fractionation vs magma mixing in wangrah suite a-type granites_lfb

8/12/2019 Fractionation vs Magma Mixing in Wangrah Suite a-type Granites_LFB

1/20

Fractionation vs. magma mixing in the Wangrah Suite A-type

granites, Lachlan Fold Belt, Australia: Experimental constraints

Kevin Klimma ,, Francois Holtz a, Penelope L. King b

aInstitut fr Mineralogie, Universitt Hannover, Welfengarten 1, D-30167 Hannover, Germany

b Department of Earth Sciences, University of Western Ontario, London ON, Canada N6A 5B7

Received 17 October 2006; accepted 25 July 2007

Available online 24 August 2007

Abstract

The Wangrah Suite granites (Lachlan Fold Belt, Australia) reflect different stages of differentiation in the magmatic history of an

A-type plutonic suite. In this study we use experimentally determined phase equilibria of four natural A-type granitic compositions of

the Wangrah Suite to constrain phases and phase compositions involved in fractionation processes. Each composition represents a

distinct granite intrusion in the Wangrah Suite. The intrusions are the Danswell Creek (DCG), Wangrah (WG), Eastwood (EG) and

Dunskeig Granite (DG), ordered from most maficto most felsicby increasing SiO2and decreasing FeOtotal.

Experimental investigation show that the initial water content in melts from DCG is between 23 w t . % H2O.If the DCG isviewed

as the parental magma for the Wangrah Suite, then (1) fractionation of magnetite, orthopyroxene and plagioclase (20 wt. %) of the

DCG composition, leads to compositions similar to that of the EG; (2) further fractionation of plagioclase, quartz, K-feldspar and

biotite (40 wt. %) from the EG composition, leads to the DG composition. These fractionation steps can occur nearly isobaricallyand are confirmed by bulk rock Ba, Sr, Rb and Zr concentrations.

In contrast, the generation of themost abundant WG compositioncannot be explainedby fractional crystallisationfrom the DCG at

isobaric conditions because of the high K2O content of this granite. Magma Mixing could be the process to explain the chemical

distinctiveness of the Wangrah Granite from all the other granites of the Wangrah Suite.

2007 Elsevier B.V. All rights reserved.

Keywords: A-type granite; Experiments; Crystallisation; Fractional crystallisation; Magma-mixing; Australia

1. Introduction

The origin of A-type granites, a term proposed by

Loiselle and Wones (1979) to distinguish high K2O +

Na2O, anhydrous and anorogenic granitic rocks from

calc-alkaline I-type granites, has been subject of much

controversial debate. Metaluminous to weakly peralu-

minous A-type granites are characterised by high SiO2,Ga/Al, Fe/Mg, Zr, Nb, Y and REE (except Eu) and low

CaO (Loiselle and Wones, 1979; Collins et al., 1982;

Whalen et al., 1987; Eby, 1990; King et al., 1997) but

more felsic, fractionated A-type granites have compo-

sitional characteristics similar to fractionated I-type

granites (King et al., 1997, 2001). A-type granite

magma can be generated by partial melting of crustal

igneous rocks of tonalitic to granodioritic compositions

(e.g.,Skjerlie and Johnston, 1992, 1993; Cullers et al.,

1993; Patio Douce, 1997; Dall'Agnol et al., 1999) and

Available online at www.sciencedirect.com

Lithos 102 (2008) 415 434www.elsevier.com/locate/lithos

Corresponding author. Department of Earth Sciences, University of

Bristol, WillsMemorialBuilding, Queen'sRoad,Bristol BS81RJ, UK.

E-mail address:[email protected](K. Klimm).

0024-4937/$ - see front matter 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.lithos.2007.07.018
mailto:[email protected]://dx.doi.org/10.1016/j.lithos.2007.07.018http://dx.doi.org/10.1016/j.lithos.2007.07.018mailto:[email protected]


2/20

the depths of melting and the prevailing oxygen fugacity

(fO2) could account for the diversity in A-type granitic

compositions (e.g.,Patio Douce, 1997; Dall'Agnol and

de Oliveira, 2007). Alternately, A-type magmas may be

derived by extreme fractionation from basaltic magma

(Turner et al., 1992; Frost and Frost, 1997; VanderAuwera et al., 2003). A-type magmatism is accompa-

nied by high heat flux caused by either deeper mafic

magmatism or crustal extension (rifting; e.g. Haapala

and Rm, 1992; Emslie and Stirling, 1993; Turner and

Foden, 1996; Frost and Frost, 1997). The parental

magmas are not strictly anhydrous and may have water

contents of several wt.%, similar to other types of

granitic magma (Clemens et al., 1986; Dall'Agnol et al.,

1999; King et al., 2001; Bogaerts et al., 2003) but the

role of changing melt H2O contents during crystal-

lisation is poorly known.

Fractional crystallisation is one of the major differen-

tiation processes in A-type granitic systems (Collins et al.,

1982; Clemens et al., 1986; Chappell et al., 1987; Whalenet al., 1987; Creaser et al., 1991; King et al., 1997, 2001).

Previous fractionation models in A-type granitic systems

(e.g., Lumbers et al., 1991; Martin et al., 1994; Rajesh,

2000; Asrat and Barbey, 2003) are based on geochemical

observations in the natural rocks without respect to

mineral compositions and abundances in the parent

magma at modelled physical conditions. This is due to

the sparse experimental data in natural A-type granitic

Table 1

Compositions of natural rocks and synthesised glasses

Wangrah granites a Starting glass compositions

AB412 AB422 AB421 AB401 AB412b AB422b AB421 AB401

n = 10 n = 11 n = 10 n = 14

wt. %

SiO2 70.45 72.53 75.32 76.67 71.89 0.40 73.85 0.44 75.65 0.39 77.56 0.44

TiO2 0.54 0.37 0.15 0.09 0.55 0.04 0.38 0.05 0.17 0.04 0.10 0.02

Al2O3 13.26 13.08 12.71 12.10 13.60 0.11 13.42 0.17 13.04 0.19 12.39 0.33

FeO 4.06 2.50 1.62 0.92 4.14 0.15 2.58 0.13 1.71 0.09 1.13 0.11

MnO 0.08 0.05 0.06 0.02 0.07 0.06 0.05 0.04 0.06 0.05 0.06 0.06

MgO 0.61 0.48 0.23 0.05 0.62 0.02 0.46 0.01 0.19 0.02 0.04 0.02

CaO 1.93 1.31 0.90 0.53 1.96 0.02 1.32 0.06 0.96 0.05 0.51 0.04

Na2O 3.39 3.32 3.36 3.25 3.38 0.10 3.34 0.14 3.40 0.16 3.45 0.17

K2O 3.98 4.88 4.61 5.12 3.79 0.11 4.62 0.10 4.80 0.15 4.76 0.16

P2O5 0.18 0.12 0.05 0.01

rest 0.23 0.17 0.12 0.09

Total 98.71 98.81 99.13 98.85 100.00 100.00 100.00 100.00

Qz 34.5 33.9 38.3 38.6

Ab 36.0 32.6 31.5 29.3

Or 29.5 33.5 30.2 32.2

Ab/An 3 4 6 10

ppm

F n.d. 1040 670 n.d.

Rb 146 228 216 187Sr 128 87 56 14

Ba 505 215 148 61

Zr 535 313 190 116

Nb 26 21 20 11

Ga 22.2 21.2 19.6 22.4

Y 75 64 77 125

La 64 53 28 24

Ce 145 121 63 64

Zn 81 52 52 28

Zr-sat.T[C] c 897 843 806 764

a Source of data:King et al. (1997, 2001): AB412 sample of Danswell Creek Granite; AB422 and AB408 samples of Wangrah Granite; AB421

sample of Eastwood Granite; AB401 sample of Dunskeig Granite.b

AfterKlimm et al. (2003).c Calculated zircon saturation temperatures afterWatson and Harrison (1983).

416 K. Klimm et al. / Lithos 102 (2008) 415434


3/20

systems (Clemens et al., 1986; Dall'Agnol et al., 1999)

giving insufficient information about crystal-melt equi-

librium as a function of temperature, pressure, volatile

content (e.g. H2O, F), oxygen fugacity (fO2) or bulk

composition.

Generally crystal-melt equilibrium experiments are

performed with the goal of constraining parameters such

as primary water content in the melt, temperature,

pressure andfO2(e.g.,Clemens et al., 1986; Dall'Agnol

et al., 1999).Clemens et al. (1986)andDall'Agnol et al.

Fig. 1. Whole rock (a) FeO vs. SiO2, (b) CaO vs. SiO2, (c) CaO vs. K2O, (d) Sr vs. SiO2, (e) Ba vs. SiO2 and (f) Rb vs. SiO2 diagrams showing

compositions representative for the various granite intrusion of the Wangrah Suite. Data sources: King et al. (1997, 2001 and unpublished data).

