The geochemistry of Lower Proterozoic mafic tofelsic igneous rocks, Rombak Window, North Nor-wayARE KORNELlUSSEN & EDWARD W. SAWYER

Korneliussen, A. & Sawyer, E.W. 1969: The geochemistry 01 Lower Proterozoic mal ic to tetsicigneous rocks, Rombak Window, North Norway. Nor. geol . unders. Buff. 415, 7-21.

The supracrustal sequence 01 the Rombak Basement Window, cons isting 01 volcanic rocks, peli-tic sediments , greywackes with minor amounts 01 carbonate rocks and quartz ites, was intrudedby mal ic dykes, malic to intermediate plutons and a variety 01 granitoid batholiths c. 1.6-1.7 Gaago. The region has experienced amphibolite-grade metamorphism, followed by retrogression togreensch ist facies along Caledonian shear-zones.

On the basis of their petrographic ano geochem ical characteristics the volcanic rocks can bedivided into 3 suites: (1) high-Mg basalts : (2) malic to leisic voicanites with lair ly high potass iumcontents and with calc-alkaline affinities ; and (3) low-potass ium, calc-alkaline felsic volcanites.

Based on major element geochemis try the evolution of the potas sic voicani tes is interpre ted tohave been controneo , in the case 01 malic -intermediate varieties, by early fractionation of Fe,Mg-rich minerals , and by plagioclase crystallisation for the lelsic varieties. Suites 2 and 3 are similarto associated granites and granodiorites in their chemical composit ion.

It is concluded that the volcano-sedimentary and intrusive rocks were formed in an Lower Prate-rozoi c mature magmatic arc environment at the southern margin of a continent composed predomi -nantly 01 Archaean tona litic granitoid rocks and Lower Proterozoic greenstone terranes .

Are Kornefiussen, Geological Survey of Norway. P.B. 3006 - Lade, N-7002 Trondneim, Norway.£dward Sawyer, Sciences de la Terre. Universite du Quebec a Chicoutimi, Chicoutimi, Quebec.G7H 2Bl , Canada.

IntroductionThe Rombak Basement Window is situatednear the southern margin of the ArchaeanDomain (Pharaoh & Pearce 1984, Ohlanderet al. 1987) of the Baltic Shield (Fig. 1), Thewindow conta ins Lower Proterozoic suprac-rust al sequences consisting of turbidites andmafic to felsic volcanites that have been in-truded by numerous, large, felsic to mafic plu-tons . The Proterozoic rocks of the windoware surrounded by the allochthonous Caledon i-an nappe comp lexes (Gustavson 1974 a & b,Tull et al. 1985), and locally by a thin sequenceof autochthonous sediments belonging to theLate Proterozoic to Cambrian Dividal Group(Vogt 1942, Gustavson 1974 a, Birkeland 1976).

On a regional scale, the Archaean of theBaltic Shield consists principally of felsic tointermediate, partly tonalitic gneisses withsubordinate greenstone belts (Witschard 1984,Gaal & Gorbatschev 1987, Ohlander et al.1987), In the earliest Proterozoic (c. 2.4 Ga)

the Archaean craton was fragmented by epi-sodes of rift ing, and greenstone terranes for -med by the submar ine erupt ion of large volu-mes of basaltic (and some komat iitic) magmain these rift s (Gaal & Gorbatsc hev 1987). Innorthern Sweden , supracrustal sequencessouth of both the Lower Proterozoic greensto-ne terranes and the Archaean craton are domi-nated by volcanites that show a cont inuouscompositional range from maf ic to felsic ty-pes, and that have ages between 1.9 Ga and1,8 Ga (Fritsch & Perdahl 1987).

The purpose of this paper is to describe thegeochem istry of volcanic rocks that are partof the Lower Proterozoic supracrustal sequen -ces exposed in the Rombak Window. Thecompositional characteristics of the volcanites ,and of spatially assoc iated plutonic rocks, arethen discussed in the context of an evolvingmagmatic arc located above a subductionzone that is postulated to have existed in theregion at some time between 1.9 and 1,7 Ga.

8 Are Korneliussen & Edward W.Sawyer GU - BULL. 415. 1989

r::::::lE2J

IEEEI

N

11

Caledonian and upper Proterozoicto Cambrian autochthonouscover success ions

Prote rozoic int rusionsmainly granites (17 50 - 1900 Ma )

Early Proterozo ic supracr ustalswith fels ic- intermediate volcanites

Early Prote rozoic supracrus alswith mafic volcanites

ITII] Lapland granulite beltD Archaean base mentSimplified from a tec onic map compiledby the geological surveys of Finland . Norwayand Sweden. Nordkalott project 1986.

Fig. 1. Major geolog ical units of the northern part of the Baltic Shield in Norway. Sweden and Finland. Simplif ied from atectonic map compiled by the geolog ical surveys of Finland. Sweden and Norway. Nordka lott Project 1986.

Geologic setting of the RombakBasement Window

Age relationsAt presen t, very few rocks in the RombakWindow have been dated . Romer (this vol-ume) has obta ined an age of 2.3 Ga (Rb-Sr)

for a suite of high-Mg . low-K,O basalts in theRuvssot-Sjangeli area. The relationship betwe-en the Ruvssot-Sjangeli supracrustal belt andthe other belts in the western part of the win-dow is not clear because the two regions areseparated by a major , N-S-trending shearzone that is well exposed at Muohtaguobla

NGU • BULL. 415. 1989 The geochemistry of Lower Proterozic rocks 9

S '

S

ROMBAK WI NDOW

D M aflc --: intermediatevctc am t eso Felsic vorcamt esr:'J Pyroclastic rock s.~ undlllerentia led

D Gramte

Fig. 2. Generalized geological map of the Aombak Windowbased on Sawyer & Korne liussen (this volume). Locat ionsmentioned in the text: S - se roai. G - Gautelis, TB -tonalitic basement, N - Norddal, SH - Stasjonsholmen , M- Muohtaguobla, MTZ - Muotaguobla Tectonic Zone, AS- Ruvssot-S jangeli. K - Klubbvatnet , R - Rombaksbotn,C - Cainhavarre.

Fig. 3. Volcanic and sedimentary units of the Serda lenSupracrustal Belt.

(Fig. 2). However, by analogy with volcanicrocks of similar compos ition and textur e fromdated supracrustal sequences of northernSweden (Fritsch & Perdahl 1987, Widenfalket al. 1987), supracrustal belts west of theMuohtaguobla Tectonic Zone probab ly haveages between 1.91 and 1.88 Ga. All the suprac-rustal sequences of the Rombak BasementWindow have been extens ively intruded bylarge plutons cons isting predom inant ly of gra-nite, but also including syenite, dior ite andgabbro . Granites have been dated at 1.78 and1.69 Ga (Rb-Sr) by Gunner (1981) and Heier& Compston (1969), respectively.

