TY - JOUR
AU - Müntener, Othmar
AB - Abstract Differentiation of hydrous primary, mantle-derived magmas is a fundamental process to generate evolved intermediate to SiO2-rich compositions forming the bulk of the continental and island arc crust. This study focuses on the results of equilibrium and fractional crystallization experiments at 1·0 GPa using two different primary magmas representing deep (90 km) and shallow (35 km) mantle extraction depths. Experiments on a hydrous high-Mg basalt were conducted at graphite-saturated and more oxidized conditions (NNO to NNO + 2, where NNO is nickel–nickel oxide buffer) to exploit the influence of fO2 on phase assemblages and the evolution of derivative liquids. The liquid line of descent (LLD) was simulated from liquidus to near-solidus conditions ranging from 1330°C to 720°C. H2O contents varied from about 2·0 to more than 10 wt %. The LLD covers the entire compositional range from high-Mg basalt to high-silica rhyolite and evolves from metaluminous to peraluminous compositions at 56–60 wt % SiO2 under oxidizing conditions. The observed crystallization sequences and the LLD reveal contrasting behavior depending on oxidation state, H2O content and equilibrium versus fractional crystallization. Equilibrium crystallization of high-Mg basalt under reducing conditions is initially dominated by olivine fractionation followed by Cr-rich spinel, clinopyroxene (cpx), and orthopyroxene (opx). Finally, between 1060 and 1000°C, amphibole formed by a peritectic reaction consuming cpx + olivine and forming amphibole + opx, resulting in 16% silica-undersaturated trachy-basaltic liquid. Equilibrium crystallization of the same composition under oxidizing conditions is characterized by strongly enhanced olivine and cpx fractionation and suppression of opx only occurring as a result of the peritectic amphibole-forming reaction at and below 1040°C. The liquid at 980°C is a peraluminous, alkali-poor, high-Al andesite representing ∼15% residual liquid. Fractional crystallization of the high-Mg basalt under oxidizing conditions evolves through fractionation of early olivine joined by cpx at 1200°C, followed by opx and hercynitic spinel (1140–1080°C) and amphibole at 1050°C coexisting with cpx (and spinel) to 980°C. At 950°C both garnet and plagioclase (plag) join amphibole as liquidus phases. This paragenesis (plus ilmenite and apatite) persists to 750°C with 16% residual liquid relative to the initial basaltic composition. Liquids evolve continuously from basalt to rhyolite, crossing the metaluminous–peraluminous boundary at about 60 wt % SiO2. Fractional crystallization of the basaltic andesite starting material differs at high temperature, where opx and cpx are the liquidus phases (1200–1080°C), followed by amphibole at the expense of opx. Below 950°C the phase assemblage is identical except at the lowest temperature (720°C), where quartz saturates in a high-silica rhyolitic liquid representing 20% of the initial basaltic andesite. Liquid compositions of the two starting compositions converge below 950°C, with the exception of K2O, which behaves incompatibly along the liquid line of descent. The H2O-undersaturated fractional crystallization experiments on the high-Mg basaltic composition under relatively reducing conditions evolve towards tholeiitic or alkalic compositions owing to early plagioclase (at 1140°C) and abundant opx (at 1110°C) crystallization followed by garnet–cpx–plag–ilmenite at 1080°C. Closed-system equilibrium crystallization under relatively oxidized conditions is characterized by significant expansion of the olivine stability field (>250°C, 1280–1000°C) relative to more reducing conditions, causing depletion of the liquid in ferrous iron and increase of ferric iron. The resulting fO2 of the coexisting basaltic liquids increases by more than 3 log units from about NNO + 2 to NNO + 5·5 over the temperature range >1200–1040°C, potentially offering an explanation for the more oxidized character of hydrous arc magmas as opposed to low-H2O tholeiitic magmas. INTRODUCTION The production of magmas at convergent plate margins is an important process forming juvenile island arc and continental crust (e.g. Taylor & White, 1965; Kushiro, 1990; Rudnick, 1995). The formation of intermediate (tonalitic–granodioritic) to granitic rocks forming large batholiths and associated volcanic provinces dominated by intermediate andesite to dacite volcanoes typically occurs in subduction-related settings such as the Patagonian batholith (Chile) (Kay & Ramos, 2006), the Peruvian batholith (Atherton, 1990), the Californian batholith (Lee et al., 2006), the Kohistan batholith (Jagoutz et al., 2009) or, on a smaller scale, the Tertiary Adamello batholith in Northern Italy (Callegari, 1983). A number of processes have been proposed to account for their genesis, including the following: (1) fractional crystallization of mantle-derived basaltic precursors (Bowen, 1928; Cawthorn et al., 1976; Green, 1980; Sisson & Grove, 1993; Grove et al., 2003); (2) additional extensive assimilation of and hybridization with crustally derived partial melts generated by heating of mantle-derived magmas over prolonged time (MASH/FARM) (Hildreth & Moorbath, 1988; Thompson et al., 2002); (3) partial melting of deep crustal lithologies (mafic amphibolites to granulites; e.g. Rapp & Watson, 1995), as well as pelitic micaschists (e.g. Clemens et al., 2011); (4) direct mantle partial melts (andesites; Kushiro, 1974); (5) slab melting (Kay, 1978; Drummond & Defant, 1990). To identify and quantify the relative significance of these processes, phase equilibria constraints, melt fractions and compositions, as well as proportions and compositions of solid residues generated by the proposed endmember processes need to be known. This contribution provides experimental constraints on both the equilibrium and fractional crystallization of primary, mantle-derived, hydrous, arc tholeiitic to calc-alkaline parental melts at pressure conditions corresponding to lower crustal levels (1·0 GPa, 30–35 km). We explored the crystallization-differentiation of two different starting compositions representing variable mantle source extraction conditions—a high-Mg basalt from the Adamello batholith (Italy) (Ulmer, 1988; Hürlimann et al., 2016) and a high-Mg basaltic andesite from Mt Shasta (California) (Baker et al., 1994)—to evaluate the potential influence of the ‘mantle input’. Performing both conventional equilibrium and fractional crystallization experiments further allowed us to quantify the relative importance of crystal–liquid separation during crystallization-differentiation. In addition to variable parental magma compositions, we also varied the fO2 and fH2O conditions to explore their influence on the liquid lines of descent (LLD). We discuss arguments in favor of, and/or in contradiction to the potential derivation of intermediate silicic magmas by fractional crystallization of precursor basaltic magmas, a subject that is still highly controversial (e.g. Sisson et al., 2005; Clemens et al., 2011; Jagoutz et al., 2011; Melekhova et al., 2013; Keller et al., 2015) but intimately linked to the generation of the continental crust. Our experimental data are relevant for the diversity of magmas fractionating in the lower crust. Upon decompression, these primary magmas will mix in most cases, as is overwhelmingly demonstrated for many arc volcanoes (e.g. Eichelberger, 1975; Hildreth & Moorbath, 1988; and many others). This study provides new experimental data to evaluate the geochemistry and modal proportions of silica-rich magmas generated by equilibrium and fractional crystallization in the lower crust, prior to decompression and mixing in shallow reservoirs. Both geochemically based calculations as well as phase equilibria (experimental and by thermodynamic modeling) constraints imply that in excess of 50% crystallization of ultramafic rocks from initial, mantle-derived magmas must have taken place to produce andesitic liquids at lower crustal conditions (e.g. Arndt & Goldstein, 1989; Kay & Kay, 1991; Müntener & Ulmer, 2006), a hypothesis that is directly tested and quantified by the presented experimental data. The diversity of experimentally derived cumulates and their relevance for the formation of arc lower crust are discussed in Müntener & Ulmer (2018). EXPERIMENTAL AND ANALYTICAL TECHNIQUES Experimental strategy This study encompasses both conventional isobaric equilibrium crystallization as well as isobaric fractional crystallization experiments. Equilibrium crystallization experiments are the classical way to explore the phase relations for a single starting material and involve varying the temperatures from the liquidus down to temperatures where the melt fractions become too small to be measured reliably (e.g. Fig. 1c). Fractional crystallization experiments are very tedious and time consuming, but liquid lines of descent may differ significantly between equilibrium and fractional crystallization. Provided the equilibrium phase diagram is known the LLD can be predicted and calculated, but for hydrous systems involving amphiboles (and potentially micas) phase equilibria are not precisely known and cannot reliably be calculated from thermodynamic data. The details of the procedure to conduct (near) perfect fractional crystallization experiments have been described in previous studies (Villiger et al., 2004; Nandedkar et al., 2014) and are briefly outlined here. The first steps consist of conventional equilibrium experiments conducted with either natural rock powders or synthetic equivalents of primary, hydrous Mg basaltic (Mb) and high-Mg basaltic andesite (ba) magmas to locate the liquidus of the respective composition at 1·0 GPa. Thereafter, the liquid composition coexisting with small to moderate amounts of solid phases was determined by electron probe microanalysis (EPMA), and for any subsequent experiment a new starting material corresponding to the liquid composition of the previous, higher temperature experiment was synthesized. The temperature steps between successive fractional crystallization experiments were mostly 30°C (rarely 50°C). All experiments were evaluated by least-squares regression analysis using the known starting composition and the compositions of all coexisting phases (glass and solid phase) to compute modal proportions, to ensure closed-system behavior, to detect potential disequilibrium, to calculate the correct amount of H2O to add to the next, lower temperature starting material, and to account for potential alkali deficits in the electron microprobe analyses (mainly caused by Na migration in silicate glasses), and iron loss to noble metal sample containers. Starting materials Both natural and synthetic starting materials were employed in this study. Two series of equilibrium crystallization experiments, EQ Mb Pt–C and EQ Mb AuPd, were conducted with a natural high-Mg basalt dyke composition from the southern part of the Tertiary Adamello batholith (RC158c; coordinates 45°57'19·1"N, 10°26'32·5"E; elevation 2366 m); the same was further used as the initial starting composition for two series of fractional crystallization experiments (FC Mb AuPd and FC Mb Pt–C) (Tables 1 and 2). The specific dyke consists of mostly pseudomorphed olivine (Fo 91·1) and clinopyroxene [Mg-number (Mg#) = Mg/(Mg + Fetot) of 0·89] phenocrysts both containing small, idiomorphic Cr-spinel inclusions in a fine-grained matrix composed of amphibole, minor plagioclase and magnetite and some secondary chlorite (Ulmer, 1986). The major element composition is Mg-rich (17 wt % on an anhydrous basis) and rather silica poor (47·4 wt %) and the Fo content and Ni contents of olivine phenocrysts (91·1 and 2800 ppm) and Mg# (0·764), Cr (1240 ppm) and Ni (330 ppm) contents of the whole-rock underline its nature as a primary mantle melt (Hürlimann et al., 2016). The trace element signature clearly reveals the subduction-related provenance (high field strength element depletion; light rare earth element and large ion lithophile element enrichment) of this dyke, which is post-plutonic, cross-cutting the gabbroic to intermediate plutonic rocks of the Southern Adamello batholith (Hürlimann et al., 2016). The original rock powder contains 3·0 ± 0·3 wt % H2O and 0·05 wt % CO2 as volatile components. Ulmer (1988) conducted a series of multiple-saturation experiments in graphite–Pt double capsules and established multiply saturated conditions of the high-Mg basalt containing 3·0 ± 0·3 wt % H2O at 2·8 GPa and 1370°C corresponding to about 90 km depth, where garnet, cpx, opx and olivine join the liquidus within experimental resolution, and identified the melting reaction as peritectic (cpx + gar + opx = ol + liquid). A similar, but less primitive composition (RC156; Hürlimann et al., 2016), an olivine-tholeiite from the same dyke-rock association, has been utilized as starting material for the lower pressure, fractional crystallization experiments at 0·4 and 0·7 GPa (Nandedkar, 2013; Nandedkar et al., 2014) to avoid a large number of experiments that solely crystallize olivine. However, the two compositions RC158c and RC156 are part of the same dyke generation and are directly linked by c. 15% olivine fractionation. Table 1: Composition of starting materials (in wt %) Starting material . SiO2 . TiO2 . Al2O3 . Cr2O3 . Fe2O3 . FeO . MnO . MgO . NiO . CaO . Na2O . K2O . P2O5 . H2O . Total . ΔNNO . xMg/xAn . FC ba AuPd 85-44 52·34 0·64 15·80 1·33 6·75 0·19 10·42 9·91 2·22 0·40 5·00 100·00 –0·12 0·700 F1b 52·07 0·65 15·67 1·47 7·51 0·15 9·60 10·16 2·32 0·41 5·12 100·00 –0·20 0·660 F2b 53·24 0·71 16·40 1·60 6·45 0·20 8·16 10·31 2·46 0·47 5·55 100·00 0·30 0·648 F3b 54·40 0·75 17·07 1·51 5·70 0·18 7·23 10·04 2·61 0·51 5·85 100·00 0·40 0·646 F4b 54·10 0·84 18·13 1·51 5·35 0·19 6·72 9·70 2·82 0·64 6·20 100·00 0·45 0·641 F5b 55·19 0·93 19·75 1·36 5·02 0·19 5·00 8·70 3·13 0·73 7·08 100·00 0·40 0·588 F6b 55·17 0·90 19·89 1·53 5·51 0·22 4·20 8·72 3·11 0·76 7·26 100·00 0·40 0·521 F7b 56·73 0·91 19·18 1·54 5·44 0·23 3·40 8·60 3·16 0·81 7·68 100·00 0·40 0·470 990 ba 55·71 0·90 20·26 1·54 5·35 0·23 3·38 8·46 3·19 0·80 0·18 8·50 100·00 0·00 0·472 950 ba 57·42 0·76 20·46 1·36 5·16 0·23 2·30 8·09 3·20 0·82 0·20 9·00 100·00 0·00 0·390 900 ba 62·19 0·52 19·79 1·03 3·48 0·13 1·06 6·79 3·70 1·05 0·27 9·50 100·01 0·00 0·300 850 ba 65·78 0·51 18·49 0·84 2·55 0·12 0·68 5·55 4·10 1·22 0·15 8·97 100·00 0·00 0·270 800 ba 70·73 0·19 16·83 0·57 1·50 0·11 0·40 3·71 4·20 1·61 0·16 8·24 100·01 0·00 0·262 EQ Mb Pt–C/EQ Mb AuPd RC158c 47·41 0·72 12·80 0·16 3·25 6·44 0·19 17·04 0·04 10·33 1·22 0·40 3·0 (3) 100·00 0·764 FC Mb Pt–C F1 48·59 0·91 15·59 0·00 9·04 0·15 11·01 12·68 1·53 0·51 3·17 100·01 0·685 F2 48·53 1·14 17·87 0·00 9·91 0·18 8·17 11·39 2·06 0·74 4·00 99·99 0·595 F3 48·44 1·19 18·65 0·00 10·24 0·18 7·57 10·74 2·18 0·80 4·32 99·99 0·568 F4 49·56 1·43 19·15 0·00 11·07 0·23 5·78 8·97 2·67 1·13 5·18 99·99 0·482 F5 49·94 2·18 17·78 0·00 14·31 0·33 3·47 6·90 3·29 1·79 10·65 99·99 0·302 FC Mb AuPd F1a 47·65 0·86 14·01 1·69 8·61 0·18 13·30 11·86 1·38 0·48 2·91 100·00 –0·17 0·701 F2a 47·39 0·64 15·21 1·71 8·23 0·22 11·24 13·41 1·38 0·58 3·13 100·00 –0·28 0·672 F3a 47·54 0·76 17·62 2·00 7·87 0·22 9·50 12·08 1·63 0·79 3·85 100·00 0·34 0·637 F4a 47·96 0·86 18·47 2·07 8·35 0·25 8·05 11·15 1·86 0·98 4·27 100·00 0·31 0·584 F5a 50·70 0·93 19·49 1·84 7·79 0·24 6·11 9·42 2·33 1·15 5·92 100·00 0·23 0·535 F5a 51·09 0·94 18·26 1·99 8·32 0·24 6·15 9·50 2·35 1·17 5·92 100·01 0·23 0·520 F6a 53·17 1·00 18·87 1·95 7·47 0·33 4·77 8·56 2·64 1·25 6·95 100·00 0·35 0·480 F7a 55·81 1·03 18·19 1·97 7·44 0·32 3·41 7·63 2·84 1·35 7·64 99·99 0·38 0·398 F8a 58·55 0·97 16·87 1·83 6·29 0·32 3·02 7·61 3·02 1·52 8·14 100·00 0·39 0·404 990 Mb 57·17 0·95 18·74 1·69 6·17 0·31 2·94 7·39 2·95 1·47 0·21 7·46 99·99 0·00 0·405 950 Mb 58·10 0·88 18·98 1·54 5·93 0·29 2·40 7·14 2·99 1·52 0·23 7·80 100·00 0·00 0·369 900 Mb 63·53 0·51 18·94 0·97 3·27 0·14 0·99 6·08 3·39 1·90 0·29 9·15 100·01 0·00 0·299 850 Mb 66·32 0·48 18·31 0·78 2·32 0·09 0·55 5·25 3·66 2·10 0·15 8·88 100·00 0·00 0·244 800 Mb 70·39 0·21 16·78 0·54 1·45 0·09 0·37 3·58 3·80 2·69 0·10 8·21 100·00 0·00 0·254 Seed crystals Plag (nom.) 44·40 35·83 19·20 0·54 0·03 100·00 0·952 Plag (EPMA) 44·70 0·04 36·12 0·64 0·10 19·66 0·60 0·01 101·85 0·948 Garnet (nom.) 37·82 0·05 19·34 37·12 1·55 2·80 0·99 99·67 0·119 Garnet (EPMA) 37·49 0·05 21·80 36·29 1·85 2·88 0·98 101·34 0·124 Starting material . SiO2 . TiO2 . Al2O3 . Cr2O3 . Fe2O3 . FeO . MnO . MgO . NiO . CaO . Na2O . K2O . P2O5 . H2O . Total . ΔNNO . xMg/xAn . FC ba AuPd 85-44 52·34 0·64 15·80 1·33 6·75 0·19 10·42 9·91 2·22 0·40 5·00 100·00 –0·12 0·700 F1b 52·07 0·65 15·67 1·47 7·51 0·15 9·60 10·16 2·32 0·41 5·12 100·00 –0·20 0·660 F2b 53·24 0·71 16·40 1·60 6·45 0·20 8·16 10·31 2·46 0·47 5·55 100·00 0·30 0·648 F3b 54·40 0·75 17·07 1·51 5·70 0·18 7·23 10·04 2·61 0·51 5·85 100·00 0·40 0·646 F4b 54·10 0·84 18·13 1·51 5·35 0·19 6·72 9·70 2·82 0·64 6·20 100·00 0·45 0·641 F5b 55·19 0·93 19·75 1·36 5·02 0·19 5·00 8·70 3·13 0·73 7·08 100·00 0·40 0·588 F6b 55·17 0·90 19·89 1·53 5·51 0·22 4·20 8·72 3·11 0·76 7·26 100·00 0·40 0·521 F7b 56·73 0·91 19·18 1·54 5·44 0·23 3·40 8·60 3·16 0·81 7·68 100·00 0·40 0·470 990 ba 55·71 0·90 20·26 1·54 5·35 0·23 3·38 8·46 3·19 0·80 0·18 8·50 100·00 0·00 0·472 950 ba 57·42 0·76 20·46 1·36 5·16 0·23 2·30 8·09 3·20 0·82 0·20 9·00 100·00 0·00 0·390 900 ba 62·19 0·52 19·79 1·03 3·48 0·13 1·06 6·79 3·70 1·05 0·27 9·50 100·01 0·00 0·300 850 ba 65·78 0·51 18·49 0·84 2·55 0·12 0·68 5·55 4·10 1·22 0·15 8·97 100·00 0·00 0·270 800 ba 70·73 0·19 16·83 0·57 1·50 0·11 0·40 3·71 4·20 1·61 0·16 8·24 100·01 0·00 0·262 EQ Mb Pt–C/EQ Mb AuPd RC158c 47·41 0·72 12·80 0·16 3·25 6·44 0·19 17·04 0·04 10·33 1·22 0·40 3·0 (3) 100·00 0·764 FC Mb Pt–C F1 48·59 0·91 15·59 0·00 9·04 0·15 11·01 12·68 1·53 0·51 3·17 100·01 0·685 F2 48·53 1·14 17·87 0·00 9·91 0·18 8·17 11·39 2·06 0·74 4·00 99·99 0·595 F3 48·44 1·19 18·65 0·00 10·24 0·18 7·57 10·74 2·18 0·80 4·32 99·99 0·568 F4 49·56 1·43 19·15 0·00 11·07 0·23 5·78 8·97 2·67 1·13 5·18 99·99 0·482 F5 49·94 2·18 17·78 0·00 14·31 0·33 3·47 6·90 3·29 1·79 10·65 99·99 0·302 FC Mb AuPd F1a 47·65 0·86 14·01 1·69 8·61 0·18 13·30 11·86 1·38 0·48 2·91 100·00 –0·17 0·701 F2a 47·39 0·64 15·21 1·71 8·23 0·22 11·24 13·41 1·38 0·58 3·13 100·00 –0·28 0·672 F3a 47·54 0·76 17·62 2·00 7·87 0·22 9·50 12·08 1·63 0·79 3·85 100·00 0·34 0·637 F4a 47·96 0·86 18·47 2·07 8·35 0·25 8·05 11·15 1·86 0·98 4·27 100·00 0·31 0·584 F5a 50·70 0·93 19·49 1·84 7·79 0·24 6·11 9·42 2·33 1·15 5·92 100·00 0·23 0·535 F5a 51·09 0·94 18·26 1·99 8·32 0·24 6·15 9·50 2·35 1·17 5·92 100·01 0·23 0·520 F6a 53·17 1·00 18·87 1·95 7·47 0·33 4·77 8·56 2·64 1·25 6·95 100·00 0·35 0·480 F7a 55·81 1·03 18·19 1·97 7·44 0·32 3·41 7·63 2·84 1·35 7·64 99·99 0·38 0·398 F8a 58·55 0·97 16·87 1·83 6·29 0·32 3·02 7·61 3·02 1·52 8·14 100·00 0·39 0·404 990 Mb 57·17 0·95 18·74 1·69 6·17 0·31 2·94 7·39 2·95 1·47 0·21 7·46 99·99 0·00 0·405 950 Mb 58·10 0·88 18·98 1·54 5·93 0·29 2·40 7·14 2·99 1·52 0·23 7·80 100·00 0·00 0·369 900 Mb 63·53 0·51 18·94 0·97 3·27 0·14 0·99 6·08 3·39 1·90 0·29 9·15 100·01 0·00 0·299 850 Mb 66·32 0·48 18·31 0·78 2·32 0·09 0·55 5·25 3·66 2·10 0·15 8·88 100·00 0·00 0·244 800 Mb 70·39 0·21 16·78 0·54 1·45 0·09 0·37 3·58 3·80 2·69 0·10 8·21 100·00 0·00 0·254 Seed crystals Plag (nom.) 44·40 35·83 19·20 0·54 0·03 100·00 0·952 Plag (EPMA) 44·70 0·04 36·12 0·64 0·10 19·66 0·60 0·01 101·85 0·948 Garnet (nom.) 37·82 0·05 19·34 37·12 1·55 2·80 0·99 99·67 0·119 Garnet (EPMA) 37·49 0·05 21·80 36·29 1·85 2·88 0·98 101·34 0·124 ΔNNO = log fO2 – log fO2(Ni–NiO) at run conditions (Table 2) calculated for the Fe2O3/FeO ratio of the starting material. xMg, Mg-number or Mg# = molar [MgO/(MgO + FeO)]; all Fe as Fe2+; xAn applies to plagioclase seed crystals and xAn = molar Ca/(Ca + Na + K). RC158c is a natural hydrous high-Mg basalt (Ulmer, 1988); 85-44 is a synthetic analogue of a high-Mg basaltic-andesite (Baker et al., 1994). Plag (nom.), nominal composition of plag crystals from Miyake-Jima volcano (Amma-Miyasaka & Nakagawa, 2002) used as seed crystals (1 wt %) in starting materials 950–800 Mb and ba. Plag (EPMA), average composition of cores of plag crystals in experiments containing plag seed crystals. Garnet (nom.), nominal composition of garnet crystals from Nijar (El Hoyazo volcanic complex, Spain; Munoz-Espadas et al., 2000) used as seed crystals (1 wt %) in starting 990–800 Mb and ba. Garnet (EPMA), average composition of cores of garnet crystals in experiments containing garnet seed crystals. Open in new tab Table 1: Composition of starting materials (in wt %) Starting material . SiO2 . TiO2 . Al2O3 . Cr2O3 . Fe2O3 . FeO . MnO . MgO . NiO . CaO . Na2O . K2O . P2O5 . H2O . Total . ΔNNO . xMg/xAn . FC ba AuPd 85-44 52·34 0·64 15·80 1·33 6·75 0·19 10·42 9·91 2·22 0·40 5·00 100·00 –0·12 0·700 F1b 52·07 0·65 15·67 1·47 7·51 0·15 9·60 10·16 2·32 0·41 5·12 100·00 –0·20 0·660 F2b 53·24 0·71 16·40 1·60 6·45 0·20 8·16 10·31 2·46 0·47 5·55 100·00 0·30 0·648 F3b 54·40 0·75 17·07 1·51 5·70 0·18 7·23 10·04 2·61 0·51 5·85 100·00 0·40 0·646 F4b 54·10 0·84 18·13 1·51 5·35 0·19 6·72 9·70 2·82 0·64 6·20 100·00 0·45 0·641 F5b 55·19 0·93 19·75 1·36 5·02 0·19 5·00 8·70 3·13 0·73 7·08 100·00 0·40 0·588 F6b 55·17 0·90 19·89 1·53 5·51 0·22 4·20 8·72 3·11 0·76 7·26 100·00 0·40 0·521 F7b 56·73 0·91 19·18 1·54 5·44 0·23 3·40 8·60 3·16 0·81 7·68 100·00 0·40 0·470 990 ba 55·71 0·90 20·26 1·54 5·35 0·23 3·38 8·46 3·19 0·80 0·18 8·50 100·00 0·00 0·472 950 ba 57·42 0·76 20·46 1·36 5·16 0·23 2·30 8·09 3·20 0·82 0·20 9·00 100·00 0·00 0·390 900 ba 62·19 0·52 19·79 1·03 3·48 0·13 1·06 6·79 3·70 1·05 0·27 9·50 100·01 0·00 0·300 850 ba 65·78 0·51 18·49 0·84 2·55 0·12 0·68 5·55 4·10 1·22 0·15 8·97 100·00 0·00 0·270 800 ba 70·73 0·19 16·83 0·57 1·50 0·11 0·40 3·71 4·20 1·61 0·16 8·24 100·01 0·00 0·262 EQ Mb Pt–C/EQ Mb AuPd RC158c 47·41 0·72 12·80 0·16 3·25 6·44 0·19 17·04 0·04 10·33 1·22 0·40 3·0 (3) 100·00 0·764 FC Mb Pt–C F1 48·59 0·91 15·59 0·00 9·04 0·15 11·01 12·68 1·53 0·51 3·17 100·01 0·685 F2 48·53 1·14 17·87 0·00 9·91 0·18 8·17 11·39 2·06 0·74 4·00 99·99 0·595 F3 48·44 1·19 18·65 0·00 10·24 0·18 7·57 10·74 2·18 0·80 4·32 99·99 0·568 F4 49·56 1·43 19·15 0·00 11·07 0·23 5·78 8·97 2·67 1·13 5·18 99·99 0·482 F5 49·94 2·18 17·78 0·00 14·31 0·33 3·47 6·90 3·29 1·79 10·65 99·99 0·302 FC Mb AuPd F1a 47·65 0·86 14·01 1·69 8·61 0·18 13·30 11·86 1·38 0·48 2·91 100·00 –0·17 0·701 F2a 47·39 0·64 15·21 1·71 8·23 0·22 11·24 13·41 1·38 0·58 3·13 100·00 –0·28 0·672 F3a 47·54 0·76 17·62 2·00 7·87 0·22 9·50 12·08 1·63 0·79 3·85 100·00 0·34 0·637 F4a 47·96 0·86 18·47 2·07 8·35 0·25 8·05 11·15 1·86 0·98 4·27 100·00 0·31 0·584 F5a 50·70 0·93 19·49 1·84 7·79 0·24 6·11 9·42 2·33 1·15 5·92 100·00 0·23 0·535 F5a 51·09 0·94 18·26 1·99 8·32 0·24 6·15 9·50 2·35 1·17 5·92 100·01 0·23 0·520 F6a 53·17 1·00 18·87 1·95 7·47 0·33 4·77 8·56 2·64 1·25 6·95 100·00 0·35 0·480 F7a 55·81 1·03 18·19 1·97 7·44 0·32 3·41 7·63 2·84 1·35 7·64 99·99 0·38 0·398 F8a 58·55 0·97 16·87 1·83 6·29 0·32 3·02 7·61 3·02 1·52 8·14 100·00 0·39 0·404 990 Mb 57·17 0·95 18·74 1·69 6·17 0·31 2·94 7·39 2·95 1·47 0·21 7·46 99·99 0·00 0·405 950 Mb 58·10 0·88 18·98 1·54 5·93 0·29 2·40 7·14 2·99 1·52 0·23 7·80 100·00 0·00 0·369 900 Mb 63·53 0·51 18·94 0·97 3·27 0·14 0·99 6·08 3·39 1·90 0·29 9·15 100·01 0·00 0·299 850 Mb 66·32 0·48 18·31 0·78 2·32 0·09 0·55 5·25 3·66 2·10 0·15 8·88 100·00 0·00 0·244 800 Mb 70·39 0·21 16·78 0·54 1·45 0·09 0·37 3·58 3·80 2·69 0·10 8·21 100·00 0·00 0·254 Seed crystals Plag (nom.) 44·40 35·83 19·20 0·54 0·03 100·00 0·952 Plag (EPMA) 44·70 0·04 36·12 0·64 0·10 19·66 0·60 0·01 101·85 0·948 Garnet (nom.) 37·82 0·05 19·34 37·12 1·55 2·80 0·99 99·67 0·119 Garnet (EPMA) 37·49 0·05 21·80 36·29 1·85 2·88 0·98 101·34 0·124 Starting material . SiO2 . TiO2 . Al2O3 . Cr2O3 . Fe2O3 . FeO . MnO . MgO . NiO . CaO . Na2O . K2O . P2O5 . H2O . Total . ΔNNO . xMg/xAn . FC ba AuPd 85-44 52·34 0·64 15·80 1·33 6·75 0·19 10·42 9·91 2·22 0·40 5·00 100·00 –0·12 0·700 F1b 52·07 0·65 15·67 1·47 7·51 0·15 9·60 10·16 2·32 0·41 5·12 100·00 –0·20 0·660 F2b 53·24 0·71 16·40 1·60 6·45 0·20 8·16 10·31 2·46 0·47 5·55 100·00 0·30 0·648 F3b 54·40 0·75 17·07 1·51 5·70 0·18 7·23 10·04 2·61 0·51 5·85 100·00 0·40 0·646 F4b 54·10 0·84 18·13 1·51 5·35 0·19 6·72 9·70 2·82 0·64 6·20 100·00 0·45 0·641 F5b 55·19 0·93 19·75 1·36 5·02 0·19 5·00 8·70 3·13 0·73 7·08 100·00 0·40 0·588 F6b 55·17 0·90 19·89 1·53 5·51 0·22 4·20 8·72 3·11 0·76 7·26 100·00 0·40 0·521 F7b 56·73 0·91 19·18 1·54 5·44 0·23 3·40 8·60 3·16 0·81 7·68 100·00 0·40 0·470 990 ba 55·71 0·90 20·26 1·54 5·35 0·23 3·38 8·46 3·19 0·80 0·18 8·50 100·00 0·00 0·472 950 ba 57·42 0·76 20·46 1·36 5·16 0·23 2·30 8·09 3·20 0·82 0·20 9·00 100·00 0·00 0·390 900 ba 62·19 0·52 19·79 1·03 3·48 0·13 1·06 6·79 3·70 1·05 0·27 9·50 100·01 0·00 0·300 850 ba 65·78 0·51 18·49 0·84 2·55 0·12 0·68 5·55 4·10 1·22 0·15 8·97 100·00 0·00 0·270 800 ba 70·73 0·19 16·83 0·57 1·50 0·11 0·40 3·71 4·20 1·61 0·16 8·24 100·01 0·00 0·262 EQ Mb Pt–C/EQ Mb AuPd RC158c 47·41 0·72 12·80 0·16 3·25 6·44 0·19 17·04 0·04 10·33 1·22 0·40 3·0 (3) 100·00 0·764 FC Mb Pt–C F1 48·59 0·91 15·59 0·00 9·04 0·15 11·01 12·68 1·53 0·51 3·17 100·01 0·685 F2 48·53 1·14 17·87 0·00 9·91 0·18 8·17 11·39 2·06 0·74 4·00 99·99 0·595 F3 48·44 1·19 18·65 0·00 10·24 0·18 7·57 10·74 2·18 0·80 4·32 99·99 0·568 F4 49·56 1·43 19·15 0·00 11·07 0·23 5·78 8·97 2·67 1·13 5·18 99·99 0·482 F5 49·94 2·18 17·78 0·00 14·31 0·33 3·47 6·90 3·29 1·79 10·65 99·99 0·302 FC Mb AuPd F1a 47·65 0·86 14·01 1·69 8·61 0·18 13·30 11·86 1·38 0·48 2·91 100·00 –0·17 0·701 F2a 47·39 0·64 15·21 1·71 8·23 0·22 11·24 13·41 1·38 0·58 3·13 100·00 –0·28 0·672 F3a 47·54 0·76 17·62 2·00 7·87 0·22 9·50 12·08 1·63 0·79 3·85 100·00 0·34 0·637 F4a 47·96 0·86 18·47 2·07 8·35 0·25 8·05 11·15 1·86 0·98 4·27 100·00 0·31 0·584 F5a 50·70 0·93 19·49 1·84 7·79 0·24 6·11 9·42 2·33 1·15 5·92 100·00 0·23 0·535 F5a 51·09 0·94 18·26 1·99 8·32 0·24 6·15 9·50 2·35 1·17 5·92 100·01 0·23 0·520 F6a 53·17 1·00 18·87 1·95 7·47 0·33 4·77 8·56 2·64 1·25 6·95 100·00 0·35 0·480 F7a 55·81 1·03 18·19 1·97 7·44 0·32 3·41 7·63 2·84 1·35 7·64 99·99 0·38 0·398 F8a 58·55 0·97 16·87 1·83 6·29 0·32 3·02 7·61 3·02 1·52 8·14 100·00 0·39 0·404 990 Mb 57·17 0·95 18·74 1·69 6·17 0·31 2·94 7·39 2·95 1·47 0·21 7·46 99·99 0·00 0·405 950 Mb 58·10 0·88 18·98 1·54 5·93 0·29 2·40 7·14 2·99 1·52 0·23 7·80 100·00 0·00 0·369 900 Mb 63·53 0·51 18·94 0·97 3·27 0·14 0·99 6·08 3·39 1·90 0·29 9·15 100·01 0·00 0·299 850 Mb 66·32 0·48 18·31 0·78 2·32 0·09 0·55 5·25 3·66 2·10 0·15 8·88 100·00 0·00 0·244 800 Mb 70·39 0·21 16·78 0·54 1·45 0·09 0·37 3·58 3·80 2·69 0·10 8·21 100·00 0·00 0·254 Seed crystals Plag (nom.) 44·40 35·83 19·20 0·54 0·03 100·00 0·952 Plag (EPMA) 44·70 0·04 36·12 0·64 0·10 19·66 0·60 0·01 101·85 0·948 Garnet (nom.) 37·82 0·05 19·34 37·12 1·55 2·80 0·99 99·67 0·119 Garnet (EPMA) 37·49 0·05 21·80 36·29 1·85 2·88 0·98 101·34 0·124 ΔNNO = log fO2 – log fO2(Ni–NiO) at run conditions (Table 2) calculated for the Fe2O3/FeO ratio of the starting material. xMg, Mg-number or Mg# = molar [MgO/(MgO + FeO)]; all Fe as Fe2+; xAn applies to plagioclase seed crystals and xAn = molar Ca/(Ca + Na + K). RC158c is a natural hydrous high-Mg basalt (Ulmer, 1988); 85-44 is a synthetic analogue of a high-Mg basaltic-andesite (Baker et al., 1994). Plag (nom.), nominal composition of plag crystals from Miyake-Jima volcano (Amma-Miyasaka & Nakagawa, 2002) used as seed crystals (1 wt %) in starting materials 950–800 Mb and ba. Plag (EPMA), average composition of cores of plag crystals in experiments containing plag seed crystals. Garnet (nom.), nominal composition of garnet crystals from Nijar (El Hoyazo volcanic complex, Spain; Munoz-Espadas et al., 2000) used as seed crystals (1 wt %) in starting 990–800 Mb and ba. Garnet (EPMA), average composition of cores of garnet crystals in experiments containing garnet seed crystals. Open in new tab Table 2: Experimental run conditions Run no. . T . Starting . Time . H2O(n) . H2O(R) . Log fO2 . fO2 . Capsule(s) . ΔFe . Run products . Phase proportions (wt %) . ∑R2 . F rel. . . (°C) . material . (h) . (wt %) . (wt %) . ΔNNO . ΔNNO . (inner/outer) . (%) . . . . (wt %) . Fractional crystallization 85-44 AuPd capsules (FC ba AuPd) rk48 1230 85-44 24·0 5·0 4·4 (3) –0·10 A5Pd5/A5Pd5 –6·9 liq 100 0·10 100·0 rk50 1200 85-44 24·0 5·1 n.a. –0·13 A9Pd1/A5Pd5 –6·0 liq, opx 97·8(6), 2·2(5) 0·11 97·8 (6) rk56 1170 F1b 29·0 5·6 n.a. –0·19 A9Pd1/A5Pd5 6·5 liq, cpx, opx 91·8(9), 3·5(12), 4·7(8) 0·12 89·8 (11) rk60 1140 F2b 51·0 5·8 n.a. 0·30 A9Pd1/A5Pd5 1·2 liq, cpx, opx 95·0(12), 3·4(15), 1·6(11) 0·22 85·3 (15) rk63 1110 F3b 72·0 6·2 n.a. 0·40 A9Pd1/A5Pd5 6·8 liq, cpx, opx 94·2(8), 4·9(11), 0·9(8) 0·14 80·3 (14) rk66 1080 F4b 100·0 7·1 n.a. 0·45 A9Pd1/A5Pd5 –3·2 liq, cpx, opx 87·4(5), 9·9(7), 2·7(5) 0·05 70·2 (9) rk69 1050 F5b 113·0 7·6 n.a. 0·40 A9Pd1/A5Pd5 –12·5 liq, cpx, amp 92·7(3), 1·2(4), 6·2(7) 0·10 65·0 (6) rk70 1020 F6b 126·0 7·8 n.a. 0·40 A9Pd1/A5Pd5 –2·0 liq, cpx, amp 92·7(11), 0·6(12), 6·7(21) 0·10 60·3 (11) rk71 990 F7b 150·0 8·6 n.a. 0·40 A9Pd1/A5Pd5 3·2 liq, cpx, amp 88·6(12), 1·5(15), 10·0(24) 0·20 53·4 (15) PU1049 990 990 ba 67·0 8·5 10·2 (4) 0·00 Au 1·9 liq, amp, mag 99·0 (8), 0·9 (6), 0·1 (1) 0·40 59·7 (14) PU1070 980 990 ba 67·5 8·5 9·7 (2) 0·00 Au/Au 11·0 liq, amp 98·2(10), 1·8(10) 0·14 59·2 (15) PU1062 950 990 ba 48·0 8·5 9·1 (3) 0·00 Au/Au 14·0 liq, amp, (gar), (plg) 90·7(20), 9·3(28), tr., tr. 0·51 54·7 (22) PU1064 900 950 ba 70·0 9·0 10·2 (2) 0·00 Au/Au 0·9 liq, amp, gar, plg, ilm 76·5(12), 10·4(28), 6·4(23), 6·4(11), 0·3(3) 0·08 41·8 (19) PU1066 850 900 ba 70·0 9·5 9·5 (3) 0·00 Au/Au 0·0 liq, gar, plg, apa, (ilm) 85·2(9), 6·7/10), 7·8(11), 0·3(3), tr. 0·21 35·6 (14) PU1068 800 850 ba 114·0 9·0 9·9 (3) 0·00 Au/Au 1·3 liq, amp, gar, plg, ilm, (apa) 75·5(20), 6·7 (33), 0·8(10), 16·6(20), 0·5(4), tr. 0·19 26·9 (21) PU1072 750 800 ba 144·0 13·2 11·6 (4) 0·00 Au/Au 0·8 liq, plg, mag, apa, (amp, gar) 89·2(7), 9·1(9), 1·3(2), 0·3(2), tr. 0·12 24·0 (19) PU1072o 720 800 ba 144·0 11·4 11·2 (3) 0·00 Au/Au 0·5 liq, amp, plg, mag, qtz, (apa, gar) 73·1(13), 2·2(4), 20·2(8), 1·2(1), 3·5(6), tr., tr. 0·02 19·7 (20) Fractional crystallization RC158c AuPd capsules (FC Mb AuPd) rk47 1230 RC158c 24·0 2·9 4·2 (2) 2·00 1·5 (3) A5Pd5/A5Pd5 –10 liq, ol, sp 89·8(9), 9·3(9), 0·9(4) 0·76 89·8 (9) rk51 1200 F1a 24·0 3·2 4·3 (2) –0·17 1·3 (2) A9Pd1/A5Pd5 5·0 liq, ol, cpx 92·0(15), 5·2(5), 2·8(15) 0·17 82·6 (17) rk52 1170 F2a 32·0 3·8 5·2 (2) –0·18 1·0 (3) A9Pd1/A5Pd5 7·6 liq, ol, cpx, (sp) 81·6(37), 2·5(10), 15·9(37), tr. 0·60 67·5 (39) rk54 1140 F3a 48·0 4·6 n.a. 0·34 A9Pd1/A5Pd5 –2·6 liq, cpx, sp 84·6(20), 13·8(18), 1·6(6) 0·24 57·1 (37) rk57 1110 F4a 71·0 5·5 n.a. 0·31 A9Pd1/A5Pd5 10·5 liq, cpx, opx, sp 76·7(4), 18·8(44), 0·7(2), 3·8(1) 1·18 43·8 (40) rk58 1080 F5a 100·0 6·7 n.a. 0·23 A9Pd1/A5Pd5 –1·0 liq, cpx, opx, sp 87·4(22), 9·5(25), 1·7(16), 1·4(5) 0·53 38·3 (39) rk73 1050 F6a 120·0 7·3 7·3 (3) 0·35 A9Pd1/A5Pd5 –3·2 liq, cpx, amp 94·9(22), 1·5(23), 3·6(41) 0·35 36·3 (30) rk 65 1020 F7a 150·0 8·1 n.a. 0·38 A9Pd1/A5Pd5 0·2 liq, cpx, amp, sp 94·2(14), 2·6(15), 1·5(30), 1·7(3) 0·12 34·2 (25) rk67 990 F8a 188·0 9·3 n.a. 0·39 Au 0·1 liq, cpx, amp, sp 85·4(9), 6·0(1), 8·3(2), 0·34(26) 0·10 29·2 (15) PU1048 990 990 Mb 67·0 8·2 8·6 (2) 0·00 Au/Au 21·0 liq, cpx 99·0(10), 1·0(8) 0·69 33·9 (17) PU1069 980 990 Mb 67·5 8·2 8·3 (2) 0·00 Au/Au 9·5 liq, cpx 99·4 (9), 0·6(6) 0·24 34·0 (14) PU1061 950 990 Mb 48·0 8·6 8·4 (2) 0·00 Au/Au 11·3 liq, amp, (gar), (cpx) 94·4(8), 5·6(11), tr., tr. 0·06 32·0 (13) PU1063 900 950 Mb 70·0 9·6 8·4 (3) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm 78·0(9), 12·0(2), 9·3(2), 1·5(9), 0·3(3) 0·09 24·9 (12) PU1065 850 900 Mb 70·0 10·1 9·3 (3) 0·00 Au/Au Pt –0·4 liq, gar, plg, apa, (ilm) 90·2(10), 6·0(11), 3·4(11), 0·4(3), tr. 0·24 22·5 (13) PU1067 800 850 Mb 114·0 10·8 9·8 (5) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm, (apa) 80·9(15), 5·1(33), 0·8(10), 12·8(15), 0·4(3), tr. 0·14 18·2 (17) PU1071 750 800 Mb 144·0 12·2 10·9 (2) 0·00 Au/Au 0·7 liq, plg, mag, (gar) 89·9(10), 8·7(13), 1·4(3), tr. 0·22 16·4 (17) Equilibrium crystallization RC158c AuPd capsules (EQ Mb AuPd) PU910 1200 RC158c 3·8 3·6 4·2 (2) 2·1 (2) A9Pd1/Pt –6·0 liq, ol, sp 91·9(9), 7·3(7), 0·9(8) 0·50 PU908 1200 RC158c 6·3 3·7 4·0 (2) 1·3 (2) A9Pd1/Pt –8·6 liq, ol, sp 90·1 (7), 9·4(5), 0·6(5) 0·30 PU926 1160 RC158c 8·0 3·9 4·1 (2) 2·2 (1) A9Pd1/Pt –2·6 liq, ol, sp 83·6(5), 16·4(4), 0·1(5) 0·14 PU899 1160 RC158c 20·0 3·9 n.p. 1·3 (2) A9Pd1/Pt –1·7 liq, ol, (cpx) 84·5(9), 15·5(3), tr. 0·08 PU909 1120 RC158c 41·0 4·0 4·4 (2) 2·9 (1) A9Pd1/Pt –8·8 liq, ol, cpx, sp 82·0(6), 16·1(2), 1·1(6), 0·8(4) 0·03 PU1006 1100 RC158c 6·2 5·1 4·3 (3) 4·1 (4) A9Pd1 –4·3 liq, ol, cpx, (sp) 64·3(12), 24·4(6), 11·4(15), tr. 0·03 PU1003 1080 RC158c 10·0 5·2 5·8 (3) 4·4 (3) A9Pd1 –2·2 liq, ol, cpx, sp 63·4(9), 25·8(4), 10·4(11), 0·3(5) 0·08 PU1004 1060 RC158c 20·6 6·4 6·0 (4) 5·1 (5) A9Pd1 2·0 liq, ol, cpx, sp 50·5(15), 25·6(8), 22·6(19), 1·3(7) 0·18 PU905 1040 RC158c 65·0 9·5 9·0 (6) 5·3 (2) A9Pd1/Pt 8·2 liq, ol, cpx, opx, amp, sp 23·5(53), 1·9(30), 14·6(17), 22·1(53), 37·6(64), 0·3(2) 0·02 PU1005 1000 RC158c 46·3 10·1 8·8 (4) 5·5 (6) A9Pd1 8·0 liq, ol, cpx, opx, amp, (sp) 20·9(104), 8·4(113), 12·1(22), 11·4(130), 47·2(98), tr. 0·02 PU906 980 RC158c 67·0 11·5 9·3 (5) Au/Pt 0·2 liq, cpx, opx, amp, sp 15·2(7), 11·3(10), 21·8 (8), 50·6(22), 1·1(2) 0·06 Fractional crystallization RC158c graphite–Pt capsules (FC Mb Pt–C) rk3 1230 Rc158c 23·0 3·1 3·1 (1) C–COH C/Pt –2·5 liq, ol, sp 82·9(5), 17·1(4), 0·01(1) 0·12 82·9 (5) rk6 1200 F1 87·0 4·1 2·6 (2) C–COH C/Pt –3·6 liq, ol, cpx 78·6(12), 3·2(5), 18·2(13) 0·11 65·2 (13) rk11 1170 F2 72·0 4·4 2·6 (2) C–COH C/Pt 5·8 liq, cpx 91·5(4), 8·5(4) 0·02 59·6 (12) rk13 1140 F3 60·0 5·8 n.a. C–COH C/Pt –0·2 liq, cpx, opx, sp, plg 74·9(23), 19·0(17), 0·1(10), 2·2(6), 3·9(23) 0·01 44·7 (23) rk55 1110 F4 85·0 9·8 2·5 (2) C–COH C/Pt –7·7 liq, cpx, opx, plg, (sp) 52·7(16), 7·3(11), 14·7(11), 25·4(10), tr. 0·04 23·5 (20) rk64 1080 F5 120·0 17·6 2·0 (3) C–COH C/Pt –3·6 liq, cpx, plg, ilm, gar 60·5(17), 8·1(8), 11·0(17), 1·2(2), 19·2(14) 0·02 14·2 (20) Equilibrium crystallization RC158c graphite–Pt capsules (EQ Mb Pt–C) PU72 1350 RC158c 4·0 2·8 2·9 (1) C–COH C/Pt 11·3 liq 100 0·31 P1330 1330 RC158c 3·5 2·9 3·2 (1) C–COH C/Pt 1·4 liq, ol 98·1(4), 1·9(3) 0·08 P1301 1300 RC158c 4·3 3·0 3·7(1) C–COH C/Pt 2·5 liq, ol 92·2(1), 7·8(1) 0·02 P1302 1300 RC158c 7·5 2·9 3·4(1) C–COH C/Pt 3·0 liq, ol 95·5(4), 4·5(3) 0·10 P1270 1270 RC158c 10·5 3·0 3·6 (1) C–COH C/Pt 1·0 liq, ol 92·7(1), 7·3(1) 0·02 P1240 1240 RC158c 16·0 3·2 n.a. C–COH C/Pt 0·4 liq, ol 88·8(2), 11·2(2) 0·03 P1210 1210 RC158c 21·7 3·4 3·7(1) C–COH C/Pt 0·9 liq, ol, (sp) 83·1(2), 16·9(1), tr 0·03 P1180 1180 RC158c 25·9 3·7 3·7 (1) C–COH C/Pt 0·6 liq, ol, cpx, (sp) 75·0(4), 19·9(2), 5·1(4), tr 0·02 P1151 1150 RC158c 30·0 3·8 4·4 (2) C–COH C/Pt 3·1 liq, ol, cpx, sp 73·1(7), 19·0(2), 7·7(7), 0·2(2) 0·03 P1152 1150 RC158c 28·8 4·4 n.a. C–COH C/Pt 9·5 liq, ol, cpx, sp 63·6(9), 20·5(3), 15·7(10), 0·2(2) 0·06 P1121 1120 RC158c 50·8 4·6 n.a. C–COH C/Pt 12·7 liq, ol, cpx, opx, (sp) 60·6(37), 18·6(33), 18·7(24), 2·2(20), tr 0·39 P1122 1120 RC158c 99·0 4·7 3·6 (1) C–COH C/Pt 20·3 liq, ol, cpx, opx, (sp) 59·8(5), 14·0(4), 17·2(3), 9·0(7), tr. 0·01 P1090 1090 RC158c 75·0 5·7 4·0 (3) C–COH C/Pt –2·0 liq, ol, cpx, opx, sp 49·4(13), 17·8(11), 25·2(6), 5·4(17), 2·2(1) 0·01 P1060 1060 RC158c 98·0 10·9 6·7 (2) C–COH C/Pt 4·1 liq, ol, cpx, opx, amp, sp 24·9(20), 5·6(17), 38·7(12), 23·0(19), 3·1(31), 4·6(2) 0·01 P1000 1000 RC158c 211·3 13·6 n.a. C–COH C/Pt 5·6 liq, cpx, opx, amp, sp 15·9(2), 27·1(3), 29·1(2), 25·5(8), 2·3(1) 0·01 Run no. . T . Starting . Time . H2O(n) . H2O(R) . Log fO2 . fO2 . Capsule(s) . ΔFe . Run products . Phase proportions (wt %) . ∑R2 . F rel. . . (°C) . material . (h) . (wt %) . (wt %) . ΔNNO . ΔNNO . (inner/outer) . (%) . . . . (wt %) . Fractional crystallization 85-44 AuPd capsules (FC ba AuPd) rk48 1230 85-44 24·0 5·0 4·4 (3) –0·10 A5Pd5/A5Pd5 –6·9 liq 100 0·10 100·0 rk50 1200 85-44 24·0 5·1 n.a. –0·13 A9Pd1/A5Pd5 –6·0 liq, opx 97·8(6), 2·2(5) 0·11 97·8 (6) rk56 1170 F1b 29·0 5·6 n.a. –0·19 A9Pd1/A5Pd5 6·5 liq, cpx, opx 91·8(9), 3·5(12), 4·7(8) 0·12 89·8 (11) rk60 1140 F2b 51·0 5·8 n.a. 0·30 A9Pd1/A5Pd5 1·2 liq, cpx, opx 95·0(12), 3·4(15), 1·6(11) 0·22 85·3 (15) rk63 1110 F3b 72·0 6·2 n.a. 0·40 A9Pd1/A5Pd5 6·8 liq, cpx, opx 94·2(8), 4·9(11), 0·9(8) 0·14 80·3 (14) rk66 1080 F4b 100·0 7·1 n.a. 0·45 A9Pd1/A5Pd5 –3·2 liq, cpx, opx 87·4(5), 9·9(7), 2·7(5) 0·05 70·2 (9) rk69 1050 F5b 113·0 7·6 n.a. 0·40 A9Pd1/A5Pd5 –12·5 liq, cpx, amp 92·7(3), 1·2(4), 6·2(7) 0·10 65·0 (6) rk70 1020 F6b 126·0 7·8 n.a. 0·40 A9Pd1/A5Pd5 –2·0 liq, cpx, amp 92·7(11), 0·6(12), 6·7(21) 0·10 60·3 (11) rk71 990 F7b 150·0 8·6 n.a. 0·40 A9Pd1/A5Pd5 3·2 liq, cpx, amp 88·6(12), 1·5(15), 10·0(24) 0·20 53·4 (15) PU1049 990 990 ba 67·0 8·5 10·2 (4) 0·00 Au 1·9 liq, amp, mag 99·0 (8), 0·9 (6), 0·1 (1) 0·40 59·7 (14) PU1070 980 990 ba 67·5 8·5 9·7 (2) 0·00 Au/Au 11·0 liq, amp 98·2(10), 1·8(10) 0·14 59·2 (15) PU1062 950 990 ba 48·0 8·5 9·1 (3) 0·00 Au/Au 14·0 liq, amp, (gar), (plg) 90·7(20), 9·3(28), tr., tr. 0·51 54·7 (22) PU1064 900 950 ba 70·0 9·0 10·2 (2) 0·00 Au/Au 0·9 liq, amp, gar, plg, ilm 76·5(12), 10·4(28), 6·4(23), 6·4(11), 0·3(3) 0·08 41·8 (19) PU1066 850 900 ba 70·0 9·5 9·5 (3) 0·00 Au/Au 0·0 liq, gar, plg, apa, (ilm) 85·2(9), 6·7/10), 7·8(11), 0·3(3), tr. 0·21 35·6 (14) PU1068 800 850 ba 114·0 9·0 9·9 (3) 0·00 Au/Au 1·3 liq, amp, gar, plg, ilm, (apa) 75·5(20), 6·7 (33), 0·8(10), 16·6(20), 0·5(4), tr. 0·19 26·9 (21) PU1072 750 800 ba 144·0 13·2 11·6 (4) 0·00 Au/Au 0·8 liq, plg, mag, apa, (amp, gar) 89·2(7), 9·1(9), 1·3(2), 0·3(2), tr. 0·12 24·0 (19) PU1072o 720 800 ba 144·0 11·4 11·2 (3) 0·00 Au/Au 0·5 liq, amp, plg, mag, qtz, (apa, gar) 73·1(13), 2·2(4), 20·2(8), 1·2(1), 3·5(6), tr., tr. 0·02 19·7 (20) Fractional crystallization RC158c AuPd capsules (FC Mb AuPd) rk47 1230 RC158c 24·0 2·9 4·2 (2) 2·00 1·5 (3) A5Pd5/A5Pd5 –10 liq, ol, sp 89·8(9), 9·3(9), 0·9(4) 0·76 89·8 (9) rk51 1200 F1a 24·0 3·2 4·3 (2) –0·17 1·3 (2) A9Pd1/A5Pd5 5·0 liq, ol, cpx 92·0(15), 5·2(5), 2·8(15) 0·17 82·6 (17) rk52 1170 F2a 32·0 3·8 5·2 (2) –0·18 1·0 (3) A9Pd1/A5Pd5 7·6 liq, ol, cpx, (sp) 81·6(37), 2·5(10), 15·9(37), tr. 0·60 67·5 (39) rk54 1140 F3a 48·0 4·6 n.a. 0·34 A9Pd1/A5Pd5 –2·6 liq, cpx, sp 84·6(20), 13·8(18), 1·6(6) 0·24 57·1 (37) rk57 1110 F4a 71·0 5·5 n.a. 0·31 A9Pd1/A5Pd5 10·5 liq, cpx, opx, sp 76·7(4), 18·8(44), 0·7(2), 3·8(1) 1·18 43·8 (40) rk58 1080 F5a 100·0 6·7 n.a. 0·23 A9Pd1/A5Pd5 –1·0 liq, cpx, opx, sp 87·4(22), 9·5(25), 1·7(16), 1·4(5) 0·53 38·3 (39) rk73 1050 F6a 120·0 7·3 7·3 (3) 0·35 A9Pd1/A5Pd5 –3·2 liq, cpx, amp 94·9(22), 1·5(23), 3·6(41) 0·35 36·3 (30) rk 65 1020 F7a 150·0 8·1 n.a. 0·38 A9Pd1/A5Pd5 0·2 liq, cpx, amp, sp 94·2(14), 2·6(15), 1·5(30), 1·7(3) 0·12 34·2 (25) rk67 990 F8a 188·0 9·3 n.a. 0·39 Au 0·1 liq, cpx, amp, sp 85·4(9), 6·0(1), 8·3(2), 0·34(26) 0·10 29·2 (15) PU1048 990 990 Mb 67·0 8·2 8·6 (2) 0·00 Au/Au 21·0 liq, cpx 99·0(10), 1·0(8) 0·69 33·9 (17) PU1069 980 990 Mb 67·5 8·2 8·3 (2) 0·00 Au/Au 9·5 liq, cpx 99·4 (9), 0·6(6) 0·24 34·0 (14) PU1061 950 990 Mb 48·0 8·6 8·4 (2) 0·00 Au/Au 11·3 liq, amp, (gar), (cpx) 94·4(8), 5·6(11), tr., tr. 0·06 32·0 (13) PU1063 900 950 Mb 70·0 9·6 8·4 (3) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm 78·0(9), 12·0(2), 9·3(2), 1·5(9), 0·3(3) 0·09 24·9 (12) PU1065 850 900 Mb 70·0 10·1 9·3 (3) 0·00 Au/Au Pt –0·4 liq, gar, plg, apa, (ilm) 90·2(10), 6·0(11), 3·4(11), 0·4(3), tr. 0·24 22·5 (13) PU1067 800 850 Mb 114·0 10·8 9·8 (5) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm, (apa) 80·9(15), 5·1(33), 0·8(10), 12·8(15), 0·4(3), tr. 0·14 18·2 (17) PU1071 750 800 Mb 144·0 12·2 10·9 (2) 0·00 Au/Au 0·7 liq, plg, mag, (gar) 89·9(10), 8·7(13), 1·4(3), tr. 0·22 16·4 (17) Equilibrium crystallization RC158c AuPd capsules (EQ Mb AuPd) PU910 1200 RC158c 3·8 3·6 4·2 (2) 2·1 (2) A9Pd1/Pt –6·0 liq, ol, sp 91·9(9), 7·3(7), 0·9(8) 0·50 PU908 1200 RC158c 6·3 3·7 4·0 (2) 1·3 (2) A9Pd1/Pt –8·6 liq, ol, sp 90·1 (7), 9·4(5), 0·6(5) 0·30 PU926 1160 RC158c 8·0 3·9 4·1 (2) 2·2 (1) A9Pd1/Pt –2·6 liq, ol, sp 83·6(5), 16·4(4), 0·1(5) 0·14 PU899 1160 RC158c 20·0 3·9 n.p. 1·3 (2) A9Pd1/Pt –1·7 liq, ol, (cpx) 84·5(9), 15·5(3), tr. 0·08 PU909 1120 RC158c 41·0 4·0 4·4 (2) 2·9 (1) A9Pd1/Pt –8·8 liq, ol, cpx, sp 82·0(6), 16·1(2), 1·1(6), 0·8(4) 0·03 PU1006 1100 RC158c 6·2 5·1 4·3 (3) 4·1 (4) A9Pd1 –4·3 liq, ol, cpx, (sp) 64·3(12), 24·4(6), 11·4(15), tr. 0·03 PU1003 1080 RC158c 10·0 5·2 5·8 (3) 4·4 (3) A9Pd1 –2·2 liq, ol, cpx, sp 63·4(9), 25·8(4), 10·4(11), 0·3(5) 0·08 PU1004 1060 RC158c 20·6 6·4 6·0 (4) 5·1 (5) A9Pd1 2·0 liq, ol, cpx, sp 50·5(15), 25·6(8), 22·6(19), 1·3(7) 0·18 PU905 1040 RC158c 65·0 9·5 9·0 (6) 5·3 (2) A9Pd1/Pt 8·2 liq, ol, cpx, opx, amp, sp 23·5(53), 1·9(30), 14·6(17), 22·1(53), 37·6(64), 0·3(2) 0·02 PU1005 1000 RC158c 46·3 10·1 8·8 (4) 5·5 (6) A9Pd1 8·0 liq, ol, cpx, opx, amp, (sp) 20·9(104), 8·4(113), 12·1(22), 11·4(130), 47·2(98), tr. 0·02 PU906 980 RC158c 67·0 11·5 9·3 (5) Au/Pt 0·2 liq, cpx, opx, amp, sp 15·2(7), 11·3(10), 21·8 (8), 50·6(22), 1·1(2) 0·06 Fractional crystallization RC158c graphite–Pt capsules (FC Mb Pt–C) rk3 1230 Rc158c 23·0 3·1 3·1 (1) C–COH C/Pt –2·5 liq, ol, sp 82·9(5), 17·1(4), 0·01(1) 0·12 82·9 (5) rk6 1200 F1 87·0 4·1 2·6 (2) C–COH C/Pt –3·6 liq, ol, cpx 78·6(12), 3·2(5), 18·2(13) 0·11 65·2 (13) rk11 1170 F2 72·0 4·4 2·6 (2) C–COH C/Pt 5·8 liq, cpx 91·5(4), 8·5(4) 0·02 59·6 (12) rk13 1140 F3 60·0 5·8 n.a. C–COH C/Pt –0·2 liq, cpx, opx, sp, plg 74·9(23), 19·0(17), 0·1(10), 2·2(6), 3·9(23) 0·01 44·7 (23) rk55 1110 F4 85·0 9·8 2·5 (2) C–COH C/Pt –7·7 liq, cpx, opx, plg, (sp) 52·7(16), 7·3(11), 14·7(11), 25·4(10), tr. 0·04 23·5 (20) rk64 1080 F5 120·0 17·6 2·0 (3) C–COH C/Pt –3·6 liq, cpx, plg, ilm, gar 60·5(17), 8·1(8), 11·0(17), 1·2(2), 19·2(14) 0·02 14·2 (20) Equilibrium crystallization RC158c graphite–Pt capsules (EQ Mb Pt–C) PU72 1350 RC158c 4·0 2·8 2·9 (1) C–COH C/Pt 11·3 liq 100 0·31 P1330 1330 RC158c 3·5 2·9 3·2 (1) C–COH C/Pt 1·4 liq, ol 98·1(4), 1·9(3) 0·08 P1301 1300 RC158c 4·3 3·0 3·7(1) C–COH C/Pt 2·5 liq, ol 92·2(1), 7·8(1) 0·02 P1302 1300 RC158c 7·5 2·9 3·4(1) C–COH C/Pt 3·0 liq, ol 95·5(4), 4·5(3) 0·10 P1270 1270 RC158c 10·5 3·0 3·6 (1) C–COH C/Pt 1·0 liq, ol 92·7(1), 7·3(1) 0·02 P1240 1240 RC158c 16·0 3·2 n.a. C–COH C/Pt 0·4 liq, ol 88·8(2), 11·2(2) 0·03 P1210 1210 RC158c 21·7 3·4 3·7(1) C–COH C/Pt 0·9 liq, ol, (sp) 83·1(2), 16·9(1), tr 0·03 P1180 1180 RC158c 25·9 3·7 3·7 (1) C–COH C/Pt 0·6 liq, ol, cpx, (sp) 75·0(4), 19·9(2), 5·1(4), tr 0·02 P1151 1150 RC158c 30·0 3·8 4·4 (2) C–COH C/Pt 3·1 liq, ol, cpx, sp 73·1(7), 19·0(2), 7·7(7), 0·2(2) 0·03 P1152 1150 RC158c 28·8 4·4 n.a. C–COH C/Pt 9·5 liq, ol, cpx, sp 63·6(9), 20·5(3), 15·7(10), 0·2(2) 0·06 P1121 1120 RC158c 50·8 4·6 n.a. C–COH C/Pt 12·7 liq, ol, cpx, opx, (sp) 60·6(37), 18·6(33), 18·7(24), 2·2(20), tr 0·39 P1122 1120 RC158c 99·0 4·7 3·6 (1) C–COH C/Pt 20·3 liq, ol, cpx, opx, (sp) 59·8(5), 14·0(4), 17·2(3), 9·0(7), tr. 0·01 