Fractionation vs. magma mixing in the Wangrah Suite A-type granites, Lachlan Fold Belt, Australia: Experimental constraints

Available online at www.sciencedirect.com

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

Fractionation vs. magma mixing in the Wangrah Suite A-type granites, Lachlan Fold Belt, Australia: Experimental constraints Kevin Klimm a,⁎, Francois Holtz a , Penelope L. King b a b

Institut für Mineralogie, Universität Hannover, Welfengarten 1, D-30167 Hannover, Germany 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 mafic” to “most felsic” by increasing SiO2 and decreasing FeOtotal. Experimental investigation show that the initial water content in melts from DCG is between 2–3 wt. % H2O. If the DCG is viewed 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 isobarically and are confirmed by bulk rock Ba, Sr, Rb and Zr concentrations. In contrast, the generation of the most abundant WG composition cannot be explained by fractional crystallisation from 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⁎ Corresponding author. Department of Earth Sciences, University of Bristol, Wills Memorial Building, Queen's Road, Bristol BS8 1RJ, 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

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 compositional 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; Patiño Douce, 1997; Dall'Agnol et al., 1999) and

416

K. Klimm et al. / Lithos 102 (2008) 415–434

the depths of melting and the prevailing oxygen fugacity (f O2) could account for the diversity in A-type granitic compositions (e.g., Patiño 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; Vander Auwera et al., 2003). A-type magmatism is accompanied by high heat flux caused by either deeper mafic magmatism or crustal extension (rifting; e.g. Haapala and Rämö, 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 crystallisation is poorly known. Fractional crystallisation is one of the major differentiation processes in A-type granitic systems (Collins et al., 1982; Clemens et al., 1986; Chappell et al., 1987; Whalen et 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 AB412

wt. % SiO2 TiO2 Al2O3 FeO MnO MgO CaO Na2O K2O P2O5 rest Total Qz Ab Or Ab/An ppm F Rb Sr Ba Zr Nb Ga Y La Ce Zn Zr-sat. T [°C] c a

70.45 0.54 13.26 4.06 0.08 0.61 1.93 3.39 3.98 0.18 0.23 98.71 34.5 36.0 29.5 3

n.d. 146 128 505 535 26 22.2 75 64 145 81 897

AB422

72.53 0.37 13.08 2.50 0.05 0.48 1.31 3.32 4.88 0.12 0.17 98.81 33.9 32.6 33.5 4

1040 228 87 215 313 21 21.2 64 53 121 52 843

Starting glass compositions AB421

75.32 0.15 12.71 1.62 0.06 0.23 0.90 3.36 4.61 0.05 0.12 99.13 38.3 31.5 30.2 6

670 216 56 148 190 20 19.6 77 28 63 52 806

AB401

76.67 0.09 12.10 0.92 0.02 0.05 0.53 3.25 5.12 0.01 0.09 98.85 38.6 29.3 32.2 10

AB412 b

AB422 b

AB421

AB401

n = 10

σ

n = 11

σ

n = 10

σ

n = 14

σ

71.89 0.55 13.60 4.14 0.07 0.62 1.96 3.38 3.79 – – 100.00

0.40 0.04 0.11 0.15 0.06 0.02 0.02 0.10 0.11

73.85 0.38 13.42 2.58 0.05 0.46 1.32 3.34 4.62 – – 100.00

0.44 0.05 0.17 0.13 0.04 0.01 0.06 0.14 0.10

75.65 0.17 13.04 1.71 0.06 0.19 0.96 3.40 4.80 – – 100.00

0.39 0.04 0.19 0.09 0.05 0.02 0.05 0.16 0.15

77.56 0.10 12.39 1.13 0.06 0.04 0.51 3.45 4.76 – – 100.00

0.44 0.02 0.33 0.11 0.06 0.02 0.04 0.17 0.16

n.d. 187 14 61 116 11 22.4 125 24 64 28 764

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 After Klimm et al. (2003). c Calculated zircon saturation temperatures after Watson and Harrison (1983).

K. Klimm et al. / Lithos 102 (2008) 415–434

systems (Clemens et al., 1986; Dall'Agnol et al., 1999) giving insufficient information about crystal-melt equilibrium as a function of temperature, pressure, volatile content (e.g. H2O, F), oxygen fugacity (f O2) or bulk composition.

417

Generally crystal-melt equilibrium experiments are performed with the goal of constraining parameters such as primary water content in the melt, temperature, pressure and f O2 (e.g., Clemens et al., 1986; Dall'Agnol et al., 1999). Clemens et al. (1986) and Dall'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 of Klimm 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 melt compositions in experiments with AB421; grey circles: residual melt compositions in experiments with AB401.

418

K. Klimm et al. / Lithos 102 (2008) 415–434

Table 2 Experimental results for the composition AB421 at 200 MPa and logf O2 ∼ NNO Run

T [°C]

Duration [days]

XH2Oin a

ΔNNO b

cH2O [wt. %] c

Results (+Gl, Fl) d

61 60 76 77 65 64 80 84 85 75 74 73 72 52 59 58 57 56

700 700 725 725 750 750 760 775 775 800 800 800 800 800 850 850 850 850

32 32 32 32 32 32 32 32 32 8 8 8 8 8 8 8 8 8

1.0 0.80 0.89 0.80 0.90 0.71 0.80 0.69 0.57 0.89 0.79 0.71 0.60 0.51 0.71 0.59 0.49 0.42

±0 − 0.19 − 0.10 − 0.19 − 0.09 − 0.30 − 0.19 − 0.32 − 0.49 − 0.10 − 0.20 − 0.30 − 0.32 − 0.59 − 0.30 − 0.46 − 0.62 − 0.75

