystal fractiona tpioroncesses and do not require slab melting. Common upper plate magmatic processehs assu cmelting-assimilation-storage- homogenization (MASH) and assimilation-fractionaryl-sctallization (AFC) affecting normal arc magmas can be demonstrated to explea idni sthtinctive compositions of most adakite-like arc rocks. 136 The Archean crust (presently representing approtxeilmy a15% of the total continental crust) is dominated by tonalites and trondhjem(iltoews -K granitoids) which are characterized by sharply stoping REE patterns aintidnigc the presence of garnet in the source residual phase (Sobolev, 1991). Duringla tthee A rchean, TTG compositions started to change (Rollinson, 2005) and by abo5u Gt 2a., K-rich granites and granodiorites started to appear and the crust be cmaaminly granodioritic (Sobolev, 1991). This, according to Sm/Nd isotope evidenccoein, cided with a major period of continental crust formation (about 60% of the pnret sbeulk). During post-Proterozoic times, the earth’s granitoids have become morei ca acnidd richer in K (representing approximately 25% of the current total continenctrauls t). The genesis of the Vioolsdrif Batholith therefore coincides with the era whichllo fwoed relatively shortly after the change of the earth’s crust bulk composition froTmG T-dominated towards K-rich, granodioritic-dominated. On the modern earth, aclakalcline magmatism such as that represented by the Vioolsdrif Batholith, is typi coaf lcontinental margin subduction tectonism of which the Andean chain representsty tphe area. Adakite magmatism is currently volumetrically minor compared to normaral ngitoids (as opposed to the Archean), not only worldwide but also in the loctiaelsi where they occur (Castillo, 2006). Archean TTG are typified by high N2Oa (>4%) and low K2O (<3%). The CaO of Archean TGG is also lower than Cordilleran TTG whh cicauses their modified alkali lime index (MALI; Frost et al., 2001) to be higher. As such, the2 ON avs. K2O and K2O vs. MALI classification diagrams of Frosett al. (op cit.) can distinguish Archean TTG suites and Mesozoic Cordilleran batholiths. In Figure3s0 4 a.nd 4.31, the Goodhouse Subsuite and Orange River Group are plotted on these diasg.r am 137 Figure 4.30: A N2aO vs. K2O plot for the Orange River Group and Goodhouses uSiuteb after Froset t al. (2001) . Figure 4.31: The Goodhouse Subsuite and Orangeer RGirvoup on a K2O vs. MALI diagram after Frosett al. (2001). 138 On both diagrams, the Vioolsdrif Batholith rockso wsh a closer affinity with Mesozoic Cordilleran batholiths than with Archean TTG. N ohtoewever that the Vioolsdrif Batholith does not coincide entirely with the fi edledfined by the Cordilleran batholiths on either of the two diagrams. Especially the2O N acontent of the Vioolsdrif Batholith is on average lower than that of the Cordilleran blaitthhso. In recent years, TTG suites have been further svuidbeddi following the recognition of sanukitoid (a high-Mg, -Ni and -Cr type TTG; Smeitsh iand Champion, 2000) as well as the Closepet-type granite, the latter differingm fr osanukitoid in having higher 2KO:Na2O ratios and being relatively enriched in Ti, Nb aZnrd ( Van Tonder and Mouri, 2010). As such, the MgO, Mg#, Cr, Ni and2 OK contents distinguish sanukitoid and Closepet- type granite from TTG (Martine t al., 2005). Sanukitoids are also strongly enrichne LdR iEE, Ba, Sr and P (Rollinson, 2005). Recognizing TTG and adakite in a granitic batho mlitahy be a matter of chance since, in the field, their physical appearance much resemnobrlme al granitoids. Their distinction is entirely based on their unique geochemical proepse.r t iThe sampling distribution on the Vioolsdrif Batholith for the current study is thohutg to include all the compositional variations. Unfortunately, the lack of REE anasly sfoer the Haib Subgroup represents a significant limitation to the study since the La/ Yrabtio is a critical factor in the recognition of adakitic rocks. This ratio bears t hoen single most distinguishing characteristic of adakitic rocks namely a deple itnio Hn REE, indicating the involvement of garnet and/or amphibole in the source. Highc oSnrt ents and Sr/Y ratios are the second prerequisite for distinguishing adakitic magmase, ftohrmer indicating the presence of plagioclase in the melt and/or its absence in othuerc se, the latter again referring to the presence of garnet and/or amphibole in the souinrce sY behaves similar to the REE in its compatibility. The behavior of plagioclase m aalsyo be displayed in the absence of an Eu anomaly. Early Archean TTG typically do not wsh ao negative Eu anomaly in their REE patterns while post-2.5 Ga TTG do (Condie, 2a0).0 5The behavior of the HREE, Y and Sr therefore represent the most unique charirsatcictse distinguishing adakitic magmas from normal granitoids and the two almost univelyrs uasl ed diagnostic tools to identify 139 adakite are the Sr/Y vs. Y and La/Yb vs. La diagsr a(Cmastillo, 2006). Most adakitic rocks also have relatively high 2AOl3 contents, indicating high-pressure melting of garnet-containing source rocks (eclogite) or amoplhitieb. The presence of a primitive component (high MgO a, nNdi Cr) in sanukitoids indicates the involvement of a peridotitic mantle in the greantieon of the melt. Since the absence of such features suggests that the peridotitic lme awnetdge was not involved in the generation of the magmas, the conclusion may bwen d trhaat normal TTG magmas originate from melting in the lower crust above mthaentle wedge, while sanukitoid magmas originate from melting in the subductingb sbleaneath the mantle wedge. This model demands the presence of a thick continernutastl cfor the generation of normal TTG magmas (high La/Yb ratios indicates meltingth ien amphibole-garnet stability parts of the mantle). Characteristics such as depleinti oNnb , Ta and Ti are also very common to adakites worldwide, however, in the light ofd sietus showing that adakites are not limited to subduction zone settings, these canen octo bnsidered prerequisites for identifying adakites in a particular suite. Aftceorn sideration of parameters set in Defant and Kepezhinskas (2001), Condie (2005a), Rollin(s2o0n0 5) and Castillo (2006), the following criteria were selected for recognition a odfakitic and TTG rocks in the Vioolsdrif Batholith: a) SiO2 > 56 wt% b) Al2O3 > 15 wt% at 70% SiO2 c) Sr > 300 ppm d) Y < 20 ppm e) Sr/Y > 20 f) Yb < 1.9 ppm g) La/Yb > 20 These parameters require derivation by hydrousia pl amrtelting of mafic rocks (garnetiferous or amphibolite) at high pressureh.e TLa/Yb and Sr/Y ratios considered to typify adakite and TTG differ significantly in lirtaeture between >40 and >20. A ratio of 140 >40 applies strictly to adakite and early ArcheaTnG T, while that of >20 also takes into consideration post-2.5 Ga and later calcalkalinitee s,u which also have overall lower Al2O3 values. Condie (2005a) states that Archean hilg ThT-AG and adakites both have steep REE patterns while most post-Archean TTGn bge tlo the calcalkaline suite and often show a fractional crystallization sequencte nedxing from diorite (or gabbro) to granite. These rocks typically have very low La /rYabtios and exhibit a large range of Yb values. These observations are applicable to itohoel sVdrif Batholith. From an inspection of the analyses for the Viooifl sBdartholith, a number of samples can be identified as adakitic. They are listed in Tea 4b.l Table 4: A list of samples from the Vioolsdrif Bhaotlith which have adakitic compositions. Nr. Unit Rock SiO2 Al2O3 Sr Y Yb Sr/Y La/Yb Al2O3 type (wt%) (wt%) (ppm) (ppm) (ppm) at SiO2 = 70% CHM95 Blockwerf Migmatitic 64.89 15.38 468 16 1.6 29.25 32.86 13.12 Migmatite tonalite CHM100 Blockwerf Migmatitic 62.61 17.77 446 8 0.8 55.75 56.43 14.19 Migmatite tonalite CHM101 Gaarseep Granodiorite 67.50 15.69 309 12 1.2 25.75 78.4 51 4.4 Granodiorite (intruding the Blockwerf Migmatite) CHM166 Gaarseep Tonalite 64.30 16.34 556 16 1.6 34.75 26.89 13.75 Granodiorite CHM56 Xaminxaip Granodiorite 69.75 14.96 387 12 1.5 32.25 23.11 841 4. River Granodiorite CHM76 Xaminxaip Granodiorite 70.49 14.44 338 13 1.5 26 22.31 14.68 River granodiorite CHM192 Gaarseep Granodiorite 70.42 14.68 402 12 1.2 33.5 30.32 91 4.8 Granodiorite CHM78 Gaarseep Granodiorite 69.86 14.57 349 10 1.7 34.9 24.21 01 4.5 Granodiorite CHM98 