Ilmenite deposits and their geological environment publication/Spec...Ilmenite deposits and their...

Ilmenite deposits and theirgeological environmentWith special reference to the Rogaland Anorthosite ProvinceIncluding a geological map at scale 1:75,000and a CD with a guide to the province

SPECIAL PUBLICATION 9

Leiv Eirikssons vei 39N-7491 Trondheim,Norway

Telephone +47 73 90 40 00Telefax +47 73 92 16 20

E-mail: [email protected]://www.ngu.no

Special Pu

blicatio

n n

o.9Ilm

enite d

epo

sits and

their g

eolo

gical en

viron

men

tN

GU

Geology for Society since 1858

The Rogaland Intrusive Massifs

An Excursion guide

A CD-presentation of theexcursion guidebook andadditional geological mapinformation,photographs and SEM images

Titanium and Iron-TitaniumDeposits in Norway

Approximately 300 titaniumand iron-titanium ore depositsare known in Norway; most ofthese occur in distinctProterozoic and Phanerozoicgeological provinces.

The Rogaland Anorthosite Province

(c.940-920 Ma) is highly important in termsof contained titanium and economicsignificance.

In a workshop in Rogaland in 2001,organised under the umbrella of theEuropean Science Foundation GEODE(Geodynamics and Ore Deposit Evolution)programme,geoscientists from manycountries came together to discuss variousaspects of the evolution of ilmenitedeposits and their host rocks.

This volume of the NGU Special Publicationseries is a follow-up of the 2001-conference,with articles and abstracts dealing with theRogaland Anorthosite Province as well aswith ilmenite deposits elsewhere in theWorld.Also included is a full-size geologicalmap of the province and a CD with ageological excursion guide.

NORGES GEOLOGISKE UNDERSØKELSESpecial Publication 9

ISSN:0801-5961ISBN:82-7385-108-7

RogalandAnorthosite Province

The NGU Special Publications series comprises consecutively numbered volumescontaining papers and proceedings from national and international symposia ormeetings dealing with Norwegian and international geology, geophysics andgeochemistry; excursion guides from such symposia; and in some cases papers ofparticular value to the international geoscience community or collections of thematicarticles.The language of the Special Publications series is English.

Editor:Forsker dr.philos.David Roberts

© 2003 Norges geologiske undersøkelsePublished byNorges geologiske undersøkelse(Geological Survey of Norway)N-7491 TrondheimNorway

All rights reserved

ISSN:0801-5961ISBN:82-7385-108-7

Graphic production:Skipnes AS

Cover illustration:Geological map of the Rogaland Anorthosite Province.The image was produced byJohn F.Dehls using geological data from Marker et al. (this volume) and topographicdata derived from the Norwegian mapping authority (Statens kartverk).Inset photograph shows the Tellnes open pit in 1964 (photo by Jean-Clair Duchesne).

I L M E N I T E D E P O S I TS A N D T H E I R G E O LO G I C A L E N V I R O N M E N T SPECIAL PUBLICATION 9 - 3

NORGES GEOLOGISKE UNDERSØKELSESPECIAL PUBLICATION NO. 9

Ilmenite deposits and theirgeological environmentWith special reference to the Rogaland Anorthosite ProvinceIncluding a geological map at scale 1:75,000 and a CD with a guide to the province

Edited by:JEAN-CLAIR DUCHESNE & ARE KORNELIUSSEN

Trondheim 2003

4 - SPECIAL PUBLICATION 9 I L M E N I T E D E P O S I TS A N D T H E I R G E O LO G I C A L E N V I R O N M E N T

Images from the Rogaland Anorthosite Province

Participants at the GEODE field workshop (2001) visiting the Piggstein anorthosite-norite localityin the Egersund-Ogna anorthosite. The inset photo shows a mega-crystal of orthopyroxene inanorthosite. For further information, see the excursion guidebook (loc. 1.2) available on the enclo-sed CD.

Modally layered norite within megacyclic unit III of the Bjerkreim-Sokndal layered intrusion atTeksevatnet (loc. 84, Fig. 1 in Schiellerup et al., this volume). Inset photo shows anorthosite xeno-liths in norite near the country rock contact of the Bjerkreim-Sokndal intrusion north of Bjerkreim.


ContentsPREFACE ................................................................................................................ 9

SHORT PAPERSGenesis of magmatic oxide deposits - a view from the Bushveld Complex: Cawthorn, G. .................................................................... 11

Titanium deposits in Ukraine focussed on the Proterozoic anorthosite-hosted massifs: Gursky, D., Nechaev, S. & Bobrov, A. .................. 21

Links between the Proterozoic anorthosite-rapakivi granite plutons and ore-forming events in the Ukrainian Shield (ores of titanium, uranium, rare metal and gold):Nechaev, S. & Pastukhov, V. ................................................................................ 27

APPROACHES AND PROCESSESDo Fe-Ti oxide magmas exist? Geology: Yes; Experiments: No!: Lindsley, D.H. .............................................. 34

Some results on the role of P, T and fO2 on ilmenite composition:Vander Auwera, J., Longhi, J. & Duchesne, J.C. ................................................ 35

The nelsonite problem: the origin by melt immiscibility:Kozlowski, A. & Wiszniewska, J. ........................................................................ 37

On ilmenite and rutile ore provinces in Norway, and the relationships between geological process and deposit type:Korneliussen, A. .................................................................................................. 40

Ilmenite–zircon relationships in meta-anorthosites.Example from the Bergen Arc (Caledonides of W Norway) and implications for zircon geochronology:Bingen, B., Austrheim, H. & Whitehouse, M. .................................................... 41

Evidence from Re-Os for the origin of sulphide concentrations in anorthosites: Hannah, J.L. & Stein, H.J. ...................................................... 42

Effect of mineralogy and texture of sand and hard rock ilmenite in TiO2 pigments production by the sulfate process,a case study on Australian ilmenite concentrate and Tellnes ilmenite concentrate, Norway: Chernet, T. .......................................... 45

Parallel Re-Os isochrons and high 187Os/188Os initial ratios:constraints on the origin of sulphide-bearing anorthosite systems: Stein, H.J., Hannah, J.L., Morgan, J.W. & Scherstén, A. ...................... 47

FENNOSCANDIASveconorwegian rejuvenation of the lower crust in South Norway:Andersen, T., Griffin, W.L., Pearson, N.J. & Andresen, A: .................................. 51

Thermal modeling of the metamorphism of the Rogaland granulites: Schumacher, J.C. & Westphal, M. .................................. 53

Re-Os dating of the Ørsdalen W–Mo district in Rogaland,S Norway, and its relationship to Sveconorwegian high-grade metamorphic events: Bingen, B. & Stein, H. ................................ 55

Tectonic setting of the Egersund-Ogna anorthosite,Rogaland, Norway, and the position of Fennoscandia in the Late Proterozoic: Brown, L., McEnroe, S. & Smethurst, M. ........................ 58

Sulfides in the Rogaland Anorthosite Province:Schiellerup, H., Lambert, D.D. & Robins, B. ...................................................... 58


Fe-Ti-V-P deposits in anorthosite complexes:the bearing of parental magma composition and crystallization conditions on the economic value:Duchesne, J.C. & Vander Auwera, J. .................................................................. 60

Whole-rock geochemistry of the Bjerkreim-Sokndal layered series: bearing on crystallization processes of cumulus rocks:Charlier, B. & Duchesne, J.C. ............................................................................ 62

Relationship between the layered series and the overlying evolved rocks in the Bjerkreim-Sokndal intrusion,Southern Norway: Wilson, J.R. & Overgaard, G. .............................................. 64

SEM elemental mapping and characterisation of ilmenite,magnetite and apatite from the Bjerkreim-Sokndal layered intrusion: Gautneb, H. .......................................................................... 65

LA-HR-ICP-MS studies of ilmenite in MCU IV of the Bjerkreim-Sokndal intrusion, SW Norway: Meyer, G.B. & Mansfeld, J. .......................... 68

Geochemistry and mineralogy of the Tellnes ilmenite ore body: Kullerud, K. ............................................................ 70

Emplacement mechanism of the Tellnes ilmenite deposit (Rogaland, Norway) revealed by a combined magnetic and petrofabric study: Bolle, O., Diot, H., Lambert, J.M.,Launeau, P. & Duchesne, J.C. .............................................................................. 71

Could the Tellnes ilmenite deposit have been produced by in-situ magma mixing?: Robinson, P., Kullerud, K.,Tegner, C., Robins, B. & McEnroe, S. .................................................................. 74

Preliminary melting experiments on the Tellnes ilmenite norite from 0.5 to 1.2 GPa, implications for the composition of intercumulus melt: Skjerlie, K.P., Kullerud, K. & Robins, B. ........................ 77

Concentration of ilmenite in the late Sveconorwegian norite-anorthosite Hakefjorden Complex, SW Sweden: Årebäck, H. .............. 77

Gabbro-hosted ilmenite deposits in Finland: Kärkkäinen, N. ........................ 80

THE GRENVILLE PROVINCE IN CANADAReview of Fe-Ti±V and Fe-Ti-P2O5±V deposits associated with anorthositic suites in the Grenville Province, Québec:Perreault, S. & Hebert, C. .................................................................................. 83

A study of mineral compositions of the Lac Mirepoix layered complex, Lac St-Jean Anorthosite Complex (Québec, Canada):Morisset, C.E. ...................................................................................................... 84

The 974 Ma Vieux-Fort Anorthosite, Lower North Shore,Québec: the youngest anorthosite in the Grenville Province:Perreault, S. & Heaman, L................................................................................... 86

Contrasting styles of Fe-Ti mineralization in the Havre-Saint-Pierre anorthosite suite, Quebec’s North Shore, Canada: Perreault, S. ...................... 87

Cambrian titaniferous paleoplacers metamorphosed during the Taconian Orogeny, Quebec Appalachians: Gauthier, M. ............................ 91

AREAS IN EASTERN EUROPEIlmenite deposits and mineralization in alkaline and subalkaline magmatic complexes of the Ukrainian Shield:Kryvdik, S.G. ...................................................................................................... 94

The gabbro-anorthosite massifs of Korosten Pluton (Ukraine) and problems concerning the evolution of the parental magmas: Mitrokhin, A.V. ...................................................................... 96


First approach to the petrology of the Kamenka peridotite-gabbro-anorthosite intrusion: Shumlyanskyy, L. ............................ 97

Ti, Mg, and O isotopic compositions in ilmenites from Ukrainian ore deposits as indicators of their crystallization conditions: Didenko, P.I., Kryvdik, S.G. & Zagnitko, V.M. ................................ 99

Ti, V, Pt, Pd and Au in Travyanaya Bay ore peridotites and their possible genetic relation with Belomorian Mobile Belt anorthosites: Stepanov, V.S. .......................................................... 101

Precambrian anorthosites in the Belomorian Mobile Belt,Eastern Fennoscandian Shield: Stepanov, V.S. & Stepanova, A.V. .................... 102

Subduction-related leucogabbro-anorthosite ilmenite-bearing series: an example of water-rich high-temperatureanatexis, platinum belt of the Urals, Russia: Fershtater, G.B. ........................ 104

The Riphean layered intrusions of the Western Urals and related ilmenite-titanomagnetite deposits:Fershtater, G.B. & Kholodnov, V.V. ...................................................................... 107

ROGALAND ANORTHOSITE PROVINCE – SYNTHESESAn introduction to the geological map of the Rogaland Anorthosite Province 1:75,000:Marker, M., Schiellerup, H., Meyer, G., Robins, B. & Bolle, O. ........................ 109

Mineral resources in the Rogaland Anorthosite Province,South Norway: origins, history and recent developments:Schiellerup, H., Korneliussen, A., Heldal, T., Marker, M.,Bjerkgård, T. & Nilsson, L.P. .............................................................................. 116

Enclosures - A paper print of the geological map of the Rogaland Anorthosite Province- A CD (Guideboook + map + other information interactive)


Recommended bibliographical reference to this volume:Duchesne, J-C. & Korneliussen, A. (eds.) 2003:Ilmenite deposits and their geological environment.Norges geologiske undersøkelse Special Publication 9, 134 pp.ISBN: 82-7385-108-7

Map reference:Marker, M., Schiellerup, H., Meyer, G.B., Robins, B. & Bolle, O.: Geological map of the Rogaland Anorthosite Province.Norges geologiske undersøkelse Special Publication 9.

Individual contributions:Fershtater, G.B. & Kholodnov, V.V. 2003: The Riphean layered intrusions of the Western Urals and related ilmenite-titanomagnetitedeposits. (Extended abstract). Norges geologiske undersøkelse Special Publication 9, 107-108.


The Rogaland Anorthosite Province isfamous for its anorthosite massifs, possiblythe most accessible in the world with be-autiful, nearly continuous outcrops, itsnumerous ilmenite deposits, which inclu-de a world-class deposit (Tellnes), and thelargest layered intrusion in Europe (theBjerkreim-Sokndal body). These geologi-cal wonders have been extensively studiedfor the last 50 years following the impulseof Paul Michot (1902-1999) who compiledthe first geological map and suggested aninterpretation of the geological evolution.Rogaland, and particularly the Sokndaldistrict, is also famous for its mining acti-vities, not to forget that it was the cradle ofthe sulphate processing of the ilmenite ore.

The purpose of the present volume is thr-eefold: 1. to present a collection of papersand abstracts given at a symposium onilmenite deposits and their geological envi-ronment; 2. to publish a geological map ofthe Rogaland Anorthosite Province at scale1:75,000; 3. to make available a CD with ageological guide-book to the province.

A field-workshop "Ilmenite deposits in theRogaland Anorthosite Province" was heldat Moi, July 8-12, 2001, as part of theEuropean Science Foundation’s GEODEproject "The Fennoscandian Shield Pre-cambrian Province". The purpose of thisworkshop was to bring together geo-scientists interested in the genesis of thevarious types of mineral deposits withinthe overall evolution of the anorthosites(in Rogaland and elsewhere), with parti-cular reference to their geochemical com-position. The occurrence of a world-classdeposit of ilmenite in the area – theTellnes mine, run by Titania A/S -– andalso the application of new geochemicalconstraints to the industrial quality ofthe ore further motivated organising ameeting where the scientific and in-dustrial worlds could interact and com-pare their experience. Fifty-five partici-pants coming from 14 countries attendedthe workshop. Forty-four communicati-ons (oral and posters) were presented in 2days. Four days were devoted to field trips.We are particularly happy to present here,besides the abstracts of the communicati-ons, a general paper dealing with the gene-

sis of oxide deposits in layered intrusionsby G. Cawthorn, and two papers on thedeposits of the Ukrainian Shield – D.Gursky, S. Nechaev & A. Bobrov and S.Nechaev & V. Pastukhov, which synthesisefor the western scientists a large amountof literature in Russian and Ukrainian.

The geological map at the scale 1:75,000 ofthe Rogaland Anorthosite Province whichis included in this volume has been com-piled, revised and locally extended by M.Marker, H. Schiellerup, G. B. Meyer, B.Robins & O. Bolle after the original workof the several groups of geologists whohave studied the area. The German groupwas led by H. Krause from Clausthal.University, the Norwegian group by B.Robins from the University of Bergen, theDanish group by J.R. Wilson from theUniversity of Aarhus, and the Belgiangroup, following Paul Michot and his sonJean, by J.C. Duchesne from the Universityof Liège. Some data by the Dutch group ofthe University of Utrecht have also beenincluded. Ø. Nordgulen (NGU) is parti-cularly thanked for having initiated thepublication project as early as 1996.

The excursion guide-book entitled "TheRogaland Intrusive Massifs – an excursionguide" (J.C. Duchesne, editor) NGUreport 2001.29, 2001, 137 pp., which ispresented on a CD, is the second revisedand enlarged version of parts of "Geologyof Southernmost Norway, an excursionguide" (C. Maijer & P. Padget, editors)Norges geologiske undersøkelse, SpecialPublication 1, 1987. The first version waspublished to illustrate a NATO AdvancedScience Institute held at Moi in 1984 onthe topic "The deep Proterozoic crust inthe North Atlantic Provinces" (A.C. Tobi& J.L.R. Touret, editors) NATO ASI Series,C 158, Reidel, Dordrecht, 1985, 603 pp.The second version was prepared for ameeting held in Rogaland in 1992 on"Magma chambers and processes in anor-thosite formation" in the framework ofIGCP programme 290 on "Anorthositesand related rocks". Several localities des-cribed by Paul Michot in the very firstgeological guide-book on the region(Norges geologiske undersøkelse, 212g,1960) are still visited in the present versi-

Preface

1 0 - SPECIAL PUBLICATION 9 I L M E N I T E D E P O S I TS A N D T H E I R G E O LO G I C A L E N V I R O N M E N T

on. The guide-book, which results fromthe interaction between several groups ofresearchers, is divided into two parts: the-matic papers introduce the reader to thestate of the art on the knowledge of thevarious units, summarising decades ofresearch up to 2001; then itineraries aredescribed to show salient field exposuresof relatively easy access together withpetro-geochemical information on theexposed rocks. A map showing the itine-raries and the localities on the geologicalbackground permits the reader to easilymove to the appropriate section of thetext.

This volume, which offers an up-to-datesynthesis on our knowledge of theRogaland Anorthosite Province, its geo-logy, petrology and metallogeny, wasmade possible through the cooperationand interaction between numerous per-sons with various specialities and inter-ests, from humble students to experiencedscientists. The editors wish to thank all ofthem for their contributions. They alsoacknowledge D.J. Blundel, Chairman ofthe Steering Committee of the GEODEprogramme, A. Bjørlykke, Director of theGeological Survey of Norway (NGU), andR. Hagen, Titania A/S, for encouragement,help and financial support. Thanks also goto NGU secretary Tove Aune for her inva-luable help in putting together this volu-me.

J.C. Duchesne and A. Korneliussen (23 03 2003)

I L M E N I T E D E P O S I TS A N D T H E I R G E O LO G I C A L E N V I R O N M E N T SPECIAL PUBLICATION 9 - 1 1

IntroductionThe Bushveld Complex (Fig. 1a) is anextremely large lopolith, covering approx-imately 65,000 km2, and up to 8 km thick(see summary by Eales & Cawthorn 1996).Of all layered intrusions, it shows the mostextreme range of differentiation (Fig. 1b).The first formed cumulates (containingolivine Fo90) suggest crystallization froman almost primitive magma, and the evol-ved cumulates contain olivine app-roaching Fo0. The rocks range from duni-te and pyroxenite at the base to ferrodiori-te at the top, and contain layering (some-times monominerallic) of olivine, ortho-pyroxene, plagioclase, chromite, magne-tite and apatite.

The Bushveld Complex contains over half the world's chromium and vanadiumdeposits, contained in numerous oxide-rich layers and, in the case of vanadiferousmagnetite, discordant pipes. Over 80% ofthe world's platinum-group elements alsooccur here, with about half of this propor-tion occurring in the uppermost thickchromitite layer. Chromitite layers can be traced for many tens of km laterally inboth eastern and western limbs of theintrusion (Fig. 1a). For the lower twopackages of layers (referred to as theLower Group 1-7 and the Middle Group

Short papersGenesis of magmatic oxide deposits - a view from theBushveld Complex

Grant Cawthorn

Cawthorn, G. 2003: Genesis of magmatic oxide deposits – a view from the Bushveld Complex. Norgesgeologiske undersøkelse Special Publication 9, 11-21.

The Bushveld Complex contains 14 and 24 significantly thick layers of chromitite and vanadiferous mag-netitite respectively, that can be traced for very many tens of kilometres. Their genesis is still enigmatic.For all models, unresolved problems and contradictory evidence still exist. For chromitite layers, a modelinvolving pressure changes is the most appropriate for explaining the enormous lateral continuity andsharp boundaries. However, the initial 87Sr/86Sr of silicate gangue in chromitite layers is anomalouslyhigh and indicates mixing with a siliceous component (roof melt?). For the magnetitite layers, extremegradients of Cr content in vertical sections preclude crystal settling from a large volume of magma, andin situ growth mechanisms may have been operative. The discordant magnetitite pipes may have formedby flushing of a different magma through an unconsolidated crystal pile, feeding into the chamber, withsimultaneous dissolution of the host and precipitation of only the liquidus mineral (magnetite). An ori-gin for lenses and bifurcations of silicate gangue in oxide layers is proposed, in which the silicates are con-sidered to be a primary accumulation during a period of non-deposition of oxide, and not due to secon-dary remobilization or injection processes.

G. Cawthorn, School of Geosciences, University of the Witwatersrand, Private Bag 3, Wits 205, South Africa.

Fig. 1. (a). Simplified map of theBushveld Complex showing thelocations of the chromitite andvanadiferous magnetitite layers.The location of the platinummineralization is coincident withthe uppermost chromitite layerin the intrusion. Structural com-plications and poor outcrop inthe northern part of the easternlimb make exposure of themagnetitite layers discontinuous.(b). Simplified stratigraphic secti-on through the BushveldComplex, showing the presenceof cumulus minerals and theircompositions. Sources of dataare summarized in Eales &Cawthorn (1996).


1-4) thicknesses remain constant for longdistances, but abrupt changes in thicknessmay occur (Schurmann et al. 1999), usual-ly related to structural features (eitherfolds in the floor or lineaments, that divi-de the lowest succession into smaller com-partments). The layers in the topmostpackage (the Upper Group 1-3) are re-markably uniform in thickness for over100 km in each limb. Chemical similarityis also very uniform even across the entireintrusion (Cawthorn & Webb 2001).Chromitite layers typically occur at thebase of cyclic units in the Critical Zone,where such can be identified by changes inthe silicate mineralogy. Cycles typicallyconsist of a basal chromitite, pyroxenite orharzburgite, norite and anorthosite.Contacts of the chromitite are generallysharp, but the Lower Group 6 chromititehas gradational contacts top and bottomover about 20 cm. These major layersrange from tens of cm to 1 m in thickness,but many thinner layers (some even less 1cm) also occur, and can be equally lateral-ly extensive. They contain from 60-90%chromite, cemented by oikocrysts of pyro-xene and/or plagioclase.

The magnetitite layers are even more late-rally extensive and uniform in thicknessthan the chromitite layers, being traced forover 100 km in each limb. Even such fea-tures as the 'feldspar parting', a 20 cm-wide zone that contains about 15% cumu-lus plagioclase, exactly in the middle of the2 m-thick Main Magnetitite Layer, areuniversally present. Almost all layers in theeastern limb have sharp basal contactswith anorthosite, and gradational topcontacts, although one layer displays thereverse relationships (Molyneux 1974).

They range in thickness from a few cm upto tens of cm, with two exceptions. TheMain Magnetitite Layer (fourth from thebottom) is 2 m thick, and the uppermostlayer is up to 6 m thick, but containsnumerous autoliths and individual grainsof silicate material. The oxide content canrange up to 95% titanomagnetite, withinterstitial altered plagioclase.

A large number of erratically distributed,discordant pipes, trending orthogonally tothe layering, are known (Scoon & Mitchell1994). They range from almost pure mag-netitite to coarse-grained, magnetite-richmelagabbro. The largest is over 300 macross and has been drilled to a depth of300 m, but the total vertical continuity ofthese pipes cannot be assessed.

Chromitite layeringThe various models that have been propo-sed in the formation of chromitite layersinclude (in very approximate chrono-logical order):1. crystal settling and sorting (Sampson

1932);2. immiscibility of a Cr-rich liquid (Sam-

pson 1932, McDonald 1967);3. increase in oxygen fugacity (Cameron

& Desborough 1969);4. contamination by siliceous compo-

nents (Irvine 1975);5. mixing between resident and new

magma (Irvine 1977);6. lateral growth within a stratified

magma column (Irvine et al. 1983);7. pressure change (Osborn 1980);8. injection of a chromite-porphyritic

magma (Eales 2000).Each model explains certain aspects of thegeology or magmatic behaviour, but equ-ally valid reasons can be presented as towhy each process can be questioned.

A quantitative assessment of the envisagedprocesses, specifically a mass-balance eva-luation, presents problems. Chromititescontain about 45% Cr2O3 or 300,000 ppmCr. Mafic magmas may contain between100 and 1,000 ppm Cr. Hence, even ifevery atom of Cr could be extracted froma magma, between 300 and 3,000 units ofmagma would be needed to make one unitof chromitite. For example, allowing forthe density of chromitite being 1.5 timesthat of a magma, a one metre-thick layerof chromitite would need between 300 and

Fig. 2. Phase diagrams in the system forste-rite - quartz - chromite. (a). Magma conta-mination model after Irvine (1975). (b).Magma mixing model after Irvine (1977).


3,000 metres of magma. However, the sup-position that some process could removeevery ion of Cr from the magma is obvi-ously invalid, and so very much greaterthicknesses of magma need to be involved.It should be recalled also that most layersof chromitite can be traced for many 10s ifnot 100 km in the east and west Bushveld,and could extend for 300 km if the princi-ple of connectivity of the two limbs isaccepted (Cawthorn & Webb 2001).Processes that can operate on such hugevolumes of magma need to be evaluated.

If crystal settling has taken place, thenStokes Law, with or without the com-plexities of non-Newtonian behaviour ofmagma, needs to be applied. Althoughchromite has a high density, its small grainsize relative to pyroxene and olivine makesis unlikely to sink more rapidly than themafic minerals. Hence, producing a chro-mitite layer at the base of a cyclic unit,overlain by ultramafic layers, is not consis-tent with Stokes Law. Further, with refe-rence to Fig. 2, the olivine to chromiteratio along the cotectic ranges from 98 to2 (for point A) to about 99.6 to 0.4 (atpoint B). To make a 1 m-thick chromititelayer from a magma at the cotectic wouldrequire the implausible situation of main-taining 50 to 250 m of olivine crystals insuspension. Hence, the question in produ-cing a chromitite layer can be defined as:why were there no silicate minerals gro-wing and sinking while the chromite wasaccumulating?

The concept of Cr-rich immiscible liquidswas developed by Sampson (1932) partlyto explain the bifurcations exhibited bychromitite layers, as seen for example atthe Dwars River locality (see below), whereit was suggested that a dense, Cr-richliquid was injected downwards into a pla-gioclase crystal mush. However, no experi-mental duplication of liquid immiscibilityfor Cr-rich liquids has ever been achieved.

The stability of oxide phases dependsupon the ferric ion activity in the magma,and hence an increase in fO2 would causechromite (and magnetite) precipitation.Upward percolation of fluid has been sug-gested as the source of the oxygen, butwhether such percolation could occurover such laterally extensive distances toproduce uniformly thick layers of chromi-tite (and magnetitite) seems dubious.

Also, fO2 changes would not explain thecommon pattern of cyclic units in whichanorthosite occurred below and pyroxeni-te above chromitite layers.

Another model for chromitite formationis that of magma mixing, and will be dis-cussed in some detail here. In its earliestform, Irvine (1975) suggested that conta-mination of magma (composition A inFig. 2a) by a siliceous component (F)could have lowered the solubility limit ofCr in a mixed magma (M), and so inducedchromite separation. In support of thismodel, the initial 87Sr/86Sr values for inter-stitial silicate minerals in several chromiti-te layers have recently been reported(Schoenberg et al. 1999). Their results areshown in Fig. 3. The high ratio observedin the chromitite layer itself compared toits footwall suggests that massive degreesof crustal contamination occurred. How-ever, what is more remarkable is that thehangingwall rocks to the layers have aratio that does not show evidence of thiscontamination. The implication is that thechromitite formed from a large volume ofcontaminated magma, perhaps generatedby mixing with molten felsic roof rocks,but that the overlying pyroxenite formedfrom a magma that did not show thiscontamination effect.

A further variation of the magma-mixingmodel was introduced by Irvine (1977), inwhich mixing between differentiated mag-ma (B or C in Fig. 2b) and primitive mag-ma (A) caused the mixed magma (M orM1) to lie in the chromite stability field.Excess chromite formed, producing thelayers. These models have the advantageover the crystal settling and sortingmodel, in that it explains why the magmalies in the chromite stability field, and sodoes not crystallize any mafic minerals.The effect on the Cr budget of the magmaas this process occurs can be seen in Fig. 4.The degree of supersaturation withrespect to Cr of the mixed magma in-creases as the difference in MgO contentsof the mixing magmas also increases. Itwas suggested (Murck & Campbell 1986)that near the base of large layered intrusi-ons chromitite layers are not developed,because the resident magma is not verydifferent from the added, primitivemagma. However, even for magmas thatare very different, the extent of super-saturation may not exceed 100 ppm Cr

Fig. 3. Plot of the initial 87Sr/86Sr in theseveral chromitite layers compared to thetypical ratio throughout the Critical Zone(from Schoenberg et al. 1999). The anoma-lously high value in the chromitite layersuggests a very large degree of contami-nation.

Fig. 4. Plot of Cr versus temperature ofmagma saturated with respect to chromi-te (Murck & Campbell 1986). Mixing lineDP shows the oversaturation in chromiteinduced by mixing. M-M1 shows theexcess Cr available to produce chromite.


(i.e. difference between M and M1 in Fig.4). If this figure represents the extractableCr as chromite prior to silicate saturation,a one m-thick chromitite layer wouldrequire the processing of 4,500 metres ofmagma.

The mineral compositions can also give aclue as to the plausibility of this process.In the case of the Bushveld Complex, theparent magma (point P in Fig. 4) con-taining 13% MgO would crystallize ortho-pyroxene with an mg# of 90. The moreevolved magma crystallizing plagioclase inthe Upper Critical Zone (point D in Fig. 4)containing about 7% MgO would crystall-ize orthopyroxene mg# 82. Mixing ofthese two magmas ought to have prod-uced a mixed magma that crystallizedorthopyroxene with mg# about 86. Thus,magma addition ought to be recognizableby reversals in mineral composition.Inspection of the mg# below and abovethe economically most important andthickest layer - Lower Group 6 chromititelayer shows that the composition actuallyshows a decrease in mg#, not a reversal(Fig. 5). The data in Fig. 5 include cumu-lus orthopyroxene in the footwall andhangingwall, and also intercumulusorthopyroxene in the chromitite layer.Within the chromitite layer there is anincrease in mg#. However, this increase is

not a primary magmatic effect indicatingaddition of primitive magma. It is asecondary effect, due to re-equilibrationbetween orthopyroxene and chromite,causing the orthopyroxene grains to incre-ase in mg#. The data of Cameron (1980,1982) show that, across all chromitite lay-ers, there is a change in mg# in orthopyro-xene of less than 1 (with an increase abovethe layer being less common than a decre-ase), changes that do not support the con-cept of addition and mixing of undiffe-rentiated magma.

There is a second aspect that also needs tobe considered, namely the small-scale fieldrelations. A very few chromitite layers havegradational contacts, but most have verysharp contacts (Fig. 6a). Any mixing pro-cess would have to be instantaneous inorder to produce such a sharp contact.The mixing of such huge thicknesses ofmagma over such large horizontal distan-ces so effectively seems extremely difficultto envisage.

The lateral growth model, proposed byIrvine et al. (1983) requires that crystall-ization only occur at the margins of a stra-tified magma column. This model is a spe-cial case of in situ growth. If in situ growthis the mechanism by which these layersform, it becomes necessary for Cr to diffu-se or be transferred to the base or edge ofthe magma column to sustain chromitegrowth. Transport of such huge quantitiesof Cr as would be needed to produce a onem-thick layer of chromitite seems implau-sible.

Yet another model was originally pro-posed by Osborn (1980) and has been ela-borated upon by Lipin (1993). In thismodel, an increase in pressure occurred inthe magma chamber to initiate chromiteprecipitation. The process can be seen interms of phase diagrams for the same sys-tem but at different pressures (Fig. 7). Thestability field of chromite increases withpressure relative to pyroxene and plagio-clase. A magma may lie at the three-phaseinvariant point (point C) at low pressure,but an increase in pressure would causethe same magma to lie in the chromitefield. Demonstration of pressure changesin magma chambers is hard to prove.However, the fountaining of basic lava,followed by draining of lava lakes over aperiod of days in active volcanoes demon-

Fig. 6. Nature of contacts and detailed fieldrelations of oxide-dominant layers. (a).Extremely sharp contacts of chromitite lay-ers in anorthosite. The sharp, but somew-hat cuspate, upper contact is a commonfeature. (b). Typical contacts of magnetititelayer, showing sharp base and gradationaltop, in both cases with anorthosite. (c).Detail of lenses and bifurcations withinchromitite layers from Dwars River localityof Upper Group 1 chromitite layer.

Fig. 5. Plot of mg# in orthopyroxene across theLG6 chromitite layer. Data are taken fromCameron (1980). Open circles refer to cumulusorthopyroxene, solid circles to interstitialorthopyroxene in chromitite layers in whichthere has been re-equilibration between thephases in terms of Mg and Fe.


strates that pressure can fluctuate at leastunder volcanoes (Pollard et al. 1983). Thephysical mechanism operative within alarge magma chamber is harder to post-ulate. However, if magma is added to thechamber, the sides must expand or thevertical height must increase. In the lattercase, the pressure at the base of the cham-ber must also increase. Alternatively, if theextra load of added magma causes thefloor to collapse (Carr et al. 1994), themagma at the base would be subjected toan increase in pressure. If the magmachamber were sealed, an attempt to addmagma would result in a transmission of apressure wave without actual magmaaddition since liquids are relatively incom-pressible. Whether a sufficient column ofmagma could be affected remains specula-tive, again emphasizing the mass balanceproblem. However, pressure changes canbe transmitted everywhere essentiallyinstantaneously within the chamber, andso crystal-sharp boundaries to layers andlateral uniformity could be envisaged.

It has also been suggested that addedmagma carried a suspension of chromitecrystals, which produced a chromitite layer(Eales 2001). However, given the density ofchromite it is hard to envisage the suspen-ded chromite being uniformly depositedlaterally, at great distances from proposedfeeder sites. This model also begs the ques-tion as to what process in the deepermagma chamber or source area permittedthe formation of such a huge mass of chro-mite. If it could happen in a deep magmachamber, why could not the same processoccur in the Bushveld chamber?

In conclusion, it appears that either thereare reasons why the apparent contrad-ictory evidence presented here for the cur-rently proposed models can be ignored, orthat some other processes need to be iden-tified to explain chromitite layering.

MagnetititesWhereas most occurrences of magnetititesare as layers in the Upper Zone of theBushveld Complex, pipes or almost puremagnetite up to 300 m in diameter andrarely narrow dykes, are also recordedfrom both the Upper and the Main Zones.Similar models to those for chromititeformation have also been applied to thegeneration of magnetitite layers. Some of

the models may be questioned on similargrounds to those listed above. Other pro-cesses and problems are summarisedbelow.

Unlike the experimental rebuttal of theexistence of Cr-rich immiscible liquids,iron-rich immiscible liquids have beensynthesized and are petrographically iden-tifiable (Philpotts & Doyle 1983). How-ever, applicability to the origin of oxidelayering is questionable. If a dense iron-rich liquid did form, it ought to sink intothe underlying plagioclase crystal much,and not form a regular planar base (Fig.6b). Further, if this dense liquid did accu-mulate into a planar body, it would not bepossible for plagioclase crystals to sinkinto it to form a gradational upper contact(Fig. 6b), that is characteristic of most lay-ers. Further, the composition (Table 1) ofsuch immiscible liquids has been shownby Philpotts & Doyle (1983) to containonly 30% FeO (total). It is hardly of amagnetitite composition, and so stillrequires a large proportion of silicate mat-erial to be removed from it to become a

pure magnetitite layer. Finally, the compo-sition of the silicate magma at whichimmiscibility occurs is highly evolved,containing little MgO and so being inca-pable of forming the composition ofmafic minerals observed in associationwith magnetitite layers. For example, theorthopyroxene at the level of the lowestmagnetitite layers has an mg# of 60 (Fig.1b). Iron-rich immiscible magmas alsostrongly partition phosphorus relative tothe silicate magma (see Table 1). It is forthis reason that nelsonites, which are

Fig. 7. Schematic phase relations in thenatural system olivine-plagioclase-quartzat two different pressures, illustrating theenlargement of the chromite stability fieldwith increasing pressure (modified fromOsborn 1980).

Natural Example Experimental Analogue

Pre- Si-Rich Fe-Rich Pre- Si-Rich Fe-Rich

immiscibility Liquid Liquid immiscibility Liquid Liquid

SiO2 79.1 73.3 35.6 58.8 72.3 44.3

TiO2 0.1 0.2 3.8 1.7 0.3 2.1

Al2O3 10.3 11.9 2.2 10.2 11.7 7.3

FeO(T) 1.7 2.6 33.7 16.6 6.7 28.6

MgO 0.2 0.0 0.5 0.6 0.0 1.0

CaO 0.8 1.4 10.1 5.4 3.8 9.5

Na2O 2.6 3.0 0.0 2.9 2.9 1.0

P2O5 0.0 0.1 3.0 0.7 0.4 0.9

Note: the column headed 'pre-immiscibility' indicates the composition of the liquid at a temperatu-re immediately above that at which liquid immiscibility appeared (Philpotts & Doyle, 1983).

Table 1. Compositions of immiscibleliquids (natural and experimental exam-ples).


oxide-apatite rocks, have been consideredto result from liquid immiscibility. Thishypothesis has also been applied to themagnetitite layers of the Bushveld Com-plex by Reynolds (1985). However, itshould be noted that the lower layers ofmagnetitite contain no apatite at all, andthat the apatite-bearing layers only appearabove the level at which apatite becomes acumulus mineral in the associated silicatelayers (Cawthorn & Walsh 1988). Thus,the appearance of apatite with magnetitedoes not suggest liquid immiscibility, butsimply the normal progression of magma-tic fractionation.

As with chromitite layer contacts, thebases of magnetitite layers are extremelysharp (Fig. 6b), although upper contactsare often gradational (Fig. 6b). The effica-cy of magma mixing to produce suchabrupt contacts is questionable. TheUpper Zone magma displays a remarkablyconstant and anomalously high initial Srisotope ratio (Kruger et al. 1987). (Anal-ogous Sr isotope data to those presentedin Fig. 3 for the interstitial components ofchromitite layers have not been compiledfor magnetitites, because the plagioclaseshows significant alteration in surface andshallow borecore samples, but such datawould be critical in view of the im-plications for chromitite genesis.) TheUpper Zone magma is inferred to be themixture from several partially crystallizedmagmas, all with different isotopic ratios,that produced the preceding zones. It isunlikely that a new magma could have hadexactly the same ratio as this complexlymixed resident magma. Hence, the con-stancy of initial ratio throughout the enti-re Upper Zone argues against addition ofsignificant volumes of distinct magma.Analyses of plagioclase in the footwall andhangingwall of several layers, shown inFig. 8, suggest that they are related by dif-ferentiation and not addition of magma.Since these rocks are quite evolved (An55-

60) addition of undifferentiated magmaought to produce significant reversals incomposition, but they are not observed.Reversals in the An content of plagioclasehave been identified in the Upper Zone(Christian Tegner, unpublished research),but they are not related to the presence ofmagnetitite layers. For example, a reversaloccurs between magnetitite layers, notacross a magnetitite layer, in Fig. 8.

There is a fairly regular upward decreasein the V content of the magnetite (Fig. 9).Again, if undifferentiated magma, withhigh V content, had been added to the sys-tem, reversals or irregular trends might beexpected.

Cr is highly compatible in magnetite (thepartition coefficient is in the order of hun-dreds) and is a sensitive indicator of frac-tionation. Its abundance in magnetite(Fig. 10) decreases extremely rapidly inshort, vertical sections through magne-titite layers, and also shows abrupt rever-sals even within layers. The rate of depleti-on is not consistent with crystal settlingfrom a large volume of magma, whichwould have produced either homogene-ous layers or subdued depletion profiles inCr. The rapid depletion profiles wereinterpreted by Cawthorn & McCarthy(1980) to be the result of in situ growth.This rapid depletion in the Cr content inthe magma seems to contrast with themass balance calculations present abovefor chromitite layers, where enormousvolumes of magma need to be processed.Despite their apparent similarities, itseems that chromitite and magnetitite lay-ering may not originate by identicalmechanisms. The occurrence of reversalsin Cr content within layers was attributedto the effect of convection cells sweepingaway magma partially depleted in Cr inthe boundary zone of crystallization andreplenishing it with Cr-undepleted mag-ma. A further consequence of the preser-vation of such sharp reversals relates tohypotheses about infiltration metasoma-tism or compositional convection. Signi-ficant mineral re-equilibration duringvertical migration of residual liquid frombelow would smooth out such sharpreversals or produce a zone within themagnetitite layer in which the Cr contentgradually increased. Such patterns havenot been observed.

Finally, it is possibly significant that thefootwall to every magnetitite layer in theeastern limb is an anorthosite (Molyneux1974). Random addition of magma wouldnot be expected to be presaged by an anor-thosite below the magnetitite layer. Acomplete solution for the genesis of mag-netitite layers needs to take such observa-tions into consideration. The successionof anorthosite underlying and frequentlyoverlying magnetitite layers begs the ques-

Fig. 8. Plot of An content in plagioclase ver-sus height across three magnetitite layersnear the base of the Upper Zone, showingabsence of reversals. Bars are range of corecompositions from several grains. There isa vertical gap of about 30 m betweenthese two sections, and a reversal in Ancontent occurs in this interval. (Author'sunpublished data.)

Fig. 9. Plot of V2O5 content in magnetititelayers versus height in the Upper Zone,Bushveld Complex (from Molyneux 1974).


tion as to why no pyroxene is present. Thedeeper footwall is gabbroic, indicating amagma lying on the plagioclase-pyroxenecotectic. If plagioclase could form andsink, so should pyroxene, hence its absen-ce means that the magma has moved outof the stability field of pyroxene. Thehypothesis involving an increase in pres-sure within the magma chamber, whichmight apply to chromitite layers, may alsohave application here. An increase in pres-sure might drive the magma into the sta-bility field of magnetite to produce a mag-netitite layer. However, pressure decreasemight have a different effect. It is well esta-blished that the thermal stability of maficminerals and plagioclase increase at diffe-rent rates with increasing pressure (Fig.11). Thus, a liquid initially saturated inplagioclase and pyroxene would crystallizeonly plagioclase if there were a reductionin pressure in the system. In this way, theapparent non-crystallization of pyroxeneduring the formation of anorthosite in thefootwall to magnetitite layers might beexplained by pressure reduction (Naslund& McBirney 1996).

The discordant, pipe-like bodies of mag-netitite cannot be attributed to pressurechanges, and different processes must beinvolved. They cannot be attributed toimmiscible liquids (Scoon & Mitchell1994) for the same reasons as discussedabove. Iron-rich liquids only containabout 30% FeO (total) and so remainingsilicate component still has to be removed.Also the immiscible liquids are highlyevolved (Table 1) and would not be able toproduce magnetite with high V contentsas are found in the ore body at Kennedy'sVale in the eastern limb (Willemse 1969).

Discordant silicate-rich pipes in the east-ern limb have been modeled as the resultof infiltration of a magma that selectivelyreplaced pre-existing non-liquidus mine-rals with liquidus minerals (Cawthorn etal. 2001). This process is essentially thesame as that proposed by Keleman et al.(1992) for the formation of discordantrefractory veins of harzburgite and dunitein mantle rocks. In this model, an injec-ting magma that is out of equilibriumwith the host rock, either because it is hot-ter or chemically undersaturated in someof the minerals in the host rock, replacedexisting host rock by precipitating theliquidus mineral of this added magma.

Location of these bodies may have a struc-tural control. For example, the Kennedy'sVale pipe occurs within the Steelpoortfault system (Willemse 1969). In terms ofthe genesis of oxide deposits, it should benoted that certain ilmenite vein depositshave been similarly interpreted byDuchesne (1996), who argued that puremonominerallic bodies did not have torepresent a liquid of that composition, butthat only the liquidus mineral separatedduring extensive through-flow of magmain an open system.

However, if this model is applied to themagnetitite pipes it implies the existenceof new magma being added to the com-plex that had magnetite on the liquidus.Such suggestions would support the con-cept of addition of magma to make mag-netitite layers, which was questionedabove on Sr isotopic considerations.

Silicate lenses and bifurcations inoxide layers Within chromitite and magnetitite layersthere occur numerous lenses and bifur-cations (Fig. 6c) containing silicate-domi-nant material. Scale may range from a fewcm upwards. A bifurcation may simply bea lens that is of such a size that the entirelens shape is not apparent on the scale ofthe exposure. The most famous bifur-cations can be seen at the Dwars Riverlocality of the Upper Group 1 (UG1)chromitite layer in the eastern limb. Infact, in all exposures the UG1 displayssuch features (Lee 1981). Apart from somerare calc-silicate xenoliths derived fromthe metasedimentary remnants from theroof of the intrusion, the silicate partingsin oxide-rich layers have the same minera-logy as adjacent layered rock types.Irregularly shaped fragments also occurthat are probably autolithic. However, thelenses and bifurcations discussed herehave a length to breadth ratio that givesthem the shapes of extremely thin disks,which may be only cm thick. It is sugges-ted that they are too thin, and hence fragi-le (Fig. 6c), to be physically transportedfragments.

This author's preference for the origin ofthese lenses is that they are primary accu-mulations of silicate minerals that did notdevelop into laterally extensive layers. Theprocess of formation is shown in Fig. 12.

Fig. 10. Plot of Cr in magnetite separates,analysed by XRF versus height from theMain Magnetitite Layer (from Cawthorn &McCarthy 1980).

Fig. 11. Plot of thermal stability of plagio-clase and pyroxene in a typical basicmagma as a function of pressure.


This diagram is constructed to explain alens of anorthosite within bifurcatingchromitite layers, but could also apply tomagnetitite layers, and also for lenses ofany silicate assemblage. A thin layer ofoxide (layer 1) formed on top of a contin-uous layer of anorthosite (stage 1, Fig. 12).The exact process of oxide formation isnot important to this model. This contin-uous layer of oxide was then followed bythe development of a partial layer of sili-cate. Two different models may be envisa-ged. If layering forms by in situ growth, itis possible that initiation of silicate forma-tion will not occur everywhere on theoxide surface instantaneously (Cawthorn1994). Instead, growth nodes may form atrandom spacing on the surface (only onenode is shown in stage 2, Fig. 12). Ifgrowth continues and the nodes coalesceinto an extensive layer, their original exi-stence will not be identifiable. However, ifgrowth is terminated before coalescenceby another period of oxide formation,these nodes will be trapped as lenses wit-hin a doublet layer of oxide (stage 3, whichshows the formation of layer 2 of oxide,Fig. 12). Alternatively, localized patches ofsilicate mineral accumulation may resultfrom convective activity. If low-densityplagioclase accumulates at the base of thechamber by being carried down on des-cending limbs of convection cells, grainsmay accumulate only at the base of suchdescending columns, and not where thecell is flowing horizontally, or ascending.Thus, discontinuous patches of plagiocla-se accumulation will result. Again, accu-mulation is terminated by initiation of thesecond oxide layer (stage 3, Fig. 12), and alens of anorthosite results. These lensescould be extremely thin relative to theirlateral extent. A repetition of oxide termi-nation, localized plagioclase accumulati-on, followed by oxide layer formation isshown as stages 4 and 5 (Fig. 12), and pro-duces the second anorthosite lens and thethird oxide layer.

The consequences of an evolution asshown in Fig. 12 are:1. Whereas their shape argues against phy-

sical transportation, extremely thin(and hence fragile) lenses of silicatematerial can form and survive as prima-ry precipitates within oxide layers.

2. If the lens is larger than the scale of out-crop, the lens might be seen as a bifur-cation only.

3. If two such lenses occur at different stra-tigraphic levels within the oxide layer, abifurcation within an oxide layer will beseen to continue laterally until it joinsanother chromitite layer (above orbelow). Thus, no terminations of oxidelayers within anorthosite bodies shouldoccur, consistent with field obser-vations.

4. Apparently thick oxide layers are notsingle layers, but the result of repeatedformation of thin oxide layers. No silica-te material accumulated between thelayers to provide a marker that a non-conformity exists within the oxide layer.

5. Reversals in chemical trends (such asthe Cr in magnetitite) are in agreementwith this hypothesis of multiple discreteoxide layers superposed without inter-vening silicate.

Numerous other features are evident insuch outcrops as the Dwars River locality,and the above discussion does not specifi-cally cover these other processes, but theprinciple of lenses and bifurcations can beexplained as a primary accumulation pro-cess rather than as a secondary remobili-zation event.

Nucleation and supercooling considerationsThe previous discussions have all focusedon processes that occur during slow cool-ing and under equilibrium conditions.General models exist for layering thatappeal to non-nucleation, oscillatorynucleation, supercooling and other dis-equilibrium processes, as reviewed byNaslund & McBirney (1996). The fact thatmagmas are dynamic systems and notalways at equilibrium is inescapable.However, in very slowly cooling magmachambers the extent of deviation fromideality may be limited. Gibb (1974) stu-died experimentally the extent of super-cooling prior to plagioclase nucleation. Inrapidly cooling systems, supercooling wassignificant. However, he pointed out thatin slowly cooled intrusions (he used theStillwater intrusion for his example) therate of cooling is so slow that supercoolingand supersaturation would have been veryrestricted. For example, Cawthorn & Wal-raven (1998) showed that in the CriticalZone of the Bushveld Complex the rate ofcooling was about 1°C every 200 years,severely restricting the opportunity for

Fig. 12. Schematic model for the formationof lenses and bifurcations in oxide layers.See text for explanation.


supersaturation and non-equilibriumprocesses. However, the sharpness of oxidelayer boundaries, and the appearance ordisappearance of associated silicate mine-rals indicates that some fundamentalchange had to have occurred in the cham-ber to produce these oxide layers. Forexample, magma mixing could causesuperheating or supercooling dependingupon the detailed topology of the liquidusof the two magma and their intermediatecompositions. Pressure changes wouldalso affect the instantaneous liquidus tem-perature. Thus, either of these processesmight lead to non-equilibrium effects, butpurely internal processes related to cool-ing within a single magmatic system can-not provide the answer to oxide layering.

SummaryThe Bushveld Complex contains manyeconomically important layers of chro-mitite and vanadiferous magnetitite. Theirlateral extent and chemical uniformitydemand processes that can operateinstantaneously over enormous distances.Numerous studies yield conflicting inter-pretations on the origin of these layers.The chromitite layers often form the basallayer of a cyclic unit; magnetitites are usu-ally underlain and overlain by anorthositewithin a gabbroic sequence. Models thatdo not encompass processes producingthe distinctive silicate rock associationmay be too myopic. Each model explainssome of the field or geochemical featuresobserved, but contradictory informationcan be presented against every hypothesis.It is even inferred here that the two typesof oxide layering may have different gene-ses, based on a study of the contrastingbehaviour of Cr in both settings.

AcknowledgementsI wish to thank the Geological Survey ofNorway, and specifically Are Korneliussen,Jean-Clair Duchesne, Brian Robins andRichard Wilson for their organisation ofthe conference and for their invitation forme to participate. The writer thanks thefollowing organisations that have longsupported his research: the NationalResearch Foundation of South Africa,Lonmin Platinum, Impala Platinum andAnglo Platinum mining companies. Fruit-ful discussions with numerous colleagueshave been essential in helping with formu-

lation of many of the problems and obser-vations synthesized here.

ReferencesCameron, E.N. 1980: Evolution of the lowerCritical Zone, central sector, eastern BushveldComplex, and its chromite deposits. EconomicGeology 75, 845-871.

Cameron, E.N. 1982: The upper Critical Zone ofthe eastern Bushveld Complex - precursor ofthe Merensky Reef. Economic Geology 77, 1307-1327.

Cameron, E.G. & Desborough, G.A. 1969:Occurrence and characteristics of chromitedeposits - eastern Bushveld Complex.Economic Geology Monograph 4, 41-71.

Carr, H., Groves, D.I. & Cawthorn, R.G. 1994: Theimportance of synmagmatic deformation inthe formation of the Merensky Reef potholesin the Bushveld Complex. Economic Geology89, 1398-1410.

Cawthorn, R.G. 1994: Growth nodes at the baseof magnetitite layers in the Upper Zone of theBushveld Complex. South African Journal ofGeology 97, 455-461.

Cawthorn, R.G. & McCarthy, T.S. 1980: Variationsin Cr content of magnetite from the UpperZone of the Bushveld Complex - evidence forheterogeneity and convection currents inmagma chambers. Earth and Planetary ScienceLetters 46, 335-343.

Cawthorn, R.G. & Walraven, F. 1998:Emplacement and crystallization time for theBushveld Complex. Journal of Petrology 39,1669-1687.

Cawthorn, R.G. & Walsh, K.L 1988: The use ofphosphorus contents in yielding estimates ofthe proportion of trapped liquid in cumulatesof the Upper Zone of the Bushveld Complex.Mineralogical Magazine 52, 81-89.

Cawthorn, R.G. & Webb, S.J. 2001: Connectivitybetween the western and eastern limbs of theBushveld Complex. Tectonophysics 330, 195-209.

Cawthorn, R.G., Harris, C. & Kruger, F.J. 2001:Discordant ultramafic pegmatoidal pipes inthe Bushveld Complex. Contribution toMineralogy and Petrology 140, 119-133.


Duchesne, J.C. 1996: Liquid ilmenite or liquidusilmenite: a comment on the nature of ilmenitevein deposits. In: Demaiffe, D. (ed) Petrology andgeochemistry of magmatic suites of rocks in thcontinental and oceanic crusts. Universite Librede Bruxelles, Royal Museum for Central Africa,Brussels, 73-82.

Eales, H.V. 2000: Implications of the chromiumbudget of the western limb of the BushveldComplex. South African Journal of Geology 103,141-150.

Eales, H.V. & Cawthorn, R.G. 1996: The BushveldComplex. In: Cawthorn, R.G. (ed) LayeredIntrusions: Elsevier, Amsterdam, Netherlands,181-229.

Gibb, F.G.F. 1974: Supercooling and the crystal-lization of plagioclase from a basaltic magma.Mineralogical Magazine 39, 641-653.

Irvine, T.N. 1975: Crystallization sequences inthe Muskox intrusion and other layered intru-sions – II. Origin of chromitite layers and simi-lar deposits of other magmatic ores.Geochimimica et Cosmochimica Acta 39, 991-1020.

Irvine, T.N. 1977: Origin of chromitite layers inthe Muskox intrusion and other stratiformintrusions: a new interpretations. Geology 5,273-277.

Irvine, T.N., Keith, D.W. & Todd, S.G. 1983: The J-Mplatinum-palladium reef of the StillwaterComplex, Montana. II. Origin by double diffusi-ve convective magma mixing and implicati-ons for the Bushveld Complex. EconomicGeology 78, 1287-1334.

Keleman, P.D., Dick, H.J.B. & Quick, J.E. 1992:Formation of harzburgite by pervasivemelt/rock reaction in the upper mantle. Nature358, 635-641.

Kruger, F.J., Cawthorn, R.G. & Walsh, K.L. 1987:Strontium isotope evidence against magmaaddition in the Upper Zone of the BushveldComplex. Earth Planetary Science Letters 84, 51-58.

Lee, C.A. 1981: Post-depositional structures inthe Bushveld Complex mafic sequence.Journal of the Geological Society, London 138,327-341.

Lipin, B.R. 1993: Pressure increases, the formati-on of chromitite seams, and the developmentof the Ultramafic Series in the Stillwater

Complex, Montana. Journal of Petrology 34,955-976.

McDonald, J.A. 1967: Evolution of part of thelower Critical Zone, Farm, Ruighoek, westernBushveld. Journal of Petrology 8, 165-209.

Molyneux, T.G. 1974: A geological investigationof the Bushveld Complex in Sekhukhunilandand part of the Steelpoort valley. Transactionsof the Geological Society of South Africa 77, 329-338.

Murck, B.W. & Campbell, I.H. 1986: The effects oftemperature, oxygen fugacity and melt com-position on the behaviour of chromium inbasic and ultrabasic melts. Geochimica etCosmochimica Acta 50, 1871-1887.

Naslund, H.R. & McBirney, A.R. 1996:Mechanisms of formation of igneous layering.In: Cawthorn, R.G. (ed) Layered Intrusions:Elsevier, Amsterdam, Netherlands, 1-44.

Osborn, E.F. 1980: On the cause of the reversalof the normal fractionation trend an adden-dum to the paper by E.N. Cameron, "Evolutionof the lower critical zone, central sector, eas-tern Bushveld Complex, and its chromitedeposits". Economic Geology 75, 872-875.

Philpotts, A.R. & Doyle, C.D. 1983: Effect ofmagma oxidation state on the extent of silica-te liquid immiscibility in a tholeiite basalt.American Journal of Sciences 283, 967-986.

Pollard, D.D., Delany, P.T., Duffield, W.A., Endo, E.T.& Okamura, A.T. 1983: Surface deformation involcanic rifts. Tectonophysics 94, 541-584.

Reynolds, I.M. 1985: The nature and origin oftitaniferous magnetite-rich layers in the UpperZone of the Bushveld Complex: a review andsynthesis. Economic Geology 80, 1089-1108.

Sampson, E. 1932: Magmatic chromite depositsin Southern Africa. Economic Geology 27, 113-144.

Schoenberg, R., Kruger, F.J., Nagler, T.F., Meisel, T.& Kramers, J.D. 1999: PGE enrichment in chro-mitite layers and the Merensky Reef of thewestern Bushveld Complex; a Re-Os and Rb-Srisotope study. Earth Planetary Science Letters172, 49-64.

Schurmann, L.W., Grabe, P.J. & Steenkamp, C.J.1998: Chromium. In: Wilson, M.G.C. &Anhaeusser, C.R. (eds) The Mineral Resources of


IntroductionTitanium deposits and prospects inUkraine are associated with Proterozoicand Palaeozoic magmatic basic rocks,Mesozoic and Cenozoic weathering crustsof basic rocks and alluvial placers, andCenozoic littoral placers. Reserves andresources of Ti-ores are a stable basis forthe Ukrainian titanium industry. Thebasis for the latter are the exogenic depo-sits of Ti-ores, whereas endogenic ores arenot mined at present. Nevertheless themain Ti resources are hosted by gabbro-anorthosite massifs.