6.3 5.2 5.8 5.4 5.8 4.9 5.1 4.2 3.9 5.1 4.5 3.9 3.2 2.7 3.8 3.0 2.5 2.1

Bt, Ox Bt, Pl (An18.3Or9.5), Kfs (An1.3Or68.3), Qz, Ox Bt, Ox Bt, Pl (An22.5Or8.1), Qz, Ox Bt, Ox Bt, Pl (An23.4Or10.6), Kfs (An1.8Or60.0), Qz, Ox Bt, Pl (An28.6Or5.5), Ox Opx (En34.9Wo2.7), Bt, Pl (An30.1Or5.4), Ox Opx (En27.4Wo2.6), Bt, Pl (An21.6Or11.2), Ox Ox Opx (En39.2Wo2.5), Ox Opx (En37.3Wo2.8), Ox Opx (En32.8Wo3.2), Pl (An27.2Or7.3), Qz, Ox Opx (En28.6Wo3.0), Pl (An24.3Or9.9), Qz, Ox Ox Ox Opx (En39.3Wo2.9), Pl (An33.4Or7.0), Ox Opx (En38.7Wo2.8), Pl (An29.2Or8.0), Qz, Ox

a

XH2Oin = H2O / (H2O + CO2) loaded in the capsule (in moles). ΔNNO = log f O2 (experiment) − log f O2 (NNO; Chou, 1987). For water saturated experiments at temperatures between 700 and 850 °C in the CSPV the f O2 = NNO. For H2O-undersaturated charges, a maximum possible f O2 is calculated as log f O2 = log f O2(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 by Kretz (1983): Fl: fluid; Gl: glass; Pl: plagioclase; Kfs: K-feldspar; Qz: quartz; Ox: Fe–Ti-oxide; Bt: biotite; Opx: orthopyroxene; Glass and fluid were present in all runs. b

(1999) performed experiments using one bulk composition, assumed to be representative of a studied pluton. Thus, these studies provide information on only one part of the crystallisation history because fractionation processes 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 compositions from the A-type granite Wangrah Suite, Lachlan Fold Belt, SE Australia, to model the fractionation processes 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 3 Experimental results for the composition AB401 at 200 MPa and logfO2 ∼ NNO Run

T [°C]

Duration [days]

XH2Oin a

ΔNNO b

cH2O [wt. %] c

Results (+Gl, Fl) d

62 63 78 79 66 67 81 82 83 86 87 71 70 69 68

700 700 725 725 750 750 760 760 760 775 775 800 800 800 800

32 32 32 32 32 32 32 32 32 32 32 8 8 8 8

1.0 0.81 0.90 0.80 0.91 0.71 0.90 0.78 0.68 0.69 0.58 0.88 0.75 0.61 0.51

±0 − 0.18 − 0.09 − 0.19 − 0.08 − 0.30 − 0.09 − 0.22 − 0.33 − 0.32 − 0.47 − 0.11 − 0.25 − 0.43 − 0.59

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

Bt, Ox Bt, Pl (An13.0Or9.7), Kfs (An1.9Or57.8), Qz, Ox Bt, Ox Bt, Pl (An12.4Or18.2), Kfs (An3.2Or54.7), Qz, Ox Bt, Pl (An12.0Or18.4), Kfs (An2.3Or51.8), Qz, Ox Bt, Pl (An11.4Or18.4), Kfs (An2.0Or56.0), Qz, Ox Ox Ox Bt, Pl (An11.3Or18.9), Kfs (An2.4Or54.9), Qz, Ox – Opx (En30.6Wo1.8), Pl (An9.4Or25.6), Kfs (An3.2Or50.3), Qz, Ox – – Ox Pl (An11.6Or18.6), Qz, Ox

a–d

see Table 2.

K. Klimm et al. / Lithos 102 (2008) 415–434

Fig. 2. Comparison of the obtained melt H2O contents of the six 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 (σ) 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 the y-axis.

four compositions (Danswell Creek, Wangrah, Eastwood and Dunskeig Granite) were interpreted to result from fractional 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 crystallisation 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 SEAustralia. 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 by King et al. (2001). They concluded that the granites were emplaced at shallow level (100–200 MPa) and f O2 below NNO. Within the LFB, the Wangrah Suite is a typical A-type granitic association. In the following section, a brief summary of

419

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 to An10. 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 pink–grey 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 groundmass of K-feldspar, quartz, plagioclase and biotite ± amphibole. K-feldspar often shows ovoid crystal form and is often rimmed (partly and completely) by plagioclase displaying local rapakivi-texture. Two generations 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 to An10. 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 Kfeldspar in the matrix is 13%. Quartz and mafic mineral abundance is relatively constant with 23–30% quartz and 5–12% 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 (25–30%). Mafic minerals are subordinate with 3–7%. Anorthite content of plagioclase vary from An25 to An5. The most evolved rock of the Wangrah Suite is the brick red equigranular Dunskeig Granite. This rock is characterised by the highest amount of K-feldspar (40– 50%) and the lowest distribution of mafic minerals b 3%. 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

420

K. Klimm et al. / Lithos 102 (2008) 415–434

Table 1. The samples AB412, AB422, AB421 and AB401 come from the DCG, WG, EG and DG, respectively. The chemical variations within the Wangrah 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 close to 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 NNO1.05 in a temperature range between 700 and 900 °C (Klimm et al., 2003). Based on plagioclase stability, plagioclase composition and Zr saturation temperatures Klimm et al. (2003) suggested that the initial water content in the primary melt was between 2–3 wt. % 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 10 ± 0.2 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 fluid. The water activity (aH 2 O) of the experimental charges was varied by adding a fluid composed of a mixture of H2O and CO2. CO2 was 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 (f O2) was monitored by adding a solid Ni–NiO buffer. The pressure vessels are made of Ni-alloy, which allows to perform long run duration with the solid buffer technique at f O2 = NNO. Because water activity and water fugacity (f H2O) varies in the experiments (as a function of the mole fraction of water in the charge), f O2 is not strictly constant at given P and T. In our experiments f O2 ranges from NNO to NNO-1.05 log units (Tables 2 and 3). After the runs, a H–O or a H–O–C 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 20 μm when possible (lowest value was 5 μm, Na was measured first). The H2O contents of the glasses were estimated with microprobe analyses following the difference method calibrated with standard glasses. Six hydrous 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. 2 shows 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 experiments. The results obtained by the difference method are identical 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 in Tables 2 and 3. The results of the

