granite production inthe delamerian orogen, south...

Journal o/Ihe Geological Soeien', London, Vol. 159,2002, pp 5"7-575. Printed in Great Britain.

Granite production in the Delamerian Orogen, South Australia

J. D. FODEN1, M. A. ELBURG1

, S. P. TURNER2, M. SANDIFORD3, J. O'CALLAGHAN1

& S. MITCHELL 1

1Departmen( of Geology and Geophysics, Adelaide University, Adelaide, S.A. 5005, Australia,(e-mail: [email protected])

2Department of Earth Sciences, University of Bristol, Bristol, UK3 School al Earth Sciences. University o{ Melhourne. Melbourne. Vie. 30 IO. Australia

Abstract: In ¡he South Australian sector of the Cambro-Ordovieian Ross-Delamerian Orogen. granites rangein age from Mid-Cambrian to Early Ordovleian. Their occurrence is largely eontined to deep, Early Cambrian,sediment-tilled basins where they arc a,soeiated with matic rocks. The syntectonic suites have compositionsforming a continuum between 1- and S-type granites. After the cessation of convergent deformation at c.490 Ma an abrupt transition to a bimodal magmatic association of matic intrusions and tèlsic granites andvolcanic rocks of S- and A-type atnnitics occurred. As exposed on the south coast of Kangaroo Island, S-typegranite originated as in silll partial melts of the Early Cambrian sediments locally intruded by either maticmagmas or I· S granite magmas. These nHgmatite complexes were mingled with intrusions from the magmasthat provided the underlying heat sources. Also on Kangaroo Island, composite S-type rhyodacite-doleritedykes indicate that crustal melting involved mantle-derived melts. Field observations, major and trace elementdata and Nd-Sr isotopic data indicate that granite magmas in this fold belt result from mixing of crustal andmantle source components, and from fractional crystallization (AFC-type processes). Whereas the Nd-Srcompositions of granite suites from the Delamerian Orogen form a continuous geochemical trend between thecrust and the mantle melts, the A- and I-types cluster towards the mantle endmember and the S-types towardsthe crustal endmember. This dichotomy reflects three granite magma production situations: (I) lower-crustalmatie magma chambers that are contaminated by, and mingled with, melts of the local metasedimentsproducing I-type magmas: (2) erustal melts formed in the heated zones above upwelling mantle or close tomatie or I-type granite intrusions producing S-type magmas; (3) upper-crustal matic intrusions where closed-system fractionation dominates to produce A-type granite. The extent of fractionation and crustal assimilationvaried progressively through the c. 30 Ma deformation history (514-485 Ma) of this orogen. Importantly inthis sector of the Ross ..Delamerian Orogen, the crustal endmember is represented only by the Cambrian basinsedimentary till (Kanmantoo Group) and expressly excludes the older Preeambrian crust.

Keywords: AFC, South Australia, Ross DcJamerian Orogen, S-type granites, neodymium isotopes.

This paper concerns the origins of granite magmas in theDelamerian Fold Belt in South Australia (Table 1). \II!henformed, this orogen was contiguous with the Ross Orogen inAntarctica, forming the earliest stag..: of development of theTasman Orogen and lying between the western edge of theyounger Palaeozoic Lachlan Fold Belt and the PrecambrianCraton (Coney el al. 1990). The dispute over the origin ofgranite, particularly in the Lachlan Fold Belt. has been clJnsider-able and encapsulates debate repeated in many other globa.1granitic terranes (Chappell & White 1974, 1992; White &Chappell 1988; Chappell 1996; Collins 1998). The terms 1- andS-type granite were coined there (Chappell & White 1974, 1992;White & Chappell 1988) and likewise the concept of restitc(Chappell el al. 1987; Chappell 1996). These models led togranites being regarded as simple geochemical 'images' of theirsources and essentially intracrustal phenomena. The contraryview, however, is that granite is a mixture of crustal and mantlesources (Gray 1984; Collins 1998).

In this paper we use the granites of the Delamerian Orogen tocontribute another dimension to this debate, particularly takingadvantage of the extensive knowledge of the rocks that form thebasement to this fold belt compared with that of the LachlanFold Belt.

Regional geological setting

The Cambro-Ordovician Delamerian Orogen in the southernAdelaide Fold Belt (Figs I and 2) (Preiss 1987; Jenkins &Sandiford 1992) lies directly to the east of the Precambriancraton and hosts granitic rocks and felsic volcanic rocks withages in the range c. 520 Ma to 480 Ma. In South Australiathis orogen underwent deformation from early Mid-Cambrian(514 ± 4 Ma; Foden el al. 1999) to Early Ordovician time(c. 490 Ma). This Cambrian activity is concurrent withsubduction-related arc volcanism c. 1000 km to the east inthe Takaka Terrane in New Zealand (Fig. I) (Münker &Cooper 1995; Münker & Crawford 2000). The southernAdelaide Fold Belt is important as it marks the initiation ofthe major tectonic transition of the eastern margin of thenewly assembled Gondwanan supercominent, from Late Neo-proterozoic passive margin, to Early-Mid-Cambrian conver-gent subduction margin (Coney et al. 1990; Powell et al.1994). The latter state has prevailed from this time to thepresent, with progressive eastwards migration of orogenieactivity through time. The Cambro-Ordovician DelamerianOrogen is bounded to its east by the Mid- to LatePalaeozoic Lachlan Fold Belt (e.g. Coney et al. 1990; FigsI and 2).

557

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

558 J. D. FüDEN ETAL.

Table I. Cambro-Ordoviciall gmnites, Adelaide Fvld Bell

Granite name Locality LaI.. long. n Rock types Retèrenees

Synleetollic /- and S-Ilpe grallilesRemarkable Rocks SW Kangaroo Is. 3ó.03, 136.45 grd-gran 2, 3Cape Younghusband SW Kangaroo is. 36.02, 136.50 grd -gran 3Vivonne Bay West of Point Ellen. KI 36.02, 137.11 grd-gran 3Stun Sail Boom River 36.02, 137.01 grd granCape Willoughby East Kangaroo I, 35.52, 138.07 gran 2. 3Victor Harbor- Port Elliot Eneounter Bay 35.35, l3R.40 dior-gran 2, 3Taratap Adamellitc SE South Australia 36.40. 139.50 adamellite 2Monarto Granitc \Vest of Murray Bridge 35.05, 139.10 gran 2. 3Reedy Creek Granodioritc Reedy Creek 34.55, 139.13 dior-gran 2Palmer Granite Palmer 3452. 139.10 gran 2. 6Rathjen Gneiss '\orth of Palmcr 34.45. 139.08 grd gran 6. 7Tanunda Granite East Barossa Valle) 34.34, 139.00 grd-gran 7Cookes Hill Granite West of Sedan 34.45.139.12 tonalite 9Anabama (jranite South of Broken Il iII 32.50. 140. I O dior-grd 9Bungalina Monzonite Peake and Denison Ranges 2830. 136.00 münz--sy 8Harrow Granite West Victona 37.18.141.42 gran 1.4,5Wando Granodionte West Victofla 3728, 141.24 diorgran 1.4,5

Posl-leclonic A-I.1l'e granile,Black Hill Cambria, West Murray Valley 34.41. 139.28 gab. 111011Z. gran 1,2,3Sedan (iranite Sedan. West :vIurrav Valley 34.30.139.21 gran IMannul11 Granite '"Iannum 34.53.139.21 gran IReedy Creek Dionte Reedy Creek 3455.139.13 dior IMurray Bridge Granite \1urnlY Bridge 35.07. 139.10 gran IMt Monster Porphyry :vit Monster. SE South Australia 36.14, 140.21 ignimbrite IPadthaway Ridge····Coonawarra \iumerous small inliers in SE South Australia 35.45, 139.30 gran-sy I

37.00, 140.45Oerholm Granite West Victoria 37.25,141.20 gran 1.2,5

References: I. Turner et ,,/. ( 19'12): 2. Füden el "l. ( 19901: 3. \1; Ine, Cl ,,/. (1977): 4. rtöttmann el a/. 11994 j; 5. Gibson & Nihili ( 1992 l: 6. Fleming & White (1984 l; ï, Füdenel al. ( 1999l: g, rv1orrison & Fod~n / 1990.1: 9, Pn:iss 1.19R7).. .i.hbre\ i~lIlt'ns: grd. granodiorite: gran. gmnilc; dior. diorite: münz, ll1onl:onite: sy. syenik: gab. gabbro.

deformation and the termination of sedimentation is provided bythe age of the earliest syndeformational Rathjcn Gneiss (514 Ma;Faden el al. 1999). These results indicate that the southernAdelaidc Fold Belt evolved very quickly from sedimentary basinto fold belt and that the pre-deformational phase of Cambriansedimentation and mafic magmatism lasted only c. 12 Ma.

The Delamerian Orogen is characterized by a low-P and high-T metamorphism (Oftler & Fleming 1968; Mancktelow 1990;Dymoke & Sandiford 1992; Sandiford el al. 1992), showinglargc variation from chlorite to sillimanite grade, with the highestgrades restricted to a narrow meridional belt, which is also thesite of syntectonie granite intrusion and structural complexity.Assemblages indicate peak metamorphic conditions in the regionof 0.3-0.4 GPa at temperatures up to 650 oC (Sandiford el al.1(90).

The termination of convcrgent dcformation of the DelamerianOrogen is marked by the transition from defonned to unde-formcd granitoid intrusions, and U-Pb zircon dating (Turner &Faden 1(96) as well as new Rb-Sr geochronology reported inthis paper indicates that this transition took place at c. 485 Ma(see Table 5, below). Significantly, this transition is marked bymajor changes in the composition of granite magmas, fromsyntectonic I-S types to highly siliceous, potassic, post-tectonicA-types (Foden el al. 1990; Turner el al. 1992; Turner 1996;Tumer & Foden 1996). These late-stage A-type felsic magmasfonn part of a high-temperature, bimodal magmatic suite gener-ated after a phase of major uplift and erosion of the DelamerianFold Belt (Turner el ul. 1996). Nd-isotope and muscovite ArlArprovenance studies of sedimcntary rocks support the observationthat uplift and erosion of the Ross--Delamerian Orogen supplied

The Delameriun Orogeny

Thc southern Adelaide Fold Belt is part of a Neoproterozoic toEarly Palaeozoic orogen that extended over 5000 km along thesoutheastern edge of the early Gondwana supercontinent (e.g.Dalziel 1991; Moores 1991). lt is composed of ma1l11y Neopro-terozoic (Adelaidean) and Lower Cambrian (the Normanville andKalUnantoo Groups) sedimcntary rocks with associatcd maficigneous and granitic rocks (e.g. van der Borch 19i'O; Prciss1987).