The Ukrainian Ti-ore province, as de-scribed by Malyshev (1957), coincides spa-tially with the outlines of the UkrainianShield (USH) Precambrian complexes andtheir Mesozoic and Cenozoic cover. Later,the borders of the province were enlarged

into the Dnieper-Donets aulacogen (Fig.1).

During the time of the former USSR, alarge Ti-mineral industry was created inthe Ukraine. The total share of theUkrainian TiO2 industrial reserves in theUSSR was 42%, including more than 90%in the European part of the country. Thebasis of the industry includes more than40 titanium deposits, 11 of which areexplored in detail. Only 5 of them, of exo-genic origin, are being exploited. Theother Ti-deposits are reserves. TiO2 prod-uction in the Ukraine is shown in Table 1.

Exogenic depositsThe exogenic Ti-ores are of residual andplacer types. Residual deposits are spatial-ly related to Mesozoic weathering of Ti-

South Africa: Council for Geoscience, Pretoria,South Africa, 90-105.

Scoon, R.N. & Mitchell, A.A. 1994: Discordantiron-rich ultramafic pegmatites in theBushveld Complex and the relationship to

iron-rich intercumulus and residual liquids.Journal of Petrology 35, 881-917.

Willemse, J. 1969: The vanadiferous magneticiron ore of the Bushveld Igneous Complex.Economic Geology Monograph 4, 187-208.

Titanium deposits in Ukraine focused on theProterozoic anorthosite-hosted massifs

Dmitry Gursky, Sergey Nechaev & Alexander Bobrov

Gursky, D.S., Nechaev, S.V. & Bobrov, A. 2003: Titanium deposits in Ukraine focused on the Proterozoicanorthosite-hosted massifs. Norges geologiske undersøkelse Special Publication 9, 21-26.

The largest titanium resources of Ukraine are related to anorthosite parts of large Proterozoic (1.8 Ga)composite plutons; these contain a variety of Fe-Ti deposits with significant differences in ore geologicaland mineralogical characteristics due to local geological variations. Some well-documented deposits,such as Stremigorod and Krapivnia within the Korosten pluton, and Nosachev within the Korsun-Novomirgorod pluton, contain very large resources of ilmenite ± titanomagnetite ± apatite, althoughthese are not economic at present. Of significant economic importance are residual deposits related toMesozoic weathering of Ti-bearing gabbro-anorthosite massives and ancient placer deposits, which arepresently being exploited and contribute the basis for a Ukrainian titanium industry. Main economiccomponents in these deposits are ilmenite ± rutile ± leucoxene; in some of these deposits other mineralssuch as apatite, zircon, kyanite, sillimanite, staurolite and diamonds represents valuable by-products.

Department of Geology, Ukrainian State Geological Research Institute, Avtozavodskaia St. 78. UA-04114Kiev, Ukraine.

Year of production 1988 1989 1990 1991 1992 1994 1995 1996 1997 1998

Ilmenite concentrate 263.0 249.3 229.6 217.2 221.4 530 359 250 250 250

Rutile concentrate 118.2 130.4 123.6 108.0 108.7 80 112 50 50 50

Total 381.2 379.7 353.2 325.2 330.1 610 471 300 300 300

Table 1. Overall production of ilmenite andrutile concentrates in Ukraine during theperiod 1988-1998 (thousands of tonsrecalculated as TiO2, after Belous 1998, andWorld Mineral Statistics 2000).


bearing gabbro-anorthosite massifs in thenorthwestern part of the USH. The avera-ge thickness of the weathering crust is bet-ween 10 and 15 m, but within linear tecto-nic zones it reaches 50-70 m and evenmore in places. The content of ilmenite inthe weathering crust depends on its con-tent in the primary crystalline rocks andnormally fluctuates from 30 kg/m3 up to130 kg/m3, reaching 200 kg/m3 in produc-tive horizons of weathered crust. Accor-ding to Tarasenko (1992), layered orebodies consist of kaolinite, montmorill-onite and hydromica, and contain 6.0-10.8% TiO2 and 0.5-2.6% P2O5 related to

ilmenite and apatite, respectively. Placerdeposits are the main industrial Ti-ores in Ukraine, particularly complex Ti-Zr lit-toral placers, although continental, alluvi-al, ilmenite placers are also important. Thelatter are developed to the northwest ofthe USH and are related to intense LateJurassic-Early Cretaceous weathering anderosion. There are both ancient andmodern continental placers. All of theancient placers are buried. These are thelargest and richest deposits: ore-bodies areup to 5 km long, up to 1 km wide and upto 10 m thick, and ilmenite contents varybetween 150 and 300 kg/m3 (Belous 1998).

Alluvial placers of the northwest USH aremonomineralic, and contain altered ilme-nite (leucoxene). The modern placers aresmall as a rule and contain fresh ilmenite.Ore bodies in alluvial placers are usuallymeandering and ribbon-like, and coincidewith river valleys. Ore minerals are con-centrated in the lower parts of placers.Both residual and alluvial placer depositsare represented by the Irsha group depo-sits, located in the northwestern part of theUSH and worked by the Irsha IntegratedMining and Dressing Combine.

Littoral placers are developed in the cen-tral and southern regions of Ukraine andrepresent ancient and modern sandybeaches and shallow shelves. The formerare buried, and the latter exposed. Theancient placers coincide with sandy-clayeysediments of the Oligocene-Miocene seas,and the modern ones formed by long-range drift on spits of land along compa-ratively short parts of northern coastlines

Fig. 1. Sketch map showing the location oftitanium deposits and prospects withinthe Ukrainian Ti-province (after Bochai etal. 1998)

Table 2. The principal geological-industrialtypes of exogenic titanium deposits (afterBelous 1998).

Deposit Geological Host rocks and their age Ore body Ore mineral Simultaneouslysetting morphology composition useful component

Irsha group Sedimentary Alluvial and alluvial- Layered Ilmenite Apatitedeposits cover of the USH deluvium deposits

(secondary kaolin andkaolin clay), Mesozoic and Cenozoic

Malyshev, Sedimentary Sandy-clayey deposits, Layered and Ilmenite, Zircon, kyanite,Volchansk, cover of the Neogene lens-like rutile, staurolite, xenotime,Tarasov, USH, littoral leucoxene sillimanite, diamondsZelenoyar placers

Krasnokut The Dnieper- Sandy-clayey deposits, Layered and Ilmenite ZirconDonetz aulacogen, Neogene lens-like leucoxenedeltaic placers


of the Azov and Black Seas. The industrialvalue of the modern placers is rather limi-ted, and the ancient types form the mainindustrial titanium deposits. The latter arerepresented by a number of deposits (Tab-le 2), among which only those at Malyshevare being mined by the VerkhnedneprovskState Mining and Metallurgical Combine.The principal products are concentrates ofrutile, ilmenite, zircon, kyanite-sillimaniteand staurolite, as well as glass, and moul-ding and building sands. The kyanite-silli-manite concentrate contains diamonds0.01-0.03 mm in size and forming as muchas 1 carat/t of concentrate.

The peculiarity of the Miocene littoral pla-cer deposits of Ukraine is the presence of alarge number of fine-grained diamonds ofkimberlite, lamproite, eclogite and impac-tite types, as well as some of undetermined(pyroclastic komatiite?) origin (Kvasnitsa& Tsymbal 1998). These placers were for-med as a result of erosion and redepositionof sands older in age than those hostingdiamonds, the native sources of whichremain unknown.

Endogenic deposits There are four areas with endogenic tita-nium ores in Ukraine (Fig. 1). Three ofthem (I, II, III) are related to anorthositemassifs and bodies within the large-scaleand composite Proterozoic plutons:Korosten, Korsun-Novomirgorod and EastAzov. The fourth area, South Donbass,coincides with Devonian volcano-plutonicrocks situated in the border zone betweenthe Azov block and Dnieper-Donets aula-cogen.

There are a number of Ti-deposits-prospects related to gabbro-anorthositewithin the Korosten and Korsun-Novo-mirgorod plutons. Three deposits mayserve as examples: Stremigorod, Krapivniaand Nosachev.

The Stremigorod Ti-deposit, situated with-in the Korosten pluton, is related to themonzonite-peridotite complex of the 830km2 Chepovichi gabbro-anorthosite mas-sif. The ore body has an irregular ovalshape (2.3 x 1 km in size), elongated in anorthwest direction along the CentralKorosten fault. In cross section, the orebody is conical with rather steep andsharp contacts with the host rocks. It is

traced to a depth of 1200 m and is charac-terised by a concentric structure: the cen-tral part is represented by plagioclase peri-dotite and melanocratic troctolite, whichpass gradually in the peripheral part toleucocratic troctolite, olivine gabbro, gab-bro-monzonite and gabbro-pegmatite(Fig. 2). Troctolite is the main rock type inthe ore body.

Ores of the Stremigorod deposit are cha-racterised by uniform compositions anddistributions of apatite and ilmenite.Concentration of ore minerals increasesfrom the peripheral to the central parts ofthe ore body (Table 3). The richest apati-te-ilmenite ores in the central part of thedeposit contain 6.9-8.17 % TiO2 and 2.8-4.9% P2O5, whereas the peripheral partcontains 3.36-5.9% and 0.65-1.5`%, andandesinite and monzonite contain nomore than 1% TiO2 and 0.5 % P2O5. Theweathering crust above the Stremigoroddeposit contains between of 8-10% TiO2

and 2.5% P2O5.

Three varieties of apatite-ilmenite ores aredistinguished: (1) poor, disseminated inolivine gabbro, (2) medium-rich, disse-minated in leucocratic troctolite, (3)medium-rich, disseminated in troctoliteand plagioclase peridotite.

Table 3. Modal and chemical characteristics ofapatite-ilmenite ores of the Stremigoroddeposit (after Tarasenko 1992).

Mineral and chemical Host rockcomposition (%) gabbro troctolite peridotite

Plagioclase (An40-55) 60.6-64.4 37.1-46.0 25.5-31.6

Pyroxene (salite of augite series) 0.6-3.2 0.4-1.4 i.g*-0.7

Olivine (hyalosiderite Fa35-45-

hortonolite Fa50-58 ) 3.3-8.4 17.5-22.7 20.1-26.4

Ilmenite 4.95-11.4 11.6-12.6 11.6-15.1

Titanomagnetite 0.4-0.85 1.2-1.9 0.55-3.87

Apatite 1.1-3.0 4.5-6.0 8.1-10.1

Amphibole, talc, chlorite 15.1-19.8 17.9-21.7 16.8-21.7

TiO2 3.36-5.99 6.8-7.2 6.9-8.17

P2O5 0.65-1.5 2.8-3.1 3.58-4.5

FeO 8.97-12.3 14.0-16.47 15.0-21.82

Fe2O3 2.82-4.35 5.27-5.86 4.56-6.31

MgO 3.38-6.02 7.37-8.17 6.87-9.01

*i.g.: Individual grains


In the ore-bearing troctolite, 80-84% of allthe TiO2 is concentrated in ilmenite, and90-93.8% of the P2O5 is in apatite. Theores are characterised by a rather highextraction rate of ilmenite and apatite andhigh contents of TiO2 and P2O5 in theconcentrates.

The Krapivnia Ti-deposit is located in thecentral part of the Volodarsk-Volynsk gab-bro-anorthosite massif within the Koro-sten pluton. The area of the massif is 1250km2.

The ore body has an oval shape; it is about350 m thick, and is characterised by a zonalstructure. From the peripheral part of theore body to the centre, the host rockschange from gabbro-anorthosite to ande-sinite, leucotroctolite and gabbro, ore-bea-ring olivine gabbro, ore-bearing olivinepyroxenite and plagioclase peridotite.

The ore minerals (ilmenite, magnetite,titanomagnetite and ulvospinel) are con-centrated in the interstices between thesilicates, forming a sideronitic texture. Ofthe whole-rock content of TiO2 in the ore,28% is concentrated in ilmenite and 60%in titanomagnetite. The latter also contains0.41% V2O5. Mineral and chemical com-positions are given in Table 4.

The Nosachev Ti-deposit, situated withinthe Korsun-Novomirgorod pluton, islocated in the southern part of the Smelagabbro-anorthosite massif. The deposit islens-shaped, with an overall length of 1.5km, and a conical section.

Ore-bearing gabbronorite is bordered bygabbronorite, monzo-gabbronorite, quar-tz andesinite and intensely K-feldspathi-sed gabbro-anorthosite (Fig. 3). Thezoned structure of the deposit is emphasi-zed by variations in the chemical compo-sition of the host rocks, including theirTiO2 contents: quartz andesinite (no morethan 1% TiO2), gabbronorite (between3.68-4.58%), ore-bearing gabbronorite(more than 7%).

In gabbronorite, forming the outer part ofdeposit, plagioclase is the dominant mine-ral (50-60%). It is granulated and showsde-anorthisation and a very fine segregati-on of K-feldspar. The content of dissemi-nated ilmenite reaches up to 7-10%.

Fig. 2. Schematic geological map and cross sections of the Stremigorod deposit (after Tarasenko1990). 1 - sedimentary cover rocks, 2 - weathering crust, 3 - K-feldspathised gabbro-anorthosite, 4 -gabbro, 5 - ore-bearing troctolite, 6 - ore-bearing plagioclase peridotite and melanocratic troctoli-te, 7 - gabbro-pegmatite, 8 - olivine gabbro-pegmatite, 9 - gabbro-monzonite and quartz andesi-nite, 10 - exploratory borehole.

Fig. 3. Schematic geological section of the Nosachev deposit (after Tarasenko 1990). 1 – gabbro-anorthosite, 2 - K-feldspathised gabbro-anorthosite, 3 - quartz andesinite, 4 – monzogabbro-nori-te, 5 - gabbronorite with disseminated ilmenite ore, 6 - rich, disseminated and massive ilmeniteore, 7 - sedimentary cover rocks, 8 - prospect drillhole.


Ore-bearing gabbronorite is characterisedby medium-rich disseminated and massi-ve ores (Table 5).

The Kalchik Ti-prospect, situated in thesouthern part of the East-Azov pluton, istypically made up of syenites, includingone hosting deposits of REE-Zr and Ta-Nb ores. Anorthosite and rapakivi graniteare less characteristic, though the East-Azov pluton is synchronous with theKorosten and Korsun-Novomirgorod plu-tons. According to Tarasenko et al. (1989),the Kalchik Ti-ores are enveloped by stee-ply-dipping bodies of gabbroid rock, vary-ing from a few dozens of metres up tohundreds of metres in thickness, and froma few hundred metres up to 4 km andmore in length. Ore-bearing gabbroid isrepresented by layered lenses and schlierenwhich have indistinct contacts with thehost rocks (composite layering of gabbro-pyroxenite, wehrlite, andesinite, gabbrodi-orite and diorite). Ore bodies from a fewup to tens of metres in thickness and up toa few hundred metres in length are cha-racterised by low and moderate amountsof ilmenite (6-10%), titanomagnetite (5-7%) and apatite (5-7%). Richer minerali-zations, but limited in thickness, occurwithin wehrlite and pyroxenite with titan-omagnetite: ilmenite ratios of 1:2 and 1:1,respectively. In spite of the low grade ofthe ores, economic products of apatite andilmenite have been obtained by concentra-tion.

The Late Palaeozoic Ti-bearing vol-cano-plutonic assemblageThe setting of this assemblage is determi-ned by the Volnovakha fault zone in theeastern part of Ukraine, on the boundaryof the Azov block and the Variscides of theDonets basin. The latter is the southeas-tern part of the Dnieper-Donets aulaco-gen, separating the USH from theVoronezh massif in Russia.

The Late Palaeozoic volcano-plutonicassemblage is developed between theNorth- and South-Volnovakha faults, as astrip 60 km long and 5 km wide, and therocks of the assemblage are regarded as apotential Ti-bearing formation (Nechaev1970).

The assemblage includes three successivecomplexes (Buturlinov at al. 1972):

1. Intrusions of alkaline-ultrabasic cha-racter (390-380 ± 20 Ma),

2. Extrusions of basaltoid (365-335 Ma),3. Intrusions of nepheline syenite (330

Ma).

The first and third complexes are develo-ped to the east of the Volnovakha faultzone, where they intrude the Precambrianbasement and form the Pokrovo-Kireyevomassif, whereas the second is distributedthroughout the Volnovakha zone. Com-plexes 1 and 2 are Ti-bearing, but richerconcentrations of TiO2 are associated withto the alkaline massifs.

Mineral and chemical Host rockscomposition (%) gabbro peridotite pyroxenite

Plagioclase An45-47 and An50-55 22 6 i.g.

Pyroxene (titanaugite) 29 40.5 45.0

Olivine 14 21 19

Ilmenite 4 6 3.5

Titanomagnetite 17 21.5 25.5

Apatite 4 5 7

Amphibole, talc, chlorite 0.5 i.g. i.g.

TiO2 5.84 6.69 7.84

P2O5 2.58 2.79 3.17

FeO 20.55 24.66 29.34

Fe2O3 4.03 5.44 5.44

MgO 5.96 7.01 7.82

Mineral and chemical Type of orescomposition Disseminated Rich disseminated Massive *

Plagioclase 30-40 25-30 5-10

Pyroxene (ortho- and clino) 20-30 20-30 10-15

Olivine 10-20 i.g. i.g.

Ilmenite 15-20 30-50 70-80

Titanomagnetite - - -

Apatite 1-3 1-3 i.g.

Amphibole, talc, chlorite 4-7 3-5 5-7

TiO2 7.0 16.08 34.2

P2O5 ≤0.26 ≤0.25 0.06

FeO 14.89 17.77 29.32

Fe2O3 0.9 1.43 1.94

MgO 4.45 3.41 3.95

* Ores of massive type contain 0.24-0.28 V2O5 and 0.08-0.12% Cr2O3.

Ores are high-technological: TiO2 extraction in concentrates reaches 84%.

Table 5. Mineral and chemical characteristics of ilmenite ores of the Nosachev Ti-deposit (afterTarasenko 1992).

Table 4. Mineral and chemical characteristics of apatite-ilmenite-titanomagnetite ores of theKrapivnia deposit (after Tarasenko 1992).


The Pokrovo-Kiryevo massif is oval in formand covers an area of 10 km2. It consists ofplagioclase pyroxenite and leucocraticgabbro. In pyroxenite along the massifborder zone, bodies of ore-bearing pyro-xenites, 400-600 m long, contain on avera-ge 9.05% TiO2, 0.049% V and 0.15% P. Ti-bearing minerals are titanomagnetite andilmenite, as well as titanaugite. Titano-magnetite is the main ore mineral.

The potentially Ti-bearing Devonianbasaltoid complex (Table 6) forms a cover,with a maximum thickness of 500 m. Ti-bearing minerals in the main basaltoidrocks are titanaugite, titanomagnetite andilmenite, with titanite and leucoxene inhydrothermally altered rocks.

There are no known economic Ti-oresrelated to the basaltoid rocks. However, itmay be seen (Table 6) that ore-bearinggabbronorite from the anorthosite massifsis similar to andesite-basalt in Si, Ti, Al,Mg and total Fe contents, as well as tobasalt in Ca, Na and K contents.

ConclusionsUkraine possesses many deposits of tita-nium ores, among which those in produc-tion at present are placer and residualdeposits related to weathering of Ti-bea-ring anorthosite massifs. Placer depositsare the main industrial type of Ti-ores,and contain considerable reserves.

The greatest resources of Ti-ores are rela-ted to the Proterozoic anorthosite massifsin the USH. The Devonian Ti-bearing vol-cano-plutonic complex is interesting fromthe standpoint of the origin of the geoche-mical peculiarities of the Proterozoicanorthosite suite.

ReferencesBelous, Ya.T. 1998: Titanium (geology-economicreview). Geoinform, 48 pp. (in Russian).

Bochai, L. V., Gursky, D.S., Veselovsky,G.S. &Lasurenko, V.I., 1998, The main geological-industrial types of Ti and Zr alluvial deposits ofUkraine and their formation creation. MineralniResursy Ukrainy 3, 10-13 (in Ukrainian).

Buturlinov, N.V., Kobelev, M.V., Nechaev, S.V. &Panov, B.S. 1972: Geological structure, volca-nism and metal potential of joint zone betwe-en the Donets ridge and Azov block of theUkrainian shield, In: The Ukrainian shields frameplatform structures and their metal potential.Kiev: Naukova Dumka, 158-187 (in Russian).

Kvasnitsa, V.N. & Tsymbal, S.N. 1998: Diamondsof Ukraine and prospects of surveying theirdeposits. Mineralogicheskii Zhurnal 20, 118-129(in Ukrainian).

Malyshev, J.J. 1957: Regularities of formation andlocation of titanium ore deposits. Moscow:Gosgeoltekhizdat, 167 pp. (in Russian).

Nechaev, S.V. 1970: Mineralization in theVolnovakha fault-zone. Kiev: Naukova Dumka,172 pp. (in Russian).

Tarasenko, V.S. 1987: Petrology of anorthositesin the Ukrainian shield, and geological-geneticmodel of phosphate-titanium ore-formation.Geologicheskii Zhurnal 4, 43-52 (in Russian).

Tarasenko, V.S., Krivonos, V.P. & Zhilenko, L.A.1989: Petrology and ore potential of theSouth-Kalchik massif gabbroids (East-Azovarea). Geologicheskii Zhurnal 5, 78-88 (inRussian).

Tarasenko, V.S. 1990: Rich titanium ores in gab-bro-anorthosite massifs of the UkrainianShield. Jzvestiya Akademii Nauk SSSR 8, 35-44 (in Russian).

Tarasenko, V.S. 1992, Mineral-raw-material baseof titanium ores in Ukraine. GeologicheskiiZhurnal 5, 92-10 (in Russian).

World Mineral Statistics, 2000: Vol.2, 1994-1998:Industrial Minerals. British Geological Survey,265.

Table 6. Chemical composition of basaltoidrocks of the Volnovakha fault zone (afterButurlinov et. al. 1972).

Rock Oxydes (%)SiO2 TiO2 Al2O3 Fe2O3 FeO MgO CaO Na2O K2O

Picrite-basalt (30)* 41.85 5.18 9.61 7.00 7.09 11.42 10.84 1.72 0.69

Basalt (197) 43.85 5.27 13.18 9.04 5.09 6.40 7.71 2.70 1.99

Andesite-basalt (74) 48.60 4.45 14.63 9.43 3.46 4.50 5.13 3.48 2.21

Trachybasalt (10) 49.79 3.85 15.89 9.10 2.97 3.66 4.08 3.23 4.32

Ore-bearing gabbro- 49.22 4.60 15.44 1.78 11.67 4.33 7.37 2.70 1.58

norite of the USH

anorthosite massif

(after Tarasenko 1987)

* in brackets- number of analyses


IntroductionTwo large Proterozoic anorthosite-rapakivigranite plutons situated within theUkrainian Shield (USH) belong to thePalaeoproterozoic Kirovograd orogen-/Central Ukrainian fold belt. Extensivereserves of Ti ores and all economic occur-rences of gemmiferous (morion, topaz,beryl) pegmatite are located within theseplutons. Numerous deposits and prospectsin adjoining plutons contain major reser-ves of U and Be ores, considerable resour-ces of REE-Zr, Li and accompanying raremetal ores including gold. An analysis ofthe USH tectonic and metallogenic deve-lopment, together with spatial and agerelationships with the anorthosite-rapakivigranite suite, reveals the leading role of theplutons as an energy/temperature factorfor economic ore formation.

Geological setting of anorthosite-rapakivi granite plutons of the USHThe Korsun-Novomirgorod (KNP) andKorosten (KP) plutons, belonging to theanorthosite-rapakivi granite suite, aresituated in the central and northwesternparts of the USH, respectively. The KNP,about 5000 km2, coincides with the centralpart of the Palaeoproterozoic Kirovogradorogen (KO). The latter separates theArchaean Podol domain (foreland) to thewest from the Dnieper domain (hinter-land) to the east. The KP, about 10000km2, is located at the intersection of the

KO and the thrust belt of the Osnitsk-Mikashevichi orogen (Fig. 1).

The Podol domain comprises an enderbi-te-gneiss complex hosting REE-Zr-Thmineralization, and the Dnieper domainis made up of a granite-greenstone com-plex with ores of Fe, Au, Ni, Cu and Mo.Borders between the Archaean domainsand the KO are represented by thrustzones in the foreland and a system of nap-pes in the hinterland. The interior part ofthe KO comprises calc-alkaline associati-ons of island arc affinity. Thrust zones inthe foreland, hinterland and correspon-ding areas of the KO contain BIF, ophioli-tic mafic-ultramafic rocks and flyschassemblages forming complexes of intra-continental rift, oceanic crust and passivecontinental margins. All of these geodyna-mic settings determine the Palaeoproter-ozoic metallogeny preceding the emplace-ment of the anorthosite-rapakivi graniteplutons.

The oldest Palaeoproterozoic sedimentarysequence in the eastern part of the KO isrepresented by coarse Au-U-bearing terri-genous rocks of the Krivoy Rog super-group. Isotopic ages of U-mineralizationfall in the range 2.6-2.45 Ga (Belevtsev1995) or 2.68-2.31 Ga (unpublished data).The Archaean ages are those of detritalradioactive minerals (erosion products),and the Palaeoproterozoic ones corres-pond to sedimentary-epigenetic U-mine-

Links between the Proterozoic anorthosite-rapakivigranite plutons and ore-forming events in theUkrainian Shield (ores of titanium, uranium, raremetals and gold)

Sergey V. Nechaev & Viktor G. Pastukhov

Nechaev, S.V. & Pastukhov, V.G. 2003: Links between the Proterozoic anorthosite-rapakivi granite plu-tons and ore-forming events in the Ukrainian Shield (ores of titanium, uranium, rare metals and gold).Norges geologiske undersøkelse Special Publication 9, 27-33.

Hydrothermal and metasomatic processes have played a major role in the evolution of ore deposits asso-ciated with the Proterozoic Korsun-Novomirgorod and Korosten anorthosite-rapakivi granite plutonson the Ukrainian Shield. Magmatic anorthosite-related Fe-Ti-P ores are strongly affected and the ore ele-ments redistributed and enriched by metasomatic processes associated with rapakivi granite intrusions.In addition, hydrothermal systems associated with the rapakivi granites as well as other late-stage acidicintrusions, have formed a large variety of metasomatic/hydrothermal ore deposits; some of these containconsiderable resources of U-Be, REE-Zr, Li and accompanying rare metals including gold.

Ukrainian State Geological Research Institute, Geological Survey of Ukraine, Avtozavodskaya st. 78,UA-04114 Kiev, Ukraine.


ralization as well as to the first phase ofPalaeoproterozoic granitoids in the range2.45-2.31 Ga (Shcherbak & Ponomarenko2000, Stepaniuk 2000).

Formation of the KO occurred about 2 Gaago as a result of collision of an island arcwith the Podol domain. This explains whythe Palaeoproterozoic granitization ismost abundant in the northeastern part ofthe domain and adjacent to orogenicareas. Ages of monazite from oresegregati-ons within the Podol domain, dated as3.32, 2.52, 2.29 and about 2 Ga, reflect theinheritance of REE-mineralization fromPalaeoarchaean to Proterozoic times.Widespread within the KO, U-ore minera-lizations, located in the Palaeoproterozoicmetamorphic rocks and represented bymetasomatic U-K (K-feldspar) and hydro-

thermal fissure-vein types, have an age ofc. 2 Ga (2.19-1.93). The U-K ore type con-tains relatively high concentrations ofREE- and Th-bearing minerals as well asmolybdenite (Belevtsev 1995). It is estima-ted that the ore-regenerative potentialitiesof K-granite within the KO is mainly dueto high U background contents of the gra-nite and to the existence of initial sedi-mentary-diagenetic U-mineralizations inthe rocks which were involved in the for-mation of the anatectic/palingenetic gra-nites. Besides, related in time and space tothese granites are subeconomic spodu-mene-bearing pegmatite and microcline-apatite- tourmaline- biotite- cordieritemetasomatites enriched in Li, Rb, Cs andBe, as well as magnesian skarn and diopsi-de-plagioclase tactite. The latterre pre-sents a metapelite-marl assemblage, initi-ally enriched in W and Au (Kobzar 1981).The spatial connection between Li, B andW-bearing rocks allows comparison withdeposits in the USA, where considerableresources of these elements are present inthe Neogene deposits of the Cordilleras(California, Nevada), as well as in brinesand salt lakes (Silver Peak, Searls andothers) (Kozlovskiy 1989). A similar pre-cursor was possible within the Palaeo-proterozoic epicontinental depressions inthe central part of the USH.

The deep border between both Archaeandomains, in accordance with seismicsounding (Chekunov 1992), is determinedby the Transregional fault correspondingto the axial zone of the KO (Fig. 1). Thezone is characterised by: (1) a decrease indepth to the Moho (between 45 and 35km only) and a sharp reduction of thebasaltic layer (up to 2 km); (2) a decreasein rock density, including those rocksidentified as belonging to the upper mant-le; (3) the most extensive developmentwithin the USH of anatectic K-granitoidsand spatially related anorthosite-rapakivigranite plutons.

Origin of the anorthosite-rapakivigranite suiteAccording to petrological studies (Tara-senko 1987) the suite formed as multipha-se intrusions with emplacement of sub-crustal high-alumina basaltic magmasduring phase I, and subsequently duringphase II, of crustal granite magmatism.The first phase at 1.8-1.78 Ga (Nechaev &

Fig.1. Tectonic divisions of the Ukrainian Shield, location of anorthosite - rapakivi granite plutonsand principal deposits and prospects (excluding numerous occurrences/manifestations of oremineralisation). 1-3 – Kirovograd orogen: 1-volcanic belt, 2-backarc basin and shelf of hinterland,3-frontal overthrust belt and parautochthon; 4-5 – plutons (A-Korosten, B-Korsun-Novomirgorod):4-rapakivi-granite, 5-anorthosite; 6-9 – Faults: 6-Transregional with anatectic granitoids (1); 7-Mainoverthrust zones of orogen: (2) Odessa-Zhitomir, (3) West Ingulets, (4) Kirovograd, (5)Zvenigorodka-Annovka, (6) Jadlov-Traktemirov, (7) Pervomaysk-Brusilov, 8 – Overthrust zones ofbackarc Osnitsk-Mikashevichi orogen: (9) Teterev, (10) Perga; 9 – Shear Zones: (11) Krivoy Rog-Kremenchug megashear, (12) Talnov, (13) Central; 10-19 – Deposits/prospects: 10-U hydrothermalvein type, 11-K-U (REE±Mo) type, 12-Na-U type, 13-Ti (P,V,Sc), 14-REE-Zr, 15-Rare metal (Li,Be et all),16-Mo, 17-W (skarn type), 18-Au, 19-graphite. The metallogenic specialization of the ArchaeanPodol and Dnieper domains is shown in brackets.


Boyko 1988, Dovbush & Skobelev 2000)resulted in the intrusion of large anortho-site massifs and related small syeniteintrusions/stocks, enriched in Ti, P andREE, Zr, respectively. The lower-crustalrocks of the Archaean-Palaeoproterozoic'basement' may have caused contaminati-on of the basaltic magma. The secondphase, at about 1.77 Ga, is represented bythe rapakivi granite unit is followed by theprincipal Ti ore-bearing complexes atabout 1.76 Ga. The granite-porphyry andrhyolite unit, at 1745-1737 Ma (Dovbush& Skobelev 2000), is followed by a veinedgranite 1685±30 Ma which cuts across therapakivi granite (Nechaev & Boyko 1988).As has been noted by Dovbush & Skobelev(2000), K-Ar data between 1.8 and 1.7 Gaon biotite and amphibole from rocks ofthe suite correspond to the same range ofages that have been obtained by U-Pb andSm-Nd methods. In accordance withPastukhov & Khvorova (1994), the suiteforms a single magmatic assemblage,which is the plutonic equivalent of thebasalt-andesite-liparite series, and fromsuch a staindpoint the suite is rather simi-lar to the Devonian basaltoid complex ofthe Volnovakha fault zone (Gursky et al.2001).

The value of εNd for the Ti ore-bearingcomplexes is negative (-0.8), and similarto the value calculated for anorthosithebelonging to the early unit (Dovbush &Skobelev 2000). The εNd for rapakivi gra-nite is below –1.8, and both the mafic andthe felsic units therefore could not bedirectly derived from depleted mantle.They are most likely derivatives of partialmelting of the lower crust and/or theupper mantle, enriched by oceanic crustalmaterial due to a previous subduction.Melting of high-alumina basalt could pro-duce the initial magmas for mafic rocksbelonging to the plutons. The rapakivigranites cannot be derived from the sameparental magma, as their isotopic signatu-res differ from those of the mafic rocks. Itseems likely that anatexis of the uppercrustal rocks stimulated rapakivi graniteformation and, potentially, the ore-bea-ring features of syenite spatially and tem-porally related to anorthosite (gabbro-sye-nite intrusive complex after Nechaev &Krivdik 1989). Isotopic data supported bygeophysical deep sounding (Chekunov1992) turn our attention to the probabili-ty that the basaltic layer was a principal

factor for the formation of anorthosite aswell as for lateral tectonic transferencewithin the KO.

The spatial and age relationshipsbetween plutons and ore formationAs noted above, the potential ore-bearingmineral complexes were formed beforeemplacement of the plutons. Mineral-che-mical ore types of economic rank whichwere generated by pluton emplacementindeed occur in those geological settingswhere their corresponding precursorswere already present. The best example ofthis kind is displayed by REE-Zr and Tiores related to the formation of the anor-thosite-rapakivi granite suite.

Rare earth element-zirconium oresA deposit of this type of ore is located inthe northwestern flank of the KP (Fig. 1).The ore-bearing syenite is characterised bya high background of Zr (0.2-0.3%), andseparate ore bodies contain more than 6%ZrO2. The zircon crystals have complexstructures, typically including cores withan age of 1.79 Ga, surrounded by an over-growth reflecting an overprint of the ore-forming episode at 1.73-1.72 Ga (Nechaevet al. 1985). REE minerals are representedby britholite, orthite and bastnaesite(accompanied by Y-fluorite) and there aretwo obvious generations of REE-minerali-zations. Similar ore prospects are locatedin the southeastern corner of the KNP(Fig. 1) and in the eastern USH, where thegreat Azov REE-Zr deposit has been disco-vered recently.

Titanium oresThese ores are located directly within theprimary Ti-bearing anorthosite massifs(1.8 Ga in age) and Ti-deposits are coinci-dent with the Central fault (Fig.1). Theprincipal Ti ore-bearing complexes (1.76Ga in age) are metasomatic in origin. Theyconsist of monzonite-norite, monzonite-peridotite and oligoclasite-andesinite(Tarasenko 1987, 1990, 1992), and wereformed as a result of the contact influenceof rapakivi granite on anorthosite duringphase II (1.77 Ga). These complexes arelocated within anorthosite massifs affec-ted by extensive cataclasis and K- feldspat-hization, and have a zoned metasomaticstructure in which ore- and rock-formingcomponents have been redistributed. Ti,Fe and P were leached from the internal


portions of metasomatic columns anddeposited at the front of the columns,resulting in ore-mineralization (i.e. notinitially magmatic impregnation). Theoligoclasite-andesinite complex essentiallyrepresents chemical and mineral transfor-mation of initial anorthosite, and zones ofbasification related to that complex inclu-de ilmenite ore containing up to 20-35%TiO2.

According to Tarasenko (1987), zonationof Ti-bearing complexes is determined bytheir temperature of formation thatdepends on the distance from the rapakivigranite contact. The monzonite-noritecomplex represents the highest tempera-ture (830°C) and is located in the periphe-ral part of the anorthosite massifs near thecontact with rapakivi granite, whereasmonzonite-peridotite formed at some dis-tance from the rapakivi granite (750-700°C). The oligoclasite-andesinite com-plex, remote from the rapakivi granite,was formed at 500-450°C, and is characte-rized by a high degree of oxidation andleucoxenization of ilmenite compared tothe Ti-ores of the monzonite-norite andmonzonite-peridotite complexes. Therelationships between epigenetic Ti-oreson the one hand, and anorthosite andrapakivi granite on the other hand, implythat anorthosite provides the source of theTi and the rapakivi granite intrusionappears to be responsible for the metaso-matism which caused the redistributionand concentration of titanium.

Epigenetic Ti-ores within anorthositemassifs are only one phase in a sequenceof ore formation episodes which spreadoutside the plutons. Emplacement ofanorthosite and extensive K-feldspathiza-tion, leading to de-anorthization of plagi-oclase, caused Ca- and Na-metasomatism,the effects of which extended around andfar from the plutons.

Rare metal oresDeposits of economic rare metal oresoccur in the southwestern part of the KNP(Fig. 1). They are represented by extensi-vely brecciated microcline-albite-spodu-mene-bearing pegmatite with superimpo-sed petalite mineralization, and are consi-dered to be similar to replacement dykes(Nechaev et al. 1991). The complexity ofthese deposits is caused by extensive meta-somatic transformations (c. 1.9 Ga), not

only in the initial pegmatite, but also inthe host gneissic metasomatite enriched inrare alkalies. The main ore stage at c. 1.8Ga was synchronous with the KNP empla-cement, which was the cause of the high-temperature Li metasomatism: thequartz-albite-petalite assemblage, typicalof ore deposits, was formed by solutions atT≤680°C (Voznyak et al. 2000).

The Perga rare metal deposit, locatednorthwest of the KP (Fig.1), contains eco-nomic Be-mineralization (genthelvite,phenacite). Ore formation in the period1.74-1.62 Ga (Shcherbak 1978) is relatedto K-Na metasomatism with Li-F speciali-zation (Metalidi & Nechaev 1983,Buchinskaya & Nechaev 1990).

Uranium oresDeposits of metasomatic U-Na (albitite)type occur in K-granites which adjoin theKNP to the south (Fig. 1), as well as in fer-ruginous rocks of the Krivoy RogSupergroup remote from the pluton to theeast. These ores have an age between 1.8and 1.7 Ga, and two stages of albitizationare separated by a cataclasis (Belevtsev1995). High-temperature (450-480°C)garnet-diopside albitite, which is onlydeveloped near the contact with the KNP(Belevtsev 1995), is actually, in our opini-on, calcareous skarn that has suffered Na-metasomatism. Their formation tempera-ture is close to the oligoclasite-andesiniteTi ore-bearing complex (450-500°C), andthe albitite halo of the KNP appears to bea continuation of Na-metasomatism wit-hin the pluton itself, whereas most of theU-Na ore type formed at 250-300°C.

A geochemical feature that links the Ti-and U-ores is the presence of accompany-ing elements such as P, Zr, V and Sc.Primary ilmenite in anorthosite containsan average of 750 ppm V and 56.5 ppm Sc,whereas ilmenite in Ti-ores contains ave-rages of 1550 ppm and 74 ppm, respecti-vely (Borisenko et al. 1980). The earlymineral paragenesis of U-bearing albititeincludes apatite, malacon and U-titanates(brannerite and davidite). Andradite fromgarnet-diopside albitite contains up to 3%TiO2 as well as about 1% V, and this albi-tite also contains tortweitite with > 40%Sc2O3 and 7.6% ZrO2. The complex Sc-V-U ore of the Zheltaya Rechka deposit isrelated to Na-metasomatism of ferrugi-nous rocks of the Krivoy Rog Supergroup,


where aegirine-acmite contains 0.08-0.1%Sc and 2-9% V (Belevtsev 1995). It may bepresumed that the Ti, V, P, Zr and Sc sour-ces belonging to the U-Na ore type werehydrothermal solutions related to the plu-ton paragenesis.

Gold oresThe main deposits and prospects of Auores known at present are situated in thesouthern part of the KNP (Fig. 1). All areassociated with U and rare metals as wellas with tourmaline-bearing fields, andcommonly contain scheelite.

Epigenetic Au ores formed after extensivedeformation, including mylonitization ofthe country rocks (amphibolite, gneiss,migmatite, granite, skarn, aplite and peg-matite). Au-bearing mineralizations arelocated within linear zones of brittle defor-mation. The interior parts of these zonesare affected by oligoclase-quartz-biotitemetasomatism and contain ore mineralassemblages with native gold and tellurides(far from the KNP) and maldonite withouttellurides (close to the pluton).

Alteration of the host rocks by hydrother-mal ore-bearing solutions is displayed byminor albitization of plagioclase and for-mation of K-feldspar, carbonates andmixed-layer silicates of illite-smectite type(Mudrovskaya 2000). The upper tempera-ture for the formation of the Au-bearingassemblage corresponds to the transfor-mation of hexapyrrhotine into clinopyr-rhotine (273°C) and the lower temperatu-re limit is fixed by the formation of mixed-layer silicates (<178°C). The thermal para-meters of Au-ore formation are similar tothose of the post-albitite assemblages ofU-Na ore deposits. These assemblages for-med at temperatures lower than 300°Cunder the influence of hydrothermal solu-tions, characterised by high activities of Kand CO2 compared to U-bearing albitite(Belevtsev 1995). The Au ore-formingprocesses therefore appear to be a contin-uation of the U-Na ore-forming event. Onthe other hand, the geochemical antago-nism between Au and U within ore-bea-ring areas, ore fields and even deposits canbe explained by the transfer of Au ashydrosulphide complexes. The formationof Au-ores after U-ores agrees with theoccurrence of Au-bearing assemblagesaccompanied by adularia-chlorite metaso-matism which cuts across U-bearing albi-

tite (Prokhorov et al. 1994). However,according to Yatsenko & Gurskiy (1998),Au-bearing mineral assemblages (550-250°C) developed after tourmaline- andscheelite-quartz (420-330°C) in the KNPregion. This temperature inversion proba-bly reflects the influence of the pluton. Audispersed in Fe and As sulphides (550-380°C) subsequently changed to an assem-blage of native gold + chalcopyrite (450-370°C) and then to native gold + bismuth.(c. 270°C). This suggests that Au minerali-zation + tellurides located far from theKNP appear to be a continuation of thehigher temperature Au-bearing assembla-ges located close to the pluton.

Waning of magmatic activity in the KNPis fixed by granite veins which cut acrossrapakivi granite at 1685±35 Ma, and alsoby nearly synchronous retrograde meta-morphism (1670±30 Ma) of gneiss sur-rounding the pluton. Epigenetic quartzveinlets including Au-Bi-As mineralizati-on occur in the granite veins, and the K-Arage of biotite from a Au-bearing oligocla-se-quartz metasomatite distant from thepluton is 1615±30 Ma. The timing ofevents related to economic Au-mineraliza-tion indicates that the KNP was a signifi-cant thermal factor for ore formation.

Untraditional occurrences of Au related toREE-Zr-bearing syenite and Nb-Sn-bea-ring granite also occur in the northwes-tern part of the KP within the Perga raremetal ore field (Nechaev et al. 1983, 1986),and precious metal mineralization is con-sidered to belong to a rare metal-poly-metallic evolution series (Nechaev 1990) .The latter mineralization could be a resultof regressive hydrothermal processes cau-sed by the influence of the KP (1.8-1.7 Ga)because syenite, granite and Be-bearingmetasomatite are synchronous and spati-ally related to the pluton.

ReferencesBelevtsev, Ya. N. (chief editor) 1995: Genetictypes and regularities of location uranium oredeposits in Ukraine. Naukova Dumka Press, Kiev,396 pp. (in Russian).

Borisenko, L.F., Tarasenko, V.S., & Proskurin, G.P.1980: Ore-bearing gabbroids of the Korostenpluton. Geologiya Rudnykh Mestorozhdeniy 6,27-36. (in Russian).


Buchinskaya, K.M. & Nechaev, S.V. 1990: On thePerga granites problem. MineralogicheskiyZhurnal 3, 22-32 (in Russian).

Chekunov, A.V. (chief editor) 1992: Scheme ofthe deep lithosphere structure in the south-western part of the East European Platform,scale 1:1 000 000. Goskomgeologiya Ukrainy,Kiev (in Russian).

Dovbush, T.I. & Skobelev, V.M. 2000: Someremarks on the origin of the Korosten anor-thosite-rapakivi granite complex as based onisotope data. Geophysical Journal 22, 84-85.

Gursky, D., Nechaev, S.V., & Bobrov A. 2001:Titanium deposits in Ukraine focused on theProterozoic anorthosite-hosted massifs. Norgesgeologiske undersøkelse Report 2001.042.Abstracts-GEODE field workshop, S.Norway,51-60.

Kobzar, V.N. 1981:The Early Proterozoic sedi-mentation and metallogenic questions of theUkrainian Shield central part. Naukova DumkaPress, Kiev, 104 pp. (in Russian).

Kozlovskiy, E.A. (chief editor) 1989: MiningEncyklopaedy, United States of America.Sovetskaya Encyklopaediya Press, Moscow, 4,580-614 (in Russian).

Metalidi, S.V. & Nechaev, S.V. 1983: TheSushchany-Perga zone: geology, mineralogy orepotential. Naukova Dumka Press, Kiev, 136 pp.(in Russian).

Mudrovskaya, I.V. 2000: The mineralogical-gene-tic model of gold mineralization in the Savran-Sinitsovka ore zone (Ukrainian Shield). Thesis fora candidat. geol. sci. degree, Lviv StateUniversity, 23 pp. (in Ukrainian).

Nechaev, S.V. 1990: Evolution of ore formationprocesses in the Ukrainian Shield structures.Geologicheskiy Zhurnal 2, 68-80 (in Russian).

Nechaev, S.V. 1992: Some peculiarities of goldand silver occurrences in the western part ofthe Ukrainian Shield. Geologicheskiy Zhurnal 4,79-88 (in Russian).

Nechaev, S.V., Semka, V.A., Shirinbekov, S.V. &Chebotarev, V.A., 1983: The first find of nativesilver in the Ukrainian Shield. Doklady AkademiiNauk SSSR 270, 418-420 (in Russian).

Nechaev, S.V., Bondarenko, S.N. & Buchinskaya,K.M. 1986: Mineral state of gold and silver in

intrusive syenite and metasomatite of theUkrainian Shield. Doklady Akademii Nauk SSSR,289, 1483-1487 (in Russian).

Nechaev, S.V. & Boyko, A.K. 1988: Indication ofthe Proterozoic ore mineralization stages inthe Ukrainian Shield. Visnyk Akademii NaukUkraine 5, 40-46 (in Ukrainian).

Nechaev, S.V., Krivdik, S.G., Krochuk, V.M. &Mitskevich, N.Yu. 1985: Zircon from syenite of the Yastrebets massif (Ukrainian Shield) - indicator of their crystallization conditions.Mineralogicheskiy Zhurnal 3, 42-56 (in Russian).

Nechaev, S.V. & Krivdik, S.G. 1989: Geologicalregularities in distribution of alkaline rocks inthe Ukrainian Shield. Geologicheskiy Zhurnal 3,113-120 (in Russian).

Nechaev, S.V., Makivchuk, O.F., Belykh, N.A.,Ivanov, B.N., Kuzmenko, A.V. & Prytkov, F.Ya.1991: New rare metal district of the UkrainianShield. Geologicheskiy Zhurnal 4, 119-123 (inRussian).

Pastukhov, V.G. & Khvorova, G.P., 1994:Paleogeodynamic of the Ukrainian Shield. In:N.V.Mezhelovskiy & G.. Gusev (eds.) Geologicalmapping of the Early Precambrian complexes.Geokart, Moscow, 318-373 (in Russian).

Prokhorov, K.V., Glagolev, A.A. & Kuznetsov, A.V.,1994: Gold-bearing metasomatites of theYurievsk ore prospect in the Ukrainian Shieldcentral part. Otechestvennaya Geologiya 6, 17-25 (in Russian).

Shcherbak, N.P. (chief editor), 1978: Catalog ofthe isotope data. Naukova Dumka Press, Kiev,223 pp.

Shcherbak, N.P. & Ponomarenko, A.N., 2000: Agesequence of the volcanism and granitoidmagmatism processes of the Ukrainian Shield.Mineralogicheskiy Zhurnal 2/3, 12-24 (inRussian).

Stepanyuk, L.M. 2000: Geochronology of thePrecambrian of western part of the UkrainianShield (Archaean-Palaeoproterozoic). Thesis fora doctoral geol.sci. degree, Institute ofGeochem., Mineral. and Ore Form., NAS ofUkraine, Kiev, 35 pp. (in Ukrainian and Russian).

Tarasenko, V.S. 1987: Petrology of anorthositesof the Ukrainian Shield and geological-geneticmodel of phosphate-titanium ore formation.Geologicheskiy Zhurnal 4, 43-52 (in Russian)


Tarasenko, V.S. 1990: Rich titanium ores in gab-bro-anorthosite massives of the UkrainianShield. Izvestiya Akademii Nauk SSSR, geol. ser., 8,35-44 (in Russian).

Tarasenko, V.S. 1992: Mineral-raw material basisof titanium ores in Ukraine. GeologicheskiyZhurnal 5, 92-103 (in Russian).

Voznyak, D.K., Bugaenko, V.N., Galaburda, Yu.A.,Melnikov, V.S., Pavlishin, V.I., Bondarenko, S.N. &Semka, V.A. 2000: Peculiarities of the mineralcomposition and conditions of formation ofrare-metal pegmatites in the western part ofthe Kirovograd block (the Ukrainian Shield).Mineralogicheskiy Zhurnal 1, 21-41 (inUkrainian).

Yatsenko, G.M. & Gurskiy, D.S. (chief editors)1998: Gold deposits in the Precambrian gneisscomplexes of the Ukrainian Shield. GeoinformPress, Kiev, 256 pp. (in Russian)


Field evidence strongly suggests that manyFe- and Fe-Ti-oxide bodies were emplacedas liquids. The evidence is strongest for thedeposits at El Laco, Chile, where magneti-te bodies appear to have been extruded aslavas, and for which the experiments ofWeidner (1982) in the system Fe-C-O pro-vide convincing proof of oxide liquids at geologically reasonable temperatures(<1000°C). Thus, it was very exciting whenwe discovered that the Fe-Ti-oxide bodiesin the Laramie Anorthosite Complex con-tain graphite. Very importantly, however,Weidner’s experiments lacked TiO2 - andthe El Laco iron-oxide lavas are nearlytitanium-free.

Field evidence of a magmatic origin for Ti-bearing oxide bodies (ilmenite and magne-tite-ilmenite ores) still seems strong: manyare dike-like, crosscutting the country rockwith sharp contacts. Many are associatedwith anorthosite complexes. Isotopic stu-dies show that many are closely akin to thehost anorthosites. Many of these bodies('nelsonites') contain abundant apatite,and it has been suggested that apatite ser-ves as a flux to stabilize Fe-Ti-O-(apatite)magmas, possibly as melts immisciblewith a complementary silicate liquid.

Unfortunately, despite much effort there isno experimental evidence to support theconcept of apatite-fluxed melts. Philpotts(1967) reported ilmenite-apatite melts at~1400°C, but this is hardly compelling.The temperature is unrealistically high forcrustal conditions, and in any case fluora-patite and ilmenite separately melt nearthat temperature, so there is little eviden-ce for a eutectic or cotectic relationshipbetween them. Kolker (1982) pointed outthat while nelsonites broadly containapproximately 1/3 apatite and 2/3 oxide,the proportions in fact vary widely, fur-

ther casting doubt on the existence of aeutectic relationship.

For many years my students and I havetried - unsuccessfully - to generate Fe-Tioxide liquids in the laboratory at geologi-cally reasonable temperatures. We havetried apatite and iron phosphate and grap-hite - separately and in combination - aswell as chlorides as possible fluxes. Usinggraphite alone we have made smallamounts of Fe-rich liquid - but impor-tantly that melt is nearly Ti-free; most ofthe titanium remains in the residual crys-tals. Using graphite and apatite we againmade small proportions of Fe-rich, Ti-poor melt - but that melt dissolved morestrontium than phosphorus from thenatural apatite used in the experiments!Despite the common association of apati-te and oxides in nature, there is no experi-mental support for the existence of apati-te-oxide melts.

The experiments of Epler on an oxide-richtroctolite pegmatite (MS thesis, StonyBrook, see also Lindsley et al. 1988) sug-gest a possible explanation. Although hefailed to produce an immiscible oxidemelt, he showed conclusively that phos-phorus and iron mutually enhance theirsolubilities in the silicate melts coexistingwith crystalline Fe-Ti oxides. For example,at 5 kbar, 1100°C he produced a silicateliquid with 26.2 weight% iron as FeO,4.2% TiO2, and 3.8% P2O5. At 1150°C thevalues go up to 32.7, 8.56, and 4.2%,respectively. Thus, the close association ofapatite and oxides probably reflects copre-cipitation from a silicate parent: once thesilicate melt becomes saturated with oneof these phases, the other is likely to copre-cipitate as well. Experiments suggest thatthe oxide bodies crystallize from Fe-Ti-P-rich silicate liquids and are emplaced eit-

Do Fe-Ti oxide magmas exist?Geology: Yes; Experiments: No!

Donald H. Lindsley

Department of Geosciences, State University of New York, Stony Brook, NY 11794-2100, USA.

Approaches and processes


her as crystal mushes or in the solid state.

ReferencesKolker, A. 1982: Mineralogy and geochemistryof Fe-Ti oxide and apatite (nelsonite) depositsand evaluation of liquid immiscibility hypothe-sis. Economic Geology 77, 114-1158.

Lindsley, D. H., N. A. Epler, N.A. & Bolsover, L.R.1988: Nature and origin of the Sybille Fe-Ti

oxide deposit, Laramie anorthosite complex,Wyoming (abstract). Penrose conference: TheOrigin and evolution of anorthosites and asso-ciated rocks, Wyoming, 14-19 August 1988.

Philpotts, A. R. 1967: Origin of certain iron-tita-nium oxide and apatite rocks. EconomicGeology 62, 303-315.

Weidner, J. R. 1982: "Fe oxide magma in the sys-tem Fe-C-O." Canadian Mineralogist 20, 555-566.

Some results on the role of P, T and fO2 on ilmenite composition

Jacqueline Vander Auwera 1, John Longhi 2 & Jean-Clair Duchesne 1

1 Université de Liège, B-4000 Sart Tilman, Belgium. 2 Lamont-Doherty Earth Observatory, Palisades, NY10964, USA.

Using experimental and geochemical dataacquired on fine-grained samples, repre-sentative of liquid compositions, VanderAuwera et al. (1998) have defined for theRogaland AMC suite, a complete liquidline of descent ranging from primitivejotunites (hypersthene-bearing monzodi-orites) to evolved jotunites and then tocharnockites. Primitive jotunites representthe least differentiated compositions (highMgO, low K2O) which in the RogalandAnorthosite Province correspond to thechilled margins of leuconoritic (Hidra:Demaiffe & Hertogen 1981) and layered(Bjerkreim-Sokndal: Wilson et al. 1996)bodies. Evolved jotunites represent moredifferentiated compositions and occur inthe dyke system which crosscuts theRogaland Anorthosite Province. Modellingof the liquid line of descent supports thehypothesis that extensive fractionation ofprimitive jotunites can produce quartzmangerites with REE concentrations in therange of jotunites, strong depletions in U,Th, Sr, Ti and P, and smaller to no relativedepletions in Hf and Zr.

Jotunites characterised by very high con-centrations of FeO, TiO2 and P2O5 (FTProcks) have also been observed in Rogal-and. Experimental and petrographic dataindicate that these rocks represent accu-mulations of a dense oxide-apatite-pigeo-nite assemblage into coexisting multisatu-rated jotunitic to mangeritic liquids.

Experimental data performed in anhy-drous conditions from 1 atm (fO2 atNNO, FMQ-1 and MW-2.5) up to 13 kb(FMQ-2, FMQ-4) on the so-called Tjörn(TJ) primitive jotunite (Vander Auwera &Longhi 1994) are used to assess the pos-sible role of pressure, temperature and fO2

on ilmenite composition (Al, Mg, Cr).Indeed, in Rogaland, it has been shownthat the parent magmas of the Ana-Siramassif-type anorthosite and the layeredmafic intrusion of Bjerkreim-Sokndal,which both contain ilmenite deposits(Duchesne 1999), are generally similar tothis primitive jotunite (Duchesne &Hertogen 1988, Vander Auwera & Longhi1994, Vander Auwera et al. 1998, Longhi etal. 1999). These experimental data canthus bring information concerning thebehaviour of poisonous elements like Mgand Cr during polybaric crystallization ofilmenite in anorthosite complexes.

In the TJ primitive jotunite, ilmenite is anear-liquidus phase and experiments cor-responding to the first appearance of thisphase at a given pressure have been selec-ted in order to limit the extent of liquidcompositional variability. These experi-mental data indicate that the most impor-tant effects are an increase in DAl2O3

(Ilm/liq), a slight increase of DFe/Mg

(Ilm/liq) and a decrease of ilmenite Cr2O3

with increasing pressure. Nevertheless,this latter effect could also result from the


Fig. 1. Partition coefficients of several elements as a function of P/T and T. Open circles: 1 atm, MW; open square: 1 atm, MW-2.5; crosses: 1 atm, FMQ-1;filled circle: 5 to 13 kb.


earlier appearance of opx with increasingpressure, which decreases the Cr contentof the liquid. Besides that, there is no sig-nificant variation of DMgO with increasingpressure and DAl2O3 only slightly increases(Fig. 1). Comparison between observed(Duchesne 1999) and experimental ilme-nite compositions show that these ilmeni-tes display similar contents in Fe3+ and Mgbut experimental ilmenites are usuallyhigher in Al and lower in Mn.

These experimental data thus suggest thatthe compositional variability observed in ilmenites from different orebodies(Duchesne 1999) probably result fromvariable parent magma compositions, asthese ilmenites may have crystallized atdifferent times along the liquid line of des-cent. Subsolidus readjustments may alsobe superimposed on this primary effect(Duchesne 1972, 1999).

ReferencesDemaiffe, D. & Hertogen, J. 1988: Rare earthgeochemistry and strontium isotopic compo-sition of a massif-type anorthositic-charnocki-tic body: the Hidra massif (Rogaland, southNorway). Geochimica et Cosmochimica Acta 45,1545-1561.

Duchesne, J.C. 1972: Iron-titanium oxide mine-rals in the Bjerkreim-Sogndal Massif, south-

western Norway. Journal of Petrology 13, 57-81.

Duchesne, J.C. 1999: Fe-Ti deposits in Rogalandanorthosites (South Norway): geochemicalcharacteristics and problems of interpretation.Mineralium Deposita 34, 182-198.