LithologyA distinct feature of the Rombak Window isthe pattern of N-S trending linear supracrustalbelts (fig. 2) preserved between extensive regi-ons of younger plutonic rocks (Vogt 1942,Gustavson 1974a & b, Birkeland 1976, Robynet al. 1985, Korneliussen et al. 1986 a & b).Small rafts and inclusions of the supracrustalrocks are locally abundant in the plutons . Allof the supracrustal rocks and the Early Protero-zoic pluton ic rocks of the Rombak BasementWindow are metamorphosed at least underPT-cond itions of the lower amphibolite facies.

The rocks within the window are variablydeformed and show a generally N-S-trending,more-or-less vert ical fol iation. The contactsbetween the supracrusta l belts and the sur-round ing gran ites are commonly sheared . Wit-hin the supracrustals, practically undeformedvolcan ic and sedimentary rocks with well -preserved primary textures are common.

The rock types present, and their relativeproportions, vary cons iderably from one sup-racrustal belt to the next across the RombakWindow . The Serda len Supracrustal Belt in thesouthwestern part of the window (Fig. 2) iscomposed mainly of predominantly porphyr itic,mafic, intermediate and felsic volcanites. Seve-ral units of mafic/intermediate amygdaloidalvolcanites together with felsic volcanites havebeen identified (Fig. 3); locally, thin units ofsediment separate distinct, mappable volcanicunits. Debris flows are interbedded with theflows, particularly on the southern side of thebelt. Clast size in the debris flows varies fromunder 1 dm to 0.5 m. and indicates a high-energy environment of deposition. The lower-most (eastern) telsic volcanite unit in the S0r-

10 Are Korneliussen & Edward W. Sawyer

dalen Supracrustal Belt is K-feldspar-bearingand closely resembles volcanites at Cain-havarre.

The Stasjonsholmen-Rombak SupracrustalBelt conta ins a thick sequence of graded pe-lite-greywacke turb idites, with tuff itic layers inplaces. Amygdaloidal lavas with associateddebris flows are developed at Klubbvatnet inthe central to northern part of the Stasjonshol-men-Rombak Supracrustal Belt (Robyn et al.1985).

In the Muohtag uobla area mafic and interme-diate lavas (containing acicular plagioclasephenocrysts), telsic tuff s, pelites and graphiticschists, are complexly intermixed with cross-bedded quartz ites and conglomerates belong-ing to the Dividal Group . The comp lexity ofoutcrop pattern in this area is of tecton ic ori-gin (Romer & Boundy 1988), since the regionprobab ly represents a Caledonian imbricationzone (terminology of Butler 1982) within theRombak Window.

In the eastern part of the window the Ruvs-sot-Sjangeli supracrustal sequence containsmafic and ultramafic volcanites, fine-grainedbiotite schists , greywackes and silicate-bandedcarbonates (Romer 1988), and generally res-embles a greensto ne assoc iation. The maficlultramafic volcanic rocks occur as amphiboli-tes (locally pillowed) and serpentinites, someof which contain up to 28% MgO.

At Gautelis (fig. 2) the supracrustal sequen-ce is dominated by a turb idite sequence, butthin hor izons of tuff itic mafic and felsic vol-canites , cong lomerates and debris flow s arelocally developed (Skonseng 1985). Pebblesin the scattered conglomeratic horizons con-sist of fine- to coarse-grained tonalite andgranodiorite that resemble a nearby body oftonalite (called the Gautelis Tonalite Comp-lex). The status of this complex is important,as it might represent older (perhaps Archae-an) basement. It is overlain by a basal cong lo-merate conta ining clasts derived from the tona-lite, and a dolomitic carbonate indicating plat-form sedimentat ion, followed by the turb iditesequence.

The individual volcanic units within the tur-biditic pelites and greywackes in different partsof th e window range in t h ickness from a few

centimetres to approximately 10 m, and arein general tuffitic. In contrast, the thick vol-canite (up to 1 km) success ions are dominant-ly lava flows. This is clearly indicated by thepresence of amygdules in some cases (Klubb-

GU . BULl. 4 5. 1989

vatnet and Serdalen), and of delicate needle-shaped plagioclase phenocrysts (Muohtaguob-la and Serdalen) in others. Flow structuresare preserved in some rhyolitic flows from theStasjonsholmen area. Our interpretation isthat the volcanites were erupted adjacent toa deep basin that was periodically receivingturbidite flows. Explosive volcanic eruptionsformed ash which spread out over a largearea. Where waterlain, the ash formed tuffitichorizons intercalated with the turbidites. Adominance of felsic over mafic volcanic pebb-les in the debris flows in Serdalen may indi-cate a larger volume of felsic volcanic materi-al near to the volcanic centres.

The oldest intrusive rocks known in theRombak Basement Window are those of themedium- to coarse-grained Gautelis TonaliteComplex. The Gautelis Tonalite Complex andthe over lying conglomerate, dolomitic car-bonate and turbidite sequence are intruded bya swarm of mafic dykes. These are in turnintruded by the numerous large plutons datedat about 1.78 Ga (Rb/Sr) by Gunner (1981).Minor mafic to felsic dykes cut the plutonsand are of unknown age, although Gunner(1981) presents some evidence that they maybe 1.3 Ga old (Rb/Sr).

MetamorphismThe Rombak window, at least in its central,western and southwestern parts , has beenmetamorp hosed under amphibolite facies con-ditions (P 6kb, T 575°C; Sawyer 1986). Eviden-ce for this is the widely preserved progrademineral zonation patterns found in the interme-diate and mafic volcanites. The age of thisprograde metamorphism has not been clearlyestab lished, but is probably Lower Proterozo-ic. A greenschist-facies metamorphism hasoverprinted the rock s of the window to varyingdegrees ; in most places its effects are minor ,or even absent. However, in the Muohtaguob-la area the greenschist-facies metamorphismhas virtually obliterated all evidence of theearlier higher temperature event. The intensityo f re t ro g re s s ion in the Muohta g uob la a rea is

spat ially related to the Caledonian defo rmat ionthat has imbricated Lower Proterozoic andDividal rocks (ct. Romer & Boundy 1988);hence the greenschist-facies metamorphismis likely to be of Caledonian age.

NGU • BULL. 415. 1989 The geochemistry ofLower Proterozic rocks 11

Table 1 (b). Major and trace element abundances inselected tetslc volcanic rocks.

MuOhUg. (SN) Ca1nhav. (SN) StasJonsh. iSH) Gautel1s (G)

Sample K301.3 K302.3 K2S'.3 KID'., KlOl.' K269.3 Kl01.5 Kl03.5

od • not detected; -. not determ1ned

GeochemistryRepresentative major and trace element ana-lyses of extrusive and intrusive rocks from theRombak Basement Window are given in Table1. The major oxides were determined by XRFusing fused glass beads. The trace elementsV to Nb were determined by XRF using pres-sed powder pellets. The rare earths (REE),and Cs, Th, U, Ta and Hf were determinedby instrumental neutron activation analysis. Acomplete list of all analysis is available fromA. Korneliussen on request. For the elementvariation diagrams presented below, analysesare recalculated to an anhydrous basis.