Blockwerf Granodiorite 63.87 16.05 535 20 1.5 26.75 83.09 291 3. Migmatite In accordance with the batholith’s calcalkalineu nrea,t Al2O3 contents are relatively low as are the La/Yb ratios. A number of samples w hcliochsely comply to the set parameters were excluded from this list (some only becauseir tahdeakitic nature could not be confirmed by available REE analyses) but the sasm ipnl eTable 4 are sufficient to indicate the overall representation of adakitick r otycpes in the Vioolsdrif Batholith as a 141 whole. Volume-wise they represent but a smallt iforanc of the batholith. They do not represent a single magmatic event and do not snhtoruws ive contact relationships with their host rocks. Rather, they represent compoonsaitl ivarieties within the calcalkaline granitoids. The Blockwerf Migmatite is clearly tdinisguished as a unique unit, containing a high concentration of adakitic rocks. Howeveort, enven this unit is entirely adakitic in composition. The only samples from the batholift hw hoich the La/Yb ratios comply to the >40 value characteristic of type-adakites,f raorme this unit. When the adakitic rocks of the Vioolsdrif Batho laithre plotted on the L(na) vs. La/Yb(n) and Y vs. Sr/Y diagrams (Castillo, 2006), only wa f(econsistently from the Blockwerf Migmatite) comply with type examples (Figure 4.3 2T)h.ese figures primarily indicate the absence of garnet in the source for most sasm, hpolewever, an amphibolitic source may be deduced. The relatively low Sr/Y ratiosth fuermore indicate the influence of plagioclase fractionation during the batholith’so leuvtion, leading to low Sr values, and is again related to its calcalkaline character. (a) (b) Figure 4.32: Adakitic rocks of the Vioolsdrif Baotlhith plotted on the Y(bn) vs. La/Yb(n) (a) and Y vs. Sr/Y (b) diagrams of Castillo (2006). While the composition of adakites does correspon tdh att of Archean TTG, they do not correspond to modern TTG. Post-2.5 Ga TGG havhee hr iKg2O:Na2O ratios than 142 Archean TTG and the two can be distinguished olno ta o pf K2O vs. Na2O (Figure 4.30). According to Castillo (2006), they may also be idnigsut ished on a plot of CaO+N2Oa vs. Sr (Figure 4.33). On both these diagrams, the lVsdioroif Batholith shows similarities to modern TTG suites. Figure 4.33: Plots of Vioolsdrif adakitic rocks ao nCaO+Na2O vs. Sr diagram after Castillo (2006). The presence or absence of a primitive componmenptl,y ing the involvement of peridotitic mantle in the generation of the magm maasy, be displayed on a variety of diagrams taking into account variations in MgO M(ogr# ), TiO2, Cr and Ni. Figure 4.34(a) shows the adakitic rocks of the VioolsdBraift holith on a diagram of Si2O vs. Mg# according to which normal TTG magmas are coenrseid products of melting in the lower crust, while adakitic magmas are considereod upcts of melting of a subducting slab (Condie, 2005a). Figure 4.34(b) shows theo lVsidorif Batholith adakitic rocks plotted on a diagram of Cr vs. Ni, which indicatthees presence of a primitive component. 143 (a) (b) Figure 4.34: The Vioolsdrif Batholith adakitic rkosc plotted on diagrams of S2i Ovs. Mg# (a) and Cr vs. Ni (b). Both diagrams are after Condie (2005a). 144 Some conclusions can be drawn from the study connincge rthe recognition of adakitic and TTG rocks in the Vioolsdrif Batholith. a) The fact that adakitic rocks in the batholitdh ndoi t intrude as unique intrusive events (showing intrusive contact relationshipsh wciotuntry rock) but merely represent compositional varieties within the calcalkaline ngirtoaids, strongly argues in favour of a magma mixing model involving the mixing of primaardy akitic magmas with non- adakitic magmas. This is supported by the infe mrreadgma mixing model for the development of the Blockwerf Migmatite (4.2.5 Blowcekrf Migmatite), in which the adakitic rocks occur most frequently. b) High La/Yb ratios are commonly related to a kth oicverlying crust (Planck and Langmuir, 1988; Mantle and Collins, 2008) as haesn b aemply shown for the Andean margin (Kaye t al., 1999, Hashcket al., 2002, Hildreth and Moorbath, 1988). La/Yb ratios for the Vioolsdrif Suite suggest that theo leuvtion of the batholith was not associated with a thick overlying crust. There,f othre anomalously high La/Yb ratios (compared to the rest of the batholith) for thec Bklwoerf Migmatite imply a very uneven crustal base for the batholith. This can be acehdie bvy a process such as mafic underplating. c) If modern adakite and Archean TTG share simgielanre tic processes then, if anything, it shows that processes of crust formna itnio the Archean cannot be compared to those on the modern earth. In the Archean tphreosceesses produced vast volumes of magmas while on the modern earth, they produceti vrellya little. The fact that the changing bulk composition in the earth’s crust f rToTmG in the Archean to granodiorite- granitic after 2.5 Ga coincided with a major pe roiof dcontinental crust formation, surely cannot be considered coincidental. It implies mr (aajond rather abrupt) changes in crust formation processes from the Archean to preseonrt. t hFe early Archean, processes involving large-scale melting of deeper parts oef mthantle (not necessarily involving any overlying crust) are implied. The fact that Arcnh eTaTG on average have higher contents of fluid-mobile elements like Ba, K and Rb than kaidteas (Condie, 2005a), attests to 145 higher rates of metasomatism in the upper manthleic, hw are today commonly thought to be active in the mantle wedge overlying the subindgu cptlate. 146 5. CONCLUSIONS 5.1 SUBDIVISIONS The proposed subdivision of the Vioolsdrif Bathho liint South Africa and within the currently accepted boundaries is shown on the mt tahpe aback of this thesis. It largely follows the subdivisions from previous studies.e Tmhajor and trace element evidence presented in this study, along with isotope andio rmadetric evidence from previous studies, indicate that the whole batholith formreodm f a common source during the Orange River orogeny which started around 2.0 Ga. This study presents geochemical evidence for thbed ivsiusion of the Orange River Group into the Haib and De Hoop Subgroups, a subdiviswihoinc h was previously based entirely on field evidence. It has been shown that the HSauibgroup shows a genetic relationship with the Vuurdood Subsuite, the latter of which m baey interpreted as remnants of the primary, mantle-derived magmas. The De Hoop Souubpg rdoes not show this genetic relationship with the Vuurdood Subsuite. Since the two subgroups of the Orange River Grocucpu poy two distinct geographical areas – the Haib and Richtersveld areas, sepabrya taend area of no outcrop (cover by younger strata) – the implication of this subdiovnis is that the two geographical areas may represent two significantly different partst hoef batholith. In the Haib area, not only the Haib Subgroup but also the Goodhouse Subsisu ioten, average more mafic than the Richtersveld part of the batholith. Both the Vuouord Subsuite and Khoromus Tonalite, the most mafic members of the Vioolsdrif Suite, laimreited to the Haib area and do not occur in the Richtersveld. The Vuurdood Subsuoitme pcrises mafic-ultramafic bodies, which can be expected to be limited to the lowertrs p oaf the crust, since such magmas would not be able to rise to higher crustal levtherlso ugh normal magmatic processes, due to their high densities. The Haib Subgroup const aviinrtually no interlayered metasediments. In the Haib area, contact relahtiopns sbetween the volcanic group and the plutonic suite suggest intrusion at the roonte z of the volcanic pile. In the 147 Richtersveld, on the other hand, intrusive relasthioipns between the volcanics and the plutonic rocks are mostly concordant. The De HSooupb group contains a relatively high proportion of interlayered metasediments and thsey Rntojieberg Formation represents a rather thick, mature metasedimentary unit, ovegrl ytihne volcanics. The Kook River Formation, an entirely porphyrytic unit, and them Xinaxaip River Granodiorite have both been considered to represent subvolcanic unitste (rR, 1it980). These observations suggest that the Richtersveld may represent an i na rwehaich the upper part of the batholith has been preserved, while the Haib aerperae sr ents a deeper crustal part of the batholith. Furthermore, tectonic