K. Klimm et al. / Lithos 102 (2008) 415–434

421

Fig. 3. Phase relations for composition AB421 and composition AB401 as a function of temperature–melt H2O content at 200 MPa and f O2 ∼ NNO. Solid dots represent experimental charges at given run conditions. Mineral abbreviations as given by Kretz (1983). Stability curves are labelled with mineral names lying inside their stability field. Uncertain portions of stability curves are indicated by dashed lines.

crystallisation experiments for AB421 and AB401 are represented in phase diagrams (Fig. 3). The phase diagrams are isobaric T–cH2Omelt sections where cH2Omelt is 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 T–aH2O, because appropriate T–aH2O conditions have not been achieved within the presented set of experiments. However, this curve has been reported in Fig. 3 using data from the Qz–Ab–Or–H2O–CO2 system (Holtz et al., 2001). The solidus temperature is lowered by ∼25 °C when compared to Holtz et al. (2001) to account for slight deviations from the synthetic Qz– Ab–Or system which may decrease the solidus temperature (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 temperature dependence of −16 × 10− 4 wt. % H2O/°C. In the most felsic composition AB401 liquidus conditions were attained at cH2Omelt ≥ 4 wt. % and temperatures ≥ 775 °C. For both compositions Fe–Ti oxides are the liquidus phases. In the less evolved composition AB421 near liquidus conditions could only be obtained at 850 °C with cH2Omelt N 3 wt. % and at 800 °C with cH2Omelt ≥ 5 wt. % H2O, respectively. With decreasing temperatures and for cH2Omelt b 3.5 wt. % H2O, the crystallisation of Fe–Ti-oxide is followed by orthopyroxene, plagioclase and quartz. The orthopyroxene stability field is restricted to high temperatures

above 750 °C and cH2Omelt b 5 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 Fe–Ti-oxide is followed by quartz + plagioclase at cH2Omelt b 3.5 wt. %. At H2O contents of the melt between 3.5 to 5 wt. %, quartz, plagioclase, K-feldspar and biotite crystallise

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 error of water content determination (horizontal bar).

422

Table 4 Compositions of experimental plagioclase (wt. %) at 200 MPa and logf O2 ∼ NNO AB421 Run

AB401

60

77

T [°C] 700

r

cH2O a 5.2

n = 6 5.4

725

64 r

750

n = 6 4.9

80 r

760

n = 3 5.1

84 r

775

n = 5 4.2

85 r

775

n = 7 3.9

72 r

800

n = 5 3.2

a b

62.65 23.07 0.34 5.33 7.07 1.20 99.65 27.2 65.4 7.3

r

800

n = 7 2.7 1.11 0.98 0.08 0.49 0.31 0.25 1.9 1.4 1.8

64.70 22.87 0.28 4.87 7.31 1.67 101.70 24.3 65.9 9.9

57 r

850

n = 1 2.5 62.30 23.51 0.36 6.50 6.40 1.14 100.20 33.4 59.6 7.0

56 r

850

63 r

79

700

r

725

n = 5 2.1

n = 4 5.1

n = 5 5.5

1.27 0.89 0.05 0.56 0.30 0.22

1.01 0.43 0.18 0.44 0.20 0.26

0.45 0.38 0.11 0.34 0.12 0.27

2.2 1.7 1.5

63.51 22.76 0.30 5.71 6.77 1.31 100.36 29.2 62.8 8.0

1.7 0.6 1.8

65.62 21.57 0.31 2.64 8.65 1.66 100.44 13.0 77.2 9.7

1.7 0.4 1.6

66 r

67

750

n = 1 5.1

65.24 21.16 0.20 2.54 7.85 3.12 100.12 12.4 69.4 18.2

65.12 20.83 0.34 2.48 7.93 3.19 99.89 12.0 69.6 18.4

r

83

87

68

r

760

r

775

r

n = 3 4.0

n=4

4.2

n=4

3.4

n = 3 2.7

n=2

0.40 66.22 0.22 20.43 0.03 0.21 0.21 2.28 0.31 7.73 0.12 3.08 99.95 1.0 11.4 1.2 70.1 0.9 18.4

0.49 0.61 0.03 0.19 0.42 0.55

66.09 20.59 0.41 2.27 7.76 3.20 100.31 11.3 69.8 18.9

0.45 0.51 0.19 0.19 0.16 0.41

66.65 20.36 0.16 1.88 7.23 4.31 100.58 9.4 65.1 25.6

0.36 0.17 0.02 0.16 0.05 0.25

0.02 0.00 0.07 0.15 0.10 0.12

750

1.0 2.9 3.5

0.9 1.5 2.4

0.9 0.3 1.2

r

800

66.38 20.72 0.18 2.29 7.65 3.10 100.32 11.6 69.8 18.6

0.6 0.3 0.9

Water contents of residual glasses determined following the difference method calibrated with six hydrous rhyolitic standard glasses. Total Fe as Fe2O3.

Table 5 Compositions of experimental K-feldspar (wt. %) at 200 MPa and logf O2 ∼ NNO AB421

AB401

Run

60

T [°C]

700

r

750

r

700

r

725

r

750

r

750

r

760

r

775

r

cH2O a

5.2

n=5

4.9

n=5

5.1

n=7

5.5

n=3

5.1

n=5

4.0

n=4

4.2

n=7

3.4

n=3

SiO2 Al2O3 Fe2O3 b CaO Na2O K2O Total An Ab Or

65.96 18.58 0.27 0.26 3.34 11.42 99.83 1.3 30.4 68.3

0.38 0.13 0.13 0.05 0.13 0.10

66.47 18.67 0.24 0.34 4.08 9.73 99.52 1.8 38.2 60.0

0.53 0.30 0.10 0.09 0.17 0.28

66.76 18.71 0.12 0.38 4.40 9.58 99.95 1.9 40.3 57.8

0.64 0.39 0.04 0.18 0.42 0.51

66.17 19.13 0.21 0.64 4.64 9.19 99.97 3.2 42.1 54.7

0.39 0.12 0.08 0.11 0.10 0.42

66.88 18.82 0.13 0.46 5.06 8.66 100.01 2.3 45.9 51.8

66.56 18.70 0.11 0.39 4.59 9.29 99.64 2.0 42.0 56.0

0.60 0.30 0.04 0.09 0.34 0.60

66.20 18.91 0.34 0.49 4.81 9.42 100.16 2.4 42.7 54.9

0.27 0.17 0.25 0.10 0.11 0.18

66.76 18.89 0.19 0.64 5.15 8.46 100.08 3.2 46.5 50.3

0.42 0.41 0.02 0.03 0.13 0.30

a b

64

0.2 1.0 1.1

63

0.5 0.9 0.5

79

0.9 3.0 3.6

66

0.6 1.4 1.9

67

0.29 0.31 0.02 0.11 0.26 0.21 1.1 22.3 28.6

Water contents of residual glasses determined following the difference method calibrated with six hydrous rhyolitic standard glasses. Total Fe as Fe2O3.