P1090 1090 RC158c 75·0 5·7 4·0 (3) C–COH C/Pt –2·0 liq, ol, cpx, opx, sp 49·4(13), 17·8(11), 25·2(6), 5·4(17), 2·2(1) 0·01 P1060 1060 RC158c 98·0 10·9 6·7 (2) C–COH C/Pt 4·1 liq, ol, cpx, opx, amp, sp 24·9(20), 5·6(17), 38·7(12), 23·0(19), 3·1(31), 4·6(2) 0·01 P1000 1000 RC158c 211·3 13·6 n.a. C–COH C/Pt 5·6 liq, cpx, opx, amp, sp 15·9(2), 27·1(3), 29·1(2), 25·5(8), 2·3(1) 0·01 Starting materials are given in Table 1. H2O(n) wt %, H2O contents in the starting material corrected for melt fraction (f) and modal amount of amp (2·1 wt %); H2O(R) wt %, H2O content of glasses (liquid) determined by Raman spectroscopy; n.a., experiment charge is no longer available (lost); log fO2 ΔNNO indicates graphite capsules ≈C–COH (see text), for double-capsule experiments = initial fO2 constrained by Fe3+/Fetot of the starting material expressed as difference from the Ni–NiO equilibrium at the same pressure and temperature; fO2 (ol) (ΔNNO) corresponds to the fO2 calculated from olivine–liquid Fe2+–Mg partitioning (see text). Capsule(s): C/Pt, graphite container sealed in Pt; A9Pd1/Pt or A5Pd5, Fe-pre-saturated inner Au50Pd50 or Au90Pd10 capsule (not for Au100) in outer capsule (Pt or Au50Pd50) containing same starting material. ΔFe % is the difference between the FeO content of the bulk starting composition and the FeO obtained by mass-balance calculations; negative values indicate relative iron gain in wt %. Phase proportions are given in wt % and were calculated by least-squares regressions; numbers in parentheses indicate standard deviations and read as follows: 98·1(4) = 98·1 ± 0·4; ∑R2 is the sum of the squared residuals; F rel. (wt %) is the amount of liquid left in the fractional crystallization experiments relative to the initial starting material (RC158c and 85-44 for Mb and ba runs respectively). Phases: liq, liquid (glass); ol, olivine; cpx, high-Ca clinopyroxene; opx, low-Ca orthopyroxene; sp, Cr–Al–Fe-spinel; ilm, ilmenite; amp, amphibole; gar, garnet, plg, plagioclase, mag, (Ti-)magnetite, apa, apatite, qtz, quartz; phases in parentheses indicate that the phase is present only in trace amounts not quantified during least-squares regression (tr.). All runs were performed at a pressure of 1·0 GPa. Open in new tab Table 2: Experimental run conditions Run no. . T . Starting . Time . H2O(n) . H2O(R) . Log fO2 . fO2 . Capsule(s) . ΔFe . Run products . Phase proportions (wt %) . ∑R2 . F rel. . . (°C) . material . (h) . (wt %) . (wt %) . ΔNNO . ΔNNO . (inner/outer) . (%) . . . . (wt %) . Fractional crystallization 85-44 AuPd capsules (FC ba AuPd) rk48 1230 85-44 24·0 5·0 4·4 (3) –0·10 A5Pd5/A5Pd5 –6·9 liq 100 0·10 100·0 rk50 1200 85-44 24·0 5·1 n.a. –0·13 A9Pd1/A5Pd5 –6·0 liq, opx 97·8(6), 2·2(5) 0·11 97·8 (6) rk56 1170 F1b 29·0 5·6 n.a. –0·19 A9Pd1/A5Pd5 6·5 liq, cpx, opx 91·8(9), 3·5(12), 4·7(8) 0·12 89·8 (11) rk60 1140 F2b 51·0 5·8 n.a. 0·30 A9Pd1/A5Pd5 1·2 liq, cpx, opx 95·0(12), 3·4(15), 1·6(11) 0·22 85·3 (15) rk63 1110 F3b 72·0 6·2 n.a. 0·40 A9Pd1/A5Pd5 6·8 liq, cpx, opx 94·2(8), 4·9(11), 0·9(8) 0·14 80·3 (14) rk66 1080 F4b 100·0 7·1 n.a. 0·45 A9Pd1/A5Pd5 –3·2 liq, cpx, opx 87·4(5), 9·9(7), 2·7(5) 0·05 70·2 (9) rk69 1050 F5b 113·0 7·6 n.a. 0·40 A9Pd1/A5Pd5 –12·5 liq, cpx, amp 92·7(3), 1·2(4), 6·2(7) 0·10 65·0 (6) rk70 1020 F6b 126·0 7·8 n.a. 0·40 A9Pd1/A5Pd5 –2·0 liq, cpx, amp 92·7(11), 0·6(12), 6·7(21) 0·10 60·3 (11) rk71 990 F7b 150·0 8·6 n.a. 0·40 A9Pd1/A5Pd5 3·2 liq, cpx, amp 88·6(12), 1·5(15), 10·0(24) 0·20 53·4 (15) PU1049 990 990 ba 67·0 8·5 10·2 (4) 0·00 Au 1·9 liq, amp, mag 99·0 (8), 0·9 (6), 0·1 (1) 0·40 59·7 (14) PU1070 980 990 ba 67·5 8·5 9·7 (2) 0·00 Au/Au 11·0 liq, amp 98·2(10), 1·8(10) 0·14 59·2 (15) PU1062 950 990 ba 48·0 8·5 9·1 (3) 0·00 Au/Au 14·0 liq, amp, (gar), (plg) 90·7(20), 9·3(28), tr., tr. 0·51 54·7 (22) PU1064 900 950 ba 70·0 9·0 10·2 (2) 0·00 Au/Au 0·9 liq, amp, gar, plg, ilm 76·5(12), 10·4(28), 6·4(23), 6·4(11), 0·3(3) 0·08 41·8 (19) PU1066 850 900 ba 70·0 9·5 9·5 (3) 0·00 Au/Au 0·0 liq, gar, plg, apa, (ilm) 85·2(9), 6·7/10), 7·8(11), 0·3(3), tr. 0·21 35·6 (14) PU1068 800 850 ba 114·0 9·0 9·9 (3) 0·00 Au/Au 1·3 liq, amp, gar, plg, ilm, (apa) 75·5(20), 6·7 (33), 0·8(10), 16·6(20), 0·5(4), tr. 0·19 26·9 (21) PU1072 750 800 ba 144·0 13·2 11·6 (4) 0·00 Au/Au 0·8 liq, plg, mag, apa, (amp, gar) 89·2(7), 9·1(9), 1·3(2), 0·3(2), tr. 0·12 24·0 (19) PU1072o 720 800 ba 144·0 11·4 11·2 (3) 0·00 Au/Au 0·5 liq, amp, plg, mag, qtz, (apa, gar) 73·1(13), 2·2(4), 20·2(8), 1·2(1), 3·5(6), tr., tr. 0·02 19·7 (20) Fractional crystallization RC158c AuPd capsules (FC Mb AuPd) rk47 1230 RC158c 24·0 2·9 4·2 (2) 2·00 1·5 (3) A5Pd5/A5Pd5 –10 liq, ol, sp 89·8(9), 9·3(9), 0·9(4) 0·76 89·8 (9) rk51 1200 F1a 24·0 3·2 4·3 (2) –0·17 1·3 (2) A9Pd1/A5Pd5 5·0 liq, ol, cpx 92·0(15), 5·2(5), 2·8(15) 0·17 82·6 (17) rk52 1170 F2a 32·0 3·8 5·2 (2) –0·18 1·0 (3) A9Pd1/A5Pd5 7·6 liq, ol, cpx, (sp) 81·6(37), 2·5(10), 15·9(37), tr. 0·60 67·5 (39) rk54 1140 F3a 48·0 4·6 n.a. 0·34 A9Pd1/A5Pd5 –2·6 liq, cpx, sp 84·6(20), 13·8(18), 1·6(6) 0·24 57·1 (37) rk57 1110 F4a 71·0 5·5 n.a. 0·31 A9Pd1/A5Pd5 10·5 liq, cpx, opx, sp 76·7(4), 18·8(44), 0·7(2), 3·8(1) 1·18 43·8 (40) rk58 1080 F5a 100·0 6·7 n.a. 0·23 A9Pd1/A5Pd5 –1·0 liq, cpx, opx, sp 87·4(22), 9·5(25), 1·7(16), 1·4(5) 0·53 38·3 (39) rk73 1050 F6a 120·0 7·3 7·3 (3) 0·35 A9Pd1/A5Pd5 –3·2 liq, cpx, amp 94·9(22), 1·5(23), 3·6(41) 0·35 36·3 (30) rk 65 1020 F7a 150·0 8·1 n.a. 0·38 A9Pd1/A5Pd5 0·2 liq, cpx, amp, sp 94·2(14), 2·6(15), 1·5(30), 1·7(3) 0·12 34·2 (25) rk67 990 F8a 188·0 9·3 n.a. 0·39 Au 0·1 liq, cpx, amp, sp 85·4(9), 6·0(1), 8·3(2), 0·34(26) 0·10 29·2 (15) PU1048 990 990 Mb 67·0 8·2 8·6 (2) 0·00 Au/Au 21·0 liq, cpx 99·0(10), 1·0(8) 0·69 33·9 (17) PU1069 980 990 Mb 67·5 8·2 8·3 (2) 0·00 Au/Au 9·5 liq, cpx 99·4 (9), 0·6(6) 0·24 34·0 (14) PU1061 950 990 Mb 48·0 8·6 8·4 (2) 0·00 Au/Au 11·3 liq, amp, (gar), (cpx) 94·4(8), 5·6(11), tr., tr. 0·06 32·0 (13) PU1063 900 950 Mb 70·0 9·6 8·4 (3) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm 78·0(9), 12·0(2), 9·3(2), 1·5(9), 0·3(3) 0·09 24·9 (12) PU1065 850 900 Mb 70·0 10·1 9·3 (3) 0·00 Au/Au Pt –0·4 liq, gar, plg, apa, (ilm) 90·2(10), 6·0(11), 3·4(11), 0·4(3), tr. 0·24 22·5 (13) PU1067 800 850 Mb 114·0 10·8 9·8 (5) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm, (apa) 80·9(15), 5·1(33), 0·8(10), 12·8(15), 0·4(3), tr. 0·14 18·2 (17) PU1071 750 800 Mb 144·0 12·2 10·9 (2) 0·00 Au/Au 0·7 liq, plg, mag, (gar) 89·9(10), 8·7(13), 1·4(3), tr. 0·22 16·4 (17) Equilibrium crystallization RC158c AuPd capsules (EQ Mb AuPd) PU910 1200 RC158c 3·8 3·6 4·2 (2) 2·1 (2) A9Pd1/Pt –6·0 liq, ol, sp 91·9(9), 7·3(7), 0·9(8) 0·50 PU908 1200 RC158c 6·3 3·7 4·0 (2) 1·3 (2) A9Pd1/Pt –8·6 liq, ol, sp 90·1 (7), 9·4(5), 0·6(5) 0·30 PU926 1160 RC158c 8·0 3·9 4·1 (2) 2·2 (1) A9Pd1/Pt –2·6 liq, ol, sp 83·6(5), 16·4(4), 0·1(5) 0·14 PU899 1160 RC158c 20·0 3·9 n.p. 1·3 (2) A9Pd1/Pt –1·7 liq, ol, (cpx) 84·5(9), 15·5(3), tr. 0·08 PU909 1120 RC158c 41·0 4·0 4·4 (2) 2·9 (1) A9Pd1/Pt –8·8 liq, ol, cpx, sp 82·0(6), 16·1(2), 1·1(6), 0·8(4) 0·03 PU1006 1100 RC158c 6·2 5·1 4·3 (3) 4·1 (4) A9Pd1 –4·3 liq, ol, cpx, (sp) 64·3(12), 24·4(6), 11·4(15), tr. 0·03 PU1003 1080 RC158c 10·0 5·2 5·8 (3) 4·4 (3) A9Pd1 –2·2 liq, ol, cpx, sp 63·4(9), 25·8(4), 10·4(11), 0·3(5) 0·08 PU1004 1060 RC158c 20·6 6·4 6·0 (4) 5·1 (5) A9Pd1 2·0 liq, ol, cpx, sp 50·5(15), 25·6(8), 22·6(19), 1·3(7) 0·18 PU905 1040 RC158c 65·0 9·5 9·0 (6) 5·3 (2) A9Pd1/Pt 8·2 liq, ol, cpx, opx, amp, sp 23·5(53), 1·9(30), 14·6(17), 22·1(53), 37·6(64), 0·3(2) 0·02 PU1005 1000 RC158c 46·3 10·1 8·8 (4) 5·5 (6) A9Pd1 8·0 liq, ol, cpx, opx, amp, (sp) 20·9(104), 8·4(113), 12·1(22), 11·4(130), 47·2(98), tr. 0·02 PU906 980 RC158c 67·0 11·5 9·3 (5) Au/Pt 0·2 liq, cpx, opx, amp, sp 15·2(7), 11·3(10), 21·8 (8), 50·6(22), 1·1(2) 0·06 Fractional crystallization RC158c graphite–Pt capsules (FC Mb Pt–C) rk3 1230 Rc158c 23·0 3·1 3·1 (1) C–COH C/Pt –2·5 liq, ol, sp 82·9(5), 17·1(4), 0·01(1) 0·12 82·9 (5) rk6 1200 F1 87·0 4·1 2·6 (2) C–COH C/Pt –3·6 liq, ol, cpx 78·6(12), 3·2(5), 18·2(13) 0·11 65·2 (13) rk11 1170 F2 72·0 4·4 2·6 (2) C–COH C/Pt 5·8 liq, cpx 91·5(4), 8·5(4) 0·02 59·6 (12) rk13 1140 F3 60·0 5·8 n.a. C–COH C/Pt –0·2 liq, cpx, opx, sp, plg 74·9(23), 19·0(17), 0·1(10), 2·2(6), 3·9(23) 0·01 44·7 (23) rk55 1110 F4 85·0 9·8 2·5 (2) C–COH C/Pt –7·7 liq, cpx, opx, plg, (sp) 52·7(16), 7·3(11), 14·7(11), 25·4(10), tr. 0·04 23·5 (20) rk64 1080 F5 120·0 17·6 2·0 (3) C–COH C/Pt –3·6 liq, cpx, plg, ilm, gar 60·5(17), 8·1(8), 11·0(17), 1·2(2), 19·2(14) 0·02 14·2 (20) Equilibrium crystallization RC158c graphite–Pt capsules (EQ Mb Pt–C) PU72 1350 RC158c 4·0 2·8 2·9 (1) C–COH C/Pt 11·3 liq 100 0·31 P1330 1330 RC158c 3·5 2·9 3·2 (1) C–COH C/Pt 1·4 liq, ol 98·1(4), 1·9(3) 0·08 P1301 1300 RC158c 4·3 3·0 3·7(1) C–COH C/Pt 2·5 liq, ol 92·2(1), 7·8(1) 0·02 P1302 1300 RC158c 7·5 2·9 3·4(1) C–COH C/Pt 3·0 liq, ol 95·5(4), 4·5(3) 0·10 P1270 1270 RC158c 10·5 3·0 3·6 (1) C–COH C/Pt 1·0 liq, ol 92·7(1), 7·3(1) 0·02 P1240 1240 RC158c 16·0 3·2 n.a. C–COH C/Pt 0·4 liq, ol 88·8(2), 11·2(2) 0·03 P1210 1210 RC158c 21·7 3·4 3·7(1) C–COH C/Pt 0·9 liq, ol, (sp) 83·1(2), 16·9(1), tr 0·03 P1180 1180 RC158c 25·9 3·7 3·7 (1) C–COH C/Pt 0·6 liq, ol, cpx, (sp) 75·0(4), 19·9(2), 5·1(4), tr 0·02 P1151 1150 RC158c 30·0 3·8 4·4 (2) C–COH C/Pt 3·1 liq, ol, cpx, sp 73·1(7), 19·0(2), 7·7(7), 0·2(2) 0·03 P1152 1150 RC158c 28·8 4·4 n.a. C–COH C/Pt 9·5 liq, ol, cpx, sp 63·6(9), 20·5(3), 15·7(10), 0·2(2) 0·06 P1121 1120 RC158c 50·8 4·6 n.a. C–COH C/Pt 12·7 liq, ol, cpx, opx, (sp) 60·6(37), 18·6(33), 18·7(24), 2·2(20), tr 0·39 P1122 1120 RC158c 99·0 4·7 3·6 (1) C–COH C/Pt 20·3 liq, ol, cpx, opx, (sp) 59·8(5), 14·0(4), 17·2(3), 9·0(7), tr. 0·01 P1090 1090 RC158c 75·0 5·7 4·0 (3) C–COH C/Pt –2·0 liq, ol, cpx, opx, sp 49·4(13), 17·8(11), 25·2(6), 5·4(17), 2·2(1) 0·01 P1060 1060 RC158c 98·0 10·9 6·7 (2) C–COH C/Pt 4·1 liq, ol, cpx, opx, amp, sp 24·9(20), 5·6(17), 38·7(12), 23·0(19), 3·1(31), 4·6(2) 0·01 P1000 1000 RC158c 211·3 13·6 n.a. C–COH C/Pt 5·6 liq, cpx, opx, amp, sp 15·9(2), 27·1(3), 29·1(2), 25·5(8), 2·3(1) 0·01 Run no. . T . Starting . Time . H2O(n) . H2O(R) . Log fO2 . fO2 . Capsule(s) . ΔFe . Run products . Phase proportions (wt %) . ∑R2 . F rel. . . (°C) . material . (h) . (wt %) . (wt %) . ΔNNO . ΔNNO . (inner/outer) . (%) . . . . (wt %) . Fractional crystallization 85-44 AuPd capsules (FC ba AuPd) rk48 1230 85-44 24·0 5·0 4·4 (3) –0·10 A5Pd5/A5Pd5 –6·9 liq 100 0·10 100·0 rk50 1200 85-44 24·0 5·1 n.a. –0·13 A9Pd1/A5Pd5 –6·0 liq, opx 97·8(6), 2·2(5) 0·11 97·8 (6) rk56 1170 F1b 29·0 5·6 n.a. –0·19 A9Pd1/A5Pd5 6·5 liq, cpx, opx 91·8(9), 3·5(12), 4·7(8) 0·12 89·8 (11) rk60 1140 F2b 51·0 5·8 n.a. 0·30 A9Pd1/A5Pd5 1·2 liq, cpx, opx 95·0(12), 3·4(15), 1·6(11) 0·22 85·3 (15) rk63 1110 F3b 72·0 6·2 n.a. 0·40 A9Pd1/A5Pd5 6·8 liq, cpx, opx 94·2(8), 4·9(11), 0·9(8) 0·14 80·3 (14) rk66 1080 F4b 100·0 7·1 n.a. 0·45 A9Pd1/A5Pd5 –3·2 liq, cpx, opx 87·4(5), 9·9(7), 2·7(5) 0·05 70·2 (9) rk69 1050 F5b 113·0 7·6 n.a. 0·40 A9Pd1/A5Pd5 –12·5 liq, cpx, amp 92·7(3), 1·2(4), 6·2(7) 0·10 65·0 (6) rk70 1020 F6b 126·0 7·8 n.a. 0·40 A9Pd1/A5Pd5 –2·0 liq, cpx, amp 92·7(11), 0·6(12), 6·7(21) 0·10 60·3 (11) rk71 990 F7b 150·0 8·6 n.a. 0·40 A9Pd1/A5Pd5 3·2 liq, cpx, amp 88·6(12), 1·5(15), 10·0(24) 0·20 53·4 (15) PU1049 990 990 ba 67·0 8·5 10·2 (4) 0·00 Au 1·9 liq, amp, mag 99·0 (8), 0·9 (6), 0·1 (1) 0·40 59·7 (14) PU1070 980 990 ba 67·5 8·5 9·7 (2) 0·00 Au/Au 11·0 liq, amp 98·2(10), 1·8(10) 0·14 59·2 (15) PU1062 950 990 ba 48·0 8·5 9·1 (3) 0·00 Au/Au 14·0 liq, amp, (gar), (plg) 90·7(20), 9·3(28), tr., tr. 0·51 54·7 (22) PU1064 900 950 ba 70·0 9·0 10·2 (2) 0·00 Au/Au 0·9 liq, amp, gar, plg, ilm 76·5(12), 10·4(28), 6·4(23), 6·4(11), 0·3(3) 0·08 41·8 (19) PU1066 850 900 ba 70·0 9·5 9·5 (3) 0·00 Au/Au 0·0 liq, gar, plg, apa, (ilm) 85·2(9), 6·7/10), 7·8(11), 0·3(3), tr. 0·21 35·6 (14) PU1068 800 850 ba 114·0 9·0 9·9 (3) 0·00 Au/Au 1·3 liq, amp, gar, plg, ilm, (apa) 75·5(20), 6·7 (33), 0·8(10), 16·6(20), 0·5(4), tr. 0·19 26·9 (21) PU1072 750 800 ba 144·0 13·2 11·6 (4) 0·00 Au/Au 0·8 liq, plg, mag, apa, (amp, gar) 89·2(7), 9·1(9), 1·3(2), 0·3(2), tr. 0·12 24·0 (19) PU1072o 720 800 ba 144·0 11·4 11·2 (3) 0·00 Au/Au 0·5 liq, amp, plg, mag, qtz, (apa, gar) 73·1(13), 2·2(4), 20·2(8), 1·2(1), 3·5(6), tr., tr. 0·02 19·7 (20) Fractional crystallization RC158c AuPd capsules (FC Mb AuPd) rk47 1230 RC158c 24·0 2·9 4·2 (2) 2·00 1·5 (3) A5Pd5/A5Pd5 –10 liq, ol, sp 89·8(9), 9·3(9), 0·9(4) 0·76 89·8 (9) rk51 1200 F1a 24·0 3·2 4·3 (2) –0·17 1·3 (2) A9Pd1/A5Pd5 5·0 liq, ol, cpx 92·0(15), 5·2(5), 2·8(15) 0·17 82·6 (17) rk52 1170 F2a 32·0 3·8 5·2 (2) –0·18 1·0 (3) A9Pd1/A5Pd5 7·6 liq, ol, cpx, (sp) 81·6(37), 2·5(10), 15·9(37), tr. 0·60 67·5 (39) rk54 1140 F3a 48·0 4·6 n.a. 0·34 A9Pd1/A5Pd5 –2·6 liq, cpx, sp 84·6(20), 13·8(18), 1·6(6) 0·24 57·1 (37) rk57 1110 F4a 71·0 5·5 n.a. 0·31 A9Pd1/A5Pd5 10·5 liq, cpx, opx, sp 76·7(4), 18·8(44), 0·7(2), 3·8(1) 1·18 43·8 (40) rk58 1080 F5a 100·0 6·7 n.a. 0·23 A9Pd1/A5Pd5 –1·0 liq, cpx, opx, sp 87·4(22), 9·5(25), 1·7(16), 1·4(5) 0·53 38·3 (39) rk73 1050 F6a 120·0 7·3 7·3 (3) 0·35 A9Pd1/A5Pd5 –3·2 liq, cpx, amp 94·9(22), 1·5(23), 3·6(41) 0·35 36·3 (30) rk 65 1020 F7a 150·0 8·1 n.a. 0·38 A9Pd1/A5Pd5 0·2 liq, cpx, amp, sp 94·2(14), 2·6(15), 1·5(30), 1·7(3) 0·12 34·2 (25) rk67 990 F8a 188·0 9·3 n.a. 0·39 Au 0·1 liq, cpx, amp, sp 85·4(9), 6·0(1), 8·3(2), 0·34(26) 0·10 29·2 (15) PU1048 990 990 Mb 67·0 8·2 8·6 (2) 0·00 Au/Au 21·0 liq, cpx 99·0(10), 1·0(8) 0·69 33·9 (17) PU1069 980 990 Mb 67·5 8·2 8·3 (2) 0·00 Au/Au 9·5 liq, cpx 99·4 (9), 0·6(6) 0·24 34·0 (14) PU1061 950 990 Mb 48·0 8·6 8·4 (2) 0·00 Au/Au 11·3 liq, amp, (gar), (cpx) 94·4(8), 5·6(11), tr., tr. 0·06 32·0 (13) PU1063 900 950 Mb 70·0 9·6 8·4 (3) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm 78·0(9), 12·0(2), 9·3(2), 1·5(9), 0·3(3) 0·09 24·9 (12) PU1065 850 900 Mb 70·0 10·1 9·3 (3) 0·00 Au/Au Pt –0·4 liq, gar, plg, apa, (ilm) 90·2(10), 6·0(11), 3·4(11), 0·4(3), tr. 0·24 22·5 (13) PU1067 800 850 Mb 114·0 10·8 9·8 (5) 0·00 Au/Au 0·8 liq, amp, gar, plg, ilm, (apa) 80·9(15), 5·1(33), 0·8(10), 12·8(15), 0·4(3), tr. 0·14 18·2 (17) PU1071 750 800 Mb 144·0 12·2 10·9 (2) 0·00 Au/Au 0·7 liq, plg, mag, (gar) 89·9(10), 8·7(13), 1·4(3), tr. 0·22 16·4 (17) Equilibrium crystallization RC158c AuPd capsules (EQ Mb AuPd) PU910 1200 RC158c 3·8 3·6 4·2 (2) 2·1 (2) A9Pd1/Pt –6·0 liq, ol, sp 91·9(9), 7·3(7), 0·9(8) 0·50 PU908 1200 RC158c 6·3 3·7 4·0 (2) 1·3 (2) A9Pd1/Pt –8·6 liq, ol, sp 90·1 (7), 9·4(5), 0·6(5) 0·30 PU926 1160 RC158c 8·0 3·9 4·1 (2) 2·2 (1) A9Pd1/Pt –2·6 liq, ol, sp 83·6(5), 16·4(4), 0·1(5) 0·14 PU899 1160 RC158c 20·0 3·9 n.p. 1·3 (2) A9Pd1/Pt –1·7 liq, ol, (cpx) 84·5(9), 15·5(3), tr. 0·08 PU909 1120 RC158c 41·0 4·0 4·4 (2) 2·9 (1) A9Pd1/Pt –8·8 liq, ol, cpx, sp 82·0(6), 16·1(2), 1·1(6), 0·8(4) 0·03 PU1006 1100 RC158c 6·2 5·1 4·3 (3) 4·1 (4) A9Pd1 –4·3 liq, ol, cpx, (sp) 64·3(12), 24·4(6), 11·4(15), tr. 0·03 PU1003 1080 RC158c 10·0 5·2 5·8 (3) 4·4 (3) A9Pd1 –2·2 liq, ol, cpx, sp 63·4(9), 25·8(4), 10·4(11), 0·3(5) 0·08 PU1004 1060 RC158c 20·6 6·4 6·0 (4) 5·1 (5) A9Pd1 2·0 liq, ol, cpx, sp 50·5(15), 25·6(8), 22·6(19), 1·3(7) 0·18 PU905 1040 RC158c 65·0 9·5 9·0 (6) 5·3 (2) A9Pd1/Pt 8·2 liq, ol, cpx, opx, amp, sp 23·5(53), 1·9(30), 14·6(17), 22·1(53), 37·6(64), 0·3(2) 0·02 PU1005 1000 RC158c 46·3 10·1 8·8 (4) 5·5 (6) A9Pd1 8·0 liq, ol, cpx, opx, amp, (sp) 20·9(104), 8·4(113), 12·1(22), 11·4(130), 47·2(98), tr. 0·02 PU906 980 RC158c 67·0 11·5 9·3 (5) Au/Pt 0·2 liq, cpx, opx, amp, sp 15·2(7), 11·3(10), 21·8 (8), 50·6(22), 1·1(2) 0·06 Fractional crystallization RC158c graphite–Pt capsules (FC Mb Pt–C) rk3 1230 Rc158c 23·0 3·1 3·1 (1) C–COH C/Pt –2·5 liq, ol, sp 82·9(5), 17·1(4), 0·01(1) 0·12 82·9 (5) rk6 1200 F1 87·0 4·1 2·6 (2) C–COH C/Pt –3·6 liq, ol, cpx 78·6(12), 3·2(5), 18·2(13) 0·11 65·2 (13) rk11 1170 F2 72·0 4·4 2·6 (2) C–COH C/Pt 5·8 liq, cpx 91·5(4), 8·5(4) 0·02 59·6 (12) rk13 1140 F3 60·0 5·8 n.a. C–COH C/Pt –0·2 liq, cpx, opx, sp, plg 74·9(23), 19·0(17), 0·1(10), 2·2(6), 3·9(23) 0·01 44·7 (23) rk55 1110 F4 85·0 9·8 2·5 (2) C–COH C/Pt –7·7 liq, cpx, opx, plg, (sp) 52·7(16), 7·3(11), 14·7(11), 25·4(10), tr. 0·04 23·5 (20) rk64 1080 F5 120·0 17·6 2·0 (3) C–COH C/Pt –3·6 liq, cpx, plg, ilm, gar 60·5(17), 8·1(8), 11·0(17), 1·2(2), 19·2(14) 0·02 14·2 (20) Equilibrium crystallization RC158c graphite–Pt capsules (EQ Mb Pt–C) PU72 1350 RC158c 4·0 2·8 2·9 (1) C–COH C/Pt 11·3 liq 100 0·31 P1330 1330 RC158c 3·5 2·9 3·2 (1) C–COH C/Pt 1·4 liq, ol 98·1(4), 1·9(3) 0·08 P1301 1300 RC158c 4·3 3·0 3·7(1) C–COH C/Pt 2·5 liq, ol 92·2(1), 7·8(1) 0·02 P1302 1300 RC158c 7·5 2·9 3·4(1) C–COH C/Pt 3·0 liq, ol 95·5(4), 4·5(3) 0·10 P1270 1270 RC158c 10·5 3·0 3·6 (1) C–COH C/Pt 1·0 liq, ol 92·7(1), 7·3(1) 0·02 P1240 1240 RC158c 16·0 3·2 n.a. C–COH C/Pt 0·4 liq, ol 88·8(2), 11·2(2) 0·03 P1210 1210 RC158c 21·7 3·4 3·7(1) C–COH C/Pt 0·9 liq, ol, (sp) 83·1(2), 16·9(1), tr 0·03 P1180 1180 RC158c 25·9 3·7 3·7 (1) C–COH C/Pt 0·6 liq, ol, cpx, (sp) 75·0(4), 19·9(2), 5·1(4), tr 0·02 P1151 1150 RC158c 30·0 3·8 4·4 (2) C–COH C/Pt 3·1 liq, ol, cpx, sp 73·1(7), 19·0(2), 7·7(7), 0·2(2) 0·03 P1152 1150 RC158c 28·8 4·4 n.a. C–COH C/Pt 9·5 liq, ol, cpx, sp 63·6(9), 20·5(3), 15·7(10), 0·2(2) 0·06 P1121 1120 RC158c 50·8 4·6 n.a. C–COH C/Pt 12·7 liq, ol, cpx, opx, (sp) 60·6(37), 18·6(33), 18·7(24), 2·2(20), tr 0·39 P1122 1120 RC158c 99·0 4·7 3·6 (1) C–COH C/Pt 20·3 liq, ol, cpx, opx, (sp) 59·8(5), 14·0(4), 17·2(3), 9·0(7), tr. 0·01 P1090 1090 RC158c 75·0 5·7 4·0 (3) C–COH C/Pt –2·0 liq, ol, cpx, opx, sp 49·4(13), 17·8(11), 25·2(6), 5·4(17), 2·2(1) 0·01 P1060 1060 RC158c 98·0 10·9 6·7 (2) C–COH C/Pt 4·1 liq, ol, cpx, opx, amp, sp 24·9(20), 5·6(17), 38·7(12), 23·0(19), 3·1(31), 4·6(2) 0·01 P1000 1000 RC158c 211·3 13·6 n.a. C–COH C/Pt 5·6 liq, cpx, opx, amp, sp 15·9(2), 27·1(3), 29·1(2), 25·5(8), 2·3(1) 0·01 Starting materials are given in Table 1. H2O(n) wt %, H2O contents in the starting material corrected for melt fraction (f) and modal amount of amp (2·1 wt %); H2O(R) wt %, H2O content of glasses (liquid) determined by Raman spectroscopy; n.a., experiment charge is no longer available (lost); log fO2 ΔNNO indicates graphite capsules ≈C–COH (see text), for double-capsule experiments = initial fO2 constrained by Fe3+/Fetot of the starting material expressed as difference from the Ni–NiO equilibrium at the same pressure and temperature; fO2 (ol) (ΔNNO) corresponds to the fO2 calculated from olivine–liquid Fe2+–Mg partitioning (see text). Capsule(s): C/Pt, graphite container sealed in Pt; A9Pd1/Pt or A5Pd5, Fe-pre-saturated inner Au50Pd50 or Au90Pd10 capsule (not for Au100) in outer capsule (Pt or Au50Pd50) containing same starting material. ΔFe % is the difference between the FeO content of the bulk starting composition and the FeO obtained by mass-balance calculations; negative values indicate relative iron gain in wt %. Phase proportions are given in wt % and were calculated by least-squares regressions; numbers in parentheses indicate standard deviations and read as follows: 98·1(4) = 98·1 ± 0·4; ∑R2 is the sum of the squared residuals; F rel. (wt %) is the amount of liquid left in the fractional crystallization experiments relative to the initial starting material (RC158c and 85-44 for Mb and ba runs respectively). Phases: liq, liquid (glass); ol, olivine; cpx, high-Ca clinopyroxene; opx, low-Ca orthopyroxene; sp, Cr–Al–Fe-spinel; ilm, ilmenite; amp, amphibole; gar, garnet, plg, plagioclase, mag, (Ti-)magnetite, apa, apatite, qtz, quartz; phases in parentheses indicate that the phase is present only in trace amounts not quantified during least-squares regression (tr.). All runs were performed at a pressure of 1·0 GPa. Open in new tab An additional series of fractional crystallization experiments were conducted on a primary, mantle-derived, high-Mg basaltic andesite from Mt Shasta (85-44; Baker et al., 1994) with a Mg# of 0·70, olivine phenocrysts with Fo > 90 but higher SiO2 (52·3 wt %) and lower MgO (10·4 wt %) contents than the high-Mg basalt; 85-44 is multiply saturated with olivine, cpx and opx at 1·0 GPa, with H2O contents in excess of 3 wt % but less than 6 wt % and a temperature between 1200 and 1270°C (Baker et al., 1994). This composition thus represents a shallow endmember of mantle-extraction depths just beneath a relatively thin continental or juvenile island arc crust (30 km) to investigate the potential influence of the primary magma compositions typically encountered in suprasubduction settings. The crystallization behavior of the same composition was already studied experimentally by Müntener et al. (2001) at 1·2 GPa under equilibrium crystallization conditions with three H2O contents of 2·5, 3·8 and 5·0 wt %. We used 5·0 wt % as the starting composition for the fractional crystallization experiments at 1·0 GPa, allowing comparison of the fractional and equilibrium crystallization experiments but keeping in mind that the pressure is not identical (1·0 versus 1·2 GPa) and the fO2 conditions might be slightly more reducing in the equilibrium crystallization experiments as these were conducted with Fe-pre-saturated Au80Pd20 inner capsules surrounded by graphite contained in an outer Pt capsule. All starting materials except the natural high-Mg basalt (i.e. initial starting composition RC158c with 3 wt % H2O) and all liquid compositions from the fractional crystallization experiments that were employed as starting material for the next fractionation experiment (Table 1) were mixed from fired (1000°C: SiO2, TiO2, Fe2O3, MgO, CaSiO3) and dried (220–350°C: γ-Al2O3, CaCO3, MnO, Fe2SiO4, NaSi2O7, KAlSi3O8/K2Si4O7) oxides and silicates and dried mixtures of AlOOH (200°C) and Al(OH)3 and Mg(OH)2 (100°C) to adjust the desired H2O content. Fayalite (Fe2SiO4) was synthesized from reagent grade SiO2 and Fe2O3 in an iron pre-saturated Pt crucible for two times 24 h with intermediate grinding at 1120°C at a log fO2 of –12·50 (1 log unit above the iron–wüstite equilibrium) in a vertical gas-mixing (CO2–H2) quench furnace and checked by powder X-ray diffraction analysis. KAlSi3O8 (potassium-feldspar) was obtained from a natural K-feldspar (Ba-bearing) that was finely ground and mixed with KCl (ratio KCl:KAlSi3O8 10:1) heated to 850°C in a Pt crucible for 48 h, quenched and the KCl washed out with deionized H2O. This produced nearly pure K-feldspar. Capsule material and treatment Graphite–Pt double capsules Two series of experiments, an equilibrium (EQ Mb Pt–C) and a fractional (FC Mb Pt–C) series starting from the high-Mg basalt RC158c (Table 1), were conducted with graphite–Pt (Pt–C) double capsules, resulting in relatively reducing conditions between the C–CO2 and the H2O maximum in the graphite buffered C–O–H system; that is, conditions of 2–3 log units below the Ni–NiO equilibrium (ΔNNO; Ulmer & Luth, 1991; Frost & Wood, 1995). Graphite–Pt double capsules have the advantage that Fe loss is minimal and that, theoretically (see below), volatile loss can be avoided. This was achieved for the equilibrium crystallization experiments, but was not the case for the fractional crystallization experiments, which resulted in H2O-poor compositions for the latter. AuPd–Pt and Au/AuPd–Au/AuPd double capsules for fractional crystallization experiments The two fractional crystallization series (high-Mg basalt starting composition FC Mb AuPd and basaltic andesite starting composition FC ba AuPd) and most equilibrium crystallization experiments on the Mg basalt (EQ Mb AuPd) were conducted in a double capsule, AuPd alloy or Au inner–Pt, AuPd alloy or Au outer capsule arrangement employing Au90Pd10 inner capsules (2·3 mm o.d., 0·15 mm wall thickness) except for the two highest temperature experiments (1230°C, rk47 and rk48), which used Au50Pd50 inner capsules, and experiments below 1000°C, which used pure Au capsules; the outer capsules were always made of Au50Pd50 (4·0 mm o.d., 0·2 mm wall thickness), except below 1000°C, where pure Au was employed. Both capsules contained the same starting material. The inner AuPd capsules (but not the pure Au capsules) were additionally pre-saturated in iron to minimize potential Fe loss using an anhydrous basaltic starting material (F3 for the FC series and RC158c for the EQ series; Table 1) tightly packed into annealed, cleaned and one-side welded capsules that were run for 48 h in a gas-mixing furnace at 1180°C and a log fO2 of –9·65 corresponding to ΔNNO – 1·7. The recovered capsules were immersed in concentrated HF for 48 h to dissolve the silicate glass and cleaned in an ultrasonic bath using distilled water. This results in negligible Fe and H2 loss, a prerequisite to maintain nearly constant fO2 conditions in fluid-undersaturated experiments (Hall et al., 2004; Kägi et al., 2005). Below 990°C the capsule arrangement was modified. A triple-capsule setup was employed with two different inner capsules, one containing the derivative composition of the basaltic andesite parental compositions (denoted ba in Tables 1 and 2) and the other containing the derivative composition of the high-Mg basaltic parental composition (Mb), both packed into an outer Au capsule containing the Mb derivative composition (these compositions are very similar except for some minor differences in CaO and K2O). Experiments PU1061 and PU1062 contained additionally 1 wt % of An95 plagioclase seeds (Table 1; Miyake-Jima volcano, Japan; Amma-Miyasaka & Nakagawa, 2002), experiments PU1063-PU1072 contained 1 wt % each of An95 plagioclase and almandine-rich garnet (Table 1; Níjar, southern Spain; Munoz-Espadas et al., 2000) seeds to prevent potential nucleation problems of these two phases and to allow measurement of overgrowth rims on seeds in the case of plagioclase, which otherwise forms very fine platelets that are very difficult to analyze reliably. Runs below 1000°C of this series contained 2 wt % of an additional trace element doped diopside glass providing 40 ppm of 32 trace elements. The addition of the diopside glass was considered calculating the final composition of the respective starting materials. However, no correction was made for the additional seed crystals (plag and garnet) as they were consistently present in the final run products armored by overgrowth rims (see Fig. 2c–f) and, thus, did not significantly contribute to the liquid composition in these experiments. The situation for the equilibrium crystallization experiments on the high-Mg basalt composition in AuPd alloys (EQ Mb AuPd) was different. Initially, a similar setup to that for the respective fractionation experiments was utilized; that is, iron pre-saturated inner capsules and Pt outer capsules. This turned out to be problematic because the natural starting material is rather oxidized, corresponding to an fO2 at its liquidus (c. 1300°C) of NNO + 2. The pre-saturation at NNO – 1·7 resulted in unconstrained conditions owing to the uptake of Fe from the Au90Pd10 capsule; repeat experiments provided inconsistent phase equilibria and phase proportions. Therefore, we abandoned Fe pre-saturation in this case and changed to untreated inner Au90Pd10 capsules, initially in a double-capsule arrangement with Pt outer capsules. This resulted in consistent, reproducible phase equilibria with minimal Fe loss or gain because Fe alloying with Au-dominated AuPd alloy is not a critical issue at oxidizing conditions, unlike under more reducing conditions where even pure Au shows considerable alloying with Fe from basaltic melts (e.g. Ratajeski & Sisson, 1999). A final modification for lower temperature runs was made by employing single Au90Pd10 capsules because the employment of double capsules (and MgO ceramic spacers around them) resulted in extensive quench crystallization in the inner capsule owing to decreased quenching rates that basically inhibited successful analysis of the melt phase; using single capsules at low temperature eliminated quench crystallization; the phase relations, however, were identical for double and single capsules at comparable conditions. For fractional crystallization experiments in AuPd capsules, oxygen fugacity was constrained by adjusting the proportions of ferrous and ferric iron in the starting material to a value close to NNO (NNO – 0·2 to NNO + 0·45) at the conditions of the experiment. The appropriate Fe2+/Fe3+ ratio was calculated with the algorithm of Kress & Carmichael (1991). This approach is strictly valid only for charges composed entirely of liquid; therefore, in the present work this is an approximation as the melt fraction (f) varied between unity and 0·75. In the absence of Fe–Ti-oxides, the fO2 most probably increases in the charges owing to the more incompatible behavior of ferric relative to ferrous iron in the presence of olivine and clinopyroxene, the main liquidus phases at high temperature. At lower temperatures, prediction is more difficult as amphibole can accommodate significant amounts of ferric relative to ferrous iron. However, in these experiments the melt fraction always exceeded 75%; therefore, we do not expect significant deviation from the target fO2 owing to crystallization of the solid phases. A prerequisite for successful application of this technique is the absence of significant hydrogen and/or iron loss during the experiment (i.e. closed-system behavior). Hydrogen diffusion was minimized by employing the double-capsule technique, packing the same material in the inner and outer capsule and thereby limiting the fH2 gradient across the inner–outer capsule interface. Fe loss was minimized by using Fe pre-saturated AuPd inner capsules (Kägi et al., 2005) and by minimizing the run duration to the minimum required for close approach to equilibrium (24–188 h). The situation is different for the ‘closed-system’ equilibrium crystallization experiments in AuPd capsules: the initial high-Mg basaltic starting material has an Fe3+/Fetot of 0·31 (molar) that corresponds to a log fO2 of NNO + 2·0 at its liquidus temperature of c. 1300°C. During crystallization down to 1040°C olivine and cpx are the dominant ferro-magnesian phases and only minor Cr-rich spinel crystallized at the highest temperatures. Thus, it is expected that ferrous iron is preferentially extracted by the solid assemblage, resulting in increased Fe3+/Fetot ratios in the residual liquid and thus resulting in an ‘auto-oxidation’ of the system with decreasing melt fraction; a result that is clearly supported by the apparent Fe–Mg partitioning between olivine and liquid that decreases with decreasing temperature (see results and discussion sections). Differences of up to 1 log unit in fO2 were observed at high temperature as a function of the assembly employed: two experiments at 1200 and 1160°C (PU908 and 899) were performed with talc–Pyrex–graphite–Pyrex–BN assemblies that revealed an fO2 of about 1 log unit more reducing than the experiments conducted at the same conditions with talc–Pyrex–graphite–MgO assemblies that were used for the remainder of this experimental series. Experimental set-up Graphite–Pt capsules The starting materials were dried at 110°C before being loaded into the graphite containers. The graphite capsules were closed with a tight-fitting lid and packed into Pt capsules that additionally contained fine-grained graphite at the bottom and top to prevent cracking of the inner capsule during initial compression in the piston cylinder. The AuPd and Au inner capsules were tightly filled with c. 15 mg of starting material, weighed, crimped, cut and welded shut using an arc welder. Sealing of the capsules was checked by weighing before and after welding by submersion in acetone. The same procedure was repeated for the outer capsule containing the inner capsule plus an additional 50 mg of the starting material around the inner capsule. The capsules of all experiments except the graphite–Pt equilibrium experiments on the high-Mg basalt (EQ Mb Pt–C; see below) and runs PU899 and 908 were embedded vertically in MgO-ceramic and contained in a NaCl–Pyrex–graphite–MgO piston cylinder assembly. The capsule was placed in the hotspot of the assembly, which was computed using the numerical model of Hernlund et al. (2006). All experiments were performed in 14 mm bore end-loaded piston-cylinder apparatus. A friction correction of –3% was applied to the nominal pressure, based on calibrations employing the fayalite + qtz = orthoferrosilite univariant reaction at 1000°C and 1·41 GPa (Bohlen et al., 1980) and the qtz = coesite transition at 1000°C and 3·07 GPa (Bose & Ganguly, 1995). The temperature was measured with a B-type Pt94Rh6–Pt70Rh30 thermocouple, with an estimated accuracy of ±10°C. No corrections for the pressure effect on the electromotive force (e.m.f.) were performed. The experiments were terminated by switching off the power supply, resulting in quenching rates in excess of 100°C s–1. The EQ Mb Pt–C experiments (denoted P1xxx runs in Table 2) and runs PU899 and PU908 were conducted in the same apparatus, but employing BN–crushable alumina–Pyrex inner assembly parts and talc–Pyrex outer parts. The friction correction for this assembly amounts to –10% based on the same calibration reactions. Temperatures in these experiments were measured with S-type Pt–Pt90Rh10 thermocouples. Analytical methods Recovered charges were embedded in epoxy resin, ground to expose a longitudinal section of the capsule(s) and polished with diamond paste down to 1 µm. For EPMA and scanning electron microscope (SEM) analyses samples were coated with 20 nm of carbon. Electron probe microanalysis and scanning electron microscopy Experimental run products were analyzed with three electron microprobes at ETH Zurich using wavelength- (and energy-) dispersive spectrometers (WDS and EDS) on a ARL SEMQ, a CAMECA SX50 and a JEOL JXA-8200. An acceleration voltage of 15 kV and a variable beam current and beam size (7 nA and 10–20 µm for glasses; 20 nA and 1 µm for minerals) were used to minimize alkali-migration during analyses of hydrous glasses. Peak and background counting times were 20–30 s. All data were corrected with either the ZAF (ARL SEMQ) or the PRZ model (Goldstein et al., 2003; Korolyuk et al., 2009). For glasses, H2O was used as additional element for ZAF or PRZ corrections, and was calculated as difference from 100 wt %. Data are consistent among the three electron microprobes because the same set of standards was employed for all analyses, and several charges have been reanalyzed in later sessions resulting in identical data within statistical error. Additionally, back-scattered electron (BSE) images and characteristic element X-ray maps were acquired with an JEOL JSM-6390LA SEM equipped with a Thermo-Scientific Noran system 7 EDS detector allowing standardized (quantitative) analyses that were particularly useful for checking the Na2O (and K2O) contents of hydrous glass with a beam current of 2·4 nA. Modal proportions were calculated using non-weighted, least-squares regression (LSR) analyses implemented in Microsoft EXCEL® 2010 and/or Origin 9.1® balancing the nominal composition of the starting material against the averages of all analyzed phases in the experimental charges (Tables 1–3). Table 3: Electron microprobe analyses of experimental phases of equilibrium crystallization experiments on composition RC158c Run # . T (°C) . Phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . Total . xMg . EQ Mb AuPd PU910 1200 glass 8 44·41 (25) 0·77 (4) 0·08 (4) 12·11 (10) 9·08 (12) 0·19 (5) n.d. 13·11 (44) 10·65 (24) 1·17 (11) 0·41 (2) 92·00 0·720 ol 8 40·70 (05) 0·02 (1) 0·05 (3) 0·05 (3) 7·89 (15) 0·14 (2) n.d. 50·78 (39) 0·24 (11) 0·01 (1) 0·00 (00) 99·88 0·920 sp 4 0·74 (93) 0·53 (7) 36·88 (483) 14·67 (367) 28·67 (80) 0·00 n.d. 15·46 (80) 0·41 (16) 0·05 (1) 0·01 (1) 97·43 0·490 PU908 1200 glass 6 44·34 (29) 0·76 (4) 0·09 (5) 12·65 (26) 9·34 (17) 0·17 (4) n.d. 12·49 (16) 10·94 (7) 1·14 (10) 0·45 (1) 92·36 0·704 ol 8 40·42 (20) 0·02 (1) 0·04 (2) 0·05 (2) 9·24 (7) 0·15 (2) n.d. 50·08 (21) 0·25 (2) 0·01 (1) 0·00 100·24 0·906 sp 5 0·86 (35) 0·34 (2) 39·42 (309) 19·35 (157) 24·09 (62) 0·00 n.d. 15·51 (31) 0·4 (8) 0·03 (2) 0·01 (1) 100·01 0·534 PU926 1160 glass 6 44·94 (13) 0·81 (3) 0·06 (3) 14·41 (11) 8·89 (8) 0·18 (2) 0·02 (2) 10·06 (15) 11·60 (10) 1·12 (4) 0·48 (1) 92·57 0·669 ol 8 40·56 (24) 0·02 (1) 0·03 (3) 0·12 (3) 9·72 (15) 0·19 (3) 0·23 (3) 49·61 (20) 0·25 (1) 0·02 (2) 0·01 (1) 100·74 0·901 sp 6 0·23 (11) 0·36 (14) 37·24 (646) 15·50 (186) 29·18 (459) 0·31 (3) 0·12 (5) 13·89 (47) 0·26 (2) 0·01 (2) 0·01 (1) 97·12 0·459 PU899 1160 glass 8 45·02 (62) 0·87 (4) 0·06 (2) 14·12 (12) 10·09 (26) 0·20 (6) n.d. 10·41 (21) 11·50 (22) 1·34 (6) 0·50 (2) 94·11 0·648 ol 8 39·22 (105) 0·01 (1) 0·03 (1) 0·04 (2) 11·55 (18) 0·16 (4) n.d. 48·12 (31) 0·21 (2) 0·00 0·01 (1) 99·36 0·881 cpx 9 51·24 (44) 0·23 (2) 0·77 (10) 4·64 (31) 5·64 (22) 0·08 (2) n.d. 17·26 (31) 20·20 (27) 0·32 (4) 0·01 (1) 100·39 0·845 PU909 1120 glass 8 45·01 (32) 0·81 (5) 0·05 (4) 13·94 (23) 9·14 (11) 0·19 (3) n.d. 9·59 (16) 11·10 (10) 1·31 (7) 0·49 (2) 91·56 0·651 ol 6 39·70 (19) 0·02 (1) 0·02 (1) 0·03 (2) 9·36 (16) 0·21 (1) n.d. 49·88 (19) 0·20 (1) 0·02 (2) 0·00 99·43 0·905 cpx 9 50·91 (18) 0·24 (3) 0·78 (13) 4·30 (25) 5·12 (23) 0·09 (3) n.d. 17·13 (29) 21·27 (38) 0·27 (2) 0·01 (1) 100·13 0·856 sp 3 0·17 (7) 0·57 (11) 19·96 (490) 16·03 (221) 44·08 (464) 0·00 n.d. 13·22 (63) 0·21 (8) 0·01 (1) 0·01 (1) 94·27 0·348 PU1006 1100 glass 12 46·87 (32) 1·02 (5) 0·02 (2) 17·31 (54) 9·08 (25) 0·17 (2) n.d. 5·55 (78) 10·63 (51) 1·46 (8) 0·57 (4) 92·70 0·522 ol 10 40·61 (13) 0·02 (1) 0·03 (2) 0·06 (2) 12·10 (13) 0·18 (2) 0·17 (2) 47·51 (31) 0·27 (2) 0·00 0·01 (1) 100·96 0·875 cpx 9 50·62 (34) 0·35 (2) 0·57 (14) 5·46 (30) 5·79 (24) 0·14 (2) 0·04 (2) 16·55 (50) 20·41 (58) 0·28 (2) 0·01 (1) 100·22 0·836 sp 6 0·32 (21) 0·54 (6) 17·75 (42) 33·96 (174) 28·88 (80) 0·21 (1) 0·19 (2) 16·41 (30) 0·30 (11) 0·01 (1) 0·01 (1) 98·57 0·503 PU1003 1080 glass 6 45·29 (44) 0·98 (4) 0·02 (3) 17·10 (25) 8·14 (16) 0·18 (1) n.d. 4·40 (42) 11·24 (14) 1·36 (5) 0·49 (4) 89·20 0·490 ol 8 40·30 (22) 0·02 (1) 0·02 (2) 0·03 (2) 12·01 (25) 0·20 (2) 0·17 (2) 47·29 (39) 0·25 (3) 0·01 (1) 0·00 100·30 0·875 cpx 9 50·38 (70) 0·35 (13) 0·59 (24) 4·88 (95) 5·56 (66) 0·14 (3) 0·02 (2) 16·33 (60) 20·87 (80) 0·24 (3) 0·01 (1) 99·36 0·839 sp 6 0·14 (10) 0·60 (10) 27·37 (563) 23·79 (333) 30·51 (215) 0·25 (1) 0·16 (5) 14·05 (65) 0·27 (4) 0·01 (2) 0·01 (1) 97·15 0·451 PU1004 1060 glass 13 46·51 (55) 1·02 (8) 0·02 (2) 19·04 (56) 6·54 (59) 0·16 (2) n.d. 2·82 (52) 10·04 (49) 1·22 (10) 0·33 (5) 87·70 0·435 ol 8 40·00 (14) 0·02 (1) 0·01 (1) 0·03 (1) 13·90 (8) 0·22 (2) 0·17 (2) 46·08 (24) 0·20 (1) 0·00 0·01 (1) 100·63 0·855 cpx 10 49·81 (96) 0·48 (14) 0·20 (12) 5·89 (101) 6·64 (35) 0·17 (2) 0·02 (2) 16·06 (91) 19·86 (96) 0·27 (5) 0·01 (1) 99·39 0·812 sp 6 0·17 (18) 0·49 (5) 7·11 (118) 44·64 (164) 29·93 (83) 0·17 (2) 0·22 (3) 16·19 (67) 0·18 (2) 0·00 0·01 (1) 99·13 0·491 PU905 1040 glass 5 46·82 (25) 0·83 (1) 0·01 (1) 18·94 (28) 5·62 (27) 0·20 (2) n.d. 2·37 (1) 9·27 (30) 1·08 (12) 0·35 (2) 85·46 0·429 ol 3 39·04 (30) 0·00 0·02 (1) 0·05 (1) 13·47 (13) 0·26 (3) n.d. 46·76 (40) 0·16 (2) 0·00 0·01 (1) 99·77 0·861 cpx 10 48·21 (48) 0·57 (24) 0·23 (10) 6·44 (53) 6·78 (27) 0·18 (2) n.d. 15·50 (47) 20·48 (47) 0·28 (3) 0·01 (1) 98·67 0·803 High-Al opx 5 51·02 (37) 0·16 (3) 0·08 (1) 6·37 (27) 10·44 (21) 0·25 (1) n.d. 29·78 (24) 1·06 (5) 0·03 (1) 0·01 (1) 99·20 0·836 Low-Al opx 5 53·85 (19) 0·11 (1) 0·04 (3) 2·93 (16) 9·65 (12) 0·27 (3) n.d. 31·21 (20) 1·25 (7) 0·03 (0) 0·01 (1) 99·33 0·852 amph 10 40·95 (57) 1·12 (7) 0·17 (8) 14·29 (32) 8·69 (42) 0·13 (4) n.d. 16·22 (16) 11·62 (32) 2·03 (4) 0·64 (13) 95·85 0·769 sp 5 0·14 (03) 1·04 (31) 9·88 (722) 22·39 (158) 49·75 (600) 0·07 (8) n.d. 11·68 (25) 0·21 (7) 0·02 (2) 0·01 (0) 95·17 0·295 PU1005 1000 glass 20 48·51 (98) 0·74 (10) 0·01 (1) 19·68 (48) 5·23 (81) 0·16 (2) n.d. 1·81 (44) 8·82 (47) 0·90 (12) 0·43 (9) 86·29 0·382 ol 10 39·90 (22) 0·02 (1) 0·01 (2) 0·03 (2) 15·87 (11) 0·25 (3) 0·15 (2) 44·96 (34) 0·19 (2) 0·00 0·00 101·38 0·835 cpx 15 49·45 (67) 0·52 (9) 0·28 (9) 6·51 (72) 6·72 (23) 0·17 (2) 0·03 (2) 15·37 (42) 20·57 (53) 0·22 (2) 0·01 (1) 99·86 0·803 opx 10 51·85 (49) 0·18 (2) 0·22 (2) 6·55 (29) 11·30 (12) 0·24 (2) 0·06 (2) 29·15 (32) 1·18 (6) 0·02 (2) 0·00 100·74 0·821 amph 15 42·34 (49) 1·13 (18) 0·32 (15) 14·29 (24) 8·63 (74) 0·10 (2) 0·05 (2) 16·33 (54) 11·40 (15) 1·95 (11) 0·60 (15) 97·13 0·771 sp 6 0·06 (2) 0·49 (3) 5·73 (13) 45·13 (62) 31·56 (54) 0·20 (2) 0·23 (2) 15·61 (7) 0·16 (3) 0·00 0·00 99·16 0·469 PU906 980 glass 10 49·39 (63) 0·53 (8) 0·03 (2) 19·03 (38) 3·82 (51) 0·20 (7) n.d. 1·66 (58) 8·47 (23) 0·68 (9) 0·52 (5) 84·33 0·437 cpx 10 48·49 (66) 0·43 (9) 0·18 (9) 5·77 (68) 6·76 (24) 0·20 (3) n.d. 15·66 (61) 20·87 (45) 0·20 (2) 0·01 (1) 98·57 0·805 opx 6 50·43 (52) 0·20 (1) 0·18 (6) 6·74 (49) 11·77 (20) 0·26 (2) n.d. 28·14 (24) 1·25 (15) 0·03 (5) 0·01 (2) 99·01 0·810 amph 6 41·73 (45) 0·92 (6) 0·26 (13) 13·61 (61) 8·35 (18) 0·13 (2) n.d. 16·43 (25) 11·77 (15) 1·62 (3) 0·62 (5) 95·43 0·778 sp 6 0·16 (07) 1·53 (13) 3·34 (246) 15·37 (46) 64·18 (255) 0·18 (11) n.d. 9·09 (35) 0·26 (4) 0·00 0·01 (1) 94·11 0·201 EQ Mb Pt–C PU72 1350 glass 5 46·19 (28) 0·67 (4) 0·15 (1) 12·80 (07) 8·14 (20) 0·15 (4) n.d. 17·06 (14) 9·93 (8) 1·19 (3) 0·40 (1) 96·68 0·789 ZP1330 1330 glass 10 46·90 (34) 0·74 (6) 0·21 (3) 12·70 (17) 9·04 (20) 0·17 (2) n.d. 16·05 (21) 10·13 (7) 1·19 (4) 0·42 (2) 97·56 0·760 ol 11 41·45 (19) 0·01 (1) 0·10 (2) 0·00 9·20 (18) 0·13 (1) 0·07 (4) 49·73 (20) 0·25 (1) 0·01 (1) 0·00 100·95 0·906 ZP1301 1300 glass 10 46·58 (39) 0·74 (6) 0·21 (3) 13·35 (27) 8·76 (21) 0·17 (2) 0·01 (2) 13·93 (28) 10·82 (13) 1·27 (4) 0·43 (1) 96·27 0·739 ol 10 41·23 (25) 0·01 (2) 0·15 (1) 0·00 10·21 (22) 0·15 (2) 0·14 (3) 49·20 (31) 0·29 (2) 0·01 (1) 0·00 101·39 0·896 ZP1302 1300 glass 10 46·93 (52) 0·74 (6) 0·20 (3) 13·00 (31) 8·84 (35) 0·16 (3) n.d. 15·12 (10) 10·33 (5) 1·20 (4) 0·44 (1) 96·96 0·753 ol 10 41·69 (19) 0·01 (1) 0·09 (2) 0·00 9·51 (22) 0·14 (2) 0·06 (5) 49·91 (15) 0·23 (2) 0·02 (1) 0·00 101·65 0·903 ZP1270 1270 glass 10 46·52 (27) 0·80 (4) 0·18 (2) 13·40 (38) 8·91 (26) 0·15 (3) 0·00 14·11 (60) 10·72 (24) 1·22 (3) 0·42 (1) 96·44 0·738 ol 10 41·09 (32) 0·00 0·12 (1) 0·00 10·22 (34) 0·13 (1) 0·04 (4) 49·17 (62) 0·24 (2) 0·01 (1) 0·00 101·02 0·896 ZP1240 1240 glass 9 46·95 (27) 0·78 (8) 0·19 (2) 14·08 (41) 8·84 (28) 0·15 (2) 0·01 (2) 12·81 (15) 11·11 (9) 1·32 (4) 0·47 (1) 96·70 0·721 ol 11 41·31 (42) 0·00 0·12 (2) 0·00 11·23 (27) 0·15 (2) 0·05 (4) 48·64 (72) 0·24 (2) 0·01 (1) 0·01 (1) 101·75 0·885 ZP1210 1210 glass 12 47·67 (38) 0·85 (5) 0·16 (2) 15·17 (45) 8·49 (23) 0·17 (3) 0·00 10·71 (39) 12·00 (13) 1·37 (6) 0·49 (2) 97·09 0·692 ol 10 41·18 (38) 0·00 0·13 (3) 0·00 12·37 (37) 0·16 (2) 0·04 (3) 48·16 (20) 0·28 (3) 0·01 (1) 0·00 102·32 0·874 sp 10 0·19 (29) 0·33 (3) 32·81 (86) 34·89 (61) 14·49 (31) 0·17 (2) 0·04 (2) 16·30 (19) 0·11 (4) 0·00 0·00 99·34 0·667 ZP1180 1180 glass 10 47·44 (71) 0·92 (8) 0·10 (4) 16·31 (56) 8·31 (34) 0·17 (2) 0·01 (2) 9·15 (52) 12·00 (24) 1·48 (6) 0·52 (2) 96·41 0·662 ol 8 40·61 (26) 0·00 0·05 (4) 0·00 13·44 (22) 0·19 (2) 0·03 (4) 46·45 (10) 0·26 (2) 0·00 0·00 101·03 0·860 High-Al cpx 9 50·56 (45) 