environment dimiscinration diagrams suggest that the Haib Subgroup was formed in a primitive arc envmiroen t, while the De Hoop Subgroup was formed in a normal continental arc environm eTnhti.s suggests that the Haib area represent the initial stages, while the Richterds vaerel a represents the later stages, in the batholith's evolution. The subdivision of the De Hoop Subgroup accordoin tgh et scheme of Ritter (1980) may be controversial since it is largely based on ao rtyh envisaging development from a volcanic centre in the northeast of the Richterds.v eTlhis study identifies the Blockwerf Migmatite as a possible volcanic centre. Howenvoenr,e of the volcanic rocks of the De Hoop Subgroup which were analyzed for REE in thtuisd ys bears the unique La/Yb signatures of the Blockwerf Migmatite. This meathnast the De Hoop Subgroup as it is currently exposed, probably did not originate frtohmis unit. However, the geochemical variations within the subgroup presented in thuisd ys tdo not motivate for any alternative subdivision scheme. Based on its clear association with the rest o fb tahteholith on variation diagrams, the Vuurdood Subsuite is considered here as remna nthtse opfrimary source magmas. Multi- element diagrams show that this subsuite was stuebdj etoc the same magmatic processes as the rest of the batholith while initial isotospigen atures and REE patterns show that it shares a similar source. Some of the ultramacfikcs r oin the subsuite are considered to represent cumulate material which formed in sha lcloruwstal magma chambers during fractional crystallization. 148 The Goodhouse Subsuite is subdivided into fours u. n Tithe Khoromus Tonalite, based on contact relationships and radiometric evidenceth, eis o ldest unit in the subsuite. It is the most mafic unit in the subsuite and shows limiteradd gation from the Vuurdood Subsuite on classification and variation diagrams. Field apnetrographic evidence of possible magma mixing processes are to be found in this. unit The Xaminxaip River Granodiorite also shows intvrues rielationships with the Gaarseep Granodiorite, showing the latter to be intrusivteo inthe former. The Xaminxaip River Granodiorite and Khoromus Tonalite are not foun dco intact with each other but a single zircon age for the Xaminxaip River Granodiorite wshso it to be of similar age as the Gaarseep Granodiorite. Therefore, the Xaminxavipe rR Gi ranodiorite is placed chronologically between the Khoromus Tonalite ahned Gt aarseep Granodiorite. The Xaminxaip River Granodiorite is thought to reprets ae nsubvolcanic unit, in agreement with Ritter (1980). The development of migmatite in places in the Xaxmaipn River Granodiorite can be explained in terms of locally existing high gradet ammorphic conditions in shallower parts of the crust during orogenic processes (sacsr idbe d by Mehnert, 1968), together with the simultaneous intrusion of the Gaarseepn Godraiorite on regional scale (in agreement with De villiers and Söhnge, 1959). Ardcicnog to Mehnert o(p cit.), much evidence suggest that migmatites, as currently seexdp oat surface, did not form at the base of the crust or at depths of 15-20 km, asl ds hboeu assumed from experimental data if a normal geothermal gradient of 30-40°C/km is tpuolated. Rather it has to be assumed that the metamorphic conditions necessary for tphaeritrial melting was dragged upwards into relatively shallow depths by orogenic process. s Tehe geothermal gradient may reach 5-150°C/km in orogenic belts, whereby all the meotrapmhic zones can occur side by side within a rather small area, including those of ipaal rotr complete melting, at about 4-10 km depths. 149 The Gaarseep Granodiorite is the main phase ocfu trhre ntly exposed Vioolsdrif Batholith and spans the entire compositional rarnegpere sented by the other units (except the ultramafic rocks of the Vuurdood Subsuite) aesll was their geographical distribution. It is the only unit which occur in both the Haibd a Rnichtersveld areas but is on average more mafic in the Haib area than in the Richterds.v el The Hoogoor Granite is the only unit consideredth ins study which is located outside the batholith type area and bears the pervasive fonlia atind metamorphism of the Namaqua orogeny. Its inclusion in the batholith may ind eremd ain controversial but based on the available radiometric and geochemical evidencies, cito nsidered here as part of the Goodhouse Subsuite. All of its major and tracem elnet variations coincide with similar compositions in the rest of the batholith. Curlrye natvailable radiometric data supports its contemporaneous development with the granodiorroitcick s of the Gaarseep Granodiorite (~1.89 Ga; Moen and Toogood, 2007). The Ramansdrif Subsuite postdates the main intnru bsyio about 300 m.y. (e.g., Cornetll al., 2007). It is regarded here as the product rotfia pl acrustal melting of the older phases of the Vioolsdrif Batholith during the final stag oefs the Orange River orogeny. Within the Ramansdrif Subsuite, the further subdivisioton tinhe Ghaams and Sout Granites can be justified on three criteria: texturally the S oGurtanite is consistently coarser grained than the Ghaams Granite; petrographically the SGoraunt ite is microcline-dominated with abundant exsolution textures while the Ghaams Gter aisn iorthoclase-dominated with only minor development of exsolution textures; gheeomcically the Ghaams Granite is on average less felsic than the Sout Granite. Peatprohgicrally and geochemically all the occurrences of the Ramansdrif Subsuite which acraet elod within the Hoogoor Granite, can be correlated with the Ghaams granite. 5.2 MAGMATIC PROCESSES AND TECTONIC ENVIRONMENT All the evidence relate the Vioolsdrif Batholith mtoodern continental margin subduction zone tectonic settings. The initial stages odf eitvse lopment, represented by the 150 Vuurdood Subsuite and part of the Haib Subgroupy, bmea comparable to a primitive continental arc setting. The absence of basaoltmicp cositions in the volcanic succession argue against an intra-ocean island arc settinhge.r eT is also no evidence to be found that any of the volcanic units were deposited under rw. a Itne fact, the characteristics of the volcanic rocks call for deposition on a contine nctraulst. The bulk chemical composition of the batholith furthermore demands the presenf caen overlying continental crust, in terms of modern tectono-magmatic theories. Le1a9ke8 3(), among others, states that granite liquids cannot be derived as primary mferoltms either mantle peridotite or subducted oceanic crust. Atherton and Tarney ()1 s9t7a9te that, over the past few years, evidence have amassed for mixed mantle-crust o froigr imn any granitic rocks. In the Vioolsdrif Batholith, the earliest phases have rlhitiyco compositions. In the Vioolsdrif Batholith, no evidence can be nfodu of pre-existing crust and isotope data argue against it. Modern continental margoilnc avnic arcs are underlain by thick continental crust and the volcanic rocks show c elevaidrence of crustal contamination. For the Vioolsdrif Batholith, La/Yb ratios indica theat the crustal thickness during its development cannot be compared to modern examepxlceesp, t in a few samples of the Blockwerf Migmatite. The Orange River Group alsho wss no evidence of crustal contamination in relation to the Vioolsdrif Sui tTe.h is may be interpreted as further evidence that tectonomagmatic processes durinAg rtchhee an-Proterozoic transition period, may not be comparable to modern exampTlehsis. is supported by studies like that of Hamilton (1998) as well as various studdieas ling with TTG and adakite. With the onset of formation at about 2.0 Ga, the Vioroifl sBdatholith falls within a geological time era when crust formation processes are cornesdid teo have been similar to those observed today (e.g., Hamiltoonp, cit.). However, the above observations call for the need of alternative theories to those based ono rumniitfarianism to explain crust formation processes in the Archean and Protero zoic. A thin continental crust must have been presentht ein a rea where the Vioolsdrif Batholith developed. Overall low La/Yb ratios bear evidetnhcaet the crust could not have been thicker than 20 km (except under the Blockwerf Maigtimte). The fact that the Vioolsdrif 151 Batholith does not show isotope signatures sugngge