83

0.5 3.0 3.5

87

0.5 0.7 0.9

0.3 0.2 0.4

K. Klimm et al. / Lithos 102 (2008) 415–434

64.93 1.13 63.19 0.61 64.78 0.61 61.91 0.14 61.57 0.43 64.10 0.84 SiO2 21.81 0.69 22.91 0.45 21.82 0.43 23.70 0.41 24.05 0.45 22.63 0.25 Al2O3 0.46 0.15 0.39 0.13 0.59 0.05 0.33 0.05 0.31 0.03 0.31 0.05 Fe2O3 b CaO 3.62 0.19 4.54 0.32 4.36 0.35 5.81 0.17 6.13 0.35 4.30 0.14 7.91 0.19 7.72 0.15 6.81 0.15 7.38 0.12 7.25 0.05 7.39 0.17 Na2O 1.58 0.19 1.38 0.21 1.66 0.31 0.94 0.05 0.92 0.08 1.88 0.16 K2O Total 100.30 100.12 100.02 100.06 100.23 100.61 An 18.3 1.1 22.5 1.5 23.4 1.4 28.6 0.8 30.1 1.4 21.6 1.0 Ab 72.2 0.6 69.4 0.8 66.0 1.9 65.8 0.6 64.5 1.0 67.1 0.5 Or 9.5 1.1 8.1 1.2 10.6 1.0 5.5 0.4 5.4 0.5 11.2 0.8

52

Table 6 Compositions of experimental orthopyroxene (wt. %) at 200 MPa and logfO2 ∼ NNO AB421 Run

84

AB401 85

74

73

72

52

57

56

775

775

r

775

r

800

r

800

r

800

r

800

r

850

r

850

r

87

r

− 0.32

n=4

− 0.49

n=4

−0.20

n=5

− 0.30

n=6

− 0.32

n=6

−0.59

n=3

− 0.62

n=4

−0.75

n=3

− 0.47

n=2

SiO2 TiO2 Al2O3 FeO b MnO MgO CaO Na2O K2O Total XMg En Fs Wo

50.05 0.11 1.06 33.44 1.82 11.12 1.21 0.08 0.15 99.04 0.37 34.9 62.1 2.7

0.14 0.02 0.05 0.22 0.08 0.11 0.02 0.01 0.02

48.32 0.13 1.18 36.51 2.16 8.52 1.13 0.04 0.12 98.11 0.29 27.4 69.8 2.6

0.45 0.02 0.21 1.17 0.04 0.30 0.33 0.01 0.08

50.18 0.12 0.66 32.27 1.70 12.87 1.13 0.05 0.06 99.04 0.42 39.2 58.1 2.5

0.28 0.03 0.08 0.32 0.12 0.15 0.09 0.02 0.01

50.00 0.13 0.92 32.57 1.67 11.98 1.24 0.05 0.08 98.66 0.40 37.3 59.7 2.8

0.59 0.02 0.21 0.95 0.08 0.43 0.10 0.03 0.05

51.04 0.12 1.81 32.23 1.61 9.88 1.35 0.17 0.37 98.58 0.35 32.8 63.2 3.2

1.15 0.01 0.57 1.61 0.13 0.77 0.15 0.18 0.22

50.38 0.19 2.29 35.58 1.66 8.84 1.27 0.19 0.32 100.73 0.31 28.6 67.6 3.0

0.47 0.01 0.11 0.76 0.06 0.11 0.19 0.01 0.02

50.50 0.11 0.90 31.74 1.60 12.74 1.30 0.04 0.07 99.01 0.42 39.3 57.7 2.9

0.36 0.03 0.21 0.33 0.04 0.27 0.13 0.01 0.01

52.33 0.13 2.01 30.48 1.64 12.12 1.23 0.24 0.40 100.58 0.41 38.7 57.5 2.8

0.23 0.02 0.28 0.66 0.03 0.51 0.12 0.10 0.07

49.06 0.10 0.99 34.47 1.54 9.27 0.76 0.17 0.23 96.60 0.32 30.6 66.8 1.8

0.37 0.02 0.12 0.66 0.02 0.40 0.07 0.10 0.10

a b

0.00 0.3 0.3 0.1

0.01 1.0 1.7 0.7

0.00 0.3 0.3 0.2

0.01 1.3 1.3 0.2

0.01 1.1 0.2 0.3

0.01 0.5 1.0 0.5

0.01 0.7 0.8 0.3

0.01 1.2 1.5 0.3

0.01 0.6 0.3 0.1

ΔNNO = log f O2 (experiment) − log f O2 (NNO; Chou, 1987). Total Fe as FeO.