The Delamerian Orogen is a compressional orogt:n developedby westward vergent folds and thrust faults. Its structural historyhas been described by Offler & Fleming (1968). Fleming &White (1984), Mancktelow (1990), Jenkins & Sandiford (1992)and flötlmann el al. (1994). Recent studies (Flöttmann el al.1998, 1995, 1994) have recognized that the Kanmantoo sedimen-tary basin with its associated basalts and dolerites (Foden el al.2002) fanned as a localized steep-sided tear basin resulting fromdextral shear on an east-west fault system south of KangarooIsland (Fig. 2). As the Cambro-Ordovician mafic and felsicmagmatism (Table I) was largely confined to this Cambriansedimentary trough it seems that early rifting was a veryimportant stage in thc evolution of the fold belt. It resulted inmantle exhumation and in the localization of zones (If thermalinteraction between crust and mantle.

The maximum age of commencement of the Normanville-Kanmantoo sedimentation and the age of onset of the Delamer-ian folding are constrained by U- Pb dating of zircons. Tuf!'layers in the basal Normanville Group yielded an Early Cambrianage of 526 Ma (Cooper er al. 1992). and [he age of onset of


GRANITES l:-.JDELAMERIAN OROGEN

/ WARBURTON

.

'."'.·'•..·•... ·t· ........•. '.· X BASINLV.. ,A;.".;' V

;X,;.~..........

Mt PainterInlier

Peake & DennisonRanges

Tasman Line

t~~.~."9-o-

'\03 x""X(1)1 Anabama

(~: ~. GranileMUrray basin

Adelaide'.x Black( - -I ,1'f7\ Hill

Lachlan FoldBelt

-~îc~xangaroo~~, Complex,

....... x ~.+"' Padtn X .~'.\t_'..c;•......+

~

awaY "X>~;:;Rldge'~-;'+

Harrow ¥.

UCambrian ~ ?

UNeo- "-...proterozoie ?~.KingIsland . ?

DelamerianFold Belt

Proterozoicbasement

Archaeanrock

u• CambrianlE. XOrdovieian Granites

))~..,\-1

~~l '"

DUNDASTROUGH

Tasmania

Cambrianlholeiilic lavas V

Cambrian Greenstones +100

--J,

the extensive Lachlan Fold Belt Ordovician quartzosc flyschdeposits (e.g. Cas 1983: Fergusson et al. 1989: Turner el al.1996).

Analytical methods

Clean fresh fragments of samples were crushed and thcn groundin a WC: mill to < 2 µm grain size. This material was used forXRF and Sr and Nd isotopic analyses. XRF analyses wcreperfonncd at the Department of Geology and Gcophysics,Adelaide University, with a Philips PWl480 100 kV spectro-meter using techniques describcd by Turner et al. (1993b) andFaden et al. (199<)). Sr and Nd isotope ratios were analysed atthe Department of Geology and Geophysics on either theFinnigan 261 single-collector, thermal ionization mass spectro-meter (TlMSj in dynamic mode. or on the Finnigan MAT 262TIMS in static mode. Methods have been described by Turner etal. (I <)93b) and Foden et al. (1999). During the intcl"\al overwhich analysis was undertaken the in-house Nd standard (l andM specpure Nd20ll gave 0.511604 ± 9 (10 of total population,n = 105), the La Jolla standard gave 0.511842 ± 15 and BeR-l yielded 0.512636 ± 16. Blanks are in the order of 1OOu 200 pgfor Nd. The average for the NBS987 Sr standard is0.710242 ± 12 (n = 56) and Sr blanks arc better than 2 ng.The 143Nd/144Nd composition of CHUR was taken as 0.512638and 147Sm/1441\das O. I967.

'Mathinna Beds Fig. lo Regional geological map showingthe pre-Australia-Antarctic breakupcontiguration of the Delamerian and RossOrogcns. MT, Manstield-MelboumeTrough; KC, Kuark Complex; CC, CoomaComplex; BCC, Bradley's Creek Complex;EYT. Eden Yalwal Trough

The Delamerian granites

This paper focuses on the Delamerian S-type granites (Tables 2and 3) and particularly uses the Kangaroo Island exposures todevelop an understanding of their origins and their relationshipto the Delamerian I-type granites (Table 2). We also emphasizethat the belt includes post-tectonic A-type granites (Turner et al.19(2). which are \ery important to the interpretation and under-standing of the geochemical variation of the entire Delameriangranites series.

The syntectonic to earliest post-tectonic Delamerian granitoidsrange from I-type to S-type granite. The I-types are biotite- and/or hornblende-bearing, often titanite-bearing, plagioclase-richdiorite, tonalite, granodiorite and granite. The S-types are biotite-rich, hornblende-free, muscovite-bearing granite to granodiorite.The granites occur as relatively small and scattered outcrops onKangaroo Island (Table 3, Figs 2 and 3) and northwards alongthe axis of the fold belt, to Mt Painter and the Peake andDenison Ranges in the far north (Fig. I). They are confined tothe Cambrian basin or to fault structures active during Cambriantime. Although exposed granite outcrops are small and scattered,recent high-resolution total magnetic intensity (TMI) imageryreveals that both syn- and post-tectonic granite intrusions aremuch more extensive to the east of the Mt Lofty Ranges in theMurray Basin, which is floored by Cambrian rocks beneathMesozoic cover. The syntectonic granites (e.g. the Rathjen

559


560 .I. D. FüDEN ET AL.

Platform Rocks (C)

• Kanmantoo Group (C)

Adelaidean (NP)

ffii] Basement

• Granites

f::"z:q Shear zone---~Thrust

J:Y Anticline

5> Syncline

\ Mafic or FelslcDykes

.Tanunda";; \ ~

Mt LoftyRanges

ReedyCreek

CARBONATE SHELFON CRATON MARGIN o 10 20 30 40 50

I I I I I I

km

Ni

Pt. Elliot

Victor Harbor

Cape Willoughby

~ "Vivonne Bay

Remarkable Rocks .. Cape Stun S~il BoomYounghusband RIver

Fig,2. A more detailed map of the southern Adelaide Fold Belt showing the granites and the locality of geographical and geological features mentionedin the text. The inset shows the postulated conliguration of lh~ Kanmantoo basin developed as a tear basin as proposed by Flöttmann el al. (1998);stippling. Precambrian bas~ment rocks; dark grey, KanmanlOobasin fill: cheeked pattern, granites.

has been involved in the evolution of I-type granites throughoutthe belt.

Gneiss; Foden el al. 1999) are 1- to S-type and haw tcctonicfabrics, metamorphic halos (with staurolite-cordicrite-andalu-site assemblages) and wcrc emplaced at middle upper-crustallevels (c. 12 km) between 516 and 490 Ma. With this series wealso include the deformed granites and associated dioritic-granodioritic rocks (the Wando granites) in the wesícrn VictorianGlenelg Province, regarding this as a geologically equivalentterrane (Gibson & Nihill 1992; Turner et al. 1993a). Thecompositional relationships between the three granite series aresummarized in Figs 4 and 5. Although dommated by I-type,there are also peraluminous S-type granites that arc sometimesmigmatitic (fig. 3). Examples of these occur at Vivonne Bay-Stun Sail Boom River in southern Kangaroo Island (Fig. 2) or atHarrow (Figs I and 2) in western Victoria (Turner e( ill. 1993a).

The Cape Willoughby granile. The Cape Willoughby granite hascontacts exposed on the northeastern and southeastern coasts andappears to be a roughly layer-parallel sheet-like intrusion into thehost Kanmantoo Group turbiditic sediments. This granite hasbeen dated at 509 ± 7 Ma by sensitive high-resolution ionmicroprobe (SHRIMP) V-Pb analysis of zircons (Fanning 1990).The granite has evolved thin, leucocratic, sill-like apophyses,which also intrude the host sequence layer parallel and which areboudinaged. The Cape Willoughby granite (Table 2) is a mildlyperaluminous, leucocratic, siliceous (> 75% Si02) S-type gran-ite. It is composed of quartz, K-rich microcline and relativelyscarce plagioclase, muscovite and biotite. It is clearly a fractio-nated granite and has low MgO, CaO, FeO*, Ti02, P20S, Zr, Srand REE, and is moderately rich in K20, Rb and U.Kangaroo Island

Southern Kangaroo Island provides exposures that give usparticularly good insight into the origin of S-type granite in theSouth Australian sector of the Delamerian belt. These data alsoprovide good evidence for the origin of a component that clearly

The Vivonne Bay-Siun Sail Boom River Migmatite Complex. Atthese localities the Kanmantoo Group is intruded by a K-feldsparmegacrystic biotite granite of transitional 1- to S-type composi-


GRANIT ES I'. DELAMERIAN OROGEN

tion (Table 2). This granite has limited on-land exposure andcrops out at a series of headlands between Vivonne Bay andRemarkable Rocks 30 km to the west (Fig. 2). At the VivonneBay-Stun Sail Boom River localities, the host Kanmantoo Groupmetasedimentary rocks are metamorphosed to upper amphibolitefacies and are partially mcltcd to produce a locally mobilizeddiatexitic biotite granodiorite (Fig. 3, Table 2). This granodinriteis fairly biotite rich and is host to numerous enclaves of unmcltedKanmantoo Group metasediments (Fig. 3). These enclaves areclearly of lithologies that were less favourable for mclt produc-tion at the prevailing temperatures and comprise both quartz-richlithologies and also very biotite-rich melanosome fragmcnts.These latter are the samc composition as some in Sitll melano-some layers bordering the biotitc granodiorite and are composedof biotite, plagioclase (An34-12), minor quartz and abundantaccessory apatite, zircon and monazite.

The locally produced biotite granodiorite melt was obviouslycontemporancous with the megacrystic granitc intrusion, as thereare good examplcs of mingling of I1ngers of the porphyriticgranite with the grey biotite granodiorite. The biotite granodioritehas a moderately developed foliation, which defines approxi-mately northerly dips of 30S0°. Development of the migmatitecomplex and emplacement of the megacrystic gramte (S03 ±4 Ma) are regarded as syn-Delamerian orogenic events.

In addition to the megacrystic granite and the biotite grano-diorite, these outcrops also expose bodies of highly leucocraticgranite (Table 2). These biotite-free rocks are composed ofquartz, K-rich microeline and muscovite and contain numerouspatches of almandine-rieh garnet gramte. These garnct leueogra-nites arc seen locally to minglc with the biotite granodiorite andthey are taken to represent extracted and fractionated nearminimum-temperature melts. The occurrence of muscovite as asolidus phase in these liquids constrains their pressure to beabove 3 kbar (e.g. Dymoke & Sandiford 1992). This granite hasbeen dated at S04 ± 8 Ma by SHRIMP U-Pb analysis of zircons(Fanning 1990). In this paper we also present new, two mica-apatite, internal Rb-Sr isochron data pointing to an age of487.4 ± 3.S Ma for this complex, probably reflecting thc age ofcooling below 3S0 DC(Table S).

The three rock types exposed at these localitics have thefollowing geochemical features.

Fig. J. Field photograph showing detail ofthe biotite granodiorite diatexite at VivonneBay- Stun Sail Boom River. Hammer forscale (35 cm).