Duchesne, J.C. & Hertogen, J. 1988: Le magmaparental du lopolithe de Bjerkreim-Sokndal(Norvège méridionale). C.R. Académie desSciences de Paris 306, 45-48.

Longhi, J., Vander Auwera, J., Fram, M. &Duchesne, J.C. 1999: Some phase equilibriumconstraints on the origin of Proterozoic(Massif ) anorthosites and related rocks.Journal of Petrology 40, 339-362.

Vander Auwera, J. & Longhi, J. 1994:Experimental study of a jotunite: constraintson the parent magma composition and crys-tallization conditions, P, T, fO2 of the Bjerkreim-Sokndal layered intrusion, Norway. Contributionto Mineralogy and Petrolology 118, 60-78.

Vander Auwera, J., Longhi, J. & Duchesne, J.C.1998: A liquid line of descent of the jotunite(hypersthene monzodiorite) suite. Journal ofPetrology 39, 439-468.

Wilson, J.R., Robins, B., Nielsen, F.M., Duchesne,J.C. & Vander Auwera, J. 1996: The Bjerkreim-Sokndal layered intrusion, southwest Norway.In: Cawthorn, R.G. (ed) Layered intrusions.Elsevier, Amsterdam, 231-255.

The origins of ore-bearing apatite rock(nelsonite) from the Proterozoic Suwalkianorthosite massif and Fe-Ti-P-rich hyp-ersthene monzodiorite (jotunite) from theSuwalki and Sejny massifs, NE Poland,have been investigated by means of meltinclusion techniques.

The investigated rock samples come froma deep borehole drilled in the crystallinePrecambrian basement in NE Poland,belonging to the East European Craton

(Fig.1). An E-W trending shear zone ofpost-collisional origin controls the occur-rence of Meso-Proterozoic intrusions –called the Mazury Complex – of bimodalcomposition, anorthosite-norite and rap-akivi-like granite. The Suwalki and Sejnyintrusions are classified as 'massif-typeanorthosites', belonging to the AMCGsuite of Proterozoic plutonism (Wiszniew-ska et al. in press). The ore-bearing apati-te dykes (nelsonites) were found for thefirst time in Poland (Lopuchowo boreho-

The nelsonite problem:an origin by melt immiscibility

Anrdrzej Kozlowski 1 & Janina Wiszniewska 2

1 Faculty of Geology of the Warsaw University, Poland. 2 Polish Geological Survey, Warsaw, Poland.

le) in the marginal part of the Suwalkianorthosite massif. Jotunite with chilledtexture has been found in the Sejny nori-tic intrusion (Sejny borehole) and also asfine-grained dykes in the Suwalki anor-thosite massif (Udryn borehole) (Fig. 1).

Nelsonites consist mainly of fluorapatite(up to 50-70 vol. %), magnetite, ilmenite,pyrrhotite, pyrite and chalcopyrite (Fig.2). The oxide/sulphide ratio is variable,ranging from 4:1 to 0.7:1. The presence ofREE-bearing minerals, like Ce, Nd, La-allanite and rhabdophane (Nd, Ce, La)-PO4·H2O, was recognised as rims aroundapatite grains or as an interstitial product(Wiszniewska 1997).

Jotunite dykes consist of calcic plagioclase,low- and high-Ca pyroxenes, biotite, apa-tite and Fe-Ti oxide ores as essential com-ponents, with local porphyritic structure.The jotunite dykes from the Suwalki mas-sif and chilled zones from the Sejny intru-sion have a high content of iron (from16.5 to 18.8% Fe2O3), titanium (from 2.3to 3. 5% TiO2) and phosphorus (from 0.6to 1.2% P2O5), a low magnesium number(# mg ca 0.36), and REE distributionsmoderately differentiated with (La/Yb)-CN c.10 and without a Eu anomaly(Wiszniewska et al. in press). These geo-chemical characteristics are very similar tothose of the Rogaland jotunites (Duch-esne 1999, Vander Auwera et al. 1998).

Melt inclusions in nelsonitesFluid inclusions were found in zoned apa-tite grains in inner cores, in intermediatezones and in outer rims (Fig. 2). Theinclusions in each zone differ in the typeof filling and in the homogenisation tem-peratures: (1) Inner core inclusions aremade up of pyroxene, apatite, carbonate,plagioclase, biotite, halite, sylvite andaqueous solution, and homogenise at 890-870°C. (2) The intermediate zones of theapatite grains bear inclusions poorer inpyroxene and plagioclase. They have car-bonate and apatite contents similar to theinner core inclusions, but show an incre-ased amount of aqueous solution. Theirhomogenisation temperatures range from820 to 770°C. (3) The outer rims contai-ned inclusions rich in solution and in hali-te, but poorer in carbonates and withoutsilicates. They homogenised at 690-550°C.(4) Gas-liquid inclusions were found inthe REE-rich interstitial product. They arefilled with aqueous solution, sometimescontaining halite and calcite. The homo-genisation temperatures ranged from 480to 230°C.

The homogenisation process in the high-temperature microscope heating stage ofinclusions in the core and intermediatezones of the apatite reveals, by increasingtemperature, the formation of two meltsseparated by a meniscus. The qualitativecomposition of the glassy phases wasdetermined by electron microprobe analy-sis. The data show that one melt is silicate-rich and the other phosphate-rich. More-over, small dispersed grains of silicateswere also identified and the silicate andphosphate melts were both found to con-tain carbon (i.e. carbonate component),the carbon content in the phosphate meltbeing distinctly higher than in the silicatemelt. Chlorine was also detected in bothmelts with an irregular distribution. Theseresults suggest that apatite and the otherminerals of nelsonite possibly crystallisedfrom a melt immiscible with a silicatemelt. Both melts neither homogenised upto temperature of 1080°C nor even dis-tinctly changed their volume proportions.

Melt inclusions in jotunitesPyroxenes from coarse-grained jotunitesin the central parts of the dykes containmelt inclusions with crystallised fillingsconsisting of pyroxene, feldspar, magneti-


Fig. 1. Geological map of the Suwalki anor-thosite intrusion (after Kubicki & Ryka1982).


Fig. 2. Cathodoluminescence zoning of apatite from the Suwalki nelsonite; the numbers refer tothe fluid inclusion types in apatite in individual growth zones, the circles indicate the typical inclu-sion positions.

te, calcite and apatite plus gas bubbles. Thecrystal aggregate starts to melt to anobservable degree at 800–820°C, the lastcrystals disappear at 980–1040°C, andsmall droplets of phosphate and carbona-te melt form. Complete homogenisationof all phases occurs at 1090–1170°C. Fine-grained pyroxenes from chilled facies con-tain melt inclusions with fillings being anaggregate of tiny grains of non-recogni-sable minerals with gas bubbles. These fil-lings also start to melt at 770–780°C, thelast crystals disappearing at 1000–1060°Cwith the appearance of small droplets ofphosphate and carbonate. Completehomogenisation of these droplets and gasbubbles with the silicate melt occurs at1090–1180°C. Some pyroxenes also con-tain rare inclusions filled with carbon dio-xide, probably in a liquid state.

ConclusionsObservation of the homogenization ofmelt inclusions point to the formation ofthe Suwalki nelsonite from the crystalliza-tion of a melt rich in phosphate and con-taining carbonate, which was immisciblewith a silicate-rich melt. As inferred fromthe melt inclusion fillings, the proportionof these two melts was close to a 1:1 ratio.This composition (50% apatite–50% dio-rite) is quite far within the miscibility gap,ant thus in the two-phase region on aPhilpotts`s (1967) diagram. The homoge-nisation temperature of the fluid inclusi-ons in zoned apatite distinctly indicatethat nelsonites formed from a melt whichprobably gradually and smoothly changedwith decreasing temperature to a very sali-ne aqueous solution, then to a moderate-saline medium-temperature aqueous sol-ution forming the latest outer rims of theapatite grains. As for the parent melt ofjotunites from Sejny, since its compositionplots in the one-phase silicate melt area ofthe Philpotts diagram, the observed silica-te and phosphate melts' immiscibility athigh temperature possibly results from ametastable process.

ReferencesDuchesne J.C. 1999: Fe-Ti deposits in Rogalandanorthosites (South Norway): geochemicalcharacteristics and problem of interpretation.Mineralium Deposita 34, 182-198.

Kubicki, S. & Ryka, W. 1982: Geological atlas ofcrystallnie basement in Polish part of the EastEuropean Platform, Wydawiniclwa Geologi-czne,Warszawa.

Philpotts A.R. 1967: Origin of certain iron-tita-nium oxide and apatite rocks: EconomicGeologist 62, 303–315.

Vander Auwera J., Longhi J. & Duchesne J.C.1998: A liquid line of descent of the jotunite(hypersthene monzodiorite) suite. Journal ofPetrology 39, 439-468.

Wiszniewska J. 1997: Suwalki nelsonites - newmineralogical and geochemical data. Przegl.Geol. 45, 883-892.

Wiszniewska, J., Claesson, S., Stein, H., VanderAuwera, J. & Duchesne, J.C. in press: The north-eastern Polish anorthosite massifs: petrologi-cal, geochemical and isotopic evidence for acrustal derivation. Terra Nova.

related to major crustal events have trans-formed ilmenite-bearing rocks into rutile-bearing rocks. Since rutile is a much morevaluable mineral than ilmenite, such trans-formation processes are of significant eco-nomic interest.

In western Norway, including the Sunn-fjord province, Proterozoic mafic and fel-sic igneous rocks were strongly affected bythe Caledonian orogeny. Two types of Tideposits occur in this region: magmaticilmenite-magnetite deposits associatedwith Proterozoic mafic intrusions, andrutile-bearing Caledonian eclogitic rocks.The rutile-bearing eclogites are Protero-zoic basic rocks that were transformedinto eclogite at high pressure during thecollision between Baltica and Laurentia c.400 Ma ago. During this process, ilmenitein the protolith broke down, with the Feentering garnet and Ti into rutile. Largevolumes of ilmenite-bearing mafic rockswere transformed into rutile-bearing eclo-gitic rocks with the same TiO2 content.From a mineral resource perspective, theilmenite deposits within the eclogite regi-ons in W. Norway are of only minor inte-rest, whereas the rutile-bearing eclogitesrepresent a major mineral resource, parti-cularly the Engebøfjell deposit (McLimanset al. 1999).

The third ore category is a variety of theProterozoic rutile-bearing, scapolitisedand albitised rocks in the Bamble -Arendal province (also called the BambleSector of the Fennoscandian Shield).These mafic rocks were extensively affec-ted by alkali-metasomatism in the form ofscapolitisation and albitisation. In thisprocess, ilmenite in the gabbroic proto-liths broke down to rutile during themetasomatic process, while iron wasremoved from the system by fluids. Thus,metasomatism creates rutile ores.

There are a considerable variety of diffe-rent types of titanium ore found inNorway (Korneliussen et al. 2000a, 2000b).There is also a close link between geodyna-mics, at all scales, and the evolution of tita-nium deposits. At a regional scale, the dis-tribution and character of the Ti-depositsare linked to distinct geological events atthe western and south-western margin ofthe Fennoscandian Shield, ranging in agefrom Proterozoic to Permian. The magma-tic, metamorphic and structural processesinvolved place distinct fingerprints on theindividual Ti-deposits.

From an economic-geological point ofview, the mineralogical variations are ofoverall importance, since mineralogy to alarge extent defines the mineability of a Tideposit. Mineralogy is a function of thegeological evolution of the deposit, whichis often very complex. In fact, a major taskin continued investigations of Ti mineraldeposits is to identify those that have ore-mineral qualities that will make it econo-mic to produce titanium mineral concen-trates of ilmenite and rutile of sufficientquality, particularly for the Ti-pigmentindustry.

The dominant Ti-ore category is that ofmagmatic deposits of ilmenite, titanom-agnetite and apatite in various proporti-ons, and deposits variably affected bysecondary, metamorphic processes. Themajor ilmenite province is the RogalandAnorthosite Province in southern Norway.It was emplaced in a Middle Proterozoic,migmatised, metamorphic terrane at inter-mediate crustal levels. Titanium depositsare widespread in the province, principal-ly as ilmenite-magnetite-rich cumulates,or as ilmenite-magnetite-rich intrusions.There is a wide variety in chemical charac-teristics and mineral textures in thesedeposits.

In some regions, metamorphic processes

On ilmenite and rutile ore provinces in Norway,and the relationships between geological processand deposit type

Are Korneliussen

Geological Survey of Norway, N-7491 Trondheim, Norway.



ReferencesKorneliussen, A., McEnroe, S. A., Nilsson, L.P.,Schiellerup, H., Gautneb, H., Meyer, G.B. &Størseth, L.R., 2000 a: An overview of titaniumdeposits in Norway. Norges geologiske undersø-kelse Bulletin 436, 27-38.

Korneliussen, A., McLimans, R., Braathen, A.,Erambert, M., Lutro, O. & Ragnhildstveit, J., 2000

b: Rutile in eclogites as a mineral resource inthe Sunnfjord region, western Norway. Norgesgeologiske undersøkelse Bulletin 436, 39-47.

McLimans, R.K, Rogers, W.T., Korneliussen, A.,Garson, M. & Arenberg, E. 1999: Norwegianeclogite: An ore of titanium. In: Stanley et al.(eds) Mineral Deposits: Processes to Processing.Balkema, Rotterdam, 1125-1127.

Ilmenite–zircon relationships in meta-anorthosites.Example from the Bergen Arc (Caledonides of WNorway) and implications for zircon geochronology

Bernard Bingen 1, Håkon Austrheim 2 & Martin Whitehouse 3

1 Geological Survey of Norway, 7491 Trondheim, Norway. 2 Mineralogisk Geologisk Museum, University ofOslo, 0562 Oslo, Norway. 3 Swedish Museum of Natural History, SE-104 05 Stockholm, Sweden.

The Lindås Nappe is situated in theCaledonides of West Norway and compo-sed mainly of a Proterozoic anorth-osite–charnockite complex. It was affectedby penetrative Sveconorwegian granulite-facies metamorphism, followed by a fluid-driven eclogite- and amphibolite-faciesCaledonian overprint. This overprint isspatially restricted along fractures andshear zones (Austrheim & Griffin 1985,Cohen et al. 1988, Boundy et al. 1992,Boundy et al. 1996, Austrheim et al. 1997,Bingen et al. 2001b). In this study, forma-tion of metamorphic zircon was investiga-ted in mafic lithologies. Petrographicobservations with optical microscopy,backscattered electron imaging and catho-doluminescence imaging were performedin thin- sections of whole-rocks and polis-hed mounts of zircon separates. SIMSanalyses (Cameca IMS1270) were realizedon zircon for age determination and esti-mation of trace-element concentrations(Th and REEs).

In mafic granulites and amphibolites, aluminescent anhedral zircon overgrowthsurrounds a magmatic zoned core. Theovergrowth gives an average age of 924 ±58 Ma with a Th/U = 0.52 and the mag-matic core gives an age of 952 ± 32 Mawith a Th/U = 1.27. The REE pattern ofzircon in a two-pyroxene granulite dis-plays a Ce positive anomaly and enrich-

ment in HREE. The REE pattern of theluminescent overgrowth is parallel to thatof the magmatic core, but is characterisedby distinctly lower concentrations. In thegranulites, a continuous rim of zircon or adiscontinuous corona of c. 10 µm roun-ded to flat zircon crystals is commonlyobserved in thin-section at the outer mar-gin of ilmenite grains. Baddeleyite (ZrO2)and srilankite (Ti2ZrO6) blebs are repor-ted around ilmenite included in feldsparor pyroxene. Baddeleyite is interpreted asan exsolution product from magmaticilmenite. Srilankite was probably formedas a reaction product between ilmeniteand baddeleyite during granulite-faciesmetamorphism in silica-deficient subsys-tems. The zircon corona around ilmeniteand the luminescent zircon overgrowthwere also formed as a reaction productduring granulite-facies metamorphism,where silica was available at grain bounda-ries. Textures suggest that magmatic ilme-nite was a main source of Zr to formmetamorphic zircon (Bingen et al. 2001a).

In massive amphibolites, relict ilmenitegrains are surrounded by a corona of tita-nite and a discontinuous corona of micro-zircons. Amphibolite-facies overprint isnot associated with any significant growthor dissolution of zircon.

An unsheared eclogite displays a zircon

population with a euhedral oscillatory-zoned overgrowth showing well-termina-ted edges and tips. The overgrowth yieldsan age of 424 ± 5 Ma (weighted average of31 analyses). It is characterised by a Th/Uratio lower than 0.13, a Ce positive ano-maly and an enrichment in HREE. TheREE pattern of the overgrowth is not sig-nificantly different from that of theProterozoic core, although some of theanalyses of the overgrowth are specificallypoor in all trace elements. In thin-section,a corona of micro-zircon grains is obser-ved at some distance around rutile, andlocally these micro-zircons show a pris-matic overgrowth. The specific low-Thzircon growth event is related to eclogite-facies forming reactions, involving break-down of a two-pyroxene + garnet + plagi-oclase + ilmenite assemblage to form agarnet + omphacite + rutile assemblage inthe presence of a fluid. The low Th contentof this zircon probably stems from thecoeval precipitation of clinozoisite. Thisoscillatory zoned zircon records fluidinfiltration and coeval Caledonian eclogi-tization in the crust.

ReferencesAustrheim, H., Erambert, M. & Engvik, A.K. 1997:Processing of crust in the root of theCaledonian continental collision zone: the roleof eclogitization. Tectonophysics 273, 129–153.

Austrheim, H. & Griffin, W.L. 1985: Shear defor-mation and eclogite formation within granuli-

te-facies anorthosites of the Bergen Arcs, wes-tern Norway. Chemical Geology 50, 267–281.

Bingen, B., Austrheim, H. & Whitehouse, M.2001a: Ilmenite as a source for zirconiumduring high-grade metamorphism? Texturalevidence from the Caledonides of W. Norwayand implications for zircon geochronology.Journal of Petrology 42, 355–375.

Bingen, B., Davis, W.J. & Austrheim, H. 2001b:Zircon U-Pb geochronology in the Bergen Arceclogites and their Proterozoic protoliths, andimplications for the pre-Scandian evolution ofthe Caledonides in western Norway. GeologicalSociety of America Bulletin 113, 640–649.

Boundy, T.M., Essene, E.J., Hall, C.M., Austrheim,H. & Halliday, A.N. 1996: Rapid exhumation oflower crust during continent-continent collisi-on and late extension: evidence from 40Ar/39Arincremental heating of hornblendes and mus-covites, Caledonian orogen, western Norway.Geological Society of America Bulletin 108,1425–1437.

Boundy, T.M., Fountain, D.M. & Austrheim, H.1992: Structural development and petrofabricsof eclogite facies shear zones in the granulitefacies complex, Bergen Arcs, W Norway: impli-cations for deep crustal deformational proces-ses. Journal of Metamorphic Geology 10, 127–146.

Cohen, A.S., O'Nions, R.K., Siegenthaler, R. &Griffin, W.L. 1988: Chronology of the pressure-temperature history recorded by a granuliteterrain. Contributions to Mineralogy andPetrology 98, 303–311.


Recent Re-Os studies of sulphide concen-trations in anorthosite-dominated pluto-nic suites consistently yield high initial187Os/188Os ratios in sulphide minerals,indicating a major component of crustalOs. These studies have led to three distincthypotheses for the origin of the sulphidemineralization, with implications for thegenesis of the hosting plutonic rocks: (1)

the source of the parental magmas is maficcontinental crust; (2) mantle-derived meltsachieve sulphide saturation and acquirecrustal Os through bulk assimilation ofsilicate crust; or (3) mantle-derived meltsachieve sulphide saturation and acquirecrustal Os through selective assimilationof crustal sulphur and metals by volatiliza-tion and/or melting of crustal sulphides.

Evidence from Re-Os for the origin of sulphide concentrations in anorthosites

Judith L. Hannah & Holly J. Stein

AIRIE Program, Colorado State University, Fort Collins, CO 80523-1482, USA.


Table 1. Some isotopic characteristics of sulphide-bearing intrusions in anorthosite-related plutonic suites.

References:1 Morgan et al., 2000; 2 Claesson & Ryka, 1999; 3 Wiszniewska et al., 1999b; 4 Wiszniewska & Jedrysek, 1998; 5 Schärer et al., 1996; 6 Schiellerup et al., 2000;7 Demaiffe et al., 1986; 8 Amelin et al., 1999; 9 Lambert et al., 2000; 10 Amelin et al., 2000; 11 Ripley et al., 1999; 12 Ripley et al., 1998; 13 Sproule et al., 1999.Notes:a For Suwalki, values are for magnetite, though microscopic sulphide inclusions may be present. For Rogaland and Voisey’s Bay, values are for related unmine-ralized intrusive rocks; negative numbers indicate disturbance of the Re-Os system (Lambert et al. 2000). For Duluth and Salay Malay, values are for weaklymineralized rocks or disseminated sulphide spatially removed from major concentrations. See text for discussion.b Schiellerup et al. (2000) report a much tighter range of εNd = 0.22 to 1.99 (n = 30) and 87Sr/86Sr = 0.705126 to 0.705822 (n = 18) c The Voisey’s Bay intrusion is hosted by troctolite; genetic association with nearby anorthosites is unclear.d Not true massif-type complexes; both are layered mafic-ultramafic intrusions.

Site γOs γOs εNd (87Sr/86Sr)i δ34S ‰(~ Age, Ma) sulfides othera

Suwalki (~1560) 1 +642 to +962 1 +743 to +961 1 -6.3 to -1.6 2 0.7046 to 0.7055 3 ~ 0 4

Rogaland (~930) 5 +419 6 b 0.0 to +4.6 7 b 0.7033 to 0.7062 7 --c Voisey’s Bay (~1330) 8 +898 to +1127 9 <0 to +84 9 -4.1 to +0.9 10 0.7034 to 0.7040 10 -4 to +2 11

d Duluth (~1100) 12 +500 to +1200 12 +3 12 -7.1 to -3.4 12 -- +10 to +14 12

d Sally Malay (~1845) 13 +950 to +1300 13 +60 to +479 13 -0.2 to +0.7 13 -- --

We favour the third hypothesis on thebasis of modelling of Re-Os data togetherwith other geologic and geochemical con-straints (Hannah & Stein 2002). Never-theless, the available data sets are limited,and we cannot yet preclude the possibilitythat real differences exist among the com-plexes examined to date.

Presently available Re-Os studies of anor-thosite-dominated plutonic suites (Table1) have focused on those with significantsulphide concentrations, including themajor ore deposits at Voisey’s Bay, Nainplutonic suite, Labrador (Lambert et al.1999, 2000), and less well-defined depositsin the Suwalki complex, Poland (Stein etal. 1998, Morgan et al. 2000), and theRogaland complex, Norway (Schiellerupet al. 2000). In each case, initial 187Os-188Os ratios in sulphides are extremelyhigh, indicating a crustal source for theOs. Re-Os data for two additional anor-thosite-bearing plutonic complexes alsoindicate crustal input: the 1845 Ma SallyMalay mafic-ultramafic intrusion in EastKimberly, Western Australia (γOs of abun-dant, disseminated sulphides = +950 to+1300; Sproule et al. 1999), and the 1100Ma Duluth layered igneous complex inMinnesota, USA (γOs of massive sulphides= +500 to +1200; Ripley et al. 1998).

In two cases, Suwalki and Rogaland, thehigh initial Os ratios have been attributedto a crustal source for the magmatic com-plex as a whole (Stein et al. 1998, Morganet al. 2000, Schiellerup et al. 2000). This

hypothesis is supported by a variety of field and mineralogical observations(Taylor et al. 1984, Duchesne 1999), oxy-gen isotope data (Peck & Valley 2000), andexperimental constraints (Longhi et al.1999). On the other hand, Re-Os datafrom Suwalki are limited to zones of mas-sive titanomagnetite containing up to afew percent sulphides; modelling demon-strates that observed Re-Os concentrati-ons and isotopic ratios can be explainedby selective contamination of a mantle-derived melt by crustal sulphides. In asubsequent and similar study at Rogaland,Schiellerup et al. (2000) call upon a crus-tal source for the parental magma basedon initial 187Os/188Os derived from a 10-point Re-Os isochron of 917 ± 22 Ma inagreement with zircon and baddeleyiteages (Schärer et al. 1996). While an iso-chron initial 187Os/188Os of 0.63 ± 0.25broadly supports this conclusion, the iso-chron has an extremely high MSWD (517)indicating variation in initial 187Os/188Osfor the samples analysed. That is, the sam-ples analysed do not form a perfectlycoherent genetic suite. Moreover, the datatable shows consistent variations in γOs forindividual samples, with sulphides gene-rally higher than oxides. It is conceivablethat, as at Suwalki, a larger data set woulddocument selective and variable crustalcontamination.

At Voisey’s Bay, chondritic γOs values forsilicates lead to a model of selective assi-milation of crustal sulphide (and Os) intoan otherwise mantle-derived magma


(Lambert et al. 2000). Similarly, both theDuluth and the Sally Malay complexesshow contrasts between the isotopic com-positions of massive sulphide bodies andminor disseminated sulphide occurrences.The massive sulphides of the Babbitt de-posit (Duluth) are strongly crustal, whiletroctolites with only disseminated sulphi-des yield epsilon Nd and δ34S of -0.2 to+0.2 and -2.5 to +2.5, respectively (Ripleyet al. 1998), typical of chondritic values.Similarly, for Sally Malay, Sproule et al.(1999) report γOs of +450 to +470 for sul-phides from ore-bearing troctolite, andonly +60 to +369 for sulphides from wea-kly mineralized troctolite and peridotite.These data support a model for selectivecontamination by crustal sulphides, withcrustal Os sequestered in an immisciblesulphide melt and not mixing broadlywith the complex as a whole.

ReferencesAmelin,Y., Li, C. & Naldrett, A.J. 1999: Geochrono-logy of the Voisey’s Bay intrusion, Labrador,Canada, by precise U-Pb dating of coexistingbaddeleyite, zircon, and apatite. Lithos 47, 33-51.

Claesson, S. & Ryka, W. 1999: Nd model ages ofthe Precambrian crystalline basement of NEPoland: Seventh Eurobridge Workshop –Between EUROBRIDGE and TESZ. AbstractVolume, Polish Geological Institute,Warsaw, 17-19.

Demaiffe, D., Weis, D., Michot, J. & Duchesne, J.C.1986: Isotopic constraints on the genesis ofthe anorthosite suite of rocks. ChemicalGeology 57, 167-179.

Duchesne, J.-C. 1999: Fe-Ti deposits inRogaland anorthosites (South Norway):geochemical characteristics and problems ofinterpretation. Mineralium Deposita 34, 182-198.

Hannah, J.L. & Stein, H.J. 2002: Re-Os model forthe origin of sulfide occurrences in Proterozoicanorthosite complexes. Economic Geology 97,371-383.

Lambert, D.D., Foster, J.G., Frick, L.R., Li, C. &Naldrett, A.J. 1999: Re-Os isotopic systematicsof the Voisey’s Bay Ni-Cu-Co magmatic oresystem, Labrador, Canada. Lithos 47, 69-88.

Lambert, D.D., Frick, L.R., Foster, J.G., Li, C. &Naldrett, A.J. 2000: Re-Os isotope systematics ofthe Voisey’s Bay Ni-Cu-Co magmatic sulfidesystem, Labrador, Canada: II. Implications for

parental magma chemistry, ore genesis, andmetal redistribution. Economic Geology 95, 867-888.

Longhi, J., Vander Auwera, J., Fram, M.S. &Duchesne, J.C. 1999: Some phase equilibriumconstraints on the origin of Proterozoic (mas-sif ) anorthosites and related rocks. Journal ofPetrology 40, 339-362.

Morgan, J.W., Stein, H.J., Hannah, J.L., Markey, R.J.& Wiszniewska, J. 2000: Re-Os study of Fe-Ti-Voxide and Fe-Cu-Ni sulfide deposits, Suwalkianorthosite massif, northeast Poland.Mineralium Deposita 35, 391-401.

Peck, W.H. & Valley, J.W. 2000: Large crustalinput to high δ18O anorthosite massifs of thesouthern Grenville Province: new evidencefrom the Morin Complex, Quebec. Contribu-tions to Mineralogy and Petrology 139, 402-417.

Ripley, E.M., Lambert, D.D. & Frick, L.R. 1998:Re-Os, Sm-Nd, and Pb isotopic constraints onmantle and crustal contributions to magmaticsulfide mineralization in the Duluth Complex.Geochimica et Cosmochimica Acta 62, 3349-3365.

Ripley, E.M., Park, Y.-R., Li, C. and Naldrett, A.J.1999: Sulfur and oxygen isotopic evidence ofcountry rock contamination in the Voisey’s BayNi-Cu-Co deposit, Labrador, Canada. Lithos 47,53-68.

Schärer, U., Wilmart, E. & Duchesne, J.-C. 1996:The short duration and anorogenic characterof anorthosite magmatism: U-Pb dating of theRogaland Complex, Norway. Earth andPlanetary Science Letters 139, 335-350.

Schiellerup, H., Lambert, D.D., Prestvik, T., Robins,B., McBride, J.S. & Larsen, R.B. 2000: Re-Os isoto-pic evidence for a lower crustal origin of mas-sif-type anorthosites. Nature 405, 781-784.

Sproule, R.A., Lambert, D.D. & Hoatson, D.M.1999: Re-Os isotopic constraints on the gene-sis of the Sally Malay Ni-Cu-Co deposit, EastKimberly, Western Australia. Lithos 47, 89-106.

Stein, H.J., Morgan, J.W., Markey, R.J. &Wiszniewska, J. 1998: A Re-Os study of theSuwalki anorthosite massif, northeast Poland.Geophysical Journal 4, 111-114.

Taylor, S.R., Campbell, I.H., McCulloch, M.T. &McLennan, S.M. 1984: A lower crustal origin formassif anorthosites. Nature 311, 372-374.


Effect of mineralogy and texture of sand and hard-rock ilmenite in TiO2 pigment production by the sulphate process, a case study on Australianilmenite concentrate and Tellnes ilmenite concentrate, Norway

Tegist Chernet

Geological Survey of Finland, 02151 Espoo, Finland.

The reactivity of the raw material withsulphuric acid is crucial in the productionof TiO2 pigments by the sulphate process.Ilmenite concentrates from sand andhard-rock deposits have been studied andthe results show that mineralogical cha-racteristics and textural features correlatewith the observed recovery rates.

Kemira pigments Oy, Pori, Finland, is tre-ating sand and hard-rock ilmenite con-centrates for TiO2 pigment production bythe sulphate route. The mineralogy andtextural relationships between the variousphases determine the reactivity of the con-centrate with sulphuric acid, the quality ofthe pigment and the separability of thephases. The study of commercial ilmeniteconcentrates provides vital informationfor the application of sulphate-route pig-ment manufacturing, by determining theinfluence on the production processes ofthe mineralogical characteristics of theraw materials.

Ilmenite concentrate from sand(Australia)The concentrate is composed of ilmeniteand its alteration products, several alter-ation phases being present in most singlegrains (Chernet 1999a). Alteration gradesfrom leached ilmenite through pseudorut-ile to leucoxene. Leached pseudorutile

decomposes into leucoxene that consistsof aggregates of very fine-grained rutilecrystals. The intensity and mode of alter-ation varies from grain to grain; mostcommonly, alteration appears along grainboundaries, fractures, crystallographicdirections and as leached holes. The con-centrate is composed of 53-55% ilmeniteplus leached ilmenite, 37% pseudorutileplus leached pseudorutile, and 8-10% leu-coxene plus primary rutile. Microprobeanalyses of ilmenite and its alteration pro-ducts show that trace element contentsincrease with alteration (Chernet 1999a).Ilmenite is soluble in sulphuric acid, whe-reas rutile is not (Sinha 1979). The degreeof solubility of ilmenite, however, dependson the degree of alteration. The concen-trate has been separated using a Frantzelectromagnetic separator. As the magne-tic susceptibility of ilmenite decreaseswith alteration, the nonmagnetic fractionswere recycled in an increasing electromag-netic field. Solubility tests show that solu-bility decreases with the extent of alterati-on (Chernet 1999a). The Th content ofthe concentrate, determined by XRF orICP, is 0.01%, whereas Ce, Th, Y, La andP2O5 contents in the <90 µm fractionshow a 6 to 10-fold increase, due to con-centration of monazite into the fine-grai-ned fractions. Uranium contents >3 ppmare only detected in the <90 µm fractions.

Wiszniewska, J., and Jedrysek, M.O. 1998:Preliminary δ13C and δ34S isotopic study onthe genesis of sulphide, carbonates and grap-hite mineralization in the mafic ore-bearingrocks of the Suwalki Massif, NE Poland (inPolish). Przeglad Geologiczny 46, no. 4, 359-364.

Wiszniewska, J., Duchesne, J.-C., Claesson, S.,Stein, H., and Morgan, J. 1999: Geochemicalconstraints on the origin of the Suwalki anor-thosite massif and related Fe-Ti-V ores, NEPoland: Seventh Eurobridge Workshop –Between EUROBRIDGE and TESZ, AbstractVolume, Polish Geological Institute,Warsaw, 89-91.

Hard-rock ilmeniteconcentrate (Norway)The concentrate is composed of 90%ilmenite (hemoilmenite with 8 to 13mol% hematite), 8% sulphides (pyrite,chalcopyrite, pyrrhotite) and silicates(hypersthene, augite, plagioclase, olivine,biotite), and rare spinel, magnetite,sphene and apatite. Textural features, i.e.exsolution and twinning (Chernet 1999b)are observed. The dissolution of ilmeniteand abundance of hematite exsolutionwere found to correlate. Ilmenite dissolvedalong grain boundaries with hematitelamellae and as a result hematite lenses arepartially leached out and left as orientedpits (Chernet 1999b). Twinning is relative-ly common in the concentrate but notobserved in the corresponding unreactedsolid. Cr and V are pigment colorants(Jalava 1992). The concentration of bothelements is closely linked to the presenceof exsolved hematite (0.25-0.45 wt%Cr2O3, 0.4-0.6 wt% V2O3). The concentra-te contains substantial amounts of ferriciron (32.9 wt% FeO and 12.3 wt% Fe2O3),which demands reduction process (Cher-net 1999b).

Discussion and conclusionsThere is a trend towards an increase inTiO2 content with decreasing solubility ofthe concentrate. Advanced alteration pro-ducts are less soluble than slightly alteredilmenite. Although microfractures andpits favour grinding and percolation ofthe leachate, the poorly soluble phases aregradually coating these openings, thusprogressively preventing the acid to reachthe soluble ilmenite. Textural relationshipswithin the different phases and continu-ous changes of mineral properties andcompositions prevent separation of highlyaltered (poorly soluble) phases from unal-tered (soluble) ones. Pigment colouringelements are generally in low contents inthe sand concentrates, and the presence ofcertain amounts of soluble Nb can correctthe colour caused by Cr and V. Th and Uare concentrated in the fine-grained frac-tion of the concentrate, mainly as monazi-te. Discarding the <90 µm fraction consi-derably reduces their accumulation in theprocess stream and in the tailings mud.

During the digestion step of the hard-rockilmenite concentrate, most of the hemati-

te particles will be leached out, and thepits give way to the percolation of the lea-chate, which resulted in a positive effecton the extraction process. No twinnedilmenite is observed in the correspondingunreacted solid, which indicates thatstructural discontinuities like twinningand fracturing favour dissolution process.The effect of trace elements on the qualityof the pigment is highly dependent onwhether it is incorporated in the lattice ofthe ilmenite or in insoluble phases. Cr andV in this concentrate are closely linked tohematite, which is insoluble.

Knowledge of Ti-bearing phases, inclu-ding their quantification and alterationproducts, can throw light on the rawmaterial treatment, enabling the metallur-gist and chemical engineers to deal withthe problems in TiO2 recovery. The con-tents and distribution of trace elementsand deleterious minerals provide a usefulguide to control the quality of both thefinal products and the tailings. To conclu-de, careful mineralogical examination ofany ilmenite concentrate is useful in eva-luating its suitability for a process.

ReferencesChernet, T. 1999a: Applied mineralogical studi-es on Australian sand ilmenite concentratewith special reference to its behaviour in thesulphate process. International Journal ofMinerals Engineering12(5), 485-495.

Chernet, T. 1999b: Effect of mineralogy and tex-ture in the TiO2 pigment production processof the Tellnes ilmenite concentrate. Mineralogyand Petrology 67, 21-32.

Jalava, J.P. 1992: Precipitation and properties ofTiO2 pigments in the sulphate process. 1.Precipitation of liquor and effects on Iron (II) inisoviscous liquor. Industrial Engineering andChemical Research 31, 608-611.

Sinha, H.N. 1979: Solubility of titanium minerals.Proceedings International Conference on Advan-ced Chemical Metallurgy 2, paper 35, 16 pp.



Parallel Re-Os isochrons and high 187Os/188Os initialratios: constraints on the origin of sulphide-bearinganorthosite systems

Holly J. Stein, Judith L. Hannah, John W. Morgan & Anders Scherstén

AIRIE Program, Department of Earth Resources, Colorado State University, Fort Collins, CO 80523-1482 USA.

In the past decade the Re-Os chronometerand the associated initial 187Os/188Os ratiohave provided powerful isotopic tools lea-ding to much improved understanding ofthe timing and source for ore-relatedmagmas as well as mineralization withouta direct magmatic association (Stein et al.2000, 2001, Arne et al. 2001). Notably, theRe-Os system applied to sulphide- andoxide-bearing anorthosite systems docu-ments unique processes involved in the for-mation and emplacement of mineralizedProterozoic anorthosite massifs (Hannah &Stein 2002).

Recognition of parallel isochrons and high187Os/188Os ratios reflecting synchronousemplacement with variable crustal conta-mination was first intimated at the Proter-ozoic Suwalki anorthosite complex inPoland (Stein et al. 1998; Morgan et al.2000). Petrologically, these parallel iso-chrons reflect complex mingling andaddition of crustally-derived Os withanorthosite and related melts. Stoping ofcrustal rocks by ascending magmas, armo-ring of crustal conduits, and influx of newbatches of melt making their way throughnew conduits result in highly variablemixtures of crustally-derived and primiti-ve Os (Hannah & Stein 2002). This resultsin highly variable initial 187Os/188Os ratiosin the intruding melts. Because of variableinitial 187Os/188Os with different melt bat-ches, it is essential that the resulting Re-Osdata be treated on a batch-by-batch basis,and not lumped together on a single iso-chron. Lumping of data from a large dataset may result in an isochron with anapproximately correct age, moderate tolow uncertainty, but high MSWD. A highMSWD, assuming assignment of appro-priate analytical errors, is a clear indicati-on of geological variation in the187Os/188Os initial ratio and that differentmagma batches with different Os isotopichistories are represented in the data popu-lation. Excess geological scatter resulting

from variable 187Os/188Os initial ratioscharacterises these data sets (e.g., Schiel-lerup et al. 2000; Lambert et al. 2000).Correct interpretation of Re-Os data forthese mineralized anorthosite systemsrequires a careful look at the geologic rela-tionships associated with the samplepopulation, and an analysis of the possiblebalance between crustally-derived Osobtained during the emplacement processand original mantle-derived Os inherentin the anorthosite melt.

Here we show Re-Os isochron plots fortwo data sets from two different sulphide-bearing anorthosite complexes. In Fig. 1,three different deposits within the Suwalkianorthosite complex in Poland yield Re-Os data that lie on two parallel isochronswith ages of ~1555-1560 Ma (Stein et al.1998; Morgan et al. 2000). Samples from asingle deposit define one isochron, whilecombined samples from two differentdeposits define the other. These isochronsare defined by magmatic magnetite (Re =0.4-1.5 ppb and Os = 0.036-0.144 ppb)and co-existing sulphide (pyrrhotite, chal-copyrite, pyrite, with Re = 30-55 ppb andOs = 1-6 ppb) hosted in anorthosite andrelated rocks. The 187Os/188Os initial ratiosfor the two isochrons are 1.16 ± 0.06 and0.87 ± 0.20, a clear indication of crustal Osand variable initial ratios. In both Figs. 1and 2, the size of the plotting symbol ismore than 100 times the analytical error.The uncertainties reported here are all atthe 2-sigma level.

In Fig. 2, the massive sulphides analysedare associated with an anorthosite-bea-ring, mafic intrusive complex. The localityis not named, as this was a collaborativeproject with a mining company. The Re-Os data for three samples containing chal-copyrite (cp) and pyrrhotite (po) areshown together with the present-day Osisotopic compositions for mantle andupper continental crust reservoirs. Since


the 187Os/188Os compositions of these tworeservoirs (mantle and crust) have evolved(increased) with time, it is important tonote that a lower 187Os/188Os would havecharacterised both reservoirs in the geolo-

gical past. The symbol shapes denote cpand po from the same hand sample; opensymbols denote cp and shaded symbolsrepresent po.

Fig. 1. Re-Os isochrons for magmatic sulphides and oxides from the Suwalki anorthosite massif,northeast Poland (from Morgan et al. 2000).

Fig. 2. Re-Os isochrons for sulphides (chalcopyrite and pyrrhotite) related to a Proterozoic anor-thosite in eastern Canada. The 187Os/188Os isotopic composition for reservoirs representing pre-sent-day mantle and average present-day eroding upper continental crust are shown by the dark shaded boxes along the vertical axis. Lighter shading represents 187Os/188Os values for otherpotential crustal contaminants which may extend to greater than 100 (e.g., see Figure 3 inHannah & Stein 2002).


For the example in Fig. 2, the Re and Osconcentrations are significantly higherthan at Suwalki (Re = 29.8-297.0 ppb andOs = 1.51-24.1 ppb). In all three samples,the Os concentrations for pyrrhotite areabout twice those of the accompanyingchalcopyrite. In two of the samples, pyr-rhotite has a higher Re concentration rela-tive to chalcopyrite by about twice, where-as in the third sample, the Re concentrati-ons for both minerals are similar. This sug-gests that at the hand-specimen scale, pyr-rhotite is more effective at taking Re andOs into its structure than chalcopyrite.

If we treat all six analyses together in aregression, we obtain an errorchron of1239 ± 77 Ma with a very high MSWD =61 and an initial 187Os/188Os of 1.81 ±0.15. A high MSWD implies excess scatterin the data set, that is, geological scatter inaddition to analytical scatter. The mostlikely source of geological scatter is a vari-able initial 187Os/188Os isotopic compositi-on; that is, the ore 'fluid' composition wasvariable. There are several ways to createthis situation, the most common beingmixing from two or more sources (e.g.,mixing of magma or magmatic fluid withOs acquired from the country rock, atdepth or at the site of emplacement).

A better resolution of the data set in Fig. 2is obtained using two regressions thatyield corresponding ages of 1209 ± 12 Maand 1201 ± 410 Ma. These regressions arenearly parallel and thus, yield essentiallythe same age. The very large uncertainty inthe 'dashed line' isochron (1201 Ma) is, ofcourse, a function of the fit of points tothe line, but also reflects the fact that thespread in these three points is half that forthe points on the 'solid line' isochron.What is most striking in these Re-Osresults is the indisputably high 187Os/188Osinitial ratios of 1.908 ± 0.020 ('solid line')and 1.86 ± 0.69 ('dashed line') associatedwith these sulphides. These very high ini-tial ratios require that the ore 'fluid' had alarge component of crustal Os. Also, thecrustal component was likely significantlyolder than 1200 Ma and may have contai-ned phases with high Re/Os ratios; this isnecessary to produce highly elevated187Os/188Os ratios by 1200 Ma, the appro-ximate time of contamination of the maficmagma. To add perspective, even the pre-sent-day 187Os/188Os ratios for 'averagecrust' cannot explain the data and, as we

move back in time to 1200 Ma, the crustal187Os/188Os ratios decrease. Thus, the ini-tial 187Os/188Os ratios associated with thesamples illustrated in Fig. 2 are extremelyhigh.

We recognize that the construction ofmultiple parallel isochrons may be neces-sary to characterise sulphide-bearing anor-thosite systems. A single regression for Re-Os data that plots as a band of scatteredanalyses leads to an approximate or per-haps erroneous age and initial 187Os/188Osresults. High MSWD values for a regressi-on are a clear indication of geological variation, likely tied to differing initial187Os/188Os ratios in samples. There ismuch to be learned by coupling the geolo-gical relationships with the multiple paral-lel isochron approach. In particular, preci-sion and accuracy are improved.

It is suggested that the apparent paucity ofsulphide in Proterozoic anorthosite sys-tems reflects the required happenstanceintersection of ascending anorthosite meltwith a crustal, sulphide-bearing rock pac-kage. In this way, sulphur with high187Os/188Os may be added to an anortho-site system of primitive derivation thro-ugh the assimilation or volatilization ofolder crustal sulphide.

ReferencesArne, D., Bierlein, F., Morgan, J.W. & Stein, H.J.2001: Re-Os dating of sulfides associated with gold mineralization in central Victoria,Australia. Economic Geology 96, 1455-1459.

Hannah, J.L. & Stein, H.J. 2002: Re-Os model forthe origin of sulfide deposits in anorthosite-associated intrusive complexes. EconomicGeology 97, 371-383.

Lambert, D.D., Frick, L.R., Foster, J.G., Li, C. &Naldrett, A.J. 2000: Re-Os isotope systematicsof the Voisey's Bay Ni-Cu-Co magmatic sulfidesystem, Labrador, Canada: II. Implications forparental magma chemistry, ore genesis, andmetal redistribution. Economic Geology 95,867-888.

Morgan, J.W., Stein, H.J., Hannah, J.L., Markey,R.J. & Wiszniewska, J. 2000: Re-Os study of oresfrom the Suwalki anorthosite massif, northeastPoland. Mineralium Deposita 35, 391-401.

Schiellerup, H., Lambert, D.D., Prestvik, T., Robins,


B., McBride, J.S. & Larsen, R.B. 2000: Re-Os isoto-pic evidence for a lower crustal origin of mas-sif-type anorthosites. Nature 405, 781-784.

Stein, H.J., Morgan, J.W., Markey, R.J. &Wiszniewska, J. 1998: A Re-Os study of theSuwalki anorthosite massif, northeast Poland.Geophysical Journal 4, 111-114.

Stein, H.J., Morgan, J.W., & Scherstén, A. 2000:Re-Os dating of low-level highly-radiogenic(LLHR) sulfides: the Harnäs gold deposit, south-west Sweden records continental scale tecto-nic events. Economic Geology 95, 1657-1671.

Stein, H.J., Markey, R.J., Morgan, J.W., Hannah, J.L.& Scherstén, A. 2001: The remarkable Re-Oschronometer in molybdenite: how and why it works. Terra Nova 13, 479-486.


Recent Re-Os isotope data on the c. 0.93Ga Rogaland Anorthosite Province (RAP)suggest an origin by partial melting of amafic, lower crustal protolith, without acontribution of mantle-derived Sveconor-wegian material (Schiellerup et al. 2000).Without necessarily challenging this conc-lusion, it should be noted that there isample evidence of mafic magmatism inSouth Norway during the Sveconorwegianperiod (e.g., Munz & Morvik 1991, Haaset al. 2000), which suggests that rejuvena-tion of the lower continental crust byunderplating of mantle-derived magmascould have taken place. The emplacementof the anorthosites and associated rocks ofthe RAP corresponds in time to a wides-pread event of post-tectonic, A-type gra-nitic magmatism in SW Scandinavia,which can be used to map the distributionof deep crustal components across thispart of the Baltic Shield (Andersen &Knudsen 2000). Sr, Nd and Pb isotopedata on late Sveconorwegian granites defi-ne mixing trends between crustal compo-nents and a component with a juvenileisotopic signature (Andersen et al. 2001a).As pointed out by Andersen (1997) andSchiellerup et al. (2000), the presence ofan end member with a positive epsilon Nddoes not necessarily indicate that mantle-derived magmas were injected into thecrust at the time of deep crustal melting.In fact, such a component may reside inmafic rocks within the deep crust for seve-ral hundred million years without totallylosing its depleted-mantle Nd isotopic sig-nature.

The hafnium isotope composition of zir-con in an igneous rock is a powerful tracerof interaction between the mantle and thelower crust in the past. New laser ablationICP-MS hafnium isotope data on zircons

in Proterozoic granitoids from SouthNorway (Andersen et al. 2002) allow a bet-ter characterization of crustal protoliths inSouth Norway than has previously beenpossible, and provide a definitive test forthe presence or absence of Sveconor-wegian, mantle-derived material in thedeep crust of the southwestern part of theBaltic Shield.

In terms of Hf isotope systematics, twodistinct pre-Sveconorwegian crustal com-ponents can be recognized in SouthNorway:

i: A Paleoproterozoic component, charac-terized by depleted-mantle Hf isotopemodel ages (tHf

M) > 1.80 Ga. This compo-nent resides in the granitoids of the 1.65-1.83 Ga Transscandinavian Igneous Belt(TIB). U-Pb ages and Hf isotope data on

FennoscandiaSveconorwegian rejuvenation of the lower crust inSouth Norway

Tom Andersen 1, William L. Griffin 1,3, Norman J. Pearson 2 & Arild Andresen 1

1 Institutt for geologi, Postboks 1047 Blindern, N-0316 Oslo, Norway. 2 GEMOC, Department of Earth andPlanetary Sciences, Macquarie University, NSW 2109, Australia. 3 CSIRO Exploration and Mining, NorthRyde, NSW 2113, Australia.

Fig. 1. Cumulative probability plots of time-corrected 176Hf/177Hf of zircons fromSveconorwegian granitoids from SouthNorway, compared to regional hafnium-isotope components. The black parts ofthe distribution curves represent zirconswhich cannot have crystallized from amagma formed only from the regionalcrustal protoliths, and which thereforerequire the presence of a Sveconor-wegian, mantle-derived component in the lower crust.


detrital zircons from clastic metasedi-ments indicate that equivalent rocks werepresent in the source region(s) ofProterozoic sediments in the Telemarkand Bamble sectors (Haas et al. 1999). Sr,Nd and Pb data on late Sveconorwegiangranites suggest that a similar componentcan be traced in the deep crust from thearea east of the Oslo Rift, across SouthNorway to the area west of the Mandal-Ustaoset shear zone (Andersen et al.2001a). In late Sveconorwegian time, thePaleoproterozoic crustal component wascharacterized by a very distinct hafniumisotope composition, with 176Hf/177Hf ≤0.2819 (Fig. 1).

ii: A Mesoproterozoic component withtHf

M in the range 1.55-1.75 Ga, residing in1.52-1.60 (-1.66 ?) Ga calc-alkaline, meta-igneous rocks. These rocks are the pro-ducts of subduction-related magmatismalong the SW margin of the Baltic Shield,and can be traced from SW Sweden, acrossthe Oslo Rift and into the southwestern-most part of the Bamble Sector (Andersenet al. 2001b). Both juvenile, mantle-deri-ved material and material with a crustalprehistory were involved in the petrogene-sis of these rocks, whose Hf isotope com-position in late Sveconorwegian timespans a narrow range at 176Hf/177Hf ≈0.28215 (Fig. 1).

A juvenile, depleted-mantle componentwas involved in the formation of theMesoproterozoic calca-lkaline rocks, andmay later have contributed to the lowercrust in one or more underplatingevent(s). The Hf isotopic composition of adepleted mantle reservoir in Sveconor-wegian time can be given by 176Hf/177Hf >0.28245 (Fig. 1).

Fig. 1 shows cumulative probability distri-bution curves of initial 176Hf/177Hf ratiosof suites of single zircons separated fromSveconorwegian granitoids. All threegroups give broad distribution curves,which suggest that the zircons crystallizedfrom isotopically inhomogeneous mag-mas. Whereas the low-176Hf/177Hf parts ofthe ranges suggest the involvement of aPaleoprotorozoic crustal component, asignificant proportion of the zircons fromthe granitoids show more radiogenic176Hf/177Hf than the Mesoprotoerozoiccrustal component. This indicates theinvolvement of Sveconorwegian mantle-

derived material in the petrogenesis ofthese rocks.

The Tromøy complex is a fragment of aSveconorwegian island-arc system, accre-ted onto the Baltic Shield prior to granuli-te-facies metamorphism at c. 1.10 Ga(Knudsen & Andersen 1999). The mantle-derived component in these rocks wasintroduced in the island-arc setting, anddoes not give evidence of underplating ofthe continental crust. On the other hand,both the 1.15 Ga augen-gneiss and the0.93-1.10 Ga granites were emplaced wit-hin the continent, and the presence of asignificant component of mantle-derivedhafnium in their zircons is a clear indica-tion that the granitic magmas did notform by simple remelting of a Mesopro-toerozoic deep crustal protolith. Input ofmantle-derived material to the lower crustduring the Sveconorwegian period is thusnecessary to account for the initial Hf iso-tope characteristics of these rocks.

ReferencesAndersen, T. 1997: Radiogenic isotope systema-tics of the Herefoss granite, South Norway: anindicator of Sveconorwegian (Grenvillian) crus-tal evolution in the Baltic Shield. ChemicalGeology 135, 139-158.

Andersen, T. & Knudsen, T.-L. 2000: Crustal con-taminants in the Permian Oslo Rift, SouthNorway: Constraints from Precambriangeochemistry. Lithos 53, 247-264.

Andersen, T., Andresen, A. & Sylvester, A.G.2001a: Nature and distribution of deep crustalreservoirs in the southwestern part of theBaltic Shield: Evidence from Nd, Sr and Pb iso-tope data on late Sveconorwegian granites.Journal of the Geological Society, London 158,253-267.

Andersen, T., Griffin, W.L., Jackson, S.E. &Knudsen, T.-L. 2001b: Timing of mid-Proterozoic calcalkaline magmatism across theOslo Rift: Implications for the evolution of thesouthwestern margin of the Baltic Shield.Geonytt 2001, 1, 25.

Andersen, T., Griffin, W.L. & Pearson, N.J. 2001:Crustal evolution in the SW part of the BalticShield: The Hf isotope evidence. Journal ofPetrology 43, 1725-1747.

Haas, G.J.L.M. de, Andersen, T. & Nijland, T.J.


2000: Mid-Proterozoic evolution of the SouthNorwegian lithosphere. Geonytt 2000, 1, 56.Haas, G.J.L.M. de, Andersen, T. & Vestin, J. 1999:Application of detrital zircon geochronologyto assembly of a Proterozoic terrain – anexample from the Baltic Shield. Journal ofGeology 107, 569-586.

Knudsen, T.L. & Andersen, T., 1999: Petrologyand geochemistry of the Tromøy calc-alkalinegneiss complex, South Norway, an allegedexample of Proterozoic depleted lower conti-nental crust. Journal of Petrology 40, 909-933.

Munz, I.A. & Morvik, R. 1991: Metagabbros inthe Modum complex, southern Norway: animportant heat source for Sveconorwegianmetamorphism. Precambrian Research 52, 97-113. Corrigendum: Precambrian Research 53,305.

Schiellerup, H., Lambert, D.D., Prestvik, T., Robins,B., McBride, J.S. & Larsen, R.B. 2000: Re-Os isoto-pic evidence for a lower crustal origin of mas-sive-type anorthosites. Nature 405, 781-784.

Thermal modeling of the metamorphism of theRogaland granulites

John C. Schumacher & Mathias Westphal

University of Bristol, Dept. of Earth Sciences, Wills Memorial Building, Queen's Road, Bristol BS8 1RJ, UK.

The Proterozoic Egersund AnorthositeComplex is composed of anorthosite andrelated rock types, which intrude interca-lated charnockitic and garnetiferous mig-matites. The highest metamorphic tempe-ratures are spatially associated with theintrusion, and previous workers recogni-zed high-temperature mineral isogradsfor pigeonite-in, osumilite-in and ortho-pyroxene-in. The P-T conditions at a dis-tance of 16 km from the contact are about700°C and 5 kbar and increase to morethan 1000°C and 5 kbar at 2.5 km.

The high-temperature metamorphism inthe country rocks found near the contactsof the northeastern part of the Rogalandintrusive complex (surface area of about1000 square km) cannot be explained byassuming the heat source was a simple,single phase of intrusion. In order toobtain the observed metamorphic tempe-ratures and isograd distribution, thermalmodeling indicates that the heat sourcemust have had at least two main phasesthat were separated by a hiatus of about 3- 4.5 m.y. In this model, emplacement andcrystallization of the anorthosite (arealextent: 30 x 40 km) produces a thermalgradient (750–600°C) in the country rocks.While this thermal gradient is developing,a second, smaller intrusion (Bjerkreim-Sokndal lopolith, areal extent: 9 x 12 km)

is emplaced at the anorthosite-countryrock contact. Because the country rocksnearest the anorthosite have undergoneappreciable heating, the second intrusioncan provide enough thermal input to

Fig. 1. The diagram depicts the thermalmodel of the northeastern part of theRogaland intrusive complex, and the distri-bution of maximum temperatures andduration isochrons, which indicate thetotal amount of time that the rock spentabove 750°C. See text for discussion.


obtain the observed high temperatures.Recent work (Schärer et al. 1996) indicatesthat the entire magmatic emplacementoccurred over a time interval of about 10m.y. (930-920 Ma), which is consistentwith the thermal model of the metamor-phism.

From the thermal modeling experiment,we have estimated the distribution ofmaximum temperatures and duration iso-chrons, which represent the length of timethe rocks spent above 750°C during hea-ting and cooling (Fig. 1). The durationisochrons roughly parallel the contact ofthe initial and larger intrusion, while themaximum temperature isotherms arestrongly deflected by the smaller and laterintrusion. This indicates that, while thesecond intrusion has a major impact onthe local temperature distribution, it doesnot locally prolong the heating and coo-ling cycle. The 750°C isotherm and theduration isochon for 0-5 m.y are roughlyparallel over the entire area. The durationisochrons of 10 to 14 m.y. crosscut theisotherms at steep angles near the smallerand later intrusion (near point C in Fig.1), while near point A, at greater distancesfrom the smaller intrusion, the isothermsand duration isochrons are nearly parallel.