Alteration processes involving relativelymobile elements such as Na20 and K20, can-not be excluded. Analyses of rocks from shearzones in sardaten indicate some mobility ofcertain elements, but this appears to be relati-vely minor (these results are not included inthis paper). There is a fairly good consistencybetween plots presented below involving ele-ments which are generally accepted to beamong the least mobile, Le. Th, Hf, Ta, Nb,Y, Zr, Ti and the REE. This indicates an insigni-ficant degree of element mobility during altera-

S102Al203Fe203"nO"90CaDHa20K20Tl02P205i.i ,SUM

VSeceerNIeuInPbRbs-BalrVNb

csThUTaHf

taeeNdSmEuTbVbtu

63.601B.'1z.zz

.06

.582.055.506.23.B'

.251.29

101.53

186

ndndndnd'0nd

107206

11002697

.76...ndnd

.6'2138173.702.80

•• 1.65.10

61.7818.5'3.12

.06

.7'2.0B5.106.11

.79

.2B

.6599.25

26ndndndndnd.01371

3802300

299

nd

203B183.303.20

.39

.66

.11

69.591'.233.82

.06

.371.573.105.76

.53

.12

.6399.78

296

nd13

713.720

2261B6B96322.516

639B'610.701.201.20'.50

.6B

68,'813.934.94

.07

.761.'64.104.99

.56

.10

.75100.14

3088

139

6.6223

22B116822337

5117

3.1021.807.741.'58.60

61.20132.0052.508.841.101.204.19

.71

71.8711.192.80

.02

.15

.672.504.80

.17

.01•• 5

100.62

ndndndndndnd3426

2591769.

8287935

'.7125.70

7.302.19

17.50

86.00191.0073.301•• 30

.12I.B'6.631.0'

76.3511.653.2B

.03

.05

.351.606.83.23.01.24

100.61

ndndndndndnd5238

3007446

6178332

1352209218.30

.292.107.601.18

75.7013.011.27

.02

.291.366.171.03

.22

.03

.5399.62

13•10nd237

1225

290B861891717

.3510.BO'.901.625.05

'2.607B.5025.904.04

.7B

.5B2.25

.32

n.n13.B7

1.75.04

1.441.924.702.1.

.26

.05

.9'99.84

2883462

221169

3331356

1BO1918

1.5610.102.192.0'' ...

3'.906B.5023.'0

3.49.78•• 2

1.63.23

RUV$sot (RS) sereeten (5"') Maflclntruslons ~~Sample RI.] R22.3 .1 ., .3 •• " .5 AR' ~~~ ~~Sample 1:511.J 11:75.4 U33.4 1C152.3 1:273.3 1:268.3 ARBI 1C55.] K5J6.J UotO.S !Cl4].55102 48.57 45.55 49.57 54.25 58.83 57.94 54.99 49.99 54.26

50.50 76.76A1ZO) 9.88 7.49 11.92 14.10 16.18 16.38 17.80 17.91 15.72 5102 46.59 47.81 48.38 47.79 55.27 "9.72 71.2" 67.29 71.9"Fe203 13.04 10.63 9.39 7.65 7.35 6.30 7.92 9.64 8.0" ,t,1203 16.03 16.U 16.32 13.53 14.29 lot.94 15.25 12.47 13.93 14.91 15.27.'0 .09 .16 .IS .15 .13 .11 .08 .13 .13 Fe203 12.63 12.12 12.36 14.71 10.83 12.19 12.47 1.36 3.54 4.47 1.31.,0 7.75 20.56 11.19 6.19 3.20 1.94 1.74 3.75 4.67 ,"0 .25 ... • 17 •20 .r , .re .rs .01 .0' .OS .04",0 6.48 6.89 6.22 5.58 4.26 6.25 5.95 .OB .40 1.54 .42C.O 6.86 8.85 9.47 6.18 5.35 5.06 4.48 7.83 6.40 cec 8.47 7.93 9.67 9.21 6.15 8.8B B.39 .72 1.06 3.60 1.74Na20 ".10 .50 2.63 4.05 3.59 4.60 4.90 3.97 3.96 Ha20 3.20 3.40 2.30 2.70 2.70 2.50 2.80 3.30 2.50 4.50 6.59

12 Are Korneliussen & Edward W. Sawyer NGU-BULL.415.19a9

tion as far as these elements are concerned.For the plots involving the more mobile ele-ments Na20 and K20 some scatter caused byalteration is likely to occur, though it is as-sumed that the igneous trend in these plotsis real since the interpretation of the majorand trace element plots is relatively consistent.

Extrusive rocksMajor elements: A plot of (Na20 + K20) versusSi02 (Fig. 4) for the volcanites of the Rombaksupracrustal belts shows that the 2.3 Ga Ruvs-sot-Sjangeli volcanites are more mafic andcontain less alkalis than volcanites from westof the Muohtaguobla Tectonic Zone. Threeof the Ruvssot-Sjangeli samples clearly repre-

* Ruvssot-Sjangelio Gautelis+ Serdalen (mafic-interm.lx Serdalen (f elsicl• Stasjonsholmen• Rombaken... Muohtaguobla• Cainhavarre

sent liquid compositions (2 samples with >28%MgO are probably cumulates) and are subalka-line. In contrast, the mafic and intermediatevolcanites from the serdaren, Muohtaguoblaand Rombak areas plot across the boundarybetween the alkaline and subalkaline fields.For rocks with >66 % Si02 the (Na20 + KP)versus Si02 plot is not a useful means of dis-tinguishing between alkaline and subalkalineseries. However, Fig. 4 shows that the vol-canites from west of the Muohtaguobla Tecto-nic Zone, Le. the Serdalen mafic-intermediateand felsic volcanites and the Stasjonsholmen,Cainhavarre and Muohtaguobla felsic volcani-tes in the Norddal area, form a continuousrange in Si02 contents from 50 to 78 %, witha preponderance of andesitic compositions.

The Na20 versus K20 plot (Fig. 5) illustratesthree important compositional differences with-in the Rombak volcanites: (a) The Ruvssot-Sjangeli extrusives are K20-deficient and havevariable, but low, Na20 contents; (b) the Gaute-lis felsic volcanites from within the GautelisTonalite Complex have a higher Na20/K20 rati-os than the other volcanites from west of theMuohtaguobla Tectonic Zone; and (c) withinthe Sordalen, Stasjonsholmen, MuohtaguobJaand Rombak volcanites the mafic members

x ~o 0

60 70Si02

*2 3 456 789

KzO

9

8 -,,/7 ~'\.o

Naz06 0~~'\.

+ ..+

....5 ..+oq.

+ +* + +4 + --. + + •++ x

x+ +

3+ .t+ x+

+ •• •

2+ •+ • •

1

80

•x

'. xx

x

•

50

+ /-If. + /

/+ +/ t

+ ++ ...+ ~~ i -.c

i+ + -+to ++ /* +~ +/ *,. +

r~/ *//0"'?"

40

o 12N

~ 10+o 8

NCO 6z

4

2

Fig. 4. (Na,O+K,O)versus SiO, plot for the Rombak Windowvolcanites. A - alkaline, SA - subalkaline.