0·40 (4) 1·28 (17) 5·76 (51) 4·20 (31) 0·13 (1) 0·00 16·46 (10) 20·76 (14) 0·25 (3) 0·00 99·81 0·875 Low-Al cpx 7 52·70 (66) 0·29 (4) 0·83 (6) 2·26 (27) 4·45 (38) 0·16 (1) 0·00 18·22 (24) 20·01 (60) 0·16 (3) 0·00 99·08 0·880 sp 8 0·00 0·38 (4) 32·60 (273) 33·77 (224) 15·69 (51) 0·18 (3) 0·00 15·45 (36) 0·16 (9) 0·00 0·00 98·24 0·637 ZP1151 1150 glass 10 47·75 (36) 0·94 (6) 0·05 (4) 16·22 (30) 8·19 (26) 0·17 (2) 0·00 9·39 (11) 11·40 (8) 1·43 (8) 0·53 (1) 96·08 0·672 ol 16 40·76 (73) 0·01 (2) 0·06 (3) 0·02 (02) 13·58 (30) 0·17 (3) 0·02 (2) 46·21 (87) 0·34 (12) 0·01 (1) 0·00 101·17 0·858 High-Al cpx 8 51·00 (49) 0·48 (10) 1·19 (12) 7·07 (68) 4·43 (49) 0·12 (1) 0·00 16·34 (43) 20·43 (25) 0·29 (2) 0·00 101·35 0·868 Low-Al cpx 8 53·20 (59) 0·23 (2) 1·06 (17) 4·08 (40) 4·42 (41) 0·12 (3) 0·00 17·97 (57) 19·89 (62) 0·26 (2) 0·00 101·24 0·879 sp 10 0·26 (33) 0·37 (9) 27·47 (260) 36·27 (298) 15·77 (49) 0·18 (4) 0·00 15·87 (52) 0·28 (15) 0·00 0·00 96·49 0·642 ZP1152 1150 glass 12 46·86 (96) 1·01 (9) 0·08 (3) 17·44 (43) 7·82 (35) 0·24 (4) 0·01 (3) 8·04 (43) 10·75 (51) 2·10 (11) 0·53 (4) 94·88 0·647 ol 8 40·70 (32) 0·00 0·08 (1) 0·00 13·33 (57) 0·28 (4) 0·11 (3) 46·31 (47) 0·26 (5) 0·00 0·00 101·07 0·861 cpx 10 50·97 (78) 0·48 (8) 1·06 (8) 6·49 (94) 4·86 (30) 0·20 (4) 0·00 16·30 (48) 19·99 (17) 0·34 (1) 0·01 (1) 100·68 0·857 sp 11 0·00 0·24 (3) 21·65 (119) 45·64 (118) 14·04 (55) 0·23 (5) 0·10 (6) 17·86 (40) 0·09 (6) 0·00 0·01 (1) 100·18 0·694 ZP1121 1120 glass 12 46·82 (125) 0·84 (6) 0·04 (3) 17·73 (38) 7·38 (49) 0·13 (2) 0·00 8·17 (28) 10·92 (36) 2·79 (12) 0·58 (5) 95·40 0·664 ol 12 41·40 (53) 0·01 (1) 0·05 (3) 0·00 14·05 (61) 0·17 (3) 0·07 (3) 45·95 (46) 0·24 (3) 0·02 (3) 0·01 (1) 101·97 0·854 cpx 10 51·76 (71) 0·52 (10) 0·34 (12) 7·81 (81) 5·54 (39) 0·14 (3) 0·02 (2) 16·50 (38) 18·89 (45) 0·43 (8) 0·03 (1) 101·97 0·842 High-Al opx 4 53·97 (38) 0·15 (4) 0·30 (9) 6·62 (22) 7·89 (32) 0·16 (2) 0·00 29·89 (53) 1·64 (2) 0·13 (1) 0·00 100·75 0·871 Low-Al opx 7 55·71 (53) 0·12 (3) 0·27 (7) 4·00 (85) 8·26 (29) 0·14 (2) 0·00 30·68 (45) 1·71 (10) 0·13 (3) 0·00 101·02 0·869 sp 8 0·97 (49) 0·14 (2) 9·73 (119) 56·32 (102) 12·18 (26) 0·12 (2) 0·06 (3) 19·76 (60) 0·16 (6) 0·00 0·00 99·44 0·743 ZP1122 1120 glass 10 46·51 (65) 0·98 (6) 0·00 17·91 (38) 6·86 (20) 0·13 (3) 0·00 8·66 (20) 11·05 (14) 1·94 (4) 0·74 (2) 94·77 0·692 ol 8 41·63 (47) 0·01 (1) 0·01 (1) 0·00 12·80 (23) 0·14 (4) 0·02 (7) 47·65 (57) 0·28 (4) 0·00 0·00 102·59 0·869 cpx 12 52·26 (62) 0·40 (8) 0·53 (19) 6·43 (58) 4·65 (38) 0·16 (4) 0·02 (4) 16·99 (47) 20·08 (58) 0·28 (3) 0·00 101·78 0·867 High-Al opx 9 53·27 (52) 0·21 (5) 0·39 (7) 7·78 (32) 8·69 (73) 0·13 (3) 0·03 (5) 29·50 (67) 1·72 (2) 0·04 (1) 0·00 101·76 0·858 Low-Al opx 4 56·62 (58) 0·12 (1) 0·29 (4) 3·38 (10) 8·81 (30) 0·15 (2) 0·02 (4) 31·12 (31) 1·86 (6) 0·03 (1) 0·00 102·40 0·863 ZP1090 1090 glass 8 49·12 (24) 1·09 (9) 0·01 (1) 18·33 (89) 8·63 (48) 0·17 (2) 0·01 (2) 7·01 (52) 10·29 (67) 2·31 (12) 0·86 (9) 97·83 0·592 ol 7 42·41 (28) 0·00 0·03 (2) 0·00 17·48 (45) 0·27 (2) 0·15 (2) 44·77 (48) 0·26 (1) 0·00 0·01 (1) 105·38 0·820 cpx 9 50·87 (38) 0·51 (4) 0·41 (8) 6·95 (43) 5·52 (17) 0·15 (3) 0·00 16·29 (13) 20·05 (23) 0·33 (2) 0·01 (1) 101·10 0·840 High-Al opx 5 52·27 (33) 0·30 (3) 0·03 (3) 8·23 (48) 11·01 (13) 0·25 (2) 0·00 27·92 (10) 1·72 (18) 0·05 (1) 0·01 (1) 101·79 0·819 Low-Al opx 5 55·25 (7) 0·14 (2) 0·03 (3) 3·70 (12) 10·99 (24) 0·25 (4) 0·00 29·68 (19) 1·67 (1) 0·05 (2) 0·01 (1) 101·76 0·828 sp 8 0·00 0·18 (4) 4·43 (291) 65·62 (298) 13·90 (54) 0·13 (4) 0·09 (3) 19·74 (50) 0·17 (6) 0·01 (2) 0·02 (1) 104·30 0·717 ZP1060 1060 glass 13 46·87 (148) 1·07 (18) 0·02 (3) 19·33 (52) 7·51 (40) 0·11 (3) 0·00 4·85 (64) 7·38 (37) 3·63 (30) 1·55 (9) 92·29 0·535 ol 4 39·12 (21) 0·05 (1) 0·02 (2) 0·00 20·11 (50) 0·23 (2) 0·11 (3) 39·55 (40) 0·28 (3) 0·00 0·00 99·48 0·778 cpx 10 51·16 (72) 0·81 (6) 0·15 (7) 7·39 (20) 5·82 (26) 0·14 (2) 0·02 (2) 15·79 (64) 20·21 (32) 0·41 (4) 0·02 (1) 101·93 0·829 opx 9 53·57 (74) 0·28 (4) 0·11 (4) 6·46 (64) 11·91 (57) 0·20 (3) 0·01 (2) 27·78 (50) 1·54 (22) 0·05 (4) 0·01 (2) 101·93 0·806 amp 5 42·28 (74) 1·94 (15) 0·15 (1) 15·32 (40) 7·51 (33) 0·11 (1) 0·03 (2) 15·61 (34) 11·15 (72) 2·40 (17) 0·71 (4) 97·20 0·787 sp 8 0·00 0·20 (4) 1·64 (134) 60·86 (168) 15·25 (33) 0·11 (2) 0·14 (3) 17·60 (42) 0·20 (2) 0·00 0·00 96·00 0·673 ZP1000 1000 glass 8 44·67 (55) 0·92 (7) 0·00 20·72 (13) 6·50 (09) 0·06 (2) 0·00 3·98 (41) 9·01 (15) 3·12 (20) 1·24 (6) 90·20 0·521 cpx 9 50·81 (82) 0·70 (20) 0·16 (4) 7·26 (84) 6·22 (39) 0·17 (2) 0·02 (3) 15·21 (43) 20·91 (31) 0·31 (7) 0·01 (1) 101·80 0·813 opx 9 52·97 (54) 0·17 (4) 0·13 (6) 6·12 (72) 13·06 (29) 0·24 (2) 0·01 (2) 27·32 (49) 1·22 (4) 0·02 (1) 0·00 101·28 0·788 amp 9 43·27 (69) 1·09 (20) 0·21 (8) 15·67 (93) 7·74 (48) 0·11 (2) 0·05 (2) 15·94 (49) 11·12 (57) 2·23 (14) 0·69 (9) 98·12 0·786 sp 8 0·00 0·14 (2) 1·68 (19) 65·05 (57) 15·12 (25) 0·15 (1) 0·13 (3) 18·61 (28) 0·24 (2) 0·00 0·00 101·12 0·687 Run # . T (°C) . Phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . Total . xMg . EQ Mb AuPd PU910 1200 glass 8 44·41 (25) 0·77 (4) 0·08 (4) 12·11 (10) 9·08 (12) 0·19 (5) n.d. 13·11 (44) 10·65 (24) 1·17 (11) 0·41 (2) 92·00 0·720 ol 8 40·70 (05) 0·02 (1) 0·05 (3) 0·05 (3) 7·89 (15) 0·14 (2) n.d. 50·78 (39) 0·24 (11) 0·01 (1) 0·00 (00) 99·88 0·920 sp 4 0·74 (93) 0·53 (7) 36·88 (483) 14·67 (367) 28·67 (80) 0·00 n.d. 15·46 (80) 0·41 (16) 0·05 (1) 0·01 (1) 97·43 0·490 PU908 1200 glass 6 44·34 (29) 0·76 (4) 0·09 (5) 12·65 (26) 9·34 (17) 0·17 (4) n.d. 12·49 (16) 10·94 (7) 1·14 (10) 0·45 (1) 92·36 0·704 ol 8 40·42 (20) 0·02 (1) 0·04 (2) 0·05 (2) 9·24 (7) 0·15 (2) n.d. 50·08 (21) 0·25 (2) 0·01 (1) 0·00 100·24 0·906 sp 5 0·86 (35) 0·34 (2) 39·42 (309) 19·35 (157) 24·09 (62) 0·00 n.d. 15·51 (31) 0·4 (8) 0·03 (2) 0·01 (1) 100·01 0·534 PU926 1160 glass 6 44·94 (13) 0·81 (3) 0·06 (3) 14·41 (11) 8·89 (8) 0·18 (2) 0·02 (2) 10·06 (15) 11·60 (10) 1·12 (4) 0·48 (1) 92·57 0·669 ol 8 40·56 (24) 0·02 (1) 0·03 (3) 0·12 (3) 9·72 (15) 0·19 (3) 0·23 (3) 49·61 (20) 0·25 (1) 0·02 (2) 0·01 (1) 100·74 0·901 sp 6 0·23 (11) 0·36 (14) 37·24 (646) 15·50 (186) 29·18 (459) 0·31 (3) 0·12 (5) 13·89 (47) 0·26 (2) 0·01 (2) 0·01 (1) 97·12 0·459 PU899 1160 glass 8 45·02 (62) 0·87 (4) 0·06 (2) 14·12 (12) 10·09 (26) 0·20 (6) n.d. 10·41 (21) 11·50 (22) 1·34 (6) 0·50 (2) 94·11 0·648 ol 8 39·22 (105) 0·01 (1) 0·03 (1) 0·04 (2) 11·55 (18) 0·16 (4) n.d. 48·12 (31) 0·21 (2) 0·00 0·01 (1) 99·36 0·881 cpx 9 51·24 (44) 0·23 (2) 0·77 (10) 4·64 (31) 5·64 (22) 0·08 (2) n.d. 17·26 (31) 20·20 (27) 0·32 (4) 0·01 (1) 100·39 0·845 PU909 1120 glass 8 45·01 (32) 0·81 (5) 0·05 (4) 13·94 (23) 9·14 (11) 0·19 (3) n.d. 9·59 (16) 11·10 (10) 1·31 (7) 0·49 (2) 91·56 0·651 ol 6 39·70 (19) 0·02 (1) 0·02 (1) 0·03 (2) 9·36 (16) 0·21 (1) n.d. 49·88 (19) 0·20 (1) 0·02 (2) 0·00 99·43 0·905 cpx 9 50·91 (18) 0·24 (3) 0·78 (13) 4·30 (25) 5·12 (23) 0·09 (3) n.d. 17·13 (29) 21·27 (38) 0·27 (2) 0·01 (1) 100·13 0·856 sp 3 0·17 (7) 0·57 (11) 19·96 (490) 16·03 (221) 44·08 (464) 0·00 n.d. 13·22 (63) 0·21 (8) 0·01 (1) 0·01 (1) 94·27 0·348 PU1006 1100 glass 12 46·87 (32) 1·02 (5) 0·02 (2) 17·31 (54) 9·08 (25) 0·17 (2) n.d. 5·55 (78) 10·63 (51) 1·46 (8) 0·57 (4) 92·70 0·522 ol 10 40·61 (13) 0·02 (1) 0·03 (2) 0·06 (2) 12·10 (13) 0·18 (2) 0·17 (2) 47·51 (31) 0·27 (2) 0·00 0·01 (1) 100·96 0·875 cpx 9 50·62 (34) 0·35 (2) 0·57 (14) 5·46 (30) 5·79 (24) 0·14 (2) 0·04 (2) 16·55 (50) 20·41 (58) 0·28 (2) 0·01 (1) 100·22 0·836 sp 6 0·32 (21) 0·54 (6) 17·75 (42) 33·96 (174) 28·88 (80) 0·21 (1) 0·19 (2) 16·41 (30) 0·30 (11) 0·01 (1) 0·01 (1) 98·57 0·503 PU1003 1080 glass 6 45·29 (44) 0·98 (4) 0·02 (3) 17·10 (25) 8·14 (16) 0·18 (1) n.d. 4·40 (42) 11·24 (14) 1·36 (5) 0·49 (4) 89·20 0·490 ol 8 40·30 (22) 0·02 (1) 0·02 (2) 0·03 (2) 12·01 (25) 0·20 (2) 0·17 (2) 47·29 (39) 0·25 (3) 0·01 (1) 0·00 100·30 0·875 cpx 9 50·38 (70) 0·35 (13) 0·59 (24) 4·88 (95) 5·56 (66) 0·14 (3) 0·02 (2) 16·33 (60) 20·87 (80) 0·24 (3) 0·01 (1) 99·36 0·839 sp 6 0·14 (10) 0·60 (10) 27·37 (563) 23·79 (333) 30·51 (215) 0·25 (1) 0·16 (5) 14·05 (65) 0·27 (4) 0·01 (2) 0·01 (1) 97·15 0·451 PU1004 1060 glass 13 46·51 (55) 1·02 (8) 0·02 (2) 19·04 (56) 6·54 (59) 0·16 (2) n.d. 2·82 (52) 10·04 (49) 1·22 (10) 0·33 (5) 87·70 0·435 ol 8 40·00 (14) 0·02 (1) 0·01 (1) 0·03 (1) 13·90 (8) 0·22 (2) 0·17 (2) 46·08 (24) 0·20 (1) 0·00 0·01 (1) 100·63 0·855 cpx 10 49·81 (96) 0·48 (14) 0·20 (12) 5·89 (101) 6·64 (35) 0·17 (2) 0·02 (2) 16·06 (91) 19·86 (96) 0·27 (5) 0·01 (1) 99·39 0·812 sp 6 0·17 (18) 0·49 (5) 7·11 (118) 44·64 (164) 29·93 (83) 0·17 (2) 0·22 (3) 16·19 (67) 0·18 (2) 0·00 0·01 (1) 99·13 0·491 PU905 1040 glass 5 46·82 (25) 0·83 (1) 0·01 (1) 18·94 (28) 5·62 (27) 0·20 (2) n.d. 2·37 (1) 9·27 (30) 1·08 (12) 0·35 (2) 85·46 0·429 ol 3 39·04 (30) 0·00 0·02 (1) 0·05 (1) 13·47 (13) 0·26 (3) n.d. 46·76 (40) 0·16 (2) 0·00 0·01 (1) 99·77 0·861 cpx 10 48·21 (48) 0·57 (24) 0·23 (10) 6·44 (53) 6·78 (27) 0·18 (2) n.d. 15·50 (47) 20·48 (47) 0·28 (3) 0·01 (1) 98·67 0·803 High-Al opx 5 51·02 (37) 0·16 (3) 0·08 (1) 6·37 (27) 10·44 (21) 0·25 (1) n.d. 29·78 (24) 1·06 (5) 0·03 (1) 0·01 (1) 99·20 0·836 Low-Al opx 5 53·85 (19) 0·11 (1) 0·04 (3) 2·93 (16) 9·65 (12) 0·27 (3) n.d. 31·21 (20) 1·25 (7) 0·03 (0) 0·01 (1) 99·33 0·852 amph 10 40·95 (57) 1·12 (7) 0·17 (8) 14·29 (32) 8·69 (42) 0·13 (4) n.d. 16·22 (16) 11·62 (32) 2·03 (4) 0·64 (13) 95·85 0·769 sp 5 0·14 (03) 1·04 (31) 9·88 (722) 22·39 (158) 49·75 (600) 0·07 (8) n.d. 11·68 (25) 0·21 (7) 0·02 (2) 0·01 (0) 95·17 0·295 PU1005 1000 glass 20 48·51 (98) 0·74 (10) 0·01 (1) 19·68 (48) 5·23 (81) 0·16 (2) n.d. 1·81 (44) 8·82 (47) 0·90 (12) 0·43 (9) 86·29 0·382 ol 10 39·90 (22) 0·02 (1) 0·01 (2) 0·03 (2) 15·87 (11) 0·25 (3) 0·15 (2) 44·96 (34) 0·19 (2) 0·00 0·00 101·38 0·835 cpx 15 49·45 (67) 0·52 (9) 0·28 (9) 6·51 (72) 6·72 (23) 0·17 (2) 0·03 (2) 15·37 (42) 20·57 (53) 0·22 (2) 0·01 (1) 99·86 0·803 opx 10 51·85 (49) 0·18 (2) 0·22 (2) 6·55 (29) 11·30 (12) 0·24 (2) 0·06 (2) 29·15 (32) 1·18 (6) 0·02 (2) 0·00 100·74 0·821 amph 15 42·34 (49) 1·13 (18) 0·32 (15) 14·29 (24) 8·63 (74) 0·10 (2) 0·05 (2) 16·33 (54) 11·40 (15) 1·95 (11) 0·60 (15) 97·13 0·771 sp 6 0·06 (2) 0·49 (3) 5·73 (13) 45·13 (62) 31·56 (54) 0·20 (2) 0·23 (2) 15·61 (7) 0·16 (3) 0·00 0·00 99·16 0·469 PU906 980 glass 10 49·39 (63) 0·53 (8) 0·03 (2) 19·03 (38) 3·82 (51) 0·20 (7) n.d. 1·66 (58) 8·47 (23) 0·68 (9) 0·52 (5) 84·33 0·437 cpx 10 48·49 (66) 0·43 (9) 0·18 (9) 5·77 (68) 6·76 (24) 0·20 (3) n.d. 15·66 (61) 20·87 (45) 0·20 (2) 0·01 (1) 98·57 0·805 opx 6 50·43 (52) 0·20 (1) 0·18 (6) 6·74 (49) 11·77 (20) 0·26 (2) n.d. 28·14 (24) 1·25 (15) 0·03 (5) 0·01 (2) 99·01 0·810 amph 6 41·73 (45) 0·92 (6) 0·26 (13) 13·61 (61) 8·35 (18) 0·13 (2) n.d. 16·43 (25) 11·77 (15) 1·62 (3) 0·62 (5) 95·43 0·778 sp 6 0·16 (07) 1·53 (13) 3·34 (246) 15·37 (46) 64·18 (255) 0·18 (11) n.d. 9·09 (35) 0·26 (4) 0·00 0·01 (1) 94·11 0·201 EQ Mb Pt–C PU72 1350 glass 5 46·19 (28) 0·67 (4) 0·15 (1) 12·80 (07) 8·14 (20) 0·15 (4) n.d. 17·06 (14) 9·93 (8) 1·19 (3) 0·40 (1) 96·68 0·789 ZP1330 1330 glass 10 46·90 (34) 0·74 (6) 0·21 (3) 12·70 (17) 9·04 (20) 0·17 (2) n.d. 16·05 (21) 10·13 (7) 1·19 (4) 0·42 (2) 97·56 0·760 ol 11 41·45 (19) 0·01 (1) 0·10 (2) 0·00 9·20 (18) 0·13 (1) 0·07 (4) 49·73 (20) 0·25 (1) 0·01 (1) 0·00 100·95 0·906 ZP1301 1300 glass 10 46·58 (39) 0·74 (6) 0·21 (3) 13·35 (27) 8·76 (21) 0·17 (2) 0·01 (2) 13·93 (28) 10·82 (13) 1·27 (4) 0·43 (1) 96·27 0·739 ol 10 41·23 (25) 0·01 (2) 0·15 (1) 0·00 10·21 (22) 0·15 (2) 0·14 (3) 49·20 (31) 0·29 (2) 0·01 (1) 0·00 101·39 0·896 ZP1302 1300 glass 10 46·93 (52) 0·74 (6) 0·20 (3) 13·00 (31) 8·84 (35) 0·16 (3) n.d. 15·12 (10) 10·33 (5) 1·20 (4) 0·44 (1) 96·96 0·753 ol 10 41·69 (19) 0·01 (1) 0·09 (2) 0·00 9·51 (22) 0·14 (2) 0·06 (5) 49·91 (15) 0·23 (2) 0·02 (1) 0·00 101·65 0·903 ZP1270 1270 glass 10 46·52 (27) 0·80 (4) 0·18 (2) 13·40 (38) 8·91 (26) 0·15 (3) 0·00 14·11 (60) 10·72 (24) 1·22 (3) 0·42 (1) 96·44 0·738 ol 10 41·09 (32) 0·00 0·12 (1) 0·00 10·22 (34) 0·13 (1) 0·04 (4) 49·17 (62) 0·24 (2) 0·01 (1) 0·00 101·02 0·896 ZP1240 1240 glass 9 46·95 (27) 0·78 (8) 0·19 (2) 14·08 (41) 8·84 (28) 0·15 (2) 0·01 (2) 12·81 (15) 11·11 (9) 1·32 (4) 0·47 (1) 96·70 0·721 ol 11 41·31 (42) 0·00 0·12 (2) 0·00 11·23 (27) 0·15 (2) 0·05 (4) 48·64 (72) 0·24 (2) 0·01 (1) 0·01 (1) 101·75 0·885 ZP1210 1210 glass 12 47·67 (38) 0·85 (5) 0·16 (2) 15·17 (45) 8·49 (23) 0·17 (3) 0·00 10·71 (39) 12·00 (13) 1·37 (6) 0·49 (2) 97·09 0·692 ol 10 41·18 (38) 0·00 0·13 (3) 0·00 12·37 (37) 0·16 (2) 0·04 (3) 48·16 (20) 0·28 (3) 0·01 (1) 0·00 102·32 0·874 sp 10 0·19 (29) 0·33 (3) 32·81 (86) 34·89 (61) 14·49 (31) 0·17 (2) 0·04 (2) 16·30 (19) 0·11 (4) 0·00 0·00 99·34 0·667 ZP1180 1180 glass 10 47·44 (71) 0·92 (8) 0·10 (4) 16·31 (56) 8·31 (34) 0·17 (2) 0·01 (2) 9·15 (52) 12·00 (24) 1·48 (6) 0·52 (2) 96·41 0·662 ol 8 40·61 (26) 0·00 0·05 (4) 0·00 13·44 (22) 0·19 (2) 0·03 (4) 46·45 (10) 0·26 (2) 0·00 0·00 101·03 0·860 High-Al cpx 9 50·56 (45) 0·40 (4) 1·28 (17) 5·76 (51) 4·20 (31) 0·13 (1) 0·00 16·46 (10) 20·76 (14) 0·25 (3) 0·00 99·81 0·875 Low-Al cpx 7 52·70 (66) 0·29 (4) 0·83 (6) 2·26 (27) 4·45 (38) 0·16 (1) 0·00 18·22 (24) 20·01 (60) 0·16 (3) 0·00 99·08 0·880 sp 8 0·00 0·38 (4) 32·60 (273) 33·77 (224) 15·69 (51) 0·18 (3) 0·00 15·45 (36) 0·16 (9) 0·00 0·00 98·24 0·637 ZP1151 1150 glass 10 47·75 (36) 0·94 (6) 0·05 (4) 16·22 (30) 8·19 (26) 0·17 (2) 0·00 9·39 (11) 11·40 (8) 1·43 (8) 0·53 (1) 96·08 0·672 ol 16 40·76 (73) 0·01 (2) 0·06 (3) 0·02 (02) 13·58 (30) 0·17 (3) 0·02 (2) 46·21 (87) 0·34 (12) 0·01 (1) 0·00 101·17 0·858 High-Al cpx 8 51·00 (49) 0·48 (10) 1·19 (12) 7·07 (68) 4·43 (49) 0·12 (1) 0·00 16·34 (43) 20·43 (25) 0·29 (2) 0·00 101·35 0·868 Low-Al cpx 8 53·20 (59) 0·23 (2) 1·06 (17) 4·08 (40) 4·42 (41) 0·12 (3) 0·00 17·97 (57) 19·89 (62) 0·26 (2) 0·00 101·24 0·879 sp 10 0·26 (33) 0·37 (9) 27·47 (260) 36·27 (298) 15·77 (49) 0·18 (4) 0·00 15·87 (52) 0·28 (15) 0·00 0·00 96·49 0·642 ZP1152 1150 glass 12 46·86 (96) 1·01 (9) 0·08 (3) 17·44 (43) 7·82 (35) 0·24 (4) 0·01 (3) 8·04 (43) 10·75 (51) 2·10 (11) 0·53 (4) 94·88 0·647 ol 8 40·70 (32) 0·00 0·08 (1) 0·00 13·33 (57) 0·28 (4) 0·11 (3) 46·31 (47) 0·26 (5) 0·00 0·00 101·07 0·861 cpx 10 50·97 (78) 0·48 (8) 1·06 (8) 6·49 (94) 4·86 (30) 0·20 (4) 0·00 16·30 (48) 19·99 (17) 0·34 (1) 0·01 (1) 100·68 0·857 sp 11 0·00 0·24 (3) 21·65 (119) 45·64 (118) 14·04 (55) 0·23 (5) 0·10 (6) 17·86 (40) 0·09 (6) 0·00 0·01 (1) 100·18 0·694 ZP1121 1120 glass 12 46·82 (125) 0·84 (6) 0·04 (3) 17·73 (38) 7·38 (49) 0·13 (2) 0·00 8·17 (28) 10·92 (36) 2·79 (12) 0·58 (5) 95·40 0·664 ol 12 41·40 (53) 0·01 (1) 0·05 (3) 0·00 14·05 (61) 0·17 (3) 0·07 (3) 45·95 (46) 0·24 (3) 0·02 (3) 0·01 (1) 101·97 0·854 cpx 10 51·76 (71) 0·52 (10) 0·34 (12) 7·81 (81) 5·54 (39) 0·14 (3) 0·02 (2) 16·50 (38) 18·89 (45) 0·43 (8) 0·03 (1) 101·97 0·842 High-Al opx 4 53·97 (38) 0·15 (4) 0·30 (9) 6·62 (22) 7·89 (32) 0·16 (2) 0·00 29·89 (53) 1·64 (2) 0·13 (1) 0·00 100·75 0·871 Low-Al opx 7 55·71 (53) 0·12 (3) 0·27 (7) 4·00 (85) 8·26 (29) 0·14 (2) 0·00 30·68 (45) 1·71 (10) 0·13 (3) 0·00 101·02 0·869 sp 8 0·97 (49) 0·14 (2) 9·73 (119) 56·32 (102) 12·18 (26) 0·12 (2) 0·06 (3) 19·76 (60) 0·16 (6) 0·00 0·00 99·44 0·743 ZP1122 1120 glass 10 46·51 (65) 0·98 (6) 0·00 17·91 (38) 6·86 (20) 0·13 (3) 0·00 8·66 (20) 11·05 (14) 1·94 (4) 0·74 (2) 94·77 0·692 ol 8 41·63 (47) 0·01 (1) 0·01 (1) 0·00 12·80 (23) 0·14 (4) 0·02 (7) 47·65 (57) 0·28 (4) 0·00 0·00 102·59 0·869 cpx 12 52·26 (62) 0·40 (8) 0·53 (19) 6·43 (58) 4·65 (38) 0·16 (4) 0·02 (4) 16·99 (47) 20·08 (58) 0·28 (3) 0·00 101·78 0·867 High-Al opx 9 53·27 (52) 0·21 (5) 0·39 (7) 7·78 (32) 8·69 (73) 0·13 (3) 0·03 (5) 29·50 (67) 1·72 (2) 0·04 (1) 0·00 101·76 0·858 Low-Al opx 4 56·62 (58) 0·12 (1) 0·29 (4) 3·38 (10) 8·81 (30) 0·15 (2) 0·02 (4) 31·12 (31) 1·86 (6) 0·03 (1) 0·00 102·40 0·863 ZP1090 1090 glass 8 49·12 (24) 1·09 (9) 0·01 (1) 18·33 (89) 8·63 (48) 0·17 (2) 0·01 (2) 7·01 (52) 10·29 (67) 2·31 (12) 0·86 (9) 97·83 0·592 ol 7 42·41 (28) 0·00 0·03 (2) 0·00 17·48 (45) 0·27 (2) 0·15 (2) 44·77 (48) 0·26 (1) 0·00 0·01 (1) 105·38 0·820 cpx 9 50·87 (38) 0·51 (4) 0·41 (8) 6·95 (43) 5·52 (17) 0·15 (3) 0·00 16·29 (13) 20·05 (23) 0·33 (2) 0·01 (1) 101·10 0·840 High-Al opx 5 52·27 (33) 0·30 (3) 0·03 (3) 8·23 (48) 11·01 (13) 0·25 (2) 0·00 27·92 (10) 1·72 (18) 0·05 (1) 0·01 (1) 101·79 0·819 Low-Al opx 5 55·25 (7) 0·14 (2) 0·03 (3) 3·70 (12) 10·99 (24) 0·25 (4) 0·00 29·68 (19) 1·67 (1) 0·05 (2) 0·01 (1) 101·76 0·828 sp 8 0·00 0·18 (4) 4·43 (291) 65·62 (298) 13·90 (54) 0·13 (4) 0·09 (3) 19·74 (50) 0·17 (6) 0·01 (2) 0·02 (1) 104·30 0·717 ZP1060 1060 glass 13 46·87 (148) 1·07 (18) 0·02 (3) 19·33 (52) 7·51 (40) 0·11 (3) 0·00 4·85 (64) 7·38 (37) 3·63 (30) 1·55 (9) 92·29 0·535 ol 4 39·12 (21) 0·05 (1) 0·02 (2) 0·00 20·11 (50) 0·23 (2) 0·11 (3) 39·55 (40) 0·28 (3) 0·00 0·00 99·48 0·778 cpx 10 51·16 (72) 0·81 (6) 0·15 (7) 7·39 (20) 5·82 (26) 0·14 (2) 0·02 (2) 15·79 (64) 20·21 (32) 0·41 (4) 0·02 (1) 101·93 0·829 opx 9 53·57 (74) 0·28 (4) 0·11 (4) 6·46 (64) 11·91 (57) 0·20 (3) 0·01 (2) 27·78 (50) 1·54 (22) 0·05 (4) 0·01 (2) 101·93 0·806 amp 5 42·28 (74) 1·94 (15) 0·15 (1) 15·32 (40) 7·51 (33) 0·11 (1) 0·03 (2) 15·61 (34) 11·15 (72) 2·40 (17) 0·71 (4) 97·20 0·787 sp 8 0·00 0·20 (4) 1·64 (134) 60·86 (168) 15·25 (33) 0·11 (2) 0·14 (3) 17·60 (42) 0·20 (2) 0·00 0·00 96·00 0·673 ZP1000 1000 glass 8 44·67 (55) 0·92 (7) 0·00 20·72 (13) 6·50 (09) 0·06 (2) 0·00 3·98 (41) 9·01 (15) 3·12 (20) 1·24 (6) 90·20 0·521 cpx 9 50·81 (82) 0·70 (20) 0·16 (4) 7·26 (84) 6·22 (39) 0·17 (2) 0·02 (3) 15·21 (43) 20·91 (31) 0·31 (7) 0·01 (1) 101·80 0·813 opx 9 52·97 (54) 0·17 (4) 0·13 (6) 6·12 (72) 13·06 (29) 0·24 (2) 0·01 (2) 27·32 (49) 1·22 (4) 0·02 (1) 0·00 101·28 0·788 amp 9 43·27 (69) 1·09 (20) 0·21 (8) 15·67 (93) 7·74 (48) 0·11 (2) 0·05 (2) 15·94 (49) 11·12 (57) 2·23 (14) 0·69 (9) 98·12 0·786 sp 8 0·00 0·14 (2) 1·68 (19) 65·05 (57) 15·12 (25) 0·15 (1) 0·13 (3) 18·61 (28) 0·24 (2) 0·00 0·00 101·12 0·687 All Fe as FeO. Units in parentheses indicate standard deviations (2σ) from averaged analysis [i.e. 44·41(25) should be read as 44·41 ± 0·25 wt %]; #, number of individual spots averaged for each phase composition reported. Abbreviations are as for Table 2; xMg = molar [MgO/(MgO + FeOtot)]. Open in new tab Table 3: Electron microprobe analyses of experimental phases of equilibrium crystallization experiments on composition RC158c Run # . T (°C) . Phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . Total . xMg . EQ Mb AuPd PU910 1200 glass 8 44·41 (25) 0·77 (4) 0·08 (4) 12·11 (10) 9·08 (12) 0·19 (5) n.d. 13·11 (44) 10·65 (24) 1·17 (11) 0·41 (2) 92·00 0·720 ol 8 40·70 (05) 0·02 (1) 0·05 (3) 0·05 (3) 7·89 (15) 0·14 (2) n.d. 50·78 (39) 0·24 (11) 0·01 (1) 0·00 (00) 99·88 0·920 sp 4 0·74 (93) 0·53 (7) 36·88 (483) 14·67 (367) 28·67 (80) 0·00 n.d. 15·46 (80) 0·41 (16) 0·05 (1) 0·01 (1) 97·43 0·490 PU908 1200 glass 6 44·34 (29) 0·76 (4) 0·09 (5) 12·65 (26) 9·34 (17) 0·17 (4) n.d. 12·49 (16) 10·94 (7) 1·14 (10) 0·45 (1) 92·36 0·704 ol 8 40·42 (20) 0·02 (1) 0·04 (2) 0·05 (2) 9·24 (7) 0·15 (2) n.d. 50·08 (21) 0·25 (2) 0·01 (1) 0·00 100·24 0·906 sp 5 0·86 (35) 0·34 (2) 39·42 (309) 19·35 (157) 24·09 (62) 0·00 n.d. 15·51 (31) 0·4 (8) 0·03 (2) 0·01 (1) 100·01 0·534 PU926 1160 glass 6 44·94 (13) 0·81 (3) 0·06 (3) 14·41 (11) 8·89 (8) 0·18 (2) 0·02 (2) 10·06 (15) 11·60 (10) 1·12 (4) 0·48 (1) 92·57 0·669 ol 8 40·56 (24) 0·02 (1) 0·03 (3) 0·12 (3) 9·72 (15) 0·19 (3) 0·23 (3) 49·61 (20) 0·25 (1) 0·02 (2) 0·01 (1) 100·74 0·901 sp 6 0·23 (11) 0·36 (14) 37·24 (646) 15·50 (186) 29·18 (459) 0·31 (3) 0·12 (5) 13·89 (47) 0·26 (2) 0·01 (2) 0·01 (1) 97·12 0·459 PU899 1160 glass 8 45·02 (62) 0·87 (4) 0·06 (2) 14·12 (12) 10·09 (26) 0·20 (6) n.d. 10·41 (21) 11·50 (22) 1·34 (6) 0·50 (2) 94·11 0·648 ol 8 39·22 (105) 0·01 (1) 0·03 (1) 0·04 (2) 11·55 (18) 0·16 (4) n.d. 48·12 (31) 0·21 (2) 0·00 0·01 (1) 99·36 0·881 cpx 9 51·24 (44) 0·23 (2) 0·77 (10) 4·64 (31) 5·64 (22) 0·08 (2) n.d. 17·26 (31) 20·20 (27) 0·32 (4) 0·01 (1) 100·39 0·845 PU909 1120 glass 8 45·01 (32) 0·81 (5) 0·05 (4) 13·94 (23) 9·14 (11) 0·19 (3) n.d. 9·59 (16) 11·10 (10) 1·31 (7) 0·49 (2) 91·56 0·651 ol 6 39·70 (19) 0·02 (1) 0·02 (1) 0·03 (2) 9·36 (16) 0·21 (1) n.d. 49·88 (19) 0·20 (1) 0·02 (2) 0·00 99·43 0·905 cpx 9 50·91 (18) 0·24 (3) 0·78 (13) 4·30 (25) 5·12 (23) 0·09 (3) n.d. 17·13 (29) 21·27 (38) 0·27 (2) 0·01 (1) 100·13 0·856 sp 3 0·17 (7) 0·57 (11) 19·96 (490) 16·03 (221) 44·08 (464) 0·00 n.d. 13·22 (63) 0·21 (8) 0·01 (1) 0·01 (1) 94·27 0·348 PU1006 1100 glass 12 46·87 (32) 1·02 (5) 0·02 (2) 17·31 (54) 9·08 (25) 0·17 (2) n.d. 5·55 (78) 10·63 (51) 1·46 (8) 0·57 (4) 92·70 0·522 ol 10 40·61 (13) 0·02 (1) 0·03 (2) 0·06 (2) 12·10 (13) 0·18 (2) 0·17 (2) 47·51 (31) 0·27 (2) 0·00 0·01 (1) 100·96 0·875 cpx 9 50·62 (34) 0·35 (2) 0·57 (14) 5·46 (30) 5·79 (24) 0·14 (2) 0·04 (2) 16·55 (50) 20·41 (58) 0·28 (2) 0·01 (1) 100·22 0·836 sp 6 0·32 (21) 0·54 (6) 17·75 (42) 33·96 (174) 28·88 (80) 0·21 (1) 0·19 (2) 16·41 (30) 0·30 (11) 0·01 (1) 0·01 (1) 98·57 0·503 PU1003 1080 glass 6 45·29 (44) 0·98 (4) 0·02 (3) 17·10 (25) 8·14 (16) 0·18 (1) n.d. 4·40 (42) 11·24 (14) 1·36 (5) 0·49 (4) 89·20 0·490 ol 8 40·30 (22) 0·02 (1) 0·02 (2) 0·03 (2) 12·01 (25) 0·20 (2) 0·17 (2) 47·29 (39) 0·25 (3) 0·01 (1) 0·00 100·30 0·875 cpx 9 50·38 (70) 0·35 (13) 0·59 (24) 4·88 (95) 5·56 (66) 0·14 (3) 0·02 (2) 16·33 (60) 20·87 (80) 0·24 (3) 0·01 (1) 99·36 0·839 sp 6 0·14 (10) 0·60 (10) 27·37 (563) 23·79 (333) 30·51 (215) 0·25 (1) 0·16 (5) 14·05 (65) 0·27 (4) 0·01 (2) 0·01 (1) 97·15 0·451 PU1004 1060 glass 13 46·51 (55) 1·02 (8) 0·02 (2) 19·04 (56) 6·54 (59) 0·16 (2) n.d. 2·82 (52) 10·04 (49) 1·22 (10) 0·33 (5) 87·70 0·435 ol 8 40·00 (14) 0·02 (1) 0·01 (1) 0·03 (1) 13·90 (8) 0·22 (2) 0·17 (2) 46·08 (24) 0·20 (1) 0·00 0·01 (1) 100·63 0·855 cpx 10 49·81 (96) 0·48 (14) 0·20 (12) 5·89 (101) 6·64 (35) 0·17 (2) 0·02 (2) 16·06 (91) 19·86 (96) 0·27 (5) 0·01 (1) 99·39 0·812 sp 6 0·17 (18) 0·49 (5) 7·11 (118) 44·64 (164) 29·93 (83) 0·17 (2) 0·22 (3) 16·19 (67) 0·18 (2) 0·00 0·01 (1) 99·13 0·491 PU905 1040 glass 5 46·82 (25) 0·83 (1) 0·01 (1) 18·94 (28) 5·62 (27) 0·20 (2) n.d. 2·37 (1) 9·27 (30) 1·08 (12) 0·35 (2) 85·46 0·429 ol 3 39·04 (30) 0·00 0·02 (1) 0·05 (1) 13·47 (13) 0·26 (3) n.d. 46·76 (40) 0·16 (2) 0·00 0·01 (1) 99·77 0·861 cpx 10 48·21 (48) 0·57 (24) 0·23 (10) 6·44 (53) 6·78 (27) 0·18 (2) n.d. 15·50 (47) 20·48 (47) 0·28 (3) 0·01 (1) 98·67 0·803 High-Al opx 5 51·02 (37) 0·16 (3) 0·08 (1) 6·37 (27) 10·44 (21) 0·25 (1) n.d. 29·78 (24) 1·06 (5) 0·03 (1) 0·01 (1) 99·20 0·836 Low-Al opx 5 53·85 (19) 0·11 (1) 0·04 (3) 2·93 (16) 9·65 (12) 0·27 (3) n.d. 31·21 (20) 1·25 (7) 0·03 (0) 0·01 (1) 99·33 0·852 amph 10 40·95 (57) 1·12 (7) 0·17 (8) 14·29 (32) 8·69 (42) 0·13 (4) n.d. 16·22 (16) 11·62 (32) 2·03 (4) 0·64 (13) 95·85 0·769 sp 5 0·14 (03) 1·04 (31) 9·88 (722) 22·39 (158) 49·75 (600) 0·07 (8) n.d. 11·68 (25) 0·21 (7) 0·02 (2) 0·01 (0) 95·17 0·295 PU1005 1000 glass 20 48·51 (98) 0·74 (10) 0·01 (1) 19·68 (48) 5·23 (81) 0·16 (2) n.d. 1·81 (44) 8·82 (47) 0·90 (12) 0·43 (9) 86·29 0·382 ol 10 39·90 (22) 0·02 (1) 0·01 (2) 0·03 (2) 15·87 (11) 0·25 (3) 0·15 (2) 44·96 (34) 0·19 (2) 0·00 0·00 101·38 0·835 cpx 15 49·45 (67) 0·52 (9) 0·28 (9) 6·51 (72) 6·72 (23) 0·17 (2) 0·03 (2) 15·37 (42) 20·57 (53) 0·22 (2) 0·01 (1) 99·86 0·803 opx 10 51·85 (49) 0·18 (2) 0·22 (2) 6·55 (29) 11·30 (12) 0·24 (2) 0·06 (2) 29·15 (32) 1·18 (6) 0·02 (2) 0·00 100·74 0·821 amph 15 42·34 (49) 1·13 (18) 0·32 (15) 14·29 (24) 8·63 (74) 0·10 (2) 0·05 (2) 16·33 (54) 11·40 (15) 1·95 (11) 0·60 (15) 97·13 0·771 sp 6 0·06 (2) 0·49 (3) 5·73 (13) 45·13 (62) 31·56 (54) 0·20 (2) 0·23 (2) 15·61 (7) 0·16 (3) 0·00 0·00 99·16 0·469 PU906 980 glass 10 49·39 (63) 0·53 (8) 0·03 (2) 19·03 (38) 3·82 (51) 0·20 (7) n.d. 1·66 (58) 8·47 (23) 0·68 (9) 0·52 (5) 84·33 0·437 cpx 10 48·49 (66) 0·43 (9) 0·18 (9) 5·77 (68) 6·76 (24) 0·20 (3) n.d. 15·66 (61) 20·87 (45) 0·20 (2) 0·01 (1) 98·57 0·805 opx 6 50·43 (52) 0·20 (1) 0·18 (6) 6·74 (49) 11·77 (20) 0·26 (2) n.d. 28·14 (24) 1·25 (15) 0·03 (5) 0·01 (2) 99·01 0·810 amph 6 41·73 (45) 0·92 (6) 0·26 (13) 13·61 (61) 8·35 (18) 0·13 (2) n.d. 16·43 (25) 11·77 (15) 1·62 (3) 0·62 (5) 95·43 0·778 sp 6 0·16 (07) 1·53 (13) 3·34 (246) 15·37 (46) 64·18 (255) 0·18 (11) n.d. 9·09 (35) 0·26 (4) 0·00 0·01 (1) 94·11 0·201 EQ Mb Pt–C PU72 1350 glass 5 46·19 (28) 0·67 (4) 0·15 (1) 12·80 (07) 8·14 (20) 0·15 (4) n.d. 17·06 (14) 9·93 (8) 1·19 (3) 0·40 (1) 96·68 0·789 ZP1330 1330 glass 10 46·90 (34) 0·74 (6) 0·21 (3) 12·70 (17) 9·04 (20) 0·17 (2) n.d. 16·05 (21) 10·13 (7) 1·19 (4) 0·42 (2) 97·56 0·760 ol 11 41·45 (19) 0·01 (1) 0·10 (2) 0·00 9·20 (18) 0·13 (1) 0·07 (4) 49·73 (20) 0·25 (1) 0·01 (1) 0·00 100·95 0·906 ZP1301 1300 glass 10 46·58 (39) 0·74 (6) 0·21 (3) 13·35 (27) 8·76 (21) 0·17 (2) 0·01 (2) 13·93 (28) 10·82 (13) 1·27 (4) 0·43 (1) 96·27 0·739 ol 10 41·23 (25) 0·01 (2) 0·15 (1) 0·00 10·21 (22) 0·15 (2) 0·14 (3) 49·20 (31) 0·29 (2) 0·01 (1) 0·00 101·39 0·896 ZP1302 1300 glass 10 46·93 (52) 0·74 (6) 0·20 (3) 13·00 (31) 8·84 (35) 0·16 (3) n.d. 15·12 (10) 10·33 (5) 1·20 (4) 0·44 (1) 96·96 0·753 ol 10 41·69 (19) 0·01 (1) 0·09 (2) 0·00 9·51 (22) 0·14 (2) 0·06 (5) 49·91 (15) 0·23 (2) 0·02 (1) 0·00 101·65 0·903 ZP1270 1270 glass 10 46·52 (27) 0·80 (4) 0·18 (2) 13·40 (38) 8·91 (26) 0·15 (3) 0·00 14·11 (60) 10·72 (24) 1·22 (3) 0·42 (1) 96·44 0·738 ol 10 41·09 (32) 0·00 0·12 (1) 0·00 10·22 (34) 0·13 (1) 0·04 (4) 49·17 (62) 0·24 (2) 0·01 (1) 0·00 101·02 0·896 ZP1240 1240 glass 9 46·95 (27) 0·78 (8) 0·19 (2) 14·08 (41) 8·84 (28) 0·15 (2) 0·01 (2) 12·81 (15) 11·11 (9) 1·32 (4) 0·47 (1) 96·70 0·721 ol 11 41·31 (42) 0·00 0·12 (2) 0·00 11·23 (27) 0·15 (2) 0·05 (4) 48·64 (72) 0·24 (2) 0·01 (1) 0·01 (1) 101·75 0·885 ZP1210 1210 glass 12 47·67 (38) 0·85 (5) 0·16 (2) 15·17 (45) 8·49 (23) 0·17 (3) 0·00 10·71 (39) 12·00 (13) 1·37 (6) 0·49 (2) 97·09 0·692 ol 10 41·18 (38) 0·00 0·13 (3) 0·00 12·37 (37) 0·16 (2) 0·04 (3) 48·16 (20) 0·28 (3) 0·01 (1) 0·00 102·32 0·874 sp 10 0·19 (29) 0·33 (3) 32·81 (86) 34·89 (61) 14·49 (31) 0·17 (2) 0·04 (2) 16·30 (19) 0·11 (4) 0·00 0·00 99·34 0·667 ZP1180 1180 glass 10 47·44 (71) 0·92 (8) 0·10 (4) 16·31 (56) 8·31 (34) 0·17 (2) 0·01 (2) 9·15 (52) 12·00 (24) 1·48 (6) 0·52 (2) 96·41 0·662 ol 8 40·61 (26) 0·00 0·05 (4) 0·00 13·44 (22) 0·19 (2) 0·03 (4) 46·45 (10) 0·26 (2) 0·00 0·00 101·03 0·860 High-Al cpx 9 50·56 (45) 0·40 (4) 1·28 (17) 5·76 (51) 4·20 (31) 0·13 (1) 0·00 16·46 (10) 20·76 (14) 0·25 (3) 0·00 99·81 0·875 Low-Al cpx 7 52·70 (66) 0·29 (4) 0·83 (6) 2·26 (27) 4·45 (38) 0·16 (1) 0·00 18·22 (24) 20·01 (60) 0·16 (3) 0·00 99·08 0·880 sp 8 0·00 0·38 (4) 32·60 (273) 33·77 (224) 15·69 (51) 0·18 (3) 0·00 15·45 (36) 0·16 (9) 0·00 0·00 98·24 0·637 ZP1151 1150 glass 10 47·75 (36) 0·94 (6) 0·05 (4) 16·22 (30) 8·19 (26) 0·17 (2) 0·00 9·39 (11) 11·40 (8) 1·43 (8) 0·53 (1) 96·08 0·672 ol 16 40·76 (73) 0·01 (2) 0·06 (3) 0·02 (02) 13·58 (30) 0·17 (3) 0·02 (2) 46·21 (87) 0·34 (12) 0·01 (1) 0·00 101·17 0·858 High-Al cpx 8 51·00 (49) 0·48 (10) 1·19 (12) 7·07 (68) 4·43 (49) 0·12 (1) 0·00 16·34 (43) 20·43 (25) 0·29 (2) 0·00 101·35 0·868 Low-Al cpx 8 53·20 (59) 0·23 (2) 1·06 (17) 4·08 (40) 4·42 (41) 0·12 (3) 0·00 17·97 (57) 19·89 (62) 0·26 (2) 0·00 101·24 0·879 sp 10 0·26 (33) 0·37 (9) 27·47 (260) 36·27 (298) 15·77 (49) 0·18 (4) 0·00 15·87 (52) 0·28 (15) 0·00 0·00 96·49 0·642 ZP1152 1150 glass 12 46·86 (96) 1·01 (9) 0·08 (3) 17·44 (43) 7·82 (35) 0·24 (4) 0·01 (3) 8·04 (43) 10·75 (51) 2·10 (11) 0·53 (4) 94·88 0·647 ol 8 40·70 (32) 0·00 0·08 (1) 0·00 13·33 (57) 0·28 (4) 0·11 (3) 46·31 (47) 0·26 (5) 0·00 0·00 101·07 0·861 cpx 10 50·97 (78) 0·48 (8) 1·06 (8) 6·49 (94) 4·86 (30) 0·20 (4) 0·00 16·30 (48) 19·99 (17) 0·34 (1) 0·01 (1) 100·68 0·857 sp 11 0·00 0·24 (3) 21·65 (119) 45·64 (118) 14·04 (55) 0·23 (5) 0·10 (6) 17·86 (40) 0·09 (6) 0·00 0·01 (1) 100·18 0·694 ZP1121 1120 glass 12 46·82 (125) 0·84 (6) 0·04 (3) 17·73 (38) 7·38 (49) 0·13 (2) 0·00 8·17 (28) 10·92 (36) 2·79 (12) 0·58 (5) 95·40 0·664 ol 12 41·40 (53) 0·01 (1) 0·05 (3) 0·00 14·05 (61) 0·17 (3) 0·07 (3) 45·95 (46) 0·24 (3) 0·02 (3) 0·01 (1) 101·97 0·854 cpx 10 51·76 (71) 0·52 (10) 0·34 (12) 7·81 (81) 5·54 (39) 0·14 (3) 0·02 (2) 16·50 (38) 18·89 (45) 0·43 (8) 0·03 (1) 101·97 0·842 High-Al opx 4 53·97 (38) 0·15 (4) 0·30 (9) 6·62 (22) 7·89 (32) 0·16 (2) 0·00 29·89 (53) 1·64 (2) 0·13 (1) 0·00 100·75 0·871 Low-Al opx 7 55·71 (53) 0·12 (3) 0·27 (7) 4·00 (85) 8·26 (29) 0·14 (2) 0·00 30·68 (45) 1·71 (10) 0·13 (3) 0·00 101·02 0·869 sp 8 0·97 (49) 0·14 (2) 9·73 (119) 56·32 (102) 12·18 (26) 0·12 (2) 0·06 (3) 19·76 (60) 0·16 (6) 0·00 0·00 99·44 0·743 ZP1122 1120 glass 10 46·51 (65) 0·98 (6) 0·00 17·91 (38) 6·86 (20) 0·13 (3) 0·00 8·66 (20) 11·05 (14) 1·94 (4) 0·74 (2) 94·77 0·692 ol 8 41·63 (47) 0·01 (1) 0·01 (1) 0·00 12·80 (23) 0·14 (4) 0·02 (7) 47·65 (57) 0·28 (4) 0·00 0·00 102·59 0·869 cpx 12 52·26 (62) 0·40 (8) 0·53 (19) 6·43 (58) 4·65 (38) 0·16 (4) 0·02 (4) 16·99 (47) 20·08 (58) 0·28 (3) 0·00 101·78 0·867 High-Al opx 9 53·27 (52) 0·21 (5) 0·39 (7) 7·78 (32) 8·69 (73) 0·13 (3) 0·03 (5) 29·50 (67) 1·72 (2) 0·04 (1) 0·00 101·76 0·858 Low-Al opx 4 56·62 (58) 0·12 (1) 0·29 (4) 3·38 (10) 8·81 (30) 0·15 (2) 0·02 (4) 31·12 (31) 1·86 (6) 0·03 (1) 0·00 102·40 0·863 ZP1090 1090 glass 8 49·12 (24) 1·09 (9) 0·01 (1) 18·33 (89) 8·63 (48) 0·17 (2) 0·01 (2) 7·01 (52) 10·29 (67) 2·31 (12) 0·86 (9) 97·83 0·592 ol 7 42·41 (28) 0·00 0·03 (2) 0·00 17·48 (45) 0·27 (2) 0·15 (2) 44·77 (48) 0·26 (1) 0·00 0·01 (1) 105·38 0·820 cpx 9 50·87 (38) 0·51 (4) 0·41 (8) 6·95 (43) 5·52 (17) 0·15 (3) 0·00 16·29 (13) 20·05 (23) 0·33 (2) 0·01 (1) 101·10 0·840 High-Al opx 5 52·27 (33) 0·30 (3) 0·03 (3) 8·23 (48) 11·01 (13) 0·25 (2) 0·00 27·92 (10) 1·72 (18) 0·05 (1) 0·01 (1) 101·79 0·819 Low-Al opx 5 55·25 (7) 0·14 (2) 0·03 (3) 3·70 (12) 10·99 (24) 0·25 (4) 0·00 29·68 (19) 1·67 (1) 0·05 (2) 0·01 (1) 101·76 0·828 sp 8 0·00 0·18 (4) 4·43 (291) 65·62 (298) 13·90 (54) 0·13 (4) 0·09 (3) 19·74 (50) 0·17 (6) 0·01 (2) 0·02 (1) 104·30 0·717 ZP1060 1060 glass 13 46·87 (148) 1·07 (18) 0·02 (3) 19·33 (52) 7·51 (40) 0·11 (3) 0·00 4·85 (64) 7·38 (37) 3·63 (30) 1·55 (9) 92·29 0·535 ol 4 39·12 (21) 0·05 (1) 0·02 (2) 0·00 20·11 (50) 0·23 (2) 0·11 (3) 39·55 (40) 0·28 (3) 0·00 0·00 99·48 0·778 cpx 10 51·16 (72) 0·81 (6) 0·15 (7) 7·39 (20) 5·82 (26) 0·14 (2) 0·02 (2) 15·79 (64) 20·21 (32) 0·41 (4) 0·02 (1) 101·93 0·829 opx 9 53·57 (74) 0·28 (4) 0·11 (4) 6·46 (64) 11·91 (57) 0·20 (3) 0·01 (2) 27·78 (50) 1·54 (22) 0·05 (4) 0·01 (2) 101·93 0·806 amp 5 42·28 (74) 1·94 (15) 0·15 (1) 15·32 (40) 7·51 (33) 0·11 (1) 0·03 (2) 15·61 (34) 11·15 (72) 2·40 (17) 0·71 (4) 97·20 0·787 sp 8 0·00 0·20 (4) 1·64 (134) 60·86 (168) 15·25 (33) 0·11 (2) 0·14 (3) 17·60 (42) 0·20 (2) 0·00 0·00 96·00 0·673 ZP1000 1000 glass 8 44·67 (55) 0·92 (7) 0·00 20·72 (13) 6·50 (09) 0·06 (2) 0·00 3·98 (41) 9·01 (15) 3·12 (20) 1·24 (6) 90·20 0·521 cpx 9 50·81 (82) 0·70 (20) 0·16 (4) 7·26 (84) 6·22 (39) 0·17 (2) 0·02 (3) 15·21 (43) 20·91 (31) 0·31 (7) 0·01 (1) 101·80 0·813 opx 9 52·97 (54) 0·17 (4) 0·13 (6) 6·12 (72) 13·06 (29) 0·24 (2) 0·01 (2) 27·32 (49) 1·22 (4) 0·02 (1) 0·00 101·28 0·788 amp 9 43·27 (69) 1·09 (20) 0·21 (8) 15·67 (93) 7·74 (48) 0·11 (2) 0·05 (2) 15·94 (49) 11·12 (57) 2·23 (14) 0·69 (9) 98·12 0·786 sp 8 0·00 0·14 (2) 1·68 (19) 65·05 (57) 15·12 (25) 0·15 (1) 0·13 (3) 18·61 (28) 0·24 (2) 0·00 0·00 101·12 0·687 Run # . T (°C) . Phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . Total . xMg . EQ Mb AuPd PU910 1200 glass 8 44·41 (25) 0·77 (4) 0·08 (4) 12·11 (10) 9·08 (12) 0·19 (5) n.d. 13·11 (44) 10·65 (24) 1·17 (11) 0·41 (2) 92·00 0·720 ol 8 40·70 (05) 0·02 (1) 0·05 (3) 0·05 (3) 7·89 (15) 0·14 (2) n.d. 50·78 (39) 0·24 (11) 0·01 (1) 0·00 (00) 99·88 0·920 sp 4 0·74 (93) 0·53 (7) 36·88 (483) 14·67 (367) 28·67 (80) 0·00 n.d. 15·46 (80) 0·41 (16) 0·05 (1) 0·01 (1) 97·43 0·490 PU908 1200 glass 6 44·34 (29) 0·76 (4) 0·09 (5) 12·65 (26) 9·34 (17) 0·17 (4) n.d. 12·49 (16) 10·94 (7) 1·14 (10) 0·45 (1) 92·36 0·704 ol 8 40·42 (20) 0·02 (1) 0·04 (2) 0·05 (2) 9·24 (7) 0·15 (2) n.d. 50·08 (21) 0·25 (2) 0·01 (1) 0·00 100·24 0·906 sp 5 0·86 (35) 0·34 (2) 39·42 (309) 19·35 (157) 24·09 (62) 0·00 n.d. 15·51 (31) 0·4 (8) 0·03 (2) 0·01 (1) 100·01 0·534 PU926 1160 glass 6 44·94 (13) 0·81 (3) 0·06 (3) 14·41 (11) 8·89 (8) 0·18 (2) 0·02 (2) 10·06 (15) 11·60 (10) 1·12 (4) 0·48 (1) 92·57 0·669 ol 8 40·56 (24) 0·02 (1) 0·03 (3) 0·12 (3) 9·72 (15) 0·19 (3) 0·23 (3) 49·61 (20) 0·25 (1) 0·02 (2) 0·01 (1) 100·74 0·901 sp 6 0·23 (11) 0·36 (14) 37·24 (646) 15·50 (186) 29·18 (459) 0·31 (3) 0·12 (5) 13·89 (47) 0·26 (2) 0·01 (2) 0·01 (1) 97·12 0·459 PU899 1160 glass 8 45·02 (62) 0·87 (4) 0·06 (2) 14·12 (12) 10·09 (26) 0·20 (6) n.d. 10·41 (21) 11·50 (22) 1·34 (6) 0·50 (2) 94·11 0·648 ol 8 39·22 (105) 0·01 (1) 0·03 (1) 0·04 (2) 11·55 (18) 0·16 (4) n.d. 48·12 (31) 0·21 (2) 0·00 0·01 (1) 99·36 0·881 cpx 9 51·24 (44) 0·23 (2) 0·77 (10) 4·64 (31) 5·64 (22) 0·08 (2) n.d. 17·26 (31) 20·20 (27) 0·32 (4) 0·01 (1) 100·39 0·845 PU909 1120 glass 8 45·01 (32) 0·81 (5) 0·05 (4) 13·94 (23) 9·14 (11) 0·19 (3) n.d. 9·59 (16) 11·10 (10) 1·31 (7) 0·49 (2) 91·56 0·651 ol 6 39·70 (19) 0·02 (1) 0·02 (1) 0·03 (2) 9·36 (16) 0·21 (1) n.d. 49·88 (19) 0·20 (1) 0·02 (2) 0·00 99·43 0·905 cpx 9 50·91 (18) 0·24 (3) 0·78 (13) 4·30 (25) 5·12 (23) 0·09 (3) n.d. 17·13 (29) 21·27 (38) 0·27 (2) 0·01 (1) 100·13 0·856 sp 3 0·17 (7) 0·57 (11) 19·96 (490) 16·03 (221) 44·08 (464) 0·00 n.d. 13·22 (63) 0·21 (8) 0·01 (1) 0·01 (1) 94·27 0·348 PU1006 1100 glass 12 46·87 (32) 1·02 (5) 0·02 (2) 17·31 (54) 9·08 (25) 0·17 (2) n.d. 5·55 (78) 10·63 (51) 1·46 (8) 0·57 (4) 92·70 0·522 ol 10 40·61 (13) 0·02 (1) 0·03 (2) 0·06 (2) 12·10 (13) 0·18 (2) 0·17 (2) 47·51 (31) 0·27 (2) 0·00 0·01 (1) 100·96 0·875 cpx 9 50·62 (34) 0·35 (2) 0·57 (14) 5·46 (30) 5·79 (24) 0·14 (2) 0·04 (2) 16·55 (50) 20·41 (58) 0·28 (2) 0·01 (1) 100·22 0·836 sp 6 0·32 (21) 0·54 (6) 17·75 (42) 33·96 (174) 28·88 (80) 0·21 (1) 0·19 (2) 16·41 (30) 0·30 (11) 0·01 (1) 0·01 (1) 98·57 0·503 PU1003 1080 glass 6 45·29 (44) 0·98 (4) 0·02 (3) 17·10 (25) 8·14 (16) 0·18 (1) n.d. 4·40 (42) 11·24 (14) 1·36 (5) 0·49 (4) 89·20 0·490 ol 8 40·30 (22) 0·02 (1) 0·02 (2) 0·03 (2) 12·01 (25) 0·20 (2) 0·17 (2) 47·29 (39) 0·25 (3) 0·01 (1) 0·00 100·30 0·875 cpx 9 50·38 (70) 0·35 (13) 0·59 (24) 4·88 (95) 5·56 (66) 0·14 (3) 0·02 (2) 16·33 (60) 20·87 (80) 0·24 (3) 0·01 (1) 99·36 0·839 sp 6 0·14 (10) 0·60 (10) 27·37 (563) 23·79 (333) 30·51 (215) 0·25 (1) 0·16 (5) 14·05 (65) 0·27 (4) 0·01 (2) 0·01 (1) 97·15 0·451 PU1004 1060 glass 13 46·51 (55) 1·02 (8) 0·02 (2) 19·04 (56) 6·54 (59) 0·16 (2) n.d. 2·82 (52) 10·04 (49) 1·22 (10) 0·33 (5) 87·70 0·435 ol 8 40·00 (14) 0·02 (1) 0·01 (1) 0·03 (1) 13·90 (8) 0·22 (2) 0·17 (2) 46·08 (24) 0·20 (1) 0·00 0·01 (1) 100·63 0·855 cpx 10 49·81 (96) 0·48 (14) 0·20 (12) 5·89 (101) 6·64 (35) 0·17 (2) 0·02 (2) 16·06 (91) 19·86 (96) 0·27 (5) 0·01 (1) 99·39 0·812 sp 6 0·17 (18) 0·49 (5) 7·11 (118) 44·64 (164) 29·93 (83) 0·17 (2) 0·22 (3) 16·19 (67) 0·18 (2) 0·00 0·01 (1) 99·13 0·491 PU905 1040 glass 5 46·82 (25) 0·83 (1) 0·01 (1) 18·94 (28) 5·62 (27) 0·20 (2) n.d. 2·37 (1) 9·27 (30) 1·08 (12) 0·35 (2) 85·46 0·429 ol 3 39·04 (30) 0·00 0·02 (1) 0·05 (1) 13·47 (13) 0·26 (3) n.d. 46·76 (40) 0·16 (2) 0·00 0·01 (1) 99·77 0·861 cpx 10 48·21 (48) 0·57 (24) 0·23 (10) 6·44 (53) 6·78 (27) 0·18 (2) n.d. 15·50 (47) 20·48 (47) 0·28 (3) 0·01 (1) 98·67 0·803 High-Al opx 5 51·02 (37) 0·16 (3) 0·08 (1) 6·37 (27) 10·44 (21) 0·25 (1) n.d. 29·78 (24) 1·06 (5) 0·03 (1) 0·01 (1) 99·20 0·836 Low-Al opx 5 53·85 (19) 0·11 (1) 0·04 (3) 2·93 (16) 9·65 (12) 0·27 (3) n.d. 31·21 (20) 1·25 (7) 0·03 (0) 0·01 (1) 99·33 0·852 amph 10 40·95 (57) 1·12 (7) 0·17 (8) 14·29 (32) 8·69 (42) 0·13 (4) n.d. 16·22 (16) 11·62 (32) 2·03 (4) 0·64 (13) 95·85 0·769 sp 5 0·14 (03) 1·04 (31) 9·88 (722) 22·39 (158) 49·75 (600) 0·07 (8) n.d. 11·68 (25) 0·21 (7) 0·02 (2) 0·01 (0) 95·17 0·295 PU1005 1000 glass 20 48·51 (98) 0·74 (10) 0·01 (1) 19·68 (48) 5·23 (81) 0·16 (2) n.d. 1·81 (44) 8·82 (47) 0·90 (12) 0·43 (9) 86·29 0·382 ol 10 39·90 (22) 0·02 (1) 0·01 (2) 0·03 (2) 15·87 (11) 0·25 (3) 0·15 (2) 44·96 (34) 0·19 (2) 0·00 0·00 101·38 0·835 cpx 15 49·45 (67) 0·52 (9) 0·28 (9) 6·51 (72) 6·72 (23) 0·17 (2) 0·03 (2) 15·37 (42) 20·57 (53) 0·22 (2) 0·01 (1) 99·86 0·803 opx 10 51·85 (49) 0·18 (2) 0·22 (2) 6·55 (29) 11·30 (12) 0·24 (2) 0·06 (2) 29·15 (32) 1·18 (6) 0·02 (2) 0·00 100·74 0·821 amph 15 42·34 (49) 1·13 (18) 0·32 (15) 14·29 (24) 8·63 (74) 0·10 (2) 0·05 (2) 16·33 (54) 11·40 (15) 1·95 (11) 0·60 (15) 97·13 0·771 sp 6 0·06 (2) 0·49 (3) 5·73 (13) 45·13 (62) 31·56 (54) 0·20 (2) 0·23 (2) 15·61 (7) 0·16 (3) 0·00 0·00 99·16 0·469 PU906 980 glass 10 49·39 (63) 0·53 (8) 0·03 (2) 19·03 (38) 3·82 (51) 0·20 (7) n.d. 1·66 (58) 8·47 (23) 0·68 (9) 0·52 (5) 84·33 0·437 cpx 10 48·49 (66) 0·43 (9) 0·18 (9) 5·77 (68) 6·76 (24) 0·20 (3) n.d. 15·66 (61) 20·87 (45) 0·20 (2) 0·01 (1) 98·57 0·805 opx 6 50·43 (52) 0·20 (1) 0·18 (6) 6·74 (49) 11·77 (20) 0·26 (2) n.d. 28·14 (24) 1·25 (15) 0·03 (5) 0·01 (2) 99·01 0·810 amph 6 41·73 (45) 0·92 (6) 0·26 (13) 13·61 (61) 8·35 (18) 0·13 (2) n.d. 16·43 (25) 11·77 (15) 1·62 (3) 0·62 (5) 95·43 0·778 sp 6 0·16 (07) 1·53 (13) 3·34 (246) 15·37 (46) 64·18 (255) 0·18 (11) n.d. 9·09 (35) 0·26 (4) 0·00 0·01 (1) 94·11 0·201 EQ Mb Pt–C PU72 1350 glass 5 46·19 (28) 0·67 (4) 0·15 (1) 12·80 (07) 8·14 (20) 0·15 (4) n.d. 17·06 (14) 9·93 (8) 1·19 (3) 0·40 (1) 96·68 0·789 ZP1330 1330 glass 10 46·90 (34) 0·74 (6) 0·21 (3) 12·70 (17) 9·04 (20) 0·17 (2) n.d. 16·05 (21) 10·13 (7) 1·19 (4) 0·42 (2) 97·56 0·760 ol 11 41·45 (19) 0·01 (1) 0·10 (2) 0·00 9·20 (18) 0·13 (1) 0·07 (4) 49·73 (20) 0·25 (1) 0·01 (1) 0·00 100·95 0·906 ZP1301 1300 glass 10 46·58 (39) 0·74 (6) 0·21 (3) 13·35 (27) 8·76 (21) 0·17 (2) 0·01 (2) 13·93 (28) 10·82 (13) 1·27 (4) 0·43 (1) 96·27 0·739 ol 10 41·23 (25) 0·01 (2) 0·15 (1) 0·00 10·21 (22) 0·15 (2) 0·14 (3) 49·20 (31) 0·29 (2) 0·01 (1) 0·00 101·39 0·896 ZP1302 1300 glass 10 46·93 (52) 0·74 (6) 0·20 (3) 13·00 (31) 8·84 (35) 0·16 (3) n.d. 15·12 (10) 10·33 (5) 1·20 (4) 0·44 (1) 96·96 0·753 ol 10 41·69 (19) 0·01 (1) 0·09 (2) 0·00 9·51 (22) 0·14 (2) 0·06 (5) 49·91 (15) 0·23 (2) 0·02 (1) 0·00 101·65 0·903 ZP1270 1270 glass 10 46·52 (27) 0·80 (4) 0·18 (2) 13·40 (38) 8·91 (26) 0·15 (3) 0·00 14·11 (60) 10·72 (24) 1·22 (3) 0·42 (1) 96·44 0·738 ol 10 41·09 (32) 0·00 0·12 (1) 0·00 10·22 (34) 0·13 (1) 0·04 (4) 49·17 (62) 0·24 (2) 0·01 (1) 0·00 101·02 0·896 ZP1240 1240 glass 9 46·95 (27) 0·78 (8) 0·19 (2) 14·08 (41) 8·84 (28) 0·15 (2) 0·01 (2) 12·81 (15) 11·11 (9) 1·32 (4) 0·47 (1) 96·70 0·721 ol 11 41·31 (42) 0·00 0·12 (2) 0·00 11·23 (27) 0·15 (2) 0·05 (4) 48·64 (72) 0·24 (2) 0·01 (1) 0·01 (1) 101·75 0·885 ZP1210 1210 glass 12 47·67 (38) 0·85 (5) 0·16 (2) 15·17 (45) 8·49 (23) 0·17 (3) 0·00 10·71 (39) 12·00 (13) 1·37 (6) 0·49 (2) 97·09 0·692 ol 10 41·18 (38) 0·00 0·13 (3) 0·00 12·37 (37) 0·16 (2) 0·04 (3) 48·16 (20) 0·28 (3) 0·01 (1) 0·00 102·32 0·874 sp 10 0·19 (29) 0·33 (3) 32·81 (86) 34·89 (61) 14·49 (31) 0·17 (2) 0·04 (2) 16·30 (19) 0·11 (4) 0·00 0·00 99·34 0·667 ZP1180 1180 glass 10 47·44 (71) 0·92 (8) 0·10 (4) 16·31 (56) 8·31 (34) 0·17 (2) 0·01 (2) 9·15 (52) 12·00 (24) 1·48 (6) 0·52 (2) 96·41 0·662 ol 8 40·61 (26) 0·00 0·05 (4) 0·00 13·44 (22) 0·19 (2) 0·03 (4) 46·45 (10) 0·26 (2) 0·00 0·00 101·03 0·860 High-Al cpx 9 50·56 (45) 0·40 (4) 1·28 (17) 5·76 (51) 4·20 (31) 0·13 (1) 0·00 16·46 (10) 20·76 (14) 0·25 (3) 0·00 99·81 0·875 Low-Al cpx 7 52·70 (66) 0·29 (4) 0·83 (6) 2·26 (27) 4·45 (38) 0·16 (1) 0·00 18·22 (24) 20·01 (60) 0·16 (3) 0·00 99·08 0·880 sp 8 0·00 0·38 (4) 32·60 (273) 33·77 (224) 15·69 (51) 0·18 (3) 0·00 15·45 (36) 0·16 (9) 0·00 0·00 98·24 0·637 ZP1151 1150 glass 10 47·75 (36) 0·94 (6) 0·05 (4) 16·22 (30) 8·19 (26) 0·17 (2) 0·00 9·39 (11) 11·40 (8) 1·43 (8) 0·53 (1) 96·08 0·672 ol 16 40·76 (73) 0·01 (2) 0·06 (3) 0·02 (02) 13·58 (30) 0·17 (3) 0·02 (2) 46·21 (87) 0·34 (12) 0·01 (1) 0·00 101·17 0·858 High-Al cpx 8 51·00 (49) 0·48 (10) 1·19 (12) 7·07 (68) 4·43 (49) 0·12 (1) 0·00 16·34 (43) 20·43 (25) 0·29 (2) 0·00 101·35 0·868 Low-Al cpx 8 53·20 (59) 0·23 (2) 1·06 (17) 4·08 (40) 4·42 (41) 0·12 (3) 0·00 17·97 (57) 19·89 (62) 0·26 (2) 0·00 101·24 0·879 sp 10 0·26 (33) 0·37 (9) 27·47 (260) 36·27 (298) 15·77 (49) 0·18 (4) 0·00 15·87 (52) 0·28 (15) 0·00 0·00 96·49 0·642 ZP1152 1150 glass 12 46·86 (96) 1·01 (9) 0·08 (3) 17·44 (43) 7·82 (35) 0·24 (4) 0·01 (3) 8·04 (43) 10·75 (51) 2·10 (11) 0·53 (4) 94·88 0·647 ol 8 40·70 (32) 0·00 0·08 (1) 0·00 13·33 (57) 0·28 (4) 0·11 (3) 46·31 (47) 0·26 (5) 0·00 0·00 101·07 0·861 cpx 10 50·97 (78) 0·48 (8) 1·06 (8) 6·49 (94) 4·86 (30) 0·20 (4) 0·00 16·30 (48) 19·99 (17) 0·34 (1) 0·01 (1) 100·68 0·857 sp 11 0·00 0·24 (3) 21·65 (119) 45·64 (118) 14·04 (55) 0·23 (5) 0·10 (6) 17·86 (40) 0·09 (6) 0·00 0·01 (1) 100·18 0·694 ZP1121 1120 glass 12 46·82 (125) 0·84 (6) 0·04 (3) 17·73 (38) 7·38 (49) 0·13 (2) 0·00 8·17 (28) 10·92 (36) 2·79 (12) 0·58 (5) 95·40 0·664 ol 12 41·40 (53) 0·01 (1) 0·05 (3) 0·00 14·05 (61) 0·17 (3) 0·07 (3) 45·95 (46) 0·24 (3) 0·02 (3) 0·01 (1) 101·97 0·854 cpx 10 51·76 (71) 0·52 (10) 0·34 (12) 7·81 (81) 5·54 (39) 0·14 (3) 0·02 (2) 16·50 (38) 18·89 (45) 0·43 (8) 0·03 (1) 101·97 0·842 High-Al opx 4 53·97 (38) 0·15 (4) 0·30 (9) 6·62 (22) 7·89 (32) 0·16 (2) 0·00 29·89 (53) 1·64 (2) 0·13 (1) 0·00 100·75 0·871 Low-Al opx 7 55·71 (53) 0·12 (3) 0·27 (7) 4·00 (85) 8·26 (29) 0·14 (2) 0·00 30·68 (45) 1·71 (10) 0·13 (3) 0·00 101·02 0·869 sp 8 0·97 (49) 0·14 (2) 9·73 (119) 56·32 (102) 12·18 (26) 0·12 (2) 0·06 (3) 19·76 (60) 0·16 (6) 0·00 0·00 99·44 0·743 ZP1122 1120 glass 10 46·51 (65) 0·98 (6) 0·00 17·91 (38) 6·86 (20) 0·13 (3) 0·00 8·66 (20) 11·05 (14) 1·94 (4) 0·74 (2) 94·77 0·692 ol 8 41·63 (47) 0·01 (1) 0·01 (1) 0·00 12·80 (23) 0·14 (4) 0·02 (7) 47·65 (57) 0·28 (4) 0·00 0·00 102·59 0·869 cpx 12 52·26 (62) 0·40 (8) 0·53 (19) 6·43 (58) 4·65 (38) 0·16 (4) 0·02 (4) 16·99 (47) 20·08 (58) 0·28 (3) 0·00 101·78 0·867 High-Al opx 9 53·27 (52) 0·21 (5) 0·39 (7) 7·78 (32) 8·69 (73) 0·13 (3) 0·03 (5) 29·50 (67) 1·72 (2) 0·04 (1) 0·00 101·76 0·858 Low-Al opx 4 56·62 (58) 0·12 (1) 0·29 (4) 3·38 (10) 8·81 (30) 0·15 (2) 0·02 (4) 31·12 (31) 1·86 (6) 0·03 (1) 0·00 102·40 0·863 ZP1090 1090 glass 8 49·12 (24) 1·09 (9) 0·01 (1) 18·33 (89) 8·63 (48) 0·17 (2) 0·01 (2) 7·01 (52) 10·29 (67) 2·31 (12) 0·86 (9) 97·83 0·592 ol 7 42·41 (28) 0·00 0·03 (2) 0·00 17·48 (45) 0·27 (2) 0·15 (2) 44·77 (48) 0·26 (1) 0·00 0·01 (1) 105·38 0·820 cpx 9 50·87 (38) 0·51 (4) 0·41 (8) 6·95 (43) 5·52 (17) 0·15 (3) 0·00 16·29 (13) 20·05 (23) 0·33 (2) 0·01 (1) 101·10 0·840 High-Al opx 5 52·27 (33) 0·30 (3) 0·03 (3) 8·23 (48) 11·01 (13) 0·25 (2) 0·00 27·92 (10) 1·72 (18) 0·05 (1) 0·01 (1) 101·79 0·819 Low-Al opx 5 55·25 (7) 0·14 (2) 0·03 (3) 3·70 (12) 10·99 (24) 0·25 (4) 0·00 29·68 (19) 1·67 (1) 0·05 (2) 0·01 (1) 101·76 0·828 sp 8 0·00 0·18 (4) 4·43 (291) 65·62 (298) 13·90 (54) 0·13 (4) 0·09 (3) 19·74 (50) 0·17 (6) 0·01 (2) 0·02 (1) 104·30 0·717 ZP1060 1060 glass 13 46·87 (148) 1·07 (18) 0·02 (3) 19·33 (52) 7·51 (40) 0·11 (3) 0·00 4·85 (64) 7·38 (37) 3·63 (30) 1·55 (9) 92·29 0·535 ol 4 39·12 (21) 0·05 (1) 0·02 (2) 0·00 20·11 (50) 0·23 (2) 0·11 (3) 39·55 (40) 0·28 (3) 0·00 0·00 99·48 0·778 cpx 10 51·16 (72) 0·81 (6) 0·15 (7) 7·39 (20) 5·82 (26) 0·14 (2) 0·02 (2) 15·79 (64) 20·21 (32) 0·41 (4) 0·02 (1) 101·93 0·829 opx 9 53·57 (74) 0·28 (4) 0·11 (4) 6·46 (64) 11·91 (57) 0·20 (3) 0·01 (2) 27·78 (50) 1·54 (22) 0·05 (4) 0·01 (2) 101·93 0·806 amp 5 42·28 (74) 1·94 (15) 0·15 (1) 15·32 (40) 7·51 (33) 0·11 (1) 0·03 (2) 15·61 (34) 11·15 (72) 2·40 (17) 0·71 (4) 97·20 0·787 sp 8 0·00 0·20 (4) 1·64 (134) 60·86 (168) 15·25 (33) 0·11 (2) 0·14 (3) 17·60 (42) 0·20 (2) 0·00 0·00 96·00 0·673 ZP1000 1000 glass 8 44·67 (55) 0·92 (7) 0·00 20·72 (13) 6·50 (09) 0·06 (2) 0·00 3·98 (41) 9·01 (15) 3·12 (20) 1·24 (6) 90·20 0·521 cpx 9 50·81 (82) 0·70 (20) 0·16 (4) 7·26 (84) 6·22 (39) 0·17 (2) 0·02 (3) 15·21 (43) 20·91 (31) 0·31 (7) 0·01 (1) 101·80 0·813 opx 9 52·97 (54) 0·17 (4) 0·13 (6) 6·12 (72) 13·06 (29) 0·24 (2) 0·01 (2) 27·32 (49) 1·22 (4) 0·02 (1) 0·00 101·28 0·788 amp 9 43·27 (69) 1·09 (20) 0·21 (8) 15·67 (93) 7·74 (48) 0·11 (2) 0·05 (2) 15·94 (49) 11·12 (57) 2·23 (14) 0·69 (9) 98·12 0·786 sp 8 0·00 0·14 (2) 1·68 (19) 65·05 (57) 15·12 (25) 0·15 (1) 0·13 (3) 18·61 (28) 0·24 (2) 0·00 0·00 101·12 0·687 All Fe as FeO. Units in parentheses indicate standard deviations (2σ) from averaged analysis [i.e. 44·41(25) should be read as 44·41 ± 0·25 wt %]; #, number of individual spots averaged for each phase composition reported. Abbreviations are as for Table 2; xMg = molar [MgO/(MgO + FeOtot)]. Open in new tab Micro-Raman spectroscopy The H2O contents of most recovered experimental glasses were quantified by micro-Raman spectroscopy. Raman scattering was excited using a 514·5 nm diode laser and measured with a Dilor Labram II confocal micro-Raman spectrometer. The laser beam was focused 5 µm below the sample surface with a spot size of 1–2 µm in confocal mode. Spectra were obtained in the 180–1500 and 2800–3900 cm-1 ranges to cover low- and high-frequency T–O stretching and vibration modes and the OH/H2O stretching regions. Two to three spectra were acquired for each measuring point and 3–5 spots were analyzed on each sample. The acquisition time was 120 s. The spectra were processed using Origin 9.1® software employing a cubic baseline correction scheme similar to that of Di Muro et al. (2006) and Mercier et al. (2009). Quantification was obtained by acquiring Raman spectra of a total of 20 synthetic glasses with known H2O concentrations (Karl-Fisher titration; KFT) ranging in composition from olivine-tholeiite to rhyolite and H2O contents from nominally anhydrous to 10 wt %. The relative peak ratios of the OH/H2O peak(s) to either the low-frequency (LF, 470–510 cm–1) or the high-frequency (HF, 960–1120 cm–1) silicate vibration modes was calibrated as a function of H2O content and the LF/HF ratio (as a proxy for composition or polymerization of the silicate glass). Two calibration curves were defined, one for the range olivine-tholeiite to andesite and a second for andesite to rhyolite, as no single, straightforward algorithm allowed fitting over the entire compositional range. Accuracy of the method is estimated at ±0·5 wt % H2O, and results are reported in Table 2. RESULTS Experimental conditions, phase assemblages, modes, relative Fe loss, estimated fO2 conditions and H2O contents of recovered glasses are presented in Table 2. Average phase compositions with standard deviations are reported in Tables 3 and 4. Table 4: Electron microprobe analyses of experimental phases of fractional crystallization experiments Run # . T (°C) . phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . P2O5 . Total . xMg . xAn . FC ba AuPd rk48 1230 glass 9 48·95 (33) 0·60 (4) n.d. 14·71 (35) 7·84 (23) 0·18 (7) n.d. 9·59 (16) 9·25 (8) 2·04 (14) 0·36 (3) n.d. 93·52 0·686 rk50 1200 glass 10 48·00 (48) 0·60 (5) n.d. 14·44 (31) 8·15 (13) 0·14 (5) n.d. 8·85 (11) 9·36 (14) 2·14 (16) 0·38 (2) n.d. 92·06 0·659 opx 10 55·15 (40) 0·08 (1) n.d. 3·11 (28) 7·74 (16) 0·18 (3) n.d. 32·37 (26) 1·48 (13) 0·03 (2) 0·00 n.d. 100·14 0·882 rk56 1170 glass 9 48·71 (27) 0·65 (3) n.d. 15·47 (18) 7·73 (29) 0·18 (4) n.d. 7·45 (19) 9·41 (21) 2·25 (14) 0·43 (3) n.d. 92·29 0·632 cpx 9 52·60 (38) 0·20 (2) n.d. 3·45 (50) 5·40 (18) 0·20 (3) n.d. 18·04 (49) 19·32 (49) 0·29 (12) 0·01 (1) n.d. 99·51 0·856 opx 10 55·39 (31) 0·11 (2) n.d. 2·81 (48) 9·17 (17) 0·25 (5) n.d. 30·94 (27) 1·78 (8) 0·05 (2) 0·01 (1) n.d. 100·50 0·857 rk60 1140 glass 10 50·09 (72) 0·69 (5) n.d. 15·71 (22) 7·22 (21) 0·18 (4) n.d. 6·64 (9) 9·23 (13) 2·11 (41) 0·47 (2) n.d. 92·34 0·621 cpx 10 52·73 (25) 0·21 (4) n.d. 3·00 (38) 5·30 (31) 0·21 (4) n.d. 17·99 (16) 20·00 (28) 0·34 (4) 0·01 (1) n.d. 99·80 0·857 opx 10 55·25 (63) 0·13 (3) n.d. 2·59 (29) 9·37 (31) 0·30 (3) n.d. 30·94 (37) 1·71 (17) 0·05 (2) 0·01 (1) n.d. 100·33 0·855 rk63 1110 glass 9 48·76 (28) 0·76 (5) n.d. 16·38 (27) 6·06 (19) 0·17 (5) n.d. 6·05 (11) 8·77 (10) 2·48 (12) 0·58 (3) n.d. 90·00 0·640 cpx 10 52·81 (44) 0·26 (4) n.d. 3·20 (47) 4·92 (35) 0·26 (5) n.d. 17·64 (46) 20·58 (65) 0·32 (3) 0·01 (1) n.d. 100·00 0·865 opx 10 54·79 (17) 0·15 (2) n.d. 3·09 (24) 9·18 (21) 0·32 (4) n.d. 30·59 (33) 1·56 (20) 0·04 (2) 0·01 (1) n.d. 99·73 0·856 rk66 1080 glass 10 49·38 (26) 0·83 (4) n.d. 17·68 (19) 6·14 (20) 0·17 (4) n.d. 4·47 (9) 7·78 (8) 2·80 (11) 0·66 (3) n.d. 89·89 0·565 cpx 8 49·12 (50) 0·60 (4) n.d. 6·99 (31) 6·26 (15) 0·22 (3) n.d. 14·78 (39) 20·44 (28) 0·48 (4) 0·02 (2) n.d. 98·90 0·808 opx 9 52·99 (37) 0·20 (3) n.d. 5·25 (73) 11·33 (22) 0·35 (3) n.d. 28·77 (34) 1·25 (9) 0·04 (2) 0·00 n.d. 100·19 0·819 rk69 1050 glass 10 49·66 (31) 0·81 (5) n.d. 17·91 (27) 6·20 (11) 0·20 (6) n.d. 3·78 (10) 7·85 (9) 2·48 (12) 0·68 (3) n.d. 89·57 0·521 cpx 6 48·88 (86) 0·69 (4) n.d. 6·86 (33) 6·86 (31) 0·26 (3) n.d. 14·44 (35) 21·22 (32) 0·46 (4) 0·01 (1) n.d. 99·68 0·789 amp 10 42·13 (65) 1·66 (5) n.d. 14·40 (40) 9·52 (42) 0·18 (3) n.d. 15·32 (38) 11·26 (19) 2·47 (5) 0·43 (2) n.d. 97·36 0·742 rk70 1020 glass 10 50·16 (65) 0·81 (3) n.d. 18·22 (34) 6·04 (16) 0·21 (5) n.d. 3·01 (9) 7·61 (17) 2·57 (13) 0·72 (2) n.d. 89·34 0·470 cpx 7 47·37 (38) 0·84 (4) n.d. 8·20 (46) 7·78 (12) 0·31 (2) n.d. 12·91 (31) 21·23 (16) 0·51 (5) 0·01 (1) n.d. 99·17 0·747 amp 8 41·59 (31) 1·86 (4) n.d. 15·72 (19) 10·32 (14) 0·24 (3) n.d. 14·02 (4) 11·13 (9) 2·47 (5) 0·42 (2) n.d. 97·76 0·708 rk71 990 glass 10 52·58 (58) 0·73 (5) n.d. 18·04 (7) 5·29 (26) 0·21 (4) n.d. 2·09 (11) 7·38 (12) 2·79 (19) 0·74 (5) n.d. 89·86 0·413 cpx 9 46·30 (61) 0·96 (9) n.d. 9·26 (67) 9·17 (14) 0·33 (5) n.d. 11·22 (34) 21·74 (13) 0·57 (3) 0·01 (1) n.d. 99·56 0·686 amp 10 40·79 (41) 2·02 (11) n.d. 15·72 (24) 12·50 (23) 0·26 (4) n.d. 12·05 (23) 11·51 (7) 2·49 (8) 0·39 (2) n.d. 97·72 0·632 PU1049 990 glass 11 49·67 (19) 0·76 (4) n.d. 18·46 (19) 5·46 (9) 0·20 (4) n.d. 3·23 (9) 7·43 (20) 2·85 (11) 0·72 (4) 0·16 (6) 88·94 0·513 amp 12 40·96 (39) 2·00 (18) n.d. 15·89 (39) 10·87 (42) 0·23 (3) n.d. 14·21 (36) 10·96 (37) 2·28 (4) 0·39 (4) 0·01 (1) 97·81 0·700 mag 8 0·13 (1) 3·11 (8) n.d. 15·83 (48) 69·13 (76) 0·45 (2) n.d. 6·16 (13) 0·16 (2) 0·01 (1) 0·01 (1) 0·01 (1) 94·98 0·137 PU1070 980 glass 22 48·58 (18) 0·84 (9) n.d. 18·20 (20) 5·20 (20) 0·23 (7) n.d. 2·79 (24) 7·58 (10) 2·74 (6) 0·74 (6) 0·24 (4) 87·14 0·488 amp 30 40·98 (69) 2·14 (9) n.d. 15·46 (32) 8·26 (25) 0·23 (5) n.d. 13·91 (35) 11·02 (28) 2·34 (15) 0·43 (6) n.d. 94·76 0·750 PU1062 950 glass 41 52·17 (60) 0·76 (7) n.d. 19·60 (39) 4·97 (19) 0·19 (2) n.d. 2·24 (21) 7·43 (10) 3·10 (22) 0·81 (5) 0·18 (3) 91·24 0·446 amp 34 40·15 (33) 1·96 (9) n.d. 16·72 (44) 9·64 (20) 0·23 (1) n.d. 13·52 (26) 10·90 (14) 2·36 (4) 0·46 (2) 0·04 (2) 95·90 0·714 gar 15 37·96 (32) 0·61 (12) n.d. 22·61 (20) 17·13 (23) 1·33 (11) n.d. 10·54 (44) 8·48 (49) 0·03 (2) 0·01 (1) 0·05 (1) 98·73 0·523 PU1064 900 glass 25 55·65 (30) 0·51 (6) n.d. 17·86 (25) 3·93 (7) 0·10 (3) n.d. 1·13 (3) 6·05 (8) 3·15 (15) 1·05 (5) 0·24 (4) 89·67 0·339 amp 26 41·87 (68) 2·00 (20) n.d. 17·07 (52) 15·04 (26) 0·25 (2) n.d. 10·21 (7) 10·47 (18) 2·29 (5) 0·49 (3) 0·02 (1) 99·71 0·548 gar 20 39·75 (36) 1·20 (7) n.d. 22·07 (22) 21·65 (35) 1·60 (5) n.d. 6·77 (25) 9·65 (42) 0·02 (1) 0·01 (1) 0·02 (1) 102·75 0·358 plg 32 48·83 (83) 0·03 (2) n.d. 34·61 (52) 0·38 (2) 0·01 (1) n.d. 0·02 (1) 17·44 (46) 1·86 (15) 0·03 (1) 0·02 (2) 103·30 0·838 ilm 17 0·04 (1) 41·44 (104) n.d. 0·78 (5) 55·33 (171) 0·35 (3) n.d. 2·23 (15) 0·14 (3) 0·01 (1) 0·01 (1) 0 100·19 0·055 PU1066 850 glass 24 58·06 (38) 0·38 (2) n.d. 16·41 (28) 2·83 (9) 0·08 (3) n.d. 0·74 (4) 4·88 (9) 3·22 (12) 1·24 (4) 0·21 (4) 88·05 0·317 gar 19 38·73 (55) 1·06 (12) n.d. 21·40 (30) 23·62 (27) 1·34 (12) n.d. 5·40 (24) 9·19 (23) n.d. n.d. n.d. 100·73 0·289 plg 23 49·25 (78) n.d. n.d. 33·32 (48) 0·39 (15) n.d. n.d. 16·05 (40) 2·45 (16) 0·03 (2) n.d. n.d. 101·49 0·784 ilm 6 0·33 (4) 56·45 (54) n.d. 0·35 (4) 42·84 (48) 0·48 (6) n.d. 2·53 (6) 0·25 (4) n.d. n.d. n.d. 103·24 0·095 PU1068 800 glass 27 62·93 (32) 0·21 (4) n.d. 15·06 (16) 1·76 (9) 0·05 (2) n.d. 0·38 (2) 3·28 (6) 3·73 (10) 1·55 (5) 0·12 (3) 89·07 0·276 amp 16 43·99 (87) 1·67 (7) n.d. 13·97 (62) 18·77 (89) 0·45 (7) n.d. 8·03 (30) 9·75 (26) 2·03 (9) 0·38 (6) n.d. 99·05 0·433 gar 24 37·96 (51) 1·05 (18) n.d. 20·94 (22) 25·05 (67) 1·99 (16) n.d. 3·47 (26) 9·47 (66) n.d. n.d. n.d. 99·92 0·198 plg 21 57·12 (69) n.d. n.d. 28·67 (66) 0·27 (8) n.d. n.d. n.d. 10·25 (35) 5·97 (16) 0·03 (2) n.d. 102·32 0·487 ilm 10 0·31 (15) 49·53 (70) n.d. 0·41 (2) 47·98 (69) 0·41 (7) n.d. 1·41 (4) 0·15 (5) n.d. n.d. n.d. 100·19 0·050 PU1072 750 glass 10 61·92 (39) 0·04 (2) n.d. 13·22 (15) 0·87 (6) 0·10 (3) n.d. 0·40 (3) 2·44 (7) 3·21 (7) 1·81 (5) 0·11 (2) 84·09 0·449 plg 14 55·26 (66) n.d. n.d. 26·02 (32) 0·24 (6) n.d. n.d. n.d. 8·81 (25) 6·05 (11) 0·03 (3) n.d. 96·41 0·446 mag 7 0·65 (23) 15·90 (242) n.d. 1·28 (26) 68·57 (130) 0·20 (6) n.d. 0·40 (8) 0·16 (4) n.d. n.d. n.d. 87·15 0·010 gar 10 35·45 (61) 0·05 (5) n.d. 20·45 (31) 30·29 (166) 3·48 (10) n.d. 2·37 (33) 2·92 (40) n.d. n.d. n.d. 95·01 0·122 PU1072-o 750 720 glass 15 64·34 (41) 0·07 (3) n.d. 13·42 (14) 0·78 (10) 0·10 (2) n.d. 0·36 (3) 2·05 (6) 3·28 (10) 1·92 (6) 0·11 (3) 86·43 0·453 plg 21 59·04 (59) n.d. n.d. 25·50 (35) 0·24 (7) n.d. n.d. n.d. 7·46 (25) 7·05 (21) 0·05 (3) n.d. 99·33 0·369 amp 22 44·27 (96) 1·33 (18) n.d. 11·83 (70) 19·02 (43) 0·96 (14) n.d. 8·1 (29) 8·83 (23) 2·08 (13) 0·31 (10) n.d. 96·72 0·432 mag 6 1·43 (24) 18·81 (357) n.d. 1·48 (49) 66·64 (238) 0·27 (5) n.d. 0·46 (9) 0·15 (2) n.d. m.d. n.d. 89·25 0·012 qtz 4 99·18 (67) 99·18 FC Mb AuPd rk47 1230 glass 16 43·86 (43) 0·79 (3) 0·08 (3) 12·11 (44) 9·32 (18) 0·16 (5) 0·02(2) 12·24 (70) 10·92 (33) 1·27 (9) 0·44 (3) n.d. 91·37 0·701 ol 10 40·72 (31) 0·01 (1) 0·05 (3) 0·04 (1) 8·82 (16) 0·17 (3) 0·21 (2) 50·53 (19) 0·21 (2) 0·01 (1) 0·00 n.d. 100·77 0·911 sp 8 0·00 0·30 (6) 42·13 (585) 13·85 (275) 24·83 (164) 0·00 0·13 (3) 14·08 (39) 0·21 (9) 0·03 (3) 0·00 n.d. 95·56 0·501 rk51 1200 glass 10 43·66 (41) 0·89 (4) n.d. 14·02 (28) 9·00 (22) 0·20 (6) n.d. 10·36 (12) 12·35 (13) 1·27 (10) 0·53 (2) n.d. 92·27 0·672 ol 10 40·84 (15) 0·01 (1) n.d. 0·03 (1) 10·33 (19) 0·19 (3) n.d. 48·74 (26) 0·24 (2) 0·01 (1) 0·00 n.d. 100·39 0·894 cpx 9 52·08 (55) 0·23 (3) n.d. 3·78 (52) 4·30 (14) 0·11 (3) n.d. 17·39 (45) 21·68 (42) 0·21 (2) 0·00 n.d. 99·79 0·878 rk52 1170 glass 10 44·61 (56) 0·71 (2) n.d. 16·53 (32) 9·07 (24) 0·21 (4) n.d. 8·92 (27) 11·33 (24) 1·53 (14) 0·74 (4) n.d. 93·65 0·637 ol 10 40·66 (19) 0·01 (1) n.d. 0·05 (1) 12·23 (27) 0·25 (4) n.d. 46·77 (40) 0·24 (2) 0·00 0·00 n.d. 100·21 0·872 cpx 10 49·35 (80) 0·34 (6) n.d. 7·49 (76) 5·68 (69) 0·16 (2) n.d. 15·57 (47) 21·24 (59) 0·32 (3) 0·01 (1) n.d. 100·17 0·830 sp 2 0·11 (1) 0·12 (1) n.d. 59·39 (5) 18·96 (23) 0·18 (6) n.d. 19·93 (20) 0·06 (1) 0·01 (1) 0·00 n.d. 98·74 0·652 rk54 1140 glass 9 44·67 (34) 0·80 (6) n.d. 17·21 (27) 9·53 (22) 0·23 (6) n.d. 7·50 (14) 10·38 (12) 1·73 (12) 0·91 (3) n.d. 92·97 0·584 cpx 10 48·73 (51) 0·42 (4) n.d. 7·89 (48) 6·39 (29) 0·20 (2) n.d. 14·85 (24) 21·49 (16) 0·32 (4) 0·01 (1) n.d. 100·31 0·805 sp 9 0·10 (3) 0·18 (3) n.d. 57·74 (100) 21·73 (49) 0·16 (3) n.d. 18·47 (32) 0·14 (4) 0·01 (1) 0·00 n.d. 98·53 0·602 rk57 1110 glass 9 47·00 (35) 0·86 (5) n.d. 18·07 (25) 8·47 (26) 0·22 (5) n.d. 5·66 (31) 8·74 (20) 2·16 (12) 1·07 (7) n.d. 92·26 0·544 cpx 10 48·87 (43) 0·60 (5) n.d. 7·95 (51) 7·03 (31) 0·26 (3) n.d. 14·63 (24) 20·15 (36) 0·41 (4) 0·01 (1) n.d. 99·90 0·788 opx 10 51·69 (38) 0·18 (3) n.d. 6·86 (65) 12·28 (19) 0·35 (3) n.d. 27·30 (49) 1·48 (22) 0·05 (3) 0·02 (3) n.d. 100·21 0·798 sp 10 0·10 (7) 0·17 (6) n.d. 60·28 (58) 20·73 (51) 0·20 (2) n.d. 17·10 (19) 0·12 (3) 0·01 (2) 0·00 n.d. 98·71 0·595 rk58 1080 glass 10 48·43 (44) 0·91 (4) n.d. 17·18 (29) 8·40 (28) 0·30 (6) n.d. 4·35 (27) 7·80 (34) 2·42 (9) 1·34 (4) n.d. 91·13 0·480 cpx 8 45·65 (60) 0·96 (13) n.d. 10·28 (72) 9·00 (32) 0·36 (5) n.d. 12·22 (31) 20·06 (25) 0·55 (3) 0·01 (1) n.d. 99·09 0·708 opx 7 48·63 (44) 0·25 (3) n.d. 8·61 (47) 13·72 (47) 0·48 (2) n.d. 25·34 (32) 1·11 (13) 0·05 (2) 0·00 n.d. 98·20 0·767 mag 5 0·13 (2) 3·10 (6) n.d. 12·23 (8) 73·29 (53) 0·36 (6) n.d. 4·97 (9) 0·04 (5) 0·00 0·00 n.d. 94·11 0·108 rk73 1050 glass 10 49·86 (51) 0·86 (4) n.d. 17·85 (17) 8·70 (17) 0·30 (6) n.d. 3·78 (9) 7·90 (15) 2·29 (9) 1·18 (4) n.d. 92·71 0·437 cpx 1 47·41 0·84 n.d. 8·07 8·34 0·34 n.d. 13·73 20·74 0·44 0·08 n.d. 99·99 0·746 amp 9 41·26 (49) 1·92 (10) n.d. 15·73 (36) 10·45 (111) 0·28 (4) n.d. 14·27 (56) 11·34 (18) 2·20 (8) 0·84 (4) n.d. 98·29 0·709 rk65 1020 glass 10 52·17 (27) 0·87 (4) n.d. 17·18 (39) 7·07 (18) 0·28 (7) n.d. 2·70 (5) 6·78 (9) 2·41 (11) 1·35 (3) n.d. 90·81 0·405 cpx 10 47·50 (99) 0·75 (13) n.d. 6·94 (84) 10·88 (60) 0·49 (4) n.d. 12·31 (56) 20·04 (40) 0·51 (3) 0·01 (1) n.d. 99·42 0·669 amp 4 41·01 (44) 2·17 (10) n.d. 14·36 (26) 14·08 (20) 0·38 (2) n.d. 12·05 (12) 10·91 (26) 2·19 (7) 0·70 (5) n.d. 97·84 0·604 mag 5 0·12 (2) 5·96 (11) n.d. 7·98 (14) 76·44 (101) 0·54 (4) n.d. 3·03 (4) 0·04 (4) 0·00 0·00 n.d. 94·11 0·066 rk67 990 glass 10 54·74 (111) 0·59 (8) n.d. 15·94 (18) 5·59 (38) 0·31 (5) n.d. 1·48 (12) 5·68 (7) 2·71 (16) 1·33 (7) n.d. 88·36 0·321 cpx 4 46·31 (103) 0·77 (5) n.d. 7·25 (51) 13·02 (7) 0·65 (4) n.d. 10·30 (42) 19·48 (97) 0·75 (14) 0·02 (2) n.d. 98·54 0·585 amp 12 40·46 (59) 1·96 (21) n.d. 13·41 (31) 16·90 (60) 0·48 (3) n.d. 10·28 (34) 10·75 (23) 2·19 (9) 0·67 (4) n.d. 97·09 0·520 mag 4 0·15 (1) 6·10 (28) n.d. 5·66 (4) 79·74 (110) 0·70 (6) n.d. 1·49 (5) 0·17 (6) 0·01 (1) 0·01 (1) n.d. 94·02 0·032 PU1048 990 glass 13 52·64 (30) 0·86 (4) n.d. 16·47 (10) 5·58 (8) 0·29 (3) n.d. 2·69 (9) 6·60 (11) 2·61 (9) 1·32 (4) 0·22 (7) 89·28 0·462 cpx 4 50·08 (84) 0·83 (16) n.d. 5·05 (75) 8·18 (22) 0·59 (3) n.d. 14·63 (31) 19·48 (44) 0·32 (3) 0·01 (1) 0·01 (1) 99·17 0·761 PU1069 980 glass 13 50·97 (33) 0·92 (3) n.d. 16·65 (11) 6·17 (8) 0·31 (7) n.d. 2·35 (3) 6·82 (8) 2·66 (5) 1·54 (3) 0·23 (5) 88·62 0·405 cpx 19 48·70 (79) 0·77 (20) n.d. 5·11 (119) 8·65 (33) 0·71 (9) n.d. 13·23 (57) 19·35 (50) 0·46 (7) n.d. n.d. 97·00 0·732 PU1061 950 glass 29 53·24 (34) 0·90 (8) n.d. 17·69 (20) 6·10 (14) 0·27 (3) n.d. 2·21 (7) 6·83 (8) 2·91 (10) 1·52 (5) 0·21 (2) 91·87 0·393 amp 25 41·57 (16) 2·26 (18) n.d. 14·79 (30) 11·45 (42) 0·33 (4) n.d. 13·13 (20) 10·48 (20) 2·17 (6) 0·85 (4) 0·04 (2) 97·09 0·671 gar 10 38·75 (21) 0·55 (17) n.d. 22·38 (21) 19·13 (33) 2·09 (21) n.d. 9·23 (43) 7·58 (38) 0·03 (2) 0·01 (1) 0·05 (1) 99·79 0·462 cpx 7 51·60 (103) 0·65 (8) n.d. 5·50 (99) 10·62 (93) 0·84 (19) n.d. 12·99 (52) 18·89 (34) 0·53 (8) 0·04 (1) 0·01 (1) 101·66 0·685 PU1063 900 glass 12 54·23 (43) 0·48 (3) n.d. 16·86 (14) 3·74 (12) 0·09 (2) n.d. 0·97 (5) 5·28 (8) 2·92 (5) 1·70 (5) 0·28 (3) 86·56 0·317 amp 5 40·62 (44) 2·14 (18) n.d. 15·94 (21) 15·18 (21) 0·26 (4) n.d. 9·93 (17) 10·41 (28) 2·14 (8) 0·92 (5) 0·03 (1) 97·57 0·538 gar 8 36·83 (29) 1·37 (11) n.d. 21·06 (18) 22·12 (41) 1·69 (9) n.d. 6·06 (50) 9·29 (22) 0·04 (1) 0·01 (1) 0·06 (3) 98·83 0·328 plg 31 48·84 (102) 0·02 (1) n.d. 34·09 (104) 0·48 (8) 0·01 (1) n.d. 0·02 (1) 16·97 (58) 2·07 (16) 0·07 (1) 0·01 (1) 102·70 0·819 ilm 13 0·04 (1) 43·46 (124) n.d. 0·65 (5) 51·50 (115) 0·40 (3) n.d. 2·34 (10) 0·17 (2) 0·01 (1) 0·02 (1) 0 98·37 0·063 PU1065 850 glass 29 58·74 (30) 0·38 (3) n.d. 16·32 (27) 2·65 (12) 0·06 (3) n.d. 0·68 (3) 4·68 (6) 2·92 (5) 2·13 (4) 0·19 (4) 88·75 0·314 gar 35 38·50 (55) 1·21 (16) n.d. 21·43 (26) 23·42 (41) 1·46 (12) n.d. 5·42 (27) 8·92 (43) 0·04 (1) 0·01 (1) 0·07 (2) 100·48 0·290 plg 22 49·20 (40) 0·03 (1) n.d. 33·29 (57) 0·23 (3) 0·01 (1) n.d. 0·01 (1) 16·10 (52) 2·39 (13) 0·08 (1) 0·03 (1) 101·37 0·788 ilm 8 0·13 (3) 51·90 (15) n.d. 0·28 (2) 43·09 (8) 0·41 (2) n.d. 2·65 (5) 0·17 (2) 0·02 (2) 0·02 (1) 0·02 (2) 98·68 0·099 apa 6 0·31 (14) n.d. n.d. 0·06 (5) 0·73 (7) n.d. n.d. 0·20 (3) 53·80 (33) n.d. n.d. 45·07 (34) 99·92 PU1067 800 glass 12 59·77 (65) 0·18 (4) n.d. 14·49 (13) 1·65 (5) 0·03 (2) n.d. 0·34 (2) 3·10 (2) 3·46 (5) 2·62 (5) 0·11 (2) 85·75 0·268 amp 17 43·21 (125) 1·68 (19) n.d. 15·05 (90) 18·77 (66) 0·37 (8) n.d. 7·63 (36) 9·92 (28) 1·90 (7) 0·81 (12) n.d. 99·35 0·420 gar 19 38·26 (84) 1·03 (14) n.d. 21·13 (28) 25·46 (57) 1·64 (17) n.d. 3·54 (27) 9·26 (62) n.d. n.d. n.d. 100·32 0·198 plg 18 56·22 (65) n.d. n.d. 29·95 (56) 0·24 (7) n.d. n.d. n.d. 11·54 (41) 5·41 (26) 0·04 (3) n.d. 103·41 0·541 ilm 10 0·38 (6) 49·45 (67) n.d. 0·41 (4) 48·45 (80) 0·30 (9) n.d. 0·30 (9) 1·37 (5) 0·16 (4) n.d. n.d. 100·51 0·048 PU1071 750 glass 18 61·50 (24) 0·06 (2) n.d. 13·64 (8) 0·86 (9) 0·08 (2) n.d. 0·37 (2) 2·56 (7) 2·96 (10) 2·94 (7) 0·10 (3) 85·01 0·434 gar 16 35·99 (83) 0·03 (3) n.d. 20·49 (36) 34·64 (50) 1·84 (14) n.d. 2·74 (9) 0·80 (5) n.d. n.d. n.d. 96·52 0·124 plg 18 54·61 (83) n.d. n.d. 27·25 (25) 0·22 (8) n.d. n.d. n.d. 10·05 (20) 5·43 (9) 0·03 (2) n.d. 97·58 0·506 mag 14 0·32 (9) 18·37 (179) n.d. 0·99 (17) 67·75 (115) 0·25 (10) n.d. 0·13 (4) 0·49 (7) n.d. n.d. n.d. 88·30 0·003 FC Mb C-Pt rk3 1230 glass 8 46·75 (29) 0·88 (2) 0·15 (2) 15·00 (10) 8·70 (16) 0·14 (2) n.d. 10·59 (11) 12·20 (8) 1·47 (3) 0·49 (2) n.d. 96·37 0·712 ol 10 40·47 (20) 0·02 (2) 0·11 (2) 0·06 (1) 12·50 (17) 0·19 (3) n.d. 47·05 (26) 0·33 (2) 0·03 (4) 0·01 (1) n.d. 100·76 0·870 sp 4 0·33 (25) 0·39 (5) 38·19 (676) 31·29 (619) 14·43 (77) 0·00 n.d. 15·68 (140) 0·15 (9) 0·02 (1) 0·01 (1) n.d. 100·48 0·659 rk6 1200 glass 10 47·79 (13) 1·13 (4) n.d. 17·60 (15) 9·76 (16) 0·18 (2) n.d. 8·05 (24) 11·22 (6) 2·02 (6) 0·73 (1) n.d. 98·49 0·595 ol 11 39·97 (19) 0·02 (1) n.d. 0·07 (2) 16·96 (13) 0·23 (2) n.d. 44·24 (24) 0·34 (3) 0·01 (1) 0·01 (1) n.d. 101·85 0·823 cpx 12 51·00 (34) 0·51 (3) n.d. 7·21 (37) 5·65 (20) 0·17 (2) n.d. 17·11 (23) 19·81 (56) 0·31 (3) 0·01 (1) n.d. 101·79 0·844 rk11 1170 glass 10 46·41 (64) 1·19 (3) n.d. 17·99 (30) 9·17 (13) 0·22 (3) n.d. 7·27 (8) 10·10 (13) 2·09 (15) 0·74 (2) n.d. 95·17 0·586 cpx 10 49·39 (92) 0·60 (5) n.d. 7·97 (60) 5·70 (24) 0·20 (2) n.d. 15·70 (43) 19·49 (43) 0·38 (2) 0·01 (1) n.d. 99·42 0·831 rk13 1140 glass 7 48·07 (37) 1·39 (4) n.d. 18·43 (37) 10·58 (15) 0·21 (2) n.d. 5·56 (10) 8·69 (10) 2·65 (27) 1·08 (4) n.d. 96·65 0·484 cpx 7 48·64 (41) 0·92 (15) n.d. 8·41 (74) 8·67 (32) 0·27 (3) n.d. 15·15 (40) 17·17 (48) 0·50 (6) 0·02 (1) n.d. 99·75 0·757 opx 10 50·20 (66) 0·41 (3) n.d. 7·71 (41) 14·37 (20) 0·30 (2) n.d. 25·31 (36) 1·95 (28) 0·07 (2) 0·01 (1) n.d. 100·32 0·758 plg 5 44·12 (127) 0·09 (5) n.d. 34·08 (99) 0·59 (6) 0·02 (2) n.d. 0·16 (4) 18·46 (66) 0·95 (30) 0·07 (3) n.d. 98·55 0·911 sp 10 0·38 (27) 0·31 (2) n.d. 62·52 (74) 17·31 (15) 0·14 (2) n.d. 16·44 (22) 0·20 (6) 0·04 (3) 0·01 (1) n.d. 97·34 0·629 rk55 1110 glass 5 48·16 (15) 2·10 (5) n.d. 17·15 (35) 13·80 (67) 0·32 (6) n.d. 3·35 (11) 6·65 (16) 3·17 (4) 1·73 (9) n.d. 96·43 0·302 cpx 9 47·88 (37) 1·34 (11) n.d. 7·79 (48) 13·83 (65) 0·51 (6) n.d. 12·49 (31) 14·86 (43) 0·57 (5) 0·05 (4) n.d. 99·32 0·617 opx 9 49·33 (75) 0·62 (7) n.d. 6·29 (83) 20·93 (49) 0·56 (4) n.d. 19·59 (48) 2·29 (32) 0·10 (3) 0·01 (1) n.d. 99·72 0·625 plg (high An) 5 43·97 (125) 0·27 (13) n.d. 35·35 (142) 0·53 (5) 0·05 (2) n.d. 0·10 (8) 18·57 (120) 0·69 (46) 0·09 (8) n.d. 99·60 0·931 plg (low An) 4 52·93 (56) 0·08 (2) n.d. 29·68 (25) 0·55 (4) 0·01 (1) n.d. 0·06 (1) 11·99 (49) 4·48 (23) 0·42 (4) n.d. 100·19 0·582 rk64 1080 glass 10 51·19 (38) 1·90 (8) n.d. 16·27 (19) 12·22 (30) 0·20 (4) n.d. 1·89 (6) 5·27 (11) 3·92 (13) 2·67 (5) n.d. 95·51 0·216 cpx 9 47·32 (60) 1·17 (14) n.d. 6·07 (80) 19·60 (69) 0·51 (5) n.d. 10·88 (44) 12·87 (62) 0·60 (3) 0·03 (1) n.d. 99·05 0·497 plg 9 56·48 (78) 0·08 (1) n.d. 26·83 (38) 0·60 (9) 0·01 (1) n.d. 0·05 (2) 9·46 (61) 5·92 (29) 0·80 (8) n.d. 100·22 0·448 ilm 4 0·06 (1) 50·25 (56) n.d. 0·49 (2) 42·05 (35) 0·49 (4) n.d. 2·90 (5) 0·19 (2) 0·03 (2) 0·05 (1) n.d. 96·50 0·109 gar 6 37·63 (22) 1·22 (37) n.d. 20·42 (41) 25·05 (43) 1·09 (8) n.d. 6·49 (37) 7·55 (32) 0·03 (2) 0·00 n.d. 99·48 0·316 Run # . T (°C) . phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . P2O5 . Total . xMg . xAn . FC ba AuPd rk48 1230 glass 9 48·95 (33) 0·60 (4) n.d. 14·71 (35) 7·84 (23) 0·18 (7) n.d. 9·59 (16) 9·25 (8) 2·04 (14) 0·36 (3) n.d. 93·52 0·686 rk50 1200 glass 10 48·00 (48) 0·60 (5) n.d. 14·44 (31) 8·15 (13) 0·14 (5) n.d. 8·85 (11) 9·36 (14) 2·14 (16) 0·38 (2) n.d. 92·06 0·659 opx 10 55·15 (40) 0·08 (1) n.d. 3·11 (28) 7·74 (16) 0·18 (3) n.d. 32·37 (26) 1·48 (13) 0·03 (2) 0·00 n.d. 100·14 0·882 rk56 1170 glass 9 48·71 (27) 0·65 (3) n.d. 15·47 (18) 7·73 (29) 0·18 (4) n.d. 7·45 (19) 9·41 (21) 2·25 (14) 0·43 (3) n.d. 92·29 0·632 cpx 9 52·60 (38) 0·20 (2) n.d. 3·45 (50) 5·40 (18) 0·20 (3) n.d. 18·04 (49) 19·32 (49) 0·29 (12) 0·01 (1) n.d. 99·51 0·856 opx 10 55·39 (31) 0·11 (2) n.d. 2·81 (48) 9·17 (17) 0·25 (5) n.d. 30·94 (27) 1·78 (8) 0·05 (2) 0·01 (1) n.d. 100·50 0·857 rk60 1140 glass 10 50·09 (72) 0·69 (5) n.d. 15·71 (22) 7·22 (21) 0·18 (4) n.d. 6·64 (9) 9·23 (13) 2·11 (41) 0·47 (2) n.d. 92·34 0·621 cpx 10 52·73 (25) 0·21 (4) n.d. 3·00 (38) 5·30 (31) 0·21 (4) n.d. 17·99 (16) 20·00 (28) 0·34 (4) 0·01 (1) n.d. 99·80 0·857 opx 10 55·25 (63) 0·13 (3) n.d. 2·59 (29) 9·37 (31) 0·30 (3) n.d. 30·94 (37) 1·71 (17) 0·05 (2) 0·01 (1) n.d. 100·33 0·855 rk63 1110 glass 9 48·76 (28) 0·76 (5) n.d. 16·38 (27) 6·06 (19) 0·17 (5) n.d. 6·05 (11) 8·77 (10) 2·48 (12) 0·58 (3) n.d. 90·00 0·640 cpx 10 52·81 (44) 0·26 (4) n.d. 3·20 (47) 4·92 (35) 0·26 (5) n.d. 17·64 (46) 20·58 (65) 0·32 (3) 0·01 (1) n.d. 100·00 0·865 opx 10 54·79 (17) 0·15 (2) n.d. 3·09 (24) 9·18 (21) 0·32 (4) n.d. 30·59 (33) 1·56 (20) 0·04 (2) 0·01 (1) n.d. 99·73 0·856 rk66 1080 glass 10 49·38 (26) 0·83 (4) n.d. 17·68 (19) 6·14 (20) 0·17 (4) n.d. 4·47 (9) 7·78 (8) 2·80 (11) 0·66 (3) n.d. 89·89 0·565 cpx 8 49·12 (50) 0·60 (4) n.d. 6·99 (31) 6·26 (15) 0·22 (3) n.d. 14·78 (39) 20·44 (28) 0·48 (4) 0·02 (2) n.d. 98·90 0·808 opx 9 52·99 (37) 0·20 (3) n.d. 5·25 (73) 11·33 (22) 0·35 (3) n.d. 28·77 (34) 1·25 (9) 0·04 (2) 0·00 n.d. 100·19 0·819 rk69 1050 glass 10 49·66 (31) 0·81 (5) n.d. 17·91 (27) 6·20 (11) 0·20 (6) n.d. 3·78 (10) 7·85 (9) 2·48 (12) 0·68 (3) n.d. 89·57 0·521 cpx 6 48·88 (86) 0·69 (4) n.d. 6·86 (33) 6·86 (31) 0·26 (3) n.d. 14·44 (35) 21·22 (32) 0·46 (4) 0·01 (1) n.d. 99·68 0·789 amp 10 42·13 (65) 1·66 (5) n.d. 14·40 (40) 9·52 (42) 0·18 (3) n.d. 15·32 (38) 11·26 (19) 2·47 (5) 0·43 (2) n.d. 97·36 0·742 rk70 1020 glass 10 50·16 (65) 0·81 (3) n.d. 18·22 (34) 6·04 (16) 0·21 (5) n.d. 3·01 (9) 7·61 (17) 2·57 (13) 0·72 (2) n.d. 89·34 0·470 cpx 7 47·37 (38) 0·84 (4) n.d. 8·20 (46) 7·78 (12) 0·31 (2) n.d. 12·91 (31) 21·23 (16) 0·51 (5) 0·01 (1) n.d. 99·17 0·747 amp 8 41·59 (31) 1·86 (4) n.d. 15·72 (19) 10·32 (14) 0·24 (3) n.d. 14·02 (4) 11·13 (9) 2·47 (5) 0·42 (2) n.d. 97·76 0·708 rk71 990 glass 10 52·58 (58) 0·73 (5) n.d. 18·04 (7) 5·29 (26) 0·21 (4) n.d. 2·09 (11) 7·38 (12) 2·79 (19) 0·74 (5) n.d. 89·86 0·413 cpx 9 46·30 (61) 0·96 (9) n.d. 9·26 (67) 9·17 (14) 0·33 (5) n.d. 11·22 (34) 21·74 (13) 0·57 (3) 0·01 (1) n.d. 99·56 0·686 amp 10 40·79 (41) 2·02 (11) n.d. 15·72 (24) 12·50 (23) 0·26 (4) n.d. 12·05 (23) 11·51 (7) 2·49 (8) 0·39 (2) n.d. 97·72 0·632 PU1049 990 glass 11 49·67 (19) 0·76 (4) n.d. 18·46 (19) 5·46 (9) 0·20 (4) n.d. 3·23 (9) 7·43 (20) 2·85 (11) 0·72 (4) 0·16 (6) 88·94 0·513 amp 12 40·96 (39) 2·00 (18) n.d. 15·89 (39) 10·87 (42) 0·23 (3) n.d. 14·21 (36) 10·96 (37) 2·28 (4) 0·39 (4) 0·01 (1) 97·81 0·700 mag 8 0·13 (1) 3·11 (8) n.d. 15·83 (48) 69·13 (76) 0·45 (2) n.d. 6·16 (13) 0·16 (2) 0·01 (1) 0·01 (1) 0·01 (1) 94·98 0·137 PU1070 980 glass 22 48·58 (18) 0·84 (9) n.d. 18·20 (20) 5·20 (20) 0·23 (7) n.d. 2·79 (24) 7·58 (10) 2·74 (6) 0·74 (6) 0·24 (4) 87·14 0·488 amp 30 40·98 (69) 2·14 (9) n.d. 15·46 (32) 8·26 (25) 0·23 (5) n.d. 13·91 (35) 11·02 (28) 2·34 (15) 0·43 (6) n.d. 94·76 0·750 PU1062 950 glass 41 52·17 (60) 0·76 (7) n.d. 19·60 (39) 4·97 (19) 0·19 (2) n.d. 2·24 (21) 7·43 (10) 3·10 (22) 0·81 (5) 0·18 (3) 91·24 0·446 amp 34 40·15 (33) 1·96 (9) n.d. 16·72 (44) 9·64 (20) 0·23 (1) n.d. 13·52 (26) 10·90 (14) 2·36 (4) 0·46 (2) 0·04 (2) 95·90 0·714 gar 15 37·96 (32) 0·61 (12) n.d. 22·61 (20) 17·13 (23) 1·33 (11) n.d. 10·54 (44) 8·48 (49) 0·03 (2) 0·01 (1) 0·05 (1) 98·73 0·523 PU1064 900 glass 25 55·65 (30) 0·51 (6) n.d. 17·86 (25) 3·93 (7) 0·10 (3) n.d. 1·13 (3) 6·05 (8) 3·15 (15) 1·05 (5) 0·24 (4) 89·67 0·339 amp 26 41·87 (68) 2·00 (20) n.d. 17·07 (52) 15·04 (26) 0·25 (2) n.d. 10·21 (7) 10·47 (18) 2·29 (5) 0·49 (3) 0·02 (1) 99·71 0·548 gar 20 39·75 (36) 1·20 (7) n.d. 22·07 (22) 21·65 (35) 1·60 (5) n.d. 6·77 (25) 9·65 (42) 0·02 (1) 0·01 (1) 0·02 (1) 102·75 0·358 plg 32 48·83 (83) 0·03 (2) n.d. 34·61 (52) 0·38 (2) 0·01 (1) n.d. 0·02 (1) 17·44 (46) 1·86 (15) 0·03 (1) 0·02 (2) 103·30 0·838 ilm 17 0·04 (1) 41·44 (104) n.d. 0·78 (5) 55·33 (171) 0·35 (3) n.d. 2·23 (15) 0·14 (3) 0·01 (1) 0·01 (1) 0 100·19 0·055 PU1066 850 glass 24 58·06 (38) 0·38 (2) n.d. 16·41 (28) 2·83 (9) 0·08 (3) n.d. 0·74 (4) 4·88 (9) 3·22 (12) 1·24 (4) 0·21 (4) 88·05 0·317 gar 19 38·73 (55) 1·06 (12) n.d. 21·40 (30) 23·62 (27) 1·34 (12) n.d. 5·40 (24) 9·19 (23) n.d. n.d. n.d. 100·73 0·289 plg 23 49·25 (78) n.d. n.d. 33·32 (48) 0·39 (15) n.d. n.d. 16·05 (40) 2·45 (16) 0·03 (2) n.d. n.d. 101·49 0·784 ilm 6 0·33 (4) 56·45 (54) n.d. 0·35 (4) 42·84 (48) 0·48 (6) n.d. 2·53 (6) 0·25 (4) n.d. n.d. n.d. 103·24 0·095 PU1068 800 glass 27 62·93 (32) 0·21 (4) n.d. 15·06 (16) 1·76 (9) 0·05 (2) n.d. 0·38 (2) 3·28 (6) 3·73 (10) 1·55 (5) 0·12 (3) 89·07 0·276 amp 16 43·99 (87) 1·67 (7) n.d. 13·97 (62) 18·77 (89) 0·45 (7) n.d. 8·03 (30) 9·75 (26) 2·03 (9) 0·38 (6) n.d. 99·05 0·433 gar 24 37·96 (51) 1·05 (18) n.d. 20·94 (22) 25·05 (67) 1·99 (16) n.d. 3·47 (26) 9·47 (66) n.d. n.d. n.d. 99·92 0·198 plg 21 57·12 (69) n.d. n.d. 28·67 (66) 0·27 (8) n.d. n.d. n.d. 10·25 (35) 5·97 (16) 0·03 (2) n.d. 102·32 0·487 ilm 10 0·31 (15) 49·53 (70) n.d. 0·41 (2) 47·98 (69) 0·41 (7) n.d. 1·41 (4) 0·15 (5) n.d. n.d. n.d. 100·19 0·050 PU1072 750 glass 10 61·92 (39) 0·04 (2) n.d. 13·22 (15) 0·87 (6) 0·10 (3) n.d. 0·40 (3) 2·44 (7) 3·21 (7) 1·81 (5) 0·11 (2) 84·09 0·449 plg 14 55·26 (66) n.d. n.d. 26·02 (32) 0·24 (6) n.d. n.d. n.d. 8·81 (25) 6·05 (11) 0·03 (3) n.d. 96·41 0·446 mag 7 0·65 (23) 15·90 (242) n.d. 1·28 (26) 68·57 (130) 0·20 (6) n.d. 0·40 (8) 0·16 (4) n.d. n.d. n.d. 87·15 0·010 gar 10 35·45 (61) 0·05 (5) n.d. 20·45 (31) 30·29 (166) 3·48 (10) n.d. 2·37 (33) 2·92 (40) n.d. n.d. n.d. 95·01 0·122 PU1072-o 750 720 glass 15 64·34 (41) 0·07 (3) n.d. 13·42 (14) 0·78 (10) 0·10 (2) n.d. 0·36 (3) 2·05 (6) 3·28 (10) 1·92 (6) 0·11 (3) 86·43 0·453 plg 21 59·04 (59) n.d. n.d. 25·50 (35) 0·24 (7) n.d. n.d. n.d. 7·46 (25) 7·05 (21) 0·05 (3) n.d. 99·33 0·369 amp 22 44·27 (96) 1·33 (18) n.d. 11·83 (70) 19·02 (43) 0·96 (14) n.d. 8·1 (29) 8·83 (23) 2·08 (13) 0·31 (10) n.d. 96·72 0·432 mag 6 1·43 (24) 18·81 (357) n.d. 1·48 (49) 66·64 (238) 0·27 (5) n.d. 0·46 (9) 0·15 (2) n.d. m.d. n.d. 89·25 0·012 qtz 4 99·18 (67) 99·18 FC Mb AuPd rk47 1230 glass 16 43·86 (43) 0·79 (3) 0·08 (3) 12·11 (44) 9·32 (18) 0·16 (5) 0·02(2) 12·24 (70) 10·92 (33) 1·27 (9) 0·44 (3) n.d. 91·37 0·701 ol 10 40·72 (31) 0·01 (1) 0·05 (3) 0·04 (1) 8·82 (16) 0·17 (3) 0·21 (2) 50·53 (19) 0·21 (2) 0·01 (1) 0·00 n.d. 100·77 0·911 sp 8 0·00 0·30 (6) 42·13 (585) 13·85 (275) 24·83 (164) 0·00 0·13 (3) 14·08 (39) 0·21 (9) 0·03 (3) 0·00 n.d. 95·56 0·501 rk51 1200 glass 10 43·66 (41) 0·89 (4) n.d. 14·02 (28) 9·00 (22) 0·20 (6) n.d. 10·36 (12) 12·35 (13) 1·27 (10) 0·53 (2) n.d. 92·27 0·672 ol 10 40·84 (15) 0·01 (1) n.d. 0·03 (1) 10·33 (19) 0·19 (3) n.d. 48·74 (26) 0·24 (2) 0·01 (1) 0·00 n.d. 100·39 0·894 cpx 9 52·08 (55) 0·23 (3) n.d. 3·78 (52) 4·30 (14) 0·11 (3) n.d. 17·39 (45) 21·68 (42) 0·21 (2) 0·00 n.d. 99·79 0·878 rk52 1170 glass 10 44·61 (56) 0·71 (2) n.d. 16·53 (32) 9·07 (24) 0·21 (4) n.d. 8·92 (27) 11·33 (24) 1·53 (14) 0·74 (4) n.d. 93·65 0·637 ol 10 40·66 (19) 0·01 (1) n.d. 0·05 (1) 12·23 (27) 0·25 (4) n.d. 46·77 (40) 0·24 (2) 0·00 0·00 n.d. 100·21 0·872 cpx 10 49·35 (80) 0·34 (6) n.d. 7·49 (76) 5·68 (69) 0·16 (2) n.d. 15·57 (47) 21·24 (59) 0·32 (3) 0·01 (1) n.d. 100·17 0·830 sp 2 0·11 (1) 0·12 (1) n.d. 59·39 (5) 18·96 (23) 0·18 (6) n.d. 19·93 (20) 0·06 (1) 0·01 (1) 0·00 n.d. 98·74 0·652 rk54 1140 glass 9 44·67 (34) 0·80 (6) n.d. 17·21 (27) 9·53 (22) 0·23 (6) n.d. 7·50 (14) 10·38 (12) 1·73 (12) 0·91 (3) n.d. 92·97 0·584 cpx 10 48·73 (51) 0·42 (4) n.d. 7·89 (48) 6·39 (29) 0·20 (2) n.d. 14·85 (24) 21·49 (16) 0·32 (4) 0·01 (1) n.d. 100·31 0·805 sp 9 0·10 (3) 0·18 (3) n.d. 57·74 (100) 21·73 (49) 0·16 (3) n.d. 18·47 (32) 0·14 (4) 0·01 (1) 0·00 n.d. 98·53 0·602 rk57 1110 glass 9 47·00 (35) 0·86 (5) n.d. 18·07 (25) 8·47 (26) 0·22 (5) n.d. 5·66 (31) 8·74 (20) 2·16 (12) 1·07 (7) n.d. 92·26 0·544 cpx 10 48·87 (43) 0·60 (5) n.d. 7·95 (51) 7·03 (31) 0·26 (3) n.d. 14·63 (24) 20·15 (36) 0·41 (4) 0·01 (1) n.d. 99·90 0·788 opx 10 51·69 (38) 0·18 (3) n.d. 6·86 (65) 12·28 (19) 0·35 (3) n.d. 27·30 (49) 1·48 (22) 0·05 (3) 0·02 (3) n.d. 100·21 0·798 sp 10 0·10 (7) 0·17 (6) n.d. 60·28 (58) 20·73 (51) 0·20 (2) n.d. 17·10 (19) 0·12 (3) 0·01 (2) 0·00 n.d. 98·71 0·595 rk58 1080 glass 10 48·43 (44) 0·91 (4) n.d. 17·18 (29) 8·40 (28) 0·30 (6) n.d. 4·35 (27) 7·80 (34) 2·42 (9) 1·34 (4) n.d. 91·13 0·480 cpx 8 45·65 (60) 0·96 (13) n.d. 10·28 (72) 9·00 (32) 0·36 (5) n.d. 12·22 (31) 20·06 (25) 0·55 (3) 0·01 (1) n.d. 99·09 0·708 opx 7 48·63 (44) 0·25 (3) n.d. 8·61 (47) 13·72 (47) 0·48 (2) n.d. 25·34 (32) 1·11 (13) 0·05 (2) 0·00 n.d. 98·20 0·767 mag 5 0·13 (2) 3·10 (6) n.d. 12·23 (8) 73·29 (53) 0·36 (6) n.d. 4·97 (9) 0·04 (5) 0·00 0·00 n.d. 94·11 0·108 rk73 1050 glass 10 49·86 (51) 0·86 (4) n.d. 17·85 (17) 8·70 (17) 0·30 (6) n.d. 3·78 (9) 7·90 (15) 2·29 (9) 1·18 (4) n.d. 92·71 0·437 cpx 1 47·41 0·84 n.d. 8·07 8·34 0·34 n.d. 13·73 20·74 0·44 0·08 n.d. 99·99 0·746 amp 9 41·26 (49) 1·92 (10) n.d. 15·73 (36) 10·45 (111) 0·28 (4) n.d. 14·27 (56) 11·34 (18) 2·20 (8) 0·84 (4) n.d. 98·29 0·709 rk65 1020 glass 10 52·17 (27) 0·87 (4) n.d. 17·18 (39) 7·07 (18) 0·28 (7) n.d. 2·70 (5) 6·78 (9) 2·41 (11) 1·35 (3) n.d. 90·81 0·405 cpx 10 47·50 (99) 0·75 (13) n.d. 6·94 (84) 10·88 (60) 0·49 (4) n.d. 12·31 (56) 20·04 (40) 0·51 (3) 0·01 (1) n.d. 99·42 0·669 amp 4 41·01 (44) 2·17 (10) n.d. 14·36 (26) 14·08 (20) 0·38 (2) n.d. 12·05 (12) 10·91 (26) 2·19 (7) 0·70 (5) n.d. 97·84 0·604 mag 5 0·12 (2) 5·96 (11) n.d. 7·98 (14) 76·44 (101) 0·54 (4) n.d. 3·03 (4) 0·04 (4) 0·00 0·00 n.d. 94·11 0·066 rk67 990 glass 10 54·74 (111) 0·59 (8) n.d. 15·94 (18) 5·59 (38) 0·31 (5) n.d. 1·48 (12) 5·68 (7) 2·71 (16) 1·33 (7) n.d. 88·36 0·321 cpx 4 46·31 (103) 0·77 (5) n.d. 7·25 (51) 13·02 (7) 0·65 (4) n.d. 10·30 (42) 19·48 (97) 0·75 (14) 0·02 (2) n.d. 98·54 0·585 amp 12 40·46 (59) 1·96 (21) n.d. 13·41 (31) 16·90 (60) 0·48 (3) n.d. 10·28 (34) 10·75 (23) 2·19 (9) 0·67 (4) n.d. 97·09 0·520 mag 4 0·15 (1) 6·10 (28) n.d. 5·66 (4) 79·74 (110) 0·70 (6) n.d. 1·49 (5) 0·17 (6) 0·01 (1) 0·01 (1) n.d. 94·02 0·032 PU1048 990 glass 13 52·64 (30) 0·86 (4) n.d. 16·47 (10) 5·58 (8) 0·29 (3) n.d. 2·69 (9) 6·60 (11) 2·61 (9) 1·32 (4) 0·22 (7) 89·28 0·462 cpx 4 50·08 (84) 0·83 (16) n.d. 5·05 (75) 8·18 (22) 0·59 (3) n.d. 14·63 (31) 19·48 (44) 0·32 (3) 0·01 (1) 0·01 (1) 99·17 0·761 PU1069 980 glass 13 50·97 (33) 0·92 (3) n.d. 16·65 (11) 6·17 (8) 0·31 (7) n.d. 2·35 (3) 6·82 (8) 2·66 (5) 1·54 (3) 0·23 (5) 88·62 0·405 cpx 19 48·70 (79) 0·77 (20) n.d. 5·11 (119) 8·65 (33) 0·71 (9) n.d. 13·23 (57) 19·35 (50) 0·46 (7) n.d. n.d. 97·00 0·732 PU1061 950 glass 29 53·24 (34) 0·90 (8) n.d. 17·69 (20) 6·10 (14) 0·27 (3) n.d. 2·21 (7) 6·83 (8) 2·91 (10) 1·52 (5) 0·21 (2) 91·87 0·393 amp 25 41·57 (16) 2·26 (18) n.d. 14·79 (30) 11·45 (42) 0·33 (4) n.d. 13·13 (20) 10·48 (20) 2·17 (6) 0·85 (4) 0·04 (2) 97·09 0·671 gar 10 38·75 (21) 0·55 (17) n.d. 22·38 (21) 19·13 (33) 2·09 (21) n.d. 9·23 (43) 7·58 (38) 0·03 (2) 0·01 (1) 0·05 (1) 99·79 0·462 cpx 7 51·60 (103) 0·65 (8) n.d. 5·50 (99) 10·62 (93) 0·84 (19) n.d. 12·99 (52) 18·89 (34) 0·53 (8) 0·04 (1) 0·01 (1) 101·66 0·685 PU1063 900 glass 12 54·23 (43) 0·48 (3) n.d. 16·86 (14) 3·74 (12) 0·09 (2) n.d. 0·97 (5) 5·28 (8) 2·92 (5) 1·70 (5) 0·28 (3) 86·56 0·317 amp 5 40·62 (44) 2·14 (18) n.d. 15·94 (21) 15·18 (21) 0·26 (4) n.d. 9·93 (17) 10·41 (28) 2·14 (8) 0·92 (5) 0·03 (1) 97·57 0·538 gar 8 36·83 (29) 1·37 (11) n.d. 21·06 (18) 22·12 (41) 1·69 (9) n.d. 6·06 (50) 9·29 (22) 0·04 (1) 0·01 (1) 0·06 (3) 98·83 0·328 plg 31 48·84 (102) 0·02 (1) n.d. 34·09 (104) 0·48 (8) 0·01 (1) n.d. 0·02 (1) 16·97 (58) 2·07 (16) 0·07 (1) 0·01 (1) 102·70 0·819 ilm 13 0·04 (1) 43·46 (124) n.d. 0·65 (5) 51·50 (115) 0·40 (3) n.d. 2·34 (10) 0·17 (2) 0·01 (1) 0·02 (1) 0 98·37 0·063 PU1065 850 glass 29 58·74 (30) 0·38 (3) n.d. 16·32 (27) 2·65 (12) 0·06 (3) n.d. 0·68 (3) 4·68 (6) 2·92 (5) 2·13 (4) 0·19 (4) 88·75 0·314 gar 35 38·50 (55) 1·21 (16) n.d. 21·43 (26) 23·42 (41) 1·46 (12) n.d. 5·42 (27) 8·92 (43) 0·04 (1) 0·01 (1) 0·07 (2) 100·48 0·290 plg 22 49·20 (40) 0·03 (1) n.d. 33·29 (57) 0·23 (3) 0·01 (1) n.d. 0·01 (1) 16·10 (52) 2·39 (13) 0·08 (1) 0·03 (1) 101·37 0·788 ilm 8 0·13 (3) 51·90 (15) n.d. 0·28 (2) 43·09 (8) 0·41 (2) n.d. 2·65 (5) 0·17 (2) 0·02 (2) 0·02 (1) 0·02 (2) 98·68 0·099 apa 6 0·31 (14) n.d. n.d. 0·06 (5) 0·73 (7) n.d. n.d. 0·20 (3) 53·80 (33) n.d. n.d. 45·07 (34) 99·92 PU1067 800 glass 12 59·77 (65) 0·18 (4) n.d. 14·49 (13) 1·65 (5) 0·03 (2) n.d. 0·34 (2) 3·10 (2) 3·46 (5) 2·62 (5) 0·11 (2) 85·75 0·268 amp 17 43·21 (125) 1·68 (19) n.d. 15·05 (90) 18·77 (66) 0·37 (8) n.d. 7·63 (36) 9·92 (28) 1·90 (7) 0·81 (12) n.d. 99·35 0·420 gar 19 38·26 (84) 1·03 (14) n.d. 21·13 (28) 25·46 (57) 1·64 (17) n.d. 3·54 (27) 9·26 (62) n.d. n.d. n.d. 100·32 0·198 plg 18 56·22 (65) n.d. n.d. 29·95 (56) 0·24 (7) n.d. n.d. n.d. 11·54 (41) 5·41 (26) 0·04 (3) n.d. 103·41 0·541 ilm 10 0·38 (6) 49·45 (67) n.d. 0·41 (4) 48·45 (80) 0·30 (9) n.d. 0·30 (9) 1·37 (5) 0·16 (4) n.d. n.d. 100·51 0·048 PU1071 750 glass 18 61·50 (24) 0·06 (2) n.d. 13·64 (8) 0·86 (9) 0·08 (2) n.d. 0·37 (2) 2·56 (7) 2·96 (10) 2·94 (7) 0·10 (3) 85·01 0·434 gar 16 35·99 (83) 0·03 (3) n.d. 20·49 (36) 34·64 (50) 1·84 (14) n.d. 2·74 (9) 0·80 (5) n.d. n.d. n.d. 96·52 0·124 plg 18 54·61 (83) n.d. n.d. 27·25 (25) 0·22 (8) n.d. n.d. n.d. 10·05 (20) 5·43 (9) 0·03 (2) n.d. 97·58 0·506 mag 14 0·32 (9) 18·37 (179) n.d. 0·99 (17) 67·75 (115) 0·25 (10) n.d. 0·13 (4) 0·49 (7) n.d. n.d. n.d. 88·30 0·003 FC Mb C-Pt rk3 1230 glass 8 46·75 (29) 0·88 (2) 0·15 (2) 15·00 (10) 8·70 (16) 0·14 (2) n.d. 10·59 (11) 12·20 (8) 1·47 (3) 0·49 (2) n.d. 96·37 0·712 ol 10 40·47 (20) 0·02 (2) 0·11 (2) 0·06 (1) 12·50 (17) 0·19 (3) n.d. 47·05 (26) 0·33 (2) 0·03 (4) 0·01 (1) n.d. 100·76 0·870 sp 4 0·33 (25) 0·39 (5) 38·19 (676) 31·29 (619) 14·43 (77) 0·00 n.d. 15·68 (140) 0·15 (9) 0·02 (1) 0·01 (1) n.d. 100·48 0·659 rk6 1200 glass 10 47·79 (13) 1·13 (4) n.d. 17·60 (15) 9·76 (16) 0·18 (2) n.d. 8·05 (24) 11·22 (6) 2·02 (6) 0·73 (1) n.d. 98·49 0·595 ol 11 39·97 (19) 0·02 (1) n.d. 0·07 (2) 16·96 (13) 0·23 (2) n.d. 44·24 (24) 0·34 (3) 0·01 (1) 0·01 (1) n.d. 101·85 0·823 cpx 12 51·00 (34) 0·51 (3) n.d. 7·21 (37) 5·65 (20) 0·17 (2) n.d. 17·11 (23) 19·81 (56) 0·31 (3) 0·01 (1) n.d. 101·79 0·844 rk11 1170 glass 10 46·41 (64) 1·19 (3) n.d. 17·99 (30) 9·17 (13) 0·22 (3) n.d. 7·27 (8) 10·10 (13) 2·09 (15) 0·74 (2) n.d. 95·17 0·586 cpx 10 49·39 (92) 0·60 (5) n.d. 7·97 (60) 5·70 (24) 0·20 (2) n.d. 15·70 (43) 19·49 (43) 0·38 (2) 0·01 (1) n.d. 99·42 0·831 rk13 1140 glass 7 48·07 (37) 1·39 (4) n.d. 18·43 (37) 10·58 (15) 0·21 (2) n.d. 5·56 (10) 8·69 (10) 2·65 (27) 1·08 (4) n.d. 96·65 0·484 cpx 7 48·64 (41) 0·92 (15) n.d. 8·41 (74) 8·67 (32) 0·27 (3) n.d. 15·15 (40) 17·17 (48) 0·50 (6) 0·02 (1) n.d. 99·75 0·757 opx 10 50·20 (66) 0·41 (3) n.d. 7·71 (41) 14·37 (20) 0·30 (2) n.d. 25·31 (36) 1·95 (28) 0·07 (2) 0·01 (1) n.d. 100·32 0·758 plg 5 44·12 (127) 0·09 (5) n.d. 34·08 (99) 0·59 (6) 0·02 (2) n.d. 0·16 (4) 18·46 (66) 0·95 (30) 0·07 (3) n.d. 98·55 0·911 sp 10 0·38 (27) 0·31 (2) n.d. 62·52 (74) 17·31 (15) 0·14 (2) n.d. 16·44 (22) 0·20 (6) 0·04 (3) 0·01 (1) n.d. 97·34 0·629 rk55 1110 glass 5 48·16 (15) 2·10 (5) n.d. 17·15 (35) 13·80 (67) 0·32 (6) n.d. 3·35 (11) 6·65 (16) 3·17 (4) 1·73 (9) n.d. 96·43 0·302 cpx 9 47·88 (37) 1·34 (11) n.d. 7·79 (48) 13·83 (65) 0·51 (6) n.d. 12·49 (31) 14·86 (43) 0·57 (5) 0·05 (4) n.d. 99·32 0·617 opx 9 49·33 (75) 0·62 (7) n.d. 6·29 (83) 20·93 (49) 0·56 (4) n.d. 19·59 (48) 2·29 (32) 0·10 (3) 0·01 (1) n.d. 99·72 0·625 plg (high An) 5 43·97 (125) 0·27 (13) n.d. 35·35 (142) 0·53 (5) 0·05 (2) n.d. 0·10 (8) 18·57 (120) 0·69 (46) 0·09 (8) n.d. 99·60 0·931 plg (low An) 4 52·93 (56) 0·08 (2) n.d. 29·68 (25) 0·55 (4) 0·01 (1) n.d. 0·06 (1) 11·99 (49) 4·48 (23) 0·42 (4) n.d. 100·19 0·582 rk64 1080 glass 10 51·19 (38) 1·90 (8) n.d. 16·27 (19) 12·22 (30) 0·20 (4) n.d. 1·89 (6) 5·27 (11) 3·92 (13) 2·67 (5) n.d. 95·51 0·216 cpx 9 47·32 (60) 1·17 (14) n.d. 6·07 (80) 19·60 (69) 0·51 (5) n.d. 10·88 (44) 12·87 (62) 0·60 (3) 0·03 (1) n.d. 99·05 0·497 plg 9 56·48 (78) 0·08 (1) n.d. 26·83 (38) 0·60 (9) 0·01 (1) n.d. 0·05 (2) 9·46 (61) 5·92 (29) 0·80 (8) n.d. 100·22 0·448 ilm 4 0·06 (1) 50·25 (56) n.d. 0·49 (2) 42·05 (35) 0·49 (4) n.d. 2·90 (5) 0·19 (2) 0·03 (2) 0·05 (1) n.d. 96·50 0·109 gar 6 37·63 (22) 1·22 (37) n.d. 20·42 (41) 25·05 (43) 1·09 (8) n.d. 6·49 (37) 7·55 (32) 0·03 (2) 0·00 n.d. 99·48 0·316 All Fe as FeO. Units in parentheses indicate standard deviations (2σ) from averaged analysis [i.e. 46·75(29) should be read as 46·75 ± 0·29 wt %]; #, number of individual spots averaged for each phase composition reported. Abbreviations are as for Table 2; xMg = molar [MgO/(MgO + FeOtot)]; xAn, anorthite content of plagioclase = molar [Ca/(Ca + Na + K)]. Open in new tab Table 4: Electron microprobe analyses of experimental phases of fractional crystallization experiments Run # . T (°C) . phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . P2O5 . Total . xMg . xAn . FC ba AuPd rk48 1230 glass 9 48·95 (33) 0·60 (4) n.d. 14·71 (35) 7·84 (23) 0·18 (7) n.d. 9·59 (16) 9·25 (8) 2·04 (14) 0·36 (3) n.d. 93·52 0·686 rk50 1200 glass 10 48·00 (48) 0·60 (5) n.d. 14·44 (31) 8·15 (13) 0·14 (5) n.d. 8·85 (11) 9·36 (14) 2·14 (16) 0·38 (2) n.d. 92·06 0·659 opx 10 55·15 (40) 0·08 (1) n.d. 3·11 (28) 7·74 (16) 0·18 (3) n.d. 32·37 (26) 1·48 (13) 0·03 (2) 0·00 n.d. 100·14 0·882 rk56 1170 glass 9 48·71 (27) 0·65 (3) n.d. 15·47 (18) 7·73 (29) 0·18 (4) n.d. 7·45 (19) 9·41 (21) 2·25 (14) 0·43 (3) n.d. 92·29 0·632 cpx 9 52·60 (38) 0·20 (2) n.d. 3·45 (50) 5·40 (18) 0·20 (3) n.d. 18·04 (49) 19·32 (49) 0·29 (12) 0·01 (1) n.d. 99·51 0·856 opx 10 55·39 (31) 0·11 (2) n.d. 2·81 (48) 9·17 (17) 0·25 (5) n.d. 30·94 (27) 1·78 (8) 0·05 (2) 0·01 (1) n.d. 100·50 0·857 rk60 1140 glass 10 50·09 (72) 0·69 (5) n.d. 15·71 (22) 7·22 (21) 0·18 (4) n.d. 6·64 (9) 9·23 (13) 2·11 (41) 0·47 (2) n.d. 92·34 0·621 cpx 10 52·73 (25) 0·21 (4) n.d. 3·00 (38) 5·30 (31) 0·21 (4) n.d. 17·99 (16) 20·00 (28) 0·34 (4) 0·01 (1) n.d. 99·80 0·857 opx 10 55·25 (63) 0·13 (3) n.d. 2·59 (29) 9·37 (31) 0·30 (3) n.d. 30·94 (37) 1·71 (17) 0·05 (2) 0·01 (1) n.d. 100·33 0·855 rk63 1110 glass 9 48·76 (28) 0·76 (5) n.d. 16·38 (27) 6·06 (19) 0·17 (5) n.d. 6·05 (11) 8·77 (10) 2·48 (12) 0·58 (3) n.d. 90·00 0·640 cpx 10 52·81 (44) 0·26 (4) n.d. 3·20 (47) 4·92 (35) 0·26 (5) n.d. 17·64 (46) 20·58 (65) 0·32 (3) 0·01 (1) n.d. 100·00 0·865 opx 10 54·79 (17) 0·15 (2) n.d. 3·09 (24) 9·18 (21) 0·32 (4) n.d. 30·59 (33) 1·56 (20) 0·04 (2) 0·01 (1) n.d. 99·73 0·856 rk66 1080 glass 10 49·38 (26) 0·83 (4) n.d. 17·68 (19) 6·14 (20) 0·17 (4) n.d. 4·47 (9) 7·78 (8) 2·80 (11) 0·66 (3) n.d. 89·89 0·565 cpx 8 49·12 (50) 0·60 (4) n.d. 6·99 (31) 6·26 (15) 0·22 (3) n.d. 14·78 (39) 20·44 (28) 0·48 (4) 0·02 (2) n.d. 98·90 0·808 opx 9 52·99 (37) 0·20 (3) n.d. 5·25 (73) 11·33 (22) 0·35 (3) n.d. 28·77 (34) 1·25 (9) 0·04 (2) 0·00 n.d. 100·19 0·819 rk69 1050 glass 10 49·66 (31) 0·81 (5) n.d. 17·91 (27) 6·20 (11) 0·20 (6) n.d. 3·78 (10) 7·85 (9) 2·48 (12) 0·68 (3) n.d. 89·57 0·521 cpx 6 48·88 (86) 0·69 (4) n.d. 6·86 (33) 6·86 (31) 0·26 (3) n.d. 14·44 (35) 21·22 (32) 0·46 (4) 0·01 (1) n.d. 99·68 0·789 amp 10 42·13 (65) 1·66 (5) n.d. 14·40 (40) 9·52 (42) 0·18 (3) n.d. 15·32 (38) 11·26 (19) 2·47 (5) 0·43 (2) n.d. 97·36 0·742 rk70 1020 glass 10 50·16 (65) 0·81 (3) n.d. 18·22 (34) 6·04 (16) 0·21 (5) n.d. 3·01 (9) 7·61 (17) 2·57 (13) 0·72 (2) n.d. 89·34 0·470 cpx 7 47·37 (38) 0·84 (4) n.d. 8·20 (46) 7·78 (12) 0·31 (2) n.d. 12·91 (31) 21·23 (16) 0·51 (5) 0·01 (1) n.d. 99·17 0·747 amp 8 41·59 (31) 1·86 (4) n.d. 15·72 (19) 10·32 (14) 0·24 (3) n.d. 14·02 (4) 11·13 (9) 2·47 (5) 0·42 (2) n.d. 97·76 0·708 rk71 990 glass 10 52·58 (58) 0·73 (5) n.d. 18·04 (7) 5·29 (26) 0·21 (4) n.d. 2·09 (11) 7·38 (12) 2·79 (19) 0·74 (5) n.d. 89·86 0·413 cpx 9 46·30 (61) 0·96 (9) n.d. 9·26 (67) 9·17 (14) 0·33 (5) n.d. 11·22 (34) 21·74 (13) 0·57 (3) 0·01 (1) n.d. 99·56 0·686 amp 10 40·79 (41) 2·02 (11) n.d. 15·72 (24) 12·50 (23) 0·26 (4) n.d. 12·05 (23) 11·51 (7) 2·49 (8) 0·39 (2) n.d. 97·72 0·632 PU1049 990 glass 11 49·67 (19) 0·76 (4) n.d. 18·46 (19) 5·46 (9) 0·20 (4) n.d. 3·23 (9) 7·43 (20) 2·85 (11) 0·72 (4) 0·16 (6) 88·94 0·513 amp 12 40·96 (39) 2·00 (18) n.d. 15·89 (39) 10·87 (42) 0·23 (3) n.d. 14·21 (36) 10·96 (37) 2·28 (4) 0·39 (4) 0·01 (1) 97·81 0·700 mag 8 0·13 (1) 3·11 (8) n.d. 15·83 (48) 69·13 (76) 0·45 (2) n.d. 6·16 (13) 0·16 (2) 0·01 (1) 0·01 (1) 0·01 (1) 94·98 0·137 PU1070 980 glass 22 48·58 (18) 0·84 (9) n.d. 18·20 (20) 5·20 (20) 0·23 (7) n.d. 2·79 (24) 7·58 (10) 2·74 (6) 0·74 (6) 0·24 (4) 87·14 0·488 amp 30 40·98 (69) 2·14 (9) n.d. 15·46 (32) 8·26 (25) 0·23 (5) n.d. 13·91 (35) 11·02 (28) 2·34 (15) 0·43 (6) n.d. 94·76 0·750 PU1062 950 glass 41 52·17 (60) 0·76 (7) n.d. 19·60 (39) 4·97 (19) 0·19 (2) n.d. 2·24 (21) 7·43 (10) 3·10 (22) 0·81 (5) 0·18 (3) 91·24 0·446 amp 34 40·15 (33) 1·96 (9) n.d. 16·72 (44) 9·64 (20) 0·23 (1) n.d. 13·52 (26) 10·90 (14) 2·36 (4) 0·46 (2) 0·04 (2) 95·90 0·714 gar 15 37·96 (32) 0·61 (12) n.d. 22·61 (20) 17·13 (23) 1·33 (11) n.d. 10·54 (44) 8·48 (49) 0·03 (2) 0·01 (1) 0·05 (1) 98·73 0·523 PU1064 900 glass 25 55·65 (30) 0·51 (6) n.d. 17·86 (25) 3·93 (7) 0·10 (3) n.d. 1·13 (3) 6·05 (8) 3·15 (15) 1·05 (5) 0·24 (4) 89·67 0·339 amp 26 41·87 (68) 2·00 (20) n.d. 17·07 (52) 15·04 (26) 0·25 (2) n.d. 10·21 (7) 10·47 (18) 2·29 (5) 0·49 (3) 0·02 (1) 99·71 0·548 gar 20 39·75 (36) 1·20 (7) n.d. 22·07 (22) 21·65 (35) 1·60 (5) n.d. 6·77 (25) 9·65 (42) 0·02 (1) 0·01 (1) 0·02 (1) 102·75 0·358 plg 32 48·83 (83) 0·03 (2) n.d. 34·61 (52) 0·38 (2) 0·01 (1) n.d. 0·02 (1) 17·44 (46) 1·86 (15) 0·03 (1) 0·02 (2) 103·30 0·838 ilm 17 0·04 (1) 41·44 (104) n.d. 0·78 (5) 55·33 (171) 0·35 (3) n.d. 2·23 (15) 0·14 (3) 0·01 (1) 0·01 (1) 0 100·19 0·055 PU1066 850 glass 24 58·06 (38) 0·38 (2) n.d. 16·41 (28) 2·83 (9) 0·08 (3) n.d. 0·74 (4) 4·88 (9) 3·22 (12) 1·24 (4) 0·21 (4) 88·05 0·317 gar 19 38·73 (55) 1·06 (12) n.d. 21·40 (30) 23·62 (27) 1·34 (12) n.d. 5·40 (24) 9·19 (23) n.d. n.d. n.d. 100·73 0·289 plg 23 49·25 (78) n.d. n.d. 33·32 (48) 0·39 (15) n.d. n.d. 16·05 (40) 2·45 (16) 0·03 (2) n.d. n.d. 101·49 0·784 ilm 6 0·33 (4) 56·45 (54) n.d. 0·35 (4) 42·84 (48) 0·48 (6) n.d. 2·53 (6) 0·25 (4) n.d. n.d. n.d. 103·24 0·095 PU1068 800 glass 27 