sat ipre-existing crust may be explained by two factors. Firstly, the fact thhaet tcrust was thin means that only a small volume of it was recycled. Secondly, the pre-einxgis ct rust must have been a primary crust which did not undergo previous recycling oedpeiss, so that its isotope ratios were not significantly raised. The age of this pre-etinxigs crust might have been 2.7 Ga, as indicated by a single inherited zircon (Nordin, 92)0.0 Furthermore, it was noted before that the Vioolsdrif Batholith does not coincide fpeecrtly with fields defined for Mesozoic (Cordilleran) batholiths on the N2Oa vs. K2O and MALI vs. K2O diagrams of Froset t al. (2001; Figures 4.30 and 4.31). This may also ere tloa tthe fact that the crust associated with the development of the Vioolsdrif Batholith sw ma uch thinner than that associated with the development of the Mesozoic-Tertiary balithso. In the light of all the evidence, the following eluvtoionary model is suggested for the Vioolsdrif Batholith and summarized in Figure 5.1: Stage 1 (Figure 5.1a): Initial dehydration andt ipaal mr elting processes in the subduction zone ced vertical heat flow and accumulation oft mate tlhe base of the original crust. This led to limited crustal melting, producing fiec lsmagmas which rose through the overlying crust and erupted at surface. Thesere aprree sented by the Tsams Formation. Stage 2 (Figure 5.1b): At this stage, the mantales MwORB-like in composition but with lower overall trace element concentrations than m thoedern MORB. In the mantle wedge, olivine-tholeiitic magmas were produced abryt ipal melting’ in accordance with modern subduction zone processes (e.g., Condie5,b 2).0 0These magmas rose to shallow magma chambers (3-6 km depth) where they underfwraecntito nal crystallization to produce tholeiitic and more evolved magmas. Thpehsaes es are represented by the Vuurdood Subsuite. During the residence time eo fm thagmas in the magma chambers and fractional crystallization, cumulates comprgis pinyroxene and olivine were also formed. With progressive subduction, volcanismd pormeinated and the Haib Subgroup was deposited at surface. Within the crust, trhset pfihases of the plutonic suite were formed. 152 Figure 5.1(a): Initial stages of the evolutiont hoef Vioolsdrif Suite. Vertical heat transfer anudil db-up of magma at the base of the crust produce limitedic f emlsagmas which erupt at surface through vents and fissures, forming the Tsams Formation. Figure 5.1(b): Early stages (primitive arc envimroennt). Formation of intra-crustal magma chambnedrs a early plutons; eruption of basaltic-andesitic la(Nvao us Formation) produced by fractional crystaltliiozna in the magma chambers; limited deposition of sedimrey nsttarata at surface. 153 Figure 5.1(c): Late stages (normal arc environm). e Enxt tensive partial melting in the mantle lea d to basaltic underplating in parts along the arc a xTihse. rest of the crust remain relatively thin dou e t continuous reworking by subduction. Early magmam cbhers and plutons crystallize within the crust (Vuurdood Subsuite and early stages of the Goodeh Sousbsuite). Remnants of the original crust ma y be preserved while the rest of the crust is comple rtewlyorked into younger, more felsic crust. Pa rmtiaellting of the crust results in magma mixing processes ebentw mantle and crustal melts. Younger plutonsu dinet r older units, including the root zone of the oldoelrc vanic units. At surface, more felsic lavas and sedimentary deposits are formed (De Hoop Subgr oup). Figure 5.1(d): Post-Orange River orogeny. Tect odneiformation, probably the initial stages of thaen -P African orogeny, produces a normal fault betweeen H thaib and Richtersveld areas, juxtapositioning a deeper part of the crust along a surface to nerafar-cseu part. Subsequent erosion and the deposoift iothne Nama Group lead to the concealment of the boun dary. Stage 3 (Figure 5.1c): Partial melting and recnygc loi f newly formed crust became increasingly important and eventually predomina pterdo,ducing mainly felsic magmas. During these later stages, magma mixing and contaatmioin played a key role in the determination of final magma compositions. TheH Doeo p Subgroup is the surface 154 representative of these stages. The various oufn tihtse Goodhouse Subsuite were produced at different stages throughout this evioonluatry history. Stage 4 (Figure 5.1d): After closure of the Ora Rngiveer orogeny, probably during the initial stages of the Pan-african orogeny, the Rteircshveld area was thrown down relative to the Haib area along a normal fault. This reesdu ilnt the juxtapositioning of a deeper part of the crust (exposed in the Haib area) tuor faa cse to near-surface part of the crust (exposed in the Richtersveld). Subsequent eroasniodn t he deposition of the Nama Group followed, concealing the Post-Orange River normaaul tf. The progressive development of a calcalkaline nea ftruorm more primitive magmas in modern arcs, relates to the thickness of the oivnegr lcyrust with the magmas becoming progressively more potassic as well as alkalic,y a fwroam the arc towards the continent (e.g., Miyashiro, 1974). For the Vioolsdrif Batihtho,l the remnant trend of an arc axis could not be established during the current stu Hdoyw. ever, there is an increase in the calcalkaline nature and2 KO contents from the Vuurdood Subsuite, throughH tahieb Subgroup to the rest of the batholith. This feea toufr modern continental arcs is often attributed to an increase in thickness of the oyvinegrl crust, but may be related to any process leading to the enrichment o2Of K including fractional crystallization, magma mixing with more felsic magmas, assimilation of tcinoenntal crust, or the metasomatic process in the mantle wedge. It may also be rde ltaot ethe source composition. At least a portion of this variation is likely to be contribeudt by a mantle-derived component because it is also observed in island arc lavaes,r ew ah continental contribution is lacking (Dickenson, 1975). No progressive increase in crustal thickening frooldme r to younger units in the Vioolsdrif Batholith can be observed in the La/Yabti ors, as can be for the Andean arc (e.g., Haschke t al., 2006). This attests to continuous recyclingn eowf ly formed crust throughout its development. The model of recyc olinf gthe same material during the evolution of the batholith is also supported by nthaeture of fragments in agglomerates, 155 pebbles in conglomerates and xenoliths in lavag aranndi toids, which resemble material from the same batholith throughout (e.g., Figur.e3s, 2.5 and 2.6). The nature of the Ramansdrif Subsuite has not sbetetlne d. Initial isotope ratios indicate that it is the product of partial melting of conetnintal crust (Reid, 1977). The fact that its geochemical variation patterns coincide closelyh wthite Vioolsdrif Batholith, as well as its close spatial association with thise batho liimthp,lies that the Goodhouse Subsuite acted as the source for the Ramansdrif Subsuitem masa.g Radiometric ages indicate it to be post-Goodhouse Subsuite but there is no straul cetuvirdence of a major orogenic event at this time. On classification diagrams of Freots at l. (2001), the Ramansdrif Subsuite shows similarity to both peraluminous leucogran aitneds cordilleran batholiths (Figures 4.26 and 4.27). On the discrimination diagrams g froarnites of Pearcet al. (1984), it plots mostly in the field for volcanic arc grani t(eFsigure 4.28). The latter observation speaks in favour of the idea that the Ramansdrbifs Suiute was derived from the Goodhouse Subsuite, which has been shown to haigvien aotred in that environment. The peraluminous leucogranites of Froest ta l. (2001), tend to be associated with overthickened orogens and are usually producedm bayll sdegrees of melting that is typically ascribed to a stage of crustal rebounidth, owut any direct mantle contribution. La/Yb ratios argue against the Vioolsdrif Batho lbitehing associated with an overthickened orogen. It is therefore suggestaetd t hthe Ramansdrif Subsuite represents an event during which the Vioolsdrif Batholith unrwdent partial melting to produce felsic magmas. The orogenic event leading top tahristi al melting has not been identified. 