Fig. 5. Na,O versus K,O plot for the Rombak Window vor-canites. Symbols as in Fig. 4.

NGU· BULL. 415.1989

have higher Na20/K20 ratios than the associa-ted felsic volcanites.

On the basis of Figs. 4 and 5 the volcanitesof the Rombak Window supracrustal sequen-ces are divided into three principal types: (1)The RS (Ruvssot-Sjangeli)-type; low-K20 maficto ultramafic subalkaline extrusives from the2.3 Ga supracrustal belt in the Ruvssot-Sjange-li area. (2) The G (Gautelis)-type; low-K20, high-Na20 rhyodacitic to rhyolitic volcanites withinthe Gautelis Tonalite Complex. (3) The SN(Serdal-Norodalj-type: a suite of mafic to fel-sic, generally K20-rich extrusives that are cha-racteristic of the Serdalen-Norddalen area, butoccur widely in the supracrustal belts westof the Muohtaguobla Tectonic Zone.

The three types of volcanites are shown ona (Na20+K20)-FeOtot-MgO plot (Fig. 6). Someof the SN-type volcanites were clearly classi-fied as alkaline on Fig. 4, and Fig. 6 confirmsthat the SN volcanltes cannot be part of atholeiitic trend, but belong to either the alkali-ne suite or the calc-alkaline suite defined byIrvine & Baragar (1971). Thus, on the basisof major elements alone the largest group ofvolcanic rocks in the Rombak Window (theSN-type) cannot be classified with certainty,but the predominance of andesitic compo-sitions favours a calc-alkaline affinity. In con-trast, the G-type rocks (three samples) areclassified directly as belonging to the calc-alkaline suite and the RS-rocks as tholeiitic(see below), though the mafic member of theRS-type plot near to the tholeiitic/calc-alkalineboundary.

FeOtot

D. SN-type volcanites

* RS-type votcamteso Gr-type votcanites

+ Hafic to felsic intrusions• Gautetis tonalite

MgO

Fig. 6. The Rombak Window suite of volcanic and intrusiverocks plotted in an AFM diagram.

The geochemistry of Lower Proterozic rocks 13

Trace elements: Chondrite-normalised REEpatterns for the Rombak volcanites are shownin Fig. 7. All the samples, except those fromthe ultramafic rocks of the RS group (Fig.7a), have similar REE patterns that are en-riched in the light rare earths (LREE), buthave essentially unfractionated heavy rareearths (HREE). The mafic RS-type volcanite(Fig. 7a) differs somewhat from either the SN-or the G-type volcanites (compare Figs. 7a,b, and c) in its lower La/SmN ratio. Neverthe-less, the LREE-enriched patterns of the SN-,G- and mafic RS-type volcanites resemblesthe REE patterns of calc-alkaline mafic andandesitic magmas (e.g. McBirney et al. 1987,Meen & Eggler 1987, Gill 1981), but contrastswith the smooth REE patterns characteristicof alkali basalts and andesites (e.g. Eiche etal. 1987, Lanphere & Frey 1987, Frey 1981,Gill 1981). Thus, the REE patterns suggestthat the Rombak mafic to felsic volcanitesbelong to the calc-alkaline suite.

In general, the felsic rocks have higher totalREE contents than the more mafic rocks. Thechange in REE abundance is accompanied bya change in the Eu anomaly present, as isdemonstrated by the telsic members of theSN-type volcanites (Fig. 7b). The samples withthe highest total REE contents have largenegative Eu anomalies, whereas the sampleswith low total REE contents have positive Euanomalies. This feature is here ascribed tolow-pressure tractionation of feldspar (probab-ly plagioclase) in the parental magma.

The REE pattern for the ultramafic extru-sives of the RS-type (Iowermost curve on Fig.7a) is LREE-depleted, and ranges from 1 to4 times chondritic values. This type of patternis interpreted as indicating that these rockswere derived from a LREE-depleted mantle.The REE pattern and low Zr content of theseultramafic rocks resembles Type I (also knownas aluminium undepleted) komatiites (Sun &Nesbitt 1978, Jahn et al. 1982), but becausethe Ruvssot-Sjangeli samples are Ti-depletedthey also have some affinities with boniniticmagmas. Boninite series volcanites, however,range from 52 to 68 % Si02 (Bloomer & Haw-kins 1987).

In order to examine the compositional variati-ons of a number of trace elements simultane-ously, normalised element plots ('spider-grams') are used (Fig. 8). In Fig. 8 the traceelements with a strong affinity for the silicatemelt - the hygromagmatophile elements (HYG)

14 Are Korneliussen & Edward W.Sawyer NGU - BULL. 415. 1989

VOLCANITES INTRUSIONS1000.-..--..--..--................................,.........,.........,........,........,...........-..............-r...,......,......,......,........,........,...............,.......,.......,.......,.......,.......,......,......,.---,1000

2'i::"0c~ 100U

.x:oo

er:10

RS- ultramallc R22.3l

A

............ --..::·::t:::>~~::~~~.~~ ...:...

DMafic intrusions

....,....'" ----- ,,"..•~..::::::::::::::~:..~~~.~~.~.....-;a

100

10

SN - felsic volcanites

---••••••••• .,===,.....c

100

10

1

NGU - BULL. 415, 1989 Thegeochemistry ofLower Proterozic rocks 15

1000',.--------------------, of Frey & Gordon (1974) - are arranged inorder of increasing D values (mineral/liquidpartition coefficients) for partial melting undermantle conditions of low PH20 and Po2. Theabundance of HYG-elements in the Rombaksamples is then normalised to the values foundin primordial mantle (Le. undepleted mantle)using the mantle values of Wood (1979). Com-pared to anororogenic basalts (Wood 1979)the Rombak basaltic andesites (representedby the average Rombak Window basaltic ande-site - ARA) are characteristically enriched inthe more HYG-elements and display a distinctnegative Ta-Nb anomaly (Fig. 8a). Thus, it isinferred that the Rombak Window basalticandesites are not of a anorogenic type. Incontrast, when compared to orogenic (or sub-duction-related) andesites (Fig. 8b) a strongsimilarity in HYG-element contents is obser-ved, suggesting a similar origin. The relativeenrichment of the large ionic Iithopile (L1L)elements such as Cs, Rb, K and Ba in sub-duction-related rocks is considered to be theresult of the dehydration, or incipient melting,of subducted lithosphere enriching the overly-ing mantle wedge (Hanson 1977, Best 1975,Hawkesworth et al. 1977). The depletion ofTa, Nb and Ti in the subduction-related igne-ous rocks is attributed to the retention of aTa-Nb-rich refractory titanium oxide phase athigh PH20 and P02 conditions in the overlyingmantle wedge (Best 1975, Hawkesworth etal. 1977, Sun 1979).

A

c

T1

........

Ta Nb

\ .... ARA

\........

ARSI\. ,

v

.......

AAA

..•....•...•.....•.....,..,ISLi ..

./ MORS.'