Crosses: starting compositions used in the experiments of this study (AB421 and AB401) and ofKlimm et al. (2003; AB412 and AB422). Additional

symbols in a, b and c: solid circles: residual melt compositions in experiments with AB412 ( Klimm et al., 2003); white circles: residual meltcompositions in experiments with AB421; grey circles: residual melt compositions in experiments with AB401.

417K. Klimm et al. / Lithos 102 (2008) 415434


4/20

(1999) performed experiments using one bulk composi-

tion, assumed to be representative of a studied pluton.

Thus, these studies provide information on only one part

of the crystallisation history because fractionation pro-cesses may lead to the formation of different bulk granitic

composition within a suite. To model fully fractionation

processes in a magmatic suite, a systematic investigation

of different compositions representing different magmatic

stages is necessary.

This study uses phase diagrams of four compositions

representative of the whole spectrum of granitic composi-

tions from the A-type granite Wangrah Suite, Lachlan

Fold Belt, SE Australia, to model the fractionationprocesses in this typical A-type granitic suite. Together

with previous experimentally determined phase diagrams

for the less evolved Danswell Creek and Wangrah Granite

(Klimm et al., 2003) phase relations of the more evolved

Eastwood and the Dunskeig Granite are presented. The

Table 2

Experimental results for the composition AB421 at 200 MPa and logfO2NNO

Run T[C] Duration [days] XH2Oina NNOb cH2O [wt. %]

c Results (+Gl, Fl) d

61 700 32 1.0 0 6.3 Bt, Ox

60 700 32 0.80 0.19 5.2 Bt, Pl (An18.3Or9.5), Kfs (An1.3Or68.3), Qz, Ox

76 725 32 0.89 0.10 5.8 Bt, Ox77 725 32 0.80 0.19 5.4 Bt, Pl (An22.5Or8.1), Qz, Ox

65 750 32 0.90 0.09 5.8 Bt, Ox


80 760 32 0.80 0.19 5.1 Bt, Pl (An28.6Or5.5), Ox

84 775 32 0.69 0.32 4.2 Opx (En34.9Wo2.7), Bt, Pl (An30.1Or5.4), Ox

85 775 32 0.57 0.49 3.9 Opx (En27.4Wo2.6), Bt, Pl (An21.6Or11.2), Ox

75 800 8 0.89 0.10 5.1 Ox

74 800 8 0.79 0.20 4.5 Opx (En39.2Wo2.5), Ox

73 800 8 0.71 0.30 3.9 Opx (En37.3Wo2.8), Ox

72 800 8 0.60 0.32 3.2 Opx (En32.8Wo3.2), Pl (An27.2Or7.3), Qz, Ox

52 800 8 0.51 0.59 2.7 Opx (En28.6Wo3.0), Pl (An24.3Or9.9), Qz, Ox

59 850 8 0.71 0.30 3.8 Ox

58 850 8 0.59 0.46 3.0 Ox

57 850 8 0.49 0.62 2.5 Opx (En39.3Wo2.9), Pl (An33.4Or7.0), Ox56 850 8 0.42 0.75 2.1 Opx (En38.7Wo2.8), Pl (An29.2Or8.0), Qz, Ox

a XH2Oin= H2O/(H2O+CO2) loaded in the capsule (in moles).b NNO = logfO2(experiment) logfO2(NNO;Chou, 1987). For water saturated experiments at temperatures between 700 and 850 C in the

CSPV the fO2=NNO. For H2O-undersaturated charges, a maximum possiblefO2is calculated as log fO2=logfO2(aH2O= 1)+2log XH2Oin.c Water contents of glasses determined following the difference method calibrated with six hydrous rhyolitic standard glasses.d Phases present in run products. Mineral abbreviations as given byKretz (1983): Fl: fluid; Gl: glass; Pl: plagioclase; Kfs: K-feldspar; Qz: quartz;

Ox: FeTi-oxide; Bt: biotite; Opx: orthopyroxene; Glass and fluid were present in all runs.

Table 3

Experimental results for the composition AB401 at 200 MPa and logfO2NNO

Run T[C] Duration [days] XH2Oina NNOb cH2O [wt. %]

c Results (+Gl, Fl) d

62 700 32 1.0 0 6.3 Bt, Ox

63 700 32 0.81 0.18 5.1 Bt, Pl (An13.0Or9.7), Kfs (An1.9Or57.8), Qz, Ox78 725 32 0.90 0.09 6.0 Bt, Ox




81 760 32 0.90 0.09 5.0 Ox

82 760 32 0.78 0.22 4.6 Ox


86 775 32 0.69 0.32 4.2

87 775 32 0.58 0.47 3.4 Opx (En30.6Wo1.8), Pl (An9.4Or25.6), Kfs (An3.2Or50.3), Qz, Ox

71 800 8 0.88 0.11 4.9

70 800 8 0.75 0.25 4.1

69 800 8 0.61 0.43 3.2 Ox

68 800 8 0.51 0.59 2.7 Pl (An11.6Or18.6), Qz, Ox

ad seeTable 2.



5/20

four compositions (Danswell Creek, Wangrah, Eastwood

and Dunskeig Granite) were interpreted to result fromfractional crystallisation processes, based on major and

trace element distributions (King et al., 2001). However,

King et al. (2001)noted that Rb concentrations showed

trends that were difficult to explain via this process. The

experimental data are used to test the fractional crystal-

lisation hypothesis and to constrain the nature, amount

and composition of fractionating minerals.

2. Granites of the Wangrah Suite and starting

material

The Wangrah Suite occurs over an area of 23 km2 in

the Lachlan Fold Belt (LFB), near Jerangle, in SE-

Australia. It consists of four major granite intrusions;

Danswell Creek Granite (DCG), Wangrah Granite (WG),

Eastwood Granite (EG) and Dunskeig Granite (DG),

ordered by increasing silica content. The geology,

petrology and geochemistry of the Wangrah Suite have

been described in more detail byKing et al. (2001). They

concluded that the granites were emplaced at shallow

level (100200 MPa) and fO2 below NNO. Within the

LFB, the Wangrah Suite is a typical A-type granitic

association. In the following section, a brief summary of

petrological features important for the understanding of

the crystallisation of the Wangrah Suite are given.

The Danswell Creek Granite is a white, equigranular

medium grained monzogranite containing amphibole,

biotite, ilmenite, zircon, apatite and minor titanite.

Anorthite contents of plagioclase vary from An30 toAn10. Among the Wangrah Suite granites the Danswell

Creek Granite has the highest modal abundance of mafic

minerals (biotite, ilmenite and amphibole; 8 to 11%.) and

the lowest abundance of K-feldspar (21 to 35%).

The Wangrah Granite is the most abundant rock of the

Wangrah Suite. It is a pinkgrey to white monzogranite

containing amphibole, biotite, ilmenite, zircon and apatite.

The rock is distinguished by a variably porphyric texture

with K-feldspar and quartz phenocrysts set in a ground-

mass of K-feldspar, quartz, plagioclase and biotite

amphibole. K-feldspar often shows ovoid crystal formand is often rimmed (partly and completely) by

plagioclase displaying local rapakivi-texture. Two gen-

erations of quartz can be identified within the rock, which

is typical for many rapakivi granites (e.g., Eklund and

Shebanov, 1999). Anorthite contents of plagioclase vary

from An30 toAn10. Modal abundance of minerals depends

on the distribution of phenocrysts. K-feldspar is enriched

up to 47% in samples displaying rapakivi-texture coupled

with the lowest plagioclase contents of 18%. In samples

with a few (or none) K-feldspar phenocrysts plagioclase is

the most abundant phase (up to 49%) and the amount of K-

feldspar in the matrix is 13%. Quartz and mafic mineralabundance is relatively constant with 2330% quartz and

512% mafic minerals.

The Eastwood Granite is a fine-grained red annite

monzogranite containing biotite, ilmenite, zircon and

apatite with sparse K-feldspar and quartz phenocrysts.

K-feldspar is the most abundant phase (up to 40%) while

plagioclase and quartz occur in approximately equal

proportions (2530%). Mafic minerals are subordinate

with 37%. Anorthite content of plagioclase vary from

An25to An5.

The most evolved rock of the Wangrah Suite is thebrick red equigranular Dunskeig Granite. This rock is

characterised by the highest amount of K-feldspar (40

50%) and the lowest distribution of mafic minerals

b3%. Biotite is conspicuous and occurs either as large

single euhedral (8 mm) or interstitial crystals. Anorthite

content of plagioclase is An13 to An5. The Dunskeig

Granite could represent a carapace of a deeper pluton

because it outcrops topographically higher than the

other granites (King et al., 2001).