Table 7 Compositions of experimental glasses (wt. %, normalised to 100 on anhydrous basis) of AB421 at 200 MPa and logf O2 ∼ NNO Run

61

σ

60

σ

76

n = 6 700

SiO2 TiO2 Al2O3 FeO a MnO MgO CaO Na2O K2O Total cH2O b Qz Ab Or

0.34 77.36 0.46 77.10 0.40 0.02 0.08 0.03 0.06 0.03 0.20 13.01 0.20 12.97 0.08 0.04 1.10 0.14 0.93 0.10 0.02 0.08 0.07 0.08 0.04 0.04 0.03 0.03 0.09 0.04 0.09 0.47 0.06 0.78 0.05 0.14 3.06 0.15 3.14 0.17 0.16 4.82 0.11 4.84 0.11 100.00 100.00 5.2 5.8 1.5 41.9 1.4 40.6 1.8 1.4 27.7 1.3 28.6 1.6 1.0 30.4 0.7 30.8 0.7

a

77

n = 8 725 76.91 0.07 13.00 1.03 0.10 0.07 0.53 3.06 5.22 100.00 5.4 39.5 27.6 32.9

σ

65

σ

64

σ

80

σ

84

σ

85

75

n = 6 750

n = 10 760

n = 7 775

0.42 0.04 0.20 0.26 0.09 0.03 0.08 0.09 0.16

0.41 0.04 0.06 0.14 0.04 0.03 0.10 0.19 0.11

0.35 0.03 0.20 0.22 0.07 0.01 0.04 0.27 0.15

0.39 77.80 0.42 76.93 0.31 76.88 0.02 0.11 0.03 0.09 0.02 0.14 0.13 12.44 0.15 12.81 0.25 12.68 0.12 0.99 0.19 1.11 0.14 1.42 0.01 0.01 0.01 0.06 0.05 0.07 0.02 0.08 0.02 0.05 0.03 0.16 0.06 0.63 0.08 0.39 0.04 0.81 0.25 2.99 0.13 2.99 0.14 3.14 0.16 4.96 0.09 5.57 0.09 4.71 100.00 100.00 100.00 4.2 3.9 5.1 1.2 41.8 0.8 38.5 1.2 40.8 2.2 27.0 1.2 26.7 1.3 28.9 1.2 31.2 0.6 34.8 0.5 30.3

0.7 0.7 1.0

78.06 0.09 12.75 0.59 0.04 0.04 0.74 3.17 4.52 100.00 5.8 42.9 28.6 28.5

1.2 1.7 0.9

76.92 0.08 13.40 1.15 0.12 0.01 0.33 3.14 4.85 100.00 4.9 41.0 28.4 30.6

2.5 2.4 0.9

77.66 0.12 12.75 0.85 0.01 0.04 0.66 3.00 4.91 100.00 5.1 41.9 27.1 30.9

n = 7 775

σ

n = 6 750

Total Fe as FeO. Water contents determined following the difference method calibrated with six hydrous rhyolitic standard glasses.

n = 6 800

σ

74

σ

73

σ

72

σ

52

σ

59

σ

58

σ

57

σ

56

σ

n = 14 800

n = 7 800

n = 9 800

n = 10 800

n = 9 850

n = 11 850

n = 14 850

n = 12 850

n=9

0.38 0.02 0.20 0.22 0.04 0.04 0.07 0.16 0.18

0.32 0.04 0.18 0.13 0.09 0.03 0.10 0.19 0.21

0.34 0.05 0.26 0.21 0.02 0.03 0.09 0.11 0.15

0.35 0.03 0.14 0.18 0.01 0.03 0.07 0.23 0.14

0.32 0.04 0.11 0.13 0.05 0.03 0.03 0.09 0.19

0.28 0.03 0.15 0.16 0.04 0.03 0.04 0.11 0.10

0.36 0.02 0.15 0.14 0.05 0.03 0.07 0.08 0.09

0.42 0.02 0.15 0.21 0.06 0.02 0.03 0.11 0.17

0.33 0.03 0.24 0.19 0.06 0.03 0.04 0.09 0.08

1.9 1.4 1.1

76.89 0.14 12.67 1.33 0.09 0.15 0.83 3.17 4.74 100.00 4.5 40.5 29.1 30.4

2.1 1.8 1.3

76.70 0.14 12.65 1.36 0.04 0.15 0.79 3.41 4.76 100.00 3.9 38.4 31.2 30.4

1.6 1.0 0.9

77.03 0.15 12.72 1.20 0.01 0.09 0.59 3.30 4.92 100.00 3.2 39.2 29.8 31.0

1.7 2.0 1.0

76.95 0.16 12.76 1.22 0.05 0.09 0.49 3.02 5.26 100.00 2.7 39.6 27.2 33.2

1.3 0.8 1.1

76.99 0.15 12.75 1.41 0.05 0.18 0.85 3.04 4.58 100.00 3.8 42.2 28.2 29.6

0.8 1.1 0.7

76.98 0.15 12.74 1.36 0.05 0.19 0.87 3.11 4.55 100.00 3.0 41.9 28.7 29.4

1.0 0.7 0.6

77.15 0.16 12.68 1.29 0.07 0.17 0.79 3.13 4.56 100.00 2.5 41.9 28.8 29.3

1.7 1.0 1.1

77.33 0.17 12.65 1.19 0.03 0.16 0.62 3.04 4.81 100.00 2.1 41.7 27.7 30.6

1.0 0.8 0.6

423

b

77.66 0.06 12.68 0.90 0.02 0.07 0.81 3.21 4.59 100.00 6.3 41.7 29.2 29.1

n = 8 725

σ

T [°C] 700

K. Klimm et al. / Lithos 102 (2008) 415–434

T [°C] ΔNNO a

424

K. Klimm et al. / Lithos 102 (2008) 415–434

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 0.5–0.7 under the prevailing experimental conditions (200 MPa). At cH2Omelt N 5 wt. % Fe–Tioxide 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 from AB421 decrease with decreasing temperature and cH2Omelt (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 after Klimm et al. (2003). Error bars: standard deviation (σ) from EPMA. Symbols with no error bar in a–k indicate that σ was smaller than the symbol. For clarity only the maximum σ is presented in l–u (for details see Tables 7 and 8). The standard error on the melt H2O content is estimated to be ±0.5 wt. % H2O for all analyses (a–u). 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 presented because it does not vary much over the Wangrah Suite granites.