10

-'"u

~ 10

100

QJ.,c:ee

::;;-.;'El 100oE~

Fig. 8. The composition of some Rombak Window rocksnormalised to the primordial mantle. The elements havebeen arranged after the scheme of Wood (1979) in theorder of increasing calculated bulk partition coefficient formantle mineralogies, Le. the more 'incompatible' elementsto the left in the diagram,(a) Comparison of average Rombak Basement Windowbasaltic andesite (ARA) with selected mafic lavas fromanorogenic tectonic environments. ARA is the average ofthe six mafic-intermediate units M1 to M6 from the serdatprofile (Fig. 3). SEA - a basanite from Victoria, SE Austra-lia; AZ - Azores basalt; ISL - an Icelandic basalt. MORB- normal mid ocean ridge basalt. Dataafter Wood et al. 1979.(b) Rombak Window andesite (ARA) compared to orogenicandesites (52-56 % SiO,). The apparent similarity sug-gests that the Rombak Window andesites are of orogenictype, i.e. subduction-related. RP - K-rich Series of Vol-canic Roman Province. Mediterranean; M - Mediterranean(excluded K-rich Series of Roman Province): WSA - Wes-tern (Andean) South America: NWP - North-Western Paci-fic. Andesite data after Ewart (1982).(c) Comparison of Rombak Window mafic dykes and minormalic plutons represented by the average of 6 analysedsamples from sercat, Gautelis and Stasjonsholmen (ARBI:see REE-plotsof the individual samples in Fig.7d)and ARA.

Intrusive rocksSome workers (e.g. McCarthy & Groves 1979,Tindle & Pearce 1981) have pointed out thatmany granitic plutons are predominantly accu-mulations of crystals, and do not necessarilyrepresent melt compositions; thus comparisonwith volcanic rocks is not straightforward. Forthe purposes of this study our primary pointin documenting the compositional characteris-tics of the Rombak Window plutonics is toshow their close compositional similarity withthe SN-volcanites.Major elements: On the (Na20+K20)-FeOl0l-MgO plot (Fig. 6) the intrusive rocks generallyplot along a calc-alkaline trend similar to theSN-type volcanites. Many mafic dykes andminor mafic plutons are, however, iron en-riched compared with the mafic SN-type vol-canites, and they plot on the tholeiite side ofthe tholeiitic-/calc-alkaline boundary. A corre-

16 Are Korneliussen & Edward W.Sawyer

sponding phosphorus and titanium enrichmenttor these rocks (ARBI) is shown in Fig. 8c.

Trace elements: Figs. 7 d-t show the chondri-te-normalized REE patterns ot mafic dykesand minor plutons and felsic plutonic rocksfrom the Rombak Window. The Rombak Win-dow intrusive rocks have REE patterns of simi-lar shape, specifically LREE-enriched and with-out significant fractionation of the HREE. TheREE patterns therefore resemble those of thecalk-alkaline rocks in the area. In general, theRombak intrusive rocks have REE pattern ofsimilar shape and level as the SN-type vol-canites which they intrude.

On the mantle-normalised hygromagmato-phile element diagram, Fig. 8c, the RombakWindow mafic dykes and minor plutons areenriched in the L1L elements in a manner simi-lar to the SN-type extrusive rocks. Further-more, they also have prominent negative Ta-Nb and Ti anomalies, indicative of subduction-related magmas.

DiscussionSeveral Lower Proterozoic volcanic terranesin North America and on the Baltic Shield beara striking resemblance to modern arc systemsin lithological and geochemical characteristics(Condie 1987, Vivallo & Claesson 1987, amongothers). A common problem is the bimodalityin the volcanic successions, with a rarity ofandesites; in proper arcs the volcanic suitesshow a continuous evolution from mafic tofelsic including large volumes of andesite. Abimodality, however, can be explained by aritting of the volcanic arc (Condie 1987, Vival-10& Claesson 1987) and is not at all contradic-tory to a hypothesis that modern-style platetectonics were active in the Lower Proterozo-ic. It is particularly interesting to observe thata convincing ophlcllte complex has been desc-ribed from northeast Finland (Kontinen 1987).giving the best evidence so far that modernplate-tectonic processes were active in theLower Proterozoic. Thus, an interpretation ofthe rocks in the Rombak Window in the con-text of modern plate tectonics is relevant.

On the basis of major and trace elementsand REE data the ultramafic rocks of theRuvssot-Sjangeli area are shown to be compa-rable to komatiites, and the SN·, G- and ma-fie RS-type volcanites all belong to the catc-alkaline suite. Potassic andesite is the most

NGU - BULL. 415. 1989

Hfl3

~ !'!~, le - :ntl"'m~!atl"vorcao.tes IS~·tY;:lel

V' f.lsic vole ISN-type)

I;) Fttsic vole IG-type)

~ ""aflc and ultra,"T'Iaflcvotcaotes (RS-ty:>e)

• Icoat.te , Gal.OtelLs

X Maflc mtrus 0'15

Fig. 9. Mafic and felsic volcanic and intrusive rocks plottedin the Th-Hf-Ta discrimination diagram after Wood et a!.(1979) and Wood (1980). Field A - N-type MORB; field B- E-type MORB; field C - within-plate basatts: field 0 -magma series at destructive plate margins.

common rock type within the most extensivevolcanites: the SN-type. Various discriminationdiagrams have been proposed to classify thetectonic settings of volcanic and plutonic rocksby means of their geochemistry, and thesehave been applied to the Rombak Windowvolcanites. The Th-Hf-Ta concentrations ofmafic to felsic volcanic and intrusive rocksfrom the Rombak Window are plotted in thediagram (Fig. 9) of Wood et al. (1979), whichhas the advantage of being able to distinguishthe tectonic settings of both mafic and felsicmagma types. The Rombak igenous rocksplot well within the field D in Fig. 9, which isthe field for magma suites formed along de-structive plate margins, Le. subduction-relatedmagmas. On the TiO, versus Zr plot of Pearceet al. (1981) the indicated tectonic setting ofthe Rombak volcanites and intrusives withSiO, contents

NGU·BULL.415,1989 Thegeochemistry ofLower Proterozic rocks 17

originating beneath the thickest crust. Thus,the SN-type volcanic rocks of the RombakWindow could represent magmas extrudedthrough a thick crust, and on the log (CaOI(Na,O + K,O)) versus SiO, diagram of Brown(1982) they do indeed plot on the «increasingarc maturity»side of normal calc-alkaline ande-sites.

A major problem in the interpretation of theearliest Proterozoic evolution of this region isthe paucity of precise age-determinations. Theconglomerate which overlies the Gautelis To-nalite Complex (with the G-type volcanites) isitself overlain by a dolomitic carbonate that isin turn overlain by the greywacke sequence.The nature of the carbonate-greywacke con-tact is not yet known. The Gautelis Tonalitecomplex represents the local basement andcould be of either Archaean or Lower Protero-zoic age. In either case the conglomerate indi-cates an erosional period that was followedby platform carbonate sedimentation. TheGautelis greywacke-tuffite sequence indicatesthe later formation of sedimentary basins thatreceived sediment derived in part from calc-alkaline volcanic rocks (SN-type). The tectonicsetting of 'the sedimentary basin was near toeither a volcanic island or a magmatic arc si-ted on continental crust, since thick piles ofvolcanic rocks indicate a position proximal tothe volcanic centres; a more dtstal position forthe greywackes and pelites is indicated by thethin interbedded tufts.