62·93 (32) 0·21 (4) n.d. 15·06 (16) 1·76 (9) 0·05 (2) n.d. 0·38 (2) 3·28 (6) 3·73 (10) 1·55 (5) 0·12 (3) 89·07 0·276 amp 16 43·99 (87) 1·67 (7) n.d. 13·97 (62) 18·77 (89) 0·45 (7) n.d. 8·03 (30) 9·75 (26) 2·03 (9) 0·38 (6) n.d. 99·05 0·433 gar 24 37·96 (51) 1·05 (18) n.d. 20·94 (22) 25·05 (67) 1·99 (16) n.d. 3·47 (26) 9·47 (66) n.d. n.d. n.d. 99·92 0·198 plg 21 57·12 (69) n.d. n.d. 28·67 (66) 0·27 (8) n.d. n.d. n.d. 10·25 (35) 5·97 (16) 0·03 (2) n.d. 102·32 0·487 ilm 10 0·31 (15) 49·53 (70) n.d. 0·41 (2) 47·98 (69) 0·41 (7) n.d. 1·41 (4) 0·15 (5) n.d. n.d. n.d. 100·19 0·050 PU1072 750 glass 10 61·92 (39) 0·04 (2) n.d. 13·22 (15) 0·87 (6) 0·10 (3) n.d. 0·40 (3) 2·44 (7) 3·21 (7) 1·81 (5) 0·11 (2) 84·09 0·449 plg 14 55·26 (66) n.d. n.d. 26·02 (32) 0·24 (6) n.d. n.d. n.d. 8·81 (25) 6·05 (11) 0·03 (3) n.d. 96·41 0·446 mag 7 0·65 (23) 15·90 (242) n.d. 1·28 (26) 68·57 (130) 0·20 (6) n.d. 0·40 (8) 0·16 (4) n.d. n.d. n.d. 87·15 0·010 gar 10 35·45 (61) 0·05 (5) n.d. 20·45 (31) 30·29 (166) 3·48 (10) n.d. 2·37 (33) 2·92 (40) n.d. n.d. n.d. 95·01 0·122 PU1072-o 750 720 glass 15 64·34 (41) 0·07 (3) n.d. 13·42 (14) 0·78 (10) 0·10 (2) n.d. 0·36 (3) 2·05 (6) 3·28 (10) 1·92 (6) 0·11 (3) 86·43 0·453 plg 21 59·04 (59) n.d. n.d. 25·50 (35) 0·24 (7) n.d. n.d. n.d. 7·46 (25) 7·05 (21) 0·05 (3) n.d. 99·33 0·369 amp 22 44·27 (96) 1·33 (18) n.d. 11·83 (70) 19·02 (43) 0·96 (14) n.d. 8·1 (29) 8·83 (23) 2·08 (13) 0·31 (10) n.d. 96·72 0·432 mag 6 1·43 (24) 18·81 (357) n.d. 1·48 (49) 66·64 (238) 0·27 (5) n.d. 0·46 (9) 0·15 (2) n.d. m.d. n.d. 89·25 0·012 qtz 4 99·18 (67) 99·18 FC Mb AuPd rk47 1230 glass 16 43·86 (43) 0·79 (3) 0·08 (3) 12·11 (44) 9·32 (18) 0·16 (5) 0·02(2) 12·24 (70) 10·92 (33) 1·27 (9) 0·44 (3) n.d. 91·37 0·701 ol 10 40·72 (31) 0·01 (1) 0·05 (3) 0·04 (1) 8·82 (16) 0·17 (3) 0·21 (2) 50·53 (19) 0·21 (2) 0·01 (1) 0·00 n.d. 100·77 0·911 sp 8 0·00 0·30 (6) 42·13 (585) 13·85 (275) 24·83 (164) 0·00 0·13 (3) 14·08 (39) 0·21 (9) 0·03 (3) 0·00 n.d. 95·56 0·501 rk51 1200 glass 10 43·66 (41) 0·89 (4) n.d. 14·02 (28) 9·00 (22) 0·20 (6) n.d. 10·36 (12) 12·35 (13) 1·27 (10) 0·53 (2) n.d. 92·27 0·672 ol 10 40·84 (15) 0·01 (1) n.d. 0·03 (1) 10·33 (19) 0·19 (3) n.d. 48·74 (26) 0·24 (2) 0·01 (1) 0·00 n.d. 100·39 0·894 cpx 9 52·08 (55) 0·23 (3) n.d. 3·78 (52) 4·30 (14) 0·11 (3) n.d. 17·39 (45) 21·68 (42) 0·21 (2) 0·00 n.d. 99·79 0·878 rk52 1170 glass 10 44·61 (56) 0·71 (2) n.d. 16·53 (32) 9·07 (24) 0·21 (4) n.d. 8·92 (27) 11·33 (24) 1·53 (14) 0·74 (4) n.d. 93·65 0·637 ol 10 40·66 (19) 0·01 (1) n.d. 0·05 (1) 12·23 (27) 0·25 (4) n.d. 46·77 (40) 0·24 (2) 0·00 0·00 n.d. 100·21 0·872 cpx 10 49·35 (80) 0·34 (6) n.d. 7·49 (76) 5·68 (69) 0·16 (2) n.d. 15·57 (47) 21·24 (59) 0·32 (3) 0·01 (1) n.d. 100·17 0·830 sp 2 0·11 (1) 0·12 (1) n.d. 59·39 (5) 18·96 (23) 0·18 (6) n.d. 19·93 (20) 0·06 (1) 0·01 (1) 0·00 n.d. 98·74 0·652 rk54 1140 glass 9 44·67 (34) 0·80 (6) n.d. 17·21 (27) 9·53 (22) 0·23 (6) n.d. 7·50 (14) 10·38 (12) 1·73 (12) 0·91 (3) n.d. 92·97 0·584 cpx 10 48·73 (51) 0·42 (4) n.d. 7·89 (48) 6·39 (29) 0·20 (2) n.d. 14·85 (24) 21·49 (16) 0·32 (4) 0·01 (1) n.d. 100·31 0·805 sp 9 0·10 (3) 0·18 (3) n.d. 57·74 (100) 21·73 (49) 0·16 (3) n.d. 18·47 (32) 0·14 (4) 0·01 (1) 0·00 n.d. 98·53 0·602 rk57 1110 glass 9 47·00 (35) 0·86 (5) n.d. 18·07 (25) 8·47 (26) 0·22 (5) n.d. 5·66 (31) 8·74 (20) 2·16 (12) 1·07 (7) n.d. 92·26 0·544 cpx 10 48·87 (43) 0·60 (5) n.d. 7·95 (51) 7·03 (31) 0·26 (3) n.d. 14·63 (24) 20·15 (36) 0·41 (4) 0·01 (1) n.d. 99·90 0·788 opx 10 51·69 (38) 0·18 (3) n.d. 6·86 (65) 12·28 (19) 0·35 (3) n.d. 27·30 (49) 1·48 (22) 0·05 (3) 0·02 (3) n.d. 100·21 0·798 sp 10 0·10 (7) 0·17 (6) n.d. 60·28 (58) 20·73 (51) 0·20 (2) n.d. 17·10 (19) 0·12 (3) 0·01 (2) 0·00 n.d. 98·71 0·595 rk58 1080 glass 10 48·43 (44) 0·91 (4) n.d. 17·18 (29) 8·40 (28) 0·30 (6) n.d. 4·35 (27) 7·80 (34) 2·42 (9) 1·34 (4) n.d. 91·13 0·480 cpx 8 45·65 (60) 0·96 (13) n.d. 10·28 (72) 9·00 (32) 0·36 (5) n.d. 12·22 (31) 20·06 (25) 0·55 (3) 0·01 (1) n.d. 99·09 0·708 opx 7 48·63 (44) 0·25 (3) n.d. 8·61 (47) 13·72 (47) 0·48 (2) n.d. 25·34 (32) 1·11 (13) 0·05 (2) 0·00 n.d. 98·20 0·767 mag 5 0·13 (2) 3·10 (6) n.d. 12·23 (8) 73·29 (53) 0·36 (6) n.d. 4·97 (9) 0·04 (5) 0·00 0·00 n.d. 94·11 0·108 rk73 1050 glass 10 49·86 (51) 0·86 (4) n.d. 17·85 (17) 8·70 (17) 0·30 (6) n.d. 3·78 (9) 7·90 (15) 2·29 (9) 1·18 (4) n.d. 92·71 0·437 cpx 1 47·41 0·84 n.d. 8·07 8·34 0·34 n.d. 13·73 20·74 0·44 0·08 n.d. 99·99 0·746 amp 9 41·26 (49) 1·92 (10) n.d. 15·73 (36) 10·45 (111) 0·28 (4) n.d. 14·27 (56) 11·34 (18) 2·20 (8) 0·84 (4) n.d. 98·29 0·709 rk65 1020 glass 10 52·17 (27) 0·87 (4) n.d. 17·18 (39) 7·07 (18) 0·28 (7) n.d. 2·70 (5) 6·78 (9) 2·41 (11) 1·35 (3) n.d. 90·81 0·405 cpx 10 47·50 (99) 0·75 (13) n.d. 6·94 (84) 10·88 (60) 0·49 (4) n.d. 12·31 (56) 20·04 (40) 0·51 (3) 0·01 (1) n.d. 99·42 0·669 amp 4 41·01 (44) 2·17 (10) n.d. 14·36 (26) 14·08 (20) 0·38 (2) n.d. 12·05 (12) 10·91 (26) 2·19 (7) 0·70 (5) n.d. 97·84 0·604 mag 5 0·12 (2) 5·96 (11) n.d. 7·98 (14) 76·44 (101) 0·54 (4) n.d. 3·03 (4) 0·04 (4) 0·00 0·00 n.d. 94·11 0·066 rk67 990 glass 10 54·74 (111) 0·59 (8) n.d. 15·94 (18) 5·59 (38) 0·31 (5) n.d. 1·48 (12) 5·68 (7) 2·71 (16) 1·33 (7) n.d. 88·36 0·321 cpx 4 46·31 (103) 0·77 (5) n.d. 7·25 (51) 13·02 (7) 0·65 (4) n.d. 10·30 (42) 19·48 (97) 0·75 (14) 0·02 (2) n.d. 98·54 0·585 amp 12 40·46 (59) 1·96 (21) n.d. 13·41 (31) 16·90 (60) 0·48 (3) n.d. 10·28 (34) 10·75 (23) 2·19 (9) 0·67 (4) n.d. 97·09 0·520 mag 4 0·15 (1) 6·10 (28) n.d. 5·66 (4) 79·74 (110) 0·70 (6) n.d. 1·49 (5) 0·17 (6) 0·01 (1) 0·01 (1) n.d. 94·02 0·032 PU1048 990 glass 13 52·64 (30) 0·86 (4) n.d. 16·47 (10) 5·58 (8) 0·29 (3) n.d. 2·69 (9) 6·60 (11) 2·61 (9) 1·32 (4) 0·22 (7) 89·28 0·462 cpx 4 50·08 (84) 0·83 (16) n.d. 5·05 (75) 8·18 (22) 0·59 (3) n.d. 14·63 (31) 19·48 (44) 0·32 (3) 0·01 (1) 0·01 (1) 99·17 0·761 PU1069 980 glass 13 50·97 (33) 0·92 (3) n.d. 16·65 (11) 6·17 (8) 0·31 (7) n.d. 2·35 (3) 6·82 (8) 2·66 (5) 1·54 (3) 0·23 (5) 88·62 0·405 cpx 19 48·70 (79) 0·77 (20) n.d. 5·11 (119) 8·65 (33) 0·71 (9) n.d. 13·23 (57) 19·35 (50) 0·46 (7) n.d. n.d. 97·00 0·732 PU1061 950 glass 29 53·24 (34) 0·90 (8) n.d. 17·69 (20) 6·10 (14) 0·27 (3) n.d. 2·21 (7) 6·83 (8) 2·91 (10) 1·52 (5) 0·21 (2) 91·87 0·393 amp 25 41·57 (16) 2·26 (18) n.d. 14·79 (30) 11·45 (42) 0·33 (4) n.d. 13·13 (20) 10·48 (20) 2·17 (6) 0·85 (4) 0·04 (2) 97·09 0·671 gar 10 38·75 (21) 0·55 (17) n.d. 22·38 (21) 19·13 (33) 2·09 (21) n.d. 9·23 (43) 7·58 (38) 0·03 (2) 0·01 (1) 0·05 (1) 99·79 0·462 cpx 7 51·60 (103) 0·65 (8) n.d. 5·50 (99) 10·62 (93) 0·84 (19) n.d. 12·99 (52) 18·89 (34) 0·53 (8) 0·04 (1) 0·01 (1) 101·66 0·685 PU1063 900 glass 12 54·23 (43) 0·48 (3) n.d. 16·86 (14) 3·74 (12) 0·09 (2) n.d. 0·97 (5) 5·28 (8) 2·92 (5) 1·70 (5) 0·28 (3) 86·56 0·317 amp 5 40·62 (44) 2·14 (18) n.d. 15·94 (21) 15·18 (21) 0·26 (4) n.d. 9·93 (17) 10·41 (28) 2·14 (8) 0·92 (5) 0·03 (1) 97·57 0·538 gar 8 36·83 (29) 1·37 (11) n.d. 21·06 (18) 22·12 (41) 1·69 (9) n.d. 6·06 (50) 9·29 (22) 0·04 (1) 0·01 (1) 0·06 (3) 98·83 0·328 plg 31 48·84 (102) 0·02 (1) n.d. 34·09 (104) 0·48 (8) 0·01 (1) n.d. 0·02 (1) 16·97 (58) 2·07 (16) 0·07 (1) 0·01 (1) 102·70 0·819 ilm 13 0·04 (1) 43·46 (124) n.d. 0·65 (5) 51·50 (115) 0·40 (3) n.d. 2·34 (10) 0·17 (2) 0·01 (1) 0·02 (1) 0 98·37 0·063 PU1065 850 glass 29 58·74 (30) 0·38 (3) n.d. 16·32 (27) 2·65 (12) 0·06 (3) n.d. 0·68 (3) 4·68 (6) 2·92 (5) 2·13 (4) 0·19 (4) 88·75 0·314 gar 35 38·50 (55) 1·21 (16) n.d. 21·43 (26) 23·42 (41) 1·46 (12) n.d. 5·42 (27) 8·92 (43) 0·04 (1) 0·01 (1) 0·07 (2) 100·48 0·290 plg 22 49·20 (40) 0·03 (1) n.d. 33·29 (57) 0·23 (3) 0·01 (1) n.d. 0·01 (1) 16·10 (52) 2·39 (13) 0·08 (1) 0·03 (1) 101·37 0·788 ilm 8 0·13 (3) 51·90 (15) n.d. 0·28 (2) 43·09 (8) 0·41 (2) n.d. 2·65 (5) 0·17 (2) 0·02 (2) 0·02 (1) 0·02 (2) 98·68 0·099 apa 6 0·31 (14) n.d. n.d. 0·06 (5) 0·73 (7) n.d. n.d. 0·20 (3) 53·80 (33) n.d. n.d. 45·07 (34) 99·92 PU1067 800 glass 12 59·77 (65) 0·18 (4) n.d. 14·49 (13) 1·65 (5) 0·03 (2) n.d. 0·34 (2) 3·10 (2) 3·46 (5) 2·62 (5) 0·11 (2) 85·75 0·268 amp 17 43·21 (125) 1·68 (19) n.d. 15·05 (90) 18·77 (66) 0·37 (8) n.d. 7·63 (36) 9·92 (28) 1·90 (7) 0·81 (12) n.d. 99·35 0·420 gar 19 38·26 (84) 1·03 (14) n.d. 21·13 (28) 25·46 (57) 1·64 (17) n.d. 3·54 (27) 9·26 (62) n.d. n.d. n.d. 100·32 0·198 plg 18 56·22 (65) n.d. n.d. 29·95 (56) 0·24 (7) n.d. n.d. n.d. 11·54 (41) 5·41 (26) 0·04 (3) n.d. 103·41 0·541 ilm 10 0·38 (6) 49·45 (67) n.d. 0·41 (4) 48·45 (80) 0·30 (9) n.d. 0·30 (9) 1·37 (5) 0·16 (4) n.d. n.d. 100·51 0·048 PU1071 750 glass 18 61·50 (24) 0·06 (2) n.d. 13·64 (8) 0·86 (9) 0·08 (2) n.d. 0·37 (2) 2·56 (7) 2·96 (10) 2·94 (7) 0·10 (3) 85·01 0·434 gar 16 35·99 (83) 0·03 (3) n.d. 20·49 (36) 34·64 (50) 1·84 (14) n.d. 2·74 (9) 0·80 (5) n.d. n.d. n.d. 96·52 0·124 plg 18 54·61 (83) n.d. n.d. 27·25 (25) 0·22 (8) n.d. n.d. n.d. 10·05 (20) 5·43 (9) 0·03 (2) n.d. 97·58 0·506 mag 14 0·32 (9) 18·37 (179) n.d. 0·99 (17) 67·75 (115) 0·25 (10) n.d. 0·13 (4) 0·49 (7) n.d. n.d. n.d. 88·30 0·003 FC Mb C-Pt rk3 1230 glass 8 46·75 (29) 0·88 (2) 0·15 (2) 15·00 (10) 8·70 (16) 0·14 (2) n.d. 10·59 (11) 12·20 (8) 1·47 (3) 0·49 (2) n.d. 96·37 0·712 ol 10 40·47 (20) 0·02 (2) 0·11 (2) 0·06 (1) 12·50 (17) 0·19 (3) n.d. 47·05 (26) 0·33 (2) 0·03 (4) 0·01 (1) n.d. 100·76 0·870 sp 4 0·33 (25) 0·39 (5) 38·19 (676) 31·29 (619) 14·43 (77) 0·00 n.d. 15·68 (140) 0·15 (9) 0·02 (1) 0·01 (1) n.d. 100·48 0·659 rk6 1200 glass 10 47·79 (13) 1·13 (4) n.d. 17·60 (15) 9·76 (16) 0·18 (2) n.d. 8·05 (24) 11·22 (6) 2·02 (6) 0·73 (1) n.d. 98·49 0·595 ol 11 39·97 (19) 0·02 (1) n.d. 0·07 (2) 16·96 (13) 0·23 (2) n.d. 44·24 (24) 0·34 (3) 0·01 (1) 0·01 (1) n.d. 101·85 0·823 cpx 12 51·00 (34) 0·51 (3) n.d. 7·21 (37) 5·65 (20) 0·17 (2) n.d. 17·11 (23) 19·81 (56) 0·31 (3) 0·01 (1) n.d. 101·79 0·844 rk11 1170 glass 10 46·41 (64) 1·19 (3) n.d. 17·99 (30) 9·17 (13) 0·22 (3) n.d. 7·27 (8) 10·10 (13) 2·09 (15) 0·74 (2) n.d. 95·17 0·586 cpx 10 49·39 (92) 0·60 (5) n.d. 7·97 (60) 5·70 (24) 0·20 (2) n.d. 15·70 (43) 19·49 (43) 0·38 (2) 0·01 (1) n.d. 99·42 0·831 rk13 1140 glass 7 48·07 (37) 1·39 (4) n.d. 18·43 (37) 10·58 (15) 0·21 (2) n.d. 5·56 (10) 8·69 (10) 2·65 (27) 1·08 (4) n.d. 96·65 0·484 cpx 7 48·64 (41) 0·92 (15) n.d. 8·41 (74) 8·67 (32) 0·27 (3) n.d. 15·15 (40) 17·17 (48) 0·50 (6) 0·02 (1) n.d. 99·75 0·757 opx 10 50·20 (66) 0·41 (3) n.d. 7·71 (41) 14·37 (20) 0·30 (2) n.d. 25·31 (36) 1·95 (28) 0·07 (2) 0·01 (1) n.d. 100·32 0·758 plg 5 44·12 (127) 0·09 (5) n.d. 34·08 (99) 0·59 (6) 0·02 (2) n.d. 0·16 (4) 18·46 (66) 0·95 (30) 0·07 (3) n.d. 98·55 0·911 sp 10 0·38 (27) 0·31 (2) n.d. 62·52 (74) 17·31 (15) 0·14 (2) n.d. 16·44 (22) 0·20 (6) 0·04 (3) 0·01 (1) n.d. 97·34 0·629 rk55 1110 glass 5 48·16 (15) 2·10 (5) n.d. 17·15 (35) 13·80 (67) 0·32 (6) n.d. 3·35 (11) 6·65 (16) 3·17 (4) 1·73 (9) n.d. 96·43 0·302 cpx 9 47·88 (37) 1·34 (11) n.d. 7·79 (48) 13·83 (65) 0·51 (6) n.d. 12·49 (31) 14·86 (43) 0·57 (5) 0·05 (4) n.d. 99·32 0·617 opx 9 49·33 (75) 0·62 (7) n.d. 6·29 (83) 20·93 (49) 0·56 (4) n.d. 19·59 (48) 2·29 (32) 0·10 (3) 0·01 (1) n.d. 99·72 0·625 plg (high An) 5 43·97 (125) 0·27 (13) n.d. 35·35 (142) 0·53 (5) 0·05 (2) n.d. 0·10 (8) 18·57 (120) 0·69 (46) 0·09 (8) n.d. 99·60 0·931 plg (low An) 4 52·93 (56) 0·08 (2) n.d. 29·68 (25) 0·55 (4) 0·01 (1) n.d. 0·06 (1) 11·99 (49) 4·48 (23) 0·42 (4) n.d. 100·19 0·582 rk64 1080 glass 10 51·19 (38) 1·90 (8) n.d. 16·27 (19) 12·22 (30) 0·20 (4) n.d. 1·89 (6) 5·27 (11) 3·92 (13) 2·67 (5) n.d. 95·51 0·216 cpx 9 47·32 (60) 1·17 (14) n.d. 6·07 (80) 19·60 (69) 0·51 (5) n.d. 10·88 (44) 12·87 (62) 0·60 (3) 0·03 (1) n.d. 99·05 0·497 plg 9 56·48 (78) 0·08 (1) n.d. 26·83 (38) 0·60 (9) 0·01 (1) n.d. 0·05 (2) 9·46 (61) 5·92 (29) 0·80 (8) n.d. 100·22 0·448 ilm 4 0·06 (1) 50·25 (56) n.d. 0·49 (2) 42·05 (35) 0·49 (4) n.d. 2·90 (5) 0·19 (2) 0·03 (2) 0·05 (1) n.d. 96·50 0·109 gar 6 37·63 (22) 1·22 (37) n.d. 20·42 (41) 25·05 (43) 1·09 (8) n.d. 6·49 (37) 7·55 (32) 0·03 (2) 0·00 n.d. 99·48 0·316 Run # . T (°C) . phase . # . SiO2 . TiO2 . Cr2O3 . Al2O3 . FeO . MnO . NiO . MgO . CaO . Na2O . K2O . P2O5 . Total . xMg . xAn . FC ba AuPd rk48 1230 glass 9 48·95 (33) 0·60 (4) n.d. 14·71 (35) 7·84 (23) 0·18 (7) n.d. 9·59 (16) 9·25 (8) 2·04 (14) 0·36 (3) n.d. 93·52 0·686 rk50 1200 glass 10 48·00 (48) 0·60 (5) n.d. 14·44 (31) 8·15 (13) 0·14 (5) n.d. 8·85 (11) 9·36 (14) 2·14 (16) 0·38 (2) n.d. 92·06 0·659 opx 10 55·15 (40) 0·08 (1) n.d. 3·11 (28) 7·74 (16) 0·18 (3) n.d. 32·37 (26) 1·48 (13) 0·03 (2) 0·00 n.d. 100·14 0·882 rk56 1170 glass 9 48·71 (27) 0·65 (3) n.d. 15·47 (18) 7·73 (29) 0·18 (4) n.d. 7·45 (19) 9·41 (21) 2·25 (14) 0·43 (3) n.d. 92·29 0·632 cpx 9 52·60 (38) 0·20 (2) n.d. 3·45 (50) 5·40 (18) 0·20 (3) n.d. 18·04 (49) 19·32 (49) 0·29 (12) 0·01 (1) n.d. 99·51 0·856 opx 10 55·39 (31) 0·11 (2) n.d. 2·81 (48) 9·17 (17) 0·25 (5) n.d. 30·94 (27) 1·78 (8) 0·05 (2) 0·01 (1) n.d. 100·50 0·857 rk60 1140 glass 10 50·09 (72) 0·69 (5) n.d. 15·71 (22) 7·22 (21) 0·18 (4) n.d. 6·64 (9) 9·23 (13) 2·11 (41) 0·47 (2) n.d. 92·34 0·621 cpx 10 52·73 (25) 0·21 (4) n.d. 3·00 (38) 5·30 (31) 0·21 (4) n.d. 17·99 (16) 20·00 (28) 0·34 (4) 0·01 (1) n.d. 99·80 0·857 opx 10 55·25 (63) 0·13 (3) n.d. 2·59 (29) 9·37 (31) 0·30 (3) n.d. 30·94 (37) 1·71 (17) 0·05 (2) 0·01 (1) n.d. 100·33 0·855 rk63 1110 glass 9 48·76 (28) 0·76 (5) n.d. 16·38 (27) 6·06 (19) 0·17 (5) n.d. 6·05 (11) 8·77 (10) 2·48 (12) 0·58 (3) n.d. 90·00 0·640 cpx 10 52·81 (44) 0·26 (4) n.d. 3·20 (47) 4·92 (35) 0·26 (5) n.d. 17·64 (46) 20·58 (65) 0·32 (3) 0·01 (1) n.d. 100·00 0·865 opx 10 54·79 (17) 0·15 (2) n.d. 3·09 (24) 9·18 (21) 0·32 (4) n.d. 30·59 (33) 1·56 (20) 0·04 (2) 0·01 (1) n.d. 99·73 0·856 rk66 1080 glass 10 49·38 (26) 0·83 (4) n.d. 17·68 (19) 6·14 (20) 0·17 (4) n.d. 4·47 (9) 7·78 (8) 2·80 (11) 0·66 (3) n.d. 89·89 0·565 cpx 8 49·12 (50) 0·60 (4) n.d. 6·99 (31) 6·26 (15) 0·22 (3) n.d. 14·78 (39) 20·44 (28) 0·48 (4) 0·02 (2) n.d. 98·90 0·808 opx 9 52·99 (37) 0·20 (3) n.d. 5·25 (73) 11·33 (22) 0·35 (3) n.d. 28·77 (34) 1·25 (9) 0·04 (2) 0·00 n.d. 100·19 0·819 rk69 1050 glass 10 49·66 (31) 0·81 (5) n.d. 17·91 (27) 6·20 (11) 0·20 (6) n.d. 3·78 (10) 7·85 (9) 2·48 (12) 0·68 (3) n.d. 89·57 0·521 cpx 6 48·88 (86) 0·69 (4) n.d. 6·86 (33) 6·86 (31) 0·26 (3) n.d. 14·44 (35) 21·22 (32) 0·46 (4) 0·01 (1) n.d. 99·68 0·789 amp 10 42·13 (65) 1·66 (5) n.d. 14·40 (40) 9·52 (42) 0·18 (3) n.d. 15·32 (38) 11·26 (19) 2·47 (5) 0·43 (2) n.d. 97·36 0·742 rk70 1020 glass 10 50·16 (65) 0·81 (3) n.d. 18·22 (34) 6·04 (16) 0·21 (5) n.d. 3·01 (9) 7·61 (17) 2·57 (13) 0·72 (2) n.d. 89·34 0·470 cpx 7 47·37 (38) 0·84 (4) n.d. 8·20 (46) 7·78 (12) 0·31 (2) n.d. 12·91 (31) 21·23 (16) 0·51 (5) 0·01 (1) n.d. 99·17 0·747 amp 8 41·59 (31) 1·86 (4) n.d. 15·72 (19) 10·32 (14) 0·24 (3) n.d. 14·02 (4) 11·13 (9) 2·47 (5) 0·42 (2) n.d. 97·76 0·708 rk71 990 glass 10 52·58 (58) 0·73 (5) n.d. 18·04 (7) 5·29 (26) 0·21 (4) n.d. 2·09 (11) 7·38 (12) 2·79 (19) 0·74 (5) n.d. 89·86 0·413 cpx 9 46·30 (61) 0·96 (9) n.d. 9·26 (67) 9·17 (14) 0·33 (5) n.d. 11·22 (34) 21·74 (13) 0·57 (3) 0·01 (1) n.d. 99·56 0·686 amp 10 40·79 (41) 2·02 (11) n.d. 15·72 (24) 12·50 (23) 0·26 (4) n.d. 12·05 (23) 11·51 (7) 2·49 (8) 0·39 (2) n.d. 97·72 0·632 PU1049 990 glass 11 49·67 (19) 0·76 (4) n.d. 18·46 (19) 5·46 (9) 0·20 (4) n.d. 3·23 (9) 7·43 (20) 2·85 (11) 0·72 (4) 0·16 (6) 88·94 0·513 amp 12 40·96 (39) 2·00 (18) n.d. 15·89 (39) 10·87 (42) 0·23 (3) n.d. 14·21 (36) 10·96 (37) 2·28 (4) 0·39 (4) 0·01 (1) 97·81 0·700 mag 8 0·13 (1) 3·11 (8) n.d. 15·83 (48) 69·13 (76) 0·45 (2) n.d. 6·16 (13) 0·16 (2) 0·01 (1) 0·01 (1) 0·01 (1) 94·98 0·137 PU1070 980 glass 22 48·58 (18) 0·84 (9) n.d. 18·20 (20) 5·20 (20) 0·23 (7) n.d. 2·79 (24) 7·58 (10) 2·74 (6) 0·74 (6) 0·24 (4) 87·14 0·488 amp 30 40·98 (69) 2·14 (9) n.d. 15·46 (32) 8·26 (25) 0·23 (5) n.d. 13·91 (35) 11·02 (28) 2·34 (15) 0·43 (6) n.d. 94·76 0·750 PU1062 950 glass 41 52·17 (60) 0·76 (7) n.d. 19·60 (39) 4·97 (19) 0·19 (2) n.d. 2·24 (21) 7·43 (10) 3·10 (22) 0·81 (5) 0·18 (3) 91·24 0·446 amp 34 40·15 (33) 1·96 (9) n.d. 16·72 (44) 9·64 (20) 0·23 (1) n.d. 13·52 (26) 10·90 (14) 2·36 (4) 0·46 (2) 0·04 (2) 95·90 0·714 gar 15 37·96 (32) 0·61 (12) n.d. 22·61 (20) 17·13 (23) 1·33 (11) n.d. 10·54 (44) 8·48 (49) 0·03 (2) 0·01 (1) 0·05 (1) 98·73 0·523 PU1064 900 glass 25 55·65 (30) 0·51 (6) n.d. 17·86 (25) 3·93 (7) 0·10 (3) n.d. 1·13 (3) 6·05 (8) 3·15 (15) 1·05 (5) 0·24 (4) 89·67 0·339 amp 26 41·87 (68) 2·00 (20) n.d. 17·07 (52) 15·04 (26) 0·25 (2) n.d. 10·21 (7) 10·47 (18) 2·29 (5) 0·49 (3) 0·02 (1) 99·71 0·548 gar 20 39·75 (36) 1·20 (7) n.d. 22·07 (22) 21·65 (35) 1·60 (5) n.d. 6·77 (25) 9·65 (42) 0·02 (1) 0·01 (1) 0·02 (1) 102·75 0·358 plg 32 48·83 (83) 0·03 (2) n.d. 34·61 (52) 0·38 (2) 0·01 (1) n.d. 0·02 (1) 17·44 (46) 1·86 (15) 0·03 (1) 0·02 (2) 103·30 0·838 ilm 17 0·04 (1) 41·44 (104) n.d. 0·78 (5) 55·33 (171) 0·35 (3) n.d. 2·23 (15) 0·14 (3) 0·01 (1) 0·01 (1) 0 100·19 0·055 PU1066 850 glass 24 58·06 (38) 0·38 (2) n.d. 16·41 (28) 2·83 (9) 0·08 (3) n.d. 0·74 (4) 4·88 (9) 3·22 (12) 1·24 (4) 0·21 (4) 88·05 0·317 gar 19 38·73 (55) 1·06 (12) n.d. 21·40 (30) 23·62 (27) 1·34 (12) n.d. 5·40 (24) 9·19 (23) n.d. n.d. n.d. 100·73 0·289 plg 23 49·25 (78) n.d. n.d. 33·32 (48) 0·39 (15) n.d. n.d. 16·05 (40) 2·45 (16) 0·03 (2) n.d. n.d. 101·49 0·784 ilm 6 0·33 (4) 56·45 (54) n.d. 0·35 (4) 42·84 (48) 0·48 (6) n.d. 2·53 (6) 0·25 (4) n.d. n.d. n.d. 103·24 0·095 PU1068 800 glass 27 62·93 (32) 0·21 (4) n.d. 15·06 (16) 1·76 (9) 0·05 (2) n.d. 0·38 (2) 3·28 (6) 3·73 (10) 1·55 (5) 0·12 (3) 89·07 0·276 amp 16 43·99 (87) 1·67 (7) n.d. 13·97 (62) 18·77 (89) 0·45 (7) n.d. 8·03 (30) 9·75 (26) 2·03 (9) 0·38 (6) n.d. 99·05 0·433 gar 24 37·96 (51) 1·05 (18) n.d. 20·94 (22) 25·05 (67) 1·99 (16) n.d. 3·47 (26) 9·47 (66) n.d. n.d. n.d. 99·92 0·198 plg 21 57·12 (69) n.d. n.d. 28·67 (66) 0·27 (8) n.d. n.d. n.d. 10·25 (35) 5·97 (16) 0·03 (2) n.d. 102·32 0·487 ilm 10 0·31 (15) 49·53 (70) n.d. 0·41 (2) 47·98 (69) 0·41 (7) n.d. 1·41 (4) 0·15 (5) n.d. n.d. n.d. 100·19 0·050 PU1072 750 glass 10 61·92 (39) 0·04 (2) n.d. 13·22 (15) 0·87 (6) 0·10 (3) n.d. 0·40 (3) 2·44 (7) 3·21 (7) 1·81 (5) 0·11 (2) 84·09 0·449 plg 14 55·26 (66) n.d. n.d. 26·02 (32) 0·24 (6) n.d. n.d. n.d. 8·81 (25) 6·05 (11) 0·03 (3) n.d. 96·41 0·446 mag 7 0·65 (23) 15·90 (242) n.d. 1·28 (26) 68·57 (130) 0·20 (6) n.d. 0·40 (8) 0·16 (4) n.d. n.d. n.d. 87·15 0·010 gar 10 35·45 (61) 0·05 (5) n.d. 20·45 (31) 30·29 (166) 3·48 (10) n.d. 2·37 (33) 2·92 (40) n.d. n.d. n.d. 95·01 0·122 PU1072-o 750 720 glass 15 64·34 (41) 0·07 (3) n.d. 13·42 (14) 0·78 (10) 0·10 (2) n.d. 0·36 (3) 2·05 (6) 3·28 (10) 1·92 (6) 0·11 (3) 86·43 0·453 plg 21 59·04 (59) n.d. n.d. 25·50 (35) 0·24 (7) n.d. n.d. n.d. 7·46 (25) 7·05 (21) 0·05 (3) n.d. 99·33 0·369 amp 22 44·27 (96) 1·33 (18) n.d. 11·83 (70) 19·02 (43) 0·96 (14) n.d. 8·1 (29) 8·83 (23) 2·08 (13) 0·31 (10) n.d. 96·72 0·432 mag 6 1·43 (24) 18·81 (357) n.d. 1·48 (49) 66·64 (238) 0·27 (5) n.d. 0·46 (9) 0·15 (2) n.d. m.d. n.d. 89·25 0·012 qtz 4 99·18 (67) 99·18 FC Mb AuPd rk47 1230 glass 16 43·86 (43) 0·79 (3) 0·08 (3) 12·11 (44) 9·32 (18) 0·16 (5) 0·02(2) 12·24 (70) 10·92 (33) 1·27 (9) 0·44 (3) n.d. 91·37 0·701 ol 10 40·72 (31) 0·01 (1) 0·05 (3) 0·04 (1) 8·82 (16) 0·17 (3) 0·21 (2) 50·53 (19) 0·21 (2) 0·01 (1) 0·00 n.d. 100·77 0·911 sp 8 0·00 0·30 (6) 42·13 (585) 13·85 (275) 24·83 (164) 0·00 0·13 (3) 14·08 (39) 0·21 (9) 0·03 (3) 0·00 n.d. 95·56 0·501 rk51 1200 glass 10 43·66 (41) 0·89 (4) n.d. 14·02 (28) 9·00 (22) 0·20 (6) n.d. 10·36 (12) 12·35 (13) 1·27 (10) 0·53 (2) n.d. 92·27 0·672 ol 10 40·84 (15) 0·01 (1) n.d. 0·03 (1) 10·33 (19) 0·19 (3) n.d. 48·74 (26) 0·24 (2) 0·01 (1) 0·00 n.d. 100·39 0·894 cpx 9 52·08 (55) 0·23 (3) n.d. 3·78 (52) 4·30 (14) 0·11 (3) n.d. 17·39 (45) 21·68 (42) 0·21 (2) 0·00 n.d. 99·79 0·878 rk52 1170 glass 10 44·61 (56) 0·71 (2) n.d. 16·53 (32) 9·07 (24) 0·21 (4) n.d. 8·92 (27) 11·33 (24) 1·53 (14) 0·74 (4) n.d. 93·65 0·637 ol 10 40·66 (19) 0·01 (1) n.d. 0·05 (1) 12·23 (27) 0·25 (4) n.d. 46·77 (40) 0·24 (2) 0·00 0·00 n.d. 100·21 0·872 cpx 10 49·35 (80) 0·34 (6) n.d. 7·49 (76) 5·68 (69) 0·16 (2) n.d. 15·57 (47) 21·24 (59) 0·32 (3) 0·01 (1) n.d. 100·17 0·830 sp 2 0·11 (1) 0·12 (1) n.d. 59·39 (5) 18·96 (23) 0·18 (6) n.d. 19·93 (20) 0·06 (1) 0·01 (1) 0·00 n.d. 98·74 0·652 rk54 1140 glass 9 44·67 (34) 0·80 (6) n.d. 17·21 (27) 9·53 (22) 0·23 (6) n.d. 7·50 (14) 10·38 (12) 1·73 (12) 0·91 (3) n.d. 92·97 0·584 cpx 10 48·73 (51) 0·42 (4) n.d. 7·89 (48) 6·39 (29) 0·20 (2) n.d. 14·85 (24) 21·49 (16) 0·32 (4) 0·01 (1) n.d. 100·31 0·805 sp 9 0·10 (3) 0·18 (3) n.d. 57·74 (100) 21·73 (49) 0·16 (3) n.d. 18·47 (32) 0·14 (4) 0·01 (1) 0·00 n.d. 98·53 0·602 rk57 1110 glass 9 47·00 (35) 0·86 (5) n.d. 18·07 (25) 8·47 (26) 0·22 (5) n.d. 5·66 (31) 8·74 (20) 2·16 (12) 1·07 (7) n.d. 92·26 0·544 cpx 10 48·87 (43) 0·60 (5) n.d. 7·95 (51) 7·03 (31) 0·26 (3) n.d. 14·63 (24) 20·15 (36) 0·41 (4) 0·01 (1) n.d. 99·90 0·788 opx 10 51·69 (38) 0·18 (3) n.d. 6·86 (65) 12·28 (19) 0·35 (3) n.d. 27·30 (49) 1·48 (22) 0·05 (3) 0·02 (3) n.d. 100·21 0·798 sp 10 0·10 (7) 0·17 (6) n.d. 60·28 (58) 20·73 (51) 0·20 (2) n.d. 17·10 (19) 0·12 (3) 0·01 (2) 0·00 n.d. 98·71 0·595 rk58 1080 glass 10 48·43 (44) 0·91 (4) n.d. 17·18 (29) 8·40 (28) 0·30 (6) n.d. 4·35 (27) 7·80 (34) 2·42 (9) 1·34 (4) n.d. 91·13 0·480 cpx 8 45·65 (60) 0·96 (13) n.d. 10·28 (72) 9·00 (32) 0·36 (5) n.d. 12·22 (31) 20·06 (25) 0·55 (3) 0·01 (1) n.d. 99·09 0·708 opx 7 48·63 (44) 0·25 (3) n.d. 8·61 (47) 13·72 (47) 0·48 (2) n.d. 25·34 (32) 1·11 (13) 0·05 (2) 0·00 n.d. 98·20 0·767 mag 5 0·13 (2) 3·10 (6) n.d. 12·23 (8) 73·29 (53) 0·36 (6) n.d. 4·97 (9) 0·04 (5) 0·00 0·00 n.d. 94·11 0·108 rk73 1050 glass 10 49·86 (51) 0·86 (4) n.d. 17·85 (17) 8·70 (17) 0·30 (6) n.d. 3·78 (9) 7·90 (15) 2·29 (9) 1·18 (4) n.d. 92·71 0·437 cpx 1 47·41 0·84 n.d. 8·07 8·34 0·34 n.d. 13·73 20·74 0·44 0·08 n.d. 99·99 0·746 amp 9 41·26 (49) 1·92 (10) n.d. 15·73 (36) 10·45 (111) 0·28 (4) n.d. 14·27 (56) 11·34 (18) 2·20 (8) 0·84 (4) n.d. 98·29 0·709 rk65 1020 glass 10 52·17 (27) 0·87 (4) n.d. 17·18 (39) 7·07 (18) 0·28 (7) n.d. 2·70 (5) 6·78 (9) 2·41 (11) 1·35 (3) n.d. 90·81 0·405 cpx 10 47·50 (99) 0·75 (13) n.d. 6·94 (84) 10·88 (60) 0·49 (4) n.d. 12·31 (56) 20·04 (40) 0·51 (3) 0·01 (1) n.d. 99·42 0·669 amp 4 41·01 (44) 2·17 (10) n.d. 14·36 (26) 14·08 (20) 0·38 (2) n.d. 12·05 (12) 10·91 (26) 2·19 (7) 0·70 (5) n.d. 97·84 0·604 mag 5 0·12 (2) 5·96 (11) n.d. 7·98 (14) 76·44 (101) 0·54 (4) n.d. 3·03 (4) 0·04 (4) 0·00 0·00 n.d. 94·11 0·066 rk67 990 glass 10 54·74 (111) 0·59 (8) n.d. 15·94 (18) 5·59 (38) 0·31 (5) n.d. 1·48 (12) 5·68 (7) 2·71 (16) 1·33 (7) n.d. 88·36 0·321 cpx 4 46·31 (103) 0·77 (5) n.d. 7·25 (51) 13·02 (7) 0·65 (4) n.d. 10·30 (42) 19·48 (97) 0·75 (14) 0·02 (2) n.d. 98·54 0·585 amp 12 40·46 (59) 1·96 (21) n.d. 13·41 (31) 16·90 (60) 0·48 (3) n.d. 10·28 (34) 10·75 (23) 2·19 (9) 0·67 (4) n.d. 97·09 0·520 mag 4 0·15 (1) 6·10 (28) n.d. 5·66 (4) 79·74 (110) 0·70 (6) n.d. 1·49 (5) 0·17 (6) 0·01 (1) 0·01 (1) n.d. 94·02 0·032 PU1048 990 glass 13 52·64 (30) 0·86 (4) n.d. 16·47 (10) 5·58 (8) 0·29 (3) n.d. 2·69 (9) 6·60 (11) 2·61 (9) 1·32 (4) 0·22 (7) 89·28 0·462 cpx 4 50·08 (84) 0·83 (16) n.d. 5·05 (75) 8·18 (22) 0·59 (3) n.d. 14·63 (31) 19·48 (44) 0·32 (3) 0·01 (1) 0·01 (1) 99·17 0·761 PU1069 980 glass 13 50·97 (33) 0·92 (3) n.d. 16·65 (11) 6·17 (8) 0·31 (7) n.d. 2·35 (3) 6·82 (8) 2·66 (5) 1·54 (3) 0·23 (5) 88·62 0·405 cpx 19 48·70 (79) 0·77 (20) n.d. 5·11 (119) 8·65 (33) 0·71 (9) n.d. 13·23 (57) 19·35 (50) 0·46 (7) n.d. n.d. 97·00 0·732 PU1061 950 glass 29 53·24 (34) 0·90 (8) n.d. 17·69 (20) 6·10 (14) 0·27 (3) n.d. 2·21 (7) 6·83 (8) 2·91 (10) 1·52 (5) 0·21 (2) 91·87 0·393 amp 25 41·57 (16) 2·26 (18) n.d. 14·79 (30) 11·45 (42) 0·33 (4) n.d. 13·13 (20) 10·48 (20) 2·17 (6) 0·85 (4) 0·04 (2) 97·09 0·671 gar 10 38·75 (21) 0·55 (17) n.d. 22·38 (21) 19·13 (33) 2·09 (21) n.d. 9·23 (43) 7·58 (38) 0·03 (2) 0·01 (1) 0·05 (1) 99·79 0·462 cpx 7 51·60 (103) 0·65 (8) n.d. 5·50 (99) 10·62 (93) 0·84 (19) n.d. 12·99 (52) 18·89 (34) 0·53 (8) 0·04 (1) 0·01 (1) 101·66 0·685 PU1063 900 glass 12 54·23 (43) 0·48 (3) n.d. 16·86 (14) 3·74 (12) 0·09 (2) n.d. 0·97 (5) 5·28 (8) 2·92 (5) 1·70 (5) 0·28 (3) 86·56 0·317 amp 5 40·62 (44) 2·14 (18) n.d. 15·94 (21) 15·18 (21) 0·26 (4) n.d. 9·93 (17) 10·41 (28) 2·14 (8) 0·92 (5) 0·03 (1) 97·57 0·538 gar 8 36·83 (29) 1·37 (11) n.d. 21·06 (18) 22·12 (41) 1·69 (9) n.d. 6·06 (50) 9·29 (22) 0·04 (1) 0·01 (1) 0·06 (3) 98·83 0·328 plg 31 48·84 (102) 0·02 (1) n.d. 34·09 (104) 0·48 (8) 0·01 (1) n.d. 0·02 (1) 16·97 (58) 2·07 (16) 0·07 (1) 0·01 (1) 102·70 0·819 ilm 13 0·04 (1) 43·46 (124) n.d. 0·65 (5) 51·50 (115) 0·40 (3) n.d. 2·34 (10) 0·17 (2) 0·01 (1) 0·02 (1) 0 98·37 0·063 PU1065 850 glass 29 58·74 (30) 0·38 (3) n.d. 16·32 (27) 2·65 (12) 0·06 (3) n.d. 0·68 (3) 4·68 (6) 2·92 (5) 2·13 (4) 0·19 (4) 88·75 0·314 gar 35 38·50 (55) 1·21 (16) n.d. 21·43 (26) 23·42 (41) 1·46 (12) n.d. 5·42 (27) 8·92 (43) 0·04 (1) 0·01 (1) 0·07 (2) 100·48 0·290 plg 22 49·20 (40) 0·03 (1) n.d. 33·29 (57) 0·23 (3) 0·01 (1) n.d. 0·01 (1) 16·10 (52) 2·39 (13) 0·08 (1) 0·03 (1) 101·37 0·788 ilm 8 0·13 (3) 51·90 (15) n.d. 0·28 (2) 43·09 (8) 0·41 (2) n.d. 2·65 (5) 0·17 (2) 0·02 (2) 0·02 (1) 0·02 (2) 98·68 0·099 apa 6 0·31 (14) n.d. n.d. 0·06 (5) 0·73 (7) n.d. n.d. 0·20 (3) 53·80 (33) n.d. n.d. 45·07 (34) 99·92 PU1067 800 glass 12 59·77 (65) 0·18 (4) n.d. 14·49 (13) 1·65 (5) 0·03 (2) n.d. 0·34 (2) 3·10 (2) 3·46 (5) 2·62 (5) 0·11 (2) 85·75 0·268 amp 17 43·21 (125) 1·68 (19) n.d. 15·05 (90) 18·77 (66) 0·37 (8) n.d. 7·63 (36) 9·92 (28) 1·90 (7) 0·81 (12) n.d. 99·35 0·420 gar 19 38·26 (84) 1·03 (14) n.d. 21·13 (28) 25·46 (57) 1·64 (17) n.d. 3·54 (27) 9·26 (62) n.d. n.d. n.d. 100·32 0·198 plg 18 56·22 (65) n.d. n.d. 29·95 (56) 0·24 (7) n.d. n.d. n.d. 11·54 (41) 5·41 (26) 0·04 (3) n.d. 103·41 0·541 ilm 10 0·38 (6) 49·45 (67) n.d. 0·41 (4) 48·45 (80) 0·30 (9) n.d. 0·30 (9) 1·37 (5) 0·16 (4) n.d. n.d. 100·51 0·048 PU1071 750 glass 18 61·50 (24) 0·06 (2) n.d. 13·64 (8) 0·86 (9) 0·08 (2) n.d. 0·37 (2) 2·56 (7) 2·96 (10) 2·94 (7) 0·10 (3) 85·01 0·434 gar 16 35·99 (83) 0·03 (3) n.d. 20·49 (36) 34·64 (50) 1·84 (14) n.d. 2·74 (9) 0·80 (5) n.d. n.d. n.d. 96·52 0·124 plg 18 54·61 (83) n.d. n.d. 27·25 (25) 0·22 (8) n.d. n.d. n.d. 10·05 (20) 5·43 (9) 0·03 (2) n.d. 97·58 0·506 mag 14 0·32 (9) 18·37 (179) n.d. 0·99 (17) 67·75 (115) 0·25 (10) n.d. 0·13 (4) 0·49 (7) n.d. n.d. n.d. 88·30 0·003 FC Mb C-Pt rk3 1230 glass 8 46·75 (29) 0·88 (2) 0·15 (2) 15·00 (10) 8·70 (16) 0·14 (2) n.d. 10·59 (11) 12·20 (8) 1·47 (3) 0·49 (2) n.d. 96·37 0·712 ol 10 40·47 (20) 0·02 (2) 0·11 (2) 0·06 (1) 12·50 (17) 0·19 (3) n.d. 47·05 (26) 0·33 (2) 0·03 (4) 0·01 (1) n.d. 100·76 0·870 sp 4 0·33 (25) 0·39 (5) 38·19 (676) 31·29 (619) 14·43 (77) 0·00 n.d. 15·68 (140) 0·15 (9) 0·02 (1) 0·01 (1) n.d. 100·48 0·659 rk6 1200 glass 10 47·79 (13) 1·13 (4) n.d. 17·60 (15) 9·76 (16) 0·18 (2) n.d. 8·05 (24) 11·22 (6) 2·02 (6) 0·73 (1) n.d. 98·49 0·595 ol 11 39·97 (19) 0·02 (1) n.d. 0·07 (2) 16·96 (13) 0·23 (2) n.d. 44·24 (24) 0·34 (3) 0·01 (1) 0·01 (1) n.d. 101·85 0·823 cpx 12 51·00 (34) 0·51 (3) n.d. 7·21 (37) 5·65 (20) 0·17 (2) n.d. 17·11 (23) 19·81 (56) 0·31 (3) 0·01 (1) n.d. 101·79 0·844 rk11 1170 glass 10 46·41 (64) 1·19 (3) n.d. 17·99 (30) 9·17 (13) 0·22 (3) n.d. 7·27 (8) 10·10 (13) 2·09 (15) 0·74 (2) n.d. 95·17 0·586 cpx 10 49·39 (92) 0·60 (5) n.d. 7·97 (60) 5·70 (24) 0·20 (2) n.d. 15·70 (43) 19·49 (43) 0·38 (2) 0·01 (1) n.d. 99·42 0·831 rk13 1140 glass 7 48·07 (37) 1·39 (4) n.d. 18·43 (37) 10·58 (15) 0·21 (2) n.d. 5·56 (10) 8·69 (10) 2·65 (27) 1·08 (4) n.d. 96·65 0·484 cpx 7 48·64 (41) 0·92 (15) n.d. 8·41 (74) 8·67 (32) 0·27 (3) n.d. 15·15 (40) 17·17 (48) 0·50 (6) 0·02 (1) n.d. 99·75 0·757 opx 10 50·20 (66) 0·41 (3) n.d. 7·71 (41) 14·37 (20) 0·30 (2) n.d. 25·31 (36) 1·95 (28) 0·07 (2) 0·01 (1) n.d. 100·32 0·758 plg 5 44·12 (127) 0·09 (5) n.d. 34·08 (99) 0·59 (6) 0·02 (2) n.d. 0·16 (4) 18·46 (66) 0·95 (30) 0·07 (3) n.d. 98·55 0·911 sp 10 0·38 (27) 0·31 (2) n.d. 62·52 (74) 17·31 (15) 0·14 (2) n.d. 16·44 (22) 0·20 (6) 0·04 (3) 0·01 (1) n.d. 97·34 0·629 rk55 1110 glass 5 48·16 (15) 2·10 (5) n.d. 17·15 (35) 13·80 (67) 0·32 (6) n.d. 3·35 (11) 6·65 (16) 3·17 (4) 1·73 (9) n.d. 96·43 0·302 cpx 9 47·88 (37) 1·34 (11) n.d. 7·79 (48) 13·83 (65) 0·51 (6) n.d. 12·49 (31) 14·86 (43) 0·57 (5) 0·05 (4) n.d. 99·32 0·617 opx 9 49·33 (75) 0·62 (7) n.d. 6·29 (83) 20·93 (49) 0·56 (4) n.d. 19·59 (48) 2·29 (32) 0·10 (3) 0·01 (1) n.d. 99·72 0·625 plg (high An) 5 43·97 (125) 0·27 (13) n.d. 35·35 (142) 0·53 (5) 0·05 (2) n.d. 0·10 (8) 18·57 (120) 0·69 (46) 0·09 (8) n.d. 99·60 0·931 plg (low An) 4 52·93 (56) 0·08 (2) n.d. 29·68 (25) 0·55 (4) 0·01 (1) n.d. 0·06 (1) 11·99 (49) 4·48 (23) 0·42 (4) n.d. 100·19 0·582 rk64 1080 glass 10 51·19 (38) 1·90 (8) n.d. 16·27 (19) 12·22 (30) 0·20 (4) n.d. 1·89 (6) 5·27 (11) 3·92 (13) 2·67 (5) n.d. 95·51 0·216 cpx 9 47·32 (60) 1·17 (14) n.d. 6·07 (80) 19·60 (69) 0·51 (5) n.d. 10·88 (44) 12·87 (62) 0·60 (3) 0·03 (1) n.d. 99·05 0·497 plg 9 56·48 (78) 0·08 (1) n.d. 26·83 (38) 0·60 (9) 0·01 (1) n.d. 0·05 (2) 9·46 (61) 5·92 (29) 0·80 (8) n.d. 100·22 0·448 ilm 4 0·06 (1) 50·25 (56) n.d. 0·49 (2) 42·05 (35) 0·49 (4) n.d. 2·90 (5) 0·19 (2) 0·03 (2) 0·05 (1) n.d. 96·50 0·109 gar 6 37·63 (22) 1·22 (37) n.d. 20·42 (41) 25·05 (43) 1·09 (8) n.d. 6·49 (37) 7·55 (32) 0·03 (2) 0·00 n.d. 99·48 0·316 All Fe as FeO. Units in parentheses indicate standard deviations (2σ) from averaged analysis [i.e. 46·75(29) should be read as 46·75 ± 0·29 wt %]; #, number of individual spots averaged for each phase composition reported. Abbreviations are as for Table 2; xMg = molar [MgO/(MgO + FeOtot)]; xAn, anorthite content of plagioclase = molar [Ca/(Ca + Na + K)]. Open in new tab General remarks The experiments reported in this study are equilibrium and fractional crystallization experiments performed over restricted temperature steps of mostly 30°C and 50°C below 1000°C; the equilibrium crystallization experiments under oxidized conditions (EQ Mb AuPd) used step sizes of 40°C with some additional steps between them (20°C) to explore phase relations, modes and compositions within steps that crystallized significant amounts of solids over a small temperature interval. Whereas fractional crystallization experiments under oxidizing conditions produced large amounts of liquids (>76%) facilitating analysis of the liquid using large electron beam sizes (10–20 μm) and reduced currents (7 nA) to minimize alkali migration in H2O-rich glasses, fractional crystallization experiments under reduced conditions (FC Mb Pt–C) resulted in lower melt fractions down to 52·7%. Equilibrium crystallization produces decreasing melts fractions with decreasing temperature resulting in very low melt fractions (<20%) at the lowest temperatures (980–1000°C), additionally complicated by quench crystallization (amphibole and cpx) that rendered analysis of the melt phase rather difficult. A prerequisite to successfully employing the capsule-in-capsule technique to constrain the fO2 conditions via the ferric–ferrous iron ratio in the starting material in the H2O-undersaturated, oxidized fractionation experiments is the preservation of a near closed system, in particular regarding potential iron and hydrogen loss or gain. Iron loss was evaluated via LSR analysis of the experimental charges. The resulting iron loss or gain (from the pre-saturated capsules) was less than 12·5% relative for all fractional crystallization experiments. Fe loss or gain of generally less than 8% is considered acceptable, thus, we consider the experiment as (nearly) closed system regarding iron. This also applies to the Pt–graphite double-capsule experiments, where all but three experiments lost less than 10% Fe relative, the three exceptions being the highest temperature run (PU72, 1350°C) and two replicates at 1120°C (equilibrium, P1121 and 1122), which lost 12 and 20% Fe relative. Hydrogen loss (or gain) by diffusion through the noble metal capsule will significantly influence the ambient fO2 conditions in the noble metal capsule experiments, either reducing it through hydrogen gain or increasing it by loss to the outer capsule or assembly (Luth, 1989). In the Pt–C experiments, hydrogen loss results in increased CO2 contents through the reaction H2O + C = H2 (lost) + CO2. A general feature of the experiments conducted under oxidizing conditions at temperatures of less than 1090°C is the presence of bubbles (Figs 1 and 2), which become more frequent and generally larger with decreasing temperature and increasing nominal H2O contents in the liquid phase (up to 11·6 wt %, PU1072). The solubility of pure H2O exceeds the amounts of H2O added to the starting material as well as the amounts calculated and determined for the residual liquids in the respective charges. The solubility of H2O in basalts exceeds 8 wt % already at 0·5 GPa and is estimated to be ∼18 wt % for basalts at 1 GPa (Newman & Lowenstern, 2002; Botcharnikov et al., 2005; Iacono-Marziano et al., 2012); for rhyolite the pure H2O solubility exceeds 15 wt % at 1 GPa, extrapolating the equation of Zhang (1999) beyond 0·8 GPa. The nominally added H2O contents amount to 3·0 ± 0·3 wt % for the initial high-Mg basalt and reach 11·6 wt % for the final rhyolite (PU1072, 750°C), taking into account the liquid and hydrous phase proportions (amphibole) in lower temperature runs. These values are considerably lower than the inferred solubility limits of pure H2O. The presence of bubbles is most probably caused by the presence of additional fluid components with much lower solubility than H2O, resulting in the exsolution of a mixed volatile phase at pressure–temperature conditions far below the H2O-saturation of the respective liquid phase. The most likely candidate is CO2, although this is not verified to date. We attempted to avoid any potential CO2 contamination of the starting material during preparation and welding by employing NaSi2O7 and KAlSi3O8 as sources for Na2O and K2O instead of commonly used carbonates, and used only CaCO3 as a calcium source initially. Using wollastonite (CaSiO3) did not change the results. The problem, however, might be related either to carbon introduced from the tip of the carbon electrode of the welder used in most of this study, the diffusion of carbon from the graphite furnace through the noble metal container (e.g. Brooker et al., 1998; Jakobsson, 2012), or absorbance of CO2 by the very-fine grained, and thus very reactive, starting material leading to the formation of hydromagnesite despite the fact that the starting material was stored in a desiccator and heated to 110°C before loading. CO2 contents required to saturate the liquids at 1·0 GPa with a mixed H2O–CO2 fluid range between 5200 ppm at 5·8 wt % H2O in a basalt and 7000 ppm at 11·6 wt % in a rhyolite based on calculations using VolatileCalc (Newman & Lowenstern, 2002), extrapolating beyond its nominal application range of 0·5 GPa. The natural high-Mg basalt RC158c contained 500 ppm CO2 (measured by coulometric titration); thus, even in this case additional CO2 had to be present during the experiments (EQ Mb AuPd series). The companion study at 0·7 GPa (Nandedkar et al., 2014) revealed bubbles in nominally H2O-undersaturated experiments, with the significant difference that in the latter study any carbonates were avoided in the starting materials for the fractional crystallization experiments and a W-based electrode has been used for welding the AuPd capsules, eliminating two of the four above-mentioned potential carbon sources. Alonso-Perez et al. (2009) determined the CO2 contents of their vitrified starting oxide mixes prepared exactly in the same way as the present synthetic starting material (they used composition F8a of this study) and obtained 9100 ± 200 ppm CO2 in their nominally CO2-free starting materials ranging in H2O content from 6·2 to 9·4 wt %. Fig. 1. Open in new tabDownload slide Back-scattered electron (BSE) images of run charges from equilibrium crystallization experiments on composition RC158c (Mb): (a) run PU926 (1160°C, AuPd capsules) composed of olivine (ol) and zoned Cr-rich spinel (spl) and quench high-Ca clinopyroxene (cpx); (b) run PU905 (1040°C, AuPd capsule) composed of ol, cpx, low-Ca orthopyroxene (opx) and Cr–Al–Fe-rich spinel; (c) run PU906 (980°C, AuPd capsule) showing cpx overgrown by amphibole (amp) coexisting with opx and spl (magnetite); (d) run ZP1210 (1210°C, Pt–C capsules) depicting Cr-rich spinel coexisting with ol that contains melt inclusions (MI); (e) run ZP1180 (1180°C, Pt–C capsules) showing sector-zoned cpx enclosing partly corroded ol; (f) run ZP1090 (1090°C, Pt–C capsules) composed of ol, cpx, opx and Al-rich (hercynitic) spinel; cpx enclosed in opx is resorbed. Fig. 1. Open in new tabDownload slide Back-scattered electron (BSE) images of run charges from equilibrium crystallization experiments on composition RC158c (Mb): (a) run PU926 (1160°C, AuPd capsules) composed of olivine (ol) and zoned Cr-rich spinel (spl) and quench high-Ca clinopyroxene (cpx); (b) run PU905 (1040°C, AuPd capsule) composed of ol, cpx, low-Ca orthopyroxene (opx) and Cr–Al–Fe-rich spinel; (c) run PU906 (980°C, AuPd capsule) showing cpx overgrown by amphibole (amp) coexisting with opx and spl (magnetite); (d) run ZP1210 (1210°C, Pt–C capsules) depicting Cr-rich spinel coexisting with ol that contains melt inclusions (MI); (e) run ZP1180 (1180°C, Pt–C capsules) showing sector-zoned cpx enclosing partly corroded ol; (f) run ZP1090 (1090°C, Pt–C capsules) composed of ol, cpx, opx and Al-rich (hercynitic) spinel; cpx enclosed in opx is resorbed. Fig. 2. Open in new tabDownload slide Back-scattered electron (BSE) images of run charges from fractional crystallization experiments. (a, b) BSE image and Ca–Kα distribution map of run RK70 (1020°C, ba) with high-Ca clinopyroxene (cpx) and amphibole (amp); where cpx is enclosed in amp it shows resorption; (c) BSE image of run PU1063 (950°C, Mb) exhibits garnet (gar) with seed cores (almandine) coexisting with amp, plagioclase (plg) and ilmenite (ilm); (d) BSE image of run PU1064 (900°C, ba) depicting gar, amp, ilm and plag; (e) BSE image of run PU1072 (750°C, ba) showing plg with seed cores (anorthite) coexisting with amp and magnetite (mag); (f) BSE image of the outer capsule of run PU1072o (720°C, ba) containing quartz (qtz) coexisting with plg, amp, ulvöspinel (usp) and unreacted garnet seeds (lacking idiomorphic overgrowth rims). Fig. 2. Open in new tabDownload slide Back-scattered electron (BSE) images of run charges from fractional crystallization experiments. (a, b) BSE image and Ca–Kα distribution map of run RK70 (1020°C, ba) with high-Ca clinopyroxene (cpx) and amphibole (amp); where cpx is enclosed in amp it shows resorption; (c) BSE image of run PU1063 (950°C, Mb) exhibits garnet (gar) with seed cores (almandine) coexisting with amp, plagioclase (plg) and ilmenite (ilm); (d) BSE image of run PU1064 (900°C, ba) depicting gar, amp, ilm and plag; (e) BSE image of run PU1072 (750°C, ba) showing plg with seed cores (anorthite) coexisting with amp and magnetite (mag); (f) BSE image of the outer capsule of run PU1072o (720°C, ba) containing quartz (qtz) coexisting with plg, amp, ulvöspinel (usp) and unreacted garnet seeds (lacking idiomorphic overgrowth rims). The case for the equilibrium crystallization experiments employing graphite capsules is different. The natural starting material contained only 500 ppm of CO2, but the starting material is rather oxidized and assuming that all ferric iron (3·25 wt % Fe2O3) is converted to ferrous iron via the reaction 2Fe2O3 + C = 4FeO + CO2 produces an additional 4500 ppm of CO2, resulting in 5000 ppm CO2 that will drive volatile saturation at 1·0 GPa to about 7·0 wt % H2O, consistent with the observation that run P1060 with 6·7 wt % H2O determined by Raman spectroscopy is just about fluid-saturated. The H2O contents calculated from the amount of H2O in the starting material with the H2O content obtained by the by-difference method using an internal mineral standard (clinopyroxene) for correction of the melt analysis totals is shown in Fig. 3. The diagram reveals a consistent variation along a 1:1 correlation up to 6 wt % corresponding to 1090°C. The highest values clearly deviate; the nominal values are higher than the calculated values. The upper limit of H2O contents around 10–11 wt % in the liquid is most probably the result of CO2 contamination limiting the amount of dissolved H2O in the silicate liquid. Fig. 3. Open in new tabDownload slide Comparison of H2O contents (wt %) determined by (a) the difference method and (b) Raman spectroscopy with nominal H2O contents calculated from the amount contained in the starting material, the melt fraction of the respective experiment corrected for amphibole modal amounts (assuming 2·1 wt % H2O in amphibole, H2O calc modal). In (a), microprobe totals of the glass phase were corrected to cpx totals using an external standard and assuming cpx totals are 99·6–99·9 wt %. In (b), numbers in parenthesis indicate H2O contents calculated for the same experiments with the Waters & Lange (2015) plagioclase–melt hygrometer. Raman result for experiment P1060 (10·9 wt % modal, 6·7 wt % Raman) is unreliable because of interference with additional (solid) phases. The fractional crystallization experiments under reducing conditions (FC Mb Pt–C) experienced significant H2O loss (for details see text). ‘Calib error’ denotes approximate accuracy of Raman H2O determination based on calibration with 20 in-house standards. Fig. 3. Open in new tabDownload slide Comparison of H2O contents (wt %) determined by (a) the difference method and (b) Raman spectroscopy with nominal H2O contents calculated from the amount contained in the starting material, the melt fraction of the respective experiment corrected for amphibole modal amounts (assuming 2·1 wt % H2O in amphibole, H2O calc modal). In (a), microprobe totals of the glass phase were corrected to cpx totals using an external standard and assuming cpx totals are 99·6–99·9 wt %. In (b), numbers in parenthesis indicate H2O contents calculated for the same experiments with the Waters & Lange (2015) plagioclase–melt hygrometer. Raman result for experiment P1060 (10·9 wt % modal, 6·7 wt % Raman) is unreliable because of interference with additional (solid) phases. The fractional crystallization experiments under reducing conditions (FC Mb Pt–C) experienced significant H2O loss (for details see text). ‘Calib error’ denotes approximate accuracy of Raman H2O determination based on calibration with 20 in-house standards. In summary, we observe small amounts of vapor or fluid bubbles in experiments conducted at temperatures <1080°C that we attribute to CO2 absorption by the very fine-grained oxide–silicate starting materials and/or the reaction of oxidized starting materials with graphite. However, the amount of calculated CO2 in our experiments is small compared with the amount of H2O (always <5 mol % of the total volatile concentration). With the exception of the fractional crystallization experiments in Pt–C double capsules, all experiments were successful in maintaining the attempted H2O contents in our charges and neither gained nor lost significant amounts of hydrogen through diffusion through the noble metal containers. This is demonstrated in Fig. 3b, which shows a comparison of the nominal H2O content calculated from the initially loaded amount of H2O (as