Four representative analyses of granites from the

Wangrah Suite DCG, WG, EG and DG, labeled AB412,

AB422, AB421 and AB401, respectively are listed in

Fig. 2. Comparison of the obtained melt H2O contents of thesix hydrous

rhyolitic standard glasses by Karl Fischer Titration (KFT; y-axis) and the

difference in total of electron microprobe analyses (EMP;x-axis) of one

representative analytical session. The error bars represent the error in

KFT-analyses (horizontal bar: 0.15 wt. % H2O;Holtz et al., 1995) and

the standard deviation () of EMP analyses (vertical bar). The regression

curve is used to correct the experimental glass H2O contents. Final water

contents of experimental glasses as reported in Tables 2 and 3 were taken

from they-axis.



6/20

Table 1. The samples AB412, AB422, AB421 and

AB401 come from the DCG, WG, EG and DG,

respectively. The chemical variations within the Wan-

grah Suite, such as decreases in FeO, CaO, Ba and Sr

with increasing SiO2 (Fig. 1), are likely the result of

crystal fractionation from a parental composition closeto that of AB412 (King et al., 2001). Zircon saturation

temperatures (calculated after Watson and Harrison,

1983) are 897 C for the more mafic composition

AB412, 843 C for AB422, 806 C for AB421 and

764 C for the most felsic composition AB401.

Phase relations for compositions representative of

DCG and WG (AB412 and AB422) were previously

determined at 200 MPa, fO2 between NNO to NNO-

1.05 in a temperature range between 700 and 900 C

(Klimm et al., 2003). Based on plagioclase stability,

plagioclase composition and Zr saturation temperaturesKlimm et al. (2003) suggested that the initial water

content in the primary melt was between 23 w t . % H2O.

Klimm et al. (2003) also emphasised that orthopyroxene,

besides plagioclase, was part of the first fractionating

mineral assemblage in the Wangrah Suite, although this

mineral was not found in the natural rock.

In this study the compositions chosen as starting

material for the experiments were AB421 and AB401,

representative of EG and DG, respectively.

3. Experimental techniques

The natural rock powders (AB421 and AB401) were

melted twice in a Pt crucible at 1600 C and 1 atm, and

the samples were ground in an agate mortar between the

two melting steps. The glass compositions were

determined by electron microprobe analysis and are

similar to the natural rock composition (Table 1). The

anhydrous glasses were ground and then 90 wt. % glass

and 100.2 wt. % fluid were loaded in welded gold

capsules. The amount of added fluid was low to avoid

undesirable effects of incongruent dissolution of silicate

in the f luid. The water activity (aH2O) of theexperimental charges was varied by adding a fluid

composed of a mixture of H2O and CO2. CO2was added

as silver oxalate (Ag2C2O4). The mole fraction of water

in the added fluid phase was varied in a range of

XH2Oinitial= 0.4 to 1.0. It is emphasised that this

XH2Oinitial is different from the final XH2O because

H2O is preferentially dissolved in silicate melts.

Crystallisation experiments with AB421 and AB401,

covering the temperature interval 700 to 850 C, were

performed at 200 MPa in cold seal pressure vessels

(CSPV) with water as the pressure medium. Run

duration varied with temperature: 32 days for runs at

700 to 775 C, 8 days at 800 and 850 C. The oxygen

fugacity (fO2) was monitored by adding a solid NiNiO

buffer. The pressure vessels are made of Ni-alloy, which

allows to perform long run duration with the solid buffer

technique at fO2= NNO. Because water activity and

water fugacity (fH2O) varies in the experiments (as afunction of the mole fraction of water in the charge),

fO2 is not strictly constant at given P and T. In our

experiments fO2 ranges from NNO to NNO-1.05 log

units (Tables 2 and 3).

After the runs, a HO or a HOC fluid phase was

detected qualitatively in all charges by weight loss and

freezing in a cold trap cooled by liquid nitrogen (for

CO2) and weight loss after heating up to 105 C (for

H2O). However, because of the small amount of added

fluid, the final XH2O of the fluid after the experiments

could not be analysed quantitatively by this method.The phases were identified and their compositions

were determined by electron microprobe (Cameca SX

100 at the University of Hannover). Crystalline phases

were analysed with 15 kV acceleration voltage, 15 nA

sample current and 10 s total counting time. To

minimise the migration of alkalis during glass analysis

the analytical conditions were 15 kV, 4 nA and 5 s total

counting time and the beam was defocused up to 20m

when possible (lowest value was 5 m, Na was

measured first). The H2O contents of the glasses were

estimated with microprobe analyses following the

difference method calibrated with standard glasses. Sixhydrous rhyolitic glasses (with 0.2 to 7.15 wt. % H2O)

whose H2O contents were determined by Karl Fischer

titration (KFT) with an uncertainty of 0.15 wt. % H2O

(Holtz et al., 1995) were used as the standard glasses.

Fig. 2shows a calibration for the measured deficit in the

total by microprobe analysis compared to the values

determined by KFT of the standard glasses. The

standard and experimental glasses were measured

during the same analytical session. The precision of

the method can be tested for water saturated experi-

ments. The results obtained by the difference method areidentical within 0.2 wt. % H2O to water solubilities at

200 MPa determined in subaluminous rhyolitic glasses

(Klimm et al., 2003). In charges with lower melt fraction

this error might be higher but is expected to be within

0.5 wt. % H2O.

4. Experimental results

4.1. Phase relations

Details concerning the experimental conditions and run

products are listed inTables 2 and 3. The results of the



7/20

crystallisation experiments for AB421 and AB401 are

represented in phase diagrams (Fig. 3). The phase

diagrams are isobaric TcH2Omeltsections where cH2Omeltis the H2O content of the glasses (wt. %) determined by

the calibrated deficit in the microprobe total. The

experiments do not allow us to draw the solidus curve

as a function of TaH2O, because appropriate TaH2O

conditions have not been achieved within the presentedset of experiments. However, this curve has been reported

in Fig. 3 using data from the QzAbOrH2OCO2system (Holtz et al., 2001). The solidus temperature is

lowered by25 C when compared toHoltz et al. (2001)

to account for slight deviations from the synthetic Qz

AbOr system which may decrease the solidus temper-

ature (e.g., weakly peraluminous compositions). The

water-saturation curve has been drawn on the basis of the

experimental results in the subaluminous granitic system

(Holtz et al., 2001) with a maximum water solubility of

6.2 wt. % at 800 C, 200 MPa and a temperaturedependence of16104 wt. % H2O/C.

In the most felsic composition AB401 liquidus

conditions were attained at cH2Omelt4 wt. % and

temperatures775 C. For both compositions FeTi

oxides are the liquidus phases. In the less evolved

composition AB421 near liquidus conditions could only

be obtained at 850 C with cH2OmeltN3 wt. % and at

800 C with cH2Omelt5 wt. % H2O, respectively. With

decreasing temperatures and for cH2Omeltb3.5 wt. %

H2O, the crystallisation of FeTi-oxide is followed by

orthopyroxene, plagioclase and quartz. The orthopyr-

oxene stability field is restricted to high temperatures

above 750 C and cH2Omeltb5 wt. %. Biotite occurs at

775 C and below. When compared to AB421 biotite

crystallises at lower temperatures in AB401 (750 to

760 C). In AB401 the crystallisation of FeTi-oxide is

followed by quartz+plagioclase at cH2Omeltb3.5 wt. %.

At H2O contents of the melt between 3.5 to 5 wt. %,

quartz, plagioclase, K-feldspar and biotite crystallise

Fig. 3. Phase relations for composition AB421 and composition AB401 as a function of temperaturemelt H2O content at 200 MPa and fO2NNO.

Solid dots represent experimental charges at given run conditions. Mineral abbreviations as given byKretz (1983). Stability curves are labelled with

mineral names lying inside their stability field. Uncertain portions of stability curves are indicated by dashed lines.

Fig. 4. An-content of plagioclase (average of analyses) as a function of

melt H2O content and temperature. Grey fields with white numbers

indicate experimental temperatures for AB412 and AB422. Black

numbers indicate run temperatures AB421 and AB401. Solid lines are

linear fits for AB412 at 800 and 850 C. Dashed lines for 700 and

750 C have been constructed considering that the linear fits for given

T are parallel in the diagram. The analytical error represents the

maximal obtained standard deviation of plagioclase analyses of

AB421 and AB401 (vertical bar, see Table 4) and the maximal errorof water content determination (horizontal bar).