Sawyer & Korneliussen (this volume) haveshown that the tectonic setting in which thegreywackes (turbidites) formed can be inferredfrom their composition by determining thepossible source-rock types. The turbidites fromRombaksbotn and Gautelis formed in an ac-tive marginal basin setting adjacent to a matu-re volcanic arc that was, in the case of Gaute-lis, probably located on a tonalitic crust ofLower Proterozoic or Archaean age. An Ande-an-type setting is proposed. From a considera-tion of the geochemistry of the SN-type vol-canites it is possible to elaborate on the histo-ry of that magmatic arc.

The high MgO content in some of the SN-type volcanites indicates that the parentalmagma originated by the partial melting of amantle source. As indicated by the REE pat-terns (Fig. 7b) of the felsic SN-volcanites, frac-tional crystallization has been a major factorin the evolution of the calc-alkaline volcanites.A negative Ta-Nb anomaly (Fig. 8b) and the

1000

WPG

ORG

Zr

- ---,-,

\\

\\\\

100

+ Intrusions Volcanites

Syn-COLG

VAG

Felsle vole.•SN-typeo Felsle vole.•G-type+ Granitess Tonallte, Gautells

/I • __,

I ".,:+.... :+\' ,+':,j' : ..

1BAre Korneliussen & Edward W. Sawyer NGU· BUll. 415. 1989

Fig. 12. Plot of log (CaO/(Na,O+K,O)) against SiO, for vol-canic and intrusive rocks in the Rombak Window (RS-typeexcluded). The trends for volcanic suites from modernmagmatic arcs and the field indicating the range for nor-mal calc-alkaline andesites are from Brown (1982). T-SS -Tonga-S.Sandwich; NZ - New Zealand; NG - New Guinea.

enrichment of the L1L-elements in the basalticandesites indicates a subduction-related mag-ma origin. Thus three stages are envisioned:(1) The parental magma may have originatedin the mantle wedge above the related sub-duction zone, followed by (2) crystal fractiona-tion by Fe-Mg silicates to generate fractiona-ted, gradually more siliceous, intermediatemagmas; and finally (3) a late stage of magmaevolution in which the felsic magmas wereformed, and which was, to a large extent,controlled by the fractional crystallization ofplagiocfase, presumably at shallow crustal lev-els. During this process significant interactionwith crustal rocks is likely to have occurred.In modern geological settings magmatic rockssimilar to those of the Rombak Window potas-sic calc-alkaline suite are believed to haveoriginated beneath a thickened crust. Thick-ening may result from magma injection into,or magma extrusion onto the crust; in eithercase potassic cale-alkaline volcanites repre-sent a late-stage mature or 'continentised'stage of magmatic arc development (Fig. 12).Since no rocks with distinct alkaline REE pat-tern were found ;n the Rombak Window thereis no reason to suppose that the magmaticarc reached the stage of rifting.

The Ruvssot turbidites are more mafic thanthose from Gautelis or Rombaksbotn, andappear to contain neither continent-derived,

nor fractionated volcanic material; thus, Saw-yer & Korneliussen (this volume) proposed anintraoceanic setting. Furthermore, they wereable to determine that the Ruvssot turbiditescontain material that could have been derivedfrom RS-type volcanites. The komatiitic affinityof the ultramafic members of the AS-typevolcanites and their associated low-K mafiecalc-atkaune pillow basalts that show only slightLREE-enrichment is consistent with a primi-tive, intra-oceanic island arc setting for thevolcanic rocks of eastern Rombak BasementWindow.

Thus, the eastern part of the Rombak Base-ment Window contains the remnants of theearly stages of an intra-oceanic arc voleanismwhich occurred at about 2.3 Ga. In contrast,the western part of the window contains theremains of a younger, but pre-1.78 Ga, ma-ture volcanic arc located on continental crustperhaps of Archaean, but probably of Protero-zoic age. The present close spatial relation-ship of these two terranes may be due to eit-her Lower Proterozoic collision, or crustalshortening of the Lower Proterozoic crustduring the Caldonian orogeny.

It is interesting to note that in northernSweden, the supracrustal rocks south of theArchaean Domain (Fig. 1) are dominated by acontinuous series of mafic to telsic voleanitesof continental affinity (Fritsch & Perdahl 1987).Their age is 1.9 Ga based on U-Pb dating ofzircons (Ski6ld 1988). In the Skellefte district(300 km south of Kiruna) this province givesway southwards to a somewhat older, predomi-nantly felsic volcanic province of marine affini-ty (Claesson 1985, Wilson et al. 1987). It hasbeen suggested that the Skellefte provincecould be related to the northward-directedsubduction of oceanic crust under a LowerProterozoic continent (e.g. Hietanen 1975,Wilson et al. 1987), and that the continentalprovince north of the Skellefte district repre-sents a continent-based magmatic arc whichdeveloped after the marine magmatic arc (Wie-denfalk et al. 1987).

In contrast to the Skellefte calc-alkaline!tholeiitic volcanic suite which shows a bimoda-lity with a scarity of andesitic rocks (Vivallo1987), the dominant type of volcanic rock inthe Rombak Window (the SN-type) shows acontinuous evolution from calc-alkaline basaltto rhyolite with a large proportion of andesite.The bimodal character of the Skellefte vol-canic suite led Vivallo (1987) to suggest that

8070% Si02

6050

Normal fieldfor talt-alkaline 0 Volcanites, SN-typeande.lt.. 0 Volcanites, G-type

............ J'. • Mane to acidic intrusions+ 0 '1'(' ~S • Gautelis tonalite~+ o. 0"""'" I

++ '0 o-tC ......+ .:-...::. .........+:0 """ ......

+ ~,+ ......

o /.O\>~:".~!'.~ ; ;- ": '0:e '

NGU - BULL. 415. 1989

the Skellefte volcanic arc was dominated byextensional forces during long periods, probab-ly producing incipient rifting. There is no rea-son to suppose that the Rombak Windowmagmatic arc reached the stage of rifting. Itis possible to make a tentative correlationthat could be tested by precise age determina-tions, that the Gautelis Tonalite Complex inthe Rombak Basement Window is equivalentto the Jorn tonalitic complex (Claesson 1985)and its associated felsic volcanic rocks in theSkellefte district of Sweden. The younger grey-wacke-volcanite sequence (SN-type) may thenhave formed in the Rombak equivalent of thecontinent-based magmatic arc of northernSweden. If such a correlation is correct, thenthe Gautelis Tonalite Complex has an age ofapproximately 1.9 Ga, and the greywackesand calc-alkaline SN-type potassic volcanismare somewhat younger.