K. Klimm et al. / Lithos 102 (2008) 415–434

425

2003) at identical temperatures and melt H2O contents (Fig. 4). In runs with AB401 the An-content of plagioclase, ranging from An09 to An12, shows no variation with temperature or water content of the melt and contains relatively high Or-contents (Or9.7–25.6) reflecting the low CaO content of the bulk composition. The XMg variation range of orthopyroxene is lower for AB421 (0.29–0.42, Table 5) than for AB412 (0.30– 0.59) and AB422 (0.30–0.56) 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 μm), especially common in water undersaturated charges, makes it difficult to obtain reliable analyses. Biotite crystals are very small in water undersaturated 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 f O2 in agreement with Dall'Agnol et al. (1999). TiO2 of experimental biotite is ≤ 3 wt. %. Quenched glass compositions of AB421 and AB401 normalised to 100 wt. % (anhydrous) are reported in Tables 6 and 7 and Fig. 5l–u, respectively.

(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 of 4–5 wt. %. K-feldspar stability decreases to lower temperature (from N800 °C to ≤ 775 °C) for cH2Omelt of 3–4 wt. % but is shifted to higher temperatures (≤700 °C to 750 °C) for cH2Omelt of ∼5 wt. % with increasing K2O content of the starting material from AB412 to AB401 (Table 1), excluding composition AB422. In experiments with AB422 K-feldspar crystallises 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 (800–850 °C for AB412, 775–800 °C for AB421 and b 775 °C for AB401). The orthopyroxene stability field decreases and is shifted towards lower temperature and aH2O with decreasing bulk FeO and MgO contents.

5. Discussion

There are several possibilities to model the differentiation 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 from AB421. Third, the differentiation results from a combination of both processes.

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 in Fig. 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

5.2. Differentiation by fractional crystallisation

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.%

426

K. Klimm et al. / Lithos 102 (2008) 415–434

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) after Klimm 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.

plagioclase. However, the mass balance calculation for K2O was not successful and did not reproduce adequately the composition AB422 (K2O in AB422 and all other EG is too high to be modelled 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

1.9 2.2 1.0 2.1 1.5 1.0 1.5 1.4 1.1 1.7 1.9 0.8 1.7 1.6 0.6 1.0 1.1 0.9 1.4 1.4 1.0 1.6 1.8 1.0 Total Fe as FeO.

b Water contents determined following the difference method calibrated with six hydrous rhyolitic standard glasses.

1.3 1.2 0.8 2.4 2.2 0.7 2.1 1.8 1.3 1.8 1.5 0.6 1.4 1.4 0.9 1.3 1.2 0.7 a

n = 13

0.44 0.03 0.20 0.14 0.05 0.03 0.06 0.25 0.14

77.68 0.09 12.45 0.71 0.05 0.06 0.33 3.49 5.14 100.00 2.7 37.7 30.7 31.6

800 n = 10

0.39 0.02 0.24 0.17 0.00 0.04 0.06 0.18 0.17

77.81 0.10 12.35 1.01 0.00 0.06 0.44 3.46 4.75 100.00 3.2 39.6 30.8 29.5

800 n=9

0.37 0.03 0.20 0.17 0.01 0.02 0.03 0.16 0.17

77.88 0.09 12.20 0.94 0.04 0.04 0.46 3.48 4.87 100.00 4.1 39.0 30.9 30.1 0.30 0.03 0.19 0.13 0.01 0.02 0.06 0.23 0.12

77.58 0.10 12.34 1.00 0.07 0.05 0.43 3.47 4.94 100.00 4.9 38.5 30.9 30.6

800 n=7 800

0.31 0.03 0.12 0.11 0.04 0.02 0.07 0.19 0.12

n = 11

77.38 0.10 12.54 0.87 0.05 0.06 0.30 3.33 5.38 100.00 3.4 37.5 29.4 33.1

n = 9 775

0.32 0.03 0.17 0.17 0.03 0.03 0.04 0.12 0.17

77.86 0.07 12.50 0.69 0.06 0.06 0.41 3.37 4.97 100.00 4.2 39.4 29.8 30.7

775 n = 10

0.30 0.04 0.20 0.13 0.04 0.02 0.07 0.17 0.17

77.38 0.08 12.62 0.86 0.06 0.03 0.34 3.45 5.19 100.00 4.2 37.5 30.5 32.0

760 n=8

0.22 0.03 0.19 0.13 0.04 0.02 0.06 0.22 0.16

77.79 0.07 12.50 0.75 0.06 0.04 0.42 3.34 5.02 100.00 4.6 39.3 29.6 31.1

760 n = 10

0.35 0.02 0.15 0.12 0.04 0.03 0.06 0.13 0.14

77.71 0.06 12.52 0.75 0.07 0.06 0.42 3.35 5.05 100.00 5.0 39.1 29.7 31.2

n = 11 760

0.41 0.04 0.18 0.06 0.00 0.00 0.05 0.25 0.12

77.78 0.09 12.45 0.70 0.00 0.00 0.35 3.62 5.00 100.00 4.0 37.6 31.8 30.6

750 n=6

0.28 0.02 0.21 0.23 0.01 0.02 0.07 0.20 0.22

78.09 0.07 12.52 0.86 0.06 0.04 0.33 3.34 4.70 100.00 5.1 41.2 29.7 29.1

750 n=7

0.25 0.02 0.13 0.09 0.04 0.04 0.09 0.17 0.10

77.66 0.09 12.38 0.84 0.04 0.04 0.43 3.44 5.07 100.00 5.5 38.3 30.4 31.3

725 n=7

0.45 0.03 0.14 0.09 0.01 0.03 0.03 0.15 0.15

77.98 0.07 12.57 0.74 0.03 0.05 0.42 3.11 5.03 100.00 6.0 41.1 27.7 31.2 0.27 0.04 0.19 0.09 0.01 0.01 0.05 0.15 0.12

n = 11 725

427

in experiments with AB412. Fig. 5a–e 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 cH2Omelt b 3.5 wt. % (Fig. 5e). However, at these H2O contents 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.