5.3 THE RICHTERSVELD SUBPROVINCE IN ITS REGIONAL SETTING There is currently general consensus that the Nuaam Maqetamorphic Province comprises a number of unique tectono-stratigraphic terranheisc hw were amalgamated during the 1.3-1.0 Ga Namaqua orogeny (e.g., Stoewt ael ., 1984; Joubert, 1986; Thomeats a l., 1994), although the exact configuration and ideicnattifion of the terrane boundaries is still controversial (e.g., Collistoent al., 1992; Cornell and Pettersson, 2007). In thcisto tneic framework, the Vioolsdrif Batholith is considereod r tepresent a relict Eburnian province 156 surrounded by the Kibaran Namaqua mobile belt, dbeodu nbetween two tectonic boundaries, viz., the Lower Fish River and Grookth toherusts (e.g., Blignauelt t al., 1983). Based on the tectono-magmatic model for the Vioroifl sBdatholith presented above, it is tempting to envisage the regional tectonic evonlu otiof the area in terms of a tectonic model in which two continents, separated by an no,c aepaproach and eventually collide. In this model, the Vioolsdrif Batholith representhtse subduction stage during closing of the ocean, while the Namaqua Metamorphic Provienpcere rsents the collisional stage. However, such a model is contradicted by the agpe b geatween the end of the Orange River orogeny (1.73 Ga, Ramansdrif Subsuite) aned o tnhset of the Namaqua orogeny (1.3 Ga, Little Namaqualand Suite). Furthermohre, Rt amansdrif magmatic event might be considered unrelated to the Vioolsdrif magma taisnmd Cornelle t al. (2007) in fact suggest that the Ramansdrif leucogranites be eexdc lufrdom the Vioolsdrif Suite. This would indicate an even older age for the end o fO thraenge River orogeny and would be indicated by the youngest ages of the Goodhousseu Situeb, which is currently around 1.8 Ga (e.g., Reid, 1977). The inability to identify Archean continental cr uasstsociated with the Vioolsdrif Batholith and a basement to the Orange River G rmouapy, suggest the complete recycling of such crust during the Orange River orogeny. r eCnutlry, the only evidence of such components in the Vioolsdrif Batholith is in a slien ginherited zircon core of 2.7 Ga (Nordin, 2009). Other inherited zircon cores tote d caoincide with the earlier stages of the batholith (U-Pb, LA-ICPMS ages of 1.9 and 2.a0 inG the Xaminxaip River Granodiorite; see Appendix 2). At a number of lioticeas, xenoliths are found in the volcanics of the Orange River Group but none osf eth heave as yet been dated. Based on evidence indicating rapid recycling of newly form medaterial throughout the batholith, these xenoliths most likely represent recycled mriaal toef earlier phases. Reid (1977) dated a xenolith of granodiorite at 1.73 Ma bust tshuiggests resetting during the Ramansdrif magmatic event. 157 If current models for the tectonic setting of thieo oVlsdrif Batholith is accepted then there is no point in searching for basement to the Or aRnigver Group outside the Richtersveld Subprovince, since all evidence related to theo bliathth would be bounded between the Lower Fish River and Groothoek thrusts. If it oist naccepted, a number of units may qualify for such a position. The model of terraancec retion currently reigns popular and alternative models are few. However, much evid eenxciset to suggest that the Vioolsdrif Batholith might not be isolated from the Namaquat aMmeorphic Province by the proposed terrane boundaries. Ward (1977) desc trhibee bsoundary with the “Bushmanland Subprovince” as fault bounded in p baurt ,“elsewhere the Vioolsdrif igneous assemblage both rests on and intrudesr ethy eg ngeisses characteristic of the south” (i.e., the Gladkop Suite, which extendsa ar ss of uth as Springbok in Namaqualand). Maraeist al. (2001) and Minnaaer t al. (2011) also present evidence that the Gladkop Suite is intruded by the Vioolsdrif tSeu. i Radiometric ages of 1.8 Ga for the Gladkop Suite also contradicts the idea that thceh tRerisveld Subprovince is unique in its early-Proterozoic age. Isotopic data and modesl asgheow extensive reworking of Vioolsdrif-age crust in the rest of the Namaquav Pinrcoe during the Namaqua orogeny (e.g., Barton, 1983; Reid, 1981). Furthermore, it seems likely that the Orange RGivreoru p grades from a volcanic- dominated unit in its type area, to a sedimentaormy-idnated unit eastwards. In the east, the pervasive foliation and metamorphism of the Naqauma orogeny have rendered correlation among stratigraphic units controve rbsuiatl various studies imply the possible correlation of the Orange River Group with NamaMqueata morphic Province supracrustal units. Moen and Toogood (2007) considered the gOer aRniver and Droëboom Groups lateral equivalents, while possible correlationtsw bee n the Orange River and Bushmanland Groups are suggested in all of Conll is(1to990), Agenbacht (2007), Bertrand (1975) and Moen and Toogood (2007). lI no fa tlhese findings, the Steinkopf Gneiss of the Gladkop Suite is implied as basemtoe tnhte Orange River Group. In Bushmanland, the “Achap Gneiss” has been consi dtehree bdasement to the Bushmanland Group by Moore (1977) and Watkeys (1986). Baeilti ea l. (2007) also state that combined earlier works have suggested that the a“Ap”c hand “Hoogoor” Gneisses (the 158 latter does not refer to the Hoogoor Granite) aldre ro than the Bushmanland Group and acted as the basement and provenance to the miemtaesnetsd and source of the mineralization (at Black Mountain and Gamsberg ms)i.n eRadiometric data contradicts the idea of a correlation between the Orange Raivnedr Bushmanland Groups with the inferred ages of deposition for Bushmanland foromnast iranging between 1.5-1.2 Ga (e.g., Bailiee t al., 2007). However, the 2.0 Ga ages for the Haibbg Srouup are representative of the early stages of Orange RGivroeur p evolution. Inferred depositional ages of post-1.76 Ga for the Rosyntjieberg metamseendtis, imply that Orange River Group deposition continued even after Ramansdrgifm maatism, suggesting a possible continuation into Bushmanland deposition, whichc perdees the Namaqua orogeny. 6. ACKNOWLEDGEMENTS Dave Cornell and Valby von Schrijndel of the Gotbhuerng University are warmly thanked for U-Pb (LA-ICPMS) age determinations oirnco zns from two samples. The author wishes to thank Prof. A.E. Schoch, nolyt ofor acting as supervisor but also for his invaluable input and positive guidance uthgrhoout the execution of this study. REFERENCES Agenbacht, A.L.D., 2007. The geology of the Pofeard adrea. Explanation, 2918 Pofadder (1:250 000), Council for Geoscience, 8 9 p. Agrawal, S., Guevara, M., Verma, S.P., 2008. Teic tdoinscrimination of basic and ultrabasic volcanic rocks through log-transformaetdio rs of immobile trace elements. International Geology Review 50, 1057-1079. Allsopp H.L., Köstlin E.O., Welke H.J., Kröner An. da Blignault H.J., 1979. Rb-Sr and U-Pb geochronology of late Precambrian-early Pazloaieco igneous activity in the Richtersveld (South Africa) and southern South W Aefrsict a. Transactions, Geological Society of South Africa, 82, 185 - 204. Anderson, J.L. 1983. Proterozoic anorogenic gera pnliut tonism of North America. In: L.G. Mederis, C.W. Byers, D.M. Mickelson and W.Ch.a Snks (editors) Proterozoic Geology, Selected Papers from an Internationale Prorozot ic Symposium. Memoir, Geological Society of America, 161, 133-154. Arth, J.G. and Hanson, G.N. 1975. Geochemistry o arnigdin of the early Precambrian crust of northern Minnesota. Geochimica et Cosmimoicha Acta, 39, 325-362. Atherton, M.P. and Tarney, J. 1979. Origin of gitrea nbatholiths: geochemical evidence. Shiva Publishing, 148 p. Bailie, R.H., Armstrong, R.A. and Reid, D.L. 200 T7h. e Bushmanland Group supracrustal succession, Aggeneys, Bushmanlandth, SAofruica: Provenance, age of deposition and metamorphism. Journal of Geolo1gy0,( 1), 59 - 86. Barker, F. 1979. trondhjemites, dacites and rde larotecks. Elsevier, New York. Barton, E.S. 1983. Reconnaissance isotopic ingvaetsiotins in the Namaqua mobile belt and implications for Proterozoic crustal evolut–io nN amaqualand Geotraverse. In: B.J.V. Botha (editor) Namaqualand Metamorphic Complexe. c Siapl Publication, Geological Society of South Africa, 10, 45-66. Bertrand, J.M. 1975. Granitoids and deformatioqnu seence in the Goodhouse - Henkries area: A new interpretation of the relationshipw beeetn rocks of the Vioolsdrif - Goodhouse area and the Namaqualand and Bushmagnnlaenisds es. Annual Report, Precambrian Research Unit, University of Cape To 1w3n, ,61 - 70. Beukes, G.J. 1973. 