8/20


9/20


10/20

nearly simultaneously in the temperature range 750

775 C, approximately 15 C below the liquidus

indicating that AB401 is close in composition to a

water undersaturated minimum melt composition at

aH2O between 0.50.7 under the prevailing experimen-

tal conditions (200 MPa). At cH2OmeltN5 wt. % FeTi-oxide crystallisation is followed by biotite and quartz+

K-feldspar. Orthopyroxene was observed in only one

experiment at 775 C and 3.5 wt. % H2O close to the

solidus. In both compositions at water saturated

conditions plagioclase, quartz and K-feldspar crystallise

below 700 C, slightly above the estimated solidus.

4.2. Phase compositions

The An-content of plagioclases synthesised fromAB421 decrease with decreasing temperature and cH2O-

melt(or aH2O). Assuming an analytical error of 0.5 wt. %

H2O for melt H2O contents, these An-contents are similar

to those obtained for AB412 and AB422 (Klimm et al.,

Fig. 5. Variations of selected element oxide concentrations (y-axis) of experimental glasses of all runs (anhydrous, normalised to 100 wt. %) as a

function of melt H2O content (x-axis), temperature (symbol legend in box) and starting composition (vertical columns). Data for AB412 and AB422

afterKlimm et al. (2003). Error bars: standard deviation () from EPMA. Symbols with no error bar in a k indicate that was smaller than the

symbol. For clarity only the maximum is presented in lu (for details seeTables 7 and 8). The standard error on the melt H2O content is estimated to

be 0.5 wt. % H2O for all analyses (au). Grey horizontal lines with grey labels indicate the starting bulk compositions of AB422, AB421 and AB401

(normalised to 100 wt. %). TiO2, MgO and MnO are not presented because of their low abundance in the bulk compositions. Na2O is not presentedbecause it does not vary much over the Wangrah Suite granites.



11/20

2003) at identical temperatures and melt H2O contents

(Fig. 4). In runs with AB401 the An-content of pla-

gioclase, ranging from An09to An12, shows no variation

with temperature or water content of the melt and contains

relatively high Or-contents (Or9.725.6) reflecting the low

CaO content of the bulk composition.The XMg variation range of orthopyroxene is lower

for AB421 (0.290.42,Table 5) than for AB412 (0.30

0.59) and AB422 (0.300.56) but XMg is similar in the

three compositions for identical T and aH2O. Thus, the

orthopyroxene composition in the studied systems

depends on the prevailing run conditions and not on

the starting bulk composition.

Oxide minerals, identified as titanomagnetite, are

present in all runs except in liquidus experiments with

composition AB401. The very small crystal size

(

2 m), especially common in water undersaturatedcharges, makes it difficult to obtain reliable analyses.

Biotite crystals are very small in water undersaturat-

ed charges with high crystal contents. Although the

analyses of biotite are of low quality, the XMg decreases

from 0.5 to 0.3 with decreasing temperature and

decreasing fO2 in agreement with Dall'Agnol et al.

(1999). TiO2of experimental biotite is 3 wt. %.

Quenched glass compositions of AB421 and AB401

normalised to 100 wt. % (anhydrous) are reported in

Tables 6 and 7andFig. 5lu, respectively.

5. Discussion

Together with the experimental results obtained for

the more mafic compositions AB412 and AB422

(Klimm et al., 2003) this study provides information

on phase relations at emplacement conditions (200 MPa,

King et al., 2001) for each granite in the Wangrah Suite.

In the following section, the experimental results and the

chemical compositions of natural bulk rocks and

minerals are used to model fractional crystallisation

processes and to understand the mechanisms leading to

chemical variations of the magmas.

5.1. Comparison of experimentally determined phase

relations

As mentioned above, besides temperature and aH2O,

the phase stabilities are a function of bulk rock

composition (increasing felsic character from AB412

to AB401). A comparison of mineral stability curves for

the main rock forming minerals is presented inFig. 6.

The plagioclase stability field decreases and is shifted

to lower temperature from AB412 to AB401 with

decreasing bulk CaO contents in the starting material

(Table 1). Quartz stability only depends slightly on

starting composition. Within the resolution of the

experiments quartz crystallises at similar temperatures

at a given aH2O in AB422, AB421 and AB401. In the

most mafic composition AB412 quartz stability is

slightly shifted to higher temperatures at cH2Omelt of45 wt. %. K-feldspar stability decreases to lower

temperature (from N800 C to 775 C) for cH2Omeltof

34 wt. % but is shifted to higher temperatures

(700 C to 750 C) for cH2Omeltof5 wt. % with

increasing K2O content of the starting material from

AB412 to AB401 (Table 1), excluding composition

AB422. In experiments with AB422 K-feldspar crystal-

lises over the widest range of temperature and aH2O

compared to the other compositions and biotite has a

large stability field crystallising above 850 C at

cH2Omelt

5 wt. % and above 800 C at lower cH2Omelt.In all other compositions the high temperature biotite

stability limit is nearly independent on the melt H2O

content and crystallisation temperatures decrease with

increasing felsic character of the starting composition

(800850 C for AB412, 775800 C for AB421 and

b775 C for AB401). The orthopyroxene stability field

decreases and is shifted towards lower temperature and

aH2O with decreasing bulk FeO and MgO contents.

5.2. Differentiation by fractional crystallisation

There are several possibilities to model the differen-tiation of the Wangrah Suite assuming that AB422

(WG), AB421 (EG) and AB401 (DG) are derived from

the parental magma AB412 (DCG). First, the three felsic

compositions can derive directly from AB412 assuming

that the residual melts (with compositions AB422,

AB421 and AB401) are removed after a certain crystal

fraction was formed. Second, fractional crystallisation

could occur if crystals are chemically isolated from the

coexisting melt after their formation (segregation

process). In this case, composition AB421 should derive

from AB422 as the parent magma and AB401 fromAB421. Third, the differentiation results from a

combination of both processes.

5.2.1. AB422 (WG)

Using experimentally determined phase relations

and zircon saturation temperatures, Klimm et al.

(2003)modelled the formation of AB422 considering

a differentiation by fractional crystallisation from a

parental magma AB412 (DG). Klimm et al. (2003)

showed that SiO2, Al2O3, FeO, MgO, CaO, Na2O

concentrations of AB422 can be reproduced after

fractionation of 4.7 wt. % orthopyroxene and 8.5 wt.%



12/20

plagioclase. However, the mass balance calculation

for K2O was not successful and did not reproduce

adequately the composition AB422 (K2O in AB422

an d al l o th er EG i s t oo h ig h to b e mo del le d

successfully).

Comparison of AB422 with the composition of

residual glasses from crystallisation experiments with

AB412 can be used to determine the conditions under

which AB422 can be derived from AB412 (T, aH2O, and

phases involved in fractionation processes). At the

appropriate conditions the residual melt composition

from AB412 has to be similar to the bulk composition

AB422. Fig. 1b demonstrates that the CaO and SiO2

contents of the WG are never obtained in residual melts

Fig. 6. Comparison of the experimental determined phase stability curves of plagioclase, quartz, K-feldspar, biotite and orthopyroxene as a function of

starting bulk composition in the temperature versus melt H2O contents phase diagram. AB412 (solid black) and AB422 (dashed grey) afterKlimm

et al. (2003); AB421 (dashed black) and AB401 (solid grey) this study. Grey arrows indicate trends for the occurrence of observed phases from the

most mafic composition AB412 to the most evolved composition AB401 excluding AB422. Note the broadest stability fields of the K2O bearing

phases K-feldspar and biotite of AB422 compared to all other compositions.



13/20

in experiments with AB412.Fig. 5ae also shows that it

is improbable that AB422 can be generated by fractional

crystallisation from a parental magma close to AB412.

Residual melt compositions of AB412 having K2O

contents similar to AB422 (5 wt. %) are only observed

at cH2Omeltb3.5 wt. % (Fig. 5e). However, at these H2Ocontents the SiO2 contents are systematically higher

(76 wt. %) than AB422 (Fig. 5a). Thus, the high K2O

content of AB422 cannot be explained by a simple

fractional crystallisation process from AB412 at the

applied isobaric conditions of 200 MPa and a fO2 of

NNO to NNO-1. Alternative models are proposed below.

5.2.2. AB421 (EG)

Using the same method as described above (com-

parison of experimental glass compositions with bulk

compositions, Fig. 5) the derivation of AB421 fromAB412 or AB422 can be tested. The examination of

Fig. 5ae shows that residual glasses from AB412 with

SiO2, Al2O3, FeO, CaO and K 2O concentrations

corresponding to the bulk AB421 have only been

obtained in experiments at temperatures between 900 to

850 C with cH2Omeltof 2 to 3 wt. %. Thus, AB421 can

derive from AB412 by crystallisation of magnetite,

orthopyroxene and plagioclase according to the phase

relations in this TaH2O range (Klimm et al., 2003).