( I) The diatexites range from biotite granodiorite to granite(Table 2) with between 62 and 7S% Si02. These show lineargeochemical relations on Harker diagrams (see Fig. 7, below),and with increasing silica content show depletion in mostelements including La, Ce, Nd Zr, MgO, FeO*, CaO, P20S,

Na20, Ti02. V, Sc, Cr and Ni. They also show enrichment inK20, Rb, Th and U. ASI values decrease with increasing silica.These granites have thc same characteristics as the S-typegranitcs and granodiorites from the Lachlan Fold Belt (White &Chappell 1988; Elburg & Nicholls 1995: Elburg 1996), withrelativcly low Na20 and CaO. Likewise, when also comparedwith I-types they have high Mg-number numbers and relativelyhigh V, Cr and Ni at givcn silica levels. In keeping with theknown clcvated phosphorus solubility in peraluminous melts(e.g. Bea el a!. 1992) they arc also characterized by somewhatelevatcd P20S levels at given silica levels. The linear array of thediatexites also includes the biotitc melanosome rocks. Comparedwith the diatexites. the intrusive porphyritic (megacrystic) graniteis more siliceous on avcrage and has lower ASI (Figs S and 7).Also in comparison with the diatexite series the intrusiveporphyritic granites are distinguished by their curving trends ofmore pronounccd depletion in MgO, CaO and most traceelements (exccpt for Rb and U) with increasing silica (see Fig. 8,below). Thesc trends have the garnet leucogranite (see Fig. 8) astheir endpoint.

(2) The biotite melanosome (Table 2) rocks occur both asenclaves distributed through the biotite granodiorite and in themarginal zones of thc granite sheets. These rocks are composedof dominant biotite, with plagioclase, quartz, apatite and acces-sory zircon, allanite and monazite. They have high K20, MgO,FeO*, Rb, Ba, V, Cr and Ni, low Si02, Na20 and CaO, and highAS! values (> I.S). Thesc rocks also contain very high lightREE (LREE) and U, reflecting the role of monazite as a sink forsome incompatible elements (Table 2).

(3) The garnet leucogranitcs (Table 2) formed as siliceous(7S% Si02), mafic-mineral-poor segregated melt bodies adjacentto the biotite granodiorite. They are very MgO and CaO poorand have ASI values close to 1.1. In comparison with the biotitegranodiorite (Figs 6 and 8) the leucogranites are depleted in Ba,Th, U, Nb, the REE, Sr, Zr, Ti, Y, V, Sc, Cr and Ni. They havesimilar Rb. K and Pb levcls, are slightly enriched in P20S, and

563


564

LOO DA ¡J. •C'Il • •

a.~oi

.l. D. FODEN ET AL.

10 .~ -I'!Job/À& •~ !

("') S j A. •O '& AA. A

C'Il • +O

¡f 6 J "J,S O I

';;!. 4 •• ~

i 2 SO

~--

'*300

Q)()

Ec.c.

0.2 -

>-Ec.c.

0.0

125

100 -•

•~

.c ,• • 4a. · ... ~·o·E 10

&. f\.#."A A .... _c. ..... ~c....;tJ

, .•1

O :~:~/·'.'IC'Il~Z~

. ~ .....o- 3 ~~ i~

O

SO 50

+----

•75

50••

have higher V/Th ratios. As illustrated by the MgO v. SiGlvariation diagram (Fig. 8) the composition of the garnet leuco-granites is not collinear with the trend of the biotite granodioriteand these melts can therefore not be considered as the end-member melt of the restitic trend that seems to control compos i-tional variation in those rocks.

25

O ~' --

zEc.c.

"------.h-----. .'. -30 -it, A (0.. ..A ~,>Â ••• Ó /

20. À ...... * .'iA •• .. }-. I... +\) i

10 u., ~/' j~ .. ...,'.*O~-_.~_-~:0.6 f----. ...~

A AA A <0; . <,.I.

A

••• 111. ".;.~.. •DA t .~ , 41

I lA •• ~ f ~!

0.2 i.&. I-type diorite-tonalite.~. ·1 i

I• I-type granod.-granite.o• J!I () S-type granite ,¡¡~

0.0 fi! A-type g,ranlte , @J~ _.

50

The Cape Hart-Cape Gantheaume composite dyke,\. More than20 dykes have been identified cutting the dcformed Kanmantoosequence on the SE coast of Kangaroo Island (Fig. 2). Thesehave a strongly clustered NW (325°) orientation and arevertically dipping. The Kangaroo Island dykes appear largely topost-date the Delamerian folding and contain no tectonic fabricor, rarely, a weak cross-cutting one. Although the early regionalfabrics that are cross-cut are in different orientations, theKangaroo Island dykes have exactly the same orientation as thedolerite dyke swarm at Woodside (Table 3) further north in thebelt, about 50 km east of Adelaide (Fig. 2). With their implica-tion of a common dilation direction, we attribute both dyke

60 70

Wt% Si02

,._--J70 SO

Fig.4. Selected Harker geochemicalvariation diagrams illustrating the generalregional geochemical relationships amongstthe Delamerian granite suites. Shadedtriangles, I-type diorit.:s and tonalite fromthc South Australian Reedy Creek andAnabama plutons and Wando Vale (nearHarrow in western Victoria); •• CambrianI-typc granites from the South Australianand Victorian Delamerian intrusions; 0, S-type granit.:s from the South Australian andVictorian Delamerian intrusions; shadedcircles, A-type granites from the SouthAustralian and Victorian late Delamerianintrusions.

60

Wt% Si02

complexes to the onset of the first stages of post-Delamerianbroadly eastward extension. These dykes have been dated atCape Gantheaume (Fig. 2) at 500 ± 7 Ma by SHRIMP V-Pbanalysis of zircons (Fanning 1990).

The Kangaroo Island dykes are between 0.5 and 4 m wide andare interesting because most are composite, with mafic and felsic(rhyodacite) components (Table 3). A few are either entirelymafic or felsic. The dykes commonly show mafic borders withfelsic interiors, implying initial injection of mafic melts. In mostcomposite dykes we also see mingling between the felsic andmafic magmas, the rhyodacite hosting convolute-amoeboidenclaves of basaltic melt, indicating that the two melts werecontemporaneous.

The felsic dyke rocks (Table 3) are porphyritic with roundedand embayed quartz phenocrysts together with phenocrysts ofplagioclase and K-feldspar. These are set in a fine, dark,recrystallized matrix. The mafic dyke rocks have plagioclaseintergrown with augite (mostly replaced by amphibole) in adoleritic texture.


1000

GRANITES 11\ DELAMERIAN OROGEN

2.0 r---.! •ppm Zr •

+++ •+ .... IIL.~...... + ~+~..\~ At. Jt.. t + _ . a..

'.~ À + .~1Il,:;1 + :

LI! ,:>..+

<"~.~~

--~----. .:J45 55 65 75

Wt% Si02

~t%Ti02!

+1.5

1.0100

0.5

10 0.045 55 65

Wt% 5i02• Cambria B .'

12 j Adelaid n asalts/ 1iI~~À e Fold Belt '---.

O ,'-WP' di""'-IO~·~ 1.3 fASi- OC-lIl81 • I ,~~-~ '1'~ 1.1 ~- 1- :\it O I

Ci' I ..... +... .... - -#fi'_. • •• 'r;: 4 • ~ • 000 O 9 ;",-1Jl0l~~y:~~~~':;:-g""¡I'0:7 _J/IN+,.... ::~

O _. A-type granite' .. =1 AO .~ ..5 0.5 --;.~

Wt% MgO 10 40 -~-- 60Wt% Si02 70 80

The felsic componcnts of the dykes have Si02 contents thatrange between 63.5 and 73.5%. At the more felsic end of thisspectrum they hme compositions that are the same as those ofthe Vivonne Bay anatectic biotite granodioritcs with about 72-73% Si02 or the megacrystic intrusive granite (Fig. 7. and seeFig. Il, below). The felsic component of the composite dykes,however, extends towards lower Si02, dacitic compositions withlinear increases in CaO, MgO, FeO*, Ti02. Ni. V, Cr and Sc. andwith decreasing ASl values (Figs 7 and Il). These trends extenddirectly towards the mafic magmas and are most probably due tomingling and hybridization. This conclusion is further borne outby the Nd and Sr isotopic compositions, which show progressivevariation as a function of changing bulk rock composition,appearing to mix between Kanmantoo Group type crustal valuesand the mantle values of the mafic rocks (Figs 7 and 9)

Nd- and Sr-isotope ratios

Nd- and Sr-isotopic data are presented in Table 4 and Figs 7, 9and 10. The combined suite of 1- and S-type granites showswidely variable isotopic compositions. They range from l-typegranodiorite and tonalite, which differ strongly from KanmantooGroup country rocks or any pre-Adelaidean basement, in havinglow initial 87Sr/86Sr ratios (c. 0.7060) and high initial (Nd (+ l to- 3), through to siliceous S-type granites including those atVictor Harbor (e.g. Milnes et al. 1977), Cape Willoughby orVivonne Bay (Kangaroo Island), which have initial X7Sr/86Srratios in excess of 0.715 and (Nd values as low as - 13.

The wide variation in Sr- and Nd-isotopie compositions of thesyn- to late orogenie 1 and S-type granites is also partlycorrelated with major and trace element variations (Fig. 10). TheNd- and Sr-isotopic composition of the I-type granites fallsbetween the contemporary mafic magmas from the belt and thefield of S-type granite (Figs 9 and 10). On a regional hasis, some

Fig. 5. Selected Harkcr diagrams and CaOv. MgO covariation illustrating therelationship between the regionalcompositional variation of the Delameriangranites and the composition of Cambrianbasalts from the belt. Symbols as in Fig. 4;., Cambrian basalts from the Delamerianbelt (Faden et al. 20(2). AS!' aluminiumsaturation index (molar AI20]/{Na20 +K20 ~ CaO)).

I-type granites define a trend of increasing silica content withdecreasing (Nd value (Fig. 10. Trend Cl) like the isotopicvariation observed in the Kangaroo lsland composite dykes (Figs7 and lO). The Kangaroo Island composite dykes (Figs 7 and 10)range from the hybridized less siliceous dacitic dyke rocks withinitial (Nd values in the range -8 to -6 to the uncontaminatedS-type rhyodacite with ENd values of -13, which are the sameas the Kanmantoo Group sediments and the Vivonne Bay-StunSail Boom River S-type granites. The composite dykes providedirect examples of crust-mantle interaction (Trend Cl. Fig. lOb)interpreted as a major regional control on granite composition.

Significantly. the Kanmantoo Group has very different (Ndsoofrom the Palaeoproterozoic Precambrian basement. which hasCambrian ( Nd values lower than -20. with the bulk closer to-25 (Fig. 9). No granites in the Delamerian Fold Belt haveinitial Nd isotopic compositions lower than those of the Kanman-too Group. None approach the values of the more ancicntbasement that forms the predominant crust only 15 km to thewest or NW of thc Cambro-Ordovician belt. In comparison withthe I-S-type granites, the post-tectonic A-types show much lesscrust-like isotopic compositions, with initial (Nd values similarto Bulk Earth (-4 to + I).