These relative orientations of the isot-herms and the duration isochrons showthe interaction of the two intrusions. Theeffects are strongest in the area betweenthe intrusions (near point C in Fig. 1), andthe effect of the smaller intrusion dissipa-tes noticeably at distances greater than 10km (near point B in Fig. 1). The areaaround point C (Fig. 1), where the isot-herms and duration isochrons are mostoblique, gives an impression of the exten-sive variation in heating and cooling ratesthat can occur over relatively small areas.

Figure 2 shows the variation of tempera-ture with time for the four lettered pointsin Figure 1. Point A is located 5 km fromthe intrusive contact along the extensionof the long axis of the large, ellipticalintrusion. Point B is located at 5 km dis-tance to the large and 10 km to the smallintrusion and lies on 10 m.y. duration iso-chrons. The initial rapid heating and longperiod of cooling for these localities (Aand B) are controlled by the anorthosite(Fig. 2). The difference in the temperatu-res at points A and B after 3 m.y. reflectsthe geometrical effect (elliptical shape) ofthe modeled intrusion (see also point C).

For point C the heating and cooling histo-ry is radically different from points A andB (Fig. 2). The variations in heating andcooling rates and the higher maximumtemperature are due to the major influen-ce of the second intrusion (modeled BSL).

At point D the thermal effects of thesecond intrusion on the country rocks(heating and cooling rates) are more dras-tic than at point C, but the maximum tem-perature is lower because of the lower ini-tial temperatures at the time of the secondintrusion. The major differences betweenthe thermal development at points C andD (Fig. 2) are due to differences in coun-try-rock temperature at the time of thesecond intrusion (i.e., the magnitude ofthe initial thermal gradient caused by theanorthosite).

The P-T estimates from 5 and 10 km fromthe intrusive contact and data from thethermal model were also used to estimatecooling rates and time intervals. Maxi-mum temperatures are 860°C and 800°Cand temperatures greater than 750°C(apparent end of new garnet growth) weremaintained for 4.5 to 8.5 m.y. About 5 km

Fig. 2. The diagram shows the variation intemperature versus time based on thethermal modeling experiment. The fourcurves are for the points (A-D) on Figure 1.


from the intrusion, calculated post-peakcooling rates indicate a decrease from 30°to 10°/m.y. over about 7 m.y., while atabout 10 km calculated post-peak coolingrates decrease from 12° to 7°/m.y. overabout 7 m.y. During this time interval,retrograde and fine-grained Gar-Qz rimsformed around orthopyroxene and pri-mary garnet (outermost 150-200 µm) pre-sent in assemblages of garnet-quartz-pla-gioclase-orthopyroxene ± spinel. Based onthe widths of preserved zoning profiles ingarnet and orthopyroxene and diffusiondata, estimates of cooling rates can bemade, and these agree well with the calcu-lated temperatures and cooling rates from

the model. Where garnet and orthopyro-xene are in direct contact, the retrogradeexchange may have continued as long as15-25 m.y., which resulted in more highlydeveloped Fe/Mg-zoning at the rims(outermost 40-240 µm) of both orthopy-roxene and garnet.

ReferenceSchärer, U., Wilmart, E. & Duchesne, J.-C. 1996:The short duration and anorogenic characterof anorthosite magmatism: U-Pb dating of the Rogaland complex, Norway. Earth andPlanetary Science Letters 139, 335-350.

Re-Os dating of the Ørsdalen W–Mo district inRogaland, S Norway, and its relationship toSveconorwegian high-grade metamorphic events

Bernard Bingen 1 & Holly Stein 1,2

1 Geological Survey of Norway, Leiv Eirikssons vei 39, 7491 Trondheim, Norway. 2 AIRIE Program,Department of Earth Resources, Colorado State University, Fort Collins, CO 80523-1482 USA.

In the Rogaland–Vest Agder sector of theSveconorwegian orogen (S Norway), agranulite-facies metamorphic domain isexposed around the 931 ± 3 Ma RogalandAnorthosite Province (Duchesne et al.1985, Schärer et al. 1996). This domain ischaracterized by a succession of threemain isograds with increasing grade to-wards the west: the orthopyroxene isogradin leucocratic rocks, the osumilite isogradin paragneiss, and the pigeonite isograd inleucocratic rocks (Tobi et al. 1985). Theosumilite and pigeonite isograds are indi-cative of very high-temperature and dryconditions. Osumilite-bearing assembla-ges provide P–T estimates of 5.5kbar–800-850°C (Holland et al. 1996).There is little doubt that osumilite-bea-ring assemblages developed as a result ofgranulite-facies contact metamorphismrelated to the intrusion of the anorthositecomplex (M2 metamorphism), as the osu-milite isograd is parallel to the externalcontact of the plutonic complex. Availablepetrological and geochronological data onmonazite in the region nevertheless pointto high-grade metamorphism, possibly ingranulite-facies conditions, before the

intrusion of the anorthosite complex (M1metamorphism; Tobi et al. 1985, Bingen &van Breemen 1998). In this work, we usethe Re-Os chronometer applied to molyb-denite hosted in small metamorphic de-posits, to further explore the geochrono-logy of polyphase metamorphism inRogaland.

Reliable and precise ages for molybdenitecrystallization can be obtained by theRe–Os method using improved analyticalmethods (Markey et al. 1998, Stein et al.2001). The Re–Os chronometer applied tomolybdenite has been shown to maintainits systematics through high-grade meta-morphism (Stein et al. 1998, Raith & Stein2000). The method is thus suitable forobtaining primary ages of molybdenitedeposition, regardless of metamorphicoverprinting. Among the small molybde-nite-bearing deposits in Rogaland, theØrsdalen district was mined for wolframi-te, scheelite, and molybdenite between1904 and 1956. It is located between theorthopyroxene and osumilite isograds. Itoccurs in a granitic to granodioritic gneissunit that commonly contains metamor-


phic garnet and/or orthopyroxene. Themineralized zone is restricted to two ca. 50m thick layers. These layers are composedof biotite gneiss with amphibolite lensesand contain glassy quartz + feldspar leu-cocratic veins up to 1 m wide, both elon-gate parallel to the gneiss foliation (Heier1955, 1956, Urban 1974, Olerud 1980).The W-Mo mineralization is closely asso-ciated with the quartz-rich veins.

Samples were collected at the Mjåvass-knuten (MJ1 and MJ2) and Stopulen(ST1) prospects, situated within the twomineralized layers. Sample MJ1 is takenfrom irregular, coarse-grained, glassyquartz + plagioclase + orthoclase leuco-cratic masses that are associated with afine-grained and dark-coloured granulite,rich in brown-red biotite. Both the leuco-cratic masses and the host contain disse-minated coarse-grained dark red garnetand orthopyroxene. Molybdenite occursas clots and patches larger than 1 cm thatare made of fine- to coarse-grained crys-tals. It is dominantly hosted in the leuco-cratic masses and associated with scheeli-te. Molybdenite, biotite and Fe-Ti oxideminerals are commonly surrounded by afine-grained symplectite of garnet +quartz. Samples MJ2 and ST1 are fromconcordant, clear, glassy quartz-rich veinsin biotite gneiss. The molybdenite is con-centrated along vein margins as mm- tocm-sized blebs to stringers.

The Re-Os data for the three molybdenitesamples are presented in Fig. 1 with onereplicate analysis for MJ2. All model agesinclude the uncertainty in the 187Re decayconstant and overlap at the 2-sigma levelof uncertainty. Their weighted mean age is971.4 ± 3.6 Ma. The four data points pro-vide a Model 3 isochron age (Ludwig1999) of 973.3 ± 8.2 Ma with the expectedzero initial 187Os intercept of -0.10 ± 0.88(2-sigma). Inclusion of the zero point,implicit in Re–Os model ages for molyb-denite and as demonstrated in the four-point regression, provides essentially thesame Model 3 age of 972.9 ± 3.8 Ma. Thefive-point regression contributes to thereduced uncertainty.

The agreement of four Re-Os ages usingthree different samples is a clear verifica-tion of a robust Re-Os chronometer inmolybdenite, and points to an event ofdeposition of molybdenite at about 973

Ma. This age is younger than any possibledeposition age or magmatic crystalliza-tion age for the amphibolite lenses associ-ated with the deposit. A presumablyMesoproterozoic stratiform volcanogenicorigin for the deposit, as proposed byUrban (1974) is thus not supported bythis new data. The field and petrographicrelations in the Mjåvassknuten localityindicate that the leucocratic molybdenite,garnet- and orthopyroxene-bearing quartz+ feldspar masses represent veins andpods of partial melt formed during anevent of granulite-facies metamorphism.By direct inference, we propose that themolybdenite age of 973 Ma is directly rela-ted to the formation of the leucocraticveins and pods. This event is related to thegranulite-facies metamorphic and defor-mation event that produced coarse-grai-ned metamorphic assemblages, produceda strong foliation in the country rocks,produced leucocratic veins and lenses, andaligned these veins within this foliation.The occurrence of garnet + quartz sym-plectite around molybdenite shows thatthe deposit was affected by an event of sta-tic high-grade metamorphism after depo-sition (M2 contact metamorphism atabout 931 Ma). The Re–Os systematics inmolybdenite were not affected by this sub-sequent event.

The age of 973 Ma for molybdenite depo-sition lies within the range of proposedtimings for Sveconorwegian metamor-phism in the orogen. In the easternmostsector of the orogen in SW Sweden, ahigh-pressure granulite-facies domainwith local occurrence of eclogite-faciesrocks is well constrained between 980 and960 Ma (Johansson et al. 2001). This meta-morphism probably resulted from thrus-ting of the allochthonous terranes onto theforeland of the orogen. In the westernmostsector of the orogen, in Rogaland, single-grain monazite U–Pb ages define a rangebetween 1024 and 970 Ma for M1 regionalmetamorphism and between 930 and 925Ma for M2 contact metamorphism(Bingen & van Breemen 1998). The occur-rence of negatively discordant monazitegrains at 970 Ma in some biotite-bearinggneiss samples suggests that the age of 970Ma is geologically significant. The newdata on the Ørsdalen district suggest for-mation of mineralized quartz + feldspargranulite-facies veins at 973 Ma, and thusgive independent evidence that M1 regio-


nal metamorphism reached granulite-faci-es conditions at 973 Ma. Available datasuggest a protracted nature for M1 meta-morphism between ca. 1025 and 970 Ma.They also point to the superposition oftwo granulite-facies metamorphic eventsin Rogaland (M1 and M2 events), and thatosumilite-bearing mineral associationsformed during the second of the twoevents. Superposition of two granulite-facies events may be required to sufficient-ly dehydrate a gneissic basement so thatosumilite-bearing assemblages are develo-ped and preserved. The necessity for twosuperimposed stages of granulite-faciesmetamorphism may explain the rarity ofosumilite-bearing assemblages in granuli-te-facies domains.

AcknowledgementsThis work was supported by theGeological Survey of Norway and by aFulbright Senior Research Fellowship toHJS in 2000. The Re-Os analyses weremade by Mr. Richard Markey at AIRIE,Colorado State University.

ReferencesBingen, B. & van Breemen, O. 1998: U–Pbmonazite ages in amphibolite- to granulite-facies orthogneisses reflect hydrous mineralbreakdown reactions: SveconorwegianProvince of SW Norway. Contributions toMineralogy and Petrology 132, 336-353.

Duchesne, J.-C., Maquil, R. & Demaiffe, D. 1985:The Rogaland anorthosites: facts and specula-tions. In Tobi, A.C. & Touret, J.L. (eds.) The deepProterozoic crust in the north Atlantic provinces.NATO-ASI C158, 449-476. Reidel, Dordrecht.

Heier, K.S. 1955: The Ørsdalen tungsten deposit.Norsk Geologisk Tidsskrift 35, 69-85.

Heier, K.S. 1956: The geology of the Ørsdalendistrict. Rogaland, S. Norway. Norsk GeologiskTidsskrift 36, 167-211.

Holland, T.J., Babu, E.V. & Waters, D.J. 1996: Phaserelations of osumilite and dehydration meltingin pelitic rocks: a simple thermodynamicmodel for the KFMASH system. Contributions toMineralogy and Petrology 124, 383-394.

Johansson, L., Möller, C. & Söderlund, U. 2001:Geochronology of eclogite facies metamor-phism in the Sveconorwegian Province of SW

Sweden. Precambrian Research 106, 261-275.

Ludwig, K.R. 1999: User's manual for Isoplot/Exversion 2.02, a geochronological toolkit forMicrosoft Excel. Berkeley GeochronologyCenter Special Pubication, Berkeley, USA.

Markey, R.J., Stein, H.J. & Morgan, J.W. 1998:Highly precise Re–Os dating for molybdeniteusing alkali fusion and NTIMS. Talanta 45, 935-946.

Olerud, S. 1980: Geologiske og geokjemiskeundersøkelser rundt Ørsdalen W–Mo fore-komst. Norges geologiske undersøkelse Rapport1650/51A, 34 pp.

Raith, J.G. & Stein, H.J. 2000: Re–Os dating andsulfur isotope composition of molybdenitefrom tungsten deposits in westernNamaqualand, South Africa: Implications forore genesis and the timing of metamorphism.Mineralium Deposita 35, 741-753.

Schärer, U., Wilmart, E. & Duchesne, J.-C. 1996:The short duration and anorogenic characterof anorthosite magmatism: U–Pb dating of theRogaland complex, Norway. Earth andPlanetary Science Letters 139, 335-350.

Stein, H.J., Sundblad, K., Markey, R.J., Morgan,J.W. & Motuza, G. 1998: Re-Os ages for Archeanmolybdenite and pyrite, Kuittila-Kivisuo,Finland and Proterozoic molybdenite, Kabeliai,Lithuania: testing the chronometer in a meta-morphic and metasomatic setting. MineraliumDeposita 33, 329-345.

Stein, H.J., Markey, R.J., Morgan, J.W., Hannah, J.L.& Scherstén, A. 2001: The remarkable Re-Oschronometer in molybdenite: how and why itworks. Terra Nova 13, 479-486.

Tobi, A.C., Hermans, G.A., Maijer, C. & Jansen,J.B.H. 1985: Metamorphic zoning in the high-grade Proterozoic of Rogaland-Vest Agder, SWNorway. In: Tobi, A.C. & Touret, J.L. (eds.) Thedeep Proterozoic crust in the north Atlantic pro-vinces. NATO-ASI C158, 477-497. Reidel,Dordrecht.

Urban, H. 1974: Zur Kenntnis der präkambris-chen, schichtgebundenen Molybdänitvor-kommen in Südnorwegen. GeologischeRundschau 63, 180-190.


A detailed magnetic study has beenundertaken on the Egersund-Ogna anor-thosite body of the Rogaland AnorthositeProvince, southeastern Norway. The anor-thosite, with published U-Pb ages of 930Ma, is part of a larger complex of massif-type anorthosites and the Bjerkreim-Sokndal layered intrusion. Thirteen pale-omagnetic sites were collected in the body,representing various parts of the anortho-site. Average susceptibilities range from0.03 to 2.24 x 10-3 SI and NRM intensiti-es range from 0.004 to 1.54 A/m.Corresponding Q values range from 3 to148 with a mean value of 36, indicatingremanent-controlled magnetic anomalies.Remanent directions from all samples arecharacterized by steep negative inclinati-ons with southwest to northeast variabledeclinations. Thermal demagnetizationreveals square shouldered demagnetiza-tion curves, with little or no loss of inten-sity until 550 or 575° C. Alternating fielddemagnetization produces a wide range ofdemagnetization behaviors with mean

destructive fields varying from less than 5mT to greater than 80 mT. There is littleevidence of overprinting or secondarycomponents, and all information pointsto a remanence gained during initial coo-ling of the anorthosite. Mean directionaldata for the 13 sites are I = -81.7° and D =326.8°, a95 = 6.0. Assuming this meandirection represents normal polarity, pale-olatitude for southern Fennoscandia atthis time is 70°S. The magnetic pole calcu-lated for Egersund-Ogna is, at -44° latitu-de and 198° longitude, in good agreementwith earlier poles determined from otherRogaland Anorthosite Province rocks.This work supports reconstructions thatplace Baltica in high (southern) latitude atapproximately 900 Ma. Apparent polarwander paths for Baltica at this time areambiguous with both clockwise or coun-ter-clockwise loops proposed. These dataconfirm the presence of mid-latitudepoles in the Late Proterozoic, but at thispoint are unable to discern between sug-gested scenarios.

Tectonic setting of the Egersund-Ogna anorthosite,Rogaland, Norway, and the position of Fennoscandiain the Late Proterozoic

Laurie Brown 1, Suzanne McEnroe 2, Mark Smethurst 2

1 Dept. Geosciences, Univ. Mass., Amherst MA 01003, USA. 2 Geological Survey of Norway, N-7491Trondheim, Norway.

Sulphides in the Rogaland Anorthosite Province

Henrik Schiellerup 1, David D. Lambert 2 & Brian Robins 3

1 Geological Survey of Norway, N-7491 Trondheim, Norway. 2 Division of Earth Sciences, National ScienceFoundation, Arlington, VA 22230, USA. 3 Department of Earth Science, The University of Bergen, N-5007Bergen, Norway.

The Rogaland Anorthosite Province inSouthwest Norway is a composite massif-type anorthosite province consisting of anumber of anorthositic plutons, mafic,dominantly jotunitic or noritic intrusions,as well as a collection of granitoid bodies.Many mafic intrusions form significantdeposits of Fe-Ti oxides.

Sulphide occurrences in Rogaland may be

divided into two types. 1: minor dissemi-nated, lumpy or semi-massive occurrenceswithin anorthosite bodies, and 2: dissemi-nated sulphides hosted by noritic ororthopyroxenitic cumulates believed toderive from jotunitic magmas. Type 2occurrences include a number of noriticintrusions that contain dispersed sulphidedroplets, but only in one instance did aconcentration process accompany the sul-


phide saturation, to form a significantlysulphide-enriched cumulate.

Type 1 deposits are hosted by anorthositeplutons, but usually occur in spatial asso-ciation with slightly more evolved noriticor even pyroxenitic cumulate pockets,pegmatite dykes or intrusions. The sulphi-de occurrences are not confined to margi-nal sites, but may be found scattered thro-ughout the anorthosite bodies. Re-Os andother chalcophile element modelling (Fig.1) suggest that all anorthosite-hosted de-posits, except one, are characterised byvery low R-factors (ratio between amo-unts of silicate melt in effective equili-brium with the sulphide melt). The low R-factors (~50) imply that little interactionbetween silicate and exsolved sulphidemelt took place. The scattered occurrence,low R-factors and association with lower-temperature cumulates, imply that thesulphide melt exsolved from the interstiti-al silicate melt at a late stage in the anor-thosite formation. Elevated R-factors(~500) in anorthosite-hosted occurrencesare only observed in the Fossfjellet depositlocated in the deformed margin of theEgersund-Ogna anorthosite.

Sulphide droplets are ubiquitous in allmajor mafic intrusions in Rogaland, whichmust have been near-saturated in sulphur

prior to emplacement. Fractional segrega-tion generally dispersed sulphide melt dro-plets evenly in the mafic cumulates, andonly in the Bjerkreim-Sokndal LayeredIntrusion did sulfides segregate to form asignificant stratiform type 2 deposit in anoritic intrusion. An elevated R-factor (3-500) is required to explain the compositi-on of this occurrence.

Isochronous relationships among Fe-Tirich norites, anorthosites and containedsulphides have been recorded by Re-Osand Sm-Nd isotopes, that yield isochronages of 917±22 and 914±35 Ma, respecti-vely (Fig. 2). These ages strongly corrobo-rate existing U-Pb dating in the province(Schärer et al., 1996), and the existence ofisochrons is good evidence that the sourcecharacteristics for the mafic intrusions,anorthosites and sulphides are similar.Rare marginal chills, correlated mineralchemistry, restricted initial Sr isotope rati-os and phase equilibrium considerationsimply that all mafic rocks are cumulatesformed at various stages during the diffe-rentiation of jotunitic magmas. In additi-on, experimental studies have recentlyshown that the Rogaland anorthositebodies may also be derived from jotuniticmagmas (Longhi et al., 1999). R-factorprocesses and modal variations are indeedable to account for all compositional vari-

Fig. 1. Common Os (Os concentration atthe time of formation) vs. Re/Os ratio inRogaland sulphides compared with majormagmatic sulphide deposits and mantlemelt compositions. Diagram and globaldata fields adopted from Lambert et al.(1999). All sulphide data have been nor-malized to 100 % sulphide. The inferredparental magma composition is shownwith an open circle.


ability among the sulphides, whetherhosted by anorthosite or mafic intrusions.

The late sulphide saturation in the anor-thosites, the appearance of sulphides inlower-temperature mafic intrusions, andthe compositional similarities, suggest

that sulphide saturation in the anorthosi-tes and mafic intrusions resulted from thedifferentiation of silicate magmas of com-parable compositions. No external sourcefor sulphur seems to be required.

ReferencesLambert, D.D., Foster, J.G., Frick, L.R. & Ripley,E.M. 1999: Re-Os isotope geochemistry ofmagmatic sulfide ore systems. In: Lambert, D.D.& Ruiz, J. (eds.) 'Application of RadiogenicIsotopes to Ore Deposit Research andExploration'. Reviews in Economic Geology(Society of Economic Geologists) 12, 29-57.


Schärer, U., Wilmart, E. & Duchesne, J.C. 1996:The short duration and anorogenic characterof anorthosite magmatism: U-Pb dating of the Rogaland Complex, Norway. Earth andPlanetary Science Letters 139, 335-350.

Fig. 2. Regional Re-Os and Sm-Nd isochrondiagrams. 10 Re-Os samples form a model3 isochron with an elevated MSWD of 519.Data from the Storgangen and Tellnesintrusions have been reset and form apoorly defined trend close to theCaledonian 425 Ma reference line. 30 Sm-Nd samples yield a model 3 isochron withan MSWD of 27. Initial ratios correspond toγOs = +419 and εNd = +1.47.

Fe-Ti-V-P deposits in anorthosite complexes:the bearing of parental magma composition andcrystallization conditions on the economic value

Jean-Clair Duchesne & Jacqueline Vander Auwera

L. A. Géologie, Pétrologie et Géochimie, University of Liège, Sart Tilman, Belgium.

Recent experimental data (Fram & Longhi1992, Longhi et al. 1993, Longhi et al. 1999,Vander Auwera et al. 1998) indicate thatparental magmas of the AMC suite proba-bly encompass a large continuum of com-positions ranging from high-Al basalts tomore ferroan and potassic compositions,represented by jotunites (hypersthenemonzodiorites). Experimental phase equi-libria show that both end member mag-mas can account for the norite serieswhich fractionates at 3-5 kb to silica-enri-ched liquids (Longhi et al. 1999).

In Rogaland, comparison between phasesexperimentally obtained on a primitivejotunite (Vander Auwera & Longhi 1994)

and natural phases from the anorthositemassifs and from the Bjerkreim-Sokndallayered intrusion suggests that a liquidgenerally similar to this jotunite wasparental to both types of intrusions.Interestingly, in these two types of intrusi-ons, Fe-Ti-V-P deposits have been recog-nized (Duchesne 1999). In massive anor-thosites, the ore-bodies occur as (defor-med) dykes or pods ranging in compositi-on from pure ilmenite (Jerneld) to ilmeni-te norite (Tellnes, Storgangen). Polybaricfractional crystallisation and syn-empla-cement deformation in rising anorthositediapirs lead to relatively (poisonous) Mg-and Cr-rich ilmenite (± V-magnetite)deposits. Fractional crystallization in laye-


red magma chambers of jotunite magmasgives rise to voluminous 'disseminated'mineralization, containing low Mg and Crilmenite + Ti-magnetite ± REE-rich apati-te, still of sub-economic value. Immiscibi-lity is not the controlling mechanism,except perhaps in some rare nelsonites(Hesnes).

Subsolidus re-equilibration leads to a tho-rough change in the oxide mineral com-position towards an enrichment in end-member compositions (Duchesne 1970).Reactions between the oxide mineralsleave conspicuous microscopical evidenceand regularly lower the hematite contentand possibly the Cr content of ilmenite,and the ilmenite and Al-spinel contents ofthe magnetite. Mg contents of ilmenitealso seem to decrease, possibly throughreactions with pyroxenes, but no reactionrims can be observed at the contact withpyroxene.

Experimental data suggest that there islittle pressure dependence on the oxidemineral compositions (Vander Auwera etal. 2001). Therefore, irrespective of pos-sible subsolidus re-adjustment, the maincontrolling factor on the primary ilmenitecomposition (and value) is the compositi-on of the melt in which it crystallized. Thelatter, in turn depends on the parentalmagma composition and on the degree offractional crystallization that the meltattained when the deposit formed. Furthercomplexity can be introduced by the factthat the parent magmas of jotunitic com-position display a wide range of trace ele-ment and isotope contents (Duchesne etal. 1989).

ReferencesDuchesne, J.C. 1970: Microstructures of Fe-Tioxide minerals in the South Rogaland anor-thositic complex (Norway). Annales de laSociété Géologique de Belgique 93, 527-544.


Duchesne, J.C., Wilmart, E., Demaiffe, D. &Hertogen, J. 1989: Monzonorites fromRogaland (Southwest Norway): a series ofrocks coeval but not comagmatic with massif-type anorthosites. Precambrian Research 45,111-128.

Fram, M.S. & Longhi, J. 1992: Phase equilibria ofdikes associated with Proterozoic anorthositecomplexes. American Mineralogist 77, 605-616.

Longhi, J., Fram, M.S., Vander Auwera, J. &Montieth, J.N. 1993: Pressure effects, kinetics,and rheology of anorthositic and related mag-mas. American Mineralogist 78, 1016-1030.

Longhi, J., Vander Auwera, J., Fram, M. &Duchesne, J.C. 1999: Some phase equilibriumconstraints on the origin of Proterozoic(Massif ) anorthosites and related rocks.Journal of Petrology 40, 339-362.

Vander Auwera, J. & Longhi, J. 1994:Experimental study of a jotunite (hypersthenemonzodiorite): constraints on the parentmagma composition and crystallization condi-tions (P, T, fO2) of the Bjerkreim-Sokndal laye-red intrusion. Contribution to Mineralogy andPetrology 118, 60-78.


Vander Auwera, J., Longhi, J. & Duchesne, J.C.2001: Some results on the role of P, T, and fO2on ilmenite composition. Norges geologiskeundersøkelse Report 2001.42, 142.


Crystallization processes of cumulaterocks from the Bjerkreim-Sokndal layeredintrusion (Rogaland, southwest Norway)(Wilson et al. 1996) have been investigatedthrough a whole-rock major and trace ele-ments geochemical approach. More than100 whole rocks from the layered series ofthe Bjerkreim lobe (MCU II to IV) havebeen analysed for major elements. Twotypes of cumulates are particularly wellrepresented in the collection: (1) phi-C orleuconorite cumulate: plag (An50 to An45)+ hemo-ilmenite + opx (± olivine ± mag-netite); (2) phimac-C or gabbronoritecumulate: plag (An45-An36) + opx + cpx +ilmenite + magnetite + apatite (Duchesne1978).

In Harker variation diagrams (Fig.1), thetwo types of cumulates show well-definedlinear trends for each element vs. SiO2.The regression coefficients are generallyexcellent: from 0.92 to 0.96 in the leucono-rites, from 0.89 to 0.97 (except for Ca, withr = 0.53) in the gabbronorites. A simplelinear 2-component mixing model canthus account for the major element com-positions. The first component is plagio-clase alone (An47 in leuconorite cumulate,and An40 in gabbronorite cumulate), andthe second one is made up of the sum ofthe other minerals, i.e. the mafics (hemo-ilmenite and opx for leuconorite cumulate;ilmenite, magnetite, opx, cpx and apatitefor gabbronorite). In other words, as a firstapproximation, any cumulate rock com-position oscillates between an anorthositepole and a mafic pole. Small deviationsfrom the linear area can be attributed toanalytical errors, to the presence of minoramounts of trapped liquid in the cumula-te, or to the effect of cryptic layering. Largedeviations occur when the simple model isnot respected, but such cases are very rare(5 cases out of 105).

This particular property has three impor-tant implications: (1) the relative propor-

tion of the different mafic minerals (pyro-xenes, oxide and apatite) remains constantduring each specific stage of evolution; (2)though cryptic layering is conspicuous inthe layered series (Duchesne 1978, Wilsonet al. 1996) and particularly in each type of cumulates compositional variations inindividual minerals tend to perfectlybalance each other, so that the plagioclaseand mafic poles remain globally constant;(3) the variation of the whole-rock com-positions along the linear trend onlyresults from variation in the relative abun-dance of plagioclase in the cumulate.

A purely gravity-controlled mechanism ofaccumulation of minerals at the bottom ofthe magma chamber, controlled by theStokes’ law, is likely to modify the relativeproportions of the mafic minerals becauseof the large spread in mineral densities(from 5.2 for magnetite to 3.2 for apatite).Furthermore, the density difference bet-ween plagioclase and apatite is lower thanthe density contrast between apatite andilmenite or magnetite, though the propor-tion of the latter minerals remains con-stant. A process of in-situ crystallization ofthe plagioclase has already been suggestedbecause of the small density differencebetween the plagioclase and the melt(Duchesne & Hertogen 1988, VanderAuwera et al. 1994). The present data sug-gest that all minerals have formed in situ,and that oscillatory nucleation of plagio-clase alone is responsible for the variationof the plagioclase content.

Whole-rock geochemistry of the Bjerkreim-Sokndallayered series: bearing on crystallization processes of cumulus rocks

Bernard Charlier & Jean-Clair Duchesne

L.A. Géologie, Pétrologie et Géochimie, Université de Liège, Belgium.

Fig. 1. Harker diagrams showing the lineartrend in the two types of cumulates: black

open circle: leuconorite cumulates (phi-C); redcross: gabbronorite cumulates (phimac-C);

blue filled square: mafic pole for leuconorite;red filled diamond: mafic pole for gabbronori-

tes; the dashed line connect the mafic polesto the average plagioclases: green triangle:

An40; inverted blue triangle: An47.


Forty selected rocks were also analysed fortrace elements. REE distributions clearlydiscriminate the two types of cumulates.Leuconorite cumulates are poor in REEand show a large positive Eu anomaly (Eu-Eu*= 3 to 10), whereas gabbronorites con-tain 8 to 10 times more REE and presentflat normalized patterns without signifi-cant Eu anomalies. The REE (except Eu)content of leuconorite is proportional tothe amount of P2O5, which itself reflectsthe abundance of trapped liquid. The flatnormalized patterns of gabbronoritesimplies that the positive Eu anomaly ofthe plagioclase is swamped by the negativeanomalies of the mafics, and particularlyby that of apatite. REE analyses of mine-rals show that cumulus apatite has indeeda small negative Eu anomaly, and REEcontents are 100 to 300 times more abun-dant than plagioclase. Cumulus apatitethus buffers the REE content and theEu/Eu* ratio of the cumulate, irrespectiveof the other mineral contents and, parti-cularly, of the plagioclase content.

ReferencesDuchesne, J.C. 1978: Quantitative modeling ofSr, Ca, Rb and K in the Bjerkrem-Sogndal laye-red lopolith (S.W. Norway). Contributions toMineralogy and Petrology 66, 175-184.

Duchesne, J.C. & Hertogen, J. 1988: Le magmaparental du lopolithe de Bjerkreim-Sokndal(Norvège méridionale). Comptes rendusAcadémie des Sciences de Paris 306, 45-48.

Vander Auwera, J. & Longhi, J. 1994: Experi-mental study of a jotunite (hypersthene mon-zodiorite): constraints on the parent magmacomposition and crystallisation conditions (P, T,fO2) of the Bjerkreim-Sokndal layered intrusion(Norway). Contributions to Mineralogy andPetrology 118, 60-78.

Wilson, J.R., Robins, B., Nielsen, F., Duchesne,J.C. & Vander Auwera, J. 1996: The Bjerkreim-Sokndal layered intrusion, Southwest Norway.In: Cawthorn, R.G. (ed) Layered Intrusions.Elsevier, Amsterdam, 231-256.

Relationship between the layered series and theoverlying evolved rocks in the Bjerkreim-Sokndalintrusion, southern Norway

J. Richard Wilson & Gitte Overgaard

Department of Earth Sciences, Aarhus Universitet, 8000 Århus C, Denmark.

The Bjerkreim-Sokndal layered intrusionconsists of a >7000 m-thick Layered Seriescomprising anorthosites, leuconorites,troctolites, norites, gabbronorites andjotunites (hypersthene monzodiorites),overlain by an unknown thickness of mas-sive, evolved rocks (mangerites (MG),quartz mangerites (QMG) and charnocki-tes (CH)). The Layered Series is subdivi-ded into 6 megacycic units (MCUs) thatrepresent the crystallisation products ofsuccessive major influxes of magma. Therelationship between the Layered Seriesand the overlying, evolved rocks has longbeen debated. Are the two parts of theintrusion comagmatic or do the evolvedrocks represent a separate magma influx?

We have studied a c. 1200 m-thick profilethat straddles the sequence from gabbro-

norites belonging to the top of the upper-most MCU to the quartz mangerites in thenorthern part of the intrusion (theBjerkreim lobe). Mineral compositions in37 samples become continuously moreevolved in the lower part of the sequenceup to the middle of the MG unit (plagio-clase An36-18; olivine Fo39-7; Ca-poor pyro-xene mg# 57-16; Ca-rich pyroxene mg# 65-22). Above this, compositions are essential-ly constant in the upper part of the MGunit and in the QMG (An20-13; Fo6-4;Mg#opx17-13; Mg#cpx25-18). There is noevidence of a compositional break thatcould reflect a magma influx event in thisprofile.

Whole-rock chemical compositions alsoshow a continuous evolutionary trend.The MG unit defines a cumulate trend


essentially controlled by the amount ofalkali-feldspar (which is mostly mesoper-thite). All of the QMG samples lie in asmall field with e.g. 0.8-0.3% MgO and 59-65%SiO2. Duchesne & Wilmart (1997)identified two compositionally distincttypes of QMG-CH in the upper part of theintrusion. One type (olivine-bearing QMGand CH) defines an olivine trend (OT)which they interpret essentially as reflec-ting fractional crystallisation of the mang-erite liquid resident in the chamber. Theother (two-pyroxene QMG and amphibo-le CH) defines the 'main liquid line of des-cent' derived from a jotunitic liquid.

The entry of quartz as a cumulus phase, amajor concentration of country rockinclusions (the 'septum xenolithique') and

two-pyroxene QMG all occur at the samestratigraphic level. This could reflect dis-ruption of the roof related to an influx ofjotunite-derived magma from an externalsource, as suggested by Duchesne &Wilmart (1997). An alternative is that thetwo-pyroxene QMG and amphibole CHrepresent a roof melt and that convectiveoverturn with the underlying, evolved,resident magma resulted in collapse of theroof.

ReferencesDuchesne, J.C. & Wilmart, E. 1997: Igneous char-nockites and related rocks from the Bjerkreim-Sokndal layered intrusion (Southwest Norway):a jotunite (hypersthene monzodiorite)-derivedA-type granitoid suite. Journal of Petrology 38,337-369.

SEM elemental mapping and characterisation of ilmenite, magnetite and apatite from theBjerkreim-Sokndal layered intrusion

Håvard Gautneb

Department of Mineral Resources, Geological Survey of Norway, N-7491, Trondheim, Norway.

The grade and grain-size distribution ofore minerals are key factors in evaluatingeconomic mineral deposits. This abstractreports the results of a mineral characteri-sation based on SEM elemental mappingand image processing. The work is part ofthe Geological Survey of Norway's ong-oing exploration for apatite, ilmenite andvanadium-bearing magnetite in theBjerkreim-Sokndal layered intrusion.

The samples for this study were collectedby several co-workers (Korneliussen et al.2001, Meyer & Mansfeld, this volume).The stratigraphical and petrogeneticinterpretations of some of the samples areby Meyer & Mansfeld (this volume).Selected thin-sections were measured bySEM using EDS XRF-RAY elemental map-ping. With our SEM instrumentation, weobtain picture files for up to 16 differentelements (one picture for each element).The images are recorded with a magnifica-tion of 16 times and the mapped areacovers about 1/3 of a standard thin-secti-

Fig. 1. Example ofcoloured backscatterimage, with modal content of apatite,ilmenite and magne-tite.


on. The x-ray element maps were proces-sed using the KS300 image processingsoftware (a commercial program fromCarl Zeiss vision). The X-ray images wereconverted to binary images, which againwere improved by morphological operati-ons (several steps combining erosion anddilation) to remove noise, artificial graincontacts, etc. Finally, by some steps of boo-lean image combination, new images weregenerated which represent the area distri-bution of selected minerals.

The image processing software KS300automatically measures the morphologi-cal features (such as area, longest andshortest axis, axis of enclosing ellipse ofeach mineral grain, etc.). In addition, thetotal modal % of all minerals in the rockand the modal % of ilmenite, magnetiteand apatite within selected size fractionsare calculated (Fig. 1). The grain-size dis-tribution is shown in Fig. 2. Table 1 showsthe total content of ilmenite, magnetiteand apatite and the content within selec-ted grain-size classes in the 35 studiedsamples.

Our analysed samples have the followingmodal and statistical data:

Ilmenite: The average mode is 14.17% andthe maximum value 54.71%. 2.7% of thetotal ilmenite in a sample is on average lessthan 50 µm . The D50 grain size (d-circle)is 2850 µm in diameter.

Magnetite: The average mode is 11.44%and the maximum content 51.95%. 4.50%of the total magnetite is on average less

than 50 µm. The D50 d-circle length is2400 µm.

Apatite: The average mode is 5.74% andthe maximum content 12.78%. 16% of thesamples total apatite is in grains on avera-ge less than 50 µm. The D50 d-circle lengthis 523 µm.

The sample (K323a) that contains the hig-hest content of apatite (12.78%) also con-tains 11.46 % of ilmenite and 6.56% ofmagnetite. This is a sample from themega-cyclic unit III, north of Vasshus.

ReferencesKorneliussen, A., Furuhaug, L., Gautneb, H.,McEnroe, S., Nilsson, L.P & Sørdal T. 2001:Forekomster med ilmenitt, vanadiumholdigmagnetitt og apatitt i norittiske bergarter iEgersundfeltet, Rogaland, Norges geologiskeundersøkelse Report 2001.014.

Meyer G.M & Mansfeld J. 2003: LA-HR-ICP-MSstudies of ilmenite in MCUIV of the Bjerkreim-Sokndal intrusion, SW Norway. Norges geologi-ske undersøkelse Special Publication 9, 68-69.

Table 1. Content of apatite, ilmenite and mag-netite in Bjerkeim-Sokndal rocks (unit refers to

the megacyclic units of the intrusion)

Fig. 2. Grain-size distribution (aggregatedata from all samples in Table 1) for apati-te, ilmenite and magnetite measured bySEM imaging and processing. (d-circle isthe diameter of a circle with the same area as the measured particle).

I L M E N I T E D E P O S I TS A N D T H E I R G E O LO G I C A L E N V I R O N M E N T SPECIAL PUBLICATION 9 - 6 7%

apa

tite

in s

ize

clas

s (m

icro

ns)

% I

lmen

ite

in s

ize

clas

ses

(mic

ron

s)%

mag

net

ite

in s

ize

clas

ses

(mic

ron

s)Sa

mpl

e .

Loca

lity

Coo

rdin

ates

Un

it%

%ap

%ap

%ap

%ap

%ap

%ap

%%

%%

%%

%%

%%

%%

%%

(utm

eas

t-n

orth

)to

tal

<50

<50

100-

150-

150-

>25

0to

tal

ilmilm

ilmilm

ilmilm

tota

lm

tm

tm

tm

tm

tm

tap

.15

020

020

0ilm

.<

50-

100-

150-

200-

>25

0m

t<

5050

-10

0-15

0-20

0>

250

5010

015

020

025

010

015

020

0-2

50

BJ9

901

Orr

esta

d33

7150

6493

625

MC

UIV

7.22

0.04

0.19

0.4

0.82

0.79

4.97

5.12

0.1

0.23

0.24

0.4

0.74

3.43

1.77

0.16

0.21

0.17

0.1

01.

12

BJ9

902

Orr

esta

d33

7200

6493

750

MC

UIV

8.15

0.05

0.24

0.71

1.11

0.72

5.32

4.65

0.1

0.1

0.29

0.2

0.24

3.69

4.64

0.04

0.11

0.16

0.31

0.19

3.82

BJ9

903

Orr

esta

d33

7100

6493

850

MC

UIV

10.7

20.

060.

160.

390.

691.

138.

294.

070.

140.

150.

230.

20.

562.

853.

010.

10.

280.

170.

380.

092

Bj9

905

Orr

esta

d33

7200

6494

000

MC

UIV

4.52

0.05

0.14

0.3

0.43

0.39

3.2

1.96

0.16

0.1

0.14

0.2

0.15

1.26

1.96

0.16

0.1

0.14

0.16

0.15

1.26

BJ9

907

Orr

esta

d33

7050

6493

425

MC

UIV

5.28

0.02

0.1

0.28

0.6

0.46

3.82

4.44

0.13

0.15

0.33

0.3

0.14

3.33

4.72

0.12

0.29

0.23

0.18

0.31

3.58

BJ9

909

Orr

esta

d33

7200

6493

250

MC

UIV

8.32

0.11

0.65

1.07

1.35

0.93

4.22

2.91

0.34

0.29

0.36

0.2

0.64

1.12

2.91

0.34

0.29

0.36

0.17

0.64

1.12

BJ9

910

Orr

esta

d33

7100

6493

150

MC

UIV

6.39

0.04

0.17

0.58

0.81

0.66

4.14

4.71

0.33

0.17

0.1

0.2

0.19

3.75

6.86

0.22

0.32

0.36

0.23

0.16

5.56

BJ9

912

Orr

esta

d33

8100

6492

950

MC

UIV

8.41

0.06

0.24

0.58

0.68

0.99

5.86

13.4

0.22

0.21

0.15

0.2

012

.64

6.21

0.34

0.32

0.16

0.08

0.44

4.87

K32

0aSv

ales

tad

3335

6064

9137

0M

CU

III

10.6

20.

421.

051.

561.

991.

853.

759.

050.

300.

350.

760.

781.

225.

649.

740.

140.

500.

530.

450.

717.

41

K32

0bSv

ales

tad

3335

6064

9137

0M

CU

III

11.2

80.

410.

911.

281.

330.

896.

466.

110.

150.

260.

190.

30.

564.

713

.70.

20.

370.

310.

330.

3212

.2

K32

2aV

assh

us

3461

0064

9100

0M

CU

III

11.3

50.

611.

992.

372.

582.

471.

338.

280.

480.

780.

991.

11.

23.

7710

.61

0.2

0.65

0.66

0.74

0.88

7.47

K32

3aV

assh

us

3461

1064

9105

0M

CU

III

12.7

80.

220.

911.

41.

81.

776.

711

.50.

260.

320.

460.

50.

399.

526.

560.

370.

360.

440.

350.

474.

56

K32

4aO

rres

tad

3364

9064

9281

0M

CU

IV11

.97

0.15

0.81

1.16

1.39

1.22

7.24

9.2

0.07

0.14

0.38

0.5

0.32

7.79

9.8

0.18

0.5

0.4

0.55

0.4

7.77

LPN

088

Bak

ka34

3350

6473

890

MC

UIV

10.1

50.

321.

241.

321.

241.

474.

5716

.50.

070.

090.

20.

50.

4915

.18

14.9

0.08

0.37

0.55

0.6

0.67

12.6

LPN

085

Bak

ka34

3390

6473

790

MC

UIV

11.2

40.

381.

492.

771.

671.

932.

9918

.90.

030.

160.

270.

20.

5117

.75

23.5

80.

130.

560.

560.

410.

5621

.4

LPN

090

Myd

lan

d34

9610

6475

800

MC

UIV

7.36

0.21

0.59

0.8

0.85

1.08

3.82

7.32

0.02

0.07

0.15

0.2

0.14

6.76

9.69

0.11

0.26

0.34

0.29

0.32

8.36

LPN

144

Hau

ge34

1090

964

7001

0M

CU

IV0.

270.

090.

170.

020

00

54.7

0.31

0.16

0.07

00

54.1

731

.39

0.05

0.13

0.04

0.04

0.34

30.8

LPN

148

Års

tad

0134

2710

6469

020

MC

UIV

00

00

00

027

.60.

460.

220.

160.

20.

2626

.32

64.3

0.06

0.06

0.02

0.04

0.1

64

LPN

166

Sokn

dal

3404

8064

6963

0M

CU

IV0.

060.

030.

020

00

051

.30.

010.

080.

090.

20.

2550

.62

00

00

00

0

LPN

167

Sokn

dal

3412

7064

6903

0M

CU

IV0.

040.

040

00

00

180.

080.

20.

310.

40.

7816

.25

8.39

0.07

0.07

0.22

0.19

0.43

7.42

LPN

184

Sokn

dal

3410

5064

7554

0M

CU

IV0.

040.

040

00

00

24.3

0.13

0.37

0.63

0.7

1.01

21.5

56.

050.

140.

190.

40.

451.

23.

66

LPN

187

Sokn

dal

3423

9064

7201

0M

CU

IV0.

070.

030.

010

0.04

00

20.6

0.18

0.12

0.15

0.2

0.15

19.8

551

.95

0.09

0.09

0.08

0.06

0.1

51.5

LPN

1921

9So

knda

l 34

3140

6472

760

MC

UIV

1.67

0.2

0.46

0.59

0.24

0.07

0.12

31.8

0.07

0.07

0.12

00

31.5

324

.90.

060.

170.

220.

340.

1424

LPN

1922

3So

knda

l 34

3140

6472

960

MC

UIV

2.42

0.25

0.73

0.6

0.46

0.24

0.14

250.

090.

20.

130.

10.

3424

.07

29.9

20.

390.

330.

330.

410.

6727

.8

LPN

223

Myd

lan

d34

3180

6473

960

MC

UIV

2.14

0.17

0.27

0.38

0.49

0.24

0.59

16.1

0.1

0.29

0.43

0.9

0.99

13.4

71.

810.

190.

290.

240.

250.

490.

35

LPN

227

Åm

dal

3462

0064

7705

0M

CU

IV9.

220.

160.

971.

571.

931.

553.

0311

.70.

130.

250.

080.

20.

2510

.79

4.6

0.24

0.22

0.27

0.55

0.33

3

LPN

231

Rek

efjo

rd34

0000

6468

950

Rek

fj.

5.3

0.13

0.75

1.1

0.98

0.4

1.94

3.98

0.07

0.19

0.56

0.3

0.44

2.47

2.27

0.04

0.12

0.13

0.04

0.37

1.57

LPN

236

Eia

3415

0064

8175

0R

ekfj

.2.

480.

330.

920.

670.

320.

230

2.04

0.12

0.3

0.31

0.5

0.07

0.78

1.94

0.24

0.25

0.34

0.4

0.29

0.42

LPN

238

Teks

evat

n

3398

3064

9364

0M

CII

I8.

250.

321.

161.

822.

041.

41.

518.

660.

210.

650.

831

0.74

5.28

2.52

0.23

0.45

0.42

0.13

0.39

0.91

LPN

248

Eig

elan

d 35

0700

6466

700

An

orth

.2.

110.

810.

910.

330

0.07

02.

60.

390.

540.

360.

30.

080.

938.

570.

291.

181.

511.

20.

883.

52

LPN

252

Dra

glan

d 34

0200

6471

900

MC

UIV

0.37

0.08

0.08

0.06

0.06

0.08

024

.40.

080.

240.

530.

61.

2221

.83.

610.

050.

080.

190.

310.

322.

66

LPN

255

Bils

tad

3442

5064

9120

0M

CU

IV0.

650.

140.

170.

090.

140

0.1

3.47

0.18

0.45

0.68

0.7

0.38

1.11

5.66

0.5

0.71

0.65

0.37

0.06

3.37

LPN

256

Bils

tad

3450

5064

9100

MC

UIV

10.9

10.

310.

951.

421.

961.

424.

859.

020.

210.

440.

610.

60.

546.

660

00

00

00

LPN

260

Bak

ka

3424

5064

7215

0M

CU

IV9.

190.

281.

131.

451.

671.

643.

0214

.40.

080.

330.

460.

20.

3813

.02

13.6

10.

160.

230.

220.

330.

3712

.3

Ave

rage

s5.

740.

190.

560.

770.

850.

752.

6314

.14

0.17

0.25

0.34

0.37

0.46

12.5

711

.44

0.16

0.29

0.32

0.30

0.38

9.99


The layered Bjerkreim-Sokndal Intrusionforms part of the Rogaland AnorthositeProvince of South Norway (Wilson et al.1996). The province is known for its volu-minous anorthosite bodies and ilmenite-rich noritic intrusions of which the TellnesDyke contributes to c. 6% of the presentworld Ti-production. Earlier studies ofilmenite compositions in the noritic bodi-es were primarily based on analyses byelectron microprobe or by XRF-analyseson ilmenite concentrates (e.g. Duchesne1972). These studies have given significantknowledge of the ilmenite compositions inmany of the noritic intrusions in the pro-vince. However, the need for high-precisi-on in situ analyses of ilmenite has grown

during the last few years. Based on thisneed we have developed a method for ana-lysing ilmenite by Laser Ablation HighResolution Inductive Coupled Plasma-Mass Spectrometer (LA-HR-ICP-MS).Here we present the method based on 24samples from a profile covering a 2400 m-thick cumulate sequence. The profile issituated in the northern part of theBjerkreim-Sokndal Intrusion.

The northern part of the Bjerkreim-Sokndal Intrusion consists of a c. 7 km-thick Layered Series, which is divided intomegacyclic units (MCU) arranged in asyncline. Each MCU formed in responseto magma replenishment in a bowl-shaped

LA-HR-ICP-MS studies of ilmenite in MCU IV of theBjerkreim-Sokndal intrusion, SW Norway

Gurli B. Meyer & Joakim Mansfeld

Geological Survey of Norway, NO-7491 Trondheim, Norway.

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

0 50 100 150 200

Nb (ppm)

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

0 200 400 600

Ni (ppm)

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

0 10 20 30

Hem (%)

MC

UII

IM

CU

IV

e

a

b

c

d

e

f

Zo

ne

s

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

Pla

g

Oliv

Op

x

Cp

x

IP Ilm

Mt

Ap


magma chamber. The megacyclic units arecharacterised by common crystallisationsequences of which the fourth and last(MCU IV) is the most complete. The pro-file we present here covers the uppermostpart of MCU III and a major portion ofMCU IV (Fig. 1).

The samples were analysed by UV-laserablation in medium resolution (M/∆M-=3000) on a Finnigan MAT ELEMENT®double focusing sector ICP-MS. Previousto the ablation, the ilmenite was analysedby electron microprobe. Continuous laserablation of the samples was performed asa 100 by 100 µm raster, giving the bulkcomposition of ilmenite plus hematiteexsolution lamella in each grain. The ele-ments Ti, Fe, Mg, Mn, Si, Ca, Al, V, Cr, Ni,Zr and Nb were analysed. The concentra-tion of the major elements (Ti, Fe, Mg andMn) was determined by stoichiometriccalculations in combination with elementcalibration. Trace elements were calibra-ted against international and syntheticilmenite and rutile standards that weremanufactured from TiFeO3 powders do-ped with trace elements in varying con-centrations. The standard powders weremelted by direct fusion in graphite elec-trodes resulting in fast melting and quenc-hing. Two or more analyses of 10 differentstandards were used to establish the cali-bration curves, which allowed outlierscaused by inhomogeneous standards oranalytical noise to be excluded. The analy-tical and calibration uncertainty is betwe-en 2-5% in total and the detection limitsare 0.1-10 ppm for trace metals and lessthan 0.1 % for major oxides. Parts of theresults are presented in Fig.1 together with

the cumulate stratigraphy for the Terlandprofile.

In the upper part of MCU III, ilmenitecontains about 10% hematite, increasingslightly across the MCU III/ IV boundaryto 13%. In zone a of MCU IV, the hemati-te content gradually falls to 7 %, and thenin zone b and c increases gradually to25%. Above this level the hematite contentgradually decreases to 5-10% in the cen-tral and uppermost part of the profile. Nishows an increase in concentration at thelowermost level of MCU IV (zone a), andthen a gradual decrease to zone d, whereNi reaches a minimum of a few ppm, cor-responding to an increasing degree offractional crystallisation. Nb shows classicincompatible element behaviour in termsof increasing concentration in response toincreasing degree of fractional crystallisa-tion. The trends described are in goodagreement with the trends described in ear-lier works for the silicate minerals (Jensenet al. 1993), which indicate a major reple-nishment event of magma reflected by thereversal in composition evolution in thelower part of MCU IV, followed by fracti-onal crystallisation of magma throughoutMCU IV. However, the hematite compo-nent seems to be strongly controlled bythe onset and offset of other crystallisingmineral phases such as magnetite, olivineand Ca-poor pyroxene (opx).

ReferencesDuchesne, J.C. 1972: Iron-titanium oxide mine-rals in the Bjerkreim-Sogndal massif, South-western Norway). Journal of Petrology 13, 57-81.

Jensen, J.C., Nielsen, F.M., Duchesne, J.C.,Demaiffe, D. & Wilson, J.R. 1993: Magma influxand mixing in the Bjerkreim-Sokndal layeredintrusion, South Norway: evidence from theboundary between two megacyclic units atStoreknuten. Lithos 29, 311-325.

Wilson, J. R., Robins, B., Nielsen, F.M., Duchesne,J.C. & Vander Auwera, J. 1996: The Bjerkreim-Sokndal layered intrusion, Southwest Norway.In: R. G. Cawthorn (Ed.) Layered Intrusions.Amsterdam, Elsevier, 231-256.

Fig. 1. Stratigraphic profile through MCU IV and the uppermost part of MCU III from theTerland area. The cumulate stratigraphy isshown on the left, where the profile is dividedinto zones based on terminology by Wilson etal. (1996). Triangles along the y-axis (stratigrap-hic height in metres) indicate each samplelocations. On the right, the results from theLaser-ICP-MS analyses of ilmenite are shownfor hematite (Hem (%)), for Ni (ppm) and forNb (ppm). Circles represent average composi-tions for each sample, while filled diamondsrepresent individual analyses within each offour grains for each sample.


The Tellnes ilmenite ore body, one of thelargest ilmenite deposits of the world, hasbeen mined by Titania A/S since 1960. Theore body, which is of magmatic origin,appears as a sickle-shaped lens about 2700m long and 400 m wide in the Åna-Siraanorthosite massif, belonging to the Roga-land Anorthosite Province.

Whole-rock geochemical analyses havebeen carried out on drill cores from theore body since the initial prospecting ofthe mine in 1955. At present, the geoche-mical database consists of about 5000 ana-

lyses including TiO2, P2O5, S, Cr andmodal magnetite. Furthermore, morethan 2000 samples have been analysed forthe remaining major elements and a selec-tion of trace elements (Fig. 1). In addition,mineral compositional data have beenanalysed from a selection of drill cores.

The ore is characterised by the lack ofmodal layering. However, the ore bodyshows a strong compositional zoning. TheTiO2 content is generally high (>18 wt %)within the central parts of the ore body,and there is a gradual decrease in the TiO2

content towards the contact to the host-rock. The variations in the ore chemistryreflect the variations in the modal con-tents of the ore-forming minerals. Basedon these variations, the ore body has beendivided in four zones (Fig. 1): the uppermarginal zone (UMZ), the upper centralzone (UCZ), the lower central zone(LCZ), and the lower marginal zone(LMZ). It should be noted that this divisi-on is approximate, because the transitionsbetween the various zones are smooth.The UMZ is characterised by high modalcontents of plagioclase and Fe-Mg silicates(clinopyroxene, orthopyroxene and olivi-ne), but low contents of ilmenite. TheUCZ shows high modal contents of ilme-nite and Fe-Mg silicates, and low contentsof plagioclase. The LCZ, which is the mostilmenite-enriched zone, is characterisedby low contents of Fe-Mg silicates (olivineis absent), and intermediate contents ofplagioclase. Through the LMZ, the ilmeni-te content decreases, while the contents ofplagioclase, clinopyroxene and orthopyro-xene increase.

The ore-forming minerals show composi-tional variations that are highly correlatedto the whole-rock chemistry (Fig. 2). Ingeneral, the Mg/(Mg+Fe)-ratios of ilme-nite and the Fe-Mg silicates are high in thecentral TiO2-rich parts of the ore body(UCZ and LCZ). Towards the margins,however, Mg/(Mg+Fe)-ratios of the ore-forming minerals decrease gradually.

Geochemistry and mineralogy of the Tellnes ilmenite ore body

Kåre Kullerud

Department of Geology, University of Tromsø, N-9037 Tromsø, Norway.

Fig. 1. Selected components of the ore vs. MgO/(MgO+FeO). The compositional variations of drillcore 9.5V-2N (grey shaded line) from top (circle) to bottom (arrow) have been indicated. Thewhite dashed lines in the upper left diagram separate the upper marginal zone (UMZ), the uppercentral zone (UCZ), the lower central zone (LCZ), and the lower marginal zone (LMZ). Note thegradual changes in composition along drill core 9.5V-2N from the upper marginal zone of the orebody, through the central part, and finally, through the lower marginal zone.


Emplacement mechanism of the Tellnes ilmenitedeposit (Rogaland, Norway) revealed by a combinedmagnetic and petrofabric study

Oliver Bolle 1, Hervé Diot 2, Jean-Marc Lambert 1, Patrick Launeau 3 & Jean-Clair Duchesne 1

1 Département de Géologie, Université de Liège, Bât. B20, B-4000 Sart Tilman, Belgium. 2 Université de LaRochelle, Av. M. Crépeau, F-17042 La Rochelle Cedex 01, France. 3 Université de Nantes, 2 rue de laHoussinière, BP 92208, F-44322 Nantes, France.

Fig. 2. Variations of selected chemical components of minerals and whole rocksalong drill cores 7V-3N, 12V-3S, 13V-10Nand 15V-12N. Note the gradual compositi-onal changes through the ore body, withrespect both to mineral chemistry and towhole-rock chemistry. The Fe/(Fe+Mg)ratios of the minerals are generally highalong the upper and lower margins of the ore body, but low in the central parts.Note that the variations in the compositi-on of the minerals are correlated to thevariations in whole-rock chemistry.