77.83 0.07 12.63 0.90 0.02 0.01 0.32 3.57 4.66 100.00 5.1 39.6 31.6 28.8

700 n=8 700 T [°C]

78.44 0.19 SiO2 0.06 0.05 TiO2 12.27 0.11 Al2O3 a 0.64 0.14 FeO MnO 0.00 0.01 MgO 0.01 0.01 CaO 0.40 0.04 3.37 0.28 Na2O 4.81 0.12 K2O Total 100.00 b 6.3 cH2O Qz 40.7 2.0 Ab 29.7 2.4 Or 29.6 0.8

σ 67 σ 66 σ 79 σ 78 σ 63 σ 62 Run

Table 8 Compositions of experimental glasses (wt. %, normalised to 100) of AB401 at 200 MPa and logfO2 ∼ NNO

81

σ

82

σ

83

σ

86

σ

87

σ 71 σ

70

σ

69

σ

68

σ

K. Klimm et al. / Lithos 102 (2008) 415–434

5.2.2. AB421 (EG) Using the same method as described above (comparison of experimental glass compositions with bulk compositions, Fig. 5) the derivation of AB421 from AB412 or AB422 can be tested. The examination of Fig. 5a–e shows that residual glasses from AB412 with SiO2, Al2O3, FeO, CaO and K2O concentrations corresponding to the bulk AB421 have only been obtained in experiments at temperatures between 900 to 850 °C with cH2Omelt of 2 to 3 wt. %. Thus, AB421 can derive from AB412 by crystallisation of magnetite, orthopyroxene and plagioclase according to the phase relations in this T–aH2O 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. % orthopyroxene and 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, calculation 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 tested by examining Fig. 5f–k. The residual glass compositions 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 cH2Omelt N 4 wt. % (Fig. 5k). In contrast the CaO content of 0.91 wt. % in AB421 is mainly determined at cH2Omelt b 4 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 reproduce

428

K. Klimm et al. / Lithos 102 (2008) 415–434

Table 9 Results of mass balance calculations for major element fractional crystallisation

a

Deduced from experimental phase relations (Fig. 2 and Klimm et al., 2003). Due 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 Ti–Fe contents to FeO = 90 wt. % and TiO2 = 10 wt.% to count for the effect of aH2O. 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%. c Calculated change of water content after fractionation assuming an initial water content of 2.5 wt. % H2O for AB412 (according to 2–3 wt. % H2O after Klimm et al., 2003). b

residual glasses from AB422 with compositions identical/close to AB421. In conclusion the experimental results indicate that EG may be derived from DCG by fractional crystallisation. Assuming that melts representative of DCG contains 2.5 wt. % H2O (2–3 wt. % 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 composition of EG would contain ∼3 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 crystallisation temperatures of tectosilicates (Dall'Agnol et al., 1999). Furthermore the Zircon saturation geothermometer 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. 5a–e, 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 cH2Omelt ≤ 3.5 wt. % (Fig. 5e).

K. Klimm et al. / Lithos 102 (2008) 415–434

However, the SiO2 contents 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. 5l–p). In fact only runs saturated with respect to K-feldspar or close to saturation yield to glasses with high K2O. This suggests that AB401 corresponds to a melt in equilibrium with Kfeldspar 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 may play a minor role (b 0.5 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 temperature 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) can be 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 Table 10 Mineral/melt partition coefficients used for fractional crystallisation modelling Orthopyroxene Rb Ba Sr a b c d

a

0.01 0.1 a 0.01 a

Plagioclase 0.011 3–4 d 4.04 b

Bacon and Druitt (1988). Ewart and Griffin (1994). Nash and Crecraft (1985). Blundy and Wood (2003).

b

Biotite c

4.1 5.6 c 0.29 c

K-feldspar 2.4 c 16 d 4.5 c

429

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 is a 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 900–850 °C and cH2Omelt of 1.8–3.4 wt. % (Klimm et al., 2003). Therefore we calculated partition coefficients for Ba in feldspars using the linear relationship given by Blundy and Wood (2003). Maximum partition coefficients for Ba in the experimental orthoclase bearing plagioclases are DBa = 3–4 (Table 9).The composition 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 crystallisation 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 a DCG 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 not follow the fractional crystallisation trend obtained. The reason for that could be a) the natural pressure and f O2 conditions differ from the experimentally applied pressure of 200 MPa and f O2 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 f O2 and pressure of crystallisation of plutonic rocks is a difficult task and differences in element concentrations (e.g. Fe, Si) that were used to discriminate parent–daughter relationship

430

K. Klimm et al. / Lithos 102 (2008) 415–434

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). Mass fractions are obtained by mass balance calculations (details in Table 9). Vertical marks indicate 10% fractional crystallisation.

of natural rocks could be explained by inappropriate choice of one or two experimental parameters (P and/or f O2). For instance, evaluating the potential parent– daughter relationships between AB412 and AB422, it is found on the panels a) to e) of Fig. 5, that the closest intersection with the isothermal liquid line of descent of AB412 at 850 °C with the horizontal lines corresponding 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 natural f O2. Slightly more reducing conditions will lead to higher FeO contents in the experiments which will in turn decrease the SiO2 content 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.

K. Klimm et al. / Lithos 102 (2008) 415–434

SiO2 contents (Fig. 1b). The experimental liquids are too CaO rich relative to WG. This is due to the higher melt water contents of N 3 wt. % H2O for most of the experimental liquids. Melts with a melt H2O content b 3 wt. % that are in agreement with the initial melt water content of 2–3 wt. % H2O for AB412 have lower CaO and higher SiO2 contents and fall in the field of EG (Klimm et al., 2003). 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 at melt water contents ≥4.5 wt. % H2O. The low water solubility of ≤4 wt. % H2O at pressures ≤100 MPa is insufficient to provide amphibole crystallisation. In addition Klimm 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 some stage 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 Kfeldspar 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 (rapakivitexture) 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 processes involving 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 fractionation 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 (Patiño Douce, 1997; Dall'Agnol et al., 1999). Patiño Douce (1997) showed that low-pressure melting of a tonalitic or granodioritic composition at 400 MPa and temperatures N900 °C leads to A-type

431

granitic residual melts with cH2Omelt b 4 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 mineral assemblage 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 f O2 ∼ NNO + 2.5 and NNO-1.5. In addition to orthopyroxene, clinopyroxene was also obtained at near liquidus conditions. 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 for WG (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 (200–400 MPa) involves the same mineral assemblage of plagioclase and orthopyroxene and therefore polybaric processes such as pressurequenching 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.