Discussion

Key issues in understanding the origins of granites include theidentity of their sources, the nature of the processes that modifythe primary melts, and the physical causes of melting. In theLachlan Fold Belt the initial recognition of the geochemical andmineralogical contrast between 1- and S-type granite led to theconcept that these granites were geochemical 'images' ofspecific, unique I-or S-type crustal source regions (Chappell etal. 1987; Chappell & White 1992). This view, however, has beencountered by a number of studies (Gray 1984; Elburg 1996; Keay

565


reach some very finn conclusions about the mechanisms andsettings of the production of S-type crusta I melts.

Comparisons hetv.'een the S- and I-type granites

The 1- and S-type suites in the Delamerian belt have exactly thesame general geochemical characteristics and distinctions asdefined in the Lachlan Fold Belt (Chappell & White 1974,1992). The I-type granites are relatively CaO. Ba and Sr rich. Asa function of increasing silica they show depletion in FeO* (total

566 J D. FODEN ET AL.

1000• mafie I-type

• fraetionated I-typediatexite biotite

O granodiorite100

10

(a) 1.0 ., /i. I

Sr P NdZrSm Ti yRb BaTh U Nb K LaCePb

1000

100

d'atexite biotiteO I 'tgranodion eKanmantoo

~ ~,. Grp. metased.\Pi,. ~ I~ '\ ,fb-M \

\/u . ~~~ \1 ~vtJ

cu-c::-ca.!EQ.cuE >ca;m·-E.¡:

Q.

10

(b)

RbBaThU NbK LaCePb Sr P NdZrSm Ti y

1000

.! 100-c::-ca.! E 10Q.cuE >ca;m·-E.¡:

Q.

díatexite biotitelO granodionte

garnet leueo-~'\¿-;"'\,,Ä ·,1 '\, • granite\ -,\......../'\::v\ r '.,/ \U---.d, \.,\, " il, i,.

\, ~ \,d, ¡ • / \\,~~/\ \ ' "D'\ , I \ '\ / " ,

: I ~ \ I ~. ~ /~

(c) 0.1RbBaTh U Nb K LaCePb Sr P NdZrSm Ti

el al. 1997; Collins 199í\), which (mostly with the benefit of aexpanded isotopic datasets) recognize that granite genesis in theLachlan Fold Belt involved both mantle and crustal components.Because of their many similarities, this paper develops theconcept that granites have similar sources and arise by similarpetrogenetic processes in both belts. A major source of debate inthe Lachlan Fold Belt arises from uncertainty about the sourcesof the granites. By contrast, in the Delamerian belt. for rcasonsdiscussed below, we can much more confidently identify thecmstal source component of the granites. and this allows us to

Fig.6. Primitive mantle-normalizedincompatible trace element diagrams.Normalizing values from Sun &McDonough (1989). (a) Comparisonbetween a matic, unfractionated I-type(sample GM153, Table 2, a tonalite fromthc Anabama intrusion), a typical I-typeti'actionated granite (sample 9 I-CYH 1,Tahle 2) and the diatexitic S-type biotitegranodiorite (VB2000-7, Table 2) from theVivonne Bay Complex. (b) Comparisonbel ween the S-type biotitc granodiorite(VB2000-7j and a representative exampleof the Kanmantoo Group metagreywaekes .(c) Comparison betwccn the S-type biotitegranodiorite (VB2000- 7) diatexite and thehighly fractionated S-type garnetleucogranite from the Vivonnc BayComplex (KI-29, Table 2).y


GRANIlT'; IN DELAMERIAN OROGEN 567

1.4 400 J.fractional crystallisationrestite unmixing R ~, \J1.2

. r I~~ ~~ ..300 fi ~ngl mixing1.0 mmg mg mlxmg ~= R >

D comp~-site dykes (felsic)E 200c.

0.8 c.~:. • composite dykes (mafic)

0,6 f>· • diatexite 100 . RaJ¡( o Woodside mafic dykes restite unmixing ~ •• ~~;

0.4 o ., , ,40 50 60 70 80 40 50 60 70 80

10- 16-mantle melt

~. ~glingl mixing

5 e ~O 12-!ti

O -:~iXing lrend

U~ 8-O....

-5 - ~_t\lno ~~ resl,le

-10 . Kanmantoo~--4- R ::J unmixing

fractional -O- Ocrystalistion [J~~ .... R

-15 -Group~,,_ _ .

O'40 50 60 70 80 40 50 60 70 80

wt% Si02 wt% Si02

Fig.7. Geochemieal variation of AS!. V, I Nd,<oo,and CaO with ,iJica. rhyodacites and dacites from the Cape Hart eomposite dykes: +, thc matic rocksfrom the Cape Hart composite dykes;-, dolerites from the similarly orientated and assumed age equivalent Woodside dyke complex (Fig. 2: Faden el al.2002): shaded circles, biotite granodiorite diatcxites from the Vi\onne Bay Stun Sail Boom River Complex. In the ,Nd v. Si02 tigure the mixing trend isthat produced between the garnet leucogranite S-type melt and the most primitive Woodside dolerite (Table 3). The shaded tield is that of the Kanmantooc1astiesedimentary rocks.

Fe as FeO). Sc. V. Cr. Ni. Ti and P20ó (Figs 4 and 5). Althoughat lower silica contents the I-type suitcs show very little variationin Mg-number, they have declining Mg-number values at highersilica levels (Fig. 4). Within the I-type suite silica enrichment isalso accompanied by K20, Rb, Ba. Th, U and Ce enrichment.

Although the most felsic granites show geochemical continuitybetween the 1- and S-ty-pe fields, the less siliceous S-types haverather high and constant Mg-number values and tend to havedecreasing ASI with increasing Si02 (Figs 5. 7 and llb (TrendR)). The I-type suites (Figs 5, 7 and lib (Trend B)) showpositive correlation between ASI and Si02. Like the LachlanFold Belt S-types (Chappell el al. 1987), and best illustrated bythe Kangaroo Island diatexites (Figs 7 and II a), the most maficS-type granites have the highest ASl values as well as the highestVand MgO. These also have the highest Pb, Yand REE (Fig. 4).In comparison with the I-types, the S-types are also more Na20and CaO poor at given silica levels (Figs 4 and 5) and aremarginally more MgO and PlO, rich. More fractiollated (IowMg-number) I-types tend 10 have lower Sr and P than siliceousS-types.

The S-types (defined to have ASI values> l.l) all have initial(Nd values of < - 7.5 and initial "Sr/g6Sr values mostly> 0.714. The 1- and A-types all have (Nd> - 7.5 and initial87Sr/g6Sr < 0.714 (Table 4, Fig. 9). Geochemically, the [-typeseries falls between the fields of mafie rocks and of fClsic truegranites (Figs 4, 5 and II). As demonstrated by Nd-isotopie v.Mg-number or Si02 variation (Fig. 10: Trends A and B), thefelsic endpoint of this trend ranges from S-type granite with localcrustal isotopic compositions to very [ractionated A-type granite

with more primitive Nd isotopic ratios. The trend towards localcrust is one of increasing 87Sr/8óSr and decreasing 143Nd/144Nd(Trend B, Fig. 9).

We regard the field of S-type granites as bracketed by threelimiting evolutionary trends. Two of these incorporate isotopicvariability with declining initial 87Sr/b6Sr and increasing initial143Nd/144Nd (Trend e, Fig. 9). Of these two trends one showscontinuity with the 1- to A-type granitoid suites (Trend Cl, Fig.lOb) and the other is towards mafic magmas (Trend C2, Fig.lOb, compare Kangaroo Island composite dykes). The third S-type granite trend (Trend R in Figs 7, 8 and Il) is one with aconstant initial (Nd value (-13.5 ± 0.5) and is typified by thediatexites from Kangaroo Island. This trend is towards lowersilica and high Mg-number and ASI values (Figs 4 and Il), anddiverges [rom the I-S granites in having higher Pb, Ni, andPlO,.

Source and origin of the S-type granites

Source 01' Ihe S-type granites. The Kangaroo Island outcropsreveal four distinct types of S-type melt behaviour: (I) fonnationby partial melting of the Kanmantoo Group metasediments underrestite-melt control (diatexite); (2) mingling of the S-typediatexite with intrusive more I-type granite magma; (3) segrega-tion and fractional crystallization of S-type melts to produceleucogranites; (4) mingling and hybridization of S-type meltswith intruding basaltic melts.

The outcrops at Vivonne Bay and Stun Sail Boom Riverprovide dircct evidence for the local production of S-type granite


Delamerian I-type granite (Faden et al. 1999) was not from theMeso- to Palaeoproterozoic basement but had age frequencypopulation patterns identical to the very distinctive patterns ofthe Kanmantoo Group (Ireland et al. 1998).

It appears that the roje of the Kanmantoo Group in theDelamerian Orogen was like that of the Ordovician flysch as themajor source of the S-type granites in the Lachlan Fold Belt(Collins 1998; Keay et al. 1997). We note that in the LachlanFold Belt, although earlier interpretations had considered theOrdovieian flysch as a geochemically inappropriate source forthe Lachlan Fold Belt S-type granites (Chappell 1996), recentzircon age frequency data indicate that the inherited zirconpopulations in both loo and S-type granites are indistinguishablefrom those that could be derived from this source (Williams etal. 1992). The combination of Nd-Sr isotopic data and field andgeochemical arguments has led recent workers (Collins 1998;Keay et al. 1997) to support the model put forward by Gray(1984), that the Lachlan Fold Belt S-type endmember is indeed apartial melt of the Ordovician flysch. These models have beendeveloped and tested at Cooma in NSW (Fig. I), where theCooma Granodiorite is undisputedly derived by partial meltingof the Ordovician flysch (Collins & Hobbs 2001) and whereprocesses and geological relations are like those we describe atVivonne Bay-Stun Sail Boom River from Kangaroo Island.

J. D. FüDEN ET AL.568

5

/R

4

Biotite -richmelanosome Calculated Restite

34% biotite, 37%An (35), 28%• ~ Qtz, apatite

Source Isediments •Il diatexite

megacrystic SIlgranitegarnet leucogranite

3

2Mixing ofanatectite andintrusive granite

1

o55 65 75 85

Wt % Si02

Fig.8. MgO Y. SiO: variation illustrating the relationships between theKangaroo Island S-type complexes. Shaded envelope: local compositionof Kanmantoo Group sedimentary rocks (psammo-pelites greywackes):large shaded circJèS,biotite granodiorite diatexite from Viv<lnncBay andStun Sail Boom River; shaded squares. the intrusive, porphyritic graniteat Vivonne Bay and Stun Sail Boom River; small shaded circles, gametleucogranite: crossed squares, biotite melanosome rocks from \"ivonneBay.The 62% restite trend shown illustrates the extent of variationresulting from extraetion of 62% of a restite composed ofbiotite (35%)+ An], plagioclase (37%¡) + quartz (28% I - minor apatite ¡Table 6).Large shaded square, bulk composition of this restite. R-R is the rcstite-controlled diatexilc trend.

melts (diatexite) from the Kanmantoo Group metasedimcntaryrocks (Fig. 3). These granites haw the same Nd Sr isotopiccomposition (initial [Nd = -13 ± 0.5) as the metasedimentsfrom which the field relations indicate they were derived, withslightly lower ;\1g-number or higher Si02 as anticipated as aresult of partial melting (Figs 7 and 8). The nonnalizedincompatible element pattern of the diatextite is also the same asthat of the Kanmantoo sediment, distinguished from the othergranites by having a significant positive Pb anomaly (Fig. 6).Thus the combination of field, geochemical and isotopic evidencepoints strongly to the Kanmantoo Group as Ihe S-type end-member in at least the southern Delamerian Fold Belt. At itsmore felsic end. the biotite granodiorite-granite diatexite serieshas compositions the same as the rhyodacite magmas from thecomposite dykes (Figs 7 and Il) and like the transitlonally 1- S-type intrusive megacrystic granite. This similarity suggests thatthe rhyodacite dyke magmas and the megacrystic intrusivegranite also originated from the same type of source as thediatexite, reftectmg similar processes of melt production andsegregation.