Mass balance calculations show that the residual melt

composition is close to composition AB421 after

extracting 2.1 wt. % magnetite, 3.5 wt. % orthopyroxeneand 13.1 wt. % plagioclase from AB412 (Model 1,

Table 9). Zircon saturation temperature for AB421 is

806 C (Table 1) suggesting that biotite may contribute

to the fractionating phases (Fig. 6). However, calcula-

tion with biotite (Model 2) show that this mineral would

play a minor role (b0.5 wt. %) in fractionation processes

leading to AB421. Calculations involving quartz and/or

K-feldspar did not yield to satisfying results indicating

that these minerals were not part of the fractionating

assemblage.

The derivation of AB421 from AB422 can be testedby examining Fig. 5fk. The residual glass composi-

tions of runs with AB422 never fit the composition

AB421 for all considered elements. The relatively low

K2O content of AB421 (4.66 wt. %), when compared to

AB422 (4.95 wt. %), is only observed in runs with

cH2OmeltN4 wt. % (Fig. 5k). In contrast the CaO content

of 0.91 wt. % in AB421 is mainly determined at

cH2Omeltb4 wt. % (Fig. 5s) except at 750 C, but at this

condition the FeO contents of residual glasses are lower

(1 wt. %) than AB421 (1.64 wt. %, Fig. 5i). Thus, at

the investigated conditions of 200 MPa and a fO2

of

NNO to NNO-1 it was not possible to reproduceTable8

Com

positionsofexperimentalglasses(wt.%,

normalisedto100)ofAB401at200MPaandlogfO2

NNO

Run

62

63

78

79

66

67

81

82

83

86

87

71

70

69

68

T[C]

700

n=8

700

n=11

725

n=772

5

n=7

750

n=6

750

n=11

760

n=10

760

n=8

760

n=10

775

n=9

775

n=11

800

n=7

800

n=9

800

n=10

800

n=13

SiO2

78.4

4

0.1

9

77.8

3

0.2

7

77.9

8

0.4

57

7.6

6

0.2

5

78.0

9

0.2

8

77.7

8

0.4

1

77.7

1

0.3

5

77.7

9

0.2

2

77.3

8

0.3

0

77.8

6

0.32

77.3

8

0.3

1

77.5

8

0.3

0

77.8

8

0.3

7

77

.81

0.3

9

77.6

8

0.4

4

TiO2

0.0

6

0.0

5

0.0

7

0.0

4

0.0

7

0.0

3

0.0

9

0.0

2

0.0

7

0.0

2

0.0

9

0.0

4

0.0

6

0.0

2

0.0

7

0.0

3

0.0

8

0.0

4

0.0

7

0.03

0.1

0

0.0

3

0.1

0

0.0

3

0.0

9

0.0

3

0

.10

0.0

2

0.0

9

0.0

3

Al2O

3

12.2

7

0.1

1

12.6

3

0.1

9

12.5

7

0.1

41

2.3

8

0.1

3

12.5

2

0.2

1

12.4

5

0.1

8

12.5

2

0.1

5

12.5

0

0.1

9

12.6

2

0.2

0

12.5

0

0.17

12.5

4

0.1

2

12.3

4

0.1

9

12.2

0

0.2

0

12

.35

0.2

4

12.4

5

0.2

0

FeOa

0.6

4

0.1

4

0.9

0

0.0

9

0.7

4

0.0

9

0.8

4

0.0

9

0.8

6

0.2

3

0.7

0

0.0

6

0.7

5

0.1

2

0.7

5

0.1

3

0.8

6

0.1

3

0.6

9

0.17

0.8

7

0.1

1

1.0

0

0.1

3

0.9

4

0.1

7

1

.01

0.1

7

0.7

1

0.1

4

MnO

0.0

0

0.0

1

0.0

2

0.0

1

0.0

3

0.0

1

0.0

4

0.0

4

0.0

6

0.0

1

0.0

0

0.0

0

0.0

7

0.0

4

0.0

6

0.0

4

0.0

6

0.0

4

0.0

6

0.03

0.0

5

0.0

4

0.0

7

0.0

1

0.0

4

0.0

1

0

.00

0.0

0

0.0

5

0.0

5

MgO

0.0

1

0.0

1

0.0

1

0.0

1

0.0

5

0.0

3

0.0

4

0.0

4

0.0

4

0.0

2

0.0

0

0.0

0

0.0

6

0.0

3

0.0

4

0.0

2

0.0

3

0.0

2

0.0

6

0.03

0.0

6

0.0

2

0.0

5

0.0

2

0.0

4

0.0

2

0

.06

0.0

4

0.0

6

0.0

3

CaO

0.4

0

0.0

4

0.3

2

0.0

5

0.4

2

0.0

3

0.4

3

0.0

9

0.3

3

0.0

7

0.3

5

0.0

5

0.4

2

0.0

6

0.4

2

0.0

6

0.3

4

0.0

7

0.4

1

0.04

0.3

0

0.0

7

0.4

3

0.0

6

0.4

6

0.0

3

0

.44

0.0

6

0.3

3

0.0

6

Na2

O

3.3

7

0.2

8

3.5

7

0.1

5

3.1

1

0.1

5

3.4

4

0.1

7

3.3

4

0.2

0

3.6

2

0.2

5

3.3

5

0.1

3

3.3

4

0.2

2

3.4

5

0.1

7

3.3

7

0.12

3.3

3

0.1

9

3.4

7

0.2

3

3.4

8

0.1

6

3

.46

0.1

8

3.4

9

0.2

5

K2O

4.8

1

0.1

2

4.6

6

0.1

2

5.0

3

0.1

5

5.0

7

0.1

0

4.7

0

0.2

2

5.0

0

0.1

2

5.0

5

0.1

4

5.0

2

0.1

6

5.1

9

0.1

7

4.9

7

0.17

5.3

8

0.1

2

4.9

4

0.1

2

4.8

7

0.1

7

4

.75

0.1

7

5.1

4

0.1

4

Total

100.0

0

100.0

0

100.0

0

10

0.0

0

100.0

0

100.0

0

100.0

0

100.0

0

100.0

0

100.0

0

100.0

0

100.0

0

100.0

0

100

.00

100.0

0

cH2

O

b

6.3

5.1

6.0

5.5

5.1

4.0

5.0

4.6

4.2

4.2

3.4

4.9

4.1

3

.2

2.7

Qz

40.7

2.0

39.6

1.3

41.1

1.4

3

8.3

1.8

41.2

2.1

37.6

2.4

39.1

1.3

39.3

1.6

37.5

1.4

39.4

1.0

37.5

1.7

38.5

1.7

39.0

1.5

39

.6

2.1

37.7

1.9

Ab

29.7

2.4

31.6

1.2

27.7

1.4

3

0.4

1.5

29.7

1.8

31.8

2.2

29.7

1.2

29.6

1.8

30.5

1.4

29.8

1.1

29.4

1.6

30.9

1.9

30.9

1.4

30

.8

1.5

30.7

2.2

Or

29.6

0.8

28.8

0.7

31.2

0.9

3

1.3

0.6

29.1

1.3

30.6

0.7

31.2

0.8

31.1

1.0

32.0

1.0

30.7

0.9

33.1

0.6

30.6

0.8

30.1

1.1

29

.5

1.0

31.6

1.0

a

TotalFeasFeO.

b

Watercontentsdeterminedfollowingthedifferencemethodcalibratedwithsixhydrousrhyoliticstandardglasses.



14/20

residual glasses from AB422 with compositions iden-

tical/close to AB421.

In conclusion the experimental results indicate

that EG may be derived from DCG by fractionalcrystallisation. Assuming that melts representative of

DCG contains 2.5 wt. % H2O (23 w t . % H2O at

liquidus conditions for AB412;Klimm et al., 2003) the

liquidus temperature for this composition should be

approximately 930 C. After fractionation of 18.7 wt. %

anhydrous minerals the residual melt with the compo-

sition of EG would contain3 wt. % H2O. The liquidus

temperature determined from the phase relations is

850 C. These experimentally determined liquidus

temperatures for DCG and EG are higher by 40 C

than the calculated Zr saturation temperatures (Table 1).

This may be explained by differences between exper-

imental and natural conditions. For example slightly

more reducing conditions may decrease the crystal-

lisation temperatures of tectosilicates (Dall'Agnol et al.,

1999). Furthermore the Zircon saturation geotherm-ometer assumes that the rocks represent pure melts that

crystallised without any further fractionation, mixing or

mingling processes.