ConclusionsThe 2.3 Ga Ruvssot-Sjangeli intra-oceanicvolcanic arc rocks and associated turbiditesequence are the oldest rocks found in theeastern part of the Rombak Window, whereasin the south rocks of the Gautelis TonaliteComplex (unknown age) are the oldest. TheGautelis Tonalite Complex is overlain by abasal conglomerate and a sequence of dolomi-tic carbonates indicating platform sedimentati-on. After carbonate deposition a sequence ofturbidites and potassic arc-related volcanites(the SN-type) developed, presumably at about1.9 Ga. The composition of these calc-alkalinevolcanic rocks ranges from that of basalt torhyollte, but is predominantly andesitic, and isconsistent with the arc being of a mature,continentised-type resting upon tonalitic conti-nental crust. At present it is not known whet-her the Ruvssot-Sjangeli and Gautelis rocksformed a continuous basement to the car-bonates, or whether the sequence was assem-bled during a Lower Proterozoic collision eventprior to 1.9 Ga.

The similarities in major element and traceelement composition between the potassicarc-related volcanites and the 1.78 Ga intru-sive rocks (Figs. 6, 7, 8 & 12) indicate a rela-ted source. Thus, the abundant plutonic rocksin the Rombak Window could, perhaps, repre-sent a later and deeper stage of mantle activi-ty as the volcanic arc thickened.

The geochemistry ofLower Proterozic rocks 19

AcknowledgementsMany geologists have been involved in ore prospecting andother geological investigations in the Rombak BasementWindow since 1983. We are especially indebted to ArneGmnlie (NGU) and Erik Skonseng (Univ. of Tromse) forhaving participated in NGU's fieldwork. and Rolf Romer(Univ. of Lule{t). Frank-Dieter Priesemann and Jan IngeTollefsrud (both Folldal Verk A/S). and Soye Flood (nowGeologiske Tjenester a.s.• previously ARea Norway Inc.)for having participated in many stimulating discussions aswell as excursions in the field. E.W. Sawyer would like tothank NTNF for a postdoctoral research fellowship (1985-1986) held at NGU, Trondheim, without which participationin this project would not have been possible. Finally. wewould like to thank G. Gaal and an anonymous reviewerfor constructive reviews on the manuscript.

20 Are Korneliussen & Edward W. Sawyer

ReferencesBest, M.G. 1975: Migration 01 hydrous fluids in the upper

mantle and potassium variation in calc-alcalic rocks.Geology 3. 429-432.

Birkeland. T. 1976:Skjomen, berggrunnsgeologisk kart N10- M 1:100000. Nor. geol. unders.

Bloomer. S.H. & Hawkins, JW. 1987: Petrology and geoche-mistry 01 boninitic series volcanic rocks Irom the Maria-nas Trench. Contr. Mineral. Petrol. 97, 361-377.

Brown, G.C. 1982: Calc-alkaline intrusive rocks: their diversi-ty. evolution, and relation to volcanic arcs. In Thorpe,R.S. (ed.) Orogenic andesites and related rocks. JohnWiley, London, 437-61.

Butler, RW.H. 1982: The terminology 01 structures in thrustbelts. J. Struct. Geol. 4, 239-246.

Condie, K.C. 1987: Early Proterozoic volcanic regimes insouthwestern North America. In Pharaoh. T.C., Beckin-sale, RD. & Rickard, D. (eos.) Geochemistry and Minera-lization of Proterozoic Volcanic Suites. Geol. Soc. Spec.Publ. 33, 211-218.

Claesson, L.A. 1985: The geochemistry 01 Early Proterozo-le metavolcanic rocks hosting massive sulphide depo-sits in the Skellelte district, northern Sweden. J. Geol.Soc. London 142. 899-909.

Eiche, G.E., Francis, D.M. & Ludden, J.N. 1987: Primaryalkaline magmas associated with the Quarternary Alliga-tor Lake volcanic complex, Yukon Territory Canada.Contr. Mineral. Petrol. 95, 191-201.

Frey, FA 1981: Application 01 neutron activation analysisin mineralogy and petrology. In Muecke. G.H. (ed) Neu-tron Activation Analysis in the Geosciences. Min. As-soc. Canada Short Course 5, 167-210.

Frietsch, R. & Perdahl. J.-A. 1987: On the nature 01 lowerProterozoic volcanic rocks in southern Norrbotten,northern Sweden. Proterozoic Geochemistry (symp.).IGCP 217. tuna, (abstr.) p. 35.

Gaal, G. & Gorbatschev, R 1987: An outline 01 the Precam-Man evolution 01 the Baltic Shield. Precamb. Res. 35,15-52.

Gill, T. 1981: Orogenic Andesites and Plate Tectonics. Sprin-ger Verlag, 390 pp.

Gunner, J.D. 1981: A reconnaissance Rb-Sr study of Precam-brtan rocks from the Sjangeli-Rombak Window and thepattern of initital "SrluSr ratios Irom northern Scandina-via. Nor. Geol. Tidsskr. 61. 281-290.

Gustavson, M. 1974 a: Narvik. Beskrivelse til det berggrunns-geologiske gradteigskart N9 - 1:100000. Nor. geol.unaers. 308, 34 pp.

Gustavson. M. 1974 b: Geologisk kart over Norge. berg-grunnskart Narvik 1:250000. Nor. geol. unoers.

Hanson, G.N. 1977: Evolution of the sub-oceanic mantle.J. Geol. Soc. London 134. 235-253.

Hawkesworth, C.J., O'Nions, R.K.. Pankhurst, R.J., Hamil-ton, P.J. & Evensen. N.M. 1977: A geochemical studyof island-arc and back-arc tnolsutes Irom the Scotiasea. Earth Planet. se: Lett. 36, 253-262.

Heier, K. & Compston, W. 1969: Interpretation 01 ae-srage pattern in high-grade metamorphic rocks, NorthNorway. Nor. Geol. Tidsskr. 49, 257-283.

Jahn, B.·M., Gruru, G. & Glikson, A.Y. 1982: Komatiites 01the Onverwacht Group S.Alrica. REE geochemistry,SmlNd age and mantle evolution. Contr. Mineral. Pet-rol. 80. 25-40.

Kontinen, A. 1987: An Early Proterozoic ophiolite - theJormua malic-ultramalic complex, northeastern Finland.Precamb. Res.• 313-341.

NGU-BUlL.415.1989

Korneliussen, A. &Sawyer, E. 1986: Berggrunns-og maim-geologi med srerlig vekt p~ muligheter lor gull, sydligedeter av Rombakvinduet, Nordland. Nor. geol. enders.Unpub/. report 86. 167. 70 pp.

Korneliussen, A., Tollefsrud, J.I., Flood, B. s Sawyer, E.1986: Precambrian volcano-sedimentary sequences andrelated ore deposits, with special relerence to the Gaute-lisljell carbonate-hosted gold deposit. Rombak base-ment window, Northern Norway. Nor. geol. unoers.Unpubl. report 86.193. 46 pp.

Lanphere, MA & Frey. FA 1987: Geochemical evolution01 Kohala Volcano, Hawaii. Contr. Mineral. Petrol. 95.100-113.