'n Geologiese ondersoek va gne dbiied suid van Warmbad, Suidwes- Afrika, met spesiale verwysing na die metamorf-matgiemse assosiasies van die Voorkambriese gesteentes. D.Sc. Thesis, Unive orsf itthye Orange Free state, Bloemfontein, 333 p. Blignault, H.J. 1974. Aspects of the Richtersverlodv Pince. In: A. Kröner (editor) Contributes to the Precambrian geology of SouthAefrinc a. Bulletin, Precambrian Research Unit, University of Cape Town, 15, 49 – 56 Blignault, H.J. 1977. Structural-metamorphic imnpt roi n part of the Namaqua mobile belt in South West Africa. Bulletin, Precambriaens Rearch Unit, University of Cape Town, 23, 197 p. Blignault, H.J., Jackson, M.P.A., Beukes, G.J. Taonodg ood, D.J. 1974. The Namaqua tectonic province in South West Africa. In: A. Kreörn (Editor) Contributions to the Precambrian geology of southern Africa. BullePtinre, cambrian Research Unit, University of Cape Town, 15, 29 – 48. Blignault, H.J., Van Aswegen, G., Van der MerweW, S. a. nd Colliston, W.P. 1983. The Namaqualand geotraverse and environs: part of rtohte rPozoic Namaqua mobile belt. In: B.J.V. Botha (editor) Namaqualand Metamorphoicm Cplex. Special Publication, Geological Society of South Africa, 10, 1-29. Bowen, N.L. 1928. The evolution of the igneousk rso. c Princeton University Press. Bowles, M. 1988. Tungsten mineralisation in them Naaqualand - Bushmanland region, Northwestern Cape, South Africa. Memoir, Geololg Sicuarvey of South Africa, 74, 75 p. Boynton, W.V. 1984. Geochemistry of the rare e aerlethments: meteorite studies. In: P. Henderson (editor) Rare earth element geochem isEtlrsye.vier, 63-114. Brown, G.C., Thorpe, R.S. and Webb, P.C. 1984. gTehoechemical characteristics of granitoids in contrasting arcs and comments on maa sgomurce. Journal of the Geological Society of London, 141, 413-426. Buddington, A.F. 1959. Granite emplacement witehc siapl reference to North America. Bulletin, Geological Society of America, 70, 6717-7. 4 Castillo, P.R. 2006. An overview of adakite peetrnoegsis. Chinese Science Bulletin, 51, 257-268. Coetzee, C.B. 1941. The petrology of the Goodh-oPuesllea area, Namaqualand, South Africa. Transactions of the Geological SocietyS ofu th Africa, 44, 167 – 206. Colliston, W.P. 1990. A stratigraphic and struactl uinr vestigation of part of the Namaqua mobile belt between Dabenoris and Steyerkraal, hS Aoufrtica. Ph.D. thesis, University of the Orange Free state, Bloemfontein, 260 p. Colliston, W.P. and Schoch, A.E. 2006. The disuttriiobn and diagnostic features of deformed plutonic rocks in two terranes of the Nqaumaa mobile belt along the Orange (Gariep) River, South Africa. South African Joulr onfa Geology, 109, 369-392. Colliston, W.P., Praekelt, H.E. and Schoch, A.E9. 21.9 The western margin of the Proterozoic Namaqua-Natal mobile belt in the nosrttweren Cape Province. Field guide book for excursion 6A, 2th4 Geocongress, Geological Society of South Africa. Condie, K.C. 2005a. TTGs and adakites: are thethy sbloab melts? Lithos, 80, 33-44. Condie, K.C. 2005b. Earth as an evolving plane stayrsytem. Elsevier Academic Press, 447 p. Cornell, D.H. and Pettersson, A. 2007. Ion proibrceo zn dating of metasediments from the Areachap and Kakamas Terrane, Namaqua-Natvailn Pcero and the stratigraphic integrity of the Areachap Group. South African rJnoaul of Geology, 110, 169 - 178. Cornell, D.H., Thomas, R.J., Moen, H.F.G., ReidL, .D, M. oore, J.M. and Gibson, R.L. 2007. The Namaqua-Natal Province. In: The Geoolofg Syo uth Africa, M.R. Johnson, C.R. Anhauesser and R.J. Thomas (editors), 3259 .– 37 Cox, K.G., Bell, J.D. and Pankhurst, R.J. 1979.e Tinhterpretation of igneous rocks. George, Allen and Unwin, London. De la Roche, H., Leterrier, J., Grande Claude,n Pd .M aarchal, M., 1980. A classification of volcanic and plutonic rocks using R1-R2 diagr aamnds major element analyses – its relationship and current nomenclature. Chemicaoll oGgey, 29, 183 – 210. De Villiers, J. and Burger, A.J. 1967. Note on mthienimum age of certain granites from the Richtersveld area. Annals, Geological Survfe Syo outh Africa, 6, pp. 83 - 84. De Villiers, J. and Söhnge, P.G. 1959. The geo loofg tyhe Richtersveld. Memoir, Geological Survey of South Africa, 48, 266 p. Defant, M.J. and Drummond, M.S. 1990. Derivatiof ns oome modern arc magmas by melting of young subducted lithosphere. Nature7,, 36462-665. Defant, M.J. and Kepenzhinskas, P. 2001. Evidesnucgeg ests slab melting in arc magmas. Eos, 82, 62-70. Dhuime, B, Bosch, D, Bodinier, J.-L., Garrido, C, .BJr.uguier, O., Hussain, S.S. and Dawood, H. 2007. Multistage evolution of the J uijaltlramafic–mafic complex (Kohistan, N Pakistan): Implications for building the roots i solfand arcs. Earth and Planetary Science Letters 261, 179–200. Dickenson, W.R. 1975. Potash-depth h()K r-elations in continental margins and intra- oceanic magmatic arcs. Geology, 3, 53-56. Drummond, M.S. and Defant, M.J. 1990. A modelt rforn dhjemite-tonalite-dacite genesis and crustal growth via slab melting: Arcahna teo modern comparisons. Journal of Geophysical Research, 95, 21503-21521. Frost, C.D. and Frost, B.R. 1997. High-K, iron-icehner d rapakivi-type granites: the tholeiite connection. Geology, 25, pp. 647-650. Frost, B.R. and Linsley, D.H. 1991. The occurre onfc Fee-Ti oxides in igneous rocks. In: D.H. Lindsley (editor) Oxide Minerals: Petrologincd a Magnetic Significance. Reviews in Mineralogy, 25, 433-486. Frost, B.R., Barnes, C.G., Collins, W.J., ArculRus.J, ., Ellis, D.J. and Frost, C.D. 2001. A geochemical classification for granitic rockso.u rJnal of Petrology, 42, 2033-2048. Germs, G.J.B. 1972. The stratigraphy and palaeloognyt oof the lower Nama Group, South West Africa. Bulletin, Precambrian ReseaUrcnhit , University of Cape Town, 12, 250 p. Gevers, T.W. 1941. Correlation of the pre-Capem faotrions in the western Cape Province: Discussion, Proceedings of the Geolo gSicoacliety of South Africa, 43, p. 17 – 30. Gevers, T.W., Partridge, F.C. and Joubert, G.K.7 1. 9 T3he pegmatite area south of the Orange River in Namaqualand. Memoir, Geologicarvl Seyu of South Africa, 31, 172 p. Gresse, P.G. and Germs, G.J.B., 1993. The Namelaa nfodr basin: sedimentation, major unconformity-bounded sequences and multisided ea cmtiavrgin advance. Precambrian Research, 63, p. 247 – 272 Griffin, W.L., O’Reilly, S.Y., Afonso, J.C. and Begg, G.C. 2009. The composition and evolution of lithospheric mantle: a re-evaluationnd aits tectonic implications. Journal of Petrology, 50, 1185-1204. Hamilton, W.B. 1998. Archean deformation and matgismma: no plate tectonics. Geocongress, Geological Society of South Afric a2,6 p-.28. Hartnady, C.J.H., Joubert, P. and Stowe, C.W. 1 9P8r5o.terozoic crustal evolution in southwestern Africa. Episodes, 8, 236 - 244. Haschke, M., Günther, A., Melnick, D, Echtler, HR.e, utter, K-J, Scheuber, E and Oncken, O. 2006. Central and southern Andean nteicc etovolution inferred from arc magmatism. In: The Central and Southern Andesin, gSeprr Verlag, p. 189-202. Haschke, M., Siebel, W., Günther, A. and ScheuEb.e 2r,0 02. Repeated crustal thickening and recycling during the Andean orogienn nyo rth Chili (21°-26°S). Journal of Geophysical Research, 107, 1-17. Hildreth, W. and Moorbath, S. 1988. Crustal cobnuttriions to arc magmatism in the Andes of central Chile. Contributions to Mineraylo agnd Petrology, 98, 455-489. Irvine, T.N. and Baragar, W.R.A. 1971. A guidet htoe chemical classification of the common volcanic rocks. Canadian Journal of Eacrtihen Sces, 8, 523-548. Jensen, L.S. 1976. A new cation plot for clasnsigfy siubalkalic volcanic rocks. Paper, Ontario Division of Mines, 66. Johnson, A.D. and Wyllie, P.J. 1988. Constraint sth oe origin of Archaean trondhjemites based on phase relationships of Nnueki sGs with H2O at 15kbar. Contributions to Mineralogy and Petrology, 100, 3p5p-.46. Joubert, P. 1986. Namaqualand - A model of Prozoteicro accretion? Transactions of the Geological Society of South Africa 89, 79-96. Joubert, P., Hartnady, C.J.H. and Von Veh, M.W.6 1. 