432

K. Klimm et al. / Lithos 102 (2008) 415–434

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 (King et 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 mechanism leading to the segregation of the crystals from the residual melt and magma mixing with a melt composition 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 main processes, 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 evolved Dunskeig Granite (AB401). The modelled differentiation 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 2.5 ± 0.5 wt. % H2O), it has been demonstrated that the residual melts evolve towards a water undersaturated minimum temperature melt composition (minimum in the system Qz–Ab–Or–An–H2O). The fractionating mineral assemblage is dominated by the tectosilicates in the order of appearance (plagioclase → quartz → Kfeldspar). The Dunskeig Granite is close in composition to a water undersaturated minimum at aH2O between 0.5–0.7 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 differentiation trends and is related to the accumulation of K-feldspar evidenced by variable porphyric textures and 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 Kfeldspar 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

a

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

K. Klimm et al. / Lithos 102 (2008) 415–434

Creek Granite. Such a combination of fractional crystallisation 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. Patiño 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 Sr–Nd isotopic geochemistry of the Konso pluton, south-western Ethiopia: implications for transition from convergence to extension in the Mozambique Belt. International Journal of Earth Sciences 92, 873–890. 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 Geochemistry 52, 59–123. 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, 149–184. Chappell, B.W., White, A.J.R., Wyborn, D., 1987. The importance of residual source material (restite) in granite petrogenesis. Journal of Petrology 28, 1111–1138. 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. 69–99. Clemens, J.D., Holloway, J.R., White, A.J.R., 1986. Origin of an A-type granite: experimental constraints. American Mineralogist 71, 317–324. Collins, W.J., Beams, S.D., White, A.J.R., Chappell, B.W., 1982. Nature and origin of A-type granites with particular reference to southeastern Australia. Contribution to Mineralogy and Petrology 80, 189–200. 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, 1037–1039. Creaser, R.A., Price, R.C., Wormland, R.J., 1991. A-type granites revisited: assessment of a residual-source model. Geology 19, 163–166. 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 midcrustal, two-mica granite of anorogenic affinity. Precambrian Research 62, 139–169. Dall'Agnol, R., de Oliveira, D.S., 2007. Oxidized, magnetite-series, rapakivi-type granites of Carajás, Brazil: implications for classification and petrogenesis of A-type granites. Lithos 93, 215–233.

433

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 petrogenesis. Lithos 26, 115–134. Eklund, O., Shebanov, A.D., 1999. The origin of rapakivi texture by subisothermal decompression. Precambrian Research 95, 129–146. 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, 821–847. Ewart, A., Griffin, W.L., 1994. Application of proton-microprobe data to trace-element partitioning in volcanic rocks. Chemical Geology 117, 251–284. Frost, C.D., Frost, B.R., 1997. Reduced rapakivi-type granites: the tholeiite connection. Geology 25, 647–650. Haapala, I., Rämö, O.T., 1992. Tectonic setting and origin of the Proterozoic rapakivi granites of southeastern Fennoscandia. Transactions of the Royal Society of Edinburgh 83, 165–171. Hibbard, M.J., 1981. The magma mixing origin of mantled feldspars. Contribution to Mineralogy and Petrology 76, 158–170. 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. 431–444. 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, 1223–1241. Holtz, F., Behrens, H., Dingwell, D.B., Johannes, W., 1995. Water solubility in haplogranitic melts. Compositional, pressure and temperature dependence. American Mineralogist 80, 94–108. 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, 1–14. Icenhower, J., London, D., 1996. Experimental partitioning of Rb, Cs, Sr, and Ba between alkali feldspar and peraluminous melt. American Mineralogist 81, 719–734. 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, 371–391. King, P.L., Chappell, B.W., Allen, C.M., White, A.J.R., 2001. Are Atype granites the high temperature felsic granites? Evidence from fractionated granites of the Wangrah Suite. Australian Journal of Earth Sciences 48, 501–514. Klimm, K., Holtz, F., Johannes, W., King, P.L., 2003. Fractionation of metaluminous A-type granites: an experimental study of the Wangrah Suite, Lachlan Fold Belt, Australia. Precambrian Research 124, 327–341. Kretz, R., 1983. Symbols for rock-forming minerals. American Mineralogist 68, 277–279. Loiselle, M.C., Wones, D.R., 1979. Characteristics and origin of anorogenic granites. Geological Society of 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, 199–231. 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, 983–1015. Nash, W.P., Crecraft, H.R., 1985. Partition coefficients for trace elements in silicic magmas. Geochimica et Cosmochimica Acta 49, 2309–2322.

434

K. Klimm et al. / Lithos 102 (2008) 415–434

Nekvasil, H., 1991. Ascent of felsic magmas and formation of rapakivi. American Mineralogist 76, 1279–1290. Patiño Douce, A.E., 1997. Generation of metaluminous A-type granites by low-pressure melting of calc-alkaline granitoids. Geology 25, 743–746. Rajesh, H.M., 2000. Characterization and origin of a compositionally zoned aluminous A-type granite from South India. Geological Magazine 137, 291–318. 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, 263–266. 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, 785–815. Turner, S.P., Foden, J.D., 1996. Magma mingling in late-Delamerian A-type granites at Mannum, South Australia. Mineralogy and Petrology 56, 147–169.

Turner, S.P., Foden, J.D., Morrison, R.S., 1992. Derivation of some Atype magmas by fractionation of basaltic magma: an example from the Padthaway Ridge, South Australia. Lithos 28, 151–179. Vander Auwera, J., Bogaerts, M., Liegeois, J.P., Demaiffe, D., Wilmart, E., Bolle, O., Duchesne, J.C., 2003. Derivation of the 1.0–0.9 Ga ferro-potassic A-type granitoids of southern Norway by extreme differentiation from basic magmas. Precambrian Research 124, 107–148. 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, 345–361. 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, 295–304. Whalen, J.B., Currie, K.L., Chappell, B.W., 1987. A-type granites: geochemical characteristics, discrimination and petrogenesis. Contribution to Mineralogy and Petrology 95, 407–419.