It is surprising that although the Kanmantoo basin is formedby faulting of the well-characterized pre-Adelaidean Palaeopro-terozoic cratonic basement rocks (Fig. 2). the latter do notcontribute to the crustal melt components identifiable in theCambro-Ordovician S-type granites (Fig. 9). This conclusion isalso supported by the observation that inherited zircon in

Petrogenesis of the S-type granites. Our results establish theKanmantoo Group as the S-type granite source and, as discussedabove, we postulate that intra-suite geochemical variation iscontrolled both by mixing or mingling with I-type and maficmagmas (Trend e in Figs <)-11) and by variable melt-residueunmixing (Trend R in Figs 7, 8 and II). This melt-residue(restite) unmixing trend is best illustrated by the diatexites fromKangaroo Island. Least-squares mixing calculations model thistrend (Figs 7 and 8) as one controlled by the assemblage quartz-biotite (Mg-number 0.75)-plagioclase (Anl7 i-apatite. As sum-marized by the MgO-Si02 relationships (Fig. 8), the range incompositions covered by the granite-granodiorite diatexite seriesrepresents about 45% restite extraction. The bulk composition ofthe calculated restite also matches that of the biotite melanosomerocks (Figs 8 and Il). These relationships imply that generationof S-tY']Jegranite magmas is controlled by a near-isothennal meltproduction and segregation from metasedimentary migmatitecomplexes (restite control).

At Vivonne Bay there are also garnet leucogranites withmarked depletions of trace elements such as Sr, Zr and Ti, whichindicate they have evolved by melt segregation and fractionalcrystallization. Least-squares mixing calculations (Table 6) in-dicate control by quartz-plagioclase (c. AnI2)-biotite fractiona-tian. This assemblage differs from the restite calculatedpreviously in having less An-rich plagioclase and a largerproportion of quartz relative 10 biotite and plagioclase. For arange of trace elements the D values required to yield theobserved concentrations in the garnet leucogranite, based on thecrystallization proportion calculated in the least-squares calcula-tion, were estimated (Table 6). These indicate that most ele-ments, including the REE, are highly compatible, confinning theinvolvement of accessory phase saturation (zircon. monazite,apatite).

In addition to the restite-dominated processes, we have alsonoted that some Delamerian S-type granite compositions aredisplaced towards the I-type field. This trend is probably one ofcontamination. It tends towards higher Nd- and lower Sr-isotopicratios. Because it is improbable that a near minimum temperatureS-type melt will assimilate any solid basalt, this contamination


GRANITES IN DELAMERIA:'>I OROGEN 569

8

• Cambrian Mantle..~'. ',0 ;'>'¡_.;~ II:

3 .

.\. • A-type granites

---~ • I-type granites-.

~ "·2

¿

e C~\' •• <> S-type granites <:

e<:

ln .~\~ • Kanmantoo Sed ~'-' """O ~,C••• • z'Z -7

(¡,J• >" .,• _,o, "*

-12,~B(l11>- ....

...-: >

0.7 0.71 0.72 0.73

87Sr í 86Sr (500 Ma)

O _l Delamerian;. granites

-5 -..-10 _~ Kanmantoo Group

-15 _..

~20 -

Gawler eralonBasement•

-25 _ Ir,¡/' ..-30, ..

0.7•

0.8 0.9

87Sr I 86Sr (500 Ma)

0.74

Fig, 9. ('ic!,SOOI v. 8'Sr!R6Srl<oo, variation of the various sub-groups of the Delamerian granites. Symbols as indicated. Trend B is calculated AFC trend(formulation of Powell 1984) modelling the contamination and rrogressive crystallization of a Cambrian basaltic parent magma by Kanmantoo Groupmetasediment (parent basalt: Nd 5 ppm, (è\d¡;OOI ~5, Sr 200 prm, 87Sr!8"Sr0.7035: crustal contaminant: Nd 30 ppm, (NdiSOlli-13, Sr 180ppm, 87Sr!86Sr0.7220; bulk distribution coefficients: ~d .,= 0.5, DST == O.S) Tick marks indicate 10%crystallization intervals. ,. == 0.45 (assimilation rate!erystallization rate). Trend C is direct mixing trend between a erustal partial melt of Kanmantoo Group metasediment with a composition of the VivonneBay garnet leucogranite and Cambrian basalt. It should be notcd that this is the samc model as illustrated in Fig. 7. Inset: (Nd¡5001v.8iSr/86Sr¡5001variation of the combined Dclamerian granites. The diagonally striped cnvelope is range of available unpublished data (45 samples) from Gawler Cratonbasement granite gneiss, granite, tèlsic and intermediate volcanic rocks and metasedimcntary samples (., data points from Turner el al. 1993b).

of the crustal melts was probably the result of mingling andhybridization caused by magma mixing. The process of directmixing is clearly illustrated by the observed hybridism in thecomposite dykes and yields a modelled close match of Sr-Ndisotopic trends (Trend C, Fig. 9), and of isotope-major clementcovariation (see ,Nd v. SiOl, Figs 7 and 10). This hybridism canresult by mixing either with mafic magmas (as in the case of theCape Hart composite dykes; Figs 7 and Il) or, as indicated by.field relations at Vivonne Bay, with granite intrusions of more 1-type composition.

We note that the Lachlan Fold Belt S-type granites also showtrends like those of the Kangaroo Island diatexites. This trend ofincreasing ASI with decreasing silica (Trend R, Fig. 11) hassometimes been used to deny the role of mixing between mantleand peraluminous crusta] melts, suggesting such a trend wouldrequire a more peraluminous mafic endmember (Chappcll eT al.1987). Our data clearly highlight that this is a restite accumula-tion trend associated with the generation of crustal melts and isseparate from, but might coexist with, mafie mingling trendstowards lower ASI (Trend C. Fig. Il).

Source and origin of the I-type granites

Trends of variation of ,Nd v. Mg-number or silica content (Fig.10) highlight two extremes in evolutionary path originating at themafic or low-silica end of the I-type series. One of these leads tothe younger, largely post-tectonic A-type granites (Foden el al.1999) and the other to the syntectonic S-type granites (Fig. 10).The trend to the A-type field is towards very low-Mg-number,silica- and incompatible element-enriched granites (Figs 4, 10)whose "Jd and Sr isotopic compositions are only slightly shiftedfrom those of the mafic magmas (Fig. 9). These trends are

produced dominantly by fractional crystallization (Turner et al.1992; Turner & Foden 1996), apparently with only limitedcrusta I assimilation. The second trend links the I-type granites tothe field of the S-type granites and of the Kanmantoo sediments.This is defined by marked decrease in ,Nd and by lesser declinein Mg-number (Figs 4, 9 and lOa). Given their convergence withthe S-type granite field, these trends indicate that the I-typegranites evolved from the assimilation of Kanmantoo-type crustalrocks either by contemporary cooling and fractionating maticmagmas, or by mixing of these mafic magmas with the S-typegranite magmas derived from these sediments. At its mafic end,the I-type granite field is in direct geochemical and isotopiccontinuity with the Cambrian mafic rocks.

Because all the I-type granites have initial (Nd values greaterthan those of any known basement rock sequences, the Ndisotope data require direct involvement of mantle melts in theproduction of the Delamerian I-type granite magmas. It is alsoclear from the inflected, curvilinear geochemical trends of the 1-type granites that fractional crystallization processes must alsobe involved. In particular, the transition from mafic tonalite tosiliceous granodiorite is one of initial enrichment of P, Sr, LREE,Zr and Pb, later followed by depletion (Figs 4 and 5). Thesetrends are explained as the result of fractional crystallization,consistent with changing Kd values as progressive (particularlyaccessory) phase saturation took place during falling tempera-tures. Minerals precipitated during the middle to late stages ofcooling are K-feldspar, zircon, apatite, titanite and allanite. Theseinflected trends cannot be explained by simple mixing of felsiccrusta I melts with mafic magmas as occurs in the KangarooIsland composite dykes. Thus our evidence supports the com-bined roles of both fractional crystallization and assimilation,favouring the evolution of the I-type suites by an AFC process


570 J. D. FODEN ET AL.

10 C-----------~ .. 1.75• • : Fractionation ASI.. Kanma~too Group

5 ~ .JIll. ~" .. sediments_ • 1.55 biotitego. • A ~,?!." restitelO '~I r Xy R melt:; ....:. ~. ./¿ 1.35 _ ~ extraction

Z -5 .. \. "•• !i.~"·Bw • I-types • Assimilation 1 15<> S-types •.".". . Ir,~¡¡ij~ R-10 • A-types C~·' X:ofcru~t t------- ,.......,...±-,,:,ij¡¡"~

• basalts . '." ."i' Y- of maflc composite dyke-15 I u Seds. ,.",':" melt 0.95 mingling tre~~ O

O 0,2 0.4 M # 0.6 0.8 1 0.75

10 Iii gil':' • • • I 145rw 55 65 75 85 ..... ~ Fractn.

~ O .. e:.:. .. . I 1.3

¡ x..' S,.. I 1.1 - _ ...-5 •• &

w y ~\KI dykes ••••

-101" .......C1· O v'!f'l~-I 0.9......

Assimilation

0.7-15 I I

40 50 60 70 SO1ii·1

Iilili~~..-=."!••'.--- .--

-8 I

+6 ENd(500) I

70 80

"-1

Fig.lO. (a) ,Nd"I'1i1 v. Mg-number variation of the various sub-groups ofth.: Delamerian granites from th.: entire b.:lt in South Australia andVietoria Shaded squar.:s with erosses and shaded ticJd, KanmantooGroup s.:diments; , S-type granites: +, I-type granitcs: ,post-teetonicA-type granites: •. Cambrian basalts. Data for thc A-type granitcs fromTurner cl al. 11992' anti Turner & Faden 11996). Trend A isfractionation-dominant trend from basaltic parent melt and kading to theA-type granites (Iow assimilation trend). Trend 13is high assllllilation andfractionation trend eommencing from basaltic parent melt and leading tothe ¡-type granites. Trend C is cl1Jstal pal1ial melt (of KanmantooGroup), contaminated by mixing with either basalt or I-type granite. (b)cNd,50o¡ v. wt% SiCh Symbols as in (a). Trend Cl is crw,talmclt mixingwith basaltic contaminant. Trend C2 iscrustal melt mixing \vith ¡-typegranitc contaminant. Trend of Kangaroo Island composite dyke'sindicated.