5.2.3. AB401 (DG)

Following the same approach as above and using

Fig. 5ae, it can be shown that AB401 cannot be

derived directly from AB412. In particular the low FeO

content of AB401 (0.93 wt. %) is only reproduced at

temperatures 800 C. At these temperatures the K2O

content of AB401 (5.18 wt. %) is only observed in

residual glasses at cH2Omelt3.5 wt. % (Fig. 5e).

Table 9

Results of mass balance calculations for major element fractional crystallisation

aDeduced from experimental phase relations (Fig. 2and Klimm et al., 2003).

bDue to small crystal size of Mt and Bt in experiments compositions are estimated considering good analyses of neighbouring charges at higher

aH2O. Mt in Run 35 was estimated with the composition of Run 1 and a decrease of TiFe contents to FeO=90 wt. % and TiO2=10 wt.% to count

for the effect ofaH2O. Biotite in Run 33 was estimated with the composition of Run 1 and a decrease of the XMg to 0.34. Biotite in Run 84 was

estimated with Run 61 and a decrease of the XMg to 0.35. Note that this procedure was only carried out to minimise the residuals and the effect on

calculated modal proportions is less than 5%.cCalculated change of water content after fractionation assuming an initial water content of 2.5 wt. % H 2O for AB412 (according to 23 wt. % H2O

afterKlimm et al., 2003).



15/20

However, the SiO2contents of these glasses are too low

when compared to AB401 (Fig. 5a). The residual glass

compositions of charges with AB421 are close to

composition AB401 for all elements except for K2O

which is too low in most experiments (Fig. 5lp). In fact

only runs saturated with respect to K-feldspar or close tosaturation yield to glasses with high K2O. This suggests

that AB401 corresponds to a melt in equilibrium with K-

feldspar in addition to quartz and plagioclase.

Considering that AB401 cannot be directly derived

from AB412, its parental composition must be already

relatively differentiated and can be assumed to be close to

AB421. Mass balance calculation show (Table 9) that

AB401 is obtained after extraction of 3.3 wt. % biotite,

13.5 wt. % plagioclase, 12.8 wt. % quartz and 11.3 wt. %

K-feldspar from AB421 (Model 3). Calculation including

orthopyroxene (Model 4) shows that this mineral mayplay a minor role (b0.5 wt. %), if any, in fractionation

processes leading to AB401, in agreement with the

decreasing orthopyroxene stability field from AB421 to

AB401 (Fig. 6). After fractionation the water content

increased from 3 wt. % H2O in EG to 4.5 to 5 wt. % H2O

in DG. The experimentally determined liquidus temper-

ature of 760 to 775 C for AB401 in the range of 4

5 wt. % H2O(Fig. 3) is in excellent agreement to 764 C

calculated from the Zr concentration of the sample.

5.3. Constraints from trace elements

The validity of the modelled mineral assemblages

fractionating from AB412 to form AB421 and AB401 as

deduced by the mass balance calculations (Table 9)canbe

tested using the trace element distribution of the Wangrah

Suite. Because we have determined phase relations for a

limited number of bulk compositions the fractionation can

be described as a two step batch fractionation. However,

under natural conditions we would expect a Rayleigh

fractionation process. Once a crystal starts to form all its

components, major and trace elements are immediately

chemically isolated from the bulk system. For that reason

we test the trace element distribution within the Wangrah

Suite with Rayleigh fractionation models. Rb, Sr and Ba

are useful trace elements to describe fractionation

processes involving feldspars and biotite (for partition

coefficients see Table 10). It has been shown that

partitioning of Ba between alkali feldspars and melts isa function of the orthoclase component (in mol%) in the

feldspars (Icenhower and London, 1996; Blundy and

Wood, 2003). Plagioclase crystallising in experiments

with AB412 contain a significant amount of orthoclase

component of 5.8 to 12.2 mol% Or at temperatures

between 900850 C and cH2Omelt of 1.83.4 wt. %

(Klimm et al., 2003). Therefore we calculated partition

coefficients for Ba in feldspars using the linear relation-

ship given by Blundy and Wood (2003). Maximum

partition coefficients for Ba in the experimental orthoclase

bearing plagioclases are DBa= 3

4(Table 9).The compo-sition of all available Wangrah Suite granites (King et al.,

1997, 2001 and unpublished data) and the calculated

fractionation trends (with mineral proportions given in

Table 9) are shown in Fig. 7. The Rb, Sr and Ba

concentrations of the EG compositions (AB421) confirm

that these granites may be derived by fractional crystal-

lisation following Model 1 and 2 (Fig. 7;Table 9). The

relatively low Rb contents of the DG can only be

explained if fractionation of K-feldspar is also involved as

confirmed by the trends corresponding to Model 3 and 4.

It has to be noted that although we excluded the WG to

be a fractionation product from a parental melt close to aDCG composition (because of the high K2O content of the

bulk rock) the Sr, Rb and Ba contents of the WG could be

explained by fractional crystallisation of plagioclase and

orthopyroxene (Fig. 7). Thus, modelling fractionation

trends from the trace element chemistry alone would blur

discrepancies in major element chemistry.

5.4. The Wangrah Granite: a result of magma mixing

not fractionation?

As mentioned above the most abundant WG does notfollow the fractional crystallisation trend obtained. The

reason for that could be a) the natural pressure and fO2conditions differ from the experimentally applied pressure

of 200 MPa and fO2 of NNO to NNO-1 or b) the WG

results from a different process such as magma mixing. In

the following we discuss chemical and petrographic

features of the WG to point out the distinctiveness of this

rock compared to the other granites of the Wangrah Suite.

Retrieving the correct fO2 and pressure of crystal-

lisation of plutonic rocks is a difficult task and

differences in element concentrations (e.g. Fe, Si) that

were used to discriminate parentdaughter relationship

Table 10

Mineral/melt partition coefficients used for fractional crystallisation

modelling

Orthopyroxene Plagioclase Biotite K-feldspar

Rb 0.01 a 0.011 b 4.1 c 2.4 c

Ba 0.1 a 34 d 5.6 c 16 d

Sr 0.01 a 4.04b 0.29 c 4.5 c

a Bacon and Druitt (1988).b Ewart and Griffin (1994).c

Nash and Crecraft (1985).d Blundy and Wood (2003).



16/20

of natural rocks could be explained by inappropriate

choice of one or two experimental parameters (P and/or

fO2). For instance, evaluating the potential parent

daughter relationships between AB412 and AB422, it is

found on the panels a) to e) ofFig. 5, that the closest

intersection with the isothermal liquid line of descent of

AB412 at 850 C with the horizontal lines correspond-

ing to AB422 can be achieved at a melt water content of

3 wt. % H2O. The higher FeO content of the natural

rock compared to the experimental liquid (Fig. 5c) can

be explained by differences of the experimental and

naturalfO2. Slightly more reducing conditions will lead

to higher FeO contents in the experiments which will in

turn decrease the SiO2content of the residual liquid, and

thus bring those two elements in closer match with the

observations.

Another problem is the upward shift of experimental

liquids produced from AB412 in terms of their CaO vs.

Fig. 7. (a) Rb vs. Sr and (b) Rb vs. Ba diagram (logarithmic scale) showing the variation of Wangrah Suite granite samples (source of data:King et al.,

1997, 2001 and unpublished data). Crosses: experimental studied compositions. Dashed vectors: Rayleigh fractionation trends for Pl, Kfs, Opx and Bt

(partition coefficients for Rb, Ba, Sr are given in Table 10). Solid lines: Rayleigh fractionation trends for experimental determined phases (Model 1 to

Model 4). For simplicity and the small differences between Model 1 and 2 and Model 3 and 4, respectively, only Model 2 and 3 are shown in (b). Massfractions are obtained by mass balance calculations (details in Table 9). Vertical marks indicate 10% fractional crystallisation.



17/20

SiO2contents (Fig. 1b). The experimental liquids are too

CaO rich relative to WG. This is due to the higher melt

water contents of N3 w t . % H2O for most of the

experimental liquids. Melts with a melt H2O content

b3 wt. % that are in agreement with the initial melt water

content of 23 w t . % H2O for AB412 have lower CaO andhigher SiO2contents and fall in the field of EG (Klimm

e t a l. , 2 00 3). Crystallisation at lower pressure

(100 MPa) enhances plagioclase crystallisation and

thus depletes the residual liquid in CaO and could

therefore account for the low CaO vs. SiO2 contents of

WG. However, crystallisation at lower pressures is not

likely in the case of WG and DCG because of the presence

of magmatic amphibole in both natural rocks. In a

previous experimental study determining the phase

relations of AB412 and AB422 Klimm et al. (2003)

observed amphibole crystallisation only for AB412 atmelt water contents 4.5 wt. % H2O. The low water

solubility of4 wt. % H2O at pressures 100 MPa is

insufficient to provide amphibole crystallisation. In

additionKlimm et al. (2003) suggested that the higher

CaO of AB412 (DCG) compared to AB422 (WG) may

explain the stability of amphibole in experiments with

AB412 and the lack of amphibole in experiments with

AB422. Thus, magmatic amphibole in the WG must have

crystallised from a melt composition with relatively high

CaO contents (probably more mafic than the bulk

AB422). This means that part of the minerals in the WG

were not in equilibrium with the surrounding melt at somestage of the magmatic history of the WG.