hydroxide, rarely as additional H2O added by micro-syringe when the amount of hydroxide was insufficient) and the modal proportions of the liquid phase (and amphibole if present) with the H2O content determined by micro-Raman spectroscopy (Table 2). Maximum H2O contents of 11·6 wt % in the presence of a fluid phase for the final rhyolite at 750°C (PU1072) are consistent with a CO2 content of c. 7000 ppm. Overall estimated minimum CO2 contents for experiments that are fluid saturated (i.e. all experiments below 1080°C under oxidized conditions) are in the range of 4500–7000 ppm, increasing with decreasing temperature. As an additional test, we calculated the amount of H2O in the liquids coexisting with plagioclase utilizing the hygrometer of Waters & Lange (2015) for all runs that contained plagioclase. All fractional crystallization experiments that contained plagioclase consistently provided H2O contents about 10–20% lower than the values obtained by Raman spectroscopy and the nominal values (Fig. 3b), but confirmed that high H2O contents close to the nominal values were present in these experiments. However, fractional crystallization experiments in Pt–C capsules have lost most H2O and will be discussed below. The attainment of equilibrium in the fractional crystallization experiments is facilitated by large liquid proportions always exceeding 75% in the oxidized experiments, very fine-grained starting materials (<10 μm grain size) and run durations optimized for equilibration versus iron and hydrogen loss that varied from 3·5 h at the highest to 212 h (9 days) at lowest temperatures. Elevated H2O concentrations further enhanced diffusional equilibration. Overheating of the starting materials to generate homogeneous liquids was avoided to circumvent the formation of metastable phases that are reluctant to dissolve and/or re-equilibrate at lower temperatures. Runs were first pressurized to 0·4 GPa and then heated at a rate of 30–50°C min–1 with concomitant increase of pressure and directly brought to run temperature. Attainment of equilibrium can further be evaluated through the following considerations. Solid and liquid phases show homogeneity. For the liquid (glass) phase 5–40 individual analyses were obtained from all experiments. Liquids were analyzed throughout the entire capsule and we did not find any statistically significant variation in any of the runs. Solid phases are generally homogeneous with the exception of some sector zoned high-Ca clinopyroxene (cpx) and low-Ca orthopyroxene (opx) with low-Al and high-Al sectors, a feature that is often observed both in experimental studies (e.g. Sisson & Grove, 1993; Nandedkar et al., 2014) and in field occurrences (e.g. Dessimoz et al., 2012; Hürlimann et al., 2016) in basaltic to andesitic systems and is not necessarily an expression of disequilibrium but might be related to relatively rapid growth of pyroxenes (e.g. Schwandt & McKay, 2006) in a specific temperature window, in the present case between 1150 and 1180°C for cpx and 1090 and 1120°C for opx. Textures in experimental charges indicate well-crystallized, unzoned, idiomorphic solid phases in most but the lowest temperature experiments where amphibole occasionally shows some limited core–rim zoning (Figs 1 and 2). Major element partitioning, in particular Fe–Mg (Table 2), is very consistent, with values of 0·30 ± 0·02 to 0·35 ± 0·03 for olivine in graphite–Pt experiments and 0·23–0·26 in fractional experiments in AuPd capsules; likewise, Fe–Mg Kd values vary very little for cpx and opx within a given series, always with the exception of the equilibrium crystallization experiments in AuPd capsules. Quench crystallization is limited in the fractional crystallization experiments, with rare feathery overgrowth of olivine, cpx and amphibole on stable crystals that are easily identified in BSE images, but becomes much more abundant in the equilibrium crystallization experiments, in particular at higher temperatures (see Fig. 1a and b), to the extent that only composition can clearly distinguish between stable and quench cpx (very high Al, and variable Ca and Fe/Mg contents). The attainment of near constant fO2 in the fractional and the equilibrium crystallization experiments in graphite–Pt capsules (which are truly buffered) at the intended values corresponding to NNO and close to the H2O maximum in the C–COH system cannot be strictly validated. However, in all experiments that contained olivine the apparent Fe2+–Mg partitioning assuming that olivine accommodates only ferrous iron provides a constraint on the relative fO2 (Fig. 4). The equilibrium and fractional crystallization experiments in Pt–C capsules result in Kd(ol–liq) values that are independent of temperature and vary within error around the preferred value of 0·324 at 1·0 GPa (Ulmer, 1989). The more oxidized experiments in AuPd alloys show significantly lower apparent Kd; the fractional ones where only a few experiments contained olivine scatter around 0·24, whereas the equilibrium crystallization experiments show a systematic decrease of the Kd with decreasing temperature. Fig. 4. Open in new tabDownload slide Fe–Mg partitioning between olivine and coexisting liquid. The distribution coefficient [Kd = (FeO/MgO)liquid/(FeO/MgO)olivine] was calculated for FeO = Fetotal. The value of 0·324 corresponds to the value of Ulmer (1989) for 1 GPa assuming all Fe = Fe2+. Experiments labelled BN were conducted with BN spacers surrounding the AuPd–Pt capsule. Error bars are propagated from averages of individual olivine and liquid analyses (Tables 3 and 4). Fig. 4. Open in new tabDownload slide Fe–Mg partitioning between olivine and coexisting liquid. The distribution coefficient [Kd = (FeO/MgO)liquid/(FeO/MgO)olivine] was calculated for FeO = Fetotal. The value of 0·324 corresponds to the value of Ulmer (1989) for 1 GPa assuming all Fe = Fe2+. Experiments labelled BN were conducted with BN spacers surrounding the AuPd–Pt capsule. Error bars are propagated from averages of individual olivine and liquid analyses (Tables 3 and 4). Phase relations and phase proportions The discussion of the phase relations and compositions is strictly limited to the results from the inner capsules in the case of a noble metal double-capsule setup, as the outer capsule was generally made of AuPd alloys with higher proportions of Pd or even of Pt and not Fe-pre-saturated, resulting in much increased Fe loss exceeding 50% in some high-temperature runs. In addition, the outer capsule was much larger and, thus, was subjected to a larger thermal gradient that in some cases resulted in the crystallization of additional (lower temperature) phases at the extremities of the capsule. Nevertheless, the phase equilibria observed in the outer capsules were consistent with and in most cases were identical to the data obtained from the inner capsules and served as control, in particular to detect potential volatile loss from the inner capsules. We first present the results of the oxidized (AuPd) experiments followed by the more reduced experiments that have less significance for the differentiation of arc-type primary magmas. The reduced fractional crystallization experiments contained low H2O and represent a case of ‘damp’ (about 2–3 wt % H2O) differentiation at fO2 conditions more probably representing arc-tholeiitic differentiation. However, near anhydrous tholeiitic differentiation trends related to H2O-poor melts generated by decompression in the sub-arc mantle have been described from several subduction-related systems (e.g. Sisson & Bronto, 1998; Grove et al., 2002, 2005; Le Voyer et al., 2010) and therefore the ‘damp’ experiments provide crucial information on the lower crustal differentiation of such magmas in subduction-zone settings. Summaries of the instantaneous solid compositions (ISC) of the fractional crystallization experiments depicting the mineral phases and their modal proportions crystallizing in every fractionation step as well as the total solid composition (TSC) are given in Figs 5–7. Fig. 5. Open in new tabDownload slide Modal proportions of solid phases in experiments on the high-Mg, basaltic andesite (85-44) initial composition. (a) ISC (instant solid composition) of the single temperature steps of the fractionation experiments (FC ba AuPd) at 1 GPa under oxidizing conditions. (b) TSC (total solid composition; summarized modal fractions of all previous steps normalized to mass of initial basaltic andesite) TSCnm = TSCn−1m + ISCnm × AMFn−1 with AMFn = IMFn × AMFn−1, where TSCnm is total of mineral m for experiment n, ISCnm is instantaneous fraction of mineral m for experiment n, IMFn is instantaneous melt fraction of experiment n, and AMF is accumulated melt fraction of experiment n. The difference from 100 corresponds to amount of liquid left relative to start. Numbers on bars indicate H2O content of liquids determined by Raman spectroscopy (bold) or nominal H2O contents calculated from starting materials (italic) (see Table 2 for details). (c) TSC of equilibrium crystallization experiments at 1·2 GPa and 5 wt % H2O taken from Müntener et al. (2001). Fig. 5. Open in new tabDownload slide Modal proportions of solid phases in experiments on the high-Mg, basaltic andesite (85-44) initial composition. (a) ISC (instant solid composition) of the single temperature steps of the fractionation experiments (FC ba AuPd) at 1 GPa under oxidizing conditions. (b) TSC (total solid composition; summarized modal fractions of all previous steps normalized to mass of initial basaltic andesite) TSCnm = TSCn−1m + ISCnm × AMFn−1 with AMFn = IMFn × AMFn−1, where TSCnm is total of mineral m for experiment n, ISCnm is instantaneous fraction of mineral m for experiment n, IMFn is instantaneous melt fraction of experiment n, and AMF is accumulated melt fraction of experiment n. The difference from 100 corresponds to amount of liquid left relative to start. Numbers on bars indicate H2O content of liquids determined by Raman spectroscopy (bold) or nominal H2O contents calculated from starting materials (italic) (see Table 2 for details). (c) TSC of equilibrium crystallization experiments at 1·2 GPa and 5 wt % H2O taken from Müntener et al. (2001). Fig. 6. Open in new tabDownload slide Modal proportions of solid phases in experiments on the high-Mg basalt (RC158c) initial composition under oxidizing conditions. (a) ISC of the single temperature steps of the fractionation experiments (FC Mb AuPd) at 1 GPa; (b) corresponding TSC (for explanation see Fig. 5); (c) TSC of equilibrium crystallization experiments at 1·0 GPa under oxidizing conditions (EQ Mb AuPd). Fig. 6. Open in new tabDownload slide Modal proportions of solid phases in experiments on the high-Mg basalt (RC158c) initial composition under oxidizing conditions. (a) ISC of the single temperature steps of the fractionation experiments (FC Mb AuPd) at 1 GPa; (b) corresponding TSC (for explanation see Fig. 5); (c) TSC of equilibrium crystallization experiments at 1·0 GPa under oxidizing conditions (EQ Mb AuPd). Fig. 7. Open in new tabDownload slide Modal proportions of solid phases in experiments on the high-Mg basalt (RC158c) initial composition under reducing conditions. (a) ISC of the single temperature steps of the fractionation experiments (FC Mb Pt–C) at 1 GPa; (b) corresponding TSC; numbers indicate H2O contents of liquids determined by Raman spectroscopy, except for the run at 1140°C, where the value was estimated by the difference method; (c) TSC of equilibrium crystallization experiments at 1·0 GPa under reducing conditions (EQ Mb Pt–C). Fig. 7. Open in new tabDownload slide Modal proportions of solid phases in experiments on the high-Mg basalt (RC158c) initial composition under reducing conditions. (a) ISC of the single temperature steps of the fractionation experiments (FC Mb Pt–C) at 1 GPa; (b) corresponding TSC; numbers indicate H2O contents of liquids determined by Raman spectroscopy, except for the run at 1140°C, where the value was estimated by the difference method; (c) TSC of equilibrium crystallization experiments at 1·0 GPa under reducing conditions (EQ Mb Pt–C). Basaltic andesite (85-44, ba) experiments The liquidus temperature of the initial starting composition (85-44, 5 wt % H2O) was bracketed at 1215 ± 15°C with orthopyroxene (opx) as the liquidus phase. This is consistent with the data of Baker et al. (1994) and Müntener et al. (2001), who conducted equilibrium crystallization experiments at 1 and 1·2 GPa under more reducing (closer to C–COH) conditions and located the liquidus at 1·0 GPa between 1260°C for 3 wt % H2O and opx as the liquidus phase, 1190°C for 6 wt % H2O and olivine (ol) as the liquidus phase, and above 1230°C at 1·2 GPa with opx and cpx. More reducing conditions increase the liquidus temperature for constant H2O contents (Kägi et al., 2005). The second phase to crystallize together with opx is high-Ca clinopyroxene (cpx) at 1170°C; this assemblage dominated by cpx with minor opx persists until 1050°C, where opx disappears and amphibole begins to crystallize; both phases form the liquidus assemblage down to 990°C, where it coexists with an andesitic liquid. Neither Fe–Ti-oxides nor plagioclase or garnet were observed in the crystallization interval from high-Mg basaltic andesite to andesite or dacite. The disappearance of opx at the onset of amphibole crystallization indicates an amphibole reaction boundary opx ( ± cpx) + liq = amph. Between 980 and 1000°C amphibole and minor Ti-bearing magnetite are the liquidus phases. At 950°C, garnet joins the assemblage, forming euhedral overgrowth rims on the almandine seeds, whereas plagioclase seeds nearly completely dissolved, implying that this composition (990 ba) is very close to plag saturation and saturated in garnet. At 900°C, amph–garnet–plag and ilmenite form the stable assemblage coexisting with an andesite or dacite liquid. This assemblage persists to 750°C, with apatite joining at 850°C. At 850°C no amphibole was observed, but at 800°C amphibole was stable again. This situation is not uncommon in stepwise fractional crystallization experiments where the (re-mixed) derivative composition for the next, lower temperature run can slightly deviate from the multiply saturated LLD, resulting in appearance and disappearance of phases. This is graphically illustrated in the discussion, where pseudoternary phase diagrams are presented. Finally, at 720°C corresponding to the outer capsule of run PU1072 in the cold end of the large capsule, quartz and ulvöspinel join the assemblage (the latter at the expense of ilmenite). At the lowest temperatures, 720–750°C, garnet seeds reveal very subtle overgrowth rims that were difficult to impossible to analyze. Amphibole was found only at 720°C but not at 750°C. The melt fractions (Fig. 8a) decrease near linearly from the liquidus to 720°C despite the fact that the crystallizing assemblages change significantly between 1080 and 1050°C where amphibole (amph) replaces opx and cpx (Fig. 5a) and below 950°C where garnet and plagioclase join the fractionating assemblage. Fractional crystallization produces a variety of different ultramafic cumulates, orthopyroxenites, followed by websterites and opx-bearing clinopyroxenites, and (cpx-bearing) hornblendites that are replaced by garnet–hornblende-gabbros below 950°C. A total of 45 wt % ultramafic cumulates (Fig. 5b) need to be extracted to produce an andesitic liquid from a primary high-Mg basaltic andesite. An additional 20% of garnet–hornblende-gabbro generates a rhyodacitic liquid at 850–800°C, and finally at 720–750°C and 80% fractionation a high-silica rhyolite is produced. For comparison, modal proportions and melt fractions of the equilibrium crystallization experiments by Müntener et al. (2001) at 1·2 GPa under more reducing conditions but identical H2O content are provided (Fig. 5c). The inferred liquidus temperature is considerably higher (estimated 70°C), and both opx and cpx form the near-liquidus assemblage, which persists to 1110°C, where it is joined by hercynitic spinel; at 1070°C amphibole (accompanied by a minute amount of garnet) crystallizes at the expense of spinel, cpx and opx, which all decrease in modal amounts implying a peritectic reaction that forms amphibole. The melt fraction versus temperature relationship displays a strong decrease below 1150°C and, with the onset of amphibole crystallization, less than 50% liquid is left (Fig. 8a). This is ∼20% less than in the fractional case at the same temperature. Fig. 8. Open in new tabDownload slide Liquid fraction as a function of run temperature. (a) High-Mg basaltic andesite (ba) series; fractional crystallization experiments show near-linear variation of melt fraction with temperature. For comparison the equilibrium crystallization experiments at 1·2 GPa for the same initial composition (85-44) with 5 wt % H2O are taken from Müntener et al. (2001). (b) High-Mg basalt (Mb) series depicting distinctly non-linear trends with characteristic shifts linked to fO2 conditions and H2O content. (For explanation see text.) EQ and FC denote equilibrium and fractional crystallization experiments. Mb and ba represent (initial) starting compositions (Mb is high-Mg basalt, RC158c, 3 ± 0·3 wt % H2O; ba is high-Mg basaltic andesite, 85-44, 5 wt % H2O; Table 1). Pt–C and AuPd denote capsule material (and fO2 conditions): Pt–C, graphite capsules in outer Pt capsules (reducing); AuPd, Au–Pd single- or double-capsule arrangement (oxidizing). Error bars correspond to standard deviations from mass-balance calculations of modal proportions (Table 2). Fig. 8. Open in new tabDownload slide Liquid fraction as a function of run temperature. (a) High-Mg basaltic andesite (ba) series; fractional crystallization experiments show near-linear variation of melt fraction with temperature. For comparison the equilibrium crystallization experiments at 1·2 GPa for the same initial composition (85-44) with 5 wt % H2O are taken from Müntener et al. (2001). (b) High-Mg basalt (Mb) series depicting distinctly non-linear trends with characteristic shifts linked to fO2 conditions and H2O content. (For explanation see text.) EQ and FC denote equilibrium and fractional crystallization experiments. Mb and ba represent (initial) starting compositions (Mb is high-Mg basalt, RC158c, 3 ± 0·3 wt % H2O; ba is high-Mg basaltic andesite, 85-44, 5 wt % H2O; Table 1). Pt–C and AuPd denote capsule material (and fO2 conditions): Pt–C, graphite capsules in outer Pt capsules (reducing); AuPd, Au–Pd single- or double-capsule arrangement (oxidizing). Error bars correspond to standard deviations from mass-balance calculations of modal proportions (Table 2). High-Mg basalt (RC158c, Mb) experiments under oxidizing conditions The liquidus temperature of the fractional crystallization experiments on the high-Mg basalt initial starting composition is located around 1280–1300°C with olivine (Fo 91·1) and minor Cr-rich spinel as the liquidus phase. As Cr was effectively depleted in the coexisting liquid by 1% spinel (sp) crystallization, Cr2O3 was not added to subsequent fractionation steps. Cpx joins olivine at 1200°C and hercynitic spinel appears at 1170°C. Olivine disappears at 1140°C whereas opx saturates at 1110°C. The cubic oxide phase changes from abundant hercynitic spinel between 1170 and 1110°C to Ti-bearing, aluminous magnetite (mag) at 1080°C. At 1050°C opx disappears and amphibole starts crystallizing; at 1020 and 990°C amphibole–cpx–Ti-magnetite form the solid assemblage coexisting with a dacitic liquid representing 30–35% of the initial high-Mg basalt (Fig. 8b). The disappearance of opx at the onset of amphibole crystallization points towards a similar amphibole reaction boundary to that in the case of the basaltic andesite series. Several experiments have been repeated in the range 980–1000°C (rk67, PU1048, PU1069) as some inconsistencies in the amount of liquid present were observed. Unfortunately, run rk67 is no longer available, and thus the true H2O content of this experiment could not be verified, but we assume that this run contained less than the target value of 8·14 wt % in the starting material that should result in 9·3 wt % in the liquid (considering 85·4 wt % melt and 8·3 wt % amphibole with 2·1 wt % H2O). We therefore repeated the experiments with a newly prepared starting material (990 pb), which gave 8·6 and 8·2 wt % H2O determined by Raman spectroscopy. Therefore, these two runs were utilized to continue the series of fractional crystallization experiments that resulted in cpx only as liquidus phases. At 950°C amphibole is the dominant mineral, crystallizing together with small amounts of garnet that formed around the almandine seeds. At 900°C plagioclase and ilmenite join the crystallizing assemblage, which persists until 800°C with the exception, as for the basaltic andesite (ba) series, that no amphibole was found at 850°C. At 750°C plagioclase and Ti-magnetite are present with unreacted garnet seeds (no rim, no resorption discernible) coexisting with a rhyolitic liquid. The following ultramafic cumulates have formed along the LLD from the liquidus to 950°C (Fig. 6a): spinel-bearing dunites, wehrlites, (ol–spl–opx-bearing) clinopyroxenites and finally hornblende-pyroxenites and cpx-hornblendites. The difference in total amount of extracted ultramafic cumulates between the basaltic andesite (45%) and the high-Mg basalt (70%) largely reflects the amount of dunite (17%) extracted from the latter. Below 950°C ilmenite-bearing garnet–hornblende-gabbros form the main cumulates with coexisting liquids evolving from dacite to rhyolite. The evolution of the melt fraction as a function of temperature is no longer near linear as for the basaltic andesite case, but shows a pronounced curvilinear decrease from the liquidus to 1100°C, with a much flatter slope towards lower temperatures when amphibole and later plagioclase and garnet dominate the solid assemblage (Fig. 8b). Phase assemblages of equilibrium crystallization experiments under oxidizing conditions are displayed in Fig. 6c. These experiments are not directly comparable with the respective fractionation experiments because fO2 was unconstrained and allowed to vary in a closed system. As mentioned above and presented below, the fO2 started at a value of c. NNO + 2 given by the Fe2O3/FeO of the starting material (except for two runs with BN spacers that resulted in about 1 log unit lower fO2) but increased to NNO + 5·5 with decreasing temperature and melt fraction; that is, it is between +1 to 2 and +3·5 to 4·5 log units higher than in the fractional crystallization experiments. The liquidus is inferred at about 1280°C and the liquidus phases are olivine (Fo 92·0) and Cr-rich spinel; this assemblage persists to 1160°C and at 1120°C cpx joins the liquidus. At 1040°C both amphibole and opx appear and the amount of cpx and olivine sharply decreases. Olivine is lost from the assemblage at 980°C, the lowest temperature at which it coexists with an andesitic liquid at a melt fraction of 15%. Phase relations and modal proportions indicate a peritectic relationship involving both the well-known ol + liq = opx reaction as well as an amphibole-forming reaction (pyx +liq = amph). The evolution of the melt fraction as a function of temperature is different, with higher initial (higher temperature) melt fractions for equilibrium than for fractional crystallization. The relationship is non-linear and similar to equilibrium crystallization experiments on a near-primary basalt from St Vincent (Lesser Antilles) with 2·3 wt % H2O (Melekhova et al., 2013). The steepest decrease in melt fraction between 1080 and 1040°C is related to the intersection of the liquidus with the peritectic amphibole reaction boundary leaving only <23% liquid. This shift towards higher melt fractions at elevated temperatures compared with the fractional crystallization experiments is related to the effect of increased fO2 shifting the onset of cpx crystallization to just above 1120°C compared with 1200°C in the fractional case at NNO. The combination of high H2O contents and high fO2 effectively depresses the cpx stability field relative to olivine, keeping the system in the olivine-only field where crystallization is not efficient. With the onset of cpx the melt fraction decreases, additionally enhanced by peritectic reactions that are suppressed by fractional crystallization, resulting in a very different T–f relationship. Cumulates produced by equilibrium crystallization under oxidizing conditions are comparable with the fractional case down to 980°C, but more abundant opx formed as a result of peritectic formation from olivine that has been extracted in the fractional case. High-Mg basalt (RC158c) experiments under reducing conditions The fractional crystallization experiments conducted in graphite–Pt capsules (FC Mb Pt–C) are distinct as they lost most of their H2O in all but the first experiment conducted with the natural high-Mg basalt RC158c. This was already suspected based on the phase equilibria and total solid compositions and the rather high totals of the quenched glass analyses (>95 wt %). The first experiment, rk3, contained the nominal value of 3·1 wt % (Table 2) whereas all other experiments were conducted with derivative starting materials (Table 1) containing H2O in the form of Al- and Mg-hydroxides. This systematically resulted in glass H2O contents of 2·0–2·6 wt % (Raman) and 1·3–2·4 wt % (plag-melt, Fig. 3b), despite the fact that with decreasing temperature the amount of H2O added in the charge increased from 3·2 to 10·7 wt %. Considering the amount of crystallized solids, the liquid should contain H2O contents ranging from 4·1 to 17·6 wt % (Fig. 3b). Numerous experiments run at identical conditions with the same starting material are very consistent regarding melt composition, melt fractions, modal mineralogy and mineral compositions. A similar behavior was observed in the equilibrium and fractional crystallization experiments of Caricchi et al. (2006) employing exactly the same technique, but the problem remained unresolved. The fact that only experiments conducted with starting materials containing H2O either in the form of Mg(OH)2 (brucite) and Al(OH)3 (gibbsite) or as free H2O added by micro-syringe showed this behavior, but none of the experiments conducted with the natural high-Mg basalt RC158c that contains H2O nearly exclusively in amphibole with some minor contribution from chlorite, points towards the availability of free H2O at low temperature when no melt phase is present. Amphibole most probably decomposes only at the onset of melting in the basaltic starting material and the water is directly incorporated into the melt phase. In the case of Al-hydroxide and possibly also brucite, they decompose partly or completely at less than 400–500°C. The most likely explanation for this unusual behavior is that free H2O at 400–500°C in equilibrium with graphite and the reduced starting material containing all Fe as ferrous iron in fayalite resulted in the production of hydrogen and methane with near instantaneous hydrogen diffusion out of the capsule. Why a near constant amount of H2O of 2·0–2·6 wt % was present in the liquid phase in these experiments cannot be resolved, but provides the base to utilize these experiments as a series of ‘damp’ fractionation experiments at low H2O content and relatively reducing conditions. These experiments might be representative for the differentiation of primary, H2O-poor mantle melts that occur in suprasubduction settings (e.g. Mt Shasta; Grove et al., 2005), but are much more widespread in mature extensional settings. The liquidus for the hydrous (3·0 ± 0·3 wt %) high-Mg basalt at 1·0 GPa is around 1330°C, the anhydrous liquidus is located about 100°C higher (at 0·1 MPa and 3·0 GPa; Ulmer, 1988). The liquidus phase at 1230°C (i.e. at least 100°C below the inferred liquidus) is olivine (Fo87) plus minor Cr-rich spinel; at 1200 and 1170°C cpx is the dominant phase, and olivine disappears at 1170°C. At 1140°C, both plagioclase (An91) and hercynitic spinel appear. At 1110°C, highly variable plag (An93 core–An58 rim and small crystals dispersed in the melt) coexists with cpx, opx and spinel, and finally at 1080°C, the lowest temperature run conducted in this series, cpx–plag–ilmenite (ilm)–Fe-rich garnet (gar) form the stable assemblage with only 14% of basaltic trachy-andesite liquid left relative to the initial high-Mg basalt. These phase relations are distinct from all other experiments by early saturation of plagioclase that has not been encountered in any other series at 1·0 GPa at temperatures in excess of 950°C, and by garnet saturation at 1080°C as opposed to the oxidized experiments with high H2O contents where both phases start crystallizing only at or below 950°C from dacitic liquids. Run rk55 at 1110°C shows clear signs of progressive H2O loss, in particular the strongly zoned plagioclase with high-An cores and low-An rims and abundant low-An microlites most probably demonstrating continuous H2O loss from the experiment at an early stage. The plag–melt hygrometer results in 2·3 and 1·6 wt % H2O for high- and low-An plagioclase, indicating that most H2O was actually lost before plagioclase even nucleated. The melt fraction versus temperature relations are consistent with low H2O contents (Fig. 8b), illustrated by a shift of the curve to higher temperatures at a given melt fraction and a high df/dT resembling the relationships of fractional crystallization of anhydrous MOR-tholeiite at 1·0 GPa (Villiger et al., 2004). The equilibrium crystallization experiments, on the other hand, were completely successful in keeping the system closed for H2O (see Fig. 3) owing to the fact that natural starting material with H2O is bound exclusively in silicates that decompose only at high temperatures under elevated pressures when a melt phase is already forming. The liquidus (Fig. 7c) is located at 1340 ± 10°C with ol (Fo 90·6) as the liquidus phase, and Cr-rich spinel saturates at 1210°C, followed by cpx at 1180°C and opx at 1120°C. Saturation in opx and decrease in olivine confirm the peritectic relationship of olivine–opx. Aluminous (hercynitic) spinel is present at 1090°C and below. At 1060°C amphibole starts crystallizing and at 1000°C olivine disappears and the modal amount of cpx decreases relative to amph + opx, implying a reaction of the type ol + cpx + liq = opx + amph. The succession of cumulates is comparable with the higher fO2 equilibrium crystallization experiments except for higher amounts of opx originating from the peritectic reaction(s). The T–f relationship (Fig. 8b) shows a similar behavior to the oxidized equilibrium crystallization experiments but shifted to higher temperatures. The crystallization rate (df/dT) is low during olivine-only crystallization but increases when cpx and opx join the liquidus. The highest crystallization rates are related to amphibole saturation through the inferred peritectic reaction, leaving 16 wt % silica-undersaturated liquid at 1000°C. Melt compositions All liquid compositions (Tables 3 and 4) were normalized to 100 wt % anhydrous and are presented as a function of SiO2 (Fig. 9) and temperature (Figs 10 and 11). Experiments under oxidized conditions evolve from basalt to basaltic andesite and andesite [total alkalis–silica (TAS) diagram, Fig. 9a]; the fractional crystallization series (FC Mb AuPd and FC ba AuPd) evolve further to dacite and finally rhyolite, reaching silica contents of nearly 75 wt % at 720–750°C. For comparison, the results of the equilibrium crystallization experiments at 1·2 GPa for 3·8 and 5·0 wt % added H2O are provided, which extend from the primary basaltic andesite to rather alkali-rich andesites (Müntener et al., 2001). The equilibrium crystallization experiments on the high-Mg basalt under oxidizing conditions (EQ Mb AuPd) show an evolution towards low total alkali contents related to (1) high melt fractions at rather high silica contents, (2) large amounts of amphibole below 1040°C and (3) analytical difficulties in obtaining correct alkali contents of the small melt pools at lowest temperatures (highest silica contents). The fractional crystallization experiments of the basaltic andesite display lower total alkali contents for a given SiO2 than the equilibrium crystallization experiments. Fig. 9. Open in new tabDownload slide Liquid compositions plotted in (a) the TAS (total alkali vs silica diagram (Le Maitre et al., 1989) and (b) the FeO/MgO vs SiO2 (wt %) diagram with the tholeiitic and calc-alkaline fields according to (Miyashiro, 1974). Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 9. Open in new tabDownload slide Liquid compositions plotted in (a) the TAS (total alkali vs silica diagram (Le Maitre et al., 1989) and (b) the FeO/MgO vs SiO2 (wt %) diagram with the tholeiitic and calc-alkaline fields according to (Miyashiro, 1974). Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 10. Open in new tabDownload slide Oxide variations of experimental liquids (glasses) as a function of temperature (°C) for the high-Mg basaltic andesite fractional series (85-44) under oxidizing conditions: (a) SiO2; (b) Al2O3; (c) TiO2; (d) FeOtot; (e) MgO; (f) CaO; (g) Na2O; (h) K2O; all in wt % normalized to 100 wt % anhydrous. Compositions of equilibrium crystallization experiments at 1·2 GPa for 3·8 and 5·0 wt % H2O [EQ ba 3.8 and EQ ba 5.0) are given for comparison and are taken from Müntener et al. (2001)]. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 10. Open in new tabDownload slide Oxide variations of experimental liquids (glasses) as a function of temperature (°C) for the high-Mg basaltic andesite fractional series (85-44) under oxidizing conditions: (a) SiO2; (b) Al2O3; (c) TiO2; (d) FeOtot; (e) MgO; (f) CaO; (g) Na2O; (h) K2O; all in wt % normalized to 100 wt % anhydrous. Compositions of equilibrium crystallization experiments at 1·2 GPa for 3·8 and 5·0 wt % H2O [EQ ba 3.8 and EQ ba 5.0) are given for comparison and are taken from Müntener et al. (2001)]. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 11. Open in new tabDownload slide Oxide variations of experimental liquids (glasses) as a function of temperature (°C) for the high-Mg basalt (RC158c) experiments: (a) SiO2; (b) Al2O3; (c) TiO2; (d) FeOtot; (e) MgO; (f) CaO; (g) Na2O; (h) K2O; all in wt % normalized to 100 wt % anhydrous. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 11. Open in new tabDownload slide Oxide variations of experimental liquids (glasses) as a function of temperature (°C) for the high-Mg basalt (RC158c) experiments: (a) SiO2; (b) Al2O3; (c) TiO2; (d) FeOtot; (e) MgO; (f) CaO; (g) Na2O; (h) K2O; all in wt % normalized to 100 wt % anhydrous. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Experiments under reducing conditions (Pt–C) reveal a different evolution: both fractional and equilibrium crystallization evolve towards high total alkali contents with restricted silica enrichment leading to trachybasalts and basaltic trachyandesite. This trend is controlled by abundant opx crystallization in the equilibrium case and the low H2O content of the fractional crystallization experiments that saturate in plagioclase and opx at 1110°C, driving the liquid to lower SiO2 at increasing alkali contents. All but the reduced, low-H2O fractional crystallization experiments (FC Mb Pt–C) straddle the tholeiitic–calc-alkaline dividing line up to 65 wt % SiO2, a trend that is rather typical for arc-related hydrous magmas. The lowest temperature fractional crystallization runs (900–720°C) show an evolution first towards higher FeO/MgO ratios, driving them deeply into the tholeiite field down to 850°C, followed by an abrupt inversion to lower ratios back into the calc-alkaline field that is mainly caused by Fe-rich garnet and ilmenite and, in the last step, by Ti-magnetite fractionation. A systematic decrease of the FeO/MgO ratio with increasing SiO2 as well as with increasing fO2 was observed at 0·4–0·7 GPa (Pichavant et al., 2002; Sisson et al., 2005) whereas a similar trend with a cross-over from tholeiitic to calc-alkaline around 55 wt % SiO2 at 0·7 GPa and experiments that cluster around the dividing line at 0·9 GPa at 62 wt % SiO2 was shown by Blatter et al. (2013). Liquid evolution of basaltic andesite (85-44) experiments The silica content steadily increases with decreasing temperature for fractional crystallization experiments, becoming steeper below 1000°C (Fig. 10a) where low-silica amphibole and garnet plus relatively anorthitic plagioclase replace opx and cpx at about 56·5 wt % SiO2. Between 950 and 720°C silica continuously increases to 75 wt %, covering the entire range from andesite through dacite and rhyodacite to rhyolite. The two equilibrium crystallization series at 1·2 GPa (Müntener et al., 2001) evolve in an identical fashion until about 1100°C but then silica strongly increases owing to earlier (1070°C) amphibole saturation and considerably lower amount of liquid left at this stage (Fig. 8a, 48 versus 70 % liquid). Al2O3 increases with decreasing temperature with some evident shifts towards higher alumina contents compared with the fractional crystallization experiments (Fig. 10b). The fractional crystallization experiments reveal a sharp bend to lower alumina contents between 900 and 950°C clearly related to plagioclase and garnet saturation effectively depleting the derivative liquids in alumina. TiO2 initially increases as it behaves incompatibly with decreasing temperature until amphibole saturates (EQ and FC) between 1070 and 1050°C (Fig. 10c). At lower temperatures TiO2 decreases in the equilibrium crystallization experiments (extensive amphibole crystallization owing to peritectic reaction) but remains constant in the FC series until 950°C followed by a sharp decrease owing to the combined effects of amphibole and ilmenite fractionation. Total FeO (Fig. 10d) for the equilibrium crystallization runs is initially nearly constant and strongly decreases below 1150°C where either hercynitic spinel (up to 5 wt % H2O) or Fe-rich garnet (3·8 wt %) effectively deplete FeO. In the fractional crystallization experiments, FeO decreases at the beginning owing to the dominance of Fe-rich opx (1200–1110°C), followed by a plateau (1110–1020°C) where cpx is the dominant phase, and then by a continuous decrease to 720°C controlled by amphibole, Fe-rich garnet and ilmenite/Ti-magnetite. MgO shows a continuous decrease with decreasing temperature that is more pronounced for the equilibrium crystallization experiments (Fig. 10e); the final rhyolitic melts for the FC experiments are nearly devoid of MgO (<0·5 wt %). A similar evolution is observed for CaO, which decreases for all but the first fractionation step (only opx on the liquidus) owing to persistent cpx and later amphibole and plagioclase crystallization (Fig. 10f). It is interesting to note that the andesite compositions obtained by fractional crystallization around 1000°C are CaO-rich (8 wt %) at low MgO (2·3–3·5 wt %) and moderate Mg# (0·42–0·51; Table 4). The evolution of Na2O is characterized by continuously increasing contents for the equilibrium crystallization runs and a ‘hump’ in the evolution of the fractional crystallization experiment at the onset of amphibole crystallization (Fig. 10g); a slowing down of Na2O enrichment is expected at that point (amphibole is pargasitic), but the decrease is most probably an experimental artifact related to Na2O loss during analyses and incomplete compensation in the next starting material. In any case, the Na2O enrichment is less pronounced than for the equilibrium runs. The final two experiments of the FC series show a slight decrease of Na2O from about 4·2 to 3·8 wt % related to the formation of albite-rich plagioclase. K2O behaves incompatibly over the entire course of differentiation (Fig. 10h); however, in amphibole-dominated experiments the increase is moderate owing to accommodation of some K2O in the amphibole structure. Unlike Na2O, K2O also increases in the last fractionation steps owing to the absence of any K2O-bearing phase such as biotite or K-feldspar, and reaches 2·22 wt % K2O at 720°C starting from an initial K2O content of 0·40 wt % (85-44; Table 1), which corresponds to a near-perfect incompatible behavior at a melt fraction just below 0·2 relative to the initial basaltic-andesite. Liquid evolution of high-Mg basalt (RC158c) experiments The SiO2 content as a function of temperature for this series reveals two contrasting trends (Fig. 11a). The SiO2 content for the oxidized experiments and to a lesser extent the fractional crystallization experiments under reducing conditions (FC Mb Pt–C) increases with decreasing temperature, although only after a prolonged plateau at around 48–50 wt % extending from the liquidus (>1300°C) down to 1120°C. At this point SiO2 increases and the liquids differentiate towards andesite, dacites, rhyodacites and finally rhyolites for the fractional crystallization series under oxidized conditions (FC Mb AuPd). The equilibrium crystallization series under oxidized conditions produces andesite with 58·5 wt % SiO2 at 980°C whereas the experiments under reduced conditions, both FC and EQ, generate liquids that are trachyandesitic with 54 wt % SiO2 and trachybasaltic with 50 wt % SiO2, respectively. The differentiation to alkaline liquids is related to both their rather silica-rich fractionating solid assemblages, dominated by opx + cpx in the EQ experiments and opx–cpx–plag in the FC experiments, and very low melt fractions at the lowest temperatures responsible for the high alkali contents of these liquids. The strong and steady increase of SiO2 in the oxidized FC experiments is related to the onset of spinel crystallization and reinforced by amphibole at and below 1050°C. Below 950°C a silica-poor solid assemblage composed of amph–plag–gar–ilm is responsible for the continuous and rapid increase of SiO2. The EQ experiments under oxidized conditions show a progressive silica enrichment over the entire crystallization interval investigated owing to the expanded olivine-only stability field followed by cpx + spinel (Cr–Al–Fe3+-spinels with decreasing Cr contents with decreasing temperature). Al2O3 concentrations increase with decreasing temperature over the entire range for all the EQ experiments. Both EQ series (reduced and oxidized) reach elevated Al2O3 of nearly 23 wt % in andesitic liquids. The reduced FC experiments show a sharp bend in the Al2O3 evolution at 1140°C (19 wt %) where plagioclase plus hercynitic spinel saturate; the final trachyandesite that also experienced garnet extraction has 17 wt % Al2O3. The oxidized fractionation experiments show a plateau at the onset of garnet + plag crystallization (950–1000°C) followed by continuous decrease to lower temperatures with final rhyolites containing ∼16 wt % Al2O3. The TiO2 contents initially increase with decreasing temperature for all series, and this increase slows down and/or inverts upon amphibole saturation between 1060 and 1040°C (Fig. 11c). The fractional crystallization experiment under reducing conditions shows a tholeiitic trend with increasing TiO2 (low melt fraction under H2O-poor conditions), except for the last run (1070°C) where ilmenite saturated. The oxidized fractionation experiments reveal strongly decreasing TiO2 concentrations below 950°C, where Ti-bearing amphibole ilmenite and, in the last step, Ti-magnetite effectively deplete TiO2 in the residual liquid. The evolutionary trend for FeOtot is initially nearly flat or even slightly increasing followed by a continuous decrease for all but the reduced fractionation experiments (Fig. 11d). The latter reveal a tholeiitic trend with strongly increasing FeOtot until ilmenite and Fe-rich garnet saturate at 1080°C. Decreasing FeOtot for all other series to lower temperatures is related to the extensive crystallization of increasingly more fayalitic olivine, hercynitic to magnetitic spinel, and, for the oxidized FC series, ilmenite and Fe-rich garnet. MgO exhibits continuous decrease with decreasing temperature that is more accentuated for the fractional crystallization experiments under reducing conditions (Fig. 11e). The melts from the oxidized equilibrium crystallization experiments have initially higher MgO contents owing to their low crystallinity, but extensive olivine (and cpx) crystallization below 1120°C drives them to the same values as the other experiments. CaO shows initial increase related to olivine (plus small amounts of Cr-rich spinel) crystallization as the liquidus phase, followed by a variable decrease for the different series (Fig. 11f): the oxidized fractional crystallization experiments show a steady decrease from 1200°C down to 750°C; the oxidized equilibrium crystallization experiments show a limited CaO decrease owing to the dominance of opx and amphibole at lower temperature resulting in considerably higher CaO contents in the coexisting solids at a given SiO2 than in the fractional case. The reduced equilibrium experiments show a decrease similar to the oxidized equilibrium experiments, but that trend inverts in the last step, where CaO increased from about 8 to 10 wt % related to the peritectic reaction that consumes (high-Ca) cpx and generates amph + opx (see Fig. 7c). The resulting liquid is rich in CaO, but of trachybasaltic composition at rather low Mg# (0·52). The reduced fractionation experiments display steeply decreasing CaO contents with decreasing temperature owing to extensive cpx and plag fractionation. The evolution of Na2O and K2O largely mimics the remaining melt fraction as they principally behave incompatibly (Fig. 11g and h), exceptions being the decrease of both Na and K in the last crystallization interval for the reduced equilibrium experiments where abundant amphibole accommodates alkalis, and likewise the behavior of both Na and K for the oxidized equilibrium crystallization experiments that reveal decreasing Na2O and basically constant K2O. At higher temperatures this reflects high melt fractions (>63% at 1080°C) but at the lowest melt fraction this is related to abundant crystallization of amphibole. As discussed below, melts are depleted in Na and also the mineral phases (amph and cpx) are unusually low in Na2O, possibly indicating that abundant amphibole (>37%) in the last three experiments (980–1040°C) accommodates more alkalis than the coexisting liquids in this high-Mg basaltic system. Mass-balance constraints on these three experiments result in a perfect match for Na and K, except for the lowest temperature experiment (PU906, 980°C) where c. 0·20 wt % Na2O is missing in the mass balance, pointing towards difficulties in measuring Na contents accurately in these highly hydrous (10 wt % H2O) melts. Mineral compositions Olivine Olivine exclusively occurred in the experiments with the high-Mg basalt starting material and the evolution of the forsterite content (fo) as a function of temperature is illustrated in Fig. 12. Olivine crystallized over nearly the entire temperature interval for the equilibrium crystallization experiments (1330–1060°C for reducing and 1200–1000°C for oxidizing conditions), but only formed a near-liquidus phase in the fractional crystallization experiments. Olivine compositions under oxidizing conditions are more forsteritic owing to the higher ferric iron content and, thus, relatively more magnesian character of the liquids (Fig. 4). The second consistent observation is that the fractional crystallization runs show a much steeper decrease of forsterite with temperature, favoring Mg extraction over Fe by olivine (e.g. Bowen & Schairer, 1935). The most magnesian olivine under oxidizing conditions is Fo 92·0, whereas under reducing conditions it is Fo 90·6, bracketing the compositions of the cores of the natural phenocrysts of the high-Mg basalt (Fo 91·1, Ulmer, 1986; Hürlimann et al., 2016) that is perfectly matched by the highest Fo content observed for the oxidized (NNO) fractional crystallization experiments (rk47, 1230°C, Fo 91·1). The Ca content of olivine is relatively low and varies between 0·21 and 0·34 wt % CaO with a clear tendency for higher values for the experiments conducted under more reducing conditions. Fig. 12. Open in new tabDownload slide Forsterite content (molar) of olivine from the high-Mg basalt (RC158c) experiments as a function of temperature (°C). Experiments labelled ‘BN’ were conducted with BN as spacers surrounding the AuPd–Pt capsules, resulting in slightly more reducing conditions. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 12. Open in new tabDownload slide Forsterite content (molar) of olivine from the high-Mg basalt (RC158c) experiments as a function of temperature (°C). Experiments labelled ‘BN’ were conducted with BN as spacers surrounding the AuPd–Pt capsules, resulting in slightly more reducing conditions. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). High-Ca clinopyroxene Clinopyroxene crystallized over a wider temperature range (1200–950°C) and displays large compositional variation, as illustrated in Fig. 13. Tetrahedrally coordinated Al (AlIV) varies negatively with Mg + Fe2+ (Fig. 13a) close to, but with a flatter slope compared to the slope resulting from an ideal Tschermak’s substitution [(Mg, Fe2+)Si ⇔ AlIVAlVI], with the exception of three cpx from the most differentiated reduced fractional crystallization experiments. The most likely reason for the deviation from ideal Tschermak’s substitution is the more pronounced Ca–Mg exchange (opx component) of these Fe-rich pyroxenes, which results in a horizontal line as depicted in the Ca–temperature diagram (Fig. 13b). Cpx shows restricted Ca variation as a function of temperature. There are notable exceptions such as the behavior of cpx from the basaltic andesite fractional experiments (FC ba AuPd), which show a constant increase of the Ca content. This is controlled by the following factors. (1) Cpx in the range 1170–1080°C is opx-buffered and its Mg# is nearly constant at 0·860 ± 0·005. Therefore, we attribute the increase in Ca content to the widening of the cpx–opx solvus with decreasing temperature at nearly constant Mg# (e.g. Lindsley, 1983). (2) Below 1080°C opx no longer coexists, and coexisting liquids of this series in the range 1080–900°C are rather Ca-rich. The equivalent experiments on the high-Mg basalt starting material (FC Mb AuPd) that are also opx-free below 1080°C do not show this increase of Ca in cpx but instead reveal a subtle decrease, most probably an expression of slightly higher silica activity in the liquid and considerably lower xMg favoring lower Ca contents at a given temperature. The reduced fractional crystallization experiments (FC Mb Pt–C) reveal strongly decreasing Ca with decreasing xMg and temperature consistent with coexisting opx except for the lowest temperature experiment. The strong decrease of Ca with decreasing xMg and temperature buffered by opx (solvus) is well known (Lindsley, 1983). Titanium in cpx is characterized by an initial increase followed by a decrease related to either ilmenite or amphibole saturation that depletes the TiO2 content in the coexisting liquid phase (Fig. 13c). In contrast, Ti in cpx continuously increases in the basaltic andesite down to the lowest temperatures. The Ti increase in cpx correlates with the highest Al2O3 and increasing Ca contents (Fig. 13b), both pointing towards decreased silica activity in hydrous, calc-alkaline magmas as already pointed out by Müntener et al. (2001). Sodium shows increasing concentrations with decreasing temperature, except for the equilibrium crystallization series on the high-Mg basalt starting material under oxidized conditions, which reveals continuous decrease (EQ Mb AuPd) over the entire temperature range (Fig. 13d). This is consistent with the low Na contents of these liquids (Fig. 11g). Both the FC and EQ experiments on the high-Mg basalt show decreasing Na contents in cpx with decreasing temperature when coexisting with amph. In contrast, the FC series on the basaltic andesite exhibits a continuous increase of Na with decreasing temperature, again pointing to lower silica activity and incorporation of Ti as NaTiAlIV component. Fig. 13. Open in new tabDownload slide Cpx compositional variations expressed as cations per formula units (p.f.u.) based on four cations and 12 positive charges. (a) Tetrahedrally coordinated Al (AlIV) as a function of the sum of Mg + Fe2+; labelled lines indicate Tschermak’s and Ti-Tschermak’s substitution respectively; (b) Ca as a function of temperature (°C); (c) Ti as a function of temperature; (d) Na as a function of temperature. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 13. Open in new tabDownload slide Cpx compositional variations expressed as cations per formula units (p.f.u.) based on four cations and 12 positive charges. (a) Tetrahedrally coordinated Al (AlIV) as a function of the sum of Mg + Fe2+; labelled lines indicate Tschermak’s and Ti-Tschermak’s substitution respectively; (b) Ca as a function of temperature (°C); (c) Ti as a function of temperature; (d) Na as a function of temperature. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Orthopyroxene Orthopyroxene (opx) forms the liquidus phase in the basaltic andesite experiments, where it persists down to 1070°C. In the fractional crystallization experiments opx disappears at 1050°C in a peritectic reaction forming amphibole. In the oxidized equilibrium experiments (EQ Mb AuPd) opx appears only at 1040°C and persists to the lowest temperatures (980°C). Compositional variations of opx are illustrated in Fig. 14. Tetrahedrally coordinated Al reveals correlation along a 1:1 vector with Mg + Fe2+ identifying the Tschermak’s substitution as the principal mechanism for Al incorporation (Fig. 14a). The variation of Mg# (expressed as xMg) as a function of temperature shows a more pronounced decrease of the Mg# with decreasing temperature for fractional than for equilibrium crystallization experiments (Fig. 14b). The behavior of Ca as a function of temperature exhibits different behavior in the different experimental series (Fig. 14c), as follows. (1) The Ca content of the fractional experiments on the basaltic andesite shows an initial increase of Ca with decreasing temperature followed by a decrease as soon as cpx joins opx in the fractionating assemblage. (2) The reduced equilibrium crystallization experiments exhibit decreasing Ca with decreasing temperature as expected for opx–cpx equilibrium and evolve along the solvus to lower Ca contents with decreasing temperature. (3) The oxidized equilibrium crystallization experiments show invariant behavior of Ca with decreasing temperature that is consistent with constant to increasing Mg# related to progressive oxidation with decreasing temperature. (4) Opx crystallization in the reduced fractionation experiments results in increasing Ca with decreasing temperature despite the fact that it is cpx-saturated. This is the effect of a strongly increasing ferrosilite (Fe2Si2O6) component expanding the pyroxene solvus on the low-Ca side to higher Ca contents (e.g. Davidson & Lindsley, 1985). Titanium incorporation into opx (Fig. 14d) shows a coherent evolution, increasing with decreasing temperature for all series correlating with liquid TiO2 contents. Excluding sector zoned opx the partitioning of TiO2 between opx and melt (Kd TiO2opx/ TiO2liq ⁠) is 0·35 ± 0·04. Fig. 14. Open in new tabDownload slide Opx compositional variations expressed as cations per formula units (p.f.u.) based on four cations and 12 positive charges. (a) Tetrahedrally coordinated Al (AlIV) as a function of the sum of Mg + Fe2+; (b) xMg(Fetot) [=Mg/(Mg + Fe2+ + Fe3+)], (c) Ca and (d) Ti as a function of temperature (°C). Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 14. Open in new tabDownload slide Opx compositional variations expressed as cations per formula units (p.f.u.) based on four cations and 12 positive charges. (a) Tetrahedrally coordinated Al (AlIV) as a function of the sum of Mg + Fe2+; (b) xMg(Fetot) [=Mg/(Mg + Fe2+ + Fe3+)], (c) Ca and (d) Ti as a function of temperature (°C). Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Amphibole Amphibole crystallized over the temperature range (1060–720°C) delimited by the upper thermal stability of amphibole for basaltic to andesitic liquid compositions at 1·0 GPa (Eggler, 1972; Holloway & Burnham, 1972; Allen & Boettcher, 1978). Equilibrium crystallization experiments were restricted to just under 1000°C and show only limited compositional variations. Amphibole compositions from fractional crystallization experiments reveal systematic changes as a function of temperature, starting composition, mode of differentiation and fO2 conditions. Amphibole at high temperature (>900°C) is pargasite or ferroan pargasites (Leake et al., 1997) that change to tschermakite (900°C), tschermakitic hornblende (800°C) and finally magnesio-hornblende (720°C). AlIV decreases for the equilibrium crystallization experiments (Fig. 15a), although with an offset between the two series, the oxidized ones plotting at higher AlIV reflecting increased ferri-Tschermak’s substitution (Fe3+AlIV ⇔ MgSi). The fractional crystallization experiments reveal initial invariant (Mb series) or even increasing (ba series) AlIV that inverts to monotonous decrease below 900°C. The latter behavior is expected as AlIV is sensitive to temperature at constant pressure in buffered systems through the edenite exchange (NaAlIV ⇔ □Si, Blundy & Holland, 1990; Holland & Blundy, 1994). Fig. 15. Open in new tabDownload slide Amphibole compositional variations expressed as cations per formula units (p.f.u.) based on 46 positive charges excluding OH, F, Cl and fixed Fe3+/Fetot (see text for details) as a function of temperature. (a) Tetrahedrally coordinated Al (AlIV); labels close to experiment indicate temperatures (°C) and pressures (GPa) calculated with the algorithms proposed by Ridolfi & Renzulli (2012); (b) Ti; (c) Na + K on A-site; labels close to experiments correspond to temperatures calculated with the plag–amph thermometer (edenite–richterite) according to Holland & Blundy (1994); (d) Na on M4-site. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). Fig. 15. Open in new tabDownload slide Amphibole compositional variations expressed as cations per formula units (p.f.u.) based on 46 positive charges excluding OH, F, Cl and fixed Fe3+/Fetot (see text for details) as a function of temperature. (a) Tetrahedrally coordinated Al (AlIV); labels close to experiment indicate temperatures (°C) and pressures (GPa) calculated with the algorithms proposed by Ridolfi & Renzulli (2012); (b) Ti; (c) Na + K on A-site; labels close to experiments correspond to temperatures calculated with the plag–amph thermometer (edenite–richterite) according to Holland & Blundy (1994); (d) Na on M4-site. Legend as in Fig. 8; error bars correspond to standard deviations of multiple analyses (Tables 3 and 4). The Ti content of amphibole (Fig. 15b) shows very similar behavior to AlIV: decrease at high temperature for the equilibrium series and initial increase followed by monotonous decrease below 950°C for the fractional series, clearly pointing towards Ti-Tschermak’s substitution (⁠ TiAl2IV ⇔ MgSi2) and most probably decreasing TiO2 content of the coexisting liquid and increasing DTi(amph/liq) with decreasing temperature (e.g. Nandedkar et al., 2016). The decrease of total alkali content on the A-site with decreasing temperature reflects decreasing edenite component with decreasing temperature (Fig. 15c). For experiments with coexisting plagioclase the edenite–richterite temperatures calculated by the equations of Holland & Blundy (1994) result in excellent agreement with the experimental temperatures within 25°C. Na(M4) increases with decreasing temperatures for the fractional crystallization experiments under oxidized conditions (Fig. 15d), whereas the equilibrium crystallization experiments plot at low Na(M4) without significant variation with decreasing temperature. The increase of Na(M4) with deceasing temperature is due to the CaAlVI ⇔ Na(M4)Si (‘plagioclase’) substitution as Ca decreases with decreasing temperature. The total Na content of amphibole is rather constant and does not show any significant variation for the fractional crystallization experiments (0·65 ± 0·05 a.p.f.u). The oxidized equilibrium crystallization experiments show lower (0·6–0·45 a.p.f.u.) total Na contents that decrease with decreasing temperature, consistent with low Na2O contents of the coexisting liquids (1·2–1·5 wt %) challenging the commonly accepted view that amphibole saturates in calc-alkaline magmas at liquid Na2O contents ≥3 wt % (Cawthorn & O’Hara, 1976). Ferric iron probably stabilizes amphibole at low Na content. Oxide phases Different spinel group minerals are present throughout most of the experiments, and their compositional variation is illustrated in Fig. 16. Ilmenite was stable in the last fractionation step (1080°C) in the reduced (Pt–C) fractionation series and between 900 and 800°C in the oxidized fractional crystallization experiments. Equilibrium crystallization experiments starting with the natural high-Mg basalt saturated with a Cr-rich to Cr-poor spinel phase over a large temperature interval, 1210–1000°C under reducing conditions and 1200–980°C under oxidizing conditions. The fractional crystallization experiments had no Cr and no spinel phase crystallized at high temperature. In the Mb-series fractional crystallization experiments, the first spinel phase was hercynitic spinel at 1170°C that persisted to 1110°C, followed by Ti-bearing magnetite at 1080°C that became replaced by ilmenite at 900°C that in turn was replaced by Ti-rich magnetite in the lowest temperature experiment at 750°C. In the basaltic andesite series, Ti-bearing magnetite first formed at 990°C and is replaced by ilmenite at 900°C and in turn by Ti-rich magnetite at 720–750°C. Zoning of Cr-rich spinel in the equilibrium crystallization experiments is mostly due to the presence of low-MgO seeds (orange star in Fig. 16) from the natural high-Mg basaltic starting material. Spinel shows rapid re-equilibration of Fe2+–Mg but seems rather reluctant to exchange its Cr versus Al and/or Fe3+, resulting in strongly zoned crystals. The variation of yAl [= Al/(Al + Cr + Fe3+)] as a function of temperature (Fig. 16a) denotes various spinel phases observed in the experiments. Spinel in the reduced equilibrium crystallization experiments evolves from Cr-rich, Fe3+-poor to Al-rich, Cr-bearing hercynitic spinel and finally to Cr-poor hercynite (Fig. 16c and d). The equilibrium crystallization experiments under oxidized conditions produced a large variety of Cr-rich to Cr-poor, Al-, but also Fe3+-rich spinels with yFe 0·15–0·66. The oxidized fractional crystallization experiments with the high-Mg basaltic starting composition produced moderate Fe3+ (0·07–0·12 yFe) aluminous spinel at high temperatures (1170–1110°C) and Ti–Al–Mg-bearing magnetite at lower temperatures (1080–990°C), as did the basaltic andesite at 990°C (PU1049, 990°C) and both series at 720–750°C. Cr-rich spinel is a common accessory phase in the equilibrium crystallization experiments, but it does not exert a significant control on the liquid line of descent (LLD) except for Cr. This is different for hercynitic spinel, which is abundant just prior to amphibole saturation (Tables 3 and 4) and is responsible for SiO2 increase in the coexisting liquids. Similar relationships have been observed at 0·7 and 1·2 GPa (Müntener et al., 2001; Nandedkar et al., 2014), underlining the importance of spinel on the differentiation path from basalt to andesite/dacite. The relative oxidation of the cubic oxide phases is related to fO2. This is important for calc-alkaline systems where ilmenite is rarely coexisting with the cubic oxide phase, preventing the use of the Fe–Ti-oxide thermo-oxybarometer (Buddington & Lindsley, 1964; Lattard et al., 2005). Ilmenite in the reduced fractional crystallization experiments contains very low recalculated ferric iron (5% hematite component) whereas ilmenite in the oxidized fractional crystallization experiments (800–900°C, ba and Mb series) contains a higher, but variable hematite component of 15–35 mol %. Fig. 16. Open in new tabDownload slide Spinel phase (cubic oxide) compositional variations expressed as x- and y-parameters based on cations per formula units (p.f.u., three cations, eight positive charges). Individual spinel analyses are plotted because of large within-sample variations for some spinel phases (see text). (a) yAl [= Al/(Al + Cr + Fe3+)] as a function of experimental temperature (°C); labels within the diagram identify various types of spinels; low-temperature magnetite spinels plot outside the compositional and temperature space of the diagram and are indicated by an arrow. (b) yCr [= Cr/(Al + Cr + Fe3+)] as a function of yAl; dashed lines indicate corresponding yFe3+ values; (c) yFe [= Fe3+/(Al + Cr + Fe3+)] as a function of temperature; (d) ternary yAl–yFe–yCr diagram summarizing spinel phase compositional variations. Legend identifies spinels from single experiments (Tables 2, 3 and 4). Fig. 16. Open in new tabDownload slide Spinel phase (cubic oxide) compositional variations expressed as x- and y-parameters based on cations per formula units (p.f.u., three cations, eight positive charges). Individual spinel analyses are plotted because of large within-sample variations for some spinel phases (see text). (a) yAl [= Al/(Al + Cr + Fe3+)] as a function of experimental temperature (°C); labels within the diagram identify various types of spinels; low-temperature magnetite spinels plot outside the compositional and temperature space of the diagram and are indicated by an arrow. (b) yCr [= Cr/(Al + Cr + Fe3+)] as a function of yAl; dashed lines indicate corresponding yFe3+ values; (c) yFe [= Fe3+/(Al + Cr + Fe3+)] as a function of temperature; (d) ternary yAl–yFe–yCr diagram summarizing spinel phase compositional variations. Legend identifies spinels from single experiments (Tables 2, 3 and 4). Feldspar Plagioclase was observed in the low-H2O reduced fractional crystallization experiments (FC Mb Pt–C) at temperatures below 1170°C and in the oxidized fractional crystallization experiments below 950°C. Compositions in equilibrium with dacitic liquids are anorthite-rich (An78–84, 900–850°C) and change sharply to lower An contents (An45–55) at 800°C and 750°C coexisting with rhyodacitic to rhyolitic liquids; the plag at 720°C saturated with qtz, mag, amph ± garnet has an An content of 37. The partitioning of Ca/Na between plag and coexisting liquid (see Supplementary Data Electronic Appendix EA1; supplementary data are available for downloading at http://www.petrology.oxfordjournals.org) expressed as distribution coefficient D(Ca/Naplag)/(Ca/Naliq) decreases from about four to two for the oxidized experiments despite high and increasing H2O contents towards low temperatures; that is, it is not simply a function of H2O content or fH2O in the system but is, additionally, strongly composition dependent as discussed by Waters & Lange (2015). Garnet Garnet was observed in the lowest temperature reducing experiment in the low-H2O fractional crystallization experiments (FC Mb Pt–C) coexisting with cpx, plagioclase, ilmenite and liquid (glass) and the oxidized FC experiments at temperatures of 950 down to 750°C. All garnets are relatively almandine-rich with >37% almandine component (for details, see Supplementary Data Electronic Appendix EA2). Garnet from the oxidized FC series shows consistent evolutionary trends with increasing almandine (Fe2+, 37–60) and grossular (Ca, 20–28) components and decreasing pyrope (Mg, 40–13) contents with decreasing temperature, except in the last step where pyrope is constant and almandine increases at the expense of grossular. The spessartine content (Mn) varies between 2·7 and 5·5% without any discernible correlation with temperature. The Ti content initially increases from 950 to 900°C followed by continuous decrease to lower temperature; the bend coincides with ilmenite saturation and, thus, strong and continuous depletion of TiO2 in the liquid indicating that Ti concentrations in garnet are controlled by partitioning between liquid and garnet. The total amount of garnet extracted between 950 and 720°C amounts to about 6·6 and 4·5 wt % relative to the initial starting material RC158c and 85-44 respectively; when normalized to andesite at 1000°C the amounts become nearly equal around 11% garnet fractionated between andesite and rhyolite. This mass fraction is likely to influence the trace element composition (rare earth elements, Sc, Y) of derivative liquids in the lower crust. The enhanced thermal stability of garnet in the reduced, low-H2O fractionation experiments is consistent with the results of Alonso-Perez et al. (2009) and Müntener et al. (2001), which both showed that lower amounts of H2O in the system increases the stability of garnet in andesitic liquids at pressures around 1·0–1·2 GPa. DISCUSSION Auto-oxidation of hydrous magmas during differentiation In this experimental study, oxygen fugacity conditions of experiments conducted in AuPd capsules were computed from the amount of ferric iron in the liquid phase that was estimated from the difference in the measured Fetot–Mg olivine–liquid partitioning coefficients and the nominal value of 0·324 at 1·0 GPa assuming all Fe is Fe2+ (Ulmer, 1989). The calculations were performed through Monte Carlo type simulations taking into account the analytical uncertainty of the olivine and liquid compositions. The fO2 values were obtained using the algorithm of Kress & Carmichael (1991) relating ferric iron content of the liquid to liquid composition, temperature and pressure conditions [Table 2; for details of the method see Kägi et al. (2005)]. The results illustrating the fO2 conditions as a function of temperature for the experiments in AuPd capsules that contained olivine are shown in Fig. 17a. The fractional crystallization experiments on the high-Mg basalt indicate fO2 conditions around NNO + 1, slightly higher than the targeted fO2 range. The behavior of the closed-system equilibrium experiments is distinctly different. The starting value around NNO + 2 corresponds to the ferric/ferrous iron ratio of the natural starting material (RC158c, Table 1). At temperatures below 1150°C, fO2 increases up to NNO + 5·5 (i.e. just above the hematite–magnetite equilibrium) at 1000°C where the remaining melt fraction is ∼20%. It has been inferred that mantle-derived, primary arc magmas are not necessarily more oxidized than, for example, mid-ocean ridge magmas, but that more evolved andesitic to dacitic liquids acquire their higher fO2 by differentiation of minerals with low ferric iron content such as olivine and clinopyroxene (e.g. Lee et al., 2005; Dauphas et al., 2009; Jenner et al., 2010). The more oxidized conditions are reflected in the abundant crystallization of magnetite and the increased ferric/ferrous iron ratio (e.g. Carmichael, 1991). Our data support this hypothesis. Using the modal proportions of phases determined by mass balance (Table 2) and solid phase compositions (Table 3) we obtained an excellent match of the Fe2O3 and FeO contents (5·95 and 3·00 wt %, respectively) of the liquid compared with the contents estimated by the Monte Carlo simulation, with a calculated difference of 0·09 in Fe2O3 and –0·17 for FeO relative to the starting high-Mg basalt [run PU1004 (1060°C, ol, cpx, sp, liq)]. This indicates that auto-oxidation driven by fractionation of low-Fe3+ ferromagnesian phases is a viable process to oxidize hydrous arc magmas. The principal reason that this is much more effective for hydrous arc magmas than for nearly dry tholeiitic magmas is probably a consequence of the expanded olivine stability field at a given pressure towards more differentiated compositions (e.g. Kushiro, 1972, 1975). The effect might be further amplified by the observation that amphibole coincides with decreased amounts of magnetite in arc-related ultramafic (hornblendite) to gabbroic (amphibole-gabbro) cumulates, suppressing the potential depletion of ferric iron that would be caused by magnetite crystallization. Two high-temperature (1200 and 1160°C, PU908 and PU899) equilibrium crystallization experiments behave slightly differently, plotting at lower fO2; these experiments were conducted in BN-bearing piston cylinder assemblies, leading to more reducing conditions (e.g. Kägi et al., 2005). The difference between reducing and oxidizing conditions in equilibrium crystallization experiments is also reflected in the cpx/amphibole ratios. Under reducing conditions (P1060, P1000) the cpx/amphibole ratio is high, whereas under oxidizing conditions (PU905, 906 and 1005) the cpx/amphibole ratio is low. Oxidizing conditions, therefore, favor the formation of hornblende-rich cumulates in the roots of island arc volcanoes. The ‘auto-oxidation’ is further recorded in the compositions of the coexisting solid phases: spinel from the oxidized equilibrium experiments displays increasingly higher yFe with decreasing temperature than spinel from the other experiments (Fig. 16c and d). Likewise, the calculated Fe3+/Fetot ratio of opx exhibits a clear distinction between the reduced experiments (Pt–C), the fractional crystallization experiments around NNO (FC Mb and ba AuPd) and the oxidizing experiments (EQ Mb AuPd) (Fig. 17b). This diagram basically mimics Fig. 4 (olivine–liquid Fe–Mg Kd), further substantiating auto-oxidation as a possible process in fractionating magmatic systems. Fig. 17. Open in new tabDownload slide (a) log fO2 (ΔNNO) vs temperature diagram illustrating the evolution of fractional and equilibrium crystallization experiments under oxidizing conditions. Oxygen fugacity conditions were calculated based on the apparent olivine–liquid Fe–Mg distribution coefficients (Fig. 4, Table 2) assuming all Fe in olivine is ferrous and using the algorithm of Kress & Carmichael (1991) to relate computed Fe3+ contents of the liquid to fO2. Data points labelled BN were conducted with BN surrounding the noble metal capsules. The data points and curves labelled Model NNO + 2 and Model FMQ are the results of model calculations using modal proportions (Table 2) and compositions of minerals (Tables 3 and 4) to compute Fe2O3 and FeO in coexisting liquids that are converted to fO2 with the algorithm of Kress & Carmichael (1991). (b) Calculated ferric/total iron ratios of experimental opx based on cation normalization (four cations, 12 positive charges) as a function of temperature for various experiments under oxidizing (AuPd) and reducing (Pt–C) conditions. Fig. 17. Open in new tabDownload slide (a) log fO2 (ΔNNO) vs temperature diagram illustrating the evolution of fractional and equilibrium crystallization experiments under oxidizing conditions. Oxygen fugacity conditions were calculated based on the apparent olivine–liquid Fe–Mg distribution coefficients (Fig. 4, Table 2) assuming all Fe in olivine is ferrous and using the algorithm of Kress & Carmichael (1991) to relate computed Fe3+ contents of the liquid to fO2. Data points labelled BN were conducted with BN surrounding the noble metal capsules. The data points and curves labelled Model NNO + 2 and Model FMQ are the results of model calculations using modal proportions (Table 2) and compositions of minerals (Tables 3 and 4) to compute Fe2O3 and FeO in coexisting liquids that are converted to fO2 with the algorithm of Kress & Carmichael (1991). (b) Calculated ferric/total iron ratios of experimental opx based on cation normalization (four cations, 12 positive charges) as a function of temperature for various experiments under oxidizing (AuPd) and reducing (Pt–C) conditions. To quantify the potential effect of extended olivine extraction from a primary liquid, we modeled the fO2 evolution of the primary high-Mg basalt (RC158c) for two initial fO2 conditions: (1) the measured NNO + 2 of the natural RC158c simulating oxidized equilibrium experiments; (2) FMQ approaching inferred ‘pristine’ MOR-like fO2 conditions during partial melting (e.g. Ballhaus, 1993). The models considered the modal proportions and compositions of the EQ Mb AuPd series solid phases including the calculated Fe3+/Fetot ratio for cpx, opx and spinel for the NNO + 2 starting model, and lower Fe3+/Fetot ratios between the values obtained for the FC Mb AuPd and EQ Mb Pt–C series. The results are illustrated in Fig. 17a. The model can reconcile the observed fO2–temperature evolution of the oxidized equilibrium crystallization experiments slightly displaced to lower fO2 values. A similar evolution is shown for the model starting at FMQ (at 1280°C and 1·0 GPa = NNO – 0·31) evolving to NNO + 1·45 at 1060°C just before amphibole saturation. This corresponds to a closed-system oxidation of 1·8 log units, resulting in fO2 values typical for arc magmas. Saturation in magnetite might buffer the further evolution of andesite to rhyolite to near constant fO2 or even decrease it (e.g. Jenner et al., 2010). A similar process to account for the elevated fO2 conditions displayed by most arc magmas has been proposed by Lee et al. (2010) based on Zn/Fe systematics and is confirmed by the fO2–temperature relationship observed in the oxidized equilibrium crystallization experiments when fO2 was not buffered externally (as typically done in experimental studies), but was exclusively controlled by phase equilibria. A comparison of the evolution of fO2 along the LLD between fractional and equilibrium crystallization experiments is not possible; the fractional experiments have been forced to differentiate along the NNO buffer by resetting the ferric/ferrous ratio in every single step to the attempted NNO value, whereas the equilibrium crystallization experiments were allowed to evolve unconstrained (closed system) in T–fO2 space resulting in strong (+3·5 log units) oxidation from basalt to andesite. Further experimental studies, in particular fractional crystallization studies not resetting the fO2 in every step, should be conducted together with spectroscopic studies (Mössbauer, X-ray absorption spectroscopy (XAS)) to evaluate a possible oxidation during differentiation. Starting materials less oxidized than in the present case would clarify whether contrasting phase relations between tholeiitic (nearly dry) and calc-alkaline (more hydrous) magmas are responsible for different fO2 conditions reported for the dominant intermediate magmas rather than different fO2 conditions in their mantle sources moderated by absence or presence of slab-derived hydrous (oxidizing?) component as alternatively proposed (e.g. Parkinson & Arculus, 1999; Kelley & Cottrell, 2009). Comparison of liquid compositions with previous experimental studies The first part of the discussion presents a comparison of the experimentally derived liquid compositions of this study with those from previous studies on hydrous calc-alkaline and anhydrous (tholeiitic) differentiation at a pressure around 1·0 GPa. Experiments conducted with similar starting materials and comparable conditions include the equilibrium and fractional crystallization experiments on an anhydrous tholeiitic starting composition at 1·0 GPa (Villiger et al., 2004), the equilibrium crystallization experiments on hydrous basaltic starting materials at 1·0 GPa (Foden & Green, 1992; Kawamoto, 1996; Pichavant & Macdonald, 2007) and 0·9 GPa (Blatter et al., 2013), as well as the fractional crystallization study of Nandedkar et al. (2014) at 0·7 GPa. To distinguish the tholeiitic from the calc-alkaline differentiation series we plotted FeO/MgO versus SiO2 (Fig. 18a). Liquid compositions of this and most other studies straddle the calc-alkaline–tholeiitic discrimination line in the range basalt to andesite and, after excursion into the tholeiite field (fractional experiments on oxidized compositions from this study), end in the calc-alkaline field for rhyodacites and rhyolites. The notable exceptions are equilibrium and fractional crystallization experiments under anhydrous (Villiger et al., 2004) and low-H2O (FC Mb Pt–C series of this study) reducing conditions and some of the data obtained by Foden & Green (1992); only the differentiated liquids obtained by Kawamoto (1996) and Pichavant & Macdonald (2007) plot within the calc-alkaline field at intermediate compositions. The observation that most experimental studies on hydrous differentiation of basaltic magmas at elevated pressures and/or fO2 < NNO plot close to or even within the tholeiitic field has been discussed in detail by Sisson et al. (2005) and Blatter et al. (2013). This is also shown by the 0·7 and 1·0 GPa fractional crystallization experiments on the high-Mg basaltic starting material (Nandedkar et al., 2014; this study FC Mb AuPd). Fig. 18. Open in new tabDownload slide Liquid compositions of this study compared with other experiments at similar conditions on comparable starting materials. (a) FeOtot/MgO vs SiO2 (wt %) diagram with the tholeiitic and calc-alkaline fields according to Miyashiro (1974); (b) TAS (total alkali vs silica, Le Maitre et al., 1989) diagram used to denote compositions with volcanic rock names; (c) K2O (wt %) vs silica diagram discriminating low-, medium-, high-K and shoshonitic rock series. Abbreviations for experiments from this study are as in Fig. 8, EQ ba M01 3.8 and 5.0 denote equilibrium crystallization experiments from Müntener et al. (2001) on basaltic andesite 85-44 with 3·8 and 5·0 wt % initial H2O; EQ B13, Blatter et al. (2013); FC Mb N14, Nandedkar et al. (2014); EQ V04, Villiger et al. (2004) equilibrium crystallization experiments; FC V04, Villiger et al. (2004) fractional crystallization experiments; EQ P&M07, Pichavant & Macdonald (2007); EQ K96, Kawamoto (1996); EQ F&G92, Foden & Green (1992). Fig. 18. Open in new tabDownload slide Liquid compositions of this study compared with other experiments at similar conditions on comparable starting materials. (a) FeOtot/MgO vs SiO2 (wt %) diagram with the tholeiitic and calc-alkaline fields according to Miyashiro (1974); (b) TAS (total alkali vs silica, Le Maitre et al., 1989) diagram used to denote compositions with volcanic rock names; (c) K2O (wt %) vs silica diagram discriminating low-, medium-, high-K and shoshonitic rock series. Abbreviations for experiments from this study are as in Fig. 8, EQ ba M01 3.8 and 5.0 denote equilibrium crystallization experiments from Müntener et al. (2001) on basaltic andesite 85-44 with 3·8 and 5·0 wt % initial H2O; EQ B13, Blatter et al. (2013); FC Mb N14, Nandedkar et al. (2014); EQ V04, Villiger et al. (2004) equilibrium crystallization experiments; FC V04, Villiger et al. (2004) fractional crystallization experiments; EQ P&M07, Pichavant & Macdonald (2007); EQ K96, Kawamoto (1996); EQ F&G92, Foden & Green (1992). Figure 18b displays the TAS diagram (Le Maitre et al., 1989) showing the respective names associated with the different liquid compositions. Most experimental studies, including 1·0 GPa anhydrous experiments (Villiger et al., 2004), show an evolution from basalt to basaltic andesite and andesite or dacite. However, the equilibrium and fractional crystallization experiments under reducing conditions (EQ and FC Mb C + Pt) evolve to alkaline compositions through trachybasalts to basaltic trachyandesite. Overall, hydrous experiments under lower crustal conditions from this and previous studies reveal the typical calc-alkaline differentiation trends from basalt to dacite to (high-silica) rhyolite. Inspection of the K2O–SiO2 diagram (Fig. 18c) reveals the importance of (1) the starting material and (2) the fraction of liquid left for a given SiO2. The anhydrous crystallization experiments (Villiger et al., 2004) were conducted with a low-K olivine tholeiite and evolve to high K contents at 1060°C with only 3·5 wt % melt left relative to the start, amplifying the initial K2O content of 0·08 wt % to over 2 wt %. The fractional experiments from this study and the equilibrium crystallization experiments from Müntener et al. (2001) employing the basaltic-andesite starting material initially straddle the low- to medium-K dividing line to andesites and then evolve towards medium-K dacites and rhyolites, whereas the higher K oxidized fractional crystallization experiments starting from the high-Mg basalt plot within the medium-K field, approaching the high-K field with increasing differentiation. In contrast, the equilibrium crystallization experiments under oxidized conditions (EQ Mb AuPd) evolve towards low-K compositions owing to extensive amphibole crystallization effectively removing alkalis. The reduced fractional and equilibrium crystallization experiments evolve towards high-K and even shoshonitic compositions owing to extensive pyroxene crystallization and late or absent amphibole crystallization leaving as little as 14 wt % liquid left relative to the start. This observation is consistent with the proposal of Meen (1987), suggesting that (anhydrous) crystallization of ‘normal-K’ calc-alkaline basalts at the base of the crust can lead to high-K or shoshonitic intermediate derivative liquids under reducing conditions. Such an evolution probably requires low initial H2O contents of 0·5–1·0 wt %. Liquid compositions obtained by various experimental studies are illustrated in oxide variation diagrams (Fig. 19). It should be noted that most previous studies on hydrous systems were crystallization experiments on fixed bulk compositions (i.e. equilibrium crystallization experiments) and thus are not strictly comparable with the fractional crystallization experimental datasets (Villiger et al., 2004; Nandedkar et al., 2014; this study). This additionally explains that these studies cover a more restricted temperature and/or composition space