(DePaolo el al. 1992). This process started with basaltic magmas(Table 3) of E-MORB affinities (Foden et al. 2002) whosecompositions e\olved towards felsic compositions as a result ofcrystallization and contamination leading to siliceous I-typegranites. These resultant felsic magmas are more depleted in ;\Ji,V, Cr and Se than the S-type granites.

AFC trends of Nd- v. Sr-isotopie variation (Fig. 9, Trend B)were calculated using the equations of DePaolo (1981) andPowell (1984). The I-type trend was modelled usíng the Sr and

0.5

40 50 60Wt% Si02

Fig. II. (a) ASI v. SiO. variation showing the Kangaroo Island S-typecomplexes and the eomposite dykes. Symbols as in Figs 7 and 8. Stripedtie Id, Cape Willoughby granite: 'basalts' is the field of dolerite dykerocks from Cape Hart and Woodside. Trend R··R is the restite-S-typemelt (diatcxite) unmixing trend (as in Figs 7 and 8). Trend C (as in Figs9 and 10) is the trend ofmatic contamination of the felsic dyke rocks. (b)Regional ASI v. Si02 variation (symbols as in Fig. 4) also summarizingNd-isotope variation in I type granite. Trcnd B (as in Figs 9 and 10) isthe AFC dominant trend producing I-type granites: Trend C' (as above) isthe mixing trend with I-typc or mantle melts displacing S-type (crostalmeh) compositions towards higher (Nd values; Trend R is the rcstite-diatexitc trend (as above and in Figs 7 and 8) showing extraction of S-type melts from their sources.

Nd concentrations and isotopíc compositions of the most primi-tive Delamerian basalt as the mantle endmember (Table 3), andthe mean composition of the Kanmantoo Group as the crustalendmember. The D values used for both Nd and Sr were < I,mimicking the crystallization of an assemblage dominated bymoderately Ca-rich plagioclase, hornblende and c1inopyroxene.The AFC equation contains a parameter r, which is the ratio of


572

5511Muscovite

.I. D. FOOEN ET AL.

Table 5. Isotope composition o/minera! sepurales: SIun Sai! Boom River; Rb-Sr isochron

Sample:Phase:

551 IApatite

5516Apatite

"Sr""'Sr SRM 087 ¡, 0.71028474 ± 0.000043; 14'l\d,'44Nd La Jalla is O.51l572R5J. 0.000021 Corrected to LaJalla 1431\d:144Nd O.Sll6~40.

" Sr/S6Sr± 1 SDRb (ppm)± I SOSr (ppm)±ISOs'Rb!shSr% SD Rb/Sr14'Nd/I44Ndl"= I SD\id (ppm)Sm (ppm)14'Sm¡I.j4Nd£ T = O((Ù;500 Ma

07211 J 70.00005950480.234

218.640.0530.0710.5121220.000052

570.26227.09

0.2407-10.1

12.9

the rate of assimilation to the rate of crystallization. Curve B inFig. 9 was determined using a high r value (0.45). With theseparameters the resultant hybrid melt reached (Nd - 7 after about50% crystallization. If a much smaller r value were to be usedthe curve would only reach the tie Id of post-tectonic A-typegranites with c. (Nd - 2 after 60'10 crystallization.

Although, in explaining the Lachlan Fold Belt granites, Keayel al. (1997) and Collins (1998) doubted the role of A fT and toa large extent emphasized mixing, the Delamerian (Nd v. Mg-number and (Nd v. wt% Si02 (Fig. 10) trends highlight the vitalrole of this process. The model we propose for the Delameriangranites is similar 10 those used in earlier studies to explain thecomposition of many Cenozoie volcanic and intrusiw rocks inthe western North and South America (Hildreth 1981; Farmer &DePaolo 1983, 1984; Hildreth & Moorbath 1988; DePaolo el al.1992). Those studies emphasized the interactive role of contem-porary mantle-derived matie magmas with isotopically distinctivecontinental crust. Likewise, exposures of very deep crustal maticmagma chamber sequences trom the Ivrea Zone (Voshage el al.1990) in the Alps provide direct support of the role of processesof interaction between mantle melts and the crust, closelymirroring those modelled by Sparks el al. (1984) and Huppcrt &Sparks (1988).

Whereas Farn1er & DePaol0 (1984) attempted to model theextent to which crust v. mantle contributions to granitic magmasvaried as a function of crustal thickness and rate of basalt magmaflux alone, DePaolo et al. (1992) emphasized that the relativeextent of these two contributions is more dependent on thethennal state of the wall roeks. High wall-rock temperatures willresult in high assimilation rates. If this reasoning is applied tothe granites in the Adelaide Fold Belt, we must conclude that theearlier synconvergent granites have been generated at sites wherecrustal wall rocks were hotter than the much less contaminatcdpost-convergent A-type granites.

We propose that at the onset of the Delamerian Orogeny,crustal deformation was focused, partly by thennal weakening(Sandiford el al 1992; Stüwe et al. 1993), at the Kanmantoobasin. This led to erustal thickening and the stalling of risingdense. mantle-derived matic melts near the Moho. The AFCprocesses we have documented are the consequence of thethennal and geochemical interaction between stalled maticintrusions and continental crust (Kanmantoo basin fill). Highcrustal temperatures resulted in high assimilation ratcs (high r

5511Biotite

0.8208520.000086

385.9900419

77.900.022

14049I

2.1897300.000119

677.720.382

10.640.003

210.91I

SSll (Rb-Sr isochron)Age = 487.4 ± 3.5 MaInitial ratio = 0.72062

MSWD=0.25

0.7216130.0000723.520.066

147.510.0280.07I0.5120670.000030

618.66227.56

0.2223-11.1-12.8

values) yielding I-S type granites. The S-types are byproducts ofthis process, being melts of the wall rocks adjacent to the coolingmatic to I-type magma bodies. Post-convergent extension thenallowed the matic melts to rise higher in the crust (as illustratedby the Kangaroo Island composite dykes) as a result of crustalthinning and dilatant fracturing. These matic magmas may thenhave ponded at shallow depths and fractionated without crustalassimilation, yielding the A-type granites (low r values). Thistrend of diminishing crustal component at the late orogenie stageis like that reported in the caldera-forming rhyolites of mid- tolate Cenozoic age of the western USA (Perry et a!. 1992) fannedduring Basin and Range extension.

Granite magmatism in a periodically extended convergentplate margin

The Kanmantoo Group sediments were rapidly deposited in adeep, narrow basin (inset, Fig. 2). Basin formation was asso-ciated with extension and deeompressional mantle melting. As aresult, the Kanmantoo sediments were intruded by Early Cam-brian matic magmas (Table 3). The onset of deformation broughtan immediate transition to granite magmatism ranging from 1- toS-type. During the syntectonic phase, when Delamerian crustalthickening was maximal, matie magmas were stalled in the lowercrust and underwent hybridization. In the subsequent post-tectonic extensional phase (Oliver & Zakowski 1995) these maticmelts rose much closer to the surface and cooled and fractionatedin contact with cold upper-crustal rocks, experiencing onlylimited crustal contamination.

The cycle of events on Kangaroo Island highlights the closeinterrelationship between evolving strain and magmatic histories.The earliest (509 ± 7 Ma) granite is the fractionated S-type atCape Willoughby, where tield relations imply this is a deformedsill that intruded layer parallel to the host Kanmantoo Groupmetasediments. This granite intruded early in the Delamerianorogenic history as a result of mantle heating of the Kanmantoosedimentary pile. The sill-like form of the intrusion implies thatthe minimum compressive stress (a3) was vertical (lithostatic),and that the NNW-directed tectonic driving force (a¡) was justdeveloping. The Vivonne Bay-Stun Sail Boom River suites(504 ± 8 Ma) were syntectonic crustal melts formed in situ as aresult of advective heat transfer by the plutonic I-S granites.These in turn must have formed by basalt-Kanmantoo Group


Table 6. Leasl-sqllares mixing models

GRANITES IN DELAMERIAN OROGEN

Least squares model for diatexitic granite

Granodioritc Granitc Modelled granite

VB2000-5 ME-37867.76 76.00 76.06 Biotite0.64 0.29 0.29 Plagioclase (An)s)

14.32 12.00 12.06 Apatite4.90 1.96 1.91 Quartz0.07 003 0.03 A % Crystals2.66 0.76 0.73 +2.04 1.03 0.84 B % Liquid=2.12 2.17 1.85 Granite3.50 4.94 4.310.12 0.10 0.16 C Biotite granodiorite

Sum of res2

Least squares model for garnet leucogranite

Granite Leueogranite :VlodelledKi-07 Ki-29 Jcucogranite

Biotite71.88 76.55 76.62 Plagioclase (Anll)0.5 I 0.04 0.04 Apatite

13.59 13.61 13.6 Quartz2.9 0.33 0.34 A % crystals0.05 0.01 0.01 +U4 0.19 0.19 B % Liquid=1.66 0.55 0.47 Garnet granite2.77 2.81 2.794.04 5.13 5.05 C Porphyritic granite0.23 0.2 0.2 Sum of res2

Si02Ti02AI20)FeOMnOMgOCaONa20K20P20;

Si02Ti02AhO)feOMnOMgOCaONa20K20P20j

% Restite

21.50 34.6522.90 36.91

0.15 0.2417.50 28.2062.05

37.95

1.000.545

% Restite

14.9017.200.40

18.3050.8

29.3033.90

0.7036.00

49.7

1000.018

Calculated bulk distribution eoeftieicnts (D) in generation of garnet leucogranite (model 2 above)

ElementRbSrBaZrNb:'-IdSmySe

DlA39573.9533062.952.83.3

Restite indicates calculated mineral assemblage normalized to 100% melt-free.

metasediment interaction at greater depth. Finally the compositedykes (500 ± 7 Ma) mark the weakening of the main tectonicstress (all, and the development of roughly east-west extension(a) now horizontal). Importantly, the composite dykes indicatethat S-type magma chambers are established deeper in themetasedimentary tectonic pile as a result of cntstal melting, butalso that these have then been invaded by mantle-derived maficmelts. The mafic melts provide clear evidence of the heat sourcefor crustal melting. Their transport in dykes to shallower crustallevels not previously invaded by mafic melts must be a responseto the transition to extension and crustal thinning. Our Rb-Srdate of 487 ± 3 Ma (Table 5) is the same as the magmatic age ofthe post-tectonic A-type granites and represents the age ofexhumation and cooling of the terrane. This uplift event is alsoobserved elsewhere in the Delamerian belt at the same time,including Tasmania (Raheim & Compston 1977) and created theerosive highland source of the Lachlan Fold Belt flysch blanket(Turner et al. 1996).

We conclude that pre-convergent rifting played a vital role inthe subsequent geological and magmatic evolution of the Dela-

merian Belt and that the S-type granite source was provided byrecently introduced basin fill in a deep, localized trough. Thisbasin fill melted in preference to the older crustal walls of thebasin probably because of its water content and fertility. Theevolution of the Delamerian belt from Kanmantoo Group deposi-tion through convergent deformation to tenninal uplift and post-tectonic extension lasted a total of 40 Ma (525-485 Ma), mark-ing a cyele of oscillation from extension to compression andback to extension.

The cycle of rapid transitions from extension to compressionand back to extension in the Delamerian belt continued to be afeature of the tectonic evolution of eastern Australia throughPhanerozoic time, particularly during the Palaeozoic evolution ofthe Lachlan Fold Belt (Gray 1997). There, examples of specificlocalized extensional (or transtensional) sedimentary basins withgeological histories very like that of the Kanmantoo basininelude the Melbourne-Mansfield Trough (O'Halloran & Cas1995), the Eden Yalwal Trough (Fergusson et al. 1979) and theMathinna trough in northeastern and western Tasmania (Burret!& Martin 1989) (Fig. 2). There are examples of linear belts of

573


574

CHAPPELL, B.W .. WHITE. AJ.R. & WYBORN, D. 1987. The importancc of residualsource material lrestire) in granite petrogene:;is. Journal ol Petrology, 28,1111-1138.

COl.LI"S. WJ. 1998. Evaluation of petrogcnctic models for Lachlan rold Beltgranitnids: implications for crusml architecture and tectonic models. Al/s!ra-lian Journal o/ Earlh Sciences. 45, 483-500.

COLL.t"S, WJ. & HOBRS. B.E. 2001. What caused the Early Silurian change frommafic to ::;ilicic (S-typc) magmatism in the ea~tcrll Lachlan Fold Belt?AusTralian ./ournu! (~lEarth Sciences. 41'1.25-41.

CO"lEY. PJ .. EDWARDS, A .. HINT-:, R .. MORRISON. E & W1NDRIM, D. 1990. Therc:gional tectonics of the Tasman orogenie system. eastern Australia. Journalo(S,rl/clural Geolo};¡', 12, 519-544.

COOPER, J.A .. JD1Kl"S. RJ .. COMPSTON. W. & WILLIAMS. LS. 1992. Ion-probezircon dating of a mid-Early Cambrian tuff in South Australia. JOl/rnal o( IheGeological Sodel\". London. 149. 185-192.

DALZ1EI .. LD. 1991. Pacific margins of Laurentia and East Antarctica-Australia as:.1 conjugate rift pair: e\'ldcnce and implications for an Eocambrian super-continent. Cie%g)", 19, 598- 601.

DEPAOLO. DJ. 1981. Trace dement and isotopic clTects of combined wall rockassimilation and fractional crystallization. Earth and Plane/m:r ScienceLelter.l. 53, l8~-202.

DEPAoLO, DJ .. PERR", EY. & B"LDRIDGE. W.S. 1992. Crustal versus mantlesources of granite magmas: a two-parameter model based on 1\d isotopicsludie~. Transactions (?( the Royal SocieZl' ql Edinhurgh: Earth Sciences, 83,439 ·446.

Dn'OKE, P. & SA"llIFORD, M. 1992. Phase relationshIps in Buchan Facies Seriespelitic assemblages: calculations with applications to andalusite-slaurolltt:parageneses in the "10unt Lofty Ranges. South Australia. COllll"ibllliollS loMine",lo".!' and Pelmlo};v. tI O, 121-132.

ELDUR(;, M.A. 19%. Evidence of isotopic equilibration between micrograniwidenclaves and host granodiorite. \Varburton Granodiorite, Lachlan Fold Belt.Australia. Ulhos. 38. 1-22.

EI.Rl1RG. M .A. & N¡CHOLLS, LA. 1995. The origin of micrograniroid enclaves inthe S-type \Vdson's Promontory Batholith, Victoria: evidence for magmamingling, Australian Journal o(Earth Sciences. 42, 423-435,

FA~:\IING, C.~'1. 1990. Single grain daling ql Ci granite sample from CapeWï/loughhy. Kangaroo Island. Prise Laboratories, Australian National Cni-vcrsity Progress Rcport 89-060. S.A. Department of Mines and Energy, OpenFile Envelope 8828, 29- 32.

FARMER. G.L. & DEPAOLO. DJ. 1983. Origin of Mesozoic and Tertiary granitc inthe western United State..; and implications for pre-~"esozoic crustal structure:I. Nd and Sr isotopic studies in the geocline of the northem Great Basin.Jo"rn,,! ,,(Geophysical Researclz, 88, 3379-3401.

FARM"R. G.L. & DrPAoLO, DJ. 1984. Origin of Mesozoic and Tertiary granite inthe western United States and implications for pre-1v1esozoic crustal structure:II. Nd and Sr isotopic studies of Cu- and Mo-mineralized granite in thePrecambrian eraton. Journal (~lGeophysical Research, 89, 141-160,

FERGllSSO"'. CL.. Co\S. RA., COLLI"S. WJ. & 5 orHERS 1979. The UpperDevonian Boyd Yolcanic Complex. Eden. New South Wales. Journal o( IheGeological SOciell" 'j'Australia, 26, 87--106.

FERGUSSON, CL., CAS, R. & SrEWART. I.R. 1989. Ordovician turbidites of thelJorham Group. easrern Vicroria: sedimentation in deep-marine channel-leveecomplexes. Auslmli,," ./ourn,,1 o( Earlh Sciences, 36. 1-12.

FLEMI"G. PD. & WHITE, AJ.R. 1984. Relationships bctwccn deformation andpartial melting in the Palmer migmatites, South Australia. A ustralian Journalo(Earrh Sc/ell,es, 31, 352· 360.

flÖTTM.\NN. T.. HAI"ES, P .. JAGD. J .. JA\lES, P., BUP!'RIO. A. & GCM. J. 1998.Formation and reactivation of the Cambrian Kanmantoo Trough, SEAustralia: implications for Early Palacozoic tectonics at eastern Gondwana'splaIe margin. Journ,,! o{lhe Geological Sodell·. London, 155,101-115.

FLÖTn1.'NN. T., J",,'s. P., MENPES, R. & 5 OTHERS 1995. Kangaroo Island SouthAustralia: :'.train and kinematic partitioning during Delamerian basin andplatform re3ctivation. Allstralian Journal (~rEarth Sciences, 42. 35-49.

FI.ÖTTYt'\NN. T .. LutEs. P., ROGERS. J. & JOHNSON, T. 1994. Early Palaeozoieforeland thrusting and basin reactivation at the Palaeo-Pacific margin of thesoutheastern Australian Prccambrian eraton: a reappraisal of the structuralevolution 01" the southern Adelaide Fold-thrust Belt. Tectollophvsics, 234,95-116.

FaDEN. J.. SANDlFORD. M., DOUCiHERTY-P"GE. J. & WILLIAMS, I. 1999. Geochem-istry and geochronology of the Rathjen Gneiss: implications for the earlytectonic evolution of the Delamerian Orogen. Australian Journal q( EarthSciences, 46, 377-389.

Fonol. J.. SONG. S.-H., Tt:RNER. S.P., ELBCR(;. M.A .. S.\ftTH. P .. VA~ DERSULDr, B. & y"" PEN(illS. D. 2002. Geochemical evolution of theIithospherie rnanlle beneath S.E. South Australia. Chemical Geology, t82.663-695

FODEN. LO., TUR"'ER. S.P. & MORR1SON. R. 1990. The teclOnic implications of theDelamerian magmatism in South l\llstralia and \vestern Victoria. In: lAGO,

.1. D. FODEN ET AL.

metamorphic complexes whose themlotectonic history is alsovery like that of the Kanmantoo. (e.g. the Late Ordovièian-EarlySilurian Bradley's Creek-Cooma Complex-Kuark Complex belt(Carson & Rickard 1998)). It may be that granite magmatism inthe Lachlan Fold Belt, as in thc Delamerian belt, is a reflectionof deep, narrow rifts, buried under a widely distributed flyschblanket.

Summary and conclusions

In the South Australian sector of the Cambro-Ordovician Ross-Delamerian Orogen granite production is largely confined todeep, Early Cambrian, sediment-filled basins where granites areassociated with mafic rocks. The syntectonic suites have compo-sitions forming a continuum between 1- and S-types. After thecessation of convergent deformation at c. 490 Ma an abrupttransition to a bimodal magmatic association of matk intrusionsand felsic granites and volcanic rocks of S- and A-type affinitiesoccurred.

The formation of S-type magma is well illustrated by exposedmigmatite complexes on the south coast of Kangaroo Island.These magmas originated as in sitll partial melts of the EarlyCambrian sediments where these were inttUded by either maticor I-S granite magmas. These migmatite complexes weremingled with the intrusive magmas that provided the heatsources for crustal melting. Our data indicate that S-type magmaproduction and migration is controlled by variable segregation ofrestite and melt. The direct role of mantle-derived melts in theprocess of crustal partial melting is indicated by composite S-type rhyodacite-dolerite dykes.

The syntectonic to early post-tectonic I-type granite magmasresulted from the assimilation of crustal melts by intruding maficmagmas. This process was coupled with fractional crystallizalion(AFC). This stage of magma production occurred when highlower-crustal temperatures wcre developed during the later stagesof convergent deformation, resulting in high crustal assimilationrates. The A-type granites cluster towards the mantle endmem-ber. These were formed when post-convergent extension allowedmafic melts to rise and pond at shallow depths and fractionatewithout crustal assimilation.

We are grateful to technical staff in the Department of Geology andGeophysics at Adelaide University. In particular, we thank D. Bruce forhis assistance in the isotope laboratory and .1. Stanley for XRF data. MEacknowledges the support of an ARC post-doetoral fellowship. The paperwas signiticantly improved by constructive reviews by R. Maas and C.Bames.

ReferencesBu. F.. FERSH 1\ 1ER. G. & CORRll(;I'. LG. 1992. The gcochemistrv of

phosphorus in granite ro<.:ksand the effect of aluminium. Litho .._ 29, 43-56.Bt'RRETT. CF. & M~RTlN, EL 1989. Geology anel Minero! Resources of

Tasmania. Geulogica~ Society of Australia, Sp~cial Publication, 15.CARSO". LJ. & RICKARD, M.L 1998. Early recumbent fold hinge III [he Bradley's

Creek T\1etamorphic Complex, near eanbara, Australia. Australia/l Journal ofEarth Sciences. 45. 501-508.

CAS. R.A. 1983. A RevicH' qllht' Palaeogeographic and Tectonic {)cl'e!opmem ofIhe Palaeo:;oÙ.' Lachlan Fold Belt cd" SoU/heustprll Al/stra!io. GeologicalSociety lJf Au~lraliJ., Special Publication. 10.

CH.-WPELL, B.W. 1996. Magma mixing and thc production (,r composirionalvariation within granite suites: evidence from the granite~ of southeasternAustralia. Journal o/Pelrology, 37. 449 ,)70.

CHAPPEl.!., B.W. & WHITE. AJ.R. 1974. Two contrasting granile Iypes. PacificGeology, 8, 1;.J -174.

CHAPPELL. B.W. & WHITE. AJ.R. 1992. I- and S-lype granites in rhe Lachlan FoldBelt. Transactions of the Roya! Sociel)'o/Edinburgh: Earlh SCll!n('('s. 83. 1-26.


GRA'<lTES IN DELAMERIAN OROGEN 575

LB. & MOORE, P.S. (eds) The Evollllioll o{ a Laie Precaml>riollEarlyPalaeo=oic Rift Complex: the Adelaide Geo;;ync/ine. Geological Snciety ofAustralia, Special Publication, 16,483-495.

GIBSO". G.M. & !"H1LL, D.N. 1992. Glenelg River Comple" western margin ofthe Lachlan Fold Belt or extension of the Delameriao Orogen into \VestemVictoria? Tectonophysics. 214, 69-91.

GRAY. C.M. 1984. An isotopic mixing model for the origin of granitic rocks insoutheastern Australia. EarTh and Planetary 5;cience Leucrs. 70. 47--60.

GRAY. D. 1997. Tectonics of the southeastern Australian Lachlan Fold Belt:structural and thermal aspects. [II: BCRG, J.-P. & FORD. M. (eds) Orogen.'"Through Time. Geological Society. London. Special Public"tions. 121.l49-ln

HILDRETH, ....'1. 1981. Gradients in silicic magma chambers: implications forlithospheric magmatism. Journal o{ Ccoph\'sical Research. 86IBII).lIltS3· tOI92.

HILDRETH, Vv', & !\ilooRBATH, S. 1l)88. Crusta] contributions to arc magmatism inthe Andes of central Chile. Contrihutions to AJineralogv mul Perr%gy. 98,455-4R9.

HUPPERT, H.E. & SPARKS, R.SJ. 1988. The genera'ion of granitic magmas byintrusion of basalt lOto cominental crust. Journal or Pelmlogt·, 29. 5~~ -624.

IRELAND. T .. FLÖTTMAN~. T., FA""tNG, .\1.. Gmso,"" G. & PRE'SS. W.V. 1998.Development of the Early Palacozoic Pacific margin of Gond\vana fromdetrital-zircon ages across the Delamerian Orogen. CeologJ' 26. 243 246

JENK'NS. RJ. & SANDl](JRD, M. 1992. Observations on the teetomc e\"lution ofthe southern Adelaide Foldbelt. TeclOllophysics, 214, 27 -36.

KEAY. S.M., COU.1NS. WJ. & MeCuLLocH. M.T. 1997. A three-component mixingmodel for granitOld genesis: Lachlan Fold Belt. eastern Austr~lia Geolog .....25,307-310.

MANCKTEt.UW. N.S. 1990. The structure of the s"u,hem Adelaide Fold Belt. Ill.JAUO. J.B. & MOORE, P.S. (eds) The Evolutioll o{ a LaIe Precalllhrilll1-EarlrPalaeo=oic R(¡i Complex: the Adelaide Geos.vnclin('. Cieological Society ofAustralia. Special Publication. 16,483 .. 495.

MlLNES. AR., CaMPsToN. W. & D,\JL Yo B. 1977. Pre- to syn-tectonic emplacememof early Palaeozoic granites in southeastern South Australia. J()urnal of ¡heGeological Soc/en o(Allstralia, 24. 87 -lOó.

MOORES. E.M. 1991. Southwest US-East Antarctica ISWEAT) connection: ahypothcsis. Geologv, 19,425-428.

MORRISON. R.S. & FaDEN, J.o. 1990. A zoned pluton in the Peakc and DennisonRanges, South Australia: implications to Mid Cambrian magmatisIll in theAdelaide Geosyncline. [Il: JAGO J.B. & MOORES, P.S. leds) The El'ollltion ufofate Precambrian,--Ear(~' Proterozoic R~ft Complex: The Adelaide (J(,osVllcline.

Geological Societ) of Australia, Special Publications. 16, 450~4M.MÜNKER. C & COOPER, R.A. 1995. The island arc setting of a f'.;ew Zealand

Cambrian volcano-sedimentary sequence: implications for the evolution ofthe SW Pacific Gondwana fragments. JOllrnal o{Geology, 103. óX7-"'OO.

MëNKER, C & CRA\\I0RD. AJ. 2000. Cambrian arc evolu'ion along thc SEGond\vana active margin: a "iynthesis from Tasmania- New Zealand~Austra-lia-·Antarctiea correlations. l,'ctollics, 19,415-432.

OFFLER, R. & hEM,,,r,, P.D. 1968. A synthesis of folding and metamorphism inthe Mt. Lofty Ranges. South Australia. Journalot rhe Geological .l'ociel.'" orAlIslralia, 15,245 ..266.

O'HALlORAN. GJ. & CAS, R.A. 1995. Evidence for syndcpositionai detnrmationfrom the La,e Devonian of the Mansfield Basin, easl-cenlral Victoria.Australian JOl/rnal ot Eanh Sciellces. 42. 5X1- 596.

OLIVER. N. & ZAKOWSK1.S. 1995. Timing and geome'ry of deformation. Iow-pressure metamorphism and anatexis in the eastern tvlt Lofty Ranges:the possible role of extension. Austmlian Journaf (~lEarth S('it'l1ces, 42.501-507.

PERRY, F., DEPAOlO, D. & Bo\[ llRlllGE, W.S. 1992. l"ltopic evidence for thedecline in crustal contributions to caldera-fonning rhyolites of th.: \vestemUnited States dUrIng the middle to late Cenozoico Geological Sociuv o/America Ahstracl.I. 23, A441

POWELL CMeA. PRElSS. W .. GATEHOUSE. C, KRAPEZ, B. & LI. Z. 19~4. SouthAustralian reeord of a Rodinian epieontinental basin and its mid-'ieoproter-

ozoic break-up (700 Ma) to form the Palaeo-Pacific ocean. Teclonophysies,237, 113-140.

POWELL R. 1984. Inversion of the assimilation and fractional crystallisation (AFC)equations: characterization of contaminants from isotope and trace elementrelationships in volcanic suites. Journal qt Ihe Geological Society~ London,141,447-452.

PREISS, \V.V. 1987. The Adelaide Geosyncline: Late Proterozoic Sfratigraph~v,sedimentation palaeontology and tectonics. Geological Survey of AustraliaBulletin, 53.

RAHEIM. A. & CaMPSTON, W. 1977. Correlations betwecn metamorphic events andRb-Sr agcs in metasediments and eclogite from western Tasmania. Lithos.10,271-289.

SANlJIFORD,M., FODEN. 1., ZHOU. S. & TUR"ER. S. 1992. Granite genesis and thcmechanics of convergent orogenie belts with application to the southernAdelaide Fold Belt. Trallsuetiom o{ the Royal Sociely of Edinburgh: EarlhSciences, 83, 83~93.

SA"D1FORD. M., OLIVER. R.L., MILLS. KJ. & ALLE", R.V. 1990. implications forthe metamorphic evolution of the Kanmantoo Group. In: JAGO, J.B. &MOORE, P.S. (eds) The Emlulion o{ a Laie Precamhrian-Early PalaeozoicRiji Complex: Ihe Adelaide Ceo.\yncline. Geologica! Society of Australia,Special Publication, 16,483-495.

SPARKS. R.S.! .. HUPPERT. H.E. & TURNER, J.S. 1984. The fluid dynamics ofevolving magma chambers. Philosophical Transactions o{ the Royal Societv,London, Series A, 310, 511-534.

ST¡;WE, K.. SA"lJIFORD, M. & POWELL. R. 1993. Episodic deformation andmetamorphism in low-?, high-T terranes. Geology, 21, 829-832.

SUN. S.-s. & MeDoNOUGH. W.F. 1989. Chemical and isotopic systematies ofoceanic basalts: implications for mantle composition and processes. [n:SACNDERS. A.D. & NORRY, MJ. leds) Magmatisl/1 in Ihe Ocean Basins.Geological SocielY. London. Special Publications, 42. 313-345.

TeRNER. S. & FaDEN, J.O. 1996. :-'1agma mingling in late-Delamerian A-typegranites at Mannum, South Australia. Mineralogy and Petrology, 56.147 -169.

TURNER, S., KELLEY. S.P., VANDE"Bl.'RG, AH.M., FaDE". J., SANDIFORD, M. &FLÖTTMAN". T. 1996. Source of the laehlan fold belt flysch linked toconvective removal of the lithospherie mantle and rapid exhumation of theDelamerian-Ross fold belt. Geologt" 24, 94t-944.

TURNER. S.P. 1996. Petrogenesis of the late Delamerian gabbroie complex at BlackHill. South Australia: implications for eonvective thinning of the lithosphericmantle. Mineralogt· and Petmlogy: 56, 51-89.

TURNER. S.P., ADAMS. CJ., hÖlTMA"". T. & FODEN. J.D. 1993b. Geoehemicaland geochronological constraints on the Glenelg River Complex, westernVic'oria. AUSlralian Journal ol Earlh Sciences, 40, 275-292.

TURNER, S.P .. FaDE", J.o. & MORRISO", R.S. 1992. Derivation of an A-typemagma by fractionation of basaltic magma: an example ¡¡-om the PadthawayRidge South Australia. Lirhos. 28.151·179.

TURNER. S.P .. FaDEN. YD., SANDlFORIJ,M. & BRUCE, D. 1993a. Sm .. Nd isotopicevidence for the provenance of scdiments from the Adelaide Foldbelt andsoutheastern Australia with implications for crustal growth models. Geochi-mica et Cosmochimica Acro, 57,1837-1856.

VON DER HORCH, e.e. 1980. Evolulion of late Proterozoie 10 early PalaeoloicAdelaide Fold Belt. Australia: comparisons with post-Permian rifts andpassive margins. TeclOnophysics, 70. I 15- 134.

VOSHAGE. H., HOFMA"". A.W .. MAZZCCCHELLI, M .. R'VALENTt. G., St"IGO'. S.,RAezEK, L & DEMARCHl, G. 1990. ISOlOpic evidence from the tvrea zone fora hybrid lower crust formed by magmatic underplating. Nalllre, 347,731-736.

WHITE, AJ.R. & CHAPPELL, B.W. 1~8X. Some supererustal (S-type) granites of theLachlan Fold Belt. Transactions of" the Royal Sociely o( Edinburgh: EarlhScience.', 79,169-182.

WtLlIAMS, IS, CHAPPElL, B.W., CHE", Y. & CROOK. K.A.W. 1992. Inherited andde'rital zireons--vital clues to the granite protoliths and early igneous historyof southeastern Australia. Transactions of" the Royal Sociely of Edinburgh:Earth Sciences, 83, 503.

Received 25 July 200 I; revised typescript accepted 4 February 2002.SCientific edi'ing by Sally Gibson


granite production inthe delamerian orogen, south...

Documents