The K2O contents of the WG are relatively high

compared to the other granites (Fig. 1c) as a result of K-

feldspar accumulation in the natural rock. King et al.

(2001) described variable porphyric textures within the

WG and a variable distribution of K-feldspar megacrysts

partly surrounded by a plagioclase mantle (rapakivi-

texture) and two generations of quartz probably related to

undercooling processes. Undercooling textures may be

the result (1) of pressure-quenching (Nekvasil, 1991;

Eklund and Shebanov, 1999) or (2) of mixing processesinvolving magmas with different composition and

temperature (Hibbard, 1981, 1991; Wark and Stimac,

1992). In the first hypothesis, crystallisation should occur

at different depths and the effects of polybaric fraction-

ation processes on differentiation can be estimated

qualitatively by comparison of our experimental results

at 200 MPa with previous experimental studies on A-type

granites at 300 and 400 MPa (Patio Douce, 1997;

Dall'Agnol et al., 1999).

Patio Douce (1997) showed that low-pressure

melting of a tonalitic or granodioritic composition at

400 MPa and temperatures N900 C leads to A-type

granitic residual melts with cH2Omeltb4 wt. %. These

melts were in equilibrium with calcic plagioclase +

orthopyroxene. Plagioclase + orthopyroxene are also the

near liquidus phases in the DCG composition (Klimm

et al., 2003). Thus, fractional crystallisation at pressures

higher than 400 MPa involves the same mineralassemblage of mainly plagioclase and orthopyroxene

and the liquid line of descent would be qualitatively

similar to those observed at 200 MPa. The high K2O

content observed in WG are not predicted from polybaric

crystallisation and differentiation of a composition similar

to DCG.

Dall'Agnol et al. (1999) performed crystallisation

experiments with an A-type granitic composition similar

to AB412 at 300 MPa and oxygen fugacities of log

fO2NNO + 2.5 and NNO-1.5. In addition to orthopyr-

oxene, clinopyroxene was also obtained at near liquidusconditions. This is most probably due to the slightly

higher CaO content of 2.2 wt. % (Dall'Agnol and de

Oliveira, 2007) and the lower FeO of 3.7 wt. % of their

starting composition when compared to AB412

(CaO= 1.96 wt. %, FeO= 4.14 wt. %; Table 1).

However, if clinopyroxene would be involved in

fractional crystallisation processes of DCG in addition

to plagioclase both minerals would fractionate high

amounts of CaO. The total amount of clinopyroxene+

plagioclase had to be lower to mass balance the CaO

content of the residual melt (AB422). This is not in

agreement with (1) the relative high K2O content forWG (Table 1), because a fractionation of 20 wt. % K2O

free crystals is needed to explain the increase from

4 wt. % K2O (AB412) to 5 wt. % K2O (AB422) and

(2) the trace element concentrations of WG (Fig. 7)

suggest that plagioclase plays a major role in the early

stage of differentiation. Therefore clinopyroxene can be

excluded to play a role in fractional crystallisation

processes in case of the WG.

Hence, any crystallisation from a parental magma of

DCG Granite composition at depths prior to the final

emplacement level (200400 MPa) involves the samemineral assemblage of plagioclase and orthopyroxene

and therefore polybaric processes such as pressure-

quenching can be ruled out to account for the chemical

distinctiveness of the WG.

We favour a magma-mixing scenario involving a

significant amount of K-feldspar to account for the

relatively high K2O content (Fig. 1c) and to explain the

variable distribution of K-feldspar megacrysts of the

WG. K-feldspar is, besides plagioclase and quartz, one

of the major phases (33 wt. % of the fractionating

mineral assemblage, Table 9) involved in fractional

crystallisation leading from AB421 to AB401.



18/20

Assuming that this fractionation occurs at shallow level

conditions this K-feldspar-rich mineral assemblage

could be accumulated and mixed with a residual melt

or a new granitic melt pulse of a slightly different

composition. Considering that DCG compositionally

overlaps with WG at the margin of the intrusion (Kinget al., 2001; Fig. 1), the mixing process may involve

melts which are compositionally close to DCG. A mass

balance calculation shows that the WG (AB422) can be

compositionally obtained by mixing of K-feldspar

(25.9 wt. %), Quartz (14 wt. %), Plagioclase (3.8 wt.

%), Biotite (1.3 wt. %) and AB412 (55 wt. %, Table 11).

Such a mixing scenario could also account for the

occurrence of rapakivi-texture within the WG (Hibbard,

1981, 1991) and the occurrence of amphibole in the

natural rock (amphibole crystallises in experiments with

AB412; Klimm et al., 2003). The exact mechanismleading to the segregation of the crystals from the

residual melt and magma mixing with a melt composi-

tion close to the primary melt composition is difficult to

constrain, but could be related to convective mixing

processes in a magma chamber (Couch et al., 2001) or

replenishment by a parental magma.

6. Conclusions

The investigation of phase relations of the Wangrah

Suite granites suggests that the diversity of granite

compositions result from interaction of two mainprocesses, fractional crystallisation and/or magma

mixing.

AB412, a representative composition from the less

evolved Danswell Creek Granite, is suitable as an

equivalent of the parental magmas, as proposed by King

et al. (2001), with an initial water content of 2.5

0.5 wt. % H2O (Klimm et al., 2003). Fractionation of

small amounts of oxide, orthopyroxene and plagioclase

(less than 20 wt. %) from AB412 leads to AB421, a

composition representative for the Eastwood Granite.

Further fractionation of biotite, plagioclase, quartz

and K-feldspar from AB421 (approximately 40 wt. %)

can lead to compositions such as the most evolvedDunskeig Granite (AB401). The modelled differentia-

tion path from AB412 to AB421 to AB401 is consistent

with the major and trace element (e.g. Rb, Sr, Ba)

distribution of the natural rocks. The experimentally

derived temperatures are in good agreement with the Zr

thermometry.

Assuming such fractional crystallisation processes

from a water undersaturated parental magma (AB412

with 2.50.5 wt. % H2O), it has been demonstrated that

the residual melts evolve towards a water undersaturated

minimum temperature melt composition (minimum inthe system QzAbOrAnH2O). The fractionating

mineral assemblage is dominated by the tectosilicates in

the order of appearance (plagioclasequartzK-

feldspar). The Dunskeig Granite is close in composition

to a water undersaturated minimum at aH2O between

0.50.7 under the prevailing experimental conditions

(200 MPa;Holtz et al. 1992).

The Wangrah Granite is enriched in K2O compared

to the other granites. This high K2O content is not

predicted from the experimentally determined differ-

entiation trends and is related to the accumulation of

K-feldspar evidenced by variable porphyric texturesand a variable distribution of K-feldspar megacrysts in

the natural rock. A mass balance calculation shows

that the most dominant granite facies of the Wangrah

Suite could result from mixing processes between

accumulated mineral assemblage dominated by K-

feldspar and Quartz due to fractional crystallisation

from a parental melt composition similar to Danswell

Table 11

Results of mass balance calculation for major element magma-mixing

aMix of 55.04 wt. % AB412 (Table 1) with the calculated proportion of crystals.



19/20

Creek Granite. Such a combination of fractional crys-

tallisation and mixing processes could also account for

the occurrence of rapakivi-texture within the Wangrah

Granite.

Acknowledgements

We thank B.W. Chappell for providing the natural

samples from the Wangrah Suite and W. Johannes for the

initiation of the project. The technical assistance of O.

Dietrich and J. Koepke is gratefully appreciated. Helpful

and constructive comments were provided by A.E. Patio

Douce and B. Landenberger on an earlier version of the

manuscript. The reviews of B. Scaillet and O. Eklund are

gratefully acknowledged. This study was supported by the

German Science Foundation (DFG, project no.: Jo64/34).

References

Asrat, A., Barbey, P., 2003. Petrology, geochronology and SrNd isotopic

geochemistry of the Konso pluton, south-western Ethiopia: implica-

tions for transitionfrom convergence to extension in the Mozambique

Belt. International Journal of Earth Sciences 92, 873890.

Bacon, C.R., Druitt, T.H., 1988. Compositional evolution of the zoned

calcalkaline magma chamber of Mount Mazana, Crater Lake,

Oregon. Contribution to Mineralogy and Petrology 98, 224 256.