The Tellnes ilmenite deposit (TID)(Sokndal district, Rogaland, Norway) isthe second most important ilmenite ore-body in production in the world today (itrepresented 5% of the world's TiO2 pro-duction in 1999). This world-class Fe-Timineralization consists of an ilmenitenorite lens-shaped body (> 400 m x 2700m), that crops out in the central part ofthe Åna-Sira anorthosite, one of the maingeological units of the Late ProterozoicRogaland Anorthosite Province. Thecumulate character of the mineralizationis generally accepted (Wilmart et al. 1989,Duchesne & Schiellerup 2001) and theTID has an obvious intrusive character. Atime gap of more than 10 Myr separates itsemplacement (at 920 ± 3 Ma) from theÅna-Sira anorthosite crystallization (at932 ± 3 Ma) (U-Pb zircon ages; Schärer et

al. 1996). The TID has not suffered anyregional tectonic reworking and meta-morphism, in sharp contrast to manyNorth American occurrences of Fe-Tideposits associated with anorthosites(Duchesne 1999). It thus provides a speci-al opportunity to study genetic and struc-tural primary relationships between anor-thosites and ilmenite deposits.

To picture the poorly constrained internalstructure of the TID, anisotropy of low-field magnetic susceptibility (AMS) (for arecent review, see Borradaile & Henry1997) was measured on samples from 49sites, mainly located in the TID open-pit.In agreement with previous petrographi-cal studies (e.g. Gierth & Krause 1973), themicroscopical description of the AMSsamples reveals a rather complex magnetic


mineralogy. However, simple considerati-ons based on the variation of the bulkmagnetic susceptibility (3.1 to 84.3 mSI inthe ore) and determination of coercivityspectra suggest that magnetic mineralogyin most samples is dominated by large(usually some tens to several hundreds ofmm large) grains of magnetite. Measure-ments of anisotropy of partial anhystereticremanence (pAARM) (Jackson et al. 1988)in representative ore samples actually sug-gest that AMS is mainly carried by thesecoarse-grained magnetites: pAARM deter-mined after magnetization in a 0-20 mTwindow, and assumed to be related to thelarge magnetite grains, is strikingly coaxi-al with AMS. An image analysis investiga-

tion (intercept method of Launeau &Robin 1996) performed on about twentysamples further demonstrates that theAMS principal axes are parallel to the cor-responding shape principal axes of theopaque subfabric (ilmenite ± magnetite ±sulphides), with an angular departure ofusually less than 30°. Since the opaquesubfabric mimics the petrofabric of theore, mainly defined by a common shape-preferred orientation of silicate prismaticcrystals (plagioclase + orthopyroxene),this implies that AMS in the TID can beequated with the petrofabric.

The pattern of the magnetic foliationsshown in the TID (Fig. 1a) depicts the 3Dshape of the orebody, i.e. a SE-plungingelongated trough whose NE and SWflanks dip SW and NE, respectively. Thisstructure presents some apophyses, andends to the southeast in a dyke-like unit towhich is connected a horsetail splay-sha-ped network of noritic dykes. The patternof the magnetic lineations (Fig. 1b) is rela-tively homogenous, from the southeasternnetwork of dykes to the northwestern endof the TID, with a calculated average valueat N159° SSE 18°. This orientation is stri-kingly similar to that of the magnetic foli-ation best axis (N156° SSE 16°) and isclose to the SE-plunging axis of the orebo-dy. Parallelism of the magnetic foliationtrajectories with the TID contacts andhomogeneity of the magnetic lineationpattern, both observed over the entire ore-body, unambiguously point to a fabricthat was acquired in a magma flow, duringplutonic emplacement of the mineralizati-on. This magma flow can be traced fromthe magnetic lineation pattern (Fig. 2): ithas an average N159° SSE 18° direction,around which some deflections are obser-ved, especially towards some apophyses.The aspect of these deflections, as well aslocal obliquity between magnetic foliati-ons and the orebody wall, are argumentsin favour of a SE to NW (rather than NWto SE) magma lateral spreading along theaverage N159° SSE 18° direction. Accor-dingly, the feeding-zone of the TID waslocated to the southeast, most probablybelow the horsetail splay-shaped networkof dykes connected to the mineralizationin that area. The sickle-shaped outcrop ofthe TID further substantiates that magmainjection from the SE feeding-zone wasfavoured by the transcurrent, dextral ope-ning of a WNW-ESE striking (N120°)

Fig. 1. (a) Map of the magnetic foliations(planes perpendicular to K3 axes). (b) Mapof the magnetic lineations (K1 axes).Schmidt stereo-plots (lower hemispheres)are shown with average values of data(and best axis for Fig. 1a) in the TID and inthe Åna-Sira anorthosite (n: number ofmeasurements).


weakness zone across the Åna-Sira anor-thosite. A previous activation of this weak-ness zone, at 931 ± 5 Ma (Schärer et al.1996), would have allowed the emplace-ment of a several km-long jotunitic dyke(the Tellnes main dyke; Wilmart et al.1989) to which the TID is associated in asame geological unit.

Two petrogenetic models have previouslybeen proposed for the TID. They take intoaccount the generally admitted cumulatecharacter of the mineralization, and varia-tions in the modal proportion and chemi-cal composition of the ore-forming mine-rals observed from the contacts towardsthe central part of the orebody (e.g.,decrease of the plagioclase content andincrease of the ilmenite one). The firstmodel (Krause et al. 1985, Force 1991)calls for in-situ crystallization of a noriticmagma, inward from the orebody flanks.The second model (Wilmart et al. 1989)envisages injection of successive crystalmush batches, coming from the progressi-ve tapping of a magma chamber contai-ning a zoned cumulate pile. The model ofWilmart et al. (1989) is constrained by:(1) the coexistence of strained, prismatic(= cumulus) and undeformed, interstitial(= intercumulus) crystals of plagioclaseand orthopyroxene in thin-sections; and(2) the existence of mixing linear trendsbetween ilmenite-rich norites and moreplagioclase-rich rocks in variation dia-grams of major element contents. Hence,the magma-flow-related fabric we havedemonstrated in the TID possibly recordsthe emplacement and the deformation inthe sub-magmatic state of successive bat-ches of noritic crystal-mush. This does notpreclude, however, that in-situ crystalliza-tion has not occurred. On the contrary,the undeformed, interstitial silicates wereobviously produced through this process.Moreover, the possibility that the strained,cumulus silicates actually crystallized andwere deformed during emplacement inthe TID cannot be completely ruled out.

AcknowledgementsK. Berge and R. Hagen from Titania A/Sare thanked for their hospitality at theTellnes mine, for valuable discussion andencouragement during this work, and alsofor providing unpublished informationconcerning the TID. Fieldwork and part ofthe AMS and image analysis studies were

financially supported by Titania A/S, whoalso authorized the publication of theresults. pAARM measurements were madeby O.B. under the supervision of R. Trin-dade during a post-doctoral stay at thePetrophysics Laboratory of J.L Bouchez(Paul-Sabatier University, Toulouse, France).

ReferencesBorradaile, G.J. & Henry, B. 1997: Tectonic appli-cations of magnetic susceptibility and its ani-sotropy. Earth-Science Reviews 42, 49-93.


Duchesne, J.C. & Schiellerup, H. 2001: The iron-titanium deposits. In: Duchesne, J.C. (ed.) TheRogaland intrusive massifs: an excursion guide.Norges geologiske undersøkelse Report2001.029, 57-78.

Force, E.R. 1991: Geology of titanium mineraldeposits. Geological Society of America SpecialPaper 259, 118 pp.

Gierth, E. & Krause, H. 1973: Die Ilmenitlager-stätte Tellnes (Süd-Norwegen). Norsk GeologiskTidsskrift Supplement 53, 359-402.

Jackson, M., Gruber, W., Marvin, J. & Banerjee,S.K. 1988: Partial anhysteretic remanence andits anisotropy: applications and grainsize-dependence. Geophysical Research Letters 15,440-443.

Krause, H., Gierth, E. & Schott, W. 1985: Ti-Fedeposits in the South Rogaland igneous com-plex with special reference to the Åna-Siraanorthosite massif. Norges geologiske undersø-kelse Bulletin 402, 25-37.

Launeau, P. & Robin, P.Y.F. 1996: Fabric analysisusing the intercept method. Tectonophysics267, 91-119.


Wilmart, E., Demaiffe, D. & Duchesne, J.C. 1989:Geochemical constraints on the genesis of theTellnes ilmenite deposit, Southwest Norway.Economic Geology 84, 1047-1056.

Fig. 2. Tectonic model proposed for empla-cement of the TID.


evolved rocks occur near margins and themost primitive in the interior (Fig. 2).Analyses of selected specimens indicate arange of mineral compositions (An37-51,En65-81, Fo71-76) similar to but exceedingthat in the ~90 m-thick regressive zoneacross the MCU III-IV boundary of theBjerkreim-Sokndal Intrusion (An44-54,En66-76, Fo74).

Equivalent mineral compositions in bothintrusions suggest an origin from similarmagmas and mixing events, but withinchambers of widely different shape. InTellnes, mixing occurred in a narrow,steep-sided chamber. Earlier we suggestedthat ilmenite-rich cumulates formed atthe base due to density sorting or in-situcrystallisation along walls as new magmaflowed from a conduit at the end or side ofthe chamber and residual liquid was dis-placed upward. This called upon jotuniticmagma and a process documented in thesame district rather than an Fe-Ti-richmagma for which there is no experimentalsupport (Skjerlie et al., Lindsley, this volu-me).

We have reconsidered the analytical dataand our petrogenetic model. The initialmodel appealed to injection of primitivejotunite magma into a chamber full of dif-ferentiated magma. Mixing forced thehybrid well into the ilmenite field, resul-ting in accumulation of ilmenite on thechamber floor. This would have two con-sequences: 1) Ilmenite would have a limi-ted composition range, 2) Ilmenite crys-tals would fall into, trap and react withmore evolved magma. We explored ilme-nite compositions for evidence of a trap-ped liquid component, taking account ofpossible subsolidus effects. The mineraldata is inferred to be incompatible withsimple mixing (Fig. 3a). Ilmenite, despitescatter from local re-equilibration, shows1-20% geikielite and correlates with the

Could the Tellnes ilmenite deposit have been produced by in-situ magma mixing?

Peter Robinson 1,4, Kåre Kullerud 2, Christian Tegner 1, Brian Robins 3 &Suzanne McEnroe 1

1 Geological Survey of Norway, N7491 Trondheim, Norway; 2 Inst. for Biology and Geology, University ofTromsø, N9037 Tromsø, Norway; 3 Dept. of Geology, University of Bergen, N5007 Bergen, Norway; 4 Dept.of Geosciences, University of Massachusetts, Amherst, MA 01003, USA.

Wilson et al. (1996) argue convincinglythat ilmenite-rich cumulates in the 930Ma Bjerkreim-Sokndal Layered Intrusion,exemplified by rocks near the base ofMegacyclic Unit IV (Fig. 1), resulted frommixing of differentiated and more primi-tive jotunitic magmas. The hybrid wastemporarily within the ilmenite phasevolume, resulting in precipitation of ilme-nite alone, after which co-saturation inplagioclase and orthopyroxene was re-established. Mixing events produced lay-ers 10-20 cm thick with 50% or moremodal ilmenite. Such layers that extendedacross the ~625 km2 floor of the magmachamber could contain >31,250,000 m3 ofilmenite.

Modal and cryptic layering in the 920 Ma,dyke-like Tellnes orebody are not obvious.Nevertheless, bulk analyses of thousandsof mine samples and cores show that more

Fig. 1. EMP analyses of orthopyroxenes(open symbols) and olivines (closed sym-bols) from the MCU III-IV boundary atStoreknuten in the Bjerkreim-SokndalIntrusion. Vertical scale gives stratigraphicthickness in meters. All data from Jensenet al. (1993) and Nielsen et al. (1996).Superimposed are EMP analyses of ortho-pyroxene from core 7V-3N through theTellnes orebody from Kullerud (1994).These are plotted using identical heightand composition scales as theStoreknuten section, but positioned to emphasize the similar variations.


most Mg-rich orthopyroxene in each rock(En67-81.5). An in plagioclase is also scatte-red, but the most sodic plagioclases (An36-

43) occur with Fe-rich pyroxenes (En67-68),and the most calcic (An44-50) with themost Mg-rich pyroxenes (En77-81). Bulkanalyses of the ore average intervalsmeters thick and may not be comparableto mineral analyses from a single thin-section within the interval. However,Kullerud (1994) showed that abrupt com-positional changes are rare in the Tellnesorebody, justifying comparison of compo-sitions of ore intervals and minerals in

thin-sections. Due to ilmenite control,there are no meaningful relationships bet-ween wt.% TiO2 vs. other oxides, but wt.%TiO2 correlates with En in orthopyroxene(7% TiO2, En65-68, to 25% TiO2, En77-81).Two intervals with 28 and 33% TiO2 bothcontain En80-81. TiO2 and Mg in ilmeniteare correlated (7-14% TiO2, 1-4% geikieli-te to 25% TiO2, 10-19% geikielite). Thesamples with 28 and 33% TiO2 containthe most uniform and most Mg-rich ilme-nite (18-20% geikielite). TiO2 correlatesalso with An in plagioclase (7% TiO2,An37-41 to 25% TiO2, An44-50). The samples

Fig. 2. Vertical NE-SW section 750 acrossthe Tellnes ore body showing compositi-ons of analyzed orthopyroxenes in cores7V-3N and 8V-2S. Compositions were con-toured by hand according to authors' pre-judice.

Fig. 3. Diagrams illustrating a hypotheticalcurved ilmenite-plagioclase cotectic andsingle a) or multiple b) magma mixing.Due to the curvature (see a), when diffe-rentiated cotectic magma A is mixed withprimitive cotectic magma B, the hybrid liesentirely within the ilmenite field. Mixedmagma, will precipitate only ilmenite,moving directly away from ilmenite com-position until reaching the cotectic whereplagioclase rejoins ilmenite. A corollary ofthe curvature is that more primitive mag-mas (such as B) must precipitate a higherproportion of ilmenite than more evolvedmagmas (such as A). Instead of a singlemixing, as in a), b) illustrates multiplemixing in a dike-like chamber of successi-ve pairs of magma batches that had beenprogressively evacuated from a chemically,thermally and density stratified subjacentmagma chamber. While the net shift intothe ilmenite primary crystallisation field isless than in a), there is a still precipitationof excess ilmenite. In addition, uponreturns to cotectic precipitation, silicatecompositions will reflect melt compositi-ons close to those that precipitated theilmenite.


with 28 and 33% TiO2 contain An46 andAn47.5, possibly due to a trapped liquidcomponent. In intervals with analyzedminerals, normative An in the bulk orecorrelates nearly linearly with En (An41,En67, to An54, En81) and wt.% TiO2. Aninterval with normative An59 is characteri-sed by En81.

These chemical and mineralogical variati-ons do not support ilmenite crystallisationfrom a single hybrid magma and evidencefor trapped evolved liquid is lacking.There is an alternative way in whichmixing may have occurred, but it requiresspecial circumstances and appropriateplumbing. We propose a stably-stratifiedmagma chamber of wide aspect ratio, withdifferentiated, light magma at the top andmore primitive, denser magma at the base(Campbell 1996). The Bjerkreim-Sokndalmagma chamber was probably stratifiedin this manner at times (Wilson et al.1996). The top of the chamber was tappedand the stratified magma injected into theTellnes dyke. Adjacent magma layersmixed as they were extracted (Fig. 3b) andthe resulting hybrids were saturated inilmenite alone, although only briefly.Otherwise the magmas were multiplysaturated. As the hybrid magmas flowedalong the Tellnes chamber, probably fromthe southeast (Bolle et al., this volume),ilmenite accumulated in higher thancotectic proportions. The density of theresulting cumulates expelled trapped melteffectively. As gradually more primitivemagmas from the subjacent chamber weretapped and mixed, the proportion ofilmenite increased. Ilmenite and otherphases became more primitive and thecentral ilmenite-rich zone precipitatedfrom the last and most primitive hybrids.Meanwhile, residual magmas were displa-ced laterally along the Tellnes dyke, tolocations not preserved. Relatively evolvedrocks at the top of the Tellnes orebodymay have crystallised in situ during injec-tion, though their moderately high ilme-nite content is conceptually difficult toexplain.

More mineralogical, textural and modaldata are required to test the mixing hypot-hesis and could provide incisive predictivetools for future mining and ore beneficia-tion at Tellnes.

ReferencesBolle, O., Diot, H., Lambert, J-M., Launeau, P. &Duchesne, J-C. 2001: Emplacement mechanismof the Tellnes ilmenite deposit (Rogaland,Norway) revealed by a combined magneticand petrofabric study (abstract). Norges geolo-giske undersøkelse Report 2001.042, 19-20.

Campbell, I.H. 1996: Fluid dynamic processes inbasaltic magma chambers. In Cawthorn, R. G.(ed.), Layered Intrusions, Elsevier Science, 45-76.

Jensen, J.C., Nielsen, F.M., Duchesne, J.C.,Demaiffe, D. & Wilson, J.R. 1993: Magma influxand mixing in the Bjerkreim-Sokndal layeredintrusion, South Norway: evidence from theboundary between two megacyclic units atStoreknuten. Lithos 29, 311-325

Kullerud, K. 1994: MgO-innhold i ilmenitt fraTellnesforekomsten. Avsluttende Rapport,Institut for Biologi og Geologi, Universitet iTromso, 14 pp., 18 figures, 41 pp. of tables.

Lindsley, D.H. 2001: Do Fe-Ti oxide magmasexist? Geology:Yes; Experiments: No! (abstract).Norges geologiske undersøkelse Report 2001.042,83.

Nielsen, F.M., Campbell, I.H., McCulloh, M. &Wilson, J.R. 1996: A strontium isotopic investi-gation of the Bjerkreim-Sokndal layered intru-sion, southwest Norway. Journal of Petrology 37,171-191.

Skjerlie, K.P., Kullerud, K. & Robins, B. 2001:Preliminary melting experiments on theTellnes ilmenite norite from 0.5 to 1.2 GPa,implications for the composition of intercu-mulus melt (abstract). Norges geologiske under-søkelse Report 2001.042, 134-135.

Wilson, J.R., Robins, B., Nielsen, F.M., Duchesne,J.C. & Vander Auwera, J. 1996: The Bjerkreim-Sokndal Layered Intrusion, Southwest Norway.In Cawthorn, R. G. (ed.), Layered Intrusions.Elsevier Science, 231-255.


There is general agreement that the Tellnesilmenite norite is a magmatic intrusionthat was emplaced at c. 0.5 GPa as a cumu-late with minor amounts of intercumulusmelt. The composition of the interstitialmelt is unknown, but it has been assumedthat it was Ti-rich. If the cumulate hypot-hesis is correct, partial-melting experi-ments should yield important informati-on on the composition of the melt andequilibrium solid phases. In accordancewith this line of reasoning we have perfor-med melting experiments in a piston-cylinder apparatus on a sample of theTellnes ilmenite norite from 0.5 to 1.2 Gpaand at various temperatures.

The experiments show that the intercumu-lus melt at 0.5 GPa and T=1050°C (solidus<950°C) is jotunite-like with SiO2 58.9wt.%, TiO2 2.2, Al2O3 17.9, FeO 7.2, MnO0.1, MgO 3.2, CaO 5.3, P2O5 1.3, Na2O 2.3,K2O 1.6, and in equilibrium with abun-dant euhedral plagioclase (An50) and ilme-nite (He15) together with some olivine(Fo72.5). The interstitial glasses in near-solidus experiments are not particularlyrich in TiO2. This is not surprising and is

observed in other melting experimentswith residual ilmenite or rutile. Theamount of a component like TiO2 will bebuffered by ilmenite or rutile, but theamount of dissolved TiO2 in a liquid isindependent of the amount of the satura-ted phase. The melt fraction increases withincreasing pressure and temperature dueto progressive dissolution of plagioclaseand ilmenite, and the melt compositionsbecome FeTi-dioritic/FeTi-basaltic. High-temperature melting yields homogeneousglasses with very high abundances of dis-solved ilmenite, showing that Ti-rich liq-uids can exist without the intervention ofliquid immiscibility.

Our experiments support the hypothesisthat the Tellnes intrusion is an ilmenite-rich, plagioclase-ilmenite cumulate, butthe intercumulus melt was jotunite-likeand not particularly rich in TiO2. Increa-sing dissolution of Fe and Ti at higher Pand T suggests that at least some cumula-tion of ilmenite may have occurred duringdecompression. Our experiments do notsupport an origin of the Tellnes ore byliquid immiscibility.

Preliminary melting experiments on the Tellnes ilmenite norite from 0.5 to 1.2 Gpa: implications forthe composition of intercumulus melt

Kjell P. Skjerlie †, Kåre Kullerud 1 & Brian Robins 2

1 Department of Geology, University of Tromsø, 9037 Tromsø. 2 Department of Earth Science, University ofBergen, Allegaten 41, 5017 Bergen.

Concentration of ilmenite in the late-Sveconorwegiannorite-anorthosite Hakefjorden Complex, SW Sweden

Hans Årebäck

Department of Geology, Earth Science Centre, Göteborg University, Box 460, SE-405 30 Göteborg, Sweden.Present address: Boliden Mineral AB, SE-936 81 Boliden, Sweden.

The c. 916 Ma Hakefjorden Complex(HFC) comprises a composite norite-monzonorite-anorthosite intrusion (Åre-bäck 1995, 2001, Scherstén et al. 2000). Itis situated c. 25 km NNW of Göteborg,SW Sweden, in the Sveconorwegian Prov-

ince and forms two E-W elongated bodies,both about 500 m wide and 3.6 and 1.2km long, respectively, intruding Mesopro-terozoic supracrustal gneisses (Fig. 1). Theintrusion is dominated by a medium-grai-ned norite-monzonorite (An54) contai-

† Deceased 7 september 2003.


ning massive anorthosite blocks, 2 cm to50 m in diameter, and 2-20 cm-largeandesine megacrysts (disaggregated anor-thosite blocks). The coarse-grained anor-thosite blocks (An45) occur in all maficrock types with a preferred occurrencealong the contacts. Subordinate ilmenite-rich leuconorite, hybrid rock and graniticdifferentiates are recorded. Gradual tran-sitions (mineralogy and geochemistry)between the norite and the marginal mon-zonorite and the granitic differentiates aswell as between the norite and the ilmeni-te-rich leuconorite, indicate their closerelationship and that they are productsfrom the same magma. The emplacementmechanism has been inferred to involvedeep-crustal (high-pressure) anorthositeand ultramafic crystallisation followed bymid-crustal norite-monzonorite crystalli-sation involving crustal contaminationand Fe-Ti enrichment during fractionalcrystallisation (Årebäck & Stigh 1997,2000).

The liquid composition from which theHFC-norite/monzonorite crystallised isnot straightforward to determine, sincemost rocks are affected by accumulationof cumulus plagioclase ± orthopyroxene ±Fe-Ti oxides ± apatite, crustal contamina-tion or both. However, a slightly morefine-grained margin (chilled margin)where the marginal monzonorite is notdeveloped is suggested to represent the

best estimate of the liquid composition ofthe exposed HFC-norite/monzonorite(Årebäck 2001). In a QAP-diagram thisrock plots in the lower left part of thequartzjotunite (quartzmonzonorite) field.Rocks that are chemically more primitivethan this composition are, indicated fromthin-section observation, affected by accu-mulation of ± plagioclase ± orthopyro-xene ± Fe-Ti oxides ± apatite, while rocksmore evolved are affected by fractionation± crustal contamination.

In the central part of the intrusion there isa limited area consisting of an extremelyilmenite-rich leuconorite (IRL; Årebäck &Stigh 2000). The IRL occurs as a ‘disc-like’,c. 10-15 m-thick, 100×300 m-wide subho-rizontal unit within the norite. The lateraland lower contacts with the surroundingnorite are gradual. The IRL matrix is fine-to medium-grained and characterised by asubhorizontal magmatic foliation definedby the preferred orientation of tabularplagioclase crystals (An55). This foliationis restricted to the IRL unit. Various typesof enclaves occur in the IRL matrix. Theycan be divided into two main groups. Thefirst consists of decimetre- to meter-sizedanorthosite and norite blocks from thehost HFC. The second group consists ofmassive to semi-massive, millimetre- todecimetre-sized Fe-Ti oxides, fine-grainedanorthosite and fine-grained, fine-scale,igneous-layered plagioclase-orthopyro-

Fig. 1. Generalised bedrock map ofHakefjorden Complex region, SW Sweden.The inset map shows the major lithotecto-nic units of southwestern Fennoscandia.CAL = Caledonian; SVF = Svecofennian;TIB = Transscandinavian Igneous Belt;WGR = Western Gneiss Region; dottedareas = rocks affected by the Sveco-norwegian Orogeny;ruled areas = Phanerozoic rocks.


xene. The second group is restricted to theIRL and is considered to represent disrup-ted layering within the IRL. The foliationof the IRL matrix is wrapped around allinclusions and interpreted to result fromdeformation in the magmatic state.

The IRL matrix contains subequal amo-unts of ilmenite (32-66 vol. %) and silica-tes (34-68 vol. %). For obvious reasons,the IRL matrix is anomalously high inmajor and trace elements entering Fe-Tioxides, i.e. Ti, Fe, Cr, Sc, Ni and V. Ilmenitegrains are typically rounded, equidimensi-onal, and either homogeneous or containthin hematite lenses arranged in the{0001} planes (hemo-ilmenite). Ilmenitein the massive Fe-Ti oxide enclaves is highin MgO, ranging from 2.17 up to 5.00%(average 3.21%), compared to the IRL-matrix ilmenite (1.35 to 2.79%; average1.97%), which in turn is more Mg-richthan ilmenite in the surrounding norite(0.03-0.3%; average 0.13%; Fig. 2). Thereason for this is interpreted as due to sub-solidus re-equilibration; Mg will decreasein ilmenite during cooling by re-equilibra-tion with the surrounding silicates. Oxide-rich units may retain relatively high con-tents of Mg, since they are more or lessisolated from silicates. This explains whythe ilmenite in the Fe-Ti oxide enclavesshows higher Mg contents than in the sur-rounding IRL matrix which, in turn, ismore Mg-rich than ilmenite in the norite(Fig. 2). Magnetite rarely exceeds 1 vol. %

in the matrix whereas the Fe-Ti oxide enc-laves contain 20-60 vol%. Ilmenite inter-growths occur as microlamellae along the{111} planes (trellis type) of magnetite oras thicker lamellae restricted to one set ofthe {111} planes (sandwich type). Thetrellis type is most frequent and occurs inat least two generations; a rather coarse setand a network of very fine-grained lamel-lae represented by submicroscopic exsolu-tions occurring as domains in the magne-tite. Exsolution of pleonaste is frequent inthe {100} planes of magnetite. The lamel-lae occur in several generations and tendto be most frequent in the centre of thehost grain. Textural evidence indicatesthat a certain amount of pleonaste wasexsolved simultaneously with ilmenite, asindicated by the presence of pleonastemicrocrystals within exsolved ilmenite. Allmagnetite has a very low Xusp component.However, this does not necessarily meanthat the magnetite phase was low in TiO2

during crystallisation, as their compositi-ons are affected by subsolidus re-equili-bration.

Crystal accumulation in a late-stage nori-tic magma residual after anorthosite crys-tallisation is put forward to be the mainmechanism forming the IRL, where theliquid from which it was produced musthave been strongly enriched in Fe and Ti.The origin of such liquids is characteristicfor residual liquids after anorthosite crys-tallisation (e.g. McLelland et al. 1994,

Fig. 2. MgO-TiO2 variation diagram forilmenites in different units of the HFC.The variation is explained by subsolidusre-equilibration.

0

10

20

30

40

50

60

0 2 4 6MgO wt%

TiO

2 w

t%

Fe-Ti oxide enclave

IRL-matrix

Trellis-type lamellae in magnetite of the Fe-Ti oxide enclave

HFC-norite


Emslie et al. 1994). Deformation in themagmatic state arranged the solid phasesaccording to the stress regime.

Textures and compositions of ilmeniteand magnetite in the Fe-Ti oxide enclavesindicate post-crystallisation modification.These enclaves were probably formed, atleast in part, by annealing, i.e. consolidati-on of loose grains into polycrystallinematerials at subsolidus temperatures.

Despite the differences in size of theHakefjorden IRL and the Tellnes ilmenitedeposit, similarities in age, petrographyand geochemistry occur. The Tellnes ilme-nite-rich norite (920 ± 2 Ma; Schärer et al.1996), which dates the end of the magma-tic activity in the Rogaland AnorthositeComplex, is interpreted as a noritic cumu-late emplaced as a crystal mush with Fe-Ti-rich intercumulus liquid (e.g. Duch-esne 1999).

ReferencesÅrebäck, H. 1995: The Hakefjorden Complex -geology and petrogenesis of a late Sveco-norwegian norite-anorthosite intrusion, south-west Sweden. Fil. Lic. Thesis. Geovetarcentrum,Göteborgs universitet, A9, Göteborg, 84 pp.

Årebäck, H. 2001: Petrography, geochemistry and geochronology of mafic to intermediate lateSveconorwegian intrusions: the HakefjordenComplex and Vinga intrusion, SW Sweden. PhD.Thesis. Geovetarcentrum, Göteborgs univer-sitet, A72 Göteborg.

Årebäck, H. & Stigh, J. 1997: Polybaric evolutionof the Hakefjorden Complex, southwestern

Sweden, deduced from partial dissolution inandesine megacrysts. Geologiska Föreningens i Stockholm Förhandlingar 119, 97-101.

Årebäck, H. & Stigh, J. 2000: Nature and originof an anorthosite associated ilmenite-rich leu-conorite, Hakefjorden Complex, SW Sweden.Lithos 51, 247-267.

Duchesne, J.C. 1999: Fe-Ti deposits in theRogaland anorthosites (South Norway):geochemical characteristics and problems ofinterpretation. Mineralium Deposita 34, 182-198.

Emslie, R.F., Hamilton, M.A. & Thériault, R.J. 1994:Petrogenesis of a Mid-proterozoic anorthosite-mangerite-charnockite-granite (AMCG) com-plex: isotopic and chemical evidence from theNain plutonic suite. Journal of Geology 102,539-558.

McLelland, J., Ashwal, L.D. & Moore, L. 1994:Composition and petrogenesis of oxide-,apatite-rich gabbronorites associated withProterozoic anorthosite massifs: example ofthe Adirondack mountains, New York.Contributions to Mineralogy and Petrology 116,225-238.


Scherstén, A., Årebäck, H., Cornell, D.H., Hoskin,P., Åberg, A. & Armstrong, R. 2000: Dating mafic-ultramafic intrusions by ion-probing contact-melt zircon: examples from SW Sweden.Contributions to Mineralogy and Petrology 139,115-125.

Ilmenite-rich gabbros in the Kälviä (Kärk-käinen et al. 1997), Kauhajoki (Kärkkäinen& Appelqvist 1999), and Kolari areas(Karvinen et al. 1988), and the OtanmäkiFe-Ti-V mine (closed 1985) (Pääkkönen1956) indicate a high-Ti potential of smallmafic intrusions in Finland (Fig. 1).

Kälviä area,the Koivusaarenneva gabbroIn the Kälviä area, several ilmenite-richgabbros have been located in a chain ofgravity and magnetic anomalies (Kärk-käinen et al., 1997, Sarapää et al. 2001).

Gabbro-hosted ilmenite deposits in Finland

Niilo Kärkkäinen

Geological Survey of Finland, 02150 Espoo, Finland.


Fig. 1. The location of some Finnish gabbro-hosted ilmenite-rich deposits

The mafic intrusions were emplaced at1881 Ma into tonalitic bedrock, and be-long to a large gabbro province interpre-ted to have been formed in tensional zonesin the vicinity or along the margin of con-vergent plate boundaries.

The Koivusaarenneva gabbro is an elonga-ted, 0.5 to 1 km thick and 3 km long sill-like intrusion. It can be divided into threedifferent zones, Lower, Middle and Upper,and the main rock type in all zones is ametamorphosed gabbro or gabbronorite(Kärkkäinen 1997). The characteristicminerals in theses zones are ilmenomag-netite, ilmenite and apatite, respectively.The Lower Zone contains minor ilmenite-rich layers with much ilmenomagnetite,and the TiO2/Fe2O3TOT ratio of 0.2 isconstant for all rock types in the LowerZone. The Middle Zone contains a 1.5km-long and from 50 to 60 m-thick mine-ralized layer that grades between 8 and48% ilmenite, 2 to 25 % magnetite andminor ilmenomagnetite. The ratio ofilmenite to magnetite is usually from 3:1to 4:1. The TiO2/Fe2O3TOT ratio of theMiddle Zone is between 0.23 and 0.50.Ilmenite is the dominant Fe-Ti oxide inthe Upper Zone that consists of P-richgabbro and leucogabbro.

The three zones are interpreted to havecrystallized from successive pulses of Ti-rich mafic magma. The parent magmasfor the Lower and Middle Zones are simi-lar, and the magma for the Upper Zonewas more evolved. The Lower and UpperZones are interpreted to have been gene-rated by relatively closed-system fractionalcrystallization. The genesis of the Fe-Tioxide-rich layers in the Middle Zone canbe explained by deposition from severalmagma pulses in an open system, becauseof the small volume of the rock in theMiddle Zone. The Middle Zone mayrepresent a channel for magma flow andthe large mass of oxides was accumulatedfrom multiple magma pulses. Suspendedoxides were removed from a Ti-saturatedmagma and sank to the floor of a shallowmagma chamber to form the mineraldeposit, composed of a Fe-Ti oxide-richmatrix between a silicate framework. TheTi- and Fe- depleted magma flowed out ofthe poorly crystallized intrusion and wasreplaced by a new pulse of Ti-saturatedparent magma.

Kauhajoki area,the Kauhajärvi gabbroThe Kauhajärvi gabbro is one of the seve-ral Fe-, Ti-, and P-rich mafic intrusionswithin a granitoid area in mid-southwestFinland (Kärkkäinen & Appelqvist 1999).It is composed of a relatively thin basalzone and a main zone. Ti- and P-rich gab-bro is the most voluminous rock type inthe 800 m-thick main zone that is diffe-rentiated with a composition varying fromperidotite to anorthosite. Layers with anaverage of 20 wt % combined ilmenite,apatite and ilmenomagnetite form a low-grade apatite-ilmenite-magnetite deposit.Apatite and Fe-Ti oxides are concentratedtogether with Fe-Mg silicates.

Early crystallization of magnetite supportsthe interpretation that the main zone crys-tallized under conditions of relatively highoxygen fugacity. There are three main fac-tors controlling the concentration of theFe-Ti oxides and apatite in the Kauhajärvigabbro: 1) an Fe-, Ti- and P-rich parentalmagma; 2) a relatively high fO2 duringcrystallization that kept the solubility ofFe and Ti low, so that these elements couldnot be enriched during crystallization;and 3) high P2O5 in the magma enabledthe crystallization of individual ilmenitecoevally with Ti-bearing magnetite (ilm-enomagnetite). The Kauhajärvi intrusionis interpreted to have crystallized underrelatively closed-system conditions andthe lack of dynamic magma flow throughthe intrusion prevented accumulation of alarge mass of Fe-Ti oxides.


Koivusaarenneva, Kauhajärvi Kauhajoki Otanmäki Karhujupukka

Kälviä Vuolijoki Kolari

Intrusion multistage gabbro 2-stage well- differentiated gabbro differentiated

with 3 zones differentiated gabbro -anorthosite sill-like gabbro

Dimensions 0.8 x 3 km2 >1 x 6 km2 15 x 1 km2 0.5 x 1 km2

Age Ga 1.881 Ga 1.875 Ga 2.06 Ga 2.1 Ga?

Country rock tonalite granodiorite granite-gneiss metasediments

Geotectonic setting arc near plate margin intracratonic rift shield (continental margin) cratonic basin

Deposit:

% TiO21 7.5 5 14 9

wt-% ilmenite 15 (8 to 48) 10 28 18

wt-% magnetite 6 (-30) 2 - 8 35 66

I / MT 3:1 to 4:1 1.5 : 1 1:2 1: 2 (varies)

% V in MT 0.7 0.3 0.6 0.57

Special apatite Co in pyrite

1) averaged TiO2 content in mineral resource estimates

Table 1. Characteristics of some Finnish gabbro-hosted ilmenite-rich deposits. (MT = magnetite,I=ilmenite).

SummaryGabbroic intrusions with ilmenite-domi-nant, Fe-Ti oxide-rich rocks occur withinvariable geotectonic settings of Palaeopro-terozoic ages between 2.1 Ga and 1.88 Ga(Table 1). The common features are thesmall size of the intrusions and the occur-rence of ilmenite and magnetite, mainly asindividual grains. Fractionation under lowoxygen fugacity conditions favours gene-ration of a Ti-rich parental magma.Shallow magmatic processes concentratedilmenite from the Ti-enriched parentmagma and furthermore, a low fO2 envi-ronment favoured ilmenite crystallizationrelative to Ti-magnetite. A magma-flow-through system in the shallow upper crus-tal environment is the critical process thatenables the concentration of large volu-mes of oxides within small intrusions, ashas been suggested for the Kälviä area(Kärkkäinen & Bornhorst 2000). Themagma-flow genesis is supported by thefact that the Koivusaarenneva gabbro is amember of a chain of small, probablyinterconnected, intrusions, three of which,Lylyneva, Peräneva, and Kairineva also,host ilmenite deposits (Sarapää et al. 2001).

ReferencesKärkkäinen, N. 1997: The Koivusaarenneva gab-bro, Finland. In: Papunen H. (ed.) Mineral depo-sits: research and exploration, where do theymeet? Proceedings of the Fourth Biennial SGAMeeting, Turku/Finland/11-13 August 1997.

Rotterdam, A. A. Balkema: 443-444.

Kärkkäinen, N. & Appelqvist, H. 1999: Genesis ofa low-grade apatite-ilmenite-magnetite depo-sit in the Kauhajärvi gabbro, western Finland.Mineralium Deposita 34, 754-769.

Kärkkäinen, N. & Bornhorst, T. J. 2000: Magmaflow genetic model for titanium ore in a smallmafic intrusion. In: Geological Society of America2000 Annual Meeting and Exposition, Reno,Nevada, November 13-16, 2000. Abstracts withPrograms 32 (7), A-426.

Kärkkäinen, N., Sarapää, O., Huuskonen, M.,Koistinen, E. & Lehtimäki, J. 1997: Ilmeniteexploration in western Finland, and the mine-ral resources of the Kälviä deposit. GeologicalSurvey of Finland Special Paper 23, 15 -24.

Karvinen, A., Kojonen, K.& Johanson, B. 1988:Geology and mineralogy of the KarhujupukkaTi-V-Fe deposit in Kolari, Northern Finland.Geological Survey of Finland Special Paper 10,95-99.

Pääkkönen, V. 1956: Otanmäki, the ilmenite -magnetite ore field in Finland: GeologicalSurvey of Finland Bulletin 171, 71 pp.

Sarapää O., Reinikainen J., Seppänen H.,Kärkkäinen N. & Ahtola T. 2001: Industrial minerals exploration in southwest and westernFinland. In: Autio, S. (ed.) Geological Survey ofFinland, Current Research 1999-2000. GeologicalSurvey of Finland Special Paper 31, 31-40.


Anorthositic suites are a characteristic fea-ture of the Grenville Province. Numerousdeposits of titaniferous magnetite, magne-tite-ilmenite-apatite and massive ilmenitehave been known since the mid-1850s.However, only a few of them have beeneconomically mined since 1870 and onlythe world-class, massive ilmenite depositof the Tio Mine has been mined since1950. We present a brief review of themost typical Fe-Ti and Fe-Ti-P2O5 depo-sits associated with anorthositic suites inthe Grenville Province in Québec.

In general, the shape of these deposits

varies from tabular intrusions to stocks,sills or dykes. Locally, some deposits arecharacterised by stratiform mineralizationin layered anorthositic rocks. Four maindistinct styles of mineralization have beendescribed and observed in the Quebec’sGrenville Province:1) Syngenetic formation for disseminated

Fe-Ti oxides in the host anorthosite,such as those observed at the west mar-gin of the Havre-Saint-Pierre anortho-sitic suite.

2) Irregular to concordant injections orintrusions, with clear-cut or diffusecontacts with the host rocks, in an older

The Grenville Province in CanadaReview of Fe-Ti ± V and Fe-Ti-P2O5 ± V deposits associated with anorthositic suites in the GrenvilleProvince, Québec

Serge Perreault 1 & Claude Hébert 2

1 Géologie Québec, 545, boul. Crémazie Est, bureau 1110, Montréal, Québec Canada H2M 2V1, 2 GéologieQuébec, 5700, 4ième avenue Ouest, Charlesbourg, Québec, Canada G1H 6R1.

Fig. 1. Location of the studied area in theGrenville Province. HSPA, Havre-Saint-Pierreanorthosite; LFA, lac Fournier anorthosite;LSJA, Lac Saint-Jean Anorthosite; RPA,Rivière Pentecôte Anorthosite; TA,Tortue Anorthosite; PMA, Petit-MécatinaAnorthosite; RAA, Atikonak River Anort-hosite; BEA, Berté Anorthosite; LA, LabrieAnorthosite; LVA, Lac Vaillant Anorthosite;PA, Pambrun Anorthosite; MA, MorinAnorthosite; SUA, Saint-UrbainAnorthosite; SILC, Sept-Iles LayeredComplex; FG, Grenville Front.


anorthosite or associated rocks, of mas-sive ilmenite, such as the Tio Mine (Fig.1 and the abstract of Perreault, thisvolume), the Ivry (typical chemical ana-lysis for the massive ore is 42.5 % Fe,33.23 % TiO2, 7.54 % SiO2), the Bignell(mineral reserves at 3.44 Mt 34.1 % Feand 34.7 % TiO2 with 2.5 Mt at 35.7 %Fe and 36.4 % TiO2 proven reserves) andthe Lac Brulé deposits (Fig. 1); massivetitaniferous magnetite, such as Puyjalon(Everett), Desgrobois and Magpie de-posits (estimation of mineral reserve at1.8 billions tons with 184 Mt of provenand 627 Mt of probable reserves at 43.1% Fe, 10.6 % TiO2, 1.55 % Cr and 0.19% V), oxide-rich norite, ferrodioriteand nelsonite, such as the Puyjalon,Desgrobois, and most deposits in the

Lac-Saint-Jean anorthositic suite.3) Rutile-bearing ilmenitite or oxide-rich

norite. Irregular to concordant injecti-ons or intrusions, such as the Coul-ombe East and West (the mean grade is36 % Fe, 39 % TiO2, 4 % SiO2, 4 %Al2O3 and 4 % CaO) and the Big Islanddeposit (typical grab samples collectedfor assays give values of 26 % FeO, 21.1% Fe2O3, 40.3 % TiO2, 3.12 % MgO,0.36 % V2O5, 4.35 Al2O3, and 6.26 %SiO2).

4) Remobilized Fe-Ti massive or semi-massive mineralizations in shear zonesand in late felsic injections cutting theanorthositic rocks, such as those obser-ved at the margin of the Havre-Saint-Pierre anorthositic suite.

A study of mineral compositions of the Lac Mirepoixlayered complex, Lac St-Jean Anorthosite Complex(Québec, Canada)

Caroline-Emmanuelle Morisset

Département des Sciences de la Terre, Université du Québec à Montréal, C.P. 8888, Succursale Centre-Ville,Montréal, PQ H3C 3P8 Canada, and L.A. Géologie, Pétrologie, Géochimie, Université de Liège, Bat. B-20,4000 Liège, Belgique

The Lac Mirepoix layered complex is loca-ted within the Lac St-Jean AnorthositeComplex of the Grenville Province, 250km north of Québec City (Canada). Themineral sequence is very similar to that ofthe Bjerkreim-Sokndal layered intrusionin Norway. In the sequence of layeredrocks, cumulus assemblages contain firstplagioclase and hemo-ilmenite, followedby the addition of orthopyroxene, clino-pyroxene and magnetite and apatite toget-her as the last phases (Fig. 1). Nelsoniteand Fe-Ti oxide mineral pure horizons (10cm thick) are present in the layered series.Cryptic layering is observed in the compo-sition of plagioclase, orthopyroxene andclinopyroxene. Reversals in the mineralo-gical sequence and in the evolution of theAn content of plagioclases and Mg# oforthopyroxenes and clinopyroxenes areinterpreted as due to a new magma influxbetween the Cyclic Units IV and V (Fig. 1).Mixing of the new magma influx with the

resident magma explains the intermediatemineral compositions at the base of thenew influx (Cyclic Unit V, Fig. 1).

Magnetite and hemo-ilmenite of the LacMirepoix layered complex have beenaffected by strong subsolidus re-equilibra-tion. Coronas of spinel pleonaste are com-monly observed at the contact of the twooxide minerals. Analyses of whole grainoxide minerals show that Al2O3 and MgOof the ilmenite vary in the same way.These analyses also show that the propor-tion of hematite in the ilmenite is usuallyless important when magnetite is present(Fig. 1). This is explained by a two-stepprocess. The first step is the exsolution ofa spinel pleonaste phase in the magnetiteat the contact of the ilmenite. The exsolu-tions are incorporated, in a second step, tothe ilmenite grain by the formation of newilmenite at the contact of the oxide mine-rals grains. This new ilmenite is formed by


oxidation of the ulvöspinel in the magne-tite. The oxidation is caused by reactionwith the hematite contained in the ilmeni-te (Buddington & Lindsley, 1964, Duch-esne, 1972). Because of this process, Al2O3

and MgO from the magnetite are incorpo-rated in the ilmenite. It also lowers theTiO2 in the magnetite and the hematitecontent of the ilmenite.

In the hemo-ilmenite of the Lac Mirepoixlayered complex, an ulvöspinel-magneti-tess phase is observed as 'exsolution' lamel-lae. These 'exsolutions' are probably pro-duced by reduction of the hematite exso-

Fig. 1. Generalized stratigraphy of the Lac Mirepoix layered complex, cryptic layering and wholegrain oxide compositions. (Pl) plagioclase; (Opx) orthopyroxene; (Cpx) clinopyroxene; (H-I) hemo-ilmenite; (Il) ilmenite; (Mt) magnetite; (Ap) apatite; (% An plag) 100*An/(An+Ab); (Mg# Opx) Mg# oforthopyroxene; (Mg# Cpx) Mg# clinopyroxene; (# Hem) 0,5*(Fe2+ + Fe3+ - Ti +Mg +Mn); (Mt/Mt+Il)ratio magnetite on magnetite plus ilmenite; (Al2O3, MgO) whole grain oxide minerals by XRF.

lution lenses. This reduction in thehemo-ilmenite can be coupled with oxi-dation in the sulfide minerals.

ReferencesBuddington, A.F & Lindsley, D.H. 1964: Iron-titanium oxide minerals and synthetic equivalents. Journal of Petrology 5, 310-357.

Duchesne, J.C. 1972 : Iron-titanium oxideminerals in the Bjerkrem-Sogndal Massif,south-western Norway. Journal of Petrology13, 57-81.


The Vieux-Fort Anorthosite is a smallcomplex (~ 60 km2) located in the northe-astern Grenville Province, in Quebec’sLower North Shore area (see Fig. 1 ofPerreault & Hébert, this volume). Thecomplex is composed mostly of massive toweakly deformed anorthosite and leuco-norite with small amounts of leucogab-bro, pegmatitic gabbro, and gabbro; it isbordered to the north by mangeritic rocks(Fig. 1). South of the NE-SW-trendingVieux-Fort fault, leuconorite at the con-tact of the pluton is foliated parallel to theregional foliation in the adjacent PinwareTerrane granitic gneiss.

Zircons recovered from a massive, pegma-titic, quartz-bearing, granophyric leuco-norite are colourless to light tan, irregularfragments and shards, with a low uraniumcontent (38-267 ppm U, usually <100ppm U) and a high Th/U ratio (0.8-0.9),

features typical of zircons crystallizingrapidly at a late stage from a mafic mag-ma. Many grains have tiny, opaque or trans-parent mineral inclusions; a few grains arefractured. The U-Pb results show a slightscatter but the 207Pb/206Pb ages for all sixanalyses are clustered between 960 and976 Ma; they are 0.6-0.9 % discordantexcept for two analyses at –1.1 % and 1.5%. A best-fit regression yields an upperintercept age of 974.5 ±1.8 Ma, which isinterpreted as a crystallization age (Fig. 1).Thus, the Vieux-Fort Anorthosite is theyoungest known anorthosite in the Gren-ville; younger than the 1018 Ma Labrie-ville Anorthosite and the 1062 Ma Havre-Saint-Pierre Anorthosite.

The massive appearance of the anorthosi-te and of the related mangerite envelope,with tectonic foliation limited to the mar-gin of the pluton, and the absence ofmetamorphic mineral assemblages in theanorthositic rocks, suggest that the Vieux-Fort Anorthosite has escaped the majorGrenvillian (1080-980 Ma) deformationand associated metamorphism that affec-ted the eastern Grenville Province. Thecrystallization of the Vieux-Fort Anorth-osite at 974 Ma was coeval with late- topost-Grenvillian gabbros, syenites, mon-zonites and granites (980-966 Ma) in theeastern Grenville. Pb loss at 950 Ma, sug-gested by a slight scatter in the distributi-on of the analyses, is coeval with a clusterof titanite U-Pb ages (970-930 Ma), whichcorrespond to titanite U-Pb closure tem-peratures and to post-peak metamorphiccooling. We suggest that the Vieux-FortAnorthosite was emplaced after the mainGrenvillian deformation and metamor-phism, in a partially cooled, thickenedcrust. The Vieux-Fort Anorthosite and itsassociated mangeritic rocks may be coevalwith the earliest components of post-tec-tonic gabbroic and granitic magmatismresulting from melting at the base of thethickened Grenvillian crust.

The 974 Ma Vieux-Fort Anorthosite, Lower NorthShore, Québec: the youngest anorthosite in theGrenville Province

Serge Perreault 1 & Larry Heaman 2

1 Géologie Québec, 545, boul. Crémazie Est, bureau 1110, Montréal, Québec Canada H2M 2V1, 2 Dept. ofEarth & Atmospheric Sciences, University of Alberta, Edmonton, Canada T6G 2E3.

Fig.1. Simplified geological map of theVieux Fort area and U/Pb diagram.


The Havre-Saint-Pierre anorthosite suite(HSP) is located in the northeastern partof the Grenville Province (see Fig. 1 ofPerreault & Hébert, this volume). It iscomposed of several massifs or lobes sepa-rated from each other by structural zones.It includes a wide range of rock units typi-cal of AMCG suites. The HSP is composedof seven massifs: the well-known LacAllard massif, as well as the Rivière Rom-aine, the Rivière Magpie-Ouest, the Nord-Ouest, the Sheldrake, the Brézel, and theRivière-au-Tonnerre massifs. Most of theFe-Ti mineralizations are associated withthe lac Allard, the Rivière-au-Tonnerreand the Rivière Romaine massifs. TheNord-Ouest massif is characterised by Cu-Ni mineralizations associated with pyro-xenite and melanorite in anorthositicrocks at the margin of the massif (Clatk2000). The structural state of the HSPranges from undeformed and weaklyrecrystallized anorthosite to highly defor-med and recrystallized anorthositic gne-iss. Partially or totally deformed and recry-stallized mangerite and charnockite enve-lop most of the anorthositic massifs.Slivers of deformed and highly metamor-phosed supracrustal rocks are commonlyfound between the massifs. The emplace-ment of the HSP was complex: in thenorth, it was thrust over the countryrocks; in places, one massif was thrust overanother; and in the south, the relativelyyoung Rivière-au-Tonnerre massif wasintruded into an older massif. U-Pb zircongeochronology on various massifs shows arange in ages. A mangerite associated withthe Allard Lake massif was dated at 1126 ±6 Ma (Emslie & Hunt 1990). An age of1130 Ma was found for the Sheldrakemassif (in Verpaelst et al. in prep.), and anage of 1062 Ma was obtained for theRivière-au-Tonnerre massif (van Breemen& Higgins 1993).

In the Havre-Saint-Pierre anorthositicsuite, two contrasting styles of Fe-Ti

mineralization are present. In the easternpart of the HSP, massive hemo-ilmeniteand Fe-Ti-P2O5 mineralizations are com-monly found. Since 1950, QIT has extrac-ted an estimated 60 Mt of ore grading 38.8% Fe and 33.6 % TiO2 and the mean com-position of the ore is 34.2 % TiO2, 27.5 %FeO, 25.2 % Fe2O3, 4.3 % SiO2, 3.5 %Al2O3, 3.1 % MgO, 0.9 % CaO, 0.1 %Cr2O3, and 0.41 % V2O5. Massive hemo-ilmenite deposits are interpreted as immi-scible Fe-Ti-rich liquid derived from aparental jotunitic or oxide-rich noriticmagma, which intruded a partially crys-tallized andesine-bearing anorthosite. Onthe opposite side of the HSP, on the south-west border, Fe-Ti mineralizations aremostly composed of titaniferous magneti-te and a lesser amount of ilmenite. Thestyle of emplacement of the mineralizati-on ranges from primary massive magneti-te layers in a labradorite-bearing anortho-site to weakly discordant magnetitite andoxide-rich leuconorite veins emplacedroughly parallel to the regional tectonicfabric in the anorthosite. Finally, remobili-zed magnetite veins, magnetite-rich mon-zonitic pegmatites and magnetite veinsemplaced in late ductile to brittle shearzones cut the HSP anorthosite.

Massive ilmenite dykes and veinsThe world-class Tio Mine deposit, exploi-ted since 1950 by QIT, is the classic exam-ple of this type. The Tio Mine is located 42km north of Havre-Saint-Pierre (Fig.1)and was discovered in 1946 by geologistsof Kennecott Copper Inc. Over 20 signifi-cant showings were discovered by J. Rettyin 1941 and by Kennecott and QIT geolo-gists between 1944 and 1950 near lakesAllard and Puyjalon.

The Tio Mine (site 1, Fig.1) and mostsatellite deposits are located on the eastside of the HSP. There, the plagioclase inthe anorthosite is andesine and most

Contrasting styles of Fe-Ti mineralization in the Havre-Saint-Pierre anorthosite suite,Quebec’s North Shore, Canada

Serge Perreault

Géologie Québec, 545, boul. Crémazie Est, bureau 1110, Montréal, Québec Canada H2M 2V1.


deposits are closely associated with jotuni-te or oxide-rich norite (Rose 1969, Berg-eron 1986, Force 1991). These depositsoccur as dykes, sills, or stratoid lenses.Most associated norite or jotunite intrusi-ons show an ilmenite enrichment at thebase. Jotunite intrusions are composed ofplagioclase (An41-44), orthopyroxene,apatite, ilmenite, and magnetite. Locally,modal proportions of apatite and oxidesreach 10 % and 20 %, respectively. Recentmodels suggest that these massive ilmeni-te deposits originated from the separationof an immiscible iron- and titanium-richliquid from a jotunitic parent liquid(Force 1991). This immiscible oxide-richmagma then moved into fractures andshear zones in the crystallizing anorthosi-te to form dykes, sills, and tabular intrusi-ons of massive ilmenite, such as thoseobserved at the Tio Mine (Bergeron 1986,Force 1991).

The Tio Mine is composed of a subhori-

zontal, massive ilmenite intrusion for-ming three deposits separated by normalfaults. The main deposit measures 1097 min the N-S trend and 1036 m E-W. Itsthickness is estimated at 110 m and thedeposit is inclined 10° to the east. Thedeposit is divided into 3 zones. The upperzone is composed of massive hemo-ilme-nite, the intermediate zone of alternatinglayers of massive to semi-massive hemo-ilmenite and anorthosite, and the lowerzone of massive hemo-ilmenite. TheNorthwest deposit forms a band 7 to 60 min thickness of massive ilmenite alterna-ting with anorthosite. The deposit is gent-ly inclined to the east. The Cliff depositforms a hill that overlooks the Tio Mine,and has an ellipsoidal shape. Both theMain and the Northwest deposits aremined. The ore is composed of a dense,coarse-grained aggregate of hemo-ilmeni-te. Accessory minerals are plagioclase, spi-nel, pyrite, and chalcopyrite. Locally, themodal proportion of sulphides reaches 2

Fig.1. Simplified geological map with themajor Fe-Ti showings and deposits of theHavre-Saint-Pierre anorthosite suite (HSP).LAS, lac Fournier anorthosite suite; LAm,lac Allard massif; NQm, Nord-Quest massif;RATm, Rivière-au-Tonnerre massif; RRm,Rivière Romaine massif; Sm, Sheldrakemassif; RMOm, Rivière Magpie-Quest mas-sif; Bm, Brézel massif.


%. The Main deposit is estimated at 125millions tons. The Northwest deposit con-tains 5 Mt at 37.4 % Fe and 32.32 % TiO2,and the Cliff deposit contains 8.4 Mt at39.2 %Fe and 33.9 % TiO2.

Other similar deposits occur around LakesAllard and Puyjalon. The most promisingdeposits include Grader (which wasmined in 1950 for a short period),Springer, and Lac-au-Vent. All these depo-sits and smaller showings are composed ofmassive ilmenite dykes emplaced in theandesine-bearing anorthosite. One excep-tion worthy of mention is the Big Islanddeposit (site 2; Fig.1). This deposit is com-posed of a rutile- and sapphirine-bearing,massive hemo-ilmenite dyke with frag-ments of oxide-rich norite and anorthosi-te. Rutile can be as high as 15 modal %.Typical grab samples collected for assaysgave values of 26 % FeO, 21.1 % Fe2O3,40.3 % TiO2, 3.12 % MgO, 0.36 % V2O5,4.35 Al2O3, and 6.26 % SiO2 for a sapphi-rine- and rutile-bearing ilmenitite, and29.1 % FeO, 29.8 % Fe2O3, 38.9 % TiO2,2.99 % MgO, 0.39 % V2O5, 1.54 % Al2O3,and 0.91 % SiO2 for massive ilmenitite.The dyke is 15 m wide and crops out over500 m.