9 T8he geology of the Alexander Bay area. Explanation, sheet 2816 (1:250 000),l oGgeicoal Survey of South Africa, 22 p. Kaur, P.K., Chaundhri, N., Kröner, A and Hofmann.W, A. 2009. Record of 1.82 Ga Andean-type continental arc magmatism in NE Raajans,t hIndia: Insights from zircon and Sm-Nd ages, combined with Nd-Sr isotope geocheym. i sGtrondwana Research, 16, 56- 71. Kay, R.W. 1978. Aleutian magnesian andesites: sm freoltm subducted Pacific Ocean crust. Journal of Volcanology and Geothermal Rrecshe, a4, 117-132. Kay, S.M. 2005. Tertiary to recent evolution of dAenan arc and back-arc magmas between 36°S and 38°S and evidence for Miocenelo swhinagl of the Nazca plate under the Neuquén basin. Int:h 6International Symposium on Andean Geodynamics, Barcelona, 420-423. Kay, S.M. and Mpodozis, C. 2002. Magmatism aso ab ep rto the Neogene shallowing of the Nazca plate beneath the modern Chilean flabt.- s Jlaournal of South American Earth Sciences, 15, 39-57. Kay, S.M., Mpodozis, C. and Coira, A.B. 1999. Nenoeg magmatism, tectonism and mineral deposits of the central Andes. In: B.Jin. nSekr (editor) Geology and Ore Deposits of the Central Andes. Special Publica, tSioonciety of Economic Geology, 7, 27- 59. King, E.M., Beard, B.L. and Valley, J.W. 2007. oSnttrium and oxygen isotope evidence for strike/slip movement of accreted terranes ein Itdhaho Batholith. Lithos, 96, 387-401. Kostitsyn, Y, A., Volkov, V.N. and Zhuravlev, D.Z2.0 07. Trace elements and evolution of granite melt as exemplified by the Raumid Plu, tSoonuthern Pamirs. Geochemistry International, 45, 971-982. Kröner, A. and Blignault, H.J. 1976. Towards ai ndietifon of some tectonic and igneous provinces in western South Africa and southern hS oWuet st Africa. Transactions, Geological Society of South Africa, 79, 232 – 238. Kröner, A., Barton, E.S., Burger, A.J., Allsopp,L H. a. nd Bertrand, J.M. 1983. The ages of the Goodhouse Granite and grey gneisses fro m athreginal zone of the Richtersveld Province and their bearing on the timing of teccto envients in the Namaqualand mobile belt. In: B.J.V. Botha (editor) Namaqualand Metarpmhoic Complex. Special Publication, Geological Society of South Africa,, 1023 - 130. Le Maitre, R.W. 1976. The geochemical variabiolitfy s ome common igneous rocks. Journal of Petrology, 17, 589-637. Le Maitre, R.W., Bateman, P., Dudek, A., Kelle,r ,L Ja.meyre Le Bas, M.J., Sabine, P.A., Schmid, R., Sorensen, H., Streckeisen, A., Woo Alle.Ry,. and Zanettin, B. 1989. A classification of igneous rocks and glossary omf tse.r Blackwell, Oxford, 193 p. Leake, B.E. 1983. Ultrametamorphism, migmatiteesl,t imng and granitoid formation. In: M.P. Atherton and C.D. Gribble (editors) Migmat,it eMselting and Metamorphism. Proceedings of the Geochemical Group of the Minoegriacal l Society, Shiva Publishing, 2-9. Maniar, P.D. and Piccoli, P.M. 1989. Tectonic driimscination of granitoids. Bulletin, Geological Society of America, 101, pp. 635-643. Mantle, G.W. and Collins, W.J. 2008. Quantifyinrgu sctal thickness variations in evolving orogens: correlation between arc basamltp coosition and Moho depth. Geology, 36, 87-90. Marais, J.A.H., Agenbacht, A.L.D., Prinsloo, M. a Bnadsson, W.A. 2001. Geology of the Springbok area. Explanation, sheet 2916 Spring(1b o: k2 50 000), Council for Geoscience, 103 p. Martin, H. 1999. Adakitic magmas: modern analog oufe Asrchean granitoids. Lithos, 46, 411-429. Martin, H., Smithies, R.H., Rapp, R., Moyen, J.a-Fnd. Champion, D. 2005. An overview of adakite, tonalite-trondhjemite-granodiorite (T)T, Gand sanukitoid: relationships and some implications for crustal evolution. Lithos9,, 71-24. Mehnert, K.R. 1968. Migmatites and the origin oraf ngitic rocks. Elsevier Publishing Company, Amsterdam, 393 p. Middlemost, E.A.K. 1965. Ultramafic rocks of thoeu sth-eastern Richtersveld. Transactions, Geological Society of South Afric8a,, 563 - 60. Miller, R.McG. 2008. The geology of Namibia. Habnodok, Geological Survey of Namibia, 3 volumes. Minnaar, H., Botha, P.M.W. and Roberts, D. 2011h.e Tgeology of the Alexander Bay area. Explanation, sheet 2816 Alexander Bay (510 : 0200), Council for Geoscience, 178 p. Minnitt, R.C.A. 1979. The geological setting ofr phoyry-type copper mineralization in the Haib River area, South West Africa. Ph.D. itsh e(usnpubl.), University of the Witwatersrand, Johannesburg, 366 p. Miyashiro, A. 1974. Volcanic rocks series in isdla anrcs and active continental margins. American Journal of Sciences, 274, 321-355. Moen, H.F.G. and Toogood, D.J. 2007. The geolof gthy eo Onseepkans area. Explanation, sheet 2818 Onseepkans (1 : 250 00o0u)n, cCil for Geoscience, 101 p. Moorbath, S. 1975. Evolution of Precambrian crfruosmt strontium evidence. Nature, 254, pp. 395-398. Moore, J.M. 1977. The geology of Namiesberg, Neorrnth Cape. Bulletin, Precambrian Research Unit, University of Cape Town, 20, 69 p. Mullen, E.D. 1983. MnO/TiO2/P2O5: a minor element discriminant for basaltic rockf s o oceanic environments and its implications for pgeetrnoesis. Earth and Planetary Science Letters, 62, 53-62. Nordin, F. 2009. U-Pb dating, Lu-Hf isotopic ansaelys and geothermobarometry of rocks in the Grünau Terrane and Richtersveld Subprov iNncaem,aqua Sector, southern Africa. M.Sc. thesis (unpublished), Göteborgs Univers5it4e tp, . Parada, M.A., Nyström, J.O. and Levi, B. 1999. tMipulel sources for the Coastal Batholith of central Chile (31-34°S): geochemicnadl aSr-Nd isotopic evidence and tectonic implications. Lithos, 46, 505-521. Patchett, P.J. 1989. Radiogenic isotope geochreym oifs rtare earth elements. In: P.H. Ribbe (editor), Geochemistry and Mineralogy of R Eaarerth Elements. Reviews in Mineralogy, 21, 348 p. Pearce, J.A. 1982. Trace element characterisft ilcasv aos from destructive plate boundaries. In: R.S. Thorpe (editor) Andesitesi.le yW, Chichester, 525-548. Pearce, J.A. 1983. Role in the sub-continenthaol slipt here in magma genesis at active continental margins. In: C.J. Hawkesworth and. MN.oJrry (editors) Continental Basalts and Mantle Xenoliths. Shiva, Nantwich, 230-249. Pearce, J.A. and Cann, J.R. 1973. Tectonic se ottfi nbgasic volcanic rocks determined using trace element analyses. Earth and PlanSetcaieryn ce Letters, 19, 290-300. Pearce, J.A. and Norry, M.J. 1979. Petrogenetpicli cimations of Ti, Zr, Y and Nb variations in volcanic rocks. Contributions to Merinalogy and Petrology, 69, 33-47. Pearce, J.A. and Parkinson, I.J. 1993. Trace enlte moedels for mantle melting: application to volcanic arc petrogenesis. In: H P.Mri.chard, T. Alabaster, N.B.W. Harris and C.R. Neary (editors) Magmatic processes antde ptelactonics. Special Publication, Geological Society, 76, 373-403. Pearce, J.A. and Peate, D.W. 1995. Tectonic imatpiolincs of the composition of volcanic arc magmas. Annual Review, Earth and Planetaryn Scec,ie 23, 251–285. Pearce, J.A., Lippard, S.J. and Roberts, S. 19h8a4r.a Ccteristics and tectonic significance of supra subduction zone ophiolites. In: B.P. Koakaer land M.R. Howells (editors) Marginal Basin Geology. Special Publication, Geoicloalg Society of London, 16, 77-94. Perugini, D. and Poli, G. 2004. Analysis and nuicmael rsimulation of chaotic advection and chemical diffusion during magma mixing: petgroicloal implications. Lithos, 78, 43- 66. Perugini, D., Poli, G. and Gatta, G.D. 2002. Ansaisl yand simulation of magma mixing processes in 3D. Lithos, 65, 313-330. Petro, W.L., Vogel, T.A. and Wilband, J.T. 1979.a jMor-element chemistry of plutonic rock suites from compressional and extensionael pblaotundaries. Chemical Geology, 26, pp. 217-235. Piper, J.D.A. 2000. The Neoproterozoic Supercoentti:n Rodinia or Palaeopangea? Earth and Planetary Science letters, 176, 131 .- 146 Planck, T and Langmuir, C.H. 1988. An evaluatiof nth oe global variations in the major element chemistry of arc basalts. Earth and Palaryn eStcience Letters, 90, 349-370. Polat, A., Appel, P.W.U. and Fryer, B.J. 2011. oAvne rview of the geochemistry of Eoarchean to Mesoarchean ultramafic to mafic voiclc raoncks, SW Greenland: Implications for mantle depletion and petrogenpertiocc esses at subduction zones in the early Earth. Gondwana Research, 4-67. Prinsloo, M.C. 1987. Notas oor die geologie vaana biel 2917BA, BB, BC en BD. Report, Council for Geoscience (unpublished). Reid, D.L. 1974. Preliminary report on petrolosgtiuc dies of volcanic and intrusive rocks in the Vioolsdrif region, lower Orange River. IAn.: Kröner (editor) Contributions to the Precambrian geology of southern Africa. BullePtinre, cambrian Research Unit, University of Cape Town, 15, 57-68. Reid, D.L. 1975. Geochronology of the Haib-Vioorlisf digneous Province. Annual Report, Precambrian Research Unit, University opf eC Taown, 13, 49 - 60. Reid, D.L. 1977. Geochemistry of Precambrian igunse ro cks in the lower Orange River region. Bulletin (P.hD. thesis), Precambrian Rrecshe aUnit, University of Cape Town, 22, 397 p. Reid, D.L. 1979. Total rock Rb-Sr and U-Th-Pb oispoict study of Precambrian metavolcanic rocks in the Lower Orange River re.g ioEnarth and Planetary Science Letters, 42, 368-378. Reid D.L. 1981. Sm-Nd ages from the Namaqua ancdh tRerisveld Provinces. Abstracts, Geocongress (1981), Pretoria, Geological Socie tSyo oufth Africa, 173 - 174. Reid, D.L. 1982. Age relationships within the Vlisodorif Batholith, lower Orange River region II. A two stage emplacement history and e txhte nt of Kibaran overprinting. Transactions, Geological Society of South Afric5a,, 1805 - 110. Reid, D.L. 1997. Sm-Nd age and REE geochemist rPyr otferozoic arc-related igneous rocks in the Richtersveld Subprovince, Namaqua Mleo Bbei lt, Southern Africa. Journal of African Earth Sciences, 24(4), 621 – 633. Richard, J.P. and Kerrich, R. 2007. Adakite-liokeck rs: their diverse origins and questionable role in metallogenesis. Economic oGgeyo, l102, 537-576. Ritter U. 1980. The Precambrian evolution of thaest ern Richtersveld. Bulletin, Precambrian Research Unit, University of Cape To 2w6n, ,276 p. Ritter U. 1983. The Mid-proterozoic magmatic aencdt ot nothermal evolution of the Richtersveld. In: B.J.V. Botha (editor) Namaqunadl aMetamorphic Complex. Special Publication, Geological Society of South Africa,, 1011 - 122. Rogers, A.W. 1911. The Kheis Series. Transac,t iGonesological Society of South Africa, 13, 93 – 104. Rogers A.W. 1916. The nature of the copper desp osf iLt ittle Namaqualand. Proceedings, Geological Society of South Africa,, 2119 - 35. Rollinson, H.R. 1993. Using geochemical data: ueavtaiol n, presentation, interpretation. Longman Group UK, 352 p. Rollinson, H.R. 2005. Geodynamic controls on atdea, kTiTG and sanukitoid genesis: implications for models of crust formation. Editaol rcolumn, Special Issue, Lithos, 79, p. ix-xii. Ross, P.S., Bédard, J.H. 2009. Magmatic affinfi tmy odern and ancient subalkaline volcanic rocks determined from trace element dmisicnrai tion diagrams. Canadian Journal of Earth Sciences, 46, 823-839. SACS (South African Committee for Stratigraphy)8. 01.9 Stratigraphy of South Africa, part 1 (compiler LE Kent), Lithostratigraphy of tRhe public of South Africa, South West Africa/Namibia and the Republics of Bophuthatsw aTnra,nskei and Venda. Handbook, Geological Survey of South Africa, 8, 690 p. Smithies, R.H. and Champion, D.C. 2000. The Arcnh heiagh-Mg diorite suite: links to tonalite-trondhjemite-granodiorite magmatism andp limcations for early Archaean crustal growth. Journal of Petrology, 41, 1653-11.6 7 Sobolev, R. 1991. Rare earth minerals in concne potfio granite genesis. In: B. Siribumrungsukha, S. Arrykul, P. Sanguansai, T.g Prausnsanni, L. Sikong and K. Kooptarnoud (editors), Proceedings of the Interonnaatli Conference on Rare Earth Minerals and Minerals for Electronic Uses, Facuolft yE ngineering, Prince of Songkla University, Thailand, 662 p. Stowe, C.W., Hartnady, C.J.H. and Joubert, P. 1 9P8r4o.terozoic tectonic provinces of Southern Africa. Precambrian Research, 25, 22391 .- 2 Streckeisen, A. 1976. To each plutonic rock itosp per name. Earth Science Reviews, 12, 1-33. Strydom, D., Colliston, W.P., Praekelt, H.E., Schh,o Ac.E., Van Aswegen, G., Pretorius, J.J., Beukes, G.J., De Villiers, B., Watkeys, Ma.Kn.d Botes, F.J. 1987. The tectonic units of parts of Namaqualand, Bushmanland andh seornu tSouth West Africa/Namibia. Map (1:250 000), Bushmanland Research Project, rDtmepeant of Geology, University of the Orange Free State, one sheet. Thomas, R.J., Agenbacht, A.L.D., Cornell, D.H. & oMreo, J.M. 1994. The Kibaran of southern Africa: tectonic evolution and metallog.e Onyre Geology Review, 9, 131-160. Thompson, R.N. 1982. British Tertiary volcanic vpirnoce. Scottish Journal of Geology, 18, 49-107. Van der Merwe, S.W. 1986. The structural develonptm oef part of the Namaqua Mobile Belt in an area between Springbok and Vioolsdroifu, tSh Africa. Ph.D. thesis, University of the Orange Free State, Bloemfontein, 373 p. Van der Merwe, S.W. and Botha, B.J.V. 1989. Thoeo Gthroek thrust belt in western Namaqualand: an example of a mid-crustal struc t uSroeu.th African Journal of Geology, 92, 156 - 166. Van Tonder, D.M. and Mouri, H. 2010. Petrology agnedochemistry of the granitoid rocks of the Johannesburg Dome, Central Kaapvaaatol Cn,r South Africa. South African Journal of Geology, 113, 257-285. Von Backström, J.W. 1953. Geological map 2817Do Vlsidorif, 2818C Goodhouse, 2818D Dabenoris, 2819C Onseepkans (1:125 000),o Ggiecoall Survey of South Africa, one sheet. Von Backström, J.W. and De Villiers, J. 1972. Tgheeo logy along the Orange River Valley between Onseepkans and the Richtersveldp.l a Enaxtion, sheets 2817D (Vioolsdrif), 2818C (Goodhouse), 2818D (Dabenoarinsd) 2819C (Onseepkans), Geological Survey of South Africa, 101 p. Von Veh, M.W., 1993. The stratigraphy and struaclt uervolution of the late Proterozoic Gariep belt in the Sendelingsdrif – Annisfonteine aa,r orthwestern Cape Province. Bulletin, Precambrian Research Unit, UniversityC oafp e Town, 174 p. Ward, J.H.W. 1977. The geology of the area souf tVhi ooolsdrif, Cape Province. Internal Report, Atomic Energy Board, PEL-257, 4. 8 p Watkeys, M.K. 1986. The Achab Gneiss: a "floor "B iunshmanland or a flaw in Namaqualand? Transactions, Geological Societyo ouft hS Africa, 89, 103 - 116. Wyllie, P.J. 1983. Experimental and thermal coaninstrs on the deep-seated parentage of some granitoid magmas in subduction zones. In:. MAt.hPerton and C.D. Gribble (editors) Migmatites, Melting and Metamorphism. iv Sah, Nantwich, UK. APPENDIX 1 ANALYSES, SAMPLE LOCALITIES, DATA TREATMENT AND PROCESSING 1. DATA PROCESSING Sample classification was done using the totall ia-slkilaica (TAS) diagram of Le Maitre t al. (1989) for the volcanic rocks, and the R1-R2 rdaiamg of De La Roche t al. (1980) for the plutonic rocks. In the thesis, data from R(e1i9d7 7) and Ritter (1980) are also used and data as to their localities, descriptions annadly ases are to be found in those studies. Samples collected during the current study werely asenda at the laboratory of the Council for Geoscience in Silverton. REE were analysedIC bPyM S (results in ppb) while all other elements were analysed by XRF (results in) .p pm All results were recalculated to 100% volatile aCnrd2O 3 free. Wherever Cr is used in this study, the total ppm values from trace elem aennatlyses are used. Total iron contents received from the laboratory were calculated to FaenOd Fe2O3 following the equations of Le Maitre (19761), i.e.: Ox = 0.93 – 0.0042Si2O – 0.022(Na2O + K2O) for volcanic rocks Ox = 0.88 – 0.0016Si2O – 0.027(Na2O + K2O) for plutonic rocks Then Ox = FeO/Fe2O3T Thus FeO = Ox(Fe2O3T) And Fe2O3 = Fe2O3T – FeO 2. ANALYSES AND SAMPLE LOCALITIES The analyses and localities of samples collecter dth feo current study are given in the next table. A sample locality map is provided. 1 Le Maitre, R.W. 1976. Some problems of the prtoiojencs of chemical data into mineralogical classifications. Contributions to Mineralogy anedt rPology, 56, 181-189. Abreviations: UNIT VD = Vuurdood Subsuite KH = Khoromus Tonalite BW = Blockwerf migmatite Complex XX = Xaminxaip River Granodiorite GSh = Gaarseep Granodiorite in the Haib area GSr = Gaarseep granodiorite in the Richtersveld GSr(bw) = Gaarseep granodiorite in the Richters vinetlrduding the Blockwerf migmatite Complex HO = Hoogoor Granite GH = Ghaams Granite ST = Sout Granite NO = Nous Formation WV = Windvlakte Formation PR = Paradys River Formation AR = Abiekwa River Formation KR = Kook River Formation KU = Kuams River Formation ROCK TYPE PY = Pyroxenite Ol-G = Olivine gabbro G = Gabbro GN = Gabbro norite D = Diorite SD = Syenodiorite T = Tonalite QzMO = Quartz monzonite GD = Granodiorite MGR = Monzogranite SGR = Syenogranite AGR = Alkali granite BA = Basaltic andesite A = Andesite TA = Trachyandesite D = Dacite RH = Rhyolite Legend to localities map (Figure 1) Figure 1: Sample localities. Figure 1 (continue). Figure 1 (Continue). APPENDIX 2 NEW GEOCHRON DATA