Blundy, J., Wood, B., 2003. Mineral-melt partitioning of uranium,

thorium and their daughters. Reviews in Mineralogy and Geochem-

istry 52, 59123.

Bogaerts, M., Scaillet, B., Liegeois, J.P., Vander Auwera, J., 2003.

Petrology and geochemistry of the Lyngdal granodiorite (Southern

Norway) and the role of fractional crystallisation in the genesis of

Proterozoic ferro-potassic A-type granites. Precambrian Research

124, 149184.

Chappell, B.W., White, A.J.R., Wyborn, D., 1987. The importance of

residual source material (restite) in granite petrogenesis. Journal of

Petrology 28, 11111138.

Chou, I.M., 1987. Oxygen buffer and hydrogen sensor techniques at

elevated pressures and temperatures. In: Barnes, H.L., Ulmer, G.C.

(Eds.), Hydrothermal Experimental Techniques. John Wiley, New

York, pp. 6999.

Clemens, J.D., Holloway, J.R., White, A.J.R., 1986. Origin of an A-type

granite:experimental constraints. American Mineralogist 71, 317324.

Collins, W.J., Beams, S.D., White, A.J.R., Chappell, B.W., 1982. Nature

and origin of A-type graniteswith particular reference to southeasternAustralia. Contribution to Mineralogy and Petrology 80, 189200.

Couch, S., Sparks, R.S.J., Carroll, M.R., 2001. Mineral disequilibrium

in lavas explained by convective self-mixing in open magma

chambers. Nature 411, 10371039.

Creaser, R.A., Price, R.C., Wormland,R.J., 1991. A-type granitesrevisited:

assessment of a residual-source model. Geology 19, 163166.

Cullers, R.L., Stone, J., Anderson, J.L., Sassarini, N., Bickford, M.E.,

1993. Petrogenesis of Mesoproterozoic Oak Creek and West

McCoy Gulch plutons: an example of cumulate unmixing of a mid-

crustal, two-mica granite of anorogenic affinity. Precambrian

Research 62, 139169.

Dall'Agnol, R., de Oliveira, D.S., 2007. Oxidized, magnetite-series,

rapakivi-type granites of Carajs, Brazil: implications for classi-

fication and petrogenesis of A-type granites. Lithos 93, 215233.

Dall'Agnol, R., Scaillet, B., Pichavant, M., 1999. An experimental

study of a lower Proterozoic A-type granite from the eastern

Amazonian craton, Brazil. Journal of Petrology 40, 1673 1698.

Eby, G.N., 1990. The A-type granitoids: a review of their occurrence

and chemical characteristics and speculations on their petrogen-

esis. Lithos 26, 115134.

Eklund, O., Shebanov, A.D., 1999. The origin of rapakivi texture bysubisothermal decompression. Precambrian Research 95, 129146.

Emslie, R.F., Stirling, J.A.R., 1993. Rapakivi and related granotoids of

the Nain Plutonic Suite: geochemistry, mineral assemblages and

fluid equilibria. Canadian Mineralogist 31, 821847.

Ewart, A., Griffin, W.L., 1994. Application of proton-microprobe data

to trace-element partitioning in volcanic rocks. Chemical Geology

117, 251284.

Frost, C.D., Frost, B.R., 1997. Reduced rapakivi-type granites: the

tholeiite connection. Geology 25, 647650.

Haapala, I., Rm, O.T., 1992. Tectonic setting and origin of the

Proterozoic rapakivi granites of southeastern Fennoscandia.

Transactions of the Royal Society of Edinburgh 83, 165171.

Hibbard, M.J., 1981. The magma mixing origin of mantled feldspars.

Contribution to Mineralogy and Petrology 76, 158170.Hibbard, M.J., 1991. Textural anomaly of twelve magma-mixed

granitoid systems. In: Didier, J., Barbarin, B. (Eds.), Enclaves and

Granite Petrology. Elsevier, Amsterdam, pp. 431444.

Holtz, F., Pichavant, M., Barbey, P., Johannes, W., 1992. Effects of

H2O on liquidus phase relations in the haplogranite system at 2 and

5 kbar. American Mineralogist 77, 12231241.

Holtz, F., Behrens, H., Dingwell, D.B., Johannes, W., 1995. Water

solubility in haplogranitic melts. Compositional, pressure and

temperature dependence. American Mineralogist 80, 94108.

Holtz, F., Johannes, W., Tamic, N., Behrens, H., 2001. Maximum and

minimum water contents of granitic melts generated in the crust: a

reevaluation and implications. Lithos 56, 114.

Icenhower, J., London, D., 1996. Experimental partitioning of Rb, Cs,

Sr, and Ba between alkali feldspar and peraluminous melt.

American Mineralogist 81, 719734.

King, P.L., White, A.J.R., Chappell, B.W., Allen, C.M., 1997.

Characterization and origin of aluminous A-type granites from

the Lachlan Fold Belt, Southeastern Australia. Journal of Petrology

38, 371391.

King, P.L., Chappell, B.W., Allen, C.M., White, A.J.R., 2001. Are A-

type granites the high temperature felsic granites? Evidence from

fractionated granites of the Wangrah Suite. Australian Journal of

Earth Sciences 48, 501514.

Klimm, K., Holtz, F., Johannes, W., King, P.L., 2003. Fractionation of

metaluminous A-type granites: an experimental study of theWangrah

Suite, Lachlan Fold Belt, Australia. Precambrian Research 124,

327341.Kretz, R., 1983. Symbols for rock-forming minerals. American

Mineralogist 68, 277279.

Loiselle, M.C., Wones, D.R., 1979. Characteristics and origin of

anorogenic granites. Geological Societyof America,Abstracts 11,468.

Lumbers, S.B., Wu, T.W., Heaman, L.M., Vertolli, V.M., MacRae, N.D.,

1991. Petrology and age of the A-type Mulock granite batholith,

northern Grenville Province, Ontario. Precambrian Research 53,

199231.

Martin, H., Bonin, B., Capdevila, R., Jahn, B.M., Lameyre, J., Wang,

Y., 1994. The Kuiqi peralkaline granitic complex (SE China):

petrology and geochemistry. Journal of Petrology 35, 9831015.

Nash, W.P., Crecraft, H.R., 1985. Partition coefficients for trace

elements in silicic magmas. Geochimica et Cosmochimica Acta

49, 23092322.



20/20

Nekvasil, H., 1991. Ascent of felsic magmas and formation of

rapakivi. American Mineralogist 76, 12791290.

Patio Douce, A.E., 1997. Generation of metaluminous A-type

granites by low-pressure melting of calc-alkaline granitoids.

Geology 25, 743746.

Rajesh, H.M., 2000. Characterization and origin of a compositionally

zoned aluminous A-type granite from South India. GeologicalMagazine 137, 291318.

Skjerlie, K.P., Johnston, A.D., 1992. Vapor-absent melting at 10 kbar

of a biotite- and amphibole-bearing tonalitic gneiss: implications

for the generation of A-type granites. Geology 20, 263266.

Skjerlie, K.P., Johnston, A.D., 1993. Fluid-absent melting behavior of an

F-rich tonalitic gneiss at mid-crustal pressures: implications for the

generation of anorogenic granites. Journal of Petrology 34, 785815.

Turner, S.P., Foden, J.D., 1996. Magma mingling in late-Delamerian

A-type granites at Mannum, South Australia. Mineralogy and

Petrology 56, 147169.

Turner, S.P., Foden, J.D., Morrison, R.S., 1992. Derivation of some A-

type magmas by fractionation of basaltic magma: an example from

the Padthaway Ridge, South Australia. Lithos 28, 151179.

Vander Auwera, J., Bogaerts, M., Liegeois, J.P., Demaiffe, D.,

Wilmart, E., Bolle, O., Duchesne, J.C., 2003. Derivation of the

1.00.9 Ga ferro-potassic A-type granitoids of southern Norway

by extreme differentiation from basic magmas. PrecambrianResearch 124, 107148.

Wark, D.A., Stimac, J.A., 1992. Origin of mantled (rapakivi) feldspars:

experimental evidence of a dissolution- and diffusion-controlled

mechanism. Contribution to Mineralogy and Petrology 111, 345361.

Watson, E.B., Harrison, T.M., 1983. Zircon saturation revisited:

temperature and composition effects in a variety of crustal magma

types. Earth and Planetary Science Letters 64, 295304.

Whalen, J.B., Currie, K.L., Chappell, B.W., 1987. A-type granites:

geochemical characteristics, discrimination and petrogenesis.

Contribution to Mineralogy and Petrology 95, 407419.


fractionation vs magma mixing in wangrah suite a-type granites_lfb

Documents