Magnetite-ilmenite-apatite mineralizationsMost ilmenite-magnetite-apatite minerali-zations are associated with jotunite on theeast side of the HSP (Fig. 1). These rocksare present at the contact between theanorthosite and mangeritic rocks. Thejotunite is usually coarse-grained, dense,and highly magnetic. Close to the contactwith the anorthosite, a primary layering iswell developed (Rose 1969, Hocq 1982,Bergeron 1986). Several deposits were dis-covered along with the hemo-ilmenitedeposits. Two typical deposits are descri-bed here. The Everett deposit (Fig. 1) cropsout along the northwest shore of LakePuyjalon. The ilmenite-magnetite-apatitemineralization is associated with a band ofjotunite 3 km long and up to 300 m thick.The mineralization is composed of 20 to30 % ilmenite, hematite, magnetite, andapatite, along with plagioclase and ortho-pyroxene. Locally, oxides and apatite makeup to 50 % of the rock. Gulf TitaniumLimited has estimated geological reservesin the deposit at 293 Mt grading 16.2 % Fe,9.75 % TiO2, and 4 % P2O5.

Other showings are located northeast ofthe Everett deposit and about 60 km nor-theast of Tio Mine (Fig. 1). The minerali-zations are similar to the Everett depositand are associated with jotunite emplacedat the contact of the anorthosite with themangerite. The main showing outcropsover 1 km and is 100 m wide. The minera-lization occurs as magnetite- and ilmeni-te-rich layers from a few centimetres tohalf a metre in thickness in the jotunite.The rock shows a well-developed bandedstructure. The jotunite is composed of 50to 60 % magnetite and ilmenite, 10 to 30% plagioclase, 10 to 30 % orthopyroxene,and accessory apatite and green spinel.

Massive ilmenite veins with ilmenite-rich leuconoriteThis type of mineralization is observed onthe southwest side of the HSP (East ofRivière-au-Tonnerre, Fig.1). It is closelyassociated with leuconoritic rocks intru-ding the labradorite-bearing anorthosite.The leuconorite forms thick dykes thatinvaded a partially crystallized anorthosi-te. The dykes are emplaced parallel to thefoliation observed in the anorthosite andcontain large enclaves of the host rocks.Contacts between the two rock types areusually sharp but are locally diffuse.Ilmenite (up to 10 %) and orthopyroxeneare disseminated in the leuconorite andlarge corroded and broken xenocrysts oflabradorite are common. The labradoritexenocrysts are fragments of the host anor-thosite. Massive ilmenite lenses and irre-gular veins are common within accumula-tions of orthopyroxene phenocrysts in thenorite. The thickness of these lenses andveins varies from 10 cm to 60 cm. Theilmenite veins are intrusive in the leuco-norite and locally they cross the contactbetween the anorthosite and the noriteand form thin veins of massive ilmenite inthe anorthosite. The ilmenite-rich veinsare composed of lamellar ilmenite (60 to80 %), plagioclase, orthopyroxene, greenspinel and locally apatite, clinopyroxeneand magnetite. Late biotite after ilmeniteis commonly observed. Values of 46.3 to59.3 % Fe2O3, 30.6 to 38.7 % TiO2, 2.06 to2.42 % MgO, 0.33 to 0.43 % V2O5, 1.64 to13.0 % SiO2, 1.35 to 5.90 % Al2O3, 0.2 to1.35 % CaO, and 0.01 to 1.1 % S wereobtained on 3 samples of ilmenitite veins.


Titaniferous magnetite mineralizationsNon-economic, titaniferous magnetitemineralizations are common in labradori-te-bearing anorthosite and are mostlyobserved in the southwest part of the HSP(West of Rivière-au-Tonerre, Fig. 1). Themineralizations occur in several forms: (1)layers of magnetitite in anorthosite; (2)layers, discordant veins, or dykes of mag-netite-rich melanorite or melagabbronori-te; (3) discordant veins and lenses of remo-bilized magnetitite; (4) magnetite-richmonzonitic pegmatite; and (5) magnetite-rich lenses or veinlets in late ductile andbrittle shear zones cutting anorthosite.

Two localities are worth mentioning. Atthe mouth of Rivière Chaloupe and at CapRond, the magnetite mineralization formscentimetre-thick layers of semi-massiveand massive magnetite in the anorthosite.Most of the layers are interpreted as pri-mary; however, some of the magnetite-rich layers are associated with the develop-ment of a tectonic foliation in ductileshear zones cutting anorthosite. This isparticularly well developed at the westernmargin of the HSP in the Rivière-au-Tonnerre lobe. Locally, in well-exposedoutcrops along the shoreline, the layers arefolded and locally reoriented in the ducti-le shear zones. Near the contact with themagnetitite layers, the anorthosite showsan increase in magnetite content. In thesame outcrop, remobilized magnetiteveins originating from the layers cross-cutthe tectonic foliation at a low angle.Fragments of undeformed labradoritexenocrysts are common in these veins.Plagioclase in the host anorthosite is high-ly deformed and recrystallized.

Veins of magnetitite or magnetite-richmelanorite from a few centimetres to 15metres thick are quite common in thispart of the HSP. These veins cut at a lowangle the tectonic and primary foliationand are oriented more or less parallel tothe regional tectonic fabric. Large xeno-crysts of labradorite and enclaves of thehost anorthosite are common. Layeredanorthosite enclaves also occur in thesemagnetite-rich veins. The magnetite con-tent varies from over 70 % in the oxide-rich veins to 30 to 50 % in the melanorite.Other minerals included plagioclase,orthopyroxene, hornblende, green spinel,

ilmenite, pyrite, chalcopyrite, clinopyro-xene, and apatite. Values of 49.5 % Fe and16.34 % TiO2 were obtained on one mag-netite-rich vein with a thickness of 15 m.Analyses of thinner veins show that theFe2O3 content varies from 50 % to 66 %,TiO2 from 10 % to 19.2 %, and V2O5 from0.24 to 0.36 %. However, due to the smallscale of these veins and also to the natureof the titaniferous magnetite, they arenon-economic.

ReferencesBergeron, R. 1986: Minéralogie et géochimie dela suite anorthositique de la région du Lac Allard,Québec : évolution des membres mafiques et ori-gine des gites massifs d’ilménite. UnpublishedPhD. Thesis, Université de Montréal, 480 pp.

Clark, T. 2000: Le potentiel en Cu-Ni ±Co ±ÉGPdu Grenville québécois : exemples de minérali-sations magmatiques et remobilisées.Chronique de la Recherche minière 539, 85-100.

Emslie, R.F. & Hunt, P.A. 1990: Ages and petroge-netic significance of igneous mangerite-char-nockite suites associated with massif anortho-sites, Grenville Province. Journal of Geology 98,213-231.

Force, E. 1991: Geology of titanium-mineraldeposits. Geological Society of America, SpecialPaper 259, 112 pp.

Hocq, M. 1982: Région du lac Allard. Ministèrede l’Énergie et des Ressources, Québec. DPV-894, 99 pp.

Rose, E.R. 1969: Geology of titanium and titanife-rous deposits of Canada. Geological Survey ofCanada, Economic Geology Report 25, 177 pp.

Van Breemen, O. & Higgins, M.D. 1993: U-Pb zircon age of the southwest lobe of the Havre-Saint-Pierre Anorthosite Complex, GrenvilleProvince, Canada. Canadian Journal of EarthSciences 30, 1453-1457.


Force (1991) stated that only two types ofore deposit account for more than 90% ofthe world production of titanium mine-rals: 1) magmatic ilmenite deposits and(2) young shoreline placer deposits. Hefurther noticed that the placer resourcesare declining while the consumption oftitanium minerals is increasing. Force(1991) concluded that "by the year 2010,we are likely to see titanium mineralsbeing produced from deposit types notcurrently exploited. This means a chal-lenge for economic geologists to definethese new ore deposit types". The purpose

of this contribution is to address this chal-lenge from a northeastern North Americanperspective. Regional metallogeny con-cepts are first applied, e.g., 1 (nature andextent of the proper source rocks, 2) weat-hering history actual (Quaternary) andpast (Eocambrian). Then, a specific locati-on where these favourable conditions seemto have occurred (e.g. Sutton, Québec) willbe examined in further detail.

The Grenville Province is a Mesoproter-ozoic orogenic belt flanking the southeas-tern margin of the Canadian Shield. It

Cambrian titaniferous paleoplacers metamorphosedduring the Taconian Orogeny, Quebec Appalachians

Michel Gauthier

Département des Sciences de la Terre, Université du Québec à Montréal, c.p. 8888, succ.‘Centre-Ville’,Montréal (Qc), Canada H3C 3P8.

Fig. 1. Geological sketch map of theGrenville Province and QuebecAppalachians.


occurs as a 500 km wide and 2 000 kmlong high-grade metamorphic belt withoutliers in the Paleozoic Appalachian oro-genic belt of New Jersey, Vermont, Maineand Quebec. The Grenville Province isknown for its large anorthosite massifsand for its numerous magmatic titanife-rous deposits. The most important ofwhich are Lac Allard in Québec andSanford Lake in the Adirondack Mount-ains of New York. Considering all this, theGrenville Province appears to be a veryprospective source region for younger tit-aniferous placers. Indeed, placers areknown to occur in Quaternary alluvialand deltaic deposits. The most importantis Natashquan, a large deposit occurringin the Natashquan river delta on the St.Lawrence North Shore. Another Quater-nary placer occurs where the Peribonkariver forms a delta in lake St. Jean. Boththe Natashquan and the Peribonka placerdeposits have been evaluated in recentyears (e.g., Natashquan by Tiomin in the1990s and Peribonka by SOQUEM in the1970s). Both deposits are large but do notseem to be economic at the present time.

Proper weathering of primary titaniumminerals seems to be a prerequisite for theformation of economic placer deposits.Low-latitude (below 30°) weathering isespecially important in increasing the per-centage of TiO2 in the ilmenite concentra-te (Force 1991). The duration of weathe-ring under favourable paleolatitude isanother important factor. Australia isquite favoured from that perspective. Onthe contrary, Natashquan and Peribonka,as well as the other Quaternary placersdeposited in the Grenvillian region, lackthese two prerequisites (i.g. the mainweathering process involved was a conti-nental glaciation and the depositionoccurred above 45°N).

In the search for new titanium mineralresources, the next question is: if modernconditions were unfavourable for the for-mation of economic placer deposits in theGrenvillian region, is there any geologicperiod in which the conditions were morefavourable? Such an approach has beensuccessfully applied to other ore deposittypes (e.g. stratiform copper deposits, cf.Kirkham 1989).

In Eocambrian (Late Neoproterozoic)time, the Grenvillian cratonic region expe-

rienced deep weathering under a low-lati-tude climate (Fig. 1). The widespreadoccurrence of red-bed sediments in riftfacies of the early Iapetus Ocean testifiesto such a paleoenvironment in the QuebecAppalachians (Gauthier et al. 1994). Ifpaleoplacers of this age are preserved inCambrian strata, they might be of bettergrade than their Quaternary equivalents.During the 1980s, such paleoplacers havebeen identified near Sutton (Fig. 1) in theQuebec Appalachians.

The Sutton paleoplacers occur at the baseof the Pinnacle Formation, in wacke andsandstone overlying alkaline volcanicrocks of the Tibbit Hill Formation. Theseheavy mineral deposits are marine placersthat were deposited on an extensive shoalwith strong drift currents at the mouth ofthe Proto-Ottawa river (Marquis & Kuma-rapeli 1993). Extending laterally for seve-ral kilometres, the horizon comprises tit-aniferous magnetite and ilmenite, repla-ced by a hematite-leucoxene assemblage,as well as primary rutile, zircon and tour-maline as heavy minerals. Massive heavy-mineral horizons average 3 m and locallyare 7 m thick. Together, the opaque mine-rals comprise 10 to 50 %, and locally 95 %,of the black sand paleoplacers.

The paleoplacer horizons were foldedduring the Mid Ordovician Taconian oro-geny. Detrital titanium minerals were she-ared and dispersed along the foliationplanes and formed fine-grained leucoxenedisseminations. These leucoxene grainsform fine intergrowths with chlorite.Local recrystallization of the detritalhematite-leucoxene assemblage to low-Timagnetite took place along Taconianthrust faults, producing an assemblage ofeuhedral magnetite and rutile with roun-ded detrital grains of zircon.

The rutile potential of these paleoplacerswas evaluated by SOQUEM in the early1990s. Although tonnages and grades wereimpressive, the fine grain size of the rutileand its intimate intermixing with hemati-te renders commercial extraction of therutile uneconomic. Although a majorsource region and favourable climate pre-vailed during the deposition of these pale-oplacers, subsequent deformation undergreenschist-facies conditions proved to bedeleterious. However, recrystallizationalong Taconian thrust faults was benefici-


al for rutile recovery.

The Sutton case history suggests that thesearch for new titanium mineral resourcesmay need the interplay of different disci-plines of the Earth sciences (e.g., igneouspetrology, paleo-climatology, sedimento-logy, metamorphic petrology and structu-ral geology), a combination which is notusual in economic geology. The deleteri-ous effects of low-grade metamophismand the positive effect of higher graderecrystallization on the Sutton titaniummineral deposits is reminiscent of theLake Superior type of iron deposits inwhich the iron-ore industry make a break-through in the 1950s by combining ametamorphic petrology approach withsedimentology (James 1954, 1955). Thismay be one of the ways of addressing thechallenge to economic geologists pointedout by Force (1991).

ReferencesForce, E.R. 1991: Geology of titanium-mineraldeposits. Geological Society of America, SpecialPaper 259, 112 pp.

Gauthier, M., Chartrand, F. & Trottier, J. 1994:Metallogenic Epochs and MetallogenicProvinces of the Estrie-Beauce Region,Southern Quebec Appalachians. EconomicGeology 89, 1322-1360.

Irving, E. 1981: Phanerozoic continental drift.Physics of the Earth and Planetary Sciences 24,197-204.

James, H.L. 1954: Sedimentary facies of iron formation. Economic Geology 76, 146-153.

James, H.L. 1955: Zones of regional metamor-phism in the Precambrian of northernMichigan. Geological Society of America Bulletin66, 1455-1488.

Kirkham, R.V. 1989: Distribution, settings andgenesis of sediment-hosted stratiform copperdeposits In: Boyle, R.W., Brown, A.C., Jefferson,C.W., Jowett, E.C. & Kirkham, R.V.(editors)Sediment-hosted Stratiform Copper deposits,Geological Association of Canada, SpecialVolume 36, 3-38.

Marquis, R. & Kumarapeli, P.S. 1993: An EarlyCambrian deltaic model for an Iapetan riftarmdrainage system, southeastern Quebec.Canadian Journal of Earth Sciences 30, 1254-1261.


Ilmenite deposits and mineralization in alkaline and subalkaline magmatic complexes of the Ukrainian Shield

Stepan G. Kryvdik

Institute of Geochemistry, Mineralogy, and Ore Formation of Ukrainian NAS, Kyiv, Ukraine.

Areas in eastern Europe

Several ilmenite deposits related to alkali-ne and subalkaline complexes are knownin the Ukrainian Shield. They usually con-tain variable contents of apatite, ilmeniteand magnetite. Proterozoic alkaline andsubalkaline complexes with ilmenite mine-ralizations predominate. A Devonian alka-line complex (Pokrovo-Kyreyevo) also occ-urs in the border zone with the Donbassfolded structure (Fig.1).

Ilmenite-bearing Proterozoic alkaline andsubalkaline complexes are subdivided intotwo types : alkaline-ultrabasic (with carbo-

natites) and gabbro-syenitic. These typeshave different geological ages: about 2.1 Gafor the first and 1.7-1.8 Ga for the second.

Ore concentrations of ilmenite associatedwith the alkaline-ultrabasic complexes arefound only in the Chernigovka (Novo-Poltavka) massif situated in the West Azovarea. Ore-bearing rocks are representedthere by alkaline pyroxenites (nepheline-free jacupirangites), which consist mainlyof alkaline pyroxene (aegirine-diopside,aegirine-salite), ilmenite and magnetite,and subordinate amphibole, biotite, apati-

Fig. 1. Massifs and occurrences of carbonatites, alkaline and subalkaline rocks in the UkrainianShield. Ultrabasic-alkaline complexes: 1 – Chernigovka (Novo-Poltavka), 2 – dykes of metajacupi-rangites, 3 – Proskurovka, 4 – Antonovka, 5 – Gorodnitsa intrusive body, 6 – fenites of BerezovaGat. Gabbro-syenitic complexes: 7 – Oktyabrsky, 8 – Mala Tersa, 9 – Pokrovo-Kyreyevo, 10 – Dav-ydky, 11 – Velika Vyska, 12 – South Kalchyk, 13 – Yelanchyk, 14 – Kalmius, 15 – Yastrubetsky, 16, 17 –aegirine syenites of Korosten and Korsun-Novomyrgorod plutons, 18 – Prymorsky, 19 – Melitopol,20 – Stremygorod apatite-ilmenite deposit. Geological blocks of the Ukrainian Shield: I –Northwestern; II – Dniepr-Bug; III – Ros-Tikich; IV – Ingulo-Ingulets; V – Middle-Dniepr; VI – Azov.


SiO2 TiO2 Al2O3 Fe2O3 FeO MnO MgO CaO Na2O K2O P2O5 S H2O- LOI CO2 Total

1 36.90 6.46 2.13 12.63 11.23 0.55 10.61 15.60 0.95 0.60 tr. 0.27 0.09 1.38 0.51 99.98

2 36.40 9.00 2.63 9.80 9.50 0.16 7.87 17.30 1.20 1.60 2.29 0.50 0.15 0.78 0.85 100.03

3 38.54 5.63 3.16 9.92 10.14 0.41 9.84 15.62 1.84 1.05 0.89 0.47 0.13 0.96 1.54 100.14

4 40.32 7.02 12.55 3.01 13.16 0.29 6.82 11.32 2.74 0.40 0.09 0.26 0.08 1.11 0.48 99.65

5 35.60 6.40 12.90 12.55 10.90 0.20 7.30 11.28 1.45 0.35 0.05 0.06 - - - 99.04

6 33.85 10.02 2.52 11.12 13.68 0.19 10.50 14.27 0.64 0.40 0.90 0.16 0.18 1.02 0.35 99.80

7 38.86 9.05 3.58 6.56 12.97 0.09 11.72 14.23 0.74 0.39 0.21 0.08 - 1.28 - 99.76

8 31.50 7.22 3.06 5.90 28.60 0.43 6.40 10.86 0.72 0.45 2.72 0.23 0.20 1.61 - 100.00

9 29.37 8.05 1.11 7.44 32.84 - 5.81 9.91 0.74 0.40 2.15 0.05 0.24 1.20 0.25 99.56

10 32.80 5.25 7.57 3.42 28.87 0.37 6.06 8.40 1.51 0.88 3.68 - 0.05 0.74 - 99.60

11 34.80 9.36 11.45 6.67 17.80 0.23 6.90 5.57 1.93 1.12 0.76 0.11 0.12 2.70 0.29 99.81

12 34.29 5.97 11.69 6.29 17.26 0.24 4.94 8.53 1.89 1.32 3.49 0.41 0.22 3.14 0.31 99.99

13 24.61 10.60 7.90 6.65 23.47 0.23 4.72 10.79 1.20 0.50 5.88 0.55 0.06 2.13 0.38 99.67

14 41.26 5.63 14.35 5.52 12.83 0.20 4.26 5.73 2.74 2.30 1.08 0.14 0.20 3.29 0.41 99.94

15 33.00 7.60 11.60 4.00 13.60 0.40 5.50 13.70 2.10 0.70 6.20 tr. 0.20 1.40 - 100.00

16 33.79 7.61 10.95 6.56 15.93 0.24 6.93 10.10 2.12 0.71 3.65 - - 1.00 - 99.59

17 25.20 10.28 5.30 4.20 12.20 0.42 7.50 14.20 1.40 0.48 7.30 0.25 - 1.20 - 99.93

18 28.53 9.66 6.94 6.20 24.43 0.25 8.85 9.42 1.10 0.57 4.58 - - 0.45 - 100.98

19 24.00 13.60 1.90 3.30 29.10 0.48 9.60 9.70 0.70 0.70 4.87 0.13 0.20 1.10 - 99.38

Table 1. Chemical composition of someilmenite- and apatite-bearing rocks fromalkaline and subalkaline magmatic com-plexes of the Ukrainian Shield.

Chernigovka (Novo-Poltavka) massif: 1, 2 - ore alkaline pyroxenites; 3 – average composition. Oktyabrsky mas-sif: 4, 5 - ore gabbro. Pokrovo-Kyreyevo massif: 6, 7 - ore pyroxenites. South-Kalchyk: 8, 9 - ore peridotites; 10 -ore gabbro. Davydky massif: 11 – troctolite; 12 - ore gabbro; 13 – ore cumulate; 14 - endocontact gabbro.Stremygorod deposit: 15, 16 - ore troctolites; 17, 18 - ore peridotites; 19 – ore cumulate.

te, calcite and sphene. The ilmenite con-tents in these rocks vary between 5 and15%, with an average at 10 % (Table 1).The ilmenite is enriched in Nb (up to 0.3%). Some varieties of carbonatites fromthese complexes have high concentrationsof ilmenite, which can thus become a use-ful accompanying mineral.

In the gabbro-syenitic complexes the ilme-nite-rich rocks are represented by ore gab-bros, troctolites, cumulate peridotites andpyroxenites. Two types of massifs are dis-tinguished. In the first one, gabbro-syeni-tic massifs are associated with spatiallyand genetically related anorthosite-rapa-kivi granite plutons, e.g., the Davydkygabbro-syenitic massif, which lies in thenorthern part of the Korosten anorthosi-te-rapakivi-granite pluton (Fig. 1). TheSouth-Kalchik massif (Azov area) repre-sents a syenitic analogue of such plutons.In the second type of gabbro-syeniticcomplexes, ore gabbroids and pyroxenitesare genetically related to compound poly-phasial massifs (Oktyabrsky, Pokrovo-Kyreyevo), whose development termina-ted in nepheline syenites (mariupolites,foyaites, juvites).

Among the ilmenite-bearing gabbroidsand peridotites of the gabbro-syeniticcomplexes, apatite-ilmenite-ore and essen-tially ilmenite-ore varieties are to be found(contents of ilmenite amount to 20%, andapatite to 5-10%). These rocks are rathersimilar to the ore troctolites of the Stre-mygorod ilmenite-apatite deposit, locali-zed in the middle of the Korosten pluton.In our view, gabbroids of this deposit dif-fer essentially from typical rocks of theKorosten pluton by their increased alkali-nity (presence of titanaugite). All theabove-mentioned deposits related to apa-tite-ilmenite gabbroids and peridotiteshave a magmatic origin and were formedmainly by crystal accumulation of ilmeni-te, magnetite and apatite. Their parentalmagmas were enriched in titanium and/orphosphorus, either at the mantle site ofgeneration (nephelinites in the alkaline-ultrabasic complexes and alkaline basaltsin the gabbro-syenitic types), or by a diffe-rentiation process in intermediate-levelmagma chambers.

Massive anorthosites represent one of thelargest occurrences of Precambrian basicmagmatism. Proterozoic anorthosite com-plexes contain numerous Fe-Ti deposits ofeconomic grade and have therefore beenstudied extensively. The Korosten Plutonoccurs in the northwest part of theUkrainian Shield, in the Volyn Block. Thepluton is a multiple anorogenic intrusion,composed of the classical association ofgabbroid - anorthosite - rapakivi-granite,which intruded in Proterozoic times(1737-1800 Ma). The rocks of the Koro-sten Complex occupy an area of 12,000km2. The largest gabbro-anorthosite mas-sifs are: Volodarsk-Volynsky (1250 km2),Chepovichsky (830 km2), Fedorovsky(104 km2), Pugachevsky (15 km2) andKrivotinsky (30 km2); they occupy 18% ofthe area. The massifs are flat, sub-horizon-tal bodies with thicknesses from 0.5 to 3km. Gabbro-anorthosite varieties of basi-tes predominate in most bodies. Anor-thosites and mesocratic gabbroids are lessabundant. Ultrabasites are rare. Anor-thosites (gabbro-anorthosites) and gab-broids have long been considered ashaving been intruded in two phases(Polkanov 1948, Bukharev et al. 1973a,b).Recent detailed field studies have, howe-ver, shown that both anorthosites andgabbroids were intruding repeatedly(Mitrokhin 2000, Zinchenko et al. 2000].Field, mineralogical, petrographical andgeochemical investigations have broughtnew evidence showing that large gabbro-anorthosite massifs are multiple intrusi-ons, composed of several age associations(suites) of basic rocks. In the Volodarsk-Volynsky, Chepovichsky, Fedorovsky andPugachovsky massifs, five age suites ofbasic rocks are distinguished: Early Anor-thosite Suite (A1), Main Anorthosite Suite(A2), Early Gabbroic Suite (G3), LateGabbroic Suite (G4) and Dyke Suite (D5).

The Early Anorthosite Suite, Main Anor-thosite Suite and leucodolerites of the

Dyke Suite form successive trends of sub-alkaline high-alumina basic rocks. Withineach trend the decreasing plagioclase con-tent with addition of mafic minerals isaccompanied by the regular descent ofSiO2, Al2O3, CaO, Na2O and increase ofTiO2, FeO and MgO. The ratios FeO:MgOand TiO2:FeO within each suite are almostconstant and correspond to fractionationand accumulation of plagioclase duringascent to the final level of emplacement.Within the high-alumina basite trend,transition from A1 anorthosites - leucono-rites to A2 anorthosites -leucogabbros-norites and to D5 leucodolerites can betraced in the variation of mineral associa-tions and compositions: Pl42-57 + Opx22-46

to Pl45-62 + Pig44-54 + Aug38-45 ± Ol45-59 ±Fsp to Pl45-60 + Aug + Fsp ± Ol57-60, withincreasing contents of Fe-Ti oxides andapatite (the indices refer to 100*An/An+Abin plagioclases and 100*Fe/Fe+Mg inmafics). This evolution is accompanied bydecreasing contents of Mg, Cr, Ni and Srand increasing contents of Ti, Fe, Mn, K, P,V, Sc, Ba, La, Ce, Zr, Y, Cu and Mo with a regular increase of FeO:MgO andTiO2:FeO ratios.

The Early Gabbroic Suite, Late GabbroicSuite and trachydiabase - trachyandesiba-salts of the Dyke Suite form trends of sub-alkaline, moderately-aluminous, basicrocks. The major chemical evolution, wit-hin each suite, is marked by noticeablechanges in FeO:MgO and TiO2:FeO rati-os. This is due to differentiation with sepa-ration of plagioclase and mafic minerals.From G3 gabbro-norites to G4 olivine gab-broids G4 and to D5 trachydiabases, thetransition is Pl43-56 + Pig45-55 + Aug16-17 toPl28-60 + Aug38-56 + Ol32-77 ± Pig56-64 ± Fspand to Pl25-62 + Aug42-72 + Fsp, with abruptaddition of Fe-Ti oxides and apatite in G4.

Taking into account the 'anorthosite sche-me' of crystallisation for high-aluminabasic suites it may be proposed that the


The gabbro-anorthosite massifs of the KorostenPluton (Ukraine) and problems concerning the evolution of the parental magmas

Alexander V. Mitrokhin

Geological Faculty of Kiev Taras Shevchenko National University, Ukraine.


depletion in Mg, Cr, Ni and Sr and enrich-ment of Ti, Fe, Mn, K, P, V, Sc, Ba, La, Ce,Zr, Y, Cu and Mo with increasingFeO:MgO, and TiO2:FeO, in row: (anor-thosite -> leuconorite A1) -> (anorthosi-te - leucogabbro-norite A2) -> (leucodole-rite D2), are connected with the deep com-positional evolution of the parental high-alumina basalt magmas. The Main Anor-thosite Suite and Late Gabbroic Suite haveinherited minerals and chemical composi-tions, which allows to consider them ascumulates (A2) with a complementaryresidual liquid (G4). The new data on theevolution of the Korosten basic magma-tism point to the possible occurrence ofTi-Fe-P magmatic deposits in the G4 andD5 formations.

References Bukharev, V.P., Stekolnikova, A.V. & Poliansky, V.D.1973a: Tectonics and deep framework of anor-

thosite massives in the north-west part ofUkrainian Shield. Geotectonics 4, 34-41 (inRussian).

Bukharev, V.P., Kolosovska, V.A. & Hvorov, M.I.1973b: The features of formation of the anor-thosite massives in the north-west part ofUkrainian Shield. Soviet Geology 6, 125-132 (in Russian).

Zinchenco, O.V., Mitrokhin A.V. & Moliavko, V.G.2000: New data on stages of basic rocks for-mation of the Korosten pluton. Reports ofNational Academy of Sciences of Ukraine 4,128-130 (in Russian).

Mitrokhin, A.V. 2000: Age relationships betweenbasic rocks of the Korosten pluton. Visnyk of KievUniversity, Geology 16, 15-20 (in Russian).

Polkanov, A.A. 1948: Gabbro-labradorite plutonof Volyn area of the USSR. Leningrad University.80 pp. (in Russian).

The Kamenka layered intrusion was disco-vered in 1965 during geological surveying.Initially it was attributed to the KorostenAMCG complex, which occurs about 50km to the southeast of Kamenka. How-ever, strong differences in chemical com-position between the leucocratic Kamenkabasites and Korosten anorthosites werelater recognised (Kogut et al. 1992), andsome characteristics in the Kamenkawhole-rock chemistry were recently used(Zinchenko et al. 1998) to argue for anaffiliation with the huge Paleoproterozoiccontinental flood-basalt (CFB) province.

The Kamenka peridotite-gabbro-anortho-site layered intrusion is situated on thenorthern edge of the Ukrainian Shield. Itis assumed from geophysical data thatKamenka is a typical batholith with a totalarea of about 500 km2 and a thickness ofup to 1300-1500 m. The main part of theintrusion is fault-bounded and down-

thrown by about 500 m and overlain byPhanerozoic sedimentary rocks; only asmall 2-3 km-wide southern strip occursnear the surface beneath a 10m-thickQuaternary sedimentary cover.

The Kamenka massif intrudes into c.1.975 Ga granites (and their comagmaticaltered basites) and into metamorphosedvolcanic and sedimentary rocks. The ageof the Kamenka intrusion, as concludedfrom thermo-ionic emission analyses onzircon is 1.98 Ga. The rocks of theKamenka intrusion are fresh, with onlyolivine-rich cumulates moderately ser-pentinized.

A Bottom zone and a Main Layered zonehave been distinguished within theKamenka intrusion.

The Bottom zone is composed of melano-cratic and mesocratic mafic rocks (ortho-

First approach to the petrology of the Kamenka peridotite-gabbro-anorthosite intrusion

Leonid Shumlyanskyy

Institute of Fundamental Studies, P.O.Box 291 Kyiv 01001, Ukraine.

clase-bearing gabbro and peridotite, mon-zonitic gabbro, monzonite, syenite, etc.),with heterogeneous, quite often inequi-granular (ataxite) fabrics. Thin graniteveins are noted in places within theserocks. The usual thickness of the Bottomzone is about 10 m. These characteristicssuggest a high degree of contamination byacidic country rock.

The Main Layered zone is composed ofperidotite, plagioclase-bearing peridotite,olivine-bearing melanocratic gabbro andmelanocratic troctolite, meso- and leuco-cratic olivine-bearing gabbroid and anor-thosite. These rocks sequentially gradeinto each others and appear to constitute anatural line of differentiation. At leastthree layered units (lower, middle andupper) are recognised within the MainLayered zone.

A high Mg-number, calculated as#Mg=Mg2+/(Mg2++Fe2++Fe3+×0.9), highAl2O3 abundances and high contents, ofNi and Sr are characteristic for theKamenka intrusion. By contrast abundan-ces of TiO2, CaO, Cu, Zn, V, Sc, Rb, Zr, Nb,Yb, and Y are relatively low compared withtypical CFB.

The Kamenka massif formed as a result ofcrystallization of high-Al tholeiitic meltwhich belonged to the PaleoproterozoicCFB province. The average compositionof the Kamenka intrusion shows that themelt underwent extended crystallizationprior to emplacement into the ultimatechamber. Although the Mg# of the avera-ge melt was rather high (0.62), at least10% of high-Mg olivine (Fo91) must beadded to it in order to bring it into equili-brium with mantle residua. Meanwhile,the high Al2O3/CaO ratio (2.15 on avera-ge) provides evidence of extended crystal-lization of clinopyroxene or clinopyroxeneplus plagioclase (in addition to olivine) enroute to the ultimate chamber. It is note-worthy that the Kamenka line of descentfollows the differentiation trend definedby the Paleoproterozoic CFB dykes. Thistrend appears to be due to precipitation ofan assemblage of An92+Fo79 in the pro-portion 1.5:1.0, which suggests shallow(up to 10 km) crystallization (Green1970). Thus, at least a two-stage evolutionmust be invoked for the Kamenka intrusi-on. The first stage involved crystallizationof high-Mg olivine along with clinopyro-

xene at depth, while the second stageinvolved precipitation of plagioclase plusolivine in a shallow intermediate chamber,and terminated with at least three injecti-ons into the Kamenka ultimate magmachamber.

Although a genetic link between theKamenka intrusion and other intrusionsof the Paleoproterozoic CFB provinceseems to be clear, the origin of the meltwith the observed chemical features is stillunexplained. Extended (up to 30-40%)partial melting of a dry pyrolite source ata depth of 70-100 km is usually accepted(Green 1970) for a tholeiite-picrite liquidthat is believed to be primary for conti-nental tholeiites. Known partition coeffi-cients and assumed mantle composition(Frey et al. 1978) predict that the abun-dances of most trace elements should bemuch higher than observed. Supposedprecipitation of an olivine plus clinopyro-xene assemblage and possible contamina-tion with crustal material en route canonly increase the abundances of the mostincompatible minor elements. Three meltcomposition models, which include 1)simple equilibrium 5% partial melting ofundepleted mantle; 2) simple equilibrium30% partial melting of depleted model 1mantle, and 3) model 2, complicated byprecipitation of 10% of olivine and conta-mination with production of 3% partialmelting of lower crust, were calculated.Compared with the Kamenka melt com-position all models reveal similar features:1) the abundances of Ba, Sr, Co, and Ni inthe Kamenka melt are much higher thanpredicted in the models; 2) abundances ofRb, Zr, Y, Sc, V, Zn, Cu, and Cr are muchlower than predicted. Among them, onlythe high content of Ba and Sr can beexplained by contamination with crustalmaterial, whereas low Cr contents may beattributed to extraction of clinopyroxeneor spinel. It should be noted here that theSr content of the Kamenka rocks is at leasta factor of two higher than that typical forother CFB rocks, which requires a specialexplanation. Neither incompatible ele-ments, such as Rb, Zr, Y, Sc, V and Zn, norcompatible (Ni) or partly-compatible(Co) elements fit the models. It is especi-ally difficult to explain the high Ni con-tent, which may appear due to the presen-ce of sulphur in the melt, coupled with anextremely low Cu abundance.



It seems to the author that no reasonableexplanation can be found at present to fitthe observed abundances of minor andmajor (especially high contents of Al andlow contents of Ti and P) elements. Moreextended geochemical investigations arerequired.

ReferencesKogut, K.V., Galiy, S.A., Skobelev, V.M., Zaionts, I.O.& Kondratenko P.A. 1992: Petrology and nickelmineralization of the Kamenka massif, north-western part of the Volynian block of theUkrainian shield. Geol. Zhurnal 6, 109-118 (inRussian).

Zinchenko, O.V., Shumlyanskyy, L.V., &Molyavko, V.G. 1998: Continental flood basaltprovince at the south of the East-Europeancraton, its composition, volume and stratigrap-hic position. In: Shcherbak, N.P. (ed.) Geologyand stratigraphy of the Precambrian deposits of

the Ukrainian shield, Institute of Mineralogy,Geochemistry and Ore Formation of theUkrainian Academy of Sciences, 102-104 (inRussian)

Skobelev, V.M., Yakovlev, B.G., Galiy, S. A., Kogut,K.V., Chernyshov, N.M., Zaionts, I.O., Homyak, T.P.,Verhoglyad, V.M. & Krochuk, V.M. 1991:Petrogenesis of the nickel-bearing gabbroic intru-sions of the Volynian megablock of the Ukrainianshield. Nauk. Dumka publisher, 140 pp.

Green, D.H. 1970: A review of experimental evi-dence on the origin of basaltic and nephelini-tic magmas. Physics of the Earth and PlanetaryInteriors 3, 221-235.

Frey, F.A., Green, D.H. & Roy, S.D. 1978: Integratedmodels of basalt petrogenesis: a study ofquartz tholeiites to olivine melilitites fromsouth-eastern Australia utilizing geochemicaland experimental petrological data. Journal ofPetrology 19, 463-513.

A Cameca IMS-4f ion microprobe hasbeen used for investigation of ilmenitesfrom different rocks of the UkrainianShield (gabbroids, alkaline pyroxenitesand carbonatites). Ilmenite crystallizes ina wide range of physico-chemical conditi-ons, which usually results in isotope frac-tionation of elements such as O, C, S, etc.Previous investigations had shown thatfractionation of Ti isotopes occurs also inmeteorites and in some terrestrial mine-rals (Fahey et al. 1987, Zinner 1989). It isespecially common in meteorites (-70 upto +273 ‰ for δ50Ti) (Fahey et al. 1985).Terrestrial minerals have not been investi-gated to the same extent as space material.

A notable fractionation of Ti isotopes hasbeen established in ilmenites from alkali-ne pyroxenites and carbonatites of the

Chernigovka complex, Azov area (Fig.1).The present authors tentatively explainthis fractionation through variations inthe physico-chemical conditions of for-mation of the ilmenite-bearing rocks(temperature, fugacity of oxygen, alkalini-ty). In the example of two pairs of ilmeni-tes from carbonatites and alkaline pyroxe-nites, an increase in heavy (49Ti and 50Ti)isotopes was discovered in rocks formedin more alkaline and oxidizing conditions.It was also supposed that the fractionationoccurred at a lower temperature than theprimary crystallization temperature (accep-ted at about 1000°C). Hence, oxidationprocesses and increase of alkalinity of themagma, and, perhaps, decreasing tempe-rature, promote enrichment of Ti heavyisotopes.

Ti, Mg, and O isotopic compositions in ilmenites fromUkrainian ore deposits as indicators of their crystalli-zation conditions

Petro I. Didenko 1, Stepan G. Kryvdik 2 & Vasyl M. Zagnitko 2

1 Institute of Environmental Geochemistry, NAS Ukraine, Kyiv, Ukraine. 2 Institute of Geochemistry,Mineralogy and Ore Formation, NAS Ukraine, Kyiv, Ukraine.

Ti isotopic compositions were also measu-red in individual ilmenites from gabbroidsof the Korosten pluton. In these samples,δ49Ti and δ50Ti have rather insignificantvalues (0 and 1.5 ‰, respectively). They aresimilar to ilmenites from carbonatite (N 1).Microprobe analyses of ilmenites from theKorosten pluton and the Stremygorod apa-tite-ilmenite deposit show that they are alsocharacterized by rather low contents of thehematite component (0.5-1.9 % Fe2O3).This testifies to rather reduced conditionsof crystallization.

Mg isotopic compositions were determi-ned in ilmenites from basic and ultrabasicrocks of the Korosten pluton and relatedcomplexes (Gorodnytza, Prutovka andVarvarovka). The ultrabasic rocks of theGorodnytza intrusion are alkaline (jacupi-rangites and melteigites) and consideredto be of deep (mantle) origin (δ26Mg=2±1‰). The Prutovka complex belongs to adifferentiated intrusion, the basic andultrabasic rocks (early differentiates) ofwhich are nickel-bearing. This complexbelongs to a Precambrian trap formation.The Varvarovka complex is supposed tohave been significantly contaminated bycrustal granitoid material. Starting fromthese characteristics, we explain the vari-ous Mg isotopic compositions in ilmenitesfrom these complexes. Thus, the δ26Mg inilmenite from the Prutovka complex isclose to that in diabases, and the δ26Mg inilmenites from the Varvarovka complex isnearer to those in granites. The data obtai-ned in the Prutovka complex suggest adefinite dependence of δ26Mg values onthe degree of oxidation of iron in ilmeni-tes (Fe2O3 contents).

The tendency for an increase in the weightof O isotopes was formerly established incarbonates from carbonatites of theChernigovka complex (Kryvdik et al.1997). It has been shown that an increasein the Fe3+/Fe2+ ratio in a rock is correla-ted to increasing 18O values in carbonates,from 5 up to 17.5 ‰. δ18O in magnetitefrom these carbonatites (+0.3 up to +5.0‰) indicates an equilibrium isotope dis-tribution between carbonates and oxides(magnetite, ilmenite) at rather high PTvalues. The isotope distribution of carbonbetween carbonates and graphite in thesame rocks also suggests a high tempera-ture of equilibrium crystallization of theseminerals.

O isotopic compositions (conventionalmass spectrometry) of ilmenites fromalkaline pyroxenites show a narrow rangeof δ18O from +1.96 up to +2.98 ‰, with atrend of increasing δ18O with increasingalkalinity and oxidation of the rocks, asalready mentioned above for Ti isotopesin ilmenites from the same rocks (δ50Ti).

In conclusion, Ti, Mg and O isotopic com-positions in ilmenites depend on the rocktype and on the crystallization conditions(temperature, alkalinity, oxygen fugacity).Our investigations point to oxygen fugaci-ty as being one of the primary factors con-trolling the isotopic composition of mag-matic ilmenite.

ReferencesFahey, A.J., Goswami, J.N., McKeegan, K.D. &Zinner, E. 1985: Evidence for extreme 50Tienrichments in primitive meteorites.Astrophysical Journal Letters 296, 17-20.

Fahey, A.J., Goswami, J.N., McKeegan, K.D. &Zinner, E. 1987: 26Al, 244Pu, 50Ti, REE, and traceelement abundances in hibonite grains fromCM and CV meteorites. Geochimica etCosmochimica Acta. 51, 329-350.

Kryvdik, S.G., Zagnitko, V.M. & Lugova, I.P. 1997:Isotopic compositions of minerals from carbo-natites of Chernigovka complex (Azov area) asindicator of crystallization conditions.Mineralogical Journal 19, 28-42.

Zinner, E. 1989: Isotopic measurements withthe ion microprobe. U.S. Geological SurveyBulletin 1890, 145-162.

1 0 0 - SPECIAL PUBLICATION 9 I L M E N I T E D E P O S I TS A N D T H E I R G E O LO G I C A L E N V I R O N M E N T

Fig.1. Dependence of δ50Ti versus (Fe2O3 +MnTiO3) contents in ilmenites from carbo-natites and alkaline pyroxenites (Sp. 1-4) ofthe Chernigovka complex, gabbroids ofthe Korosten pluton (Sp. 5), ore troctolitesof the apatite-ilmenite Stremygorod depo-sit (Sp. 6) and Lunar rocks (Sp. 7) (Fahey etal. 1987).

I L M E N I T E D E P O S I TS A N D T H E I R G E O LO G I C A L E N V I R O N M E N T SPECIAL PUBLICATION 9 - 1 0 1

Ore peridotites are exposed near Travya-naya Bay on Lake Keret, North Karelia(Russia), in the Belomorian Mobile Belt(BMB), eastern Fennoscandian Shield, as apart of the differentiated Palojarvi massif.Mineralization, composed of ilmenite andmagnetite with, locally, traces of pyriteand chalcosine, occurs in host ore rockssuch as harzburgite, lherzolite, websteriteand amphibolite. Harzburgite shows asideronite texture and consists of 25 % oli-vine (Fa41), 21% orthopyroxene (Fs34),41% magnetite and 10% ilmenite andwith secondary minerals such as horn-blende (f = Fe2++Fe3+/Fe2++Fe3++Mg=0.25), chlorite, serpentine, garnet, carbo-nate and spinel. Augite is commonly pre-sent in lherzolite and websterite. Theserocks show gradual transitions and, obvi-ously, form a continuous differentiatedsequence. During amphibolization theywere converted into ore amphibolites,most similar in chemical composition towebsterite.

The valuable components of the ore rocksare represented by TiO2 (3.0-7.4 wt.%),V2O5 (0.21-0.36 wt.%), Pt (up to 0.52ppm), Pd (up to 2.8 ppm) and Au (up to2.5 ppm). Ilmenite concentrates Ti, andmagnetite, clinopyroxene and amphiboleconcentrate V (in decreasing order). Sepa-rate mineral phases of Pt and Pd are unk-nown, but the maximum contents of theseelements are observed in olivine and ilme-nite. Au contents are maximum in ilmeni-te and much smaller in magnetite and oli-

vine. The average amount of Pt + Pd is1.23 ppm in ore peridotite and olivinewebsterite, 1.48 ppm in amphibolizedwebsterite and 1.022 ppm in ore amphi-bolite.

Ti, V, Pt, Pd and Au in Travyanaya Bay ore peridotitesand their possible genetic relation with BelomorianMobile Belt anorthosites

Vladimir S. Stepanov

Institute of Geology, Karelian Research Center, RAS, Petrazavodsk, 185610 Russia.

Fig 1. Geological map of the "Guba Travya-naya" area. Legend: 1 – quaternary rocks;2 – pegmatitic veins ; 3 – leucocratic, mostlyplagioclasic amphibolites; 4 – garnet- and gar-net feldspathic amphibolites; 5 – ore amphi-bolites (Opx & Cpx); 6 – ore websterites; 7 –ore peridotites with microareas of symplecti-tes; 8 – ore peridotites; 9-11 – structural ele-ments ; 12 – outcrops

Ore rocks form small bodies in a Fe-rich (f> 0.5) garnet-feldspathic amphibolite zone.About 10 bodies composed of ore rockshave been discovered. The largest andmost thoroughly studied body has a visiblethickness of 30 m, dips steeply east and hasbeen traced for 100 m along strike (Fig. 1).The boundary between the body and pla-gio-amphibolite extends across a 30-40cm-thick thin-laminated zone formed bycompositionally variable amphibolite. Theore rock-body wedges out northwards andsouthwards. The other ore rock- bodiesare much smaller.

The association between ore rocks andgarnet plagio-amphibolites is interpretedas a metamorphosed differentiated seriesin the fairly large (about 8 km2) Palojärvimassif, which has been tectonically divi-ded into three large fragments, markedlydiffering in chemical composition. Thelargest, western fragment is built up bygarnet-feldspathic amphibolite, withzones exhibiting well-preserved gabbrostructures. The eastern fragment is struc-turally similar, but its relict zones containboth gabbronorite and plagioperidotite.The southern fragment is richer in Fe (f >0.5) than the western and eastern frag-ments, and is presumably composed oflate differentiates. Ore peridotites haveonly been found in this fragment, and areinterpreted as melanocratic cumulates.

Leucocratic plagio-amphibolite, similar toapoanorthositic rocks, occur in the eas-tern portion of the fragment.

The Palojärvi massif thus comprises diffe-rentiated rocks from Mg-rich plagioperi-dotite (eastern fragment) to rocks similarto ferrogabbro and ore peridotite. Anintermediate position is occupied bymesocratic and leucocratic amphibolites(f < 0.5), corresponding in chemical com-position to gabbro and leucogabbro.

Comparison of the Palojärvi rocks withother BMB intrusive rocks shows that theore rocks have no counterparts in the pro-vince. The Mg-rich portion of the massifis most similar to the mesocratic units inthe gabbro-anorthosite complexes (Nizhne-popovsky, Boyarsky, Severopezhostrovskyand other massifs). Both gabbro-anortho-sites and Palojärvi rocks are intenselymetamorphosed and foliated; therefore,gabbroids pass into plagio-amphibolitesthroughout the studied area. Ilmenite,which is characteristic of Travyanaya Bayrocks, is a major ore mineral in gabbro-anorthosites.

To sum up, ore rocks similar to those inTravyanaya Bay are likely to be found inconnection with gabbro-anorthosite mas-sifs elsewhere in the BMB.


Precambrian anorthosites in the Belomorian MobileBelt, eastern Fennoscandian Shield

Vladimir S. Stepanov & Alexandra V. Stepanova

Institute of Geology, Karelian Research Centre, RAS, Petrazavodsk, 185610 Russia.

In the Belomorian Mobile Belt (BMB)(western White Sea region), anorthositesare known in at least three geological set-tings:

1. The most strongly differentiated leuco-cratic portion of an Early Proterozoic (2.4Ga), drusitic, lherzolite-gabbronorite com-plex (Stepanov 1981) is represented byanorthosites. They are subordinate andform small lenses up to tens of metres inthickness and schlieren in olivine gabbro-

norites and plagioclase lherzolites. Massifsof this complex have primary intrusivecontacts with chilled zones (1.5-2 m),apophyses and enclosed rock xenoliths. Inaddition to these intrusive forms, dykes arealso characteristic of the complex. Theparent magma of these intrusions corres-ponds to high-MgO, high-Cr tholeiite, richin SiO2, determined as a mean of chilledmargins with the following composition(wt.%, ppm, n=31): (50.13 SiO2, 0.69 TiO2 ,11.41 Al2O3, 1.59 Fe2O3, 9.02 FeO, 0.19


MnO, 14.89 MgO, 8.61 CaO, 1.70 Na2O,0.53 K2O, 0.07 P2O5, 0.11 H2O, 172 Cr, 80Ni, 47 Co, 243 V, 4 Zr, 9 Y. This type ofanorthosite is described as stratiform.

2. Anorthosites are an essential compo-nent of gabbro-anorthosite massifs (gab-bro-anorthosite complex; Stepanov 1981).In these bodies, anorthosites as well as leu-cocratic and mesocratic gabbro stronglydominate over melanocratic rocks that dis-play a cumulate structure and commonlyoccur in the lower part of the intrusion.Judging by the composition of chilled rockautoliths in the Boyarsky complex, theparent magma of this complex seemed tocorrespond in composition to high-alumi-na gabbro with (wt.%, ppm) 52.35 SiO2,0.17 TiO2, 16.30 Al2O3, 1.82 Fe2O3, 5.62FeO, 0.16 MnO, 9 MgO, 12.02 CaO, 2.28Na2O, 0.15 K2O, 0.03 P2O5, 0.11 H2O, 172Cr, 80 Ni, 47 Co, 243 V, 4 Zr and 9 Y .

The isotopic age of the gabbro-anorthosi-tes is 2450 Ma in the Kolvitsa massif(Mitrofanov et al. 1993) and 2452 Ma inthe Severopezhostrovsky massif (Alexejevet al. 2000). These ages show that their for-mation was as close as possible in time tothe formation of the lherzolite-gabbrono-rite complex. Based on this evidence, someauthors combine these units in one singledrusitic complex. However, the abovegroups of intrusion (complexes) differsubstantially in the composition of parentmelts and in relative geological ages. Dykesof the lherzolite-gabbronorite complex cutthe gabbronorite massifs. The latter aretypically strongly reworked. Their primarycontacts are unknown. Intrusive rocksusually pass gradually into banded garnetamphibolites. Lherzolite-gabbronorite bo-dies have crosscutting contacts with simi-lar amphibolites. Therefore, differentiatedgabbro-anorthosite massifs should bemore closely studied isotopically.

3. Anorthosites also occur in thin (severalmetres to tens of metres) lenticular bedsin banded amphibolites that are widespre-ad in the BMB and form thick horizons.Anorthosites are typically coarse grainedand foliated. Their boundaries with hos-ting amphibolites have no magmatic rela-tionships, suggesting that anorthosites arethe easiest recognizable fragment of a dif-ferentiated series in which the gabbroidportion is fully transformed into amphi-bolites. Small anorthosite bodies, rimmed

by alumina-rich feldspathic amphibolites,occur near Lake Skalnye, Lake Nigrozero,Kivguba and Point Kartesh in the WhiteSea and in other parts of the BMB. Thelargest massif of this type is the crescent-shaped Kotozero massif, with a visiblethickness of 1.5-2 km. Its axial part iscomposed of foliated meta-anorthositesand less common massive anorthositeswith relics of magmatic structure. Themost strongly altered anorthosite in theKotozero massif has the following compo-sition (wt.%, ppm): 50.90 SiO2, 0.25 TiO2,29.04 Al2O3, 0.75 Fe2O3, 1.32 FeO, 0.03MnO, 0.61 MgO, 12.76 CaO, 3.53 Na2O,0.21 K2O, 0.06 H2O, 0.42 LOI, 0.05 P2O5,15 Cr, 28 V, 32 Ni, 16 Co and 40 Cu. Onthe periphery, the foliated anorthosites arerimmed by a feldspathic amphiboliteband – presumably a mesocratic constitu-ent of the massif. Amphibolites consist of49.4 SiO2, 1.20 TiO2, 15.77 Al2O3, 3.42Fe2O3, 8.62 FeO, 0.20 MnO, 6.87 MgO,10.15 CaO, 2.15 Na2O, 0.14 K2O, 0.07H2O, 0.16 P2O5, 171 Cr, 111 Ni, 63 Co, 280V. Occurring in meta-anorthosites aredykes of fine-grained norites with thecomposition (wt.%, ppm) 50.86 SiO2, 0.56TiO2, 10.93 Al2O3, 1.37 Fe2O3, 9.08 FeO,0.17 MnO, 15.22 MgO, 0.83 CaO, 0.15Na2O, 0.42 K2O, H2O < 0.07, 0.91 LOI,0.09 P2O5, 1183 Cr, 190 V, 150 Ni, 63 Coand 80 Cu and garnet-diopside-plagiocla-se rocks with the composition 51.40 SiO2,1.90 TiO2, 12.75 Al2O3, 2.88 Fe2O3, 12.93FeO, 0.22 MnO, 4.35 MgO, 8.70 CaO, 1.46Na2O, 1.04 K2O, 0.10 H2O, 1.54 LOI, 0.28P2O5, 102 Cr, 302 V, 79 Ni and 120 Cu, andpossibly dykes of a complex composed oflherzolites, gabbronorites and garnet gab-bro. They are deformed together withanorthosites, but the degree of deformati-on is lower in the dykes than in the rocksof the massif.

The structural line traced by the Kotozeromassif and by the smaller gabbro-anor-thosite bodies that lie on its extension,seems to correspond to a dislocation zonewhich evolved in an extensional regime inLate Archaean and Early Proterozoictimes, and in a compressional regime inSvecofennian times. The close relationshipbetween these gabbro-anorthosites andamphibolites appears to reflect theirArchaean age. At the same time, the extentof amphibolitization is not a major crite-rion to distinguish them from the gabbro-anorthosite massifs described above (point

2). Based on this evidence (Stepanov 1981,Stepanov & Slabunov 1989), they were for-merly described as part of one complex.The isotopic age of this variety of gabbro-anorthosites has not yet been estimated.They are morphologically similar to auto-nomous anorthosites (massive type).

4. The formation of the lherzolite-gabbro-norite complex is commonly attributed toEarly Proterozoic rifting. The position ofgabbro-anorthosite magmatism in theBMB is more obscure. We (Stepanov &Slabunov 1989) assumed its formation asa manifestation of Late Archaean reactiva-tion (post-collisional stage of evolution).The available isotopic dates, indicating anEarly Proterozoic age for several massifstypical of the complex, have led us toconclude that some of the generally accep-ted views should be revised and that thecomplex should be investigated more tho-roughly.

5. Associated with the gabbro-anorthositemassifs of the BMB are the occurrences ofAu-Pt mineralization in two parageneses:1) in association with sulphide minerali-zation (Ileiki Islands, White Sea) and 2) in

association with ilmenite-magnetite miner-alization (Travyanaya Guba ore peridotites).

The study was supported by RFRF, Project 00-05-64295.

ReferencesAlexeev, N.L., Zinger, T.F., Belyatsky, B.V. &Balagansky, V.V. 2000: Age of crystallization and metamorphism of the Pezhostrov gab-bro-anorthosites, northern Karelia (Russia).5th Workshop, Lammi, Finland. Abstracts, 3.

Mitrofanov, F.P., Balagansky, V.V., Balashov, Yu.A.,Dokuchaeva, V.S., Gannibal, L.F., Nerovich, L.I.,Radchenko, M.K. & Ryunenen, G.I. 1995: U-Pbage of gabbro-anorthosite massifs in theLapland Granulite Belt. Norges geologiskeundersøkelse Special Publication 7, 179-183.

Stepanov, V.S. 1981: Precambrian mafic magma-tism of the Western White Sea region, NaukaPublishing House, Leningrad, 216 (in Russian).

Stepanov, V.S. & Slabunov, A.I. 1989:Amphibolites and Earlier Ultrabasites of thePrecambrian in Northern Karelia, Nauka Pub-lishing House, Leningrad, 175 pp. (in Russian).


The platinum belt of the Urals is compo-sed of zonal massifs belonging to the Ural-Alaskan type of the dunite-clinopyroxeni-te-gabbro rock series (420-430 Ma), withfamous platinum and magnetite depositsrelated to dunites and clinopyroxenites -hornblendites, respectively. The belt is ass-umed to be associated with a Late Ordo-vician-Silurian subduction zone dippingto the east. In the eastern part of the belta suite of hornblende leucogabbro, anor-thosite and plagiogranite dykes occurs insome zonal massifs together with a leu-cogabbro-anorthosite-plagiogranite, the

Chernoistochinsk massif, which is descri-bed below.

The root zone is composed of migmatizedhornblende gabbro and is exposed in thenorthern part of this massif. Migmati-zation (partial melting) has producednon-uniform rocks consisting of leuco-cratic patches full of gabbro relics and

Subduction-related leucogabbro-anorthosite ilmenite-bearing series: an example of water-rich high-temperature anatexis, platinum belt of the Urals, Russia

German. B. Fershtater

Institute of Geology and Geochemistry, Pochtovy per., 7. Ekaterinburg, 620151, Russia.

Table 1: Major (wt.%) and trace (ppm) elementcomposition of rocks from the Chernoistochinsk

massif.


N 1 2 3 4 5 6 7 8 9 10 11 12 13

Sample 571 578 566 565 562 117 124 115 567 568 566 562 117

SiO2 46,76 43,98 46,35 52,17 53,05 65,94 70,54 73,67 44,50 44,80 44,08 44,08 44,36

TiO2 0,79 2,23 1,09 0,70 0,66 0,20 0,14 0,13 1,25 1,28 1,28 1,55 0,99

Al2O3 12,98 15,85 17,14 22,90 23,16 18,59 15,98 14,19 8,27 7,92 10,48 10,46 8,70

Fe2O3 6,18 7,41 4,87 4,00 2,91 0,99 0,90 0,56 8,65 8,36 5,61 5,67 5,99

FeO 6,46 8,61 7,67 2,15 3,88 1,53 1,76 2,04 10,41 11,13 11,61 10,75 12,57

MnO 0,24 0,26 0,21 0,10 0,09 0,03 0,03 0,03 0,39 0,39 0,30 0,48 0,42

MgO 8,03 4,99 5,11 2,10 1,85 0,70 0,52 0,28 10,75 10,80 11,08 11,31 11,73

CaO 11,96 10,06 11,72 8,37 8,26 6,08 5,44 4,28 11,09 11,19 11,38 11,05 11,12

Na2O 3,08 2,69 3,10 6,16 4,97 4,78 3,73 3,78 1,39 1,39 1,52 1,24 1,22

K2O 0,12 0,12 0,12 0,10 0,14 0,08 0,12 0,08 0,16 0,18 0,20 0,20 0,25

P2O5 0,07 0,28 0,22 0,28 0,26 0,08 0,03 0,04 0,24 0,09 0,02 0,06 nd.

LOI 2,47 1,23 1,61 0,63 0,76 2,04 1,94 1,68 1,77 1,83 2,39 2,46 nd.

Li 2,44 2,20 1,83 2,07 2,37 0,00 0,00 0,00 2,18 1,31 13,52 0,00 1,74

Rb 0,63 0,35 0,00 0,00 0,00 0,13 0,11 0,04 0,38 2,00 1,91 0,09 0,31

Be 0,63 0,69 1,00 0,93 0,98 1,13 0,82 0,92 0,71 0,72 0,80 0,66 0,66

Sr 493 611 1366 1250 1515 725 768 763 166 94 128 139 93

Ba 64 71 110 119 126 88 114 103 48 66 128 65 56

Sc 67 62 9 10,3 9,6 8,6 5,4 4,6 50 57 58 65 64

V 362 475 145 179 155 66 52 26 326 406 468 451 500

Cr 72 1 3 2,2 9,2 2,6 1,7 82,9 729 704 34 69 22

Co 46 46 8 9,9 8,4 4,8 4,6 3,4 56 54 43 45 58

Ni 29 5 2 5,9 3,8 51,4 65,6 77,6 260 241 18 21 42

Cu 78,2 159,1 54,0 57 49 58 21 26 190 69 52 23 96

Zn 112,4 176,5 45,5 68,3 56,7 6,0 9,3 3,8 195 198 253 259 219

Ga 16,1 22,3 23,2 24,0 24,5 14,8 12,4 10,6 19 20 26 25 21

Y 21,8 25,2 6,3 9,0 7,7 1,5 1,3 0,8 42,6 47,7 38,3 48,1 36,2

Nb 1,51 4,18 0,95 1,20 1,36 0,21 0,23 0,21 5,40 3,26 4,23 2,59 4,33

Ta 0,11 0,28 0,07 0,08 0,09 0,28 0,36 0,26 0,69 0,19 0,05 0,00 1,71

Zr 20,06 17,12 6,11 6,21 7,09 1,56 0,69 0,30 42,5 38,1 57,5 40,6 27,0

Hf 0,70 0,71 0,27 0,27 0,23 0,14 0,07 0,04 1,56 1,58 2,40 1,68 1,26

Pb 3,18 7,90 3,46 4,29 4,01 0,00 0,23 0,00 4,14 4,14 10,04 4,26 10,55

U 0,02 0,03 0,00 0,00 0,04 0,00 0,00 0,00 0,06 0,10 0,49 0,00 0,03

Th 0,05 0,06 0,03 0,03 0,07 0,00 0,00 0,00 0,00 0,96 4,21 1,42 0,09

La 4,33 5,98 2,74 3,76 3,21 0,72 1,24 0,62 6,70 6,79 11,18 8,00 3,97

Ce 12,69 16,66 6,61 9,80 8,95 1,71 2,25 1,15 27,47 26,78 33,57 32,22 16,38

Pr 2,14 2,76 1,04 1,59 1,49 0,23 0,23 0,15 5,42 5,31 5,73 6,85 3,56

Nd 10,31 13,29 5,01 7,53 7,05 1,16 1,13 0,67 27,08 26,54 30,22 40,49 19,41

Sm 2,81 3,73 1,23 1,90 1,78 0,33 0,32 0,14 7,26 7,39 8,70 11,38 5,90

Eu 1,10 1,59 0,73 0,94 0,89 0,23 0,24 0,14 2,56 2,43 2,53 2,91 1,84

Gd 3,08 4,07 1,30 1,90 1,72 0,36 0,25 0,15 7,29 7,61 6,96 7,69 6,03

Tb 0,52 0,65 0,19 0,27 0,25 0,05 0,04 0,02 1,13 1,23 1,03 1,35 0,96

Dy 3,60 4,40 1,11 1,62 1,31 0,30 0,25 0,15 7,51 8,08 6,89 8,66 6,45

Ho 0,79 0,98 0,23 0,34 0,29 0,07 0,05 0,03 1,57 1,77 1,46 1,79 1,36

Er 2,14 2,62 0,62 0,88 0,79 0,16 0,10 0,08 4,34 4,78 3,76 4,81 3,63

Tm 0,33 0,39 0,09 0,13 0,11 0,02 0,01 0,01 0,64 0,72 0,54 0,69 0,53

Yb 2,19 2,49 0,53 0,79 0,67 0,15 0,13 0,08 4,06 4,81 3,45 4,21 3,44

Lu 0,31 0,35 0,08 0,11 0,09 0,02 0,02 0,01 0,56 0,66 0,49 0,60 0,49

1,2 - hornblende gabbro; 3 - partially melted gabbro; 4 - leucogabbro; 5 - anorthosite, 6 - quartzanorthosite; 7, 8 - plagiogranites; 9, 10 - hornblendites; 11-13 - hornblende from gabbro (11), anorthosite (12) and plagiogranite (13).


hornblende segregati-ons. Direct field inve-stigations show that leu-cocratic segregationsforming the hornblendeleucogabbro representthe initial anatectic melt.Crystallization of thismelt produced relativelyhomogeneous veins andsmall intrusive bodies inpartially melted horn-blende gabbros. Crystalfractionation of this leu-cogabbro melt gave riseto a leucogabbro-anor-

thosite-plagiogranite rock series mainlyforming veins with gabbro xenoliths (Fig.1). The hornblendite inclusions and schli-eren represent the restitic phases. All therocks have the same mineralogy, which isas follows: hornblende + plagioclase +magnetite + ilmenite. Magnetite is the

main opaque mineral ingabbro, while ilmenite ispredominant in anor-thosite. Quartz appearsin plagiogranite. Thevolume of these rocks isvery small, not morethan 3-5%. The amountof hornblende and theAn contents in plagio-clase decrease from gab-bro to granites from 50to 3-5 % and from 45-50to 25-30%, respectively.The hornblende con-tents are the best visibleindex of partial melting

and the degree of crystal fractionation, asshown in Fig. 1. The range of chemicalcompositions of the rock series (Table 1)represents the products of partial meltingreactions from protolith (anal. 1 - 3) toinitial melt (anal. 4) and products of itsfractional crystallization (anal. 5-8) to res-titic phases (anal. 9,10). Partial meltingproduces rocks impoverished in all theelements which are concentrated in femicminerals, such as Mg, Fe, Ti, and in mostrare elements, with the exception of Sr, Baand some others related to plagioclase. Themagnitude of the positive Eu anomaly cle-arly increases in the range from gabbro toplagiogranite. At the same time, the ilme-nite/magnetite ratio increases from 0.5 ingabbro to 10 in anorthosite.

The conditions of anatexis are shown inFig. 2. This took place in the field of horn-blende stability, while plagioclase with ananorthite content of less than An60 hasundergone melting (shaded area in Fig. 2).Mass-balance calculations show thatmajor- and trace-element compositions ofleucogabbro correspond to approximately60% of partial melting of hornblende gab-bro. It is necessary to melt about 10%hornblende and nearly 100% plagioclaseto produce the leucogabbro composition.A large body of experimental and geologi-cal work confirms the well-known factthat the usual quartz-bearing compositionof anatectic melts has resulted from parti-al melting of amphibolite. The mechanismdiscussed above can occur when high tem-peratures (maybe because of a short timeinterval between gabbro intrusion andanatectic events) are combined with highconcentrations of water coming from thesubducted slab. It is believed that suchquartz-absent anatexis was common inthe early stages of the Earth's evolutionand that it can be the source of some auto-nomous anorthosites.

ReferencesFershtater, G.B., Bea, F., Borodina, N.S., &Montero, P. 1998: Anatexis of basites in paleo-subduction zone and the origin of anorthosi-te-plagiogranite series of the Ural Platinum-bearing belt. Geochemistry International 36,684-697.

Rushmer, T. 1991: Partial melting of two amphi-bolites: contrasting experimental results underfluid absent conditions. Contributions toMineralogy & Petrology 107, 41-59.

Yoder, H.S. & Tilley, C.E. 1962: Origin of basaltmagmas: an experimental study of natural andsynthetic rock systems. Journal of Petrology 3,342-532.

Fig. 2. P-T conditions of anatexis. Wet soli-dus and wet liquidus as well as An60 liqui-dus are represented for high-Al basalt(Yoder and Tilley, 1962), which is close incomposition to the hornblende gabbro ofthe Chernoistochinsk massif. Amphibolestability lines are after Yoder & Tilley (1962)and Rushmer (1991). Dashed area repre-sents the assumed field of anatexis.

Fig. 1. Detail from a outcrop of migmatite;migmatized gabbro with anorthosite vein(light) and hornblendite restite (black). Thenumbers indicate the percentage of horn-blende in the rock.


Massif SiO2 TiO2 Al2O3 Fe2O3 FeO MnO MgO CaO Na2O K2O P2O5 LOI Total

Matkal (3) 46.91 2.19 14.89 3.89 8.57 0.11 6.66 11.21 3.00 0.42 0.06 1.58 99.49

Kopan (29) 44.50 2.90 15.47 5.72 10.25 0.09 5.66 11.4 2.36 0.34 0.14 0.57 99.64

Medvedevka (11) 44.81 2.26 14.69 5.25 11.76 0.17 7.76 9.34 2.40 0.50 0.07 0.89 99.90

Kusa (21) 43.90 3.14 13.24 5.61 11.17 0.19 7.07 11.54 1.85 0.22 0.23 1.41 99.57

Kusa, ilmenite gabbro-norite (11) 42.20 6.07 13.11 5.01 13.29 0.24 6.13 9.49 2.14 0.33 0.16 1.87 100.04

Kusa, gabbro-amphibolite (19) 46.44 2.16 16.26 4.57 9.16 0.15 5.53 10.66 2.96 0.43 0.06 1.74 100.12

Element Ilmenite type Titanomagnetite type

South group North group South group North group

Fe 14-15 14-15 22-23 23-24

TiO2 6-7 7 7 6-7

V2O5 0.12 0.12 0.25 0.25

Table 1. Average chemical compositions of the gabbros

Table 2. Average chemical compositions of the ores (wt.%)

The Riphean layered gabbro intrusions arelocalized inside the graben facies of sedi-mentary and volcanic rocks of the paleo-continental sector of the South Urals. Vol-canic rocks overlie the gabbro intrusionsand are represented mainly by tholeiiticbasaltand by dykes and small intrusions ofrhyolite. The intrusions are underlain bysedimentary carbonate rocks and containxenoliths of both sediments and volcanics.The Rb-Sr age of the volcanic rocks cor-responds to 1346±41 Ma, and the initialratio 87Sr/86Sr is 0.7098±0.001 (Krasno-baev & Semikhatov 1986). The gabbroshave yielded an age of 1300 ±42 Ma andthe initial ratio 87Sr/86Sr is 0.7093±0.0067(Gorozhanin 1998).

Gabbro intrusions form a 70 km-long beltconsisting of 4 isolated bodies dipping at60° to the east. In the two southern massifs(Matkal and Kopan), gabbros are repre-sented by the two-pyroxene variety (gab-bro-norites) and contain high-Ti titanom-agnetite ore bodies up to 10-15 km longand 1-2 m thick, parallel to the gabbro lay-ering. These massive ores are surroundedby an aureole of impregnation. The nor-thern massifs (Medvedevka and Kusa)

consist of hornblende gabbros and gab-bro-amphibolites in the upper part andilmenite gabbro-norites in the lower part(Table 1).

The ore bodies are concentrated in thegabbro-amphibolites and consist of ilme-nite or low-Ti titanomagnetite+ilmenite.Ilmenite ores mainly belong to the imp-regnation type, while Ti-magnetite+ilme-nite ores form both impregnation andmassive types (Table 2).

The first ore type is situated mainly in theupper part of the massifs, the second typein the central part. Most Ural geologistsexplain the clear differences in compositi-on of gabbros and related ores betweenthe southern and northern group of mas-sifs by the metamorphic overprint whichhas affected the northern group.Our data, however, suggest that these dif-ferences are due to different P-T conditi-ons of formation. Rocks and ores of thesouthern group were formed at a depth of4-6 km (Ptot = 1-2 kb, PH2O=0.5Ptot,T°C=900°-1100°, fluid enriched in F),while in the northern group these para-meters are as follows: depth of 20-25 km,

The Riphean layered intrusions of the Western Uralsand related ilmenite-titanomagnetite deposits

German B. Fershtater & V.V. Kholodnov

Institute of Geology and Geochemistry, Pochtovy per., 7. Ekaterinburg, 620151, Russia.


Ptot = 8-4 kb, PH2O=0.7-0.9Ptot, T°C=850°-650°, fluid enriched in Cl (Figs. 1,2). Fluidenrichment in the massifs of the northerngroup is responsible for hornblende gab-bro crystallization and its postmagmatictransformation into gabbro-amphibolites.Low temperature has caused the separatecrystallization of ilmenite and low-Tititanomagnetite, that has led to the forma-tion of the different are deposits.

The upper part of the Kusa deposit wasmined before the 1990s. Proven reserves ofboth ore types in the Matkal, Kopan andMedvedevka deposits range up to 7 billiontons.

ReferencesBuddington, A.F. & Lindsley, D.H. 1964: Iron-tita-nium oxide minerals and synthetic equiva-lents. Journal of Petrology 5, 310-357.

Fershtater, G.B. 1990: An empirical plagioclase-hornblende barometer. Geokhimia. 3 (inRussian).

Gorozhanin, V.M. 1998: Initial isotope composi-tion of strontium in magmatic complexes ofthe South Urals. In: V.A. Koroteev (ed.)Magmatism and Geodynamics. Ekaterinburg.98-108 (in Russian).

Krasnobaev, A.A. & Semikhatov, M.A. 1986:Geochronology of Upper Proterozoic of theUSSR. In: M. Nauka. Methods of isotope geologyand geochronological chart. 159-183 (in Russian).

Fig. 1. Plot of Al/Si in hornblende versus Al/Si in plagioclase for determina-tion of PH2O conditions of plagioclase-hornblende equilibrium. Numbers incircles are the pressure in kb (after Fershtater 1990).

Fig. 2. Recalculated Buddington-Lindsley thermometer. 1- 4 symbols markthe magnetite-ilmenite pairs from the same massifs as in Fig. 1. Solid linesrepresent the temperature (°C) and dashed lines mark log of fO2 (bars).


PreambleThe geological map (Plate 1) has beencompiled from available sources (see liston geological map) supplemented withmapping and field checks made by theGeological Survey of Norway during thesummers 1997-2002. Key contributionsfor the map are provided by Michot(1960), Michot (1961), Michot & Michot(1969) and Rietmeijer (1979). For the ÅnaSira anorthosite massif the mapping byKrause et al. (1985) has been used. Theirwork was compiled and digitised byKarlsen et al. (1998). Robins and Wilson(unpublished material) provided data forthe Bjerkreim-Sokndal Layered Intrusion,Bolle (1998, unpublished thesis) for theApophysis of the Bjerkreim-Sokndalintrusion, Maquil (1980, unpublishedmap reproduced in Duchesne & Maquil(1987)) for the Egersund-Ogna anorthosi-te massif, and Michot & Michot (1969)and Demaiffe et al. (1973) for the Hidraanorthosite massif. Large parts of theHåland-Helleren anorthosite complex,originally investigated by Michot (1961),have been re-mapped (M. Marker, R.B.Edland) as have the outer zone of theEgersund-Ogna anorthosite massif andthe surrounding gneisses up to the contact(M. Marker). The area of jotunitic tomangeritic intrusive rocks in the western-most part of the province northwest ofBjårvatnet has been re-investigated (H.Schiellerup), and the areas immediatelysouth of the Bjerkreim-Sokndal LayeredIntrusion have also been re-mapped (G.Meyer). The host-rock Sveconorwegiangneisses have been covered by new map-ping between the Apophysis and Lunde-vatnet (south of Heskestad) in the easternpart of the area (M. Marker). The mapdata for the Sveconorwegian gneissesnorth of the Bjerkreim-Sokndal Layered

Intrusion have been adopted from Her-mans et al. (1975). All the re-mapping andnumerous field checks throughout theprovince also aimed to ensure a precisetransformation of data from an old to amodern topographic basis.

Subdivision of the Rogaland Anorthosite ProvinceThe Rogaland Anorthosite Province com-prises a number of massif-type anorthosi-te bodies, the Bjerkreim-Sokndal LayeredIntrusion and a range of other noritic andjotunitic to charnockitic intrusions. Allunits were emplaced into Sveconorwegiangneisses that were deformed under granu-lite-facies conditions. The subdivision intorock bodies (see inset map on the geologi-cal map) generally follows the previouslyestablished subdivision of the province.An exception is the Håland-Helleren mas-sif, which has traditionally been treated asa single anorthosite unit. This massif hasnow been separated into two bodies, theHåland and the Helleren anorthosites,which show cross-cutting relationshipsand contrasting deformation features.

Based on intrusive relationships andstructural features, the following units andrelative chronology can be distinguishedin the Rogaland Anorthosite Province.The Egersund-Ogna Anorthosite in thewest forms the oldest body, with a foliatedmarginal zone that conforms to the laye-ring in the surrounding Sveconorwegiangneisses. The body is separated from thesimilarly foliated Håland Anorthosite by aseptum of highly strained, well-banded,Sveconorwegian gneisses. The relativelyyounger and undeformed Helleren Anor-thosite in the central part of the provinceintrudes both the Egersund-Ogna and the

Introduction to the geological map of the RogalandAnorthosite Province 1:75,000

Mogens Marker, Henrik Schiellerup, Gurli Meyer, Brian Robins & Oliver Bolle

Rogaland AnorthositeProvince – Syntheses

Håland bodies. The Helleren Anorthositebody may form the continuation of thelikewise undeformed Åna-Sira Anorthositein the eastern part of the province with alobe of younger magmatic rocks (theBjerkreim-Sokndal Layered Intrusion andjotunitic rocks) in between. The Bjerk-reim-Sokndal Layered Intrusion is lying inan irregular trough structure on top ofboth the anorthosite bodies and theSveconorwegian gneisses in the northeas-tern part of the province. The intrusion iscapped by mangerite and quartz mangeri-te, which extends southeastwards as theso-called Apophysis Intrusion along theeastern margin of the Åna-Sira Anortho-site body. Late-magmatic jotunitic tomangeritic rocks occur as dykes and bodiesin most of the Rogaland AnorthositeProvince. The largest of these are thebranched Eia-Rekefjord jotunite intrusionin the trough between the Helleren andÅna-Sira Anorthosite bodies, and theHadland jotunitic intrusion at the westernmargin of the Egersund-Ogna Anor-thosite body. Another group consists ofnoritic intrusions that form smaller bodiesin the southeastern half of the province(map legend 16-19). Important amongthese is the Tellnes body in the Åna-SiraAnorthosite, which contains a world-classilmenite deposit. Two smaller bodies ofanorthosite and a granite body in the ea-stern part of the Rogaland AnorthositeProvince lie isolated within the Sveconor-wegian gneisses, and their ages relative tothe other magmatic units, and to eachother, thus cannot be determined fromfield relations alone. These bodies are theGarsaknatt Anorthosite east of Lunde-vatnet and the Hidra Anorthosite andFarsund Charnockite in the southeastern-most part of the province. The RogalandAnorthosite Province was long after itsconsolidation transected by WNW-ESE-trending dykes. The 616 Ma old doleritedykes form the Egersund dyke swarm, andwere injected during the pre-Caledonianrifting just prior to the initial opening ofthe Iapetus Ocean (Bingen et al. 1998).

Though a sequential evolution can bedemonstrated for most of the magmaticunits in the Rogaland Anorthosite Prov-ince from the field relationships, agedeterminations have shown that the dura-tion of magmatism in the province wasabout 10 million years, which is surpri-singly short. U-Pb dating of zircon and

baddeleyite extracted from orthopyroxenemegacrysts (Schärer et al. 1996) yieldedemplacement ages of 929± 2, 932 ± 3 and932 ± 3 Ma for the Egersund-Ogna,Helleren and Åna-Sira Anorthosites,respectively, while late-magmatic injecti-on of jotunitic dykes occurred at 931±5Ma and the intrusion of the Tellnes noriticbody at 920 ± 3 Ma.

The Egersund-Ogna AnorthositeThe Egersund-Ogna Anorthosite body isthe largest in the Rogaland AnorthositeProvince with a more or less dome-sha-ped, concentric, lithological structure(Maquil 1980 (unpublished map reprodu-ced in Duchesne & Maquil (1987)), newmapping by NGU). The outer part of thebody shows a variably developed foliationwhich follows the concentric structure,and which is concordant with the foliationin the Sveconorwegian gneiss envelope. Inthe south, this gneiss envelope continuesas a thin septum of highly strained, shea-red gneisses that separate the Egersund-Ogna Anorthosite from the Håland Anor-thosite. The anorthosites at both sides ofthe septum show a penetrative planar foli-ation that is best developed on the nor-thern side. The foliated marginal zone ofthe body has been interpreted as havingformed during diapiric rise of the anor-thosite as a crystal mush from mid- andlower crustal levels (Barnichon et al. 1999).

The central part of the Egersund-OgnaAnorthosite body consists of medium-grained homogeneous anorthosite that, inits centre northwest of Hellvik, containsphenocrysts and megacrysts of orthopy-roxene and plagioclase. The central part ofthe body is also the part where labradori-sing plagioclase, both in matrix and phe-nocrysts, occurs most extensively in theprovince (Heldal & Lund 1995). Outwardsfollows a 1-4 km-wide zone of medium-grained leuconorite, which again is follo-wed by homogeneous anorthosite thatlocally may contain scattered orthopyro-xene and smaller, <5-10 cm-long, com-monly rounded phenocrysts of plagiocla-se. On Nordre Eigerøya (in the west), thiszone includes an area that is intruded byrather coarse-grained leuconorite formingan agmatitic structure with, on average,20% leuconorite. Then follows a 0.5-1.2km-wide, strongly foliated zone of anor-thosite and leuconorite with 2-10 cm-



long, drawn out streaks and 10-30 cm-long, fish-like phenocrysts of orthopyro-xene forming a well-developed planarstructure. In places, trails of elongated lar-ger orthopyroxene pheneocrysts alsooccur. The foliated zone has, in the north,sharp boundaries to little strained anor-thositic units on each side, but transgres-ses to the margin of the Egersund-Ognabody in the eastern and southern parts.The two zones, which occur in the nor-thern marginal part of the body, wedgeout west of Helleland in the east andnorthwest of Bjårvatnet in the west.Immediately north of the foliated zonethere is a zone of quite massive anorthosi-te with a varying density of generally <30-40 cm-long, orthopyroxene megacrysts ofirregular shape. In places the massif anor-thosite contains spectacular orthopyro-xene megacrysts or clusters of megacryststhat are up to 1-2 m in length. The outer-most zone in the north consists of homo-geneous, medium-grained leuconoriticanorthosite with 5-10% of evenly distri-buted orthopyroxene. Locally, it containsscattered, <5 cm-long, rounded pheno-cryst of plagioclase. As in the orthopyro-xene megacrystic zone, a foliated structureis weakly developed or absent, except in anarrow zone at the western margin, 3-8km north of Bjårvatnet.

A jotunitic intrusion, the Hadland Intru-sion, is located in the northwesternmostpart of the Rogaland Anorthosite Provincealong the contact between Sveconor-wegian granulite-facies gneisses and theEgersund-Ogna Anorthosite body. Theintrusion consists essentially of jotuniteand is mostly modally layered with a laye-ring striking parallel to the contacts anddipping to the northwest.

The Håland AnorthositeThe Håland Anorthosite body lies to thesouth of the Egersund-Ogna Anorthosite,and is separated from this body by astrongly foliated contact zone that dipssteeply to the north. The contact zoneincludes a septum of similarly highly strai-ned Sveconorwegian gneisses, which sepa-rates the two bodies. In the east, theHåland Anorthosite is intruded and cut bythe undeformed Helleren Anorthosite.The Håland Anorthosite is composed ofanorthosite and leuconorite with compli-cated relationships between the different

rock phases. These have, in addition, suf-fered varying degrees of deformation, thusmaking the relationships even more com-plex.

The Håland Anorthosite body can basical-ly be divided into two units. The northernunit is dominated by anorthosite withpseudo-enclaves of leuconorite (Michot1961), which show varying degrees ofdeformation. The least strained, leucono-rite pseudo-enclaves, as at Eigerøya, are10-40 cm long and bun-shaped with ellip-tical cross sections. With increasing defor-mation, the pseudo-enclaves are stretchedand thinned and eventually form a ban-ded anorthositic rock with, discontinu-ous, dark orthopyroxene-bearing bands,up to some centimetres thick, which isoften seen in the eastern part of the body.Along the northern margin of the body,immediately south of the gneiss septumfrom Koldal to Kydlandsvatnet, there is a100 m-thick zone of seemingly undefor-med leuconorite, which appears to hostmost of the ilmenite ore deposits, whi-chare found in this part of the province.The leuconorite zone widens in the east tobecome 300-400 m thick north of Kyd-landsvatnet.

The southern unit of the Håland Anor-thosite consists of interlayered anorthositeand leuconorite with enclaves of phasesshowing modal-like layering. The domi-nant rock type is a coarsely spotted leuco-norite with 2-4 cm-long, elongated ortho-pyroxene clusters that define a distinct,planar, foliated structure. The spotted leu-conorite has local, 0.5-2 m-thick, foliati-on-parallel layers of anorthosite in additi-on to plentiful diffuse veins of anorthosite(several generations?), which seem tointrude all other phases. Quite commonlythere are fish-like enclaves, up to severalmetres in length, of a modally bandedphase with close-lying mafic layers, up to afew cm thick. This phase looks much likethe high-strain banded anorthositic rockin the northern unit of the body and mayrepresent inclusions of this unit in thespotted leuconorite. The foliated structureof the spotted leuconorite is folded intotight folds in the eastern part of the body,where it is cut by the undeformedHelleren Anorthosite. It thus appears like-ly that the southern part of the HålandAnorthosite was, for the most part, intru-ded and deformed after the emplacement

and deformation of the northern unit.

The Helleren AnorthositeThe Helleren Anorthosite body cuts acrossthe boundaries and internal deformationstructure of both the Egersund-Ogna andthe Håland Anorthosite bodies. It hasclear intrusive contacts which, in thenorth and east, are transected by theBjerkreim-Sokndal Layered Intrusion.The Helleren Anorthosite consists of mas-sive, rather coarse- to medium-grainedanorthosite and leuconorite, which showlittle or no signs of structural deformati-on. Leuconorite occurs first of all in a 1-2km-wide zone at the western margin ofthe body, but patches of leuconorite alsooccur in the anorthosite farther east. Theleuconorite is usually medium-grainedwith evenly dispersed ortopyroxene (spot-ted). Finer-grained varieties occur in thewest-facing embayment into the HålandAnorthosite (Michot 1961). The anortho-site may locally contain phenocrysts ofplagioclase while orthopyroxene pheno-crysts are relatively rare.

The Åna-Sira AnorthositeThe Åna-Sira Anorthosite is a fairlyhomogeneous anorthosite massif whichconsists essentially of anorthosite andanorthositic leuconorite (Krause et al.1985). These rock types are irregularlyinterspersed and have not been differenti-ated on the map. The composition andtexture of the anorthosite appear to besimilar to those of the Helleren Anor-thosite. Both the Åna-Sira and the Hel-leren anorthosites have been dated to 932± 3 Ma by U-Pb isotopes (Schärer et al.1996). Plagioclase chemistry reveals aNW-SE elongated pattern with the mostprimitive compositions in the central partof the anorthosite body (Zeino-Mahmalat& Krause 1976). To some extent, a similarNW-SE trending pattern is displayed bymineral lamination and sporadic modallayering (Knorn & Krause 1977). TheÅna-Sira Anorthosite is the most signifi-cant anorthosite massif in Rogaland interms of Fe-Ti oxide ore deposits andhosts two giant ilmenite ore bodies, theTellnes and Storgangen deposits, as well asa number of smaller deposits. The anor-thosite is cross-cut by a several km-longnorite pegmatite (the Blåfjell-Laksedalpegmatite) which is associated with a suite

of coarse-grained Fe-Ti oxide mineralisa-tions. The Hogstad Intrusion is located inthe southeastern part of the Åna-SiraAnorthosite and consists of an eastern leu-conoritic portion and a western noriticportion. It displays sub-vertical layeringand a cumulate zoning equivalent to partsof the Bjerkreim-Sokndal Layered Intru-sion (Vander Auwera & Duchesne 1996).The Åna-Sira Anorthosite also contains asmall, layered noritic body (the BøstølenIntrusion) as well as a number of dykes,ranging in composition from jotunite tonorite and mangerite. Megacrystic Ca-poor pyroxenes occur only sporadically inthe Åna-Sira anorthosite.

The Bjerkreim-Sokndal LayeredIntrusionThe Bjerkreim-Sokndal Layered Intrusionconsists of a layered cumulate sequenceranging in composition from anorthositeand troctolite to norite, gabbronorite, jotu-nite and mangerite. The intrusion com-prises a Layered Series, more than 7000 mthick, suggested to be derived by fractionalcrystallization of jotunitic parental mag-mas (Duchesne & Hertogen 1988, Robinset al. 1997). It is capped by mangeritic,quartz-mangeritic and charnockitic rocks.The Layered Series may be divided into 6megacyclic units (MCUs), each resultingfrom magma replenishment episodes, andwithin each MCU a characteristic cumula-te sequence is partly or fully repeated. Withdifferentiation of the magma, the cumulusminerals crystallized in the general order:plagioclase - ilmenite (or plagioclase + oli-vine + ilmenite + magnetite) - Ca-poorpyroxene (olivine and magnetite out) -magnetite - Ca-rich pyroxene and apatite -fayalitic olivine, but with minor variations(Wilson et al. 1996). On the map the cum-ulate nomenclature of Wilson et al. (1996)has been maintained in order to provide asdetailed an account as possible of theintrusion petrology.

The intrusion is traditionally divided intothree exposed lobes; the Bjerkreim,Sokndal and Mydland lobes (Michot1960), although the Mydland lobe hasbeen shown to be an enclave within thejotunitic-mangeritic Apophysis Intrusion(Bolle et al. 1997). The other lobes havebeen deformed into a roughly N-S elonga-ted syncline or trough, due to gravitatio-nal subsidence of the core of the intrusion



at a late magmatic stage (Paludan et al.1994, Bolle et al. 2002). The cumulatesequence is thickest along the south-plunging axis of the northern Bjerkreimlobe, where the most complete sequenceof MCUs is found. The acidic rocks, whichcap the Layered Series, comprise quartzmangerites and charnockites (Duchesne &Wilmart 1997), which seem to be contin-uous with the Apophysis Intrusion(Rietmeijer 1979, Bolle et al. 1997) exten-ding along the northern and eastern con-tacts of the Åna-Sira Anorthosite.

The Apophysis IntrusionThe Apophysis Intrusion (Bolle 1996)extends from the quartz-mangeritic cap ofthe Bjerkreim-Sokndal Layered Intrusiontowards the southeast along the contactbetween Sveconorwegian gneisses and theÅna-Sira Anorthosite. The intrusion iscomposed of diverse, intermingled, jotu-nitic to quartz-mangeritic rock types.

The Hidra and Garsaknatt anorthositic-leuconoritic intrusionsHidra and Garsaknatt are two smaller,dominantly anorthositic to leuconoriticbodies, intruded into the gneissic envelope(Michot & Michot 1969). The Hidra bodycontains a marginal chill zone that provi-des important evidence for an origin byfractional crystallization of a jotuniticparental magma (Demaiffe & Hertogen1981). In addition, the body is cross-cutby a stockwork of charnockitic dykes(Weis & Demaiffe 1983). The Garsaknattbody shows clear intrusive relationshipswith the host Sveconorwegian gneissesand comprises anorthosite and leuconori-te as well as segments of modally layerednorite. Both the Garsaknatt and the Hidrabodies contain resources of iridescent pla-gioclase-bearing anorthosite and arepotential candidates for dimension stoneextraction (Heldal & Lund 1995).

The Eia-Rekefjord Intrusion andjotunite dykesThe Eia-Rekefjord Intrusion (Michot1960) is the most prominent jotuniticintrusive body in the province. It is foundprimarily along the contact between thesouthern lobe of the Bjerkreim-SokndalLayered Intrusion (the Sokndal lobe) andthe neighbouring anorthosite massifs, but

also transects the quartz mangerites cap-ping the intrusion. The Eia-RekefjordIntrusion intruded the Sokndal lobe cum-ulates and the Åna-Sira Anorthosite, andrelated jotunitic rocks are most probablyfound as a marginal intrusion in theMydland lobe. The Eia-Rekefjord Intru-sion generally consists of rather monoto-nous medium grained jotunite.

Other jotunitic intrusive rocks are presentthroughout the province as dykes cross-cutting all the major anorthosites (Michot1960). The most significant jotunite dykeis the Lomland dyke, which originates in anorite body in the Håland Anorthosite,cuts the Håland and Egersund-Ognaanorthosites and forms a complex branc-hing pattern upon entering the Bjerkreimlobe of the Bjerkreim-Sokndal LayeredIntrusion. The compositions of most ofthe jotunite intrusive rocks are controlledby a well defined liquid line of descent,giving rise to a continuous compositionalrange from primitive jotunites to quartzmangerites (Vander Auwera et al. 1998).

The Farsund CharnockiteThe Farsund Charnockite is one of theacidic intrusions associated with theRogaland Anorthosite Province. It is adark olive-green rock dated by U-Pb iso-topes to between 900 and 940 Ma (Falkum& Petersen 1987), with clear potassic affi-nities and characterised by distinctly hig-her Fe/Mg ratios than the hornblende-biotite granitoids elsewhere in SouthNorway.

Sveconorwegian gneissesThe granulite-facies Sveconorwegiangneiss complex in Southwest Rogaland(Hermans et al. 1975) constitutes the hostrocks of the Rogaland Anorthosite Pro-vince and is exposed in the northeast ofthe province and in a septum between theEgersund-Ogna and Håland anorthosites.Individual gneiss units generally showlarge lateral continuity which may be part-ly a primary intrusive feature but is alsodue, not least, is due to attenuation duringthe deformation. The gneisses are predo-minantly of igneous origin and can bebroadly divided into two groups. Theolder group consists of granodioritic togranitic orthogneisses showing partial-melt veins while the younger group con-

sists of various types of granites withoutanatectic veining. Both are deformed andfoliated. The gneisses of the older group inparticular show strong deformation featu-res including, quite commonly, a markedplanar fabric in which primary relationsbetween the different rock componentsare more or less obliterated and melt veinshave formed parallel to the foliation.These well-foliated gneisses are usuallyreferred to as banded gneisses despite oftenshowing a banding composed only ofstretched, parallel melt veins. In the maparea, the banded gneisses include a fewminor layers or bands of highly anatecti-cally melted, garnet-sillimanite-bearingpelite and, less frequently, orthopyroxene-bearing amphibolite. The pelitic gneisscommonly contains ossumilite that for-med during contact metamorphism. Thinquartzite bands are only important in oneplace, near Gjersdal, in the southeasternpart of the area. The last phase of the ban-ded gneisses is a light grey metagranitewhich occurs as deformed, almost confor-mable veins or bands of varying density.The rocks and high degree of strain in thegneiss septum at the Egersund-Ogna-Håland Anorthosite boundary are verysimilar to those in the banded gneisses.

The younger group of granitic gneisseswithout partial-melt veins compriseseven-grained and porphyritic types. Moreor less foliated, medium-grained graniticgneiss is the most common even-grainedtype. A rather coarse-grained metagraniterich in white pegmatites occurs near Hovs-vatnet. The porphyritic types are weaklyto well foliated. Most common, and parti-cularly westwards from Hovsvatnet, is aporhyritic metagranite with a varying den-sity of deformed K-feldspar phenocrysts.West of Lundevatnet occurs a characteris-tic coarse-porphyritic metagranite belong-ing to the Feda suite (Bingen & van Bree-men 1998a, 1998b). Both types of porphy-ritic granite contain variable amounts ofscattered, deformed veins of metagranite.

The degree of deformation is generallyhigh in the Sveconorwegian gneisses withthe highest strains recorded in the bandedgneisses, which may have evolved in a duc-tile shear regime at a deep crustal level. Asis common in granulite-facies terrains,lineations are poorly developed, but whenobserved they mostly plunge gently to theeast or west. An exception is the highly

strained gneisses at the contact to theApophysis Intrusion where lineations aresub-vertical and quite common. Also, themangeritic rocks at the eastern margin ofthe Apophysis Intrusion show a distinctfoliation.

The age of the Sveconorwegian gneisses isnot fully resolved, particularly for the ban-ded gneisses. The intrusion of the porphy-ritic granite of the Feda suite has beendated at 1051 Ma (Bingen & van Breemen1998a), while granitic gneisses give mag-matic ages around 1035-1056 Ma (Mölleret al. 2002). The age of peak metamor-phism, as determined from U-Pb monazi-te dating (Bingen & van Breemen 1998b),is around 1025-970 Ma, and thus at least40 million years older than the intrusionof the anorthosite plutons. Möller et al.(2002) obtained U-Pb ages for metamor-phic zircons at c. 1015-1000 to c. 910-930Ma in granitic gneisses north of Tekse-vatnet.

ReferencesBarnichon, J.-D., Havenith, H., Hoffer, B., Charlier,B., Jongmans, D. & Duchesne, J.C. 1999: Thedeformation of the Egersund Ogna massif,South Norway: Finite element modelling ofdiapirism. Tectonophysics 303, 109-130.

Bingen, B., Demaiffe, D. & van Breemen, O. 1998:The 616 Ma old Egersund Basaltic dike swarm,SW Norway, and Late Neoproterozoic openingof the Iapetus Ocean. Journal of Geology 106,565-574.

Bingen, B. & van Breemen, O. 1998a: Tectonicregimes and terrane boundaries in the high-grade Sveconorwegian belt of SW Norway,inferred from U-Pb zircon geochronology andgeochemical signature of augen gneiss suites.Journal of the Geological Society, London 155,143-154.

Bingen, B. & van Breemen, O. 1998b: U-Pbmonazite ages in amphibolite- to granulite-facies orthogneiss reflect hydrous mineral bre-akdown reactions: Sveconorwegian Provinceof SW Norway. Contributions to Mineralogy and Petrology 132, 336-353.

Bolle, O. 1996: L'apophyse de massif stratiformede Bjerkreim-Sokndal (Rogaland, Norvege):une intrusion composite de la suite charnocki-tique. In: Demaiffe, D. (ed.) Petrology andgeochemistry of magmatic suites of rocks in



the continental and oceanic crusts. UniversitéLibre de Bruxelles, 129-144.

Bolle, O., Diot, H. & Duchesne, J.C. 1997:Anisotropie de la susceptibilté magneétiquedans une intrusion composite de la suite charnockitique: l'apophyse du massif stratifor-me de Bjerkreim-Sokndal (Rogaland, Norvègeméridionale). Les Comptes rendus de l'Académiedes sciences, Paris 325, 799-805.

Bolle, O., Trinidada, R.I.F., Bouchez, J.L. &Duchesne, J.C. 2002: Imaging downward grani-tic magma transport in the Rogaland IgneousComplex, SW Norway. Terra Nova 14, 87-92.

Demaiffe, D., Duchesne, J.C., Michot, J. &Pasteels, P. 1973: Le massif anorthosito-leuco-noritique d'Hidra et son faciès de bordure. LesComptes rendus de l'Académie des sciences,Paris 277, 14-20.

Demaiffe, D. & Hertogen, J. 1981: Rare earth element geochemistry and strontium isotopiccomposition of a massif-type anorthositic-charnockitic body: the Hidra Massif (Rogaland,SW Norway). Geochimica et Cosmochimica Acta45, 1545-1561.

Duchesne, J.C. & Hertogen, J. 1988: Le magmaparental du lopolithe de Bjerkreim-Sokndal(Norvège Méridionale). Les Comptes rendus de l'Académie des sciences, Paris 306, 45-48.

Duchesne, J.C. & Maquil, R. 1987: The Egersund-Ogna massif. In: Maijer, C and Padget, P. (eds.)The geology of southernmost Norway: anexcursion guide. Norges geologiske undersøkelseSpecial Publication 1, 50-56.

Duchesne, J.C. & Wilmart, E. 1997: Igneous char-nockites and related rocks from the Bjerkreim-Sokndal layered intrusion (southwest Norway):a jotunite (hypersthene monzodiorite)-derivedA-type granitoid. Journal of Petrology 38, 337-369.

Falkum, T. & Petersen, J.S. 1987: The Agder migmatitic gneiss complex. In: Maijer, C. andPadget, P. (eds.) The geology of southernmostNorway: an excursion guide. Norges geologiskeundersøkelse Special Publication 1, 40-47.

Heldal, T. & Lund, B. 1995: A regional study ofthe dimension-stone potential in labradorite-bearing anorthositic rocks in the RogalandIgneous Complex. Norges geologiske undersø-kelse Bulletin 427, 123-126.

Hermans, G.A.E.M., Tobi, A.C., Poorter, R.P.E. &Maijer, C. 1975: The high-grade metamorphicPrecambrian of the Sirdal-Ørsdal area,Rogaland/Vest Agder, SW Norway. Norges geologiske undersøkelse 318, 51-74.

Karlsen, T.A., Gautneb, H., Korneliussen, A., Lund,B., Marker, M. & Nilsson, L.P. 1999: Undersøkelseav mineralressurser i Sokndals-området,Rogaland. Norges geologiske undersøkelseReport 99.056, 68 pp.

Karlsen, T.A., Nilsson, L.P., Schiellerup, H., Marker,M. & Gautneb, H. 1998: Berggrunnsgeologiskkart over Åna-Sira anortosittmassiv m. omgi-velser. Målestokk 1:25000. Norges geologiskeundersøkelse (also enclosure in NGU Rapport98.068).

Knorn, H. & Krause, H. 1977: DieVerbandsverhältnisse südlich von Tellnes imZentralteil des Åna-Sira-Massivs(Südnorwegen). Norsk Geologisk Tidsskrift 57,85-95.

Krause, H., Gierth, E. & Schott, W. 1985: Ti-Fedeposits in the South Rogaland IgneousComplex, with special reference to the Åna-Sira anorthosite massif. Norges geologiskeundersøkelse Bulletin 402, 25-37.

Michot, J. 1961: The anorthositic complex ofHaaland-Helleren. Norsk Geologisk Tidsskrift 41,157-172.

Michot, J. & Michot, P. 1969: The problem ofanorthosites: the South Rogaland igneouscomplex, southern Norway. In: Isachsen, Y.W.(ed.) Origin of anorthosite and related rocks.New York State Museum Science ServiceMemoir 18, 399-410.

Michot, P. 1960: La géologie de la catazone: leproblème des anorthosites, la paligenèse basi-que et la tectonique catazonale dans leRogaland méridional (Norvège méridionale).Excursion guide A9, 21st International GeologicalCongress 1960, 54 pp.

Möller, A., O'Brien P.J., Kennedy, A. & Kröner, A.2002: Polyphase zircon in ultrahigh-temperatu-re granulites (Rogaland, SW Norway): con-straints for Pb diffusion in zircon. Journal ofMetamorphic Geology 20, 727-740.

Paludan, J., Hansen, U.B. & Olesen, N.Ø. 1994:Structural evolution of the PrecambrianBjerkreim-Sokndal Intrusion, South Norway.Norsk Geologisk Tidsskrift 74, 185-198.

Rietmeijer, F.J.M. 1979: Pyroxenes from iron-richigneous rocks in Rogaland, SW Norway.Geologica Ultrajectina 21, 341 pp.

Robins, B., Tumyr, O., Tysseland, M. & Garmann,L.B. 1997: The Bjerkreim-Sokndal layered intru-sion, Rogaland, SW Norway: Evidence frommarginal rocks for a jotunite parent magma.Lithos 39, 121-133.

Schärer, U., Wilmart, E. & Duchesne, J.C. 1996:The short duration and anorogenic characterof anorthosite magmatism: U-Pb dating of the Rogaland complex, Norway. Earth andPlanetary Science Letters 139, 335-350.


Vander Auwera, J. & Duchesne, J.C. 1996:Petrology and geochemistry of the noritic

Hogstad layered body (Rogaland, SW Norway):implications for the nature of andesine anor-thosite parent magma. In: Demaiffe, D. (ed.)Petrology and geochemistry of magmatic suitesof rocks in the continental and oceanic crust.Université Libre de Bruxelles, 111-128.

Weis, D. & Demaiffe, D. 1983: Pb isotopegeochemistry of a massif-type anorthositic-charnockitic body: the Hidra Massif (Rogaland,S.W. Norway). Geochimica et CosmochimicaActa 47, 1405-1413.

Wilson, J.R., Robins, B., Nielsen, F.M., Duchesne,J.C. & Vander Auwera, J. 1996: The Bjerkreim-Sokndal layered intrusion, southwest Norway.In: Cawthorn, R.G. (ed.) Layered intrusions,Elsevier Science B.V., Amsterdam, 231-256.

Zeino-Mahmalat, R. & Krause, H. 1976: Plagio-klase im anorthosit-komplex von Åna-Sira, SW-Norwegen. Petrologische und chemische unter-suchungen. Norsk Geologisk Tidsskrift 56, 51-94.


Mineral resources in the Rogaland AnorthositeProvince, South Norway: origins, history and recentdevelopments

Henrik Schiellerup, Are Korneliussen, Tom Heldal, Mogens Marker, TerjeBjerkgård & Lars-Petter Nilsson

Schiellerup, H., Korneliussen, A., Heldal, T., Marker, M., Bjerkgård, T., Nilsson, L,-P.: 2003. Mineralresources in the Rogaland Anorthosite Province, South Norway: origins, history and recent develop-ments. Norges geologiske undersøkelse Special Publication 9, 117-135.

The Neoproterozoic Rogaland Anorthosite Province (RAP) in Southwest Norway contains significantmineral resources. Most important are the ilmenite ores, of which the world-class Tellnes orebody is cur-rently in production. Tellnes supplies some 7% of the total world titanium consumption. During the lastdecades ilmenite refiners have been increasingly demanding with respect to trace elements, which havewidened the targets for ilmenite exploration in the area. A review of resources, quality and the currentstate of knowledge on their genesis is presented.Sulphide mining in the RAP was abandoned by the end of the 19th century, but 11 occurrences of Fe-Ni-Cu sulphides are known. Only one ore body, the Homsevatnet deposit, is of significant size, but all occur-rences are similar in terms of paragenesis and composition. All known deposits are poor in noble metals,implying that little interaction took place between immiscible silicate and sulphide melts. In addition,most mafic intrusions contain subsidiary amounts of dispersed sulphide droplets, but only in theBjerkreim-Sokndal layered intrusion did a segregation process lead to increased grade and tenor.Current resources also include dimension stones. Anorthosite containing blue-iridescent plagioclase(Labrador Antique) is at the moment the most valuable dimension stone produced in Norway. Ongoingexploration is demonstrating the existence of large quantities of exploitable dimension stone resources.Aggregate is extracted at four quarries from white altered anorthosite as well as from jotunite, and his-torically, kaolin has been mined on a limited scale.The identified resources of ilmenite (± apatite, ± vanadian magnetite), dimension stone and aggregateare all very large. In terms of current extraction, exploration and future prospects, the RAP ranks as oneof the most important mineral provinces in Norway.

Henrik Schiellerup, Are Korneliussen, Tom Heldal, Mogens Marker, Terje Bjerkgård & Lars-Petter Nilsson,Norges geologiske undersøkelse, N-7491 Trondheim, Norway.


IntroductionThe Neoproterozoic massif-type anortho-site province in Rogaland county, SouthNorway, has been the target of mining andexploration for more than 200 years. Theprovince is currently being exploited for awide selection of mineral commodities.The economically most important pro-duct is ilmenite. The province contains avariety of ilmenite deposits, the bestknown being the Tellnes, Storgangen andBlåfjell deposits. Since 1965, Tellnes hasbeen the only ilmenite mine in productionin the province and presently the depositsupplies around 7% of the world’s tita-nium consumption. More than 80 minesand prospects are known (Fig. 1).

Changing processes in the production ofTiO2 as well as environmental concernshave focussed attention on the chemicalspecifications of titanium raw materials.For ilmenite, magnesium is the mainissue, because magnesium basically con-trols which recovery processes can beapplied to a given ilmenite feed stock. InTiO2-pigment production, a chlorinatablefeed stock must contain less than 1.5%MgO (Korneliussen et al. 2000). Otherdeleterious elements in pigment producti-on include Cr and Ni, which may reside inilmenite, and pollutants such as Ca, whichmainly derive from silicate minerals in theconcentrates. In Rogaland, much efforthas recently been expended to locatepotential deposits of low-MgO ilmenite inconjunction with vanadian magnetite andapatite. Several noritic rock bodies con-tain in excess of 10% apatite, and signifi-cant resources with 30-35% ilmenite +apatite + vanadian magnetite have beenidentified.

Several semi-massive to massive sulphidedeposits located within the anorthositebodies have been mined or explored for Niand Cu. The largest is the Homsevatnetdeposit, located 12 km northeast of Eger-sund (Fig. 1), where mining took placeduring the 1870s. Several noritic intru-sions in Rogaland contain dispersed basemetal sulphides in amounts of around1%. In the Bjerkreim-Sokndal LayeredIntrusion there is a stratiform sulphideoccurrence with sulphide contents of upto around 10%. Disseminated sulphidesare recovered as a by-product of ilmeniteexploitation at Tellnes, primarily for theirnickel content. Today, the economic poten-

tial of the Rogaland sulphide depositsseems to be insignificant.

Anorthosite, in different varieties, andassociated jotunitic rocks are quarried foraggregate, and the presence of iridescentplagioclase in the anorthosites haverecently given rise to a dimension stoneindustry. Anorthosite from Rogaland con-taining blue-iridescent plagioclase is soldunder the name 'Labrador Antique'.Mining for kaolin has taken place at seve-ral localities in the anorthosite bodies, butwas abandoned at the beginning of the20th century.

The purpose of this paper is to review thehistory, occurrence and potential of theRogaland mineral deposits and summar-ize current ideas on their formation.

Geological settingThe RAP (Fig. 1) is located along the coastof southwestern Norway, roughly betweenOgnabukta in the north and Lindesnes inthe south. It covers an onshore area ofabout 1750 km2, and a similar area offsho-re (Sigmond 1992). The province is domi-nated by four major massif-type anortho-site plutons; the Egersund-Ogna, Håland,Helleren and Åna-Sira anorthosites,which are also the oldest intrusive units inthe province. The anorthosites consistmostly of monotonous anorthosite or leu-conorite. The Egersund-Ogna massif con-tains a marginal zone of foliated leucono-rites, presumably formed through syn-emplacement deformation (Duchesne &Michot 1987), and the Håland massifmainly comprises foliated leuconorites.Metre-sized aggregates of megacrystic alu-minous and Cr-rich orthopyroxene arefound to variable extent in all massifs.They bear evidence for a polybaric originof the anorthosites and indicate initialpressures of crystallization of 11-13 kbar(Longhi et al. 1999, Vander Auwera et al.2000). Final emplacement of the anortho-sites and associated intrusions probablytook place at around 5 kbar (VanderAuwera & Longhi 1994). The whole pro-vince was formed during a brief period oftime bracketed between the emplacementof the major anorthosites at 932-929 Maand the emplacement of the Tellnes ilme-nite deposit in the Åna-Sira anorthosite at920 ± 3 Ma (Schärer et al. 1996).


Fig. 1. Map of the western part of the Rogaland Anorthosite Province (Marker et al. 2003) showing registered prospects and deposits of ilmenite(±magnetite, ±apatite) and Fe-Ni-Cu sulphides. Anorthositic rocks are shown in blue colours, jotunitic and noritic rocks in yellow-green and granitoidrocks in red (see Marker et al. (2003) full map legend). Inset map shows the major intrusive units; BKSK: Bjerkreim-Sokndal Layered Intrusion, QMG:quartz-mangeritic rocks, E-R: Eia-Rekefjord Intrusion, G: Garsaknatt Intrusion. Numbers refer to deposits. Ilmenite deposits: 1: Saksekjær, 2: Veisdal,3: Tinnbakkneset, 4: Vardåsen, 5: Hilleråsen, 6: Grasjenåsen, 7: Skolla skjerp, 8: Kvannvikåsen, 9: Dalestøa skjerp, 10: Strandbakken, 11: Gjersdal, 12: Raunslid,13: Tellnes, 14: Brombo, 15: Mydland, 16: Bøstølen, 17: Sletthei, 18: Laksedal, 19: Dalens gang, 20: Bryns skjerp, 21: Blåfjelldalen, 22: Blåfjell (several adits),23: Li gruve, 24: Aursland skjerp, 25: Brekke gruve, 26-29: Storgangen gruver, 30: Sidegangen, 31: Åmodt gruve, 32: Kjørfjell, 33: Blåfjellskaret, 34: Frøytlog,35: Ålgård, 36: Florklev, 37: Bakka, 38: Ørsland, 39: Årstad gruve, 40: Prestbro skjerp, 41: Bø utmark, 42: Årstadøy, 43: Hauge gruve, 44: Drageland, 45: Vat-land, 46: Svånes, 47: Pramstø, 48: Pramsknuten/Jerneld, 49: Snøringdalen, 50: Vaksviken, 51: Eigerøy, 52: Hestnes, 53: Tysslandsvad, 54: Rødemyr gruve,55: Fisketjern gruve, 56: Berkjedalen, 57: Thorsdalen gruve, 58: Skåra gruve, 59: Raaes gruve, 60: Kaknuten gruve, 61: Bredeskaret, 62: Sørskaret,63: Omdalsskaret, 64: Vallbakken, 65: Langemyr, 66: Karen gruve, 67: Knuts gruve, 68: Lyngnes gruve, 69: Lundetangen, 70: Peder Anker, 71: Ankershus,72: Egebakkskaret (W), 73: Egebakkskaret (E), 74-75: Lielven gruver, 76-77: Furufjellet gruver, 78: Jentofts gruve, 79-80: Sandknuten gruver,81: Beinskinnvatnet, 82: Åsen, 83: Helleland, 84: Teksevatnet, 85: Tekse-Bilstad, 86: Vasshus. Sulphide deposits: 87: Urdal, 88: Grastveit, 89: Koltjørn,90: Helland, 91: Skora, 92: Myklebust, 93: Hellvik, 94: Piggstein, 95-96: Ualand, 97: Bjørndalsnipa, 98: Homsevatnet, 99: Fossfjellet, 100: Rudleknuten.


A number of slightly younger intrusionsare found in close spatial association withthe main anorthosite bodies. These inclu-de two smaller anorthositic/leuconoriticbodies (the Hidra and Garsaknatt intru-sions), the 230 km2 Bjerkreim-SokndalLayered Intrusion and the Eia-Rekefjordjotunitic intrusion (Fig. 1). In addition,the province is cross-cut by numerousnoritic, jotunitic and mangeritic dykes ofvariable width and lateral persistence.

Acidic rocks in the province are represen-ted by quartz mangerites and charnockiteswhich cap the Bjerkreim-Sokndal Intru-sion. Their status has been interpreted ascontaminated differentiates of the Bjerk-reim-Sokndal Layered Intrusion (Nielsenet al. 1996), possibly mingled with exter-nally-derived acidic magma (Duchesne &Wilmart 1997). The Farsund Charnockitein the eastern part of the province (Fig. 1)is also considered a part of the anorthositeprovince (Duchesne et al. 1985).

IlmeniteMining historyThe existence of significant Fe-Ti oxidedeposits in Rogaland was recognised morethan 200 years ago, when a number of ironmines were opened in ilmenite depositsaround Lake Kydlandsvatnet in the wes-tern part of the province in 1785 (Krauseet al. 1985). However, Moss Jernverk, whooperated the mines, never managed tosmelt the Ti-rich ore and closed downoperations after only a few years of activi-ty. During the following century, explora-tion located Fe-Ti oxide deposits atStorgangen, Blåfjell and Flordalen (Dahll1863), and in the mid-1800s mining wasre-started here and at several localities inthe southern (Sokndal) lobe of the Bjerk-reim-Sokndal Layered Intrusion (Carlson1945), as well as in the Kydlandsvatn area.However, the high titanium content againlimited the value of the deposits as ironore, and by 1881 industrial-scale produc-tion had ceased in the province.

In 1910, JHL Vogt, after decades of dedica-ted research on the Rogaland Fe-Ti pro-spects, regrettably concluded that theRogaland deposits would have little eco-nomic value until further use for titaniumhad been found (Vogt 1910). Only a fewyears later, however, pioneering researchby G. Jebsen and P. Farup resulted in the

development of a new industrial methodfor producing a white TiO2 pigment (nowcalled the sulphate process), and demandfor titanium minerals dramatically increa-sed. By 1917, Titan Co. A/S, Frederikstad,had acquired the majority of the shares inthe mining company Titania A/S (foun-ded in 1902), that owned the rights to theRogaland deposits, in order to secure theraw materials for their new titanium pig-ment plant.

A beneficiation plant was constructed atSandbekk near the Storgangen orebody,then the largest known prospect in theprovince, and industrial-scale miningcommenced in 1916. Reserves at Stor-gangen were estimated to at least 200,000tof ore – enough for 10 years of producti-on. During the following two decadesStorgangen was one of the largest tita-nium mineral producers in the world, andwhen mining finally ceased in 1965 thedeposit had yielded more than 10 Mt ofore. More than 60 Mt of reserves remain(Hagen 1992).

By 1965, all production had shifted to anopen-pit mine at Tellnes also within theÅna-Sira anorthosite body. Here, three kmsoutheast of Storgangen, the existence ofan elongated noritic lens had been knownsince the late 19th century (Kolderup1896), but its potential as an ilmeniteresource was only recognized followingaeromagnetic surveys in the mid-1950s.Exploitation continues at Tellnes today –the Tellnes ore reserve of 380 Mt withmore than 60 Mt contained TiO2 repre-sents around 15% of the world's titaniummineral reserves, and production at Tellnescurrently accounts for 7% of the total glo-bal production of titanium minerals.

Some of the ilmenite deposits and ilmenite-rich rocks in Rogaland carry significantamounts of vanadian magnetite and apa-tite. Recently, attention has been directedtowards the large, but low-grade, gabbro-norite-hosted, ilmenite occurrences inwhich ilmenite may be exploited in con-junction with apatite and vanadian mag-netite.

Occurrence and genesisThe RAP contains a wide range of Fe-Tideposits, from small bodies of massiveilmenite to larger volumes of modally lay-ered rocks in layered intrusions. A cumu-


late origin is usually accepted for ilmenite-rich rocks found as concordant layers orzones in layered intrusions, such as in theBjerkreim-Sokndal (Fig. 1), Hogstad (Fig.1, #7), Storgangen (#26-29) and Bøstølen(#16) intrusions. However, within theanorthosite massifs oxide-rich rock com-monly occurs in minor bodies of more orless homogeneous composition, usuallycharacterized by a simple mineralogy. Inthese cases the origin is much moreobscure. Nevertheless, a continuous rangeof compositions does exist, from pureoxide rocks to oxide-plagioclase rocks andoxide-rich norites grading into multi-phase assemblages which are generally lay-ered (Table 1). Duchesne (1999) inferred acumulate origin for all non-nelsoniticdeposits in the region.

'Primitive' ilmenite-rich intrusions ofmore or less mono- or bi-mineralic com-position are widespread in Rogaland.Modal layering in these rocks is either rareor rudimentary, and the intrusions arefound mainly as irregular dykes or lenti-cular bodies. Most of these primitive ilme-nite-rich rocks consist of masses of almostpure ilmenite, some covering an area of upto a few 100 m2. Texturally, the ilmenite isthoroughly equilibrated, forming coarse-grained polygonal aggregates, with inter-stitial green spinel (pleonast) and locallysubsidiary amounts of magnetite. How-ever, isolated monomineralic ilmenite bodi-es are rare, and most primitive deposits are,in fact, spatially associated with noriticdykes or lenses of variable dimensions.

The Pramsknuten (Fig. 1, #48) (orJerneld) deposit (Duchesne 1999) is oneof a small number of deposits which donot seem to contain silicates at all. Thedeposit is composed of a string of smalloxide lenses in an elongated zone approxi-mately 200 m long within the Hellerenanorthosite. The deposit is more or lessconfined to a zone of 'white' (altered)anorthosite. Like the other massive oxidebodies, the ilmenites are texturally equili-brated and a cumulate origin is speculati-ve. Sulphides are abundant and consistpredominantly of pyrite. As in all primiti-ve ilmenite deposits ilmenite grains con-tain a relatively high amount of exsolvedhematite (Duchesne 1999) and are moreproperly termed hemo-ilmenite. The ilm-enite composition is characterized by anMgO content of around 5%, which is

among the highest MgO contents in ilme-nite recorded in Rogaland. The Cr2O3

content of ilmenite is generally positivelycorrelated with MgO and exceeds 0.4% inthe Pramsknuten ilmenite (Duchesne1999, Duchesne & Schiellerup 2001).

Bi-mineralic oxide deposits contain plagi-oclase in addition to hemo-ilmenite andare found, for instance, at Florklev andÅlgård in the Åna-Sira Anorthosite (Fig. 1,#35,36). The mineralization here is relatedto a narrow dyke in which grains of little-deformed, tabular, and generally lamina-ted plagioclase are dispersed in a recrystal-lized ilmenite matrix. In this case the pre-sence of significant amounts of euhedralplagioclase (10-50%) indicates that thedeposits have a cumulate origin, and thatall subsequent deformation was accom-modated by oxide recrystallization. In thedyke, modal variations delineate subtlelayering. This type of occurrence qualifyas plagioclase-ilmenite cumulates (piCs).Plagioclase-ilmenite cumulates also makeup large parts of the Bjerkreim-SokndalLayered Intrusion (Wilson et al. 1996,Robins & Wilson 2001). Generally, howe-ver, these cumulates are anorthosites orleuconorites with only a few percent ilme-nite. Only when magma influx and mixingin the Bjerkreim-Sokndal magma chambertook place did more oxide-rich plagiocla-se-ilmenite (± Ca-poor pyroxene) cumula-tes form (Jensen et al. 1993) (see below).

Most massive oxide deposits are found inassociation with noritic rocks, characteri-zed by the assemblage: hemo-ilmenite,plagioclase and orthopyroxene. The Blå-fjell-Laksedal field (Fig. 1, #18-23) con-tains some of the most renowned depositsin the province, all of which are related tothe same norite pegmatite (Krause et al.1985). Seven adits at Blåfjell into the prin-cipal mineralizations of the pegmatiteproduced around 100,000 tons of oreduring the second half of the 19th century.Though the exact relationship betweenthe massive oxides and the norite pegma-tite is indistinct, thick modal layers ofmonomineralic ilmenite in the pegmatiteare developed in the vicinity of the Blåfjellmines. Local gradational contacts areobserved between unmineralized noriteand plagioclase-bearing massive ore. Theseobservations seem to suggest a cumulateorigin for the ilmenite ore bodies in theBlåfjell-Laksedal field.


In other classic deposits, such as Vardåsen,in the eastern part of the Åna-SiraAnorthosite (Fig. 1, #4), and at least local-ly in the Koldal field (#61-67), massiveilmenite ores are also spatially associatedwith norite dykes. Whereas the associationwith norite dykes in the Koldal field isunclear, the Vardåsen deposit displays atransitional contact in a north-south stri-king dyke, between a poorly mineralizedlayered noritic cumulate to the west andalmost massive ilmenite to the east. Indeposits where ilmenite is not concentra-ted into massive ore bodies but intersper-sed with plagioclase and orthopyroxene innoritic rock, the ilmenite ore is typicallymodally layered. This is the case in theRaunsli, Frøytlog and Svånes deposits(Fig. 1, #12,34,46).

In the Bjerkreim-Sokndal Layered Intru-sion, plagioclase-ilmenite (± orthopyro-xene) cumulates with up to 50% hemo-ilmenite are found at the base of twomegacyclic units, where influx of freshmagma caused cryptic regressions in theLayered Series. Though these ilmenite-rich rocks have never been exploited theyconstitute a large titanium resource. Theregressive sequence at the base of megacy-clic unit IV (zone 'A' in Fig. 2) is up to 30m thick and may be followed for morethan 20 km along strike (Jensen et al.1993). A single ilmenite analysis yielded2.7% MgO, whereas ilmenite immediatelyabove the zone contains 4.4% MgO and690 ppm Cr2O3 (Duchesne 1999). Theprimitive basal oxide-rich cumulates alter-nate with anorthositic layers and crystalli-zed from a hybrid magma, produced bymixing of evolved resident magma withinflowing primitive magma (Wilson et al.1996).

In all the ilmenite deposits that containilmenite with or without cumulus plagio-clase and orthopyroxene, the ilmenitecontains extensive hematite exsolutionsand is generally characterized by MgOcontents of more than 4% and Cr2O3contents above 4-500 ppm. Coexistingplagioclase has anorthite contents of 47-51% and orthopyroxene Mg-numbersbetween 73 and 78%.

TellnesThe Tellnes deposit (Fig. 1, #13) is by farthe largest deposit in Rogaland with reser-ves amounting to 380 MT with more than

60 MT contained TiO2. Cross-cuttingrelationships, as well as U-Pb isotopicdating, imply that the emplacement of theTellnes deposit is one of the youngestevents in the Rogaland AnorthositeProvince. The emplacement has beendated by the U-Pb method to 920 ± 3 Ma(Schärer et al. 1996).

Tellnes is a large, sickle-shaped, Ti-richnoritic ore-body almost three km longand more than 400 m wide. The intrusionis fairly homogeneous and modal layeringis rare. The average modal composition ofthe Tellnes ore body is 53% plagioclase(An45-42), 29% hemo-ilmenite, 10%orthopyroxene (En77-75) and 4% biotite(Gierth & Krause 1973, Krause et al. 1985,Duchesne 1999). The remaining ~4% ismade up by the accessory phases: olivine(which is commonly replaced by orthopy-roxene-oxide symplectites), magnetite,Ca-rich pyroxene, apatite and sulphides.Though mineral lamination does occur,cumulate textures are not obvious. Themineralogy is essentially constant withmore than 90% of any sample made up bythe three phases plagioclase, orthopyro-xene and ilmenite. However, the modalproportions of the minerals may vary con-siderably which results in a strong chemi-cal zoning across the orebody (Robinsonet al., this volume). The content of TiO2

varies between 10 and 25% with somesamples containing up to 35% TiO2.

The texture is to a large extent dominatedby subhedral to euhedral plagioclase laths.Ca-rich pyroxene is often present only asrims on orthopyroxene grains. The Tellnesdeposit may therefore most appropriatelybe considered a three-phase cumulate ofplagioclase, hemo-ilmenite and orthopy-roxene (pihC), with the local presence ofcumulus olivine. The minor phases areentirely of intercumulus origin.

Hemo-ilmenites from the central part ofthe orebody contain around 15% hemati-te component. They typically have MgOcontents in excess of 3-4% and Cr2O3 con-tents of 400-550 ppm (Table 1, Krause etal. 1985, Duchesne 1999, Kullerud 1994).The MgO, Cr2O3 and hematite contents ofilmenite in the central part of Tellnes arethus typical for Rogaland deposit typeswith few coexisting mineral phases, suchas the Raunsli, Frøytlog, Flordalen andSvånes deposits mentioned above. The


MgO content of ilmenite does, however,show a systematic variation with lowervalues generally found at increasinglymarginal sites, and ilmenite and silicatecompositions show an overall correlationwith the modal amount of ilmenite(Robinson et al., this volume).

The ore deposit has been interpreted ashaving been emplaced as a noritic crystal

mush wet by an interstitial (Ti-rich) jotu-nitic melt (Wilmart et al. 1989). Recentexperimental results have now confirmedthat the late-stage melt must have beenjotunitic in composition, but suggest thatit was relatively Ti-poor (Skjerlie et al., thisvolume). Based on field relations, theTellnes deposit was also thought to begenetically related to an associated mang-eritic dyke, resulting in complex geneticmodels (Krause & Pedall 1980, Wilmart etal. 1989). However, U-Pb dating has reve-aled that the two intrusive units are sepa-rated in time by more than 10 Ma, rende-ring a genetic link improbable (Schärer etal. 1996). Similarly, a suggested relations-hip between the Tellnes orebody and the10 Ma older Åna-Sira anorthosite, basedon strontium and oxygen isotopes (Wil-mart et al. 1989, 1994), also have to beabandoned. Robinson and coworkers (inthis volume) suggest that mixing betweena relatively evolved magma and a more pri-mitive magma may have been importantin the evolution of the chemical and modalcomposition of the Tellnes ore body.

StorgangenThe Storgangen (Norwegian for 'largedyke') Intrusion is a relatively thin modal-ly-layered sheet emplaced into the Åna-Sira anorthosite (Fig. 1, #26-29). Stor-gangen is the second largest Fe-Ti orebodyin the RAP. In contrast to Tellnes, theintrusion contains magnetite as an impor-tant mineral phase. The deposit is a laye-red cumulate consisting essentially of pla-gioclase, ilmenite, orthopyroxene andmagnetite (phimC). The intrusion has anarcuate shape and a length of approxima-tely 4 km. Its thickness varies from 2-3 mat its eastern and western extremities up toa maximum of about 50-60 m in its cen-tral part. The strike of the modal layeringis parallel to the country rock contacts anddips around 50° to the north in the centralpart of the intrusion. The dip graduallydecreases with depth. Cumulate texturesare overprinted by a tectonic fabric, whichimplies that the Storgangen intrusion hadsolidified prior to the doming of the Åna-Sira anorthosite. Layering is generallyparallel with the contacts to the anortho-site and in the western extremity also withthe layering in the Bjerkreim-SokndalLayered Intrusion. It is therefore inferredthat Storgangen was originally emplacedas a sub-horizontal sill-like body.

Fig. 2. Generalized stratigraphy and zonesequences in the Bjerkreim-SokndalLayered Intrusion according to Wilson etal. (1996). The full sequence of MCUs, asshown here, is developed only in the (nor-thern) Bjerkreim-lobe of the intrusion.The total thickness of the Layered Series is close to 7000 m. 'TZ': Transition Zone.


Mining has produced a 70 m-deep gorgethat provides sections across the intrusi-on. The rock types range from anorthosi-te and leuconorite to norite and melanori-te with oxide contents varying from near-zero to more than 50%. Massive ilmeniteore with anorthosite xenoliths is found atthe footwall contact. The average TiO2

content of the Storgangen cumulates isaround 16%.

A slight cryptic layering defined by plagi-oclase and orthopyroxene compositionswas noted by Krause et al. (1985) indica-ting more-differentiated cumulates to-wards the hanging-wall contact. However,the tendency for an up-section decrease inboth anorthite and enstatite contents issubtle due to large compositional variabi-lity. Plagioclase and orthopyroxene com-positions range from An52 and En72 to An46

and En66, respectively. MgO contents ofilmenite closely mimic the compositionalvariation of orthopyroxene and vary from2 to 3.5%. Chromium contents vary betwe-en 230 and 690 ppm Cr2O3 (Duchesne1999, Duchesne & Schiellerup 2001). Inaddition, the amount of hematite exsolvedfrom ilmenite is significantly reduced com-pared with the more primitive deposits.

The Storgangen intrusion, as a four-phasecumulate assemblage, cannot be an imme-diate differentiation product of the meltthat formed the mono-mineralic anortho-site, unless the intermediate cumulates arehidden at depth. However, the presence ofsimilar plagioclase megacrysts in bothStorgangen and the Åna-Sira anorthosite,the occurrence of anorthositic layers inStorgangen, the contact relations and thejoint post-magmatic deformation historysuggest a close temporal and genetic linkbetween the two units.

From 1916 to 1965, when production hadirrevocably been transferred to the Tellnesdeposit, Storgangen produced around 10Mt of ore. However, with 60 Mt of docu-mented reserves, Storgangen remains adeposit of world-class significance.

Other phimC depositsOther intrusions in Rogaland that containilmenite-rich cumulates consisting of pla-gioclase, orthopyroxene, ilmenite andmagnetite are present in the southern partof the Bøstølen Intrusion in the centralpart of the Åna-Sira anorthosite (Fig. 1,

#16) and parts of the Bjerkreim-Sokndalintrusion. Bøstølen is probably the oldestmajor intrusion in the Åna-Sira anortho-site, and is cross-cut by both norite peg-matites of the Blåfjell-Laksedal system andlater mangerite dykes. The oxide-rich sec-tions of the layered Bøstølen Intrusionhave never been exploited, but averagearound 14% TiO2 in the southern part ofthe intrusion. A small ilmenite deposit wasexploited at Årstad in the southern part ofthe Bjerkreim-Sokndal Layered Intrusion(Fig. 1, #39) at the turn of the 19th centu-ry. This deposit contains plagioclase,orthopyroxene, ilmenite and magnetite,although it is located in a zone of phiCclose to the base of the Layered Series.

In the Håland anorthosite along its con-tact with the Egersund-Ogna anorthosite,there are a string of deposits in a zoneextending from the south end of Eigerøyeastwards for almost 10 km (Fig. 1, #50-80). These deposits are variably deformedbut generally interpreted as cumulates(Duchesne 1999). In some cases, such asthe deposits along the north bank of lakeKydlandsvatnet (#68-80), the orebodiesmay be parts of deformed layered noritesills, while mineralizations south of Koldalfarm are more clearly related to noritedykes. More than 30 deposits have beenexploited along this zone and the majoritycontain magnetite. Most of these depositscan be interpreted as phimCs. Some of thedeposits north of lake Kydlandsvatnet donot contain magnetite and classify asphiCs while others (e.g. Rødemyr, Kaknu-ten) contain apatite as well magnetite.Duchesne (1999) reported ilmenite com-positions in the phi±m deposits of thisarea with MgO ranging from 1.9 to 3.7%and Cr2O3 contents between 100 and 690ppm. Further deposits with similar cha-racteristics include some of the deposits atLaksedal in the Åna-Sira anorthosite (Fig.1, #18) and Vatland in the Helleren anor-thosite (Fig. 1,#45).

Apatite-bearing depositsApatite-bearing ilmenite deposits can beroughly divided into either nelsoniticoccurrences (consisting principally of oxi-des and apatite) or layered cumulates ofplagioclase, two pyroxenes, ilmenite, mag-netite and apatite (phcimaC).

Allthough significant apatite is reported in a number of deposits (e.g. Rødemyr,


Kaknuten (Duchesne 1999)) and strayapatite grains may be found even inoccurrences of pure ilmenite, nelsonite isrelatively rare in Rogaland compared withother anorthosite provinces. Two deposits(Eigerøy and Hestnes) close to the contactbetween the Håland-Helleren and Eger-sund-Ogna anorthosites south of Eger-sund contain insignificant amounts ofsilicates and have been classified as nelso-nites (Duchesne 1999). Duchesne (1999)reported MgO contents of ilmenite inthese deposits of 1.1-1.4%.

Apatite-bearing oxide-rich rocks areotherwise confined to the evolved gabbro-noritic portions of the Bjerkreim-SokndalLayered Intrusion. Other apatite-bearingrocks, such as the persistent jotuniticdykes transecting the province (VanderAuwera et al. 1998), various layered jotu-nitic bodies, as well as the uppermost partof the Bjerkreim-Sokndal Layered Intru-sion, are generally relatively poor in ilme-nite.

The Bjerkreim-Sokndal Layered IntrusionThe Bjerkreim-Sokndal Layered Intrusionhas been studied almost continuouslysince the late 19th century (Michot 1960,1965, Duchesne 1987, Wilson et al. 1996).The intrusion is exposed in three differentlobes; the Bjerkreim, Sokndal and Myd-land lobes. All lobes contain ilmenite pro-spects and exploitation has taken place inat least two localities in the Sokndal-lobe(Årstad and Hauge – Fig. 1, #39,43). Themost extensive stratigraphy is present inthe Bjerkreim-lobe where a 7 km-thickLayered Series is developed. The cumulatesequence may be divided into a number ofmega-cyclic units (MCUs) each resultingfrom fractional crystallization after rechar-ge of the magma chamber. Seven suchMCUs have been defined (from base totop): 0, IA, IB, II, III and IV (Wilson et al.1996). In addition, each MCU may be divi-ded into a number of zones (a-f) based oncertain index mineral phases (Fig. 2).

Only the uppermost MCU (MCU IV) dis-plays the complete a-f zonal sequence, butapatite-bearing e-zones are found inMCUs IB, III and IV. In the Sokndal-lobeall cumulates belong to MCU IV, and inthe Mydland lobe, which has been envelo-ped by a subsequent intrusion (Bolle et al.1997), the cumulate sequence is at leastequivalent to MCU IV. The mining opera-

tion at Hauge in the Sokndal-lobe tookplace in very oxide-rich cumulates ofMCU IVe and resulted in a 25 m-deepshaft. Ore samples have been reported tocontain up to 26% TiO2 (Carlson 1945).Other unexploited occurrences in theSokndal-lobe include Bakka, Årstadøyand Drageland (Fig. 1, #37,43,44), whichare all located in the apatite-bearing MCUIVe zone, and contain up to 14% TiO2 inselected samples (Carlson 1945).

In the Bjerkreim-lobe two large prospectshave been identified as possible combinedresources for low-MgO ilmenite, apatiteand vanadian magnetite. One is locatedroughly between Åsen farm (Fig. 1, #82)and the village of Bjerkreim, following thestrike of the layering for up to 3000 m.Stratigraphically, these cumulates belongto MCU IBe. At Åsen farm this zone ave-rages 8% TiO2, 3.5% P2O5, and carriesaround 11% magnetite with 0.93% V2O3

in a 45 m section (Schiellerup et al. 2001,Meyer et al. 2002). The other prospect islocated in the basal part of MCU IVe onthe eastern flank of the lobe (Tekse-Bilstad– Fig. 1, #85), and constitutes a 30-90 m-thick section which may be followed alongstrike for several kilometres. Average com-positions are on the order of 7-8% TiO2

and 4% P2O5 (Schiellerup et al. 2001). Theoccurrence contain about 8% magnetitewith 0.9% V2O3. Similar occurrences arealso found on a smaller scale in the MCUIIIe zone. The apatite-bearing ilmenitedeposits and prospects generally containilmenite with low contents of MgO andCr2O3 and the ilmenite is poor in hemati-te. In all the Bjerkreim-lobe e-zone occur-rences the MgO content of ilmenite avera-ges between 1 and 1.8% (Meyer et al.2002).

GenesisThe Rogaland ilmenite deposits have beenascribed to a variety of origins from meta-somatic (Michot 1956, Hubaux 1960) tomagmatic (Duchesne 1972, Roelandts &Duchesne 1979, Duchesne 1999). Themagmatic hypotheses have includedliquid immiscibility and the existence ofFe-Ti-rich melts (e.g. Force 1991).Currently, with the possible exception ofthe nelsonitic occurrences, all the ilmeniteconcentrations are considered to becumulates (Duchesne 1996, 1999, Duch-esne & Schiellerup 2001). As indicatedabove there is a systematic variation in


ilmenite composition with the mineralo-gical paragenesis of the individual orebodies. Generally, the content of exsolvedhematite, Cr2O3 and MgO in ilmenitedecreases as the number of associatedminerals increases and their compositionsbecome more evolved (Duchesne &Schiellerup 2001). The composition ofilmenite from deposits region-wide corre-late with the compositions of ilmenite incumulates of similar mineralogical cha-racter in the Bjerkreim-Sokndal LayeredIntrusion. These observations support thesuggestion of cumulus processes to ex-plain the formation of most ilmenitedeposits in Rogaland as well as a derivati-on from similar types of melt. Except forthe Bjerkreim-Sokndal Layered Intrusion,which formed from melts differentiatingin situ, parental melts must have evolvedthrough fractional crystallization at depth.Discrepancies do exist among certaintrace elements, most notably chromium,which may to some extent reflect variablepressure conditions during the genesis ofthe province and its ilmenite resources(Vander Auwera et al. 2000, Duchesne &Schiellerup 2001). The correlation ofMgO in ilmenite with the magmatic evol-ution is important in prospecting for spe-cific qualities of ilmenite, though otherprocesses may also affect the compositi-ons. Mg is the principal pollutant in ilme-nite and the request for better qualityilmenite has led to the identification oflower-grade deposits containing low-MgO ilmenite with apatite and vanadianmagnetite as additional resources (e.g. theBjerkreim-Sokndal Layered Intrusion).

Whereas the mineralogical composition ofthe Rogaland ilmenite deposits is evident-ly controlled by fractional crystallization,the ilmenite-rich nature of the deposits ismore difficult to explain. The compositionof the Rogaland ilmenite deposits sensustrictu with TiO2 contents in excess of15%, such as Tellnes, Storgangen, Blåfjell,Vardåsen and other primitive deposits,seems to be strongly offset from the cotec-tic modal proportions of the crystallizingphases. The cotectic proportion of ilmen-ite (or TiO2) in plagioclase-orthopyro-xene-ilmenite cumulates may be estima-ted from the composition of the c-zonesin the Bjerkreim-Sokndal Layered Intru-sion. The c-zone cumulates are typicallyad-cumulates and make up a considerablepart of the Layered Series (Fig. 2). Eigh-

teen whole-rock analyses from the c-zonesof MCU II and III yield an average of4.4% TiO2. This TiO2 content is only mar-ginally higher than in the inferred paren-tal melt (3.46%) (Duchesne & Hertogen1988), and is in agreement with the relati-vely small quantity of ilmenite producedin experimental runs on the parental com-position (Vander Auwera & Longhi 1994).

Cumulates which are offset from cotecticproportions have been explained by mag-ma mixing, either through the forcing ofmagma compositions away from cotecticequilibrium (e.g. Meyer & Wilson 1999)or by resorption-crystallization processes(Tegner & Robins 1996, Lundgaard et al.2002). Similarly, the model for the forma-tion of the Tellnes ilmenite deposit advan-ced by Robinson et al. (this volume) invol-ves mixing of variably evolved magmas. Incases where extensive layering is present inthe intrusions, such as in Storgangen,wholesale mixing processes are more diffi-cult to envisage and probably requiremixing in an open system with highmagma throughput. Additional work sho-uld be undertaken in order to verify thenature of the processes that resulted in thepronounced ilmenite enrichment obser-ved.

SulphidesAlthough Fe-Ti ore is the foremost econo-mic resource of the RAP, base metal sul-phide deposits are found locally. The Fe-Cu-Ni sulphide occurrences may be divi-ded into two types. 1: minor anorthosite-hosted, disseminated to semi-massivedeposits, and 2: disseminated occurrenceshosted by noritic or orthopyroxeniticcumulates.

In Rogaland, Fe-Cu-Ni sulphide depositshosted by massif-type anorthosites aregenerally small and disseminated, alt-hough massive ores are also found. Thelargest known deposit is the Homsevatnetdeposit (Fig. 1, #98) estimated at 75,000-100,000 tons of sulphide ore (Hovland1975). Several other deposits were exploi-ted on a small scale at the end of the 19th

century, and during the 1970s and late1990s there was renewed interest in thearea and airborne geophysical prospectingwas carried out. The anorthosite-hostedmineralizations are widely scattered inthree of the major anorthosite bodies; the


Egersund-Ogna, Håland and Hellerenanorthosites (Fig. 1).

Although there is a variation in chemicalcomposition, morphology and to someextent mineralogy among the deposits,they are remarkably similar in a numberof ways. All investigated anorthosite-hosted sulphide occurrences are found inclose spatial association with noritic orpyroxenitic dykes, pockets or lenses withinthe anorthosites. They are dominated byan assemblage of pyrrhotite, chalcopyrite,pentlandite and pyrite in accordance witha magmatic origin. No hydrous silicateminerals are associated with any of thedeposits in the province.

A compilation of the known sulphideoccurrences in the Rogaland province ispresented in Table 2 and their locationsare shown in Fig. 1.

Dispersed sulphides are ubiquitous in vir-tually all mafic (jotunitic-noritic) intrusi-ons associated with the Rogaland anor-thosites. Oxide-rich layered intrusionssuch as the Storgangen and Bøstølenintrusions as well as the Tellnes dyke gene-rally contain 0.5-2 vol% sulphides, and Cuand Ni sulphides are recovered as a bipro-duct in ilmenite beneficiation. In theBjerkreim-Sokndal Layered Intrusion sul-phides are found as dispersed accessoryconstituents throughout the stratigraphy,but they also appear in a thin stratiformenrichment, more than 30 km in lateralextent (Fig. 1, #100). The sulphides arehosted by an up to 3 m-thick layer oforthopyroxenite or melanorite located atthe boundary between MCU II and III(Fig. 2). The sulphides are interstitial andthe sulphide content is up to about 9%. Itsformation has been attributed to a mixingevent between a bouyant primitive mag-

Fig. 3. Dimension stone quarries andoccurrences of anorthosite with labrado-rescent plagioclase in the western partof the RAP. Anorthositic rocks are shownin blue, jotunitic rocks in green, basalticdykes in red and the surrounding gneis-ses in pink and brown (see Marker et al.(2003) for details). Above right is showna polished surface of typical coarse-grai-ned anorthosite with iridescent plagio-clase.


ma and a more evolved resident magma,both saturated in sulphide. The resultinghybrid magma crystallized an assemblagedominated by orthopyroxene and ilmeni-te which sank to the magma chamberfloor as dense plumes containing exsolvedsulphide melt (Jensen et al. 2003).

The sulphide paragenesis is similar to thatobserved in the anorthosite-hosted depo-sits. The sulphide assemblage is domina-ted by pyrrhotite and euhedral pyritegrains are often embedded in chalcopyri-te. As at Homsevatnet, violaritization ofprecursor pentlandite is extensive and inseveral samples pentlandite has been com-pletely replaced. 'Birds eye' replacementsof pyrrhotite by marcasite-pyrite aggrega-tes are common.

Sulphides contained in other mafic, domi-nantly noritic intrusions are different onlyin modal amounts and the extent of repla-cement and alteration. The sulphideassemblage is always found to be pyrrhoti-te, pentlandite, chalcopyrite and com-monly, but not invariably, pyrite. In seve-ral intrusions, including the Bjerkreim-Sokndal Layered Intrusion, violaritizationmay be accompanied by a more severealteration of pyrrhotite to fine-grainedaggregates of euhedral pyrite and magne-tite. In intrusions such as Tellnes, Stor-gangen or Bøstølen, this reaction has crea-ted a sulphide paragenesis commonlydominated by fine-grained pyrite andlocally almost devoid of pyrrhotite. Be-cause the pyrite-magnetite aggregateshave never been observed in pentlandite-bearing samples and only in violaritizedsamples, it is inferred that the breakdownof pyrrhotite succeeds the breakdown ofpentlandite.

In parts of Tellnes and locally in Stor-gangen, millerite is the nickel-bearing sul-phide phase. In these cases millerite isinvariably closely associated with grains ofsiegenite as the cobaltian phase. Becausemillerite-bearing samples are devoid ofpyrrhotite (this having been replaced bypyrite and magnetite), the millerite-siege-nite assemblage is considered part of alate-stage, low-temperature (non-super-gene) paragenesis. The occurrence of mil-lerite as the nickel-bearing sulphide phasein Tellnes results in a high Ni content inthe recovered sulphide concentrate.

Recalculated to 100% sulphides, theRogaland occurrences contain up to 2.7%Cu and 4.2% Ni and no more than 0.5%Co. Noble metal concentrations are verylow, generally below 10 ppb. Pt concentra-tions are lower than Pd, and common Ostenors (Os concentration in the sulphidesat the time of formation) range from 0.06to 14.7 ppb (Schiellerup et al. 2000).

GenesisThe sulphide assemblage and the absenceof hydrous mineral phases associated withthe sulphides strongly suggest a magmaticorigin for all the sulphide occurrences.Schiellerup et al. (2000) argued that theRogaland sulphide occurrences are isoto-pically indistinguishable from both thehosting anorthosites and the associatedmafic intrusions, and that a commonsource had to be invoked. As discussed bySchiellerup et al. (this volume), the nobleand chalcophile element variation alsosuggests a common heritage of sulphideshosted by anorthosites and mafic intrusi-ons, probably involving similar parentalmagmas. In their model all chalcophileelement variation may be explained bydifferent R-factors (R-factor: ratio betwe-en amounts of silicate melt in effectiveequilibrium with the sulphide melt) and afairly well-defined parental sulphide com-position. In particular, the low noble ele-ment contents suggest a derivation fromnoble element depleted magmas, and thatlow R-factor processes were at play in thegeneration of most anorthosite-hosteddeposits. This implies a more or less staticexsolution of a sulphide melt, and that theexsolution most likely resulted from thedifferentiation of the parental magma. Inthe mafic intrusions the fairly constantsulphide content in the cumulates indica-tes that continuous segregation from sul-phide-saturated melts took place and R-factors are therefore only slightly elevated.The main exception in terms of grade isthe Bjerkreim-Sokndal Layered Intrusion,where the stratiform sulphide enrichmentis related to a magma replenishment eventinvolving magma mixing (Jensen et al.2003).

Dimension stoneThe first exploitation of anorthosite fordimension stone within the RAP tookplace at the beginning of the 20th century.At that time quarrying was mainly aimed


at local housing. Industrial-scale produc-tion did not start until the mid 1990s,based on regional and detailed geologicalmapping carried out by the GeologicalSurvey of Norway in co-operation withthe local governments of the region(Heldal & Lund 1994). Dimension-stonequality anorthosite contains iridescentplagioclase (Bøggild intergrowths), dis-playing a green, yellow or blue play ofcolour (Fig. 3). This type of anorthosite ishighly attractive on the international mar-ket, and can be compared with the betterknown larvikite in the Permian Oslo-rift.

The attractive anorthosites occur predo-

minantly in the central part of theEgersund-Ogna Massif, in the Hellvikarea. In addition to being enriched in iri-descent feldspar, these anorthosites con-tain megacrysts of orthopyroxene. Thedeposits occur as scattered, irregularblocks, surrounded by megacryst-pooranorthosite containing a less calcic plagio-clase. The clustering of the more maficfragments in the central and thus lowerpart of the dome-shaped Egersund-OgnaMassif may indicate that their origin rela-tes to the earliest stage of crystallization ofthe massif. At a later stage of the magma-tic evolution, these early-crystallized rockshave been disrupted and emplaced during

Fig. 4. Map of the occurrence of whitealtered anorthosite in the RAP, with thelocations of current hard-rock aggregateproduction indicated. Inset map shows the Lædre-Mong alteration zone in detail.


the diapiric uprise of the partially consoli-dated magma.

The highly irregular nature of the occur-rences makes volumetric interpretationand reserve estimates difficult withoutcoring and trial extraction. In addition tothe size of the deposits, the market value,and thus the exploitability, depends onseveral factors; it must be possible to pro-duce large blocks (lack of penetratingfracture systems), hydrothermal alterationthat causes discolouring of the rockshould not be present, the colour andstructure should be homogeneous (pre-dictable high quality) and finally the geo-metry of the deposits must be suitable foran efficient quarry layout.

In recent years, detailed mapping has beencarried out of the valuable anorthositetypes and has defined a large number ofoccurrences within five to six principalareas (Fig. 3). Two companies are nowproducing high-value blocks in the area,and it is expected that production will ste-adily increase in the future. The currentexport of anorthosite blocks from the twoquarries amounts to around 6.5 million _per year. At the present time, an effort isbeing made to secure the anorthositeresources for the future by integrating theresource maps into land use planning.

Hard-rock aggregateAggregate exploitation in Rogaland istaking place in zones of altered anorthosi-te, leuconorite or jotunite. The Rogalandprovince contains several zones in whichthe primary purple-brown anorthositehas been partly or completely altered byhydrothermal solutions to white anortho-site, which is attractive for the hard-rockaggregate industry. Alteration tends to fol-low certain prominent topographic linea-ments and affects all rock types crossed bythe lineaments. However, it is primarilythe altered anorthosite and leuconoritewhich is attractive because of the impro-ved mechanical properties (in particularthe wearing and polishing properties)combined with pale greyish to whitecolours and lack of quartz. The degree ofalteration varies within the zones, buteven partial alteration improves themechanical properties considerably. Mostattractive is completely-altered whiteanorthosite without mafic minerals,

which also has a potential for more valu-able industrial products such as fillers. Anadditional advantage for the aggregateindustry is the localization in the coastalzone with good discharging possibilities.

Anorthosite and leuconorite occupy largeareas of the western and southern parts ofthe RAP. Recent mapping has shown thatwithin this area, altered white anorthositeoccurs in six principal zones that are sepa-rated by large areas of primary anorthositewith insignificant alteration (Fig. 4). All ofthe main zones are steeply dipping andhave a NE-SW orientation following mar-ked valleys in the terrain. Associated withthe main zones there are subordinatealteration zones, some with a different ori-entation making the internal structure ofthe zones very complex. There is a tenden-cy that the extent and intensity of thealteration are largest where subordinatezones join or cross a main zone. Fromsoutheast to northwest, the six main zonesare: The Åna-Sira, Rekefjord, Lædre,Hellvik, Ogna and Brusand zones (Fig. 4).Extraction currently takes place in theRekefjord and Hellvik zones with a totalannual extraction of 1.3 million tons(2001) of which most is exported. Theproduction value amounts to around 8million € per year.

The white anorthosite in the alterationzones was formed by hydrothermal activi-ty at a time when the rocks in the RAPwere solidified and the structure establis-hed. This is evident because the originaltextures and structures are preserved inthe hydrothermally altered anorthosites.No destruction of the original texturesand structures is observed in the steep,cross-cutting alteration zones showingthat alteration was not accompanied bydisplacement.

As shown in Fig. 4, the principal alterationzones generally trend NE-SW. Anotherimportant trend is NW-SE, as in theLædre zone, while N-S trends are less pro-minent. These trends coincide with themain joint directions that are observed inthe province. Because of the lack of evi-dence for movement, it is likely that thehydrothermal alteration propagated alongzones of weakness in the solidified anor-thosite body and was later overprinted byjointing in the same directions. The alter-ation and introduction of hydrothermal


fluids must be older than the basalticdykes of the 616 Ma old Egersund dykeswarm (Bingen at al. 1998) that cut thealtered anorthosites in the Lædre area butare themselves unaffected.

During the hydrothermal alteration theplagioclase is albitized with formation ofepidote, clinozoisite and small amounts ofwhite mica (sericite). These secondaryminerals form a fine-grained aggregatewithin the outlines of the original plagio-clases. The primary orthopyroxene in theunaltered anorthosite is transformed intoamphibole and eventually decomposedinto chlorite and epidote/clinozoisite inthe highly altered anorthosite. At the sametime ilmenite is progressively transformedinto rutile, or further into titanite.

The altered white anorthosites show mar-kedly improved mechanical propertiescompared to the primary purple-brown

anorthosite. The standard test methodsshow that the white anorthosites have pro-perties that fall into the best class of hard-rock aggregates (Erichsen & Marker2000). White anorthosites are being quar-ried for aggregate in the Hellvik andRekefjord zones, but the aggregate poten-tial is much larger in the area. This con-cerns not least the, as yet, unexploitedLædre zone (Fig. 4), where reserves ofwhite anorthosite are among the largest inthe Rogaland Anorthosite Province.

KaolinAn occurrence of kaolin was discoveredon the farmland at Dydland, southeast ofSokndal, around 1880, and in 1898 a newcompany, Den Norske Chamottefabrik,acquired the rights to exploit the occur-rence. The kaolin is located in a narroweast-west striking gully and, according toReusch (1900), can be followed for alength of about 1 km with a width ofapproximately 10-20 m. JHL Vogt (inCarlson 1945) and H. Reusch (1900) asso-ciated the kaolinitization with one of thebasaltic Egersund dykes striking parallelwith the occurrence, and interpreted thealteration as a result of hydrothermal acti-vity along the contacts between dyke andanorthosite. The spatial association withthe ESE-WNW striking basalt dykesimplies that the kaolinitization is not rela-ted to the alteration that resulted in 'whiteanorthosite'.

The main site of exploitation was about 20m long, 5 m wide and up to 3 m deep.Today the deposit is exhausted and thekaolinitization mainly evident throughthe occurrence of a light coloured soil.

ConclusionsThe Rogaland Anorthosite Province inSouthwest Norway contains some of thelargest mineral resources in Norway withon-going mining of ilmenite, dimensionstone and aggregate. In all, the three typesof geo-resources represent a yearly reve-nue exceeding 75 million . Active explo-ration has located several additional pro-spects of aggregate and dimension stone,and alternative titanium resources may, inthe future, be mined in conjunction withapatite and vanadian magnetite. A tita-nium resource in the Storgangen intrusi-on has been abandoned for several deca-

Deposit Major phases Significanceilm plag opx mt ap cpx

Pramsknuten/Jerneld x Minor

Svånes x x Minor

St. Ålgård x x Minor

Florklev x x (x) Minor

Blåfjell x (x) (x) Significant

Vardåsen x x x Minor

Raunsli x x x Minor

Frøytlog x x x Minor

Beinskinnvatnet x x x Minor

Årstad x x x Minor

Storgangen x x x x Major

Bøstølen x x x x Minor

Tellnes x x x (x) Major

Laksedal x x x (x) Significant

Koldal field x x x (x) Minor

Vatland x x x (x) Minor

Fisketjern x x x x Minor

Kydlandsvatnet field x x x (x) (x) Minor

Eigerøy x x x Minor

Hestnes x x x Minor

Kaknuden x x x x x Minor

Rødemyr x x x x x Minor

Drageland x x x x x x Minor

Hauge x x x x x x Significant

Bakka x x x x x x Significant

Åsen x x x x x x Significant

Tekse-Bilstad x x x x x x Significant

Table 1. Key ilmenite deposits in the RAPaccording to phases present and signifi-cance. Partly based on Duchesne (1999).


des but still contains large reserves.Historically, small-scale mining for Ni andCu sulphides as well as for kaolin, has beenof local importance.

Rogaland, as a classic massif-type anortho-site province, thus provides the completerange of resource types usually associatedwith this type of setting. The ilmeniteresources are primarily controlled bycumulate processes and emplacementmechanisms, and to a large extent have pre-dictable compositions. The occurrence ofhigh-quality dimension stone with irides-cent plagioclase is related to the compositi-on of the plagioclase and to the structuralhabit of the host anorthosite. The locationof desirable aggregate prospects andextraction sites, as well as of kaolin, are

controlled by zones of alteration alongmajor fracture zones or dykes. Today, thecurrent extraction, exploration and futureprospects rank the Rogaland AnorthositeProvince as one of the foremost mineralresource provinces in Norway.

AcknowledgementsHelpful reviews were provided by BrianRobins and Jean-Clair Duchesne.

ReferencesBingen, B., Demaiffe, D. & van Breemen, O. 1998:The 616 Ma old Egersund basaltic dyke swarm,SW Norway, and late Neoproterozoic openingof the Iapetus ocean. Journal of Geology 106,565-574.

Host unit Occurrence Characteristics Host rock

Egersund-Ogna Homsevatnet Massive dyke-like or lenticular body Norite pegmatite.

dominated by po.

Høgatuva1 Coarse-grained sulphides in irregular Not known.

(Ualand) lenses or laminae up to 3 m long.

Fossåsen1 Sulphide impregnation. Not known.

(Ualand)

Tunggårdsbakken1 Irregular lens or dyke up to 1.60 m wide Not known.

(Ualand) containing coarse-grained sulphides.

Bjørndalsnipa Three cm/dm-sized zones of semi-massive Mineralizations hosted by anorthosite,

po-dominated ore, surrounded by cp- but pegmatoid norite is present in

dominated veins or disseminations. cored secton.

Myklebust Disseminated sulphides and irregular Host rock mostly noritic or pyroxenitic.

veins dominated by po.

Fossfjellet Dyke or vein type po-dominated ore Strongly foliated norites and anorthosites

surrounded by cp-dominated disseminated in the deformed margin of the anorthosite.

sulphides.

Piggstein2 Disseminated po-dominated sulphides. Mafic pyroxenitic (opx+cpx) part of

noritic lens or dyke.

Helland3 Contains Ni-pyrrhotite. Details not known. Not known.

Hellvik3 Contains po-pn, cp, py. Details not known. Not known.

Skora3 Contains Ni-pyrrhotite. Details not known. Not known.

Håland-Helleren Urdal Dyke-like sulphide body consisting of a Main mineralization hosted by

central massive po-dominated zone surroun-anorthosite, but located within metres

ded by veinlets of sulphides richer in cp. of a norite dyke.

Koltjørn3 Contains pyrrhotite (Ni). Details not known. Not known.

Grastveit3 Contains Ni-pyrrhotite. Details not known. Not known.

Bjerkreim-lobe Rudleknuten4 Stratiform disseminated sulphide Orthopyroxenite or melanorite.

concentration in layered cumulates of the

Bjerkreim-Sokndal intrusion.

Table 2. Compilation of known anorthosite-hosted sulphide deposits in the RAP. 1) Stadheim (1938). 2) Henriette (1986). 3) Geological Survey of Norway,'Bergarkivet'. 4) Jensen et al. (2003).The Bjerkreim-lobe is the northern lobe of the Bjerkreim-Sokndal Layered Intrusion. See Fig. 1 for locations.


Bolle, O., Diot, H. & Duchesne, J.C. 1997:Anisotropie de la susceptibilité magnetiquedans une intrusion composite de la suite char-nockitique: l'apophyse du massif stratiformede Bjerkreim-Sokndal (Rogaland, Norvègeméridionale). C.R. Acad. Sci. Paris 325, 799-805.

Carlson, A. 1945: Malmforekomster, Sokndal-Egersund området. Internal report, A/S Titania.44 pp.

Dahll, T. 1863: Om de geologiske undersøgelseri Kristiansands stift. Förhandlinger vidNaturforskermødet i Stockholm 1863, 269-276.

Duchesne, J.C. 1972: Iron-titanium oxide mine-rals in the Bjerkrem-Sogndal Massif, south-western Norway. Journal of Petrology 13, 57-81.

Duchesne, J.C. 1987: The Bjerkreim-Sokndalmassif. In: Maijer, C. & Padget, P. The geology ofsouthernmost Norway. Norges geologiskeundersøkelse Special Publication 1, 56-59.

Duchesne, J.C. 1996: Liquid ilmenite or liquidusilmenite: a comment on the nature of ilmenitevein deposits. In: Demaiffe, D. (ed) Petrologyand geochemistry of magmatic suites of rocks in the continental and oceanic crusts. A volumededicated to Professor Jean Michot. UniversitéLibre de Bruxelles, 73-82.

Duchesne, J.C. 1999: Fe-Ti deposits in Rogalandanorthosites (South Norway): geochemicalcharacteristics and problems of interpretation.Mineraleum Deposita 34, 182-198.

Duchesne, J.C., Maquil, R. & Demaiffe, D. 1985:The Rogaland anorthosites: facts and specula-tions. In: Toby, A.C. & Touret, J.L.R. (eds.) The deepProterozoic crust in the North Atlantic provinces.NATO Adv. Stud. Inst., Reidel, Dordrecht, C158,449-476.

Duchesne, J.C. & Maquil, R. 1987: The Egersund-Ogna massif. In: Maijer, C. & Padget, P. The geolo-gy of southernmost Norway. Norges geologiskeundersøkelse Special Publication 1, 50-56.

Duchesne, J.C. & Hertogen, J. 1988: Le magmaparental du lopolithe de Bjerkreim-Sokndal(Norvège méridionale). Les Comptes Rendus de l'Academie des Sciences Paris 306, 45-48.

Duchesne, J.C. & Wilmart, E. 1997: Igneous char-nockites and related rocks from the Bjerkreim-Sokndal layered intrusion (southwest Norway):a jotunite (hypersthene monzodiorite)-derivedA-type granitoid suite. Journal of Petrology 38,

337-369.

Duchesne, J.C. & Schiellerup, H. 2001: TheRogaland iron-titanium deposits. In: J.C.Duchesne (ed.) The Rogaland Intrusive Massifs –an excursion guide. Norges geologiske under-søkelse Report 2001.029, 56-69.

Erichsen, E. & Marker, M. 2000: Anortositt i"Egersundfeltet" – Pukkpotensialet. Norges geologiske undersøkelse Report 2000.016.

Hagen, R. 1992: Titania – 90 år med ilmenitt.Geonytt 4/92, 12-14.

Heldal, T. & Lund, B. 1995: A regional study ofthe dimension-stone potential in labradorite-bearing anorthositic rocks in the RogalandIgneous Complex. Norges geologiske undersø-kelse Bulletin 427, 123-126.

Henriette, A. 1986: Les indices de sulfures dans le massif anorthositique d'Egersund-Ogna(Rogaland - Norvège méridionale). Unpublishedthesis. Université de Liège. 125 pp.

Hovland, R. 1975: Nikkelprospektering i Egersund- anorthositfelt. Malmgeologisk symposiumBVLI Bergforskningen, referat fra møte iTrondheim 22-23 april 1975, "Nikkelpros-pektering i Norge", 94-99.

Hubaux, A. 1960: Les gisements de fer titané dela région d’Egersund, Norvège. Neues Jahrbuchfür Mineralogie Abhandlungen 94, 926-992.

Jensen, J.C., Nielsen, F.M., Duchesne, J.C.,Demaiffe, D. & Wilson, J.R. 1993: Magma influxand mixing in the Bjerkreim-Sokndal layeredintrusion, South Norway: evidence from theboundary between two megacyclic units atStoreknuten. Lithos 29, 311-325.

Jensen, K.K., Wilson, J.R., Robins, B. & Chiodoni, F.2003: A sulphide-bearing orthopyroxenitelayer in the Bjerkreim-Sokndal Intrusion,Norway: implications for processes duringmagma-chamber replenishment. Lithos 67,15-37.

Kolderup, C.F. 1896: Die Labradorfelse deswestlichen Norwegens. I. Das Labrador-felsgebiet bei Ekersund und Soggendal.Bergens Museums Aarbog 1896 V. 224 pp.

Korneliussen, A., McEnroe, S.A., Nilsson, L.P.,Schiellerup, H., Gautneb, H., Meyer, G.B. &Størseth, L.R. 2000: An overview of titaniumdeposits in Norway. Norges geologiske


undersøkelse Bulletin 436, 27-38.

Krause, H. & Zeino-Mahmalat, R. 1970:Untersuchungen an Erz und Nebengesteinder Grube Blåfjell in SW-Norwegen. NorskGeologisk Tidsskrift 50, 45-88.

Krause, H. & Pedall, G. 1980: Fe-Ti-mineralisati-ons in the Åna-Sira anorthosite, southernNorway. Geological Survey of Finland Bulletin307, 56-83.

Krause, H., Gierth, E. & Schott, W. 1985: Ti-Fedeposits in the South Rogaland IgneousComplex, with special reference to the Åna-Sira anorthosite massif. Norges geologiskeundersøkelse Bulletin 402, 25-37.

Kullerud, K. 1994: MgO-innhold i ilmenitt fraTellnesforekomsten. Unpublished report.Department of biology and geology,University of Tromsø, Norway. 14 pp.


Lundgaard, K.L., Robins, B., Tegner, C. & Wilson,J.R. 2002: Formation of hybrid cumulates: mela-troctolites in Intrusion 4 of the HonningsvågIntrusive Suite, northern Norway. Lithos 61, 1-19.

Marker, M., Schiellerup, H., Meyer, G.B., Robins, B.& Bolle, O. 2003: The Rogaland AnorthositeProvince. Geological map 1:75000. Norges geologiske undersøkelse Special Publication 9,109-116.

Meyer, G.B. & Wilson, J.R. 1999: Olivine-rich unitsin the Fongen-Hyllingen Intrusion, Norway:implications for magma chamber processes.Lithos 47, 157-179.

Meyer, G.B., Schiellerup, H. & Tegner, C. 2002:Chemical characterisation of ilmenite, magne-tite and apatite in the Bjerkreim-SokndalLayered Intrusion, South Norway. Norges geologiske undersøkelse Report 2002.042. 18 pp.

Michot, P. 1956: Les gisement des minerais noirde la région d’Egersund. Annales de la SociétéGéologique de Belgique 79, 183-202.

Michot, P. 1960: La géologie de la catazone:le problème des anorthosites, la paligenèsebasique et la tectonique catazonale dans leRogaland méridional (Norvège méridionale).

Norges geologiske undersøkelse 212g, 1-54.

Michot, P. 1965: Le magma plagioclasique.Geol. Rundschau 54, 956-976

Nielsen, F.M., Campbell, I.H., McCulloch, M. &Wilson, J.R. 1996: A strontium isotopic investi-gation of the Bjerkreim-Sokndal layered intru-sion, Southwest Norway. Journal of Petrology37, 171-193.

Reusch, H.H. 1900: En forekomst af kaolin ogildfast ler ved Dydland nær Flekkefjord. Norgesgeologiske undersøkelse 32, 99-112.

Robins, B. & Wilson, J.R. 2001: The Bjerkreim-Sokndal Layered Intrusion. In: Duchesne, J.C.(ed.) The Rogaland Intrusive Massifs – an excursi-on guide. Norges geologiske undersøkelseReport 2001.029, 35-47.

Robinson, P., Kullerud, K., Tegner, C., Robins, B. &McEnroe, S.A. 2003. Could the Tellnes ilmenitedeposit have been produced by in-situmagma mixing. Norges geologiske undersøkelseSpecial Publication 9, 74-76.

Roelandts, I. & Duchesne, J.C. 1979: Rare-earthelements in apatite from layered norites andiron-titanium oxide ore-bodies related to anor-thosites (Rogaland, S.W. Norway). In: Ahrens(ed.) Origin and distribution of the elements.Pergamon Press, Oxford, 199-212.

Schiellerup, H., Lambert, D.D., Prestvik, T., Robins,B., McBride, J.S. & Larsen, R.B. 2000: Re-Os isoto-pic evidence for a lower crustal origin of mas-sif-type anorthosites. Nature 405, 781-784.

Schiellerup, H., Meyer, G.B., Tegner, C., Robins, B.& Korneliussen , A. 2001: Resources of apatite,ilmenite and magnetite in the Bjerkreim-Sokndal Layered Intrusion, Rogaland, SouthNorway. Norges geologiske undersøkelse Report2001.092. 21 pp.

Schiellerup, H., Lambert, D.D. & Robins, B. 2003.Sulphides in the Rogaland AnorthositeProvince. Norges geologiske undersøkelse SpecialPublication 9, 58-60.

Schärer, U., Wilmart, E. & Duchesne, J.C. 1996:The short duration and anorogenic characterof anorthosite magmatism: U-Pb dating of the Rogaland complex, Norway. Earth andPlanetary Science Letters 139, 335-350.

Sigmond, E.M.O. 1992: Bedrock map of Norwayand adjacent ocean areas. Scale 1:3 million.


Norges geologiske undersøkelse.

Skjerlie, K.P., Kullerud, K. & Robins, B. 2003:Preliminary melting experiments on theTellnes ilmenite norite from 0.5 to 1.2 GPa,implications for the composition of intercu-mulus melt. Norges geologiske undersøkelseSpecial Publication 9, 77.

Stadheim, J. 1938: J. Stadheims dagbok (Uelandv/ Ogna og Tunggårdsbakken) 1938. Norgesgeologiske undersøkelse Report 2757. 3pp.

Tegner, C & Robins, B. 1996: Picrite sills andcrystal-melt reactions in the HonningsvågIntrusive Suite, northern Norway. MineralogicalMagazine 60, 53-66.

Vander Auwera, J. & Longhi, J. 1994:Experimental study of a jotunite (hypersthenemonzodiorite): constraints on the parentmagma composition and crystallization condi-tions (P, T, fO2) of the Bjerkreim-Sokndal laye-red intrusion. Contributions to Mineralogy andPetrology 118, 60-78.


Vander Auwera, J., Longhi, J. & Duchesne, J.C.2000: The effect of pressure on DSr (plag/melt)and DCr (opx/melt): implications for anorthosi-te petrogenesis. Earth and Planetary ScienceLetters 178, 303-314.

Vogt, J.H.L. 1910: Norges jernmalmforekomster.Norges geologiske undersøkelse 51, 132-138.

Wilmart, E., Demaiffe, D. & Duchesne, J.C. 1989:Geochemical constraints on the genesis of theTellnes ilmenite deposit, Southwest Norway.Economic Geology 84, 1047-1056.

Wilmart, E., Pineau, F., Réjou-Michel, A. &Duchesne, J.C. 1994: Fluid transfer in anorthosi-tes and related rocks from Rogaland (South-west Norway): Evidence from stable isotopes.Earth and Planetary Science Letters 125, 55-70.

Wilson, J.R., Robins, B., Nielsen, F.M., Duchesne,J.C. & Vander Auwera, J. 1996: The Bjerkreim-Sokndal layered intrusion, Southwest Norway.In: Cawthorn, R.G. (ed.) Layered intrusions,Elsevier Science B.V., Amsterdam, 231-256.

Ilmenite deposits and theirgeological environmentWith special reference to the Rogaland Anorthosite ProvinceIncluding a geological map at scale 1:75,000and a CD with a guide to the province

SPECIAL PUBLICATION 9

Leiv Eirikssons vei 39N-7491 Trondheim,Norway

Telephone +47 73 90 40 00Telefax +47 73 92 16 20

E-mail: [email protected]://www.ngu.no

Special Pu

blicatio

n n

o.9Ilm

enite d

epo

sits and

their g

eolo

gical en

viron

men

tN

GU

Geology for Society since 1858

The Rogaland Intrusive Massifs

An Excursion guide

A CD-presentation of theexcursion guidebook andadditional geological mapinformation,photographs and SEM images

Titanium and Iron-TitaniumDeposits in Norway

Approximately 300 titaniumand iron-titanium ore depositsare known in Norway; most ofthese occur in distinctProterozoic and Phanerozoicgeological provinces.

The Rogaland Anorthosite Province

(c.940-920 Ma) is highly important in termsof contained titanium and economicsignificance.

In a workshop in Rogaland in 2001,organised under the umbrella of theEuropean Science Foundation GEODE(Geodynamics and Ore Deposit Evolution)programme,geoscientists from manycountries came together to discuss variousaspects of the evolution of ilmenitedeposits and their host rocks.

This volume of the NGU Special Publicationseries is a follow-up of the 2001-conference,with articles and abstracts dealing with theRogaland Anorthosite Province as well aswith ilmenite deposits elsewhere in theWorld.Also included is a full-size geologicalmap of the province and a CD with ageological excursion guide.

NORGES GEOLOGISKE UNDERSØKELSESpecial Publication 9

ISSN:0801-5961ISBN:82-7385-108-7

RogalandAnorthosite Province

Ilmenite deposits and their geological environment publication/Spec...Ilmenite deposits and their...

Documents

Transcript of Ilmenite deposits and their geological environment publication/Spec...Ilmenite deposits and their...