McBirney, A.R .• Taylor, H.P. & Armstrong, R.L. 1987: Paricu-tin re-examined: a classic example 01 crustal assimila-tion in cale-alkaline magmas. Contr. Mineral. Petrol. 95.4-20.

McCarthy, T.S. & Groves, D. I. 1979: The Blue Tier batho-lite. northeastern Tasmania. Contr. Mineral. Petrol. 71.193-209.

Meen, J.K. 1987: Formation 01 shoshonite Irom cetc-elkali-ne basalt magmas: geochemical and experimental cons-traints Irom the type locality. Contr. Mineral. Petrol. 97,333-351.

Meen, J.K. & Eggler. D.H. 1987: Petrology and geochemis-try 01 the Cretaceous Independence volcanic suite.Absaroka Mountains, Montana: Clues to the composi-tion 01 the Archaean sub-Montana mantle. Bull. Geol.Soc. Am. 98, 238-247.

Chlander, B., Ski6ld. T., Hamilton, P.J. & Claesso-;', L.A.1987: The western border 01 the Archaean province01 the Baltic Shield: evidence Irom northern Sweden.Contr. Mineral. Petrol. 95, 437-450.

Pearce, J.A. 1980: Geochemical evidence lor the genesisand eruptive setting 01 tavas lrom Tethyan ophiolites.In Panayiotou, A. (ed.) Opbtotites Proceedings, Interna-tional Ophiolite Symposium, Nicosia. Cypros Geol.Surv., 261-272.

Pearce, J.A .• Alabaster, T., Shelton, AW. and Searle, M.P.1981: The Oman ophiolite as a Cretaceous arc-basincomplex: evidence and implications. Phil. Trans. R.Soc. A 300. 299-317.

Pearce, J.A., Harrts, N.BW. & Tindle, A.G. 1984: Traceelement discrimination diagrams lor the tectonic inter-pretation 01granitic rocks. Journ. Petrol. 25, 956-983.

Pharaoh, T.C. & Pearce, J.A. 1984: Geochemical evidencefor the geotectonic setting of Early Proterozoic metavol-canic sequences in Lapland. Precamb. Res. 10,283-309.

Hobyn, T.L., FlOOd, B.• & Thomassen, B. 1985: A compari-son of geophysical data and geological mapping, Rom-bak Precambrian window, northern Norway. 47 tn. E.A.E.G. meeting in Budapest. 26 pp.

Romer, R.L. 1988: Disturbance of Pb. Sr and Nd isotopesystems related to medium and low grade metamor-phism illustrated with data from the polymetamorphicEarly Proterozoic (Rombak)-Sjangeli basement windowof the Caledonides of Northern Sweden. Doctoral The-sis, Lule~ Univ. of Technology, 1988:65 D, 124 pp.

Romer. R.L. & Boundy, T.M. 1988: Lithologic and tectonicprofile across the Muohtaguobla area, Rombak base-ment window, Northern Norway. Nor. geol. unders.

. Unpubl. report 88.116. 11 pp.Sawyer. E, 1986: Metamorphic assemblages and conditions

in the Rombak basement window. Nor. geol. uncers.Unpubl. report 86.168, 25 pp.

NGU - BULL. 415, 1989

Skiald, T. 1988: Implications of new U-Pb zircon chronolo-gy to Early Proterozoic crustal accretion in northernSweden. Precamb. Res. 38, 147-164.

Skonseng, E. 1985: 8erggrunnsgeologisk kartlegging iGautelisomrMet, Skjomen, Nordland. Feltrapport. Nor.geol. unders. Unpubl. report 85.214, 16 pp.

Sun, S.S. 1980: Lead isotope study of young volcanic rocksfrom mid-ocean ridges, ocean islands and island arcs.Phi!. Trans. R. Soc. London, Ser. A. 297, 409-445.

Sun, S.S. & Nesbitt, R. 1978: Petrogenesis of Archaeanultrabasic and basic volcanics: Evidence from rare earthelements. Contrib. Mineral. Petrol. 65, 301-325.

Taylor, S.R. & Gorton, M.P. 1977: Geochemical applicationof spark source mass spectrography - Ill. Elementsensitivity, precision and accuracy. Geochem. Cos-moch. Acta 41, 1375-1380.

Tindle, A.G. & Pearce, J.A. 1981: Petrogenetic modelltnq ofin situ fractional crystallization in the zoned Loch Doonpluton, Scotland. Contrib. Miner. Petrol. 78, 196-207.

Tull, J.F., 8artley, J.M. Hodges, K.V., Andresen, A., Stelten-pohl, M.G. & White, J.M. 1985: The Caledonides in theOfoten Region (68-69 N), North Norway: key aspectsof tectonic evolution. In D.G. Gee & 8.S. Sturt (eds.)The Caledonide Orogen - Scandinavia and RelatedAreas. John Wiley & Sons, Chichester, 553-568.

Vivallo, W. 1987: Early Proterozoic bimodal volcanism, hyd-rothermal activity, and massive sulfide deposition in theBoliden-LAngdal area, Skellefte district, Sweden. Econ.Geol. 82, 440-456.

Vivallo, W. & Claesson, L.-A. 1987: Intra-rifting and massive

Thegeochemistry ofLower Proterozic rocks 21

sulphide mineralization in an Early Proterozoic volcanicarc, Skellefte district, northern Sweden. In Pharaoh,T.e., Beckinsale, A.D. & Rikard, D. (eds.) Geoehemistryand mineralization of Proterozoic volcanic suites. Geol.Soc. Spec. Publ. 33, 69-79.

Vogt, T. 1942: Trekk av Narvik-Ofoten.traktens geologi.Nor. Geol. Tidsskr. 21, 198-213.' .

Widenfalk, L., Claesson, L.-A. & Skiald, 1987: Rocks' asso-ciated with uplift of a Proterozoic continental margin,N. Sweden. Proterozoic Geoehemistry (symp.), IGCP217, tuna, (sbetr.), p. 92.

Wilson, M.A., Claesson, L.A., Sehlstedt, S., Smellie, J.A.T.,Aftalion, M., Hamilton, P.J. & Fallic, A.E. 1987: Jom:an early Svecokarelian granitoid complex from northernSweden. Precamb. Res. 36, 201-225.

Witschard, F. 1984: The geological and tectonic evolutionof the Precambrian of northern Sweden - a case forbasement reactivation? Precamb. Res. 23, 275-315.

Wood, D.A. 1979: A variably veined suboceanic uppermantle - Genetic significance for mid-ocean ridge ba-salts from geochemical evidence. Geology 7,499-503.

Wood, DA, Joron, J.-L. & Treuil, M. 1979: A re-appraisalof the use of trace elements to classify and discrimina-te between magma series erupted in different tectonicsettings. Earth Planet. Sei. Left. 45, 326-336.

Wood, D.A. 1980: The application of the Th-Hf-Ta diagramto problems of tectonomagmatic classification and toestablishing the nature of crustal contamination of basal-tic lavas of the British tertiary volcanic province. EarthPlanet. Sci. Left. 50, 11-30.