as the amount of residual liquids becomes very small. Fig. 19. Open in new tabDownload slide Liquid compositions of this study compared with other experiments at similar conditions on comparable starting materials. (a) xMg [=MgO/(FeOtot+MgO) molar] versus temperature (°C) diagram; (b) SiO2 (wt %) versus temperature; (c) TiO2 (wt %) versus temperature; (d) Al2O3 (wt %) versus temperature; (e) MgO (wt %) versus SiO2 (wt %); and (f) CaO (wt %) versus MgO (wt %). All symbols and abbreviations as in Figure 18. Fig. 19. Open in new tabDownload slide Liquid compositions of this study compared with other experiments at similar conditions on comparable starting materials. (a) xMg [=MgO/(FeOtot+MgO) molar] versus temperature (°C) diagram; (b) SiO2 (wt %) versus temperature; (c) TiO2 (wt %) versus temperature; (d) Al2O3 (wt %) versus temperature; (e) MgO (wt %) versus SiO2 (wt %); and (f) CaO (wt %) versus MgO (wt %). All symbols and abbreviations as in Figure 18. The highest liquidus temperatures (1330°C) observed are those of the high-Mg basalt under reducing conditions followed by the anhydrous tholeiitic system (1300°C) (Villiger et al., 2004), which has lower liquidus temperature because its normative olivine (and MgO) content is lower. Most other compositions have liquidus temperatures around 1200–1250°C; Foden & Green (1992) and Kawamoto (1996) did not delimit the liquidus and did not crystallize olivine in their highest temperature runs, indicating that their starting compositions were either not (near) primary and/or their respective H2O contents (5 and 1 wt %) were too low. The Mg# as a function of temperature reveals a systematic decrease from 0·70–0·77 consistent with primary mantle derived magmas to below 0·4 for derivative andesitic or dacitic liquids, and a further decrease to ∼0·2 for rhyodacites (Fig. 19a). This systematics inverts for the last steps towards rhyolites, consistently exhibiting Mg# of about 0·4 [the somewhat different behavior of the Nandedkar et al. (2014) fractional crystallization experiments is explained in their paper]. Exceptions are anhydrous and low-H2O fractional crystallization experiments under reducing conditions (Villiger et al., 2004; this study) that evolve to much lower Mg# at temperatures below 1150°C related to extensive pyroxene (and plagioclase) fractionation, and some of the lower temperature experiments of Foden & Green (1992) that crystallized amphibole and pyroxene but lack Fe–Ti-oxides. The latter experiments (containing 5 wt % H2O) conducted in AgPd sample containers enclosed in BN might have resulted in very low fO2 conditions [for a discussion see Kägi et al. (2005)], thus resembling the experiments conducted under reduced conditions in this work. The evolution of SiO2 as a function of temperature displays a fan-shaped pattern between basaltic and andesitic compositions (Fig. 19b). The equilibrium and fractional crystallization experiments under oxidizing conditions start at different levels but converge at lower temperatures together with the equilibrium crystallization experiments at 0·9 GPa of Blatter et al. (2013), all leading to andesite or dacite compositions around 1000°C. The fractional series subsequently continue towards rhyolite and are virtually identical, indicating phase equilibria control towards the granite minimum. The fractional experiments at 0·7 GPa (Nandedkar et al., 2014) evolve in parallel but are shifted to higher temperatures by about 50°C; that is, they are more silica-rich for a given temperature below 1000°C consistent with considerably higher An contents of fractionating plagioclase more efficiently enriching the liquid in silica. The compositions of Foden & Green (1992) and Kawamoto (1996) show a similar evolution, but with larger scatter. The equilibrium crystallization experiments of Müntener et al. (2001) and Pichavant & Macdonald (2007) at 1·2 and 1·0 GPa with lower H2O contents evolve towards more silica-rich compositions at higher temperatures. TiO2 as a function of temperature displays a bell-shaped evolution for different experimental series dictated by (1) the initial TiO2 content, (2) fO2 conditions controlling Fe–Ti-oxide saturation, and (3) H2O content influencing amphibole saturation. The higher fO2 experiments on high-Mg basalt and basaltic andesite (this study) display slightly increasing TiO2 until Fe–Ti-oxide and/or amphibole saturates around 1050°C, and thereafter TiO2 monotonically decreases (Fig. 19c). The evolution is similar for the equilibrium crystallization experiments of Blatter et al. (2013), although shifted to higher TiO2 values owing to their higher initial TiO2. Anhydrous and low-H2O experiments under reducing conditions evolve to higher TiO2 contents owing the absence of amphibole and delayed Fe–Ti-oxide saturation. Al2O3 reveals very contrasting behavior as a function of temperature dictated by the presence or absence of plagioclase and hercynitic spinel along the liquid line of descent. Fractional crystallization experiments under oxidized conditions at 1·0 GPa are characterized by an increase of Al2O3 to about 950°C and a continuous decrease thereafter, with the onset of plagioclase and garnet crystallization effectively depleting Al2O3 in the liquid. The 0·7 GPa experiments reach a maximum at 1040°C owing to earlier saturation in plagioclase with decreasing pressure. All equilibrium crystallization experiments (EQ Mb AuPd and Pt–C) do not saturate in plagioclase down to temperatures of 980–1000°C, thus Al2O3 increases with increasing differentiation. The Blatter et al. (2013) experiments crystallize plagioclase at 0·9 GPa at temperatures below 1095°C, consistent with their relatively low H2O content (2 wt %) and slightly lower pressure. The evolution of liquid Al2O3 underlines the important role of H2O controlling the relative stabilities of plagioclase and amphibole. At lower temperatures, investigated only in this study, garnet exerts a significant control on Al2O3 liquid concentration. MgO as a function of SiO2 displays a monotonic decrease principally controlled by the initial MgO and SiO2 contents; only the low-H2O experiments on calc-alkaline compositions (FC Mb Pt–C and Foden & Green, 1992) deviate somewhat from the general trend, plotting at lower MgO contents for a given SiO2 consistent with their extensive and early crystallization of plagioclase limiting silica increase. CaO shows a consistent increase at high MgO controlled by olivine fractionation, followed by a decrease along a single trajectory for most studies controlled by cpx, amphibole and plagioclase crystallization (Fig. 19f). At the lowest temperatures both fractional series converge but show a displacement towards lower MgO for a given CaO content compared with the experiments of Blatter et al. (2013) and Nandedkar et al. (2014), which is most probably the effect of later plagioclase saturation and overall lower An contents. A significant displacement from the common evolution is shown by equilibrium crystallization of the high-Mg basalt, which is shifted towards higher CaO at lower MgO contents consistent with olivine (+ Cr-spinel) crystallization and delayed cpx saturation. A similar evolution is observed for some of the experiments of Foden & Green (1992) and Pichavant & Macdonald (2007). These experiments are characterized by large amounts of amphibole and very minor cpx. A large amount of dissolved H2O in the liquid and/or high fO2 may favor high abundances of amphibole that drive equilibrium crystallization to rather unusual compositions. In summary, the fractionation experiments on primitive, hydrous magmas show similar evolutions at lower crustal conditions, generally differentiating to intermediate andesitic to dacitic derivative liquids around 1000°C. Major controls are exerted by the absolute (or relative) H2O contents as well as fO2, ultimately leading to a variety of LLDs. The most discriminant oxides are TiO2, Al2O3 and CaO, which are mainly controlled by the phases olivine, cpx (opx), amphibole, plagioclase, garnet and Fe–Ti-oxide. Thus, to predict LLDs for a specific plutonic and/or volcanic suite requires information on the H2O content of parental magmas, fO2 conditions and pressure. This study and that of Nandedkar et al. (2014) provide additional constraints on the evolution from intermediate to SiO2-rich compositions showing systematic differences as a function of pressure. The LLD at 1·0 GPa shows delayed plagioclase crystallization and occurrence of garnet and ilmenite, the latter being replaced by ulvöspinel–magnetite solid-solution at 0·7 GPa. Importantly, the LLDs for compositions from dacite to rhyolite completely converge for all elements, probably controlled by amphibole, as also shown for Mt St Helens magmas by Blatter et al. (2017). Potassium behaves nearly completely incompatibly down to the lowest temperatures investigated and is a key element to identify sources of arc magmas. The metaluminous to peraluminous evolutionary trend of derivative liquids with increasing differentiation starting from hydrous calc-alkaline to arc-tholeiitic basaltic to basaltic andesite starting material is illustrated in Fig. 20. At 0·9–1·2 GPa, andesite and more evolved compositions (54–62 wt % SiO2) all become peraluminous, confirming that peraluminous dacitic to rhyolitic liquids are not only the product of anatectic melting of alumina-rich (e.g. metapelitic) sources, but can also be derived from metaluminous basalts by differentiation at elevated pressures dominated by early cpx and amphibole (e.g. Cawthorn & O’Hara, 1976; Cawthorn et al., 1976; Müntener et al., 2001; Blatter et al., 2013, 2017; Nandedkar et al., 2014). Previous experimental studies used for comparison terminate before reaching the peraluminous field, but the trend towards peraluminous compositions is also clearly indicated. Distinctly different are experiments under reducing conditions that remain metaluminous irrespective of their H2O contents—an expression of their more abundant opx crystallization driving them away from the metaluminous–peraluminous boundary—as well as the experiments by Foden & Green (1992) that exclusively crystallize amphibole and no pyroxenes. Fig. 20. Open in new tabDownload slide ASI [alumina saturation index, Al2O3/(CaO+Na2O+K2O) molar] as a function of SiO2 (wt %) of experimental liquids from this study and literature data. Symbols and sources of literature data as in Figure 18. Andesitic and more evolved derivative liquids in hydrous systems at 0.7–1.2 GPa are peraluminous. Fig. 20. Open in new tabDownload slide ASI [alumina saturation index, Al2O3/(CaO+Na2O+K2O) molar] as a function of SiO2 (wt %) of experimental liquids from this study and literature data. Symbols and sources of literature data as in Figure 18. Andesitic and more evolved derivative liquids in hydrous systems at 0.7–1.2 GPa are peraluminous. Liquid lines of descent and phase relations in pseudo-ternary projections Compositions of synthesized liquids for equilibrium and fractional crystallization experiments are plotted in a pseudoternary Ol–Cpx–Qtz diagram (Fig. 21) projected from plagioclase, orthoclase and Fe–Ti–Cr-oxides (Grove et al., 1992; Grove, 1993). Equilibrium and fractional crystallization experiments show significant differences for both starting compositions. The first steps of fractional crystallization of the basaltic andesite experiments are dominated by crystallization of opx and subsequent opx + cpx. The liquid line of descent (LLD) moves towards the Cpx-corner but turns away from the Ol–Cpx side-line when cpx joins the liquidus. Increasing cpx fractionation drives the LLD toward the Ol–Qtz sideline, and concomitant cpx and amph fractionation drives the liquids into the corundum-normative field (negative cpx component). The onset of amphibole crystallization coincides with a branching peritectic relationship of the form opx (+ cpx) + liquid → amphibole marked by the disappearance of opx and reduced cpx crystallization (Fig. 5a). Further evolution is dictated by amphibole and garnet fractionation driving the liquids towards the Qtz apex where a slightly peraluminous rhyolite is obtained. The liquid compositions of the equilibrium crystallization experiments (Müntener et al., 2001) are first driven towards the Ol–Cpx sideline by opx crystallization and then turn towards the Ol–Qtz sideline when cpx becomes the dominant solid phase, crossing the Ol–Qtz tieline around 1110°C. Pyroxene and finally amph crystallization drives the liquids towards intermediate corundum-normative compositions. Fig. 21. Open in new tabDownload slide Normative molecular, pseudoternary olivine (Ol)–clinopyroxene (Cpx)–quartz (Qtz) diagram (Grove et al., 1992; Grove, 1993) projected from plagioclase, orthoclase and Fe–Ti–Cr oxide for experimental liquids (glass). Labels denote run temperatures and coexisting solid assemblages for (a) experiments on high-Mg basaltic andesite (ba series, 85-44) including equilibrium crystallization experiments at 1·2 GPa with 3·8 and 5·0 wt % H2O after Müntener et al. (2001), and (b) high-Mg basalt series (Mb, RC158c). Negative cpx coordinates correspond to corundum-normative (peraluminous) liquids. Symbols and sources of literature data are as in Fig. 18. (For discussion see text.) Fig. 21. Open in new tabDownload slide Normative molecular, pseudoternary olivine (Ol)–clinopyroxene (Cpx)–quartz (Qtz) diagram (Grove et al., 1992; Grove, 1993) projected from plagioclase, orthoclase and Fe–Ti–Cr oxide for experimental liquids (glass). Labels denote run temperatures and coexisting solid assemblages for (a) experiments on high-Mg basaltic andesite (ba series, 85-44) including equilibrium crystallization experiments at 1·2 GPa with 3·8 and 5·0 wt % H2O after Müntener et al. (2001), and (b) high-Mg basalt series (Mb, RC158c). Negative cpx coordinates correspond to corundum-normative (peraluminous) liquids. Symbols and sources of literature data are as in Fig. 18. (For discussion see text.) The liquid evolution of the high-Mg basaltic starting material is shown in Fig. 20b. Three evolution paths can be identified: (1) fractional and equilibrium crystallization under oxidizing conditions lead to an evolution towards silica-rich intermediate (and SiO2-rich, fractional) liquids driven by olivine, cpx and later amphibole crystallization/fractionation. The difference between the two LLDs is the persistence of olivine as a liquidus phase in the equilibrium crystallization experiments to 1000°C with cpx (<1160°C) and amphibole (<1060°C), the final liquids being corundum-normative andesite. The fractional crystallization path evolves along lower normative cpx because of early cpx saturation (>1200°C). The peritectic reaction forming amphibole is cpx + liquid → amph not involving any opx. The lack of opx in this series is most probably a combination of high H2O content (low silica activity; see also Sisson & Grove, 1993) and high fO2 reducing opx stability relative to olivine and cpx. The crossover to peraluminous compositions plots at ∼75% normative qtz relative to ∼65% for the basaltic andesite, most probably owing to enhanced amphibole fractionation. Co-precipitation of amphibole and garnet drives liquids to less corundum-normative compositions between 900 and 750°C. The reduced experiments display very different liquid trends. The equilibrium crystallization experiments (EQ Mb Pt–C) first evolve away from olivine and then sharply turn away from cpx when cpx saturation (1180°C) followed by opx occurs. Olivine is replaced by opx, consistent with the reaction olivine + liquid → opx (Fig. 7c). Because the liquid composition is within the Ol–Opx–Cpx triangle while two pyroxenes fractionate, the liquids are driven away from the Cpx–Opx join towards the Ol–Cpx sideline and become silica-undersaturated trachybasalts to trachyandesites between 1090 and 1060°C, despite the fact that the last two crystallization steps involve amphibole that forms through a different reaction where mainly cpx and olivine is consumed and opx + amph form (cpx + ol + liq → amph + opx). This is mainly attributed to the low fO2 coupled to elevated CO2 contents from the reaction of the oxidized starting material with graphite. Elevated silica activity in the liquid stabilizes opx relative to olivine and cpx. Such phase relations have been identified in cumulates in the Southern Adamello massif where large amphiboles in hornblendites contain resorbed olivine and cpx primocrysts coexisting with newly formed (intercumulus) opx and strongly zoned amphibole overgrowth rims (Ulmer et al., 1983). Plagioclase is a late phase in these rocks and is mainly attributed to lower pressure intercumulus crystallization when the crystal mush was brought to shallower levels (Ulmer et al., 1983; Blundy & Sparks, 1992; Nimis & Ulmer, 1998). The fractional crystallization experiments under reducing, H2O-poor (2 wt %) conditions first show an evolution away from the Ol-apex when olivine is the liquidus phase, turn 180° when cpx and minor opx dominate and move towards the Ol–Qtz side-line and sharply turn in the last fractionation step when abundant garnet occurs moving away from Ol. Phase diagrams for fractionating hydrous magmas at lower crustal conditions Compositions of synthesized liquids of fractional crystallization experiments along with cpx, opx, hbl and garnet are plotted on the pseudoternary Ol–Cpx–Qtz diagram (Grove et al., 1992; Grove, 1993). Because each experiment uses new starting material, each melt coexisting with crystals is connected with a tieline to its starting composition. Arrows within mineral stability fields indicate compositional variations with decreasing temperature. The two different experimental series show some systematic differences with respect to their evolution along the liquid line of descent. The early evolution of the basaltic andesite (Fig. 22a) is tightly constrained by the two-pyroxene cotectic down to 1080°C. At 1050°C and lower temperatures two pyroxenes cease to coexist and the further crystallization path is determined by a distributary peritectic forming amphibole: cpx + liq → amph or opx + liq → amph. The cpx–amphibole cotectic line was mapped out by slightly varying the starting material and several runs between 990 and 950° formed amphibole only as the liquidus phase. But as the liquids evolve further garnet + amph + plg was stable instead of opx + amphibole at temperatures lower than 900°C, indicating that in the corundum-normative field the lower branch of the amphibole stability field is defined by amphibole + garnet. This means that the field of opx + amph is rather restricted despite the poorly defined grt + amph cotectic. Finally, at the lowest temperature, quartz is stable together with amphibole. Fig. 22. Open in new tabDownload slide Pseudo-ternary diagrams illustrating the liquid line of descent of basaltic andesite (a) and high-Mg basalt (b) together with coexisting minerals. Ol, olivine; Cpx, clinopyroxene; Qtz, quartz; Opx, orthopyroxene; gar, garnet; amp, amphibole. Arrows in mineral compositional fields indicate evolution with decreasing temperature. Open circles (starting material) and filled circles (experimental liquids) are connected with a tie-line. Dotted lines indicate approximate phase boundaries. Experiments with Fe loss >20% are not reported. Fig. 22. Open in new tabDownload slide Pseudo-ternary diagrams illustrating the liquid line of descent of basaltic andesite (a) and high-Mg basalt (b) together with coexisting minerals. Ol, olivine; Cpx, clinopyroxene; Qtz, quartz; Opx, orthopyroxene; gar, garnet; amp, amphibole. Arrows in mineral compositional fields indicate evolution with decreasing temperature. Open circles (starting material) and filled circles (experimental liquids) are connected with a tie-line. Dotted lines indicate approximate phase boundaries. Experiments with Fe loss >20% are not reported. The evolution of the high-Mg basalt is slightly different (Fig. 22b). During the early steps between 1200 and 1140°C the olivine–cpx cotectics is poorly constrained, probably owing to a change of the Mg# from 0·6 to 0·48 during the early stages of fractionation. Decreasing Mg# favors cpx over olivine, thereby expanding the clinopyroxene phase field. Coprecipitation of cpx + spl followed by cpx + opx + spl between 1110 and 1080°C drives the liquids to the right of the cpx–opx tieline, but as liquids further evolve at 1050°C opx disappears and is replaced by hornblende, indicating a similar reaction relationship opx + liq → hbl to that for the basaltic andesite. The further evolution between 1020 and 950°C is characterized by cpx + hbl-saturated liquids. Although there is some variability with respect to starting compositions and measured liquids (because of a combination of some Fe loss and compositional heterogeneity of the starting materials) the cpx–amph boundary is similar to the one from the basaltic andesite. As liquids develop down temperature to dacitic compositions, garnet appears together with amphibole + plagioclase, again suggesting that the amph + opx cotectics is rather restricted. It should be noted that experiments below 950°C saturated in garnet + plagioclase + amphibole are weakly corundum-normative dacitic liquids. The lack of hornblende in some of the experiments below 900°C could be due to small variations in the synthesized material and to different H2O contents in the liquids. The results from the two different starting materials show that the formation of hornblende at lower crustal conditions is mainly controlled by the distributary peritectic reaction cpx + opx +liq → hbl. In both cases the cpx + hbl cotectic line is substantially larger than the opx + amph cotectics, which is replaced by amph + garnet at temperatures below ∼900°C. This also indicates that garnets are likely to coexist with weakly corundum-normative dacitic liquids at 1·0 GPa conditions corresponding to the lowermost crust. A comparison of garnet compositions indicates that they are compositionally similar for both starting materials, illustrating that fractionation at 1·0 GPa in the lower crust produces dacites that might be garnet saturated. At the onset of garnet saturation, liquids have Mg# between 0·4 and 0·45. A comparison with other experimental studies in the pressure range of 0·9–1·2 GPa indicates that garnet saturation is common at 1·2 GPa (e.g. Müntener et al., 2001; Alonso-Perez et al., 2009), but rare at pressures of 0·9 and lower. Liquid compositions and/or fO2 might be critical to stabilize garnet at pressure below 1·0 GPa (e.g. Alonso-Perez et al., 2009; Blatter et al., 2017). Garnet saturation at 1·0 GPa might provide a natural upper limit for the peraluminosity of silicate liquids derived from fractionation in the lower crust. A comparison of experimentally derived liquids with natural data from the Cascades and Adamello Most plutonic and volcanic systems that cover a range of compositions from basalt to rhyolite provide evidence for complex histories with multiple generations of normally and reversely zoned crystals and mingling and mixing between diverse magmas (e.g. Eichelberger, 1975; Blundy & Sparks, 1992; Grove et al., 2005). It is therefore instructive to compare natural datasets with experimentally derived liquids. We illustrate the chemical data from the Cascades volcanic rocks (Du Bray et al., 2006) and the Adamello plutonic rocks (Ulmer et al., 1983; Hürlimann et al., 2016) in binary diagrams using MgO and the major oxides SiO2, CaO, Al2O3 and TiO2 (Fig. 23). The highly systematic variations in melt composition are evidence that the experimentally derived liquids are sufficiently close to natural compositions. There are subtle differences between the natural liquids and the experimentally derived liquids. In MgO vs SiO2 and CaO diagrams the natural liquids closely follow the fractional crystallization experiments to about 4–5 wt % MgO. Below 4 wt % MgO the fractional crystallization experiments at 1·0 GPa follow different trajectories and the natural data are better represented by lower pressure experiments (e.g. Blatter et al., 2013; Nandedkar et al., 2014). It should be noted that none of the natural datasets follow simple linear trends between 10 and 0 wt % MgO, precluding simple mixing of basalt and rhyolitic liquids to explain intermediate compositions. The similarity between experimental liquids and natural data suggests that most liquids or rock providing unequivocal evidence for mixing are formed from rather similar compositions, so that magma mixing at shallow pressures cannot be resolved by CaO and SiO2. Fig. 23. Open in new tabDownload slide Experimental liquid compositions compared with volcanic rocks from the Cascades (Du Bray et al., 2006) and plutonic rocks from Adamello (Macera et al., 1983; Ulmer et al., 1983; Blundy & Sparks, 1992; Hürlimann et al., 2016; P. Ulmer, unpublished data) plotted in wt % oxides. (a) MgO vs SiO2; (b) MgO vs CaO; (c) MgO vs TiO2; (d) MgO vs Al2O3. (For discussion, see text.) Fig. 23. Open in new tabDownload slide Experimental liquid compositions compared with volcanic rocks from the Cascades (Du Bray et al., 2006) and plutonic rocks from Adamello (Macera et al., 1983; Ulmer et al., 1983; Blundy & Sparks, 1992; Hürlimann et al., 2016; P. Ulmer, unpublished data) plotted in wt % oxides. (a) MgO vs SiO2; (b) MgO vs CaO; (c) MgO vs TiO2; (d) MgO vs Al2O3. (For discussion, see text.) A different picture emerges from the evaluation of MgO vs TiO2 (Fig. 23c). Whereas the Adamello data do not show any values exceeding 1·5 wt % TiO2, the Cascades data reach values up to 3 wt %. This can be explained by the ubiquitous evidence for (H2O-poor) arc tholeiites in the Cascades, which follow differentiation trends dominated by early plagioclase saturation, similar to anhydrous experiments (e.g. Grove et al., 1992; Villiger et al., 2004, 2007). Although experimental liquids synthesized in this study evolve similarly to subduction-zone magmas in many respects, additional processes may be discussed by evaluating MgO vs Al2O3 (Fig. 23d). At 5 wt % MgO, the Al2O3 values of a part of the Adamello data (Val Fredda series) closely follow the high-pressure experimental trajectories, whereas all other Adamello data follow the lower pressure experimental data. This difference between the Val Fredda series and all other Adamello data was noted previously (e.g. Ulmer et al., 1983; Blundy & Sparks, 1992) and indicates that both high- and low-pressure fractionation control the evolution of Adamello plutonic rocks. The Cascades data show a different evolution and the majority of data plot below the experimental trends, raising the possibility that mixing plays a more important role than illustrated by CaO and SiO2. It is important to note, however, that all the experimental series represent isobaric crystallization experiments, whereas natural liquids such as those for the Cascades probably result from polybaric crystallization (e.g. Grove et al., 2005). Near-liquidus phase relations in a decompressing magma (or decreasing H2O) indicate that the stability of plagioclase increases (e.g. Blundy & Cashman, 2001). This will probably lead to saturation in plagioclase preventing further enrichment of Al2O3 and could explain why peraluminous liquids are rare in natural datasets (e.g. Blatter et al., 2013). In addition to mixing, the Cascades trend thus could also indicate an important role for polybaric fractionation during decompression, an interesting further avenue for experimental studies. CONCLUSIONS This study reports several series of experiments conducted at 1·0 GPa to evaluate and quantify the liquid lines of descent of calc-alkaline to tholeiitic, hydrous, subduction-related primary mantle magmas by equilibrium and fractional crystallization. The most pertinent results are as follows. To produce intermediate andesitic/tonalitic derivative magmas from primary mantle melts, 40–80% of ultramafic cumulates have to be extracted. The amount is controlled by the extraction depth and composition of primary mantle melt and the process of crystallization. Fractional crystallization of Mg-rich basalt requires 15% of dunitic cumulates, followed by 45% of pyroxenite (cpx dominated) and 5% of cpx-hornblendite, whereas a primary basaltic andesite requires 30% of websterite followed by 10% cpx-hornblendite. Equilibrium crystallization under both oxidizing (NNO + 2) and reducing (graphite-saturated) conditions reaches intermediate andesite (high-fO2) to trachy-basaltic (low-fO2) liquids close to 1000°C after 80% crystallization of dunite, pyroxenites (cpx-dominated) and hornblendites. Oxygen fugacity (fO2) of hydrous, primitive, mantle-derived magmas increases through crystallization-differentiation owing to the significant expansion of the olivine (and to a lesser extent the cpx) stability field (>250°C, 1280–1000°C) relative to less hydrous magmas, resulting in efficient depletion in ferrous iron and a concomitant increase of ferric iron. The resulting fO2 of the coexisting basaltic liquids increases by more than 3 log units from about NNO + 2 to NNO + 5·5 over the temperature range >1200–1040°C in the case of an oxidized starting composition, and increases by about 1·8 log units from NNO – 0·3 (= FMQ) to NNO + 1·5 for a primary magma in equilibrium with (pristine) asthenospheric mantle, offering an explanation for the more oxidized character of hydrous arc magmas as opposed to low-H2O tholeiitic magmas where the olivine-stability field at similar conditions is much reduced (<70°C, 1300–1230°C). Enhanced amphibole fractionation at high pressure with respect to magnetite is consistent with the generally higher oxidation state of hydrous arc magmas leaving the lower crust. Crystallization from andesite to rhyolite is controlled by amphibole + garnet + plagioclase extraction, resulting in a similar LLD for both initial starting compositions below 900°C, leaving 16–20 wt % residual liquid at the lowest temperatures (720–750°C). Crystallization under more oxidizing conditions generates peraluminous intermediate to acidic differentiates crossing the metaluminous–peraluminous divide between 56 and 60 wt % SiO2. Crystallization of up to 11% garnet in the interval andesite–rhyolite will exert a strong control on some trace elements, in particular heavy rare earth elements, characteristic of such highly evolved magmas generated in the lower crust. Melt fractions as a function of temperature display a near linear decrease for Mg basalts and a strongly non-linear evolution for the high-Mg basalt. However, rescaled for basaltic andesite at 1050°C the evolution from andesite to dacite, rhyodacite and rhyolite is very similar with crystallization rates of 14·2 ± 0·6 and 13·5 ± 1·0% per 100°C. The contrasting behavior at higher temperature is controlled by extensive olivine and pyroxenite fractionation with a rate of 36 ± 3% per 100°C followed by a low rate of 7·5 ± 0·8% per 100°C relative to the initial mass for the Mg-rich basalt. In contrast, equilibrium crystallization of the high-Mg basalt under reduced and oxidizing conditions reveals an evolution in three steps: a low rate of olivine-only crystallization (∼10 ± 2% per 100°C), followed by a high crystallization rate (30–50% per 100°C) where pyroxenites and pyx-hornblendites form, and a final segment at lowest investigated temperatures (1060–980°C) with again a low crystallization rate (∼14% per 100°C). The contrasting behavior of the melt fraction evolution of the different primary magmas potentially controls the compositions of liquids extracted from lower crustal magma reservoirs. The extraction of liquids is most efficient when the rigid percolation threshold (Vigneresse et al., 1996) is reached, where dilute magmatic suspensions transform to crystal mushes around 50% crystallization (Caricchi et al., 2007; Pistone et al., 2013). This indicates that the high-Mg basaltic andesite will reach this threshold around 1000°C where the liquid is a cpx + amph-saturated peraluminous andesite (58·5 wt % SiO2), whereas the high-Mg basalt primary magma will reach this threshold around 1100°C with a two-pyroxene + spinel-saturated metaluminous, basaltic-andesite composition. Extraction of basic to intermediate magmas and ascent (decompression) accompanied by further crystallization will lead to a diversity of rising andesitic magmas. The further crystallization is controlled by lower pressure phase equilibria and mixing processes, widely documented in crystallizing plutonic complexes and erupted arc magmas. The compositions of intermediate to SiO2-rich magmas generated by crystallization differentiation at lower crustal conditions deviate from the bulk of plutonic (e.g. Peninsular batholith, California: Morton et al., 2014); Adamello: Macera et al., 1983) or volcanic (e.g. Cascades: Blatter et al., 2013) rocks typically forming shallow-level complexes at convergent plate margins, in particular by their distinctly peraluminous character. Additional processes are required to obtain the typical, mostly metaluminous compositions of calc-alkaline magmas solidifying to tonalite–granodiorite–granite plutonic rocks or reaching the surface as andesite–dacite–rhyolite. Several mutually not exclusive processes are required, including the following. (a) Mixing of mantle-derived basic to intermediate differentiates (basaltic to basaltic andesite) with acidic magmas originating either from crystallization-differentiation from the former or from partial melting of (lower) crustal lithologies (e.g. Hildreth & Moorbath, 1988; Blatter et al., 2013). (b) Ascent-driven crystallization of derivative basaltic to andesitic magmas (e.g. Blundy & Cashman, 2001) controlled by plag–opx–amph-dominated crystallization and solidification of the cumulates, ultimately forming amphibole-gabbros that are typical of many batholiths worldwide. Near adiabatic decompression of basaltic andesitic magmas inevitably results in saturation in plagioclase, thereby keeping the ascending magmas in the metaluminous field. (c) Crystallization of primary mantle-derived magmas at shallow levels (≤0·4 GPa). Existing phase equilibria on basaltic compositions (e.g. Sisson & Grove, 1993; Blatter et al., 2013; Nandedkar, 2013) reveal that derivative compositions remain metaluminous and compositions generally evolve towards the observed intermediate to acidic rocks. However, the paucity of mafic–ultramafic cumulates in the upper crust that would be necessary to balance the observed tonalite to granite compositions constituting the bulk of the upper crust indicates that this process is volumetrically minor. (d) It has been proposed that magmatic differentiation in arcs dominantly occurs in the lower crust (e.g. Müntener & Ulmer, 2006; in so-called ‘deep crustal hot zones’: Annen et al., 2006). This model explains the general lack of mafic to ultramafic cumulates in the crust by foundering of the lower arc crust directly back into the upper mantle without requiring them to sink through the highly viscous continental crust (Kay & Kay, 1991; Glazner, 1994; Jull & Kelemen, 2001), but requires a fundamental understanding of the processes modifying the basaltic to andesitic derivative liquids upon extraction from their crystal mush in the lower crust and ascent to shallow crustal levels to obtain the observed compositions. Decompression and mixing during magma ascent (e.g. Eichelberger, 1975; Blundy & Cashman, 2001) in transcrustal plumbing systems are the most likely processes that explain arc plutonic and volcanic diversity. ACKNOWLEDGEMENTS We would like to acknowledge Eric Reusser for his help in EPMA and SEM analysis and micro-Raman spectroscopy. Constructive reviews by Michel Pichavant, Renat Almeev and Yoshi Tamura, and the editorial handling of Gerhard Wörner helped to sharpen the presentation and are gratefully acknowledged. SUPPLEMENTARY DATA Supplementary data for this paper are available at Journal of Petrology online. FUNDING This work has been supported by several Swiss National Science Foundation grants (20-50661.97, PDFMP2-123097/1, PDTMP2-123074 and 20020-156408/1) and the support of the Herbette foundation. REFERENCES Allen J. C. , Boettcher A. L. ( 1978 ). Amphiboles in andesites and basalts: II. Stability as a function of P–T–fO2–fH2O . American Mineralogist 63 , 1074 – 1087 . Google Scholar OpenURL Placeholder Text WorldCat Alonso-Perez R. , Müntener O., Ulmer P. ( 2009 ). Igneous garnet and amphibole fractionation in the roots of island arcs: experimental constraints on andesitic liquids . Contributions to Mineralogy and Petrology 157 , 541 – 558 . Google Scholar Crossref Search ADS WorldCat Amma-Miyasaka M. , Nakagawa M. ( 2002 ). Origin of anorthite and olivine megacrysts in island-arc tholeiites: petrological study of 1940 and 1962 ejecta from Miyake-jima volcano, Izu–Mariana arc . Journal of Volcanology and Geothermal Research 117 , 263 – 283 . Google Scholar Crossref Search ADS WorldCat Annen C. , Blundy J. D., Sparks R. S. J. ( 2006 ). The genesis of intermediate to silicic magmas in deep crustal hot zones . Journal of Petrology 47 , 505 – 539 . Google Scholar Crossref Search ADS WorldCat Arndt N. T. , Goldstein S. L. ( 1989 ). An open boundary between lower continental crust and mantle: its role in crust formation and crustal recycling . Tectonophysics 161 , 201 – 212 . Google Scholar Crossref Search ADS WorldCat Atherton M. P. ( 1990 ). The coastal batholith of Peru: the product of rapid recycling of ‘new’ crust formed within rifted continental margin . Geological Journal 25 , 337 – 349 . Google Scholar Crossref Search ADS WorldCat Baker M. B. , Grove T. L., Price R. ( 1994 ). Primitive basalts and andesites from the Mt. Shasta region, N. California: products of varying melt fraction and water content . Contributions to Mineralogy and Petrology 118 , 111 – 129 . Google Scholar Crossref Search ADS WorldCat Ballhaus C. ( 1993 ). Redox state of lithospheric and asthenospheric upper mantle . Contributions to Mineralogy and Petrology 114 , 331 – 348 . Google Scholar Crossref Search ADS WorldCat Blatter D. L. , Sisson T. W., Hankins W. B. ( 2013 ). Crystallization of oxidized, moderately hydrous arc basalt at mid- to lower-crustal pressures: implications for andesite genesis . Contributions to Mineralogy and Petrology 166 , 861 – 886 . Google Scholar Crossref Search ADS WorldCat Blatter D. L. , Sisson T. W., Hankins W. B. ( 2017 ). Voluminous arc dacites as amphibole reaction boundary liquids . Contributions to Mineralogy and Petrology 172 , 27 . Google Scholar Crossref Search ADS WorldCat Blundy J. D. , Cashman K. ( 2001 ). Ascent-driven crystallization of dacite magma at Mount St Helens, 1980–1986 . Contributions to Mineralogy and Petrology 140 , 631 – 650 . Google Scholar Crossref Search ADS WorldCat Blundy J. D. , Holland T. J. B. ( 1990 ). Calcic amphibole equilibria and a new amphibole–plagioclase geothermometer . Contributions to Mineralogy and Petrology 104 , 208 – 224 . Google Scholar Crossref Search ADS WorldCat Blundy J. D. , Sparks R. S. J. ( 1992 ). Petrogenesis of mafic inclusions in granitoids of the Adamello Massif, Italy . Journal of Petrology 33 , 1039 – 1104 . Google Scholar Crossref Search ADS WorldCat Bohlen S. R. , Essene E. J., Boettcher A. L. ( 1980 ). Reinvestigation and application of olivine–quartz–orthopyroxene barometry . Earth and Planetary Science Letters 47 , 1 – 10 . Google Scholar Crossref Search ADS WorldCat Bose K. , Ganguly J. ( 1995 ). Quartz–coesite transition revisited—reversed experimental-determination at 500–1200°C and retrieved thermochemical properties . American Mineralogist 80 , 231 – 238 . Google Scholar Crossref Search ADS WorldCat Botcharnikov R. , Freise M., Holtz F., Behrens H. ( 2005 ). Solubility of C–O–H mixtures in natural melts: new experimental data and application range of recent models . Annals of Geophysics 48 , 633 – 646 . Google Scholar OpenURL Placeholder Text WorldCat Bowen N. L. ( 1928 ). The Evolution of the Igneous Rocks . Princeton, NJ : Princeton University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Bowen N. L. , Schairer J. F. ( 1935 ). The system MgO–FeO–SiO2 . American Journal of Science 29 , 151 – 217 . Google Scholar Crossref Search ADS WorldCat Brooker R. A. , Holloway J. R., Hervig R. L. ( 1998 ). Reduction in piston-cylinder experiments: the detection of carbon infiltration into platinum capsules . American Mineralogist 83 , 985 – 994 . Google Scholar Crossref Search ADS WorldCat Buddington A. F. , Lindsley D. H. ( 1964 ). Iron–titanium oxide minerals and synthetic equivalents . Journal of Petrology 5 , 310 – 357 . Google Scholar Crossref Search ADS WorldCat Callegari E. ( 1983 ). Geological and petrological aspects of the magmatic activity at Adamello (Northern Italy) . Memorie della Società Geologica Italiana 26 , 83 – 104 . Google Scholar OpenURL Placeholder Text WorldCat Caricchi L. , Ulmer P., Peccerillo A. ( 2006 ). A high-pressure experimental study on the evolution of the silicic magmatism of the main Ethiopian Rift . Lithos 91 , 46 – 58 . Google Scholar Crossref Search ADS WorldCat Caricchi L. , Burlini L., Ulmer P., Gerya T., Vassalli M., Papale P. ( 2007 ). Non-Newtonian rheology of crystal-bearing magmas and implications for magma ascent dynamics . Earth and Planetary Science Letters 264 , 402 – 419 . Google Scholar Crossref Search ADS WorldCat Carmichael I. S. E. ( 1991 ). The redox states of basic and silicic magmas: a reflection of their source regions? Contributions to Mineralogy and Petrology 106 , 129 – 141 . Google Scholar Crossref Search ADS WorldCat Cawthorn R. G. , O’Hara M. J. ( 1976 ). Amphibole fractionation in calc-alkaline magma genesis . American Journal of Science 276 , 309 – 329 . Google Scholar Crossref Search ADS WorldCat Cawthorn R. G. , Strong D. F., Brown P. A. ( 1976 ). Origin of corundum-normative intrusive and extrusive magmas . Nature 259 , 102 – 104 . Google Scholar Crossref Search ADS WorldCat Clemens J. D. , Stevens G., Farina F. ( 2011 ). The enigmatic sources of I-type granites: the peritectic connexion . Lithos 126 , 174 – 181 . Google Scholar Crossref Search ADS WorldCat Dauphas N. , Craddock P. R., Asimow P. D., Bennett V. C., Nutman A. P., Ohnenstetter D. ( 2009 ). Iron isotopes may reveal the redox conditions of mantle melting from Archean to present . Earth and Planetary Science Letters 288 , 255 – 267 . Google Scholar Crossref Search ADS WorldCat Davidson P. M. , Lindsley D. H. ( 1985 ). Thermodynamic analysis of quadrilateral pyroxenes Part II. Model calculation from experiments and applications to geothermometry . Contributions to Mineralogy and Petrology 91 , 390 – 404 . Google Scholar Crossref Search ADS WorldCat Dessimoz M. , Müntener O., Ulmer P. ( 2012 ). A case for hornblende dominated fractionation of arc magmas: the Chelan Complex (Washington Cascades) . Contributions to Mineralogy and Petrology 163 , 567 – 589 . Google Scholar Crossref Search ADS WorldCat Di Muro A. , Villemant B., Montagnac G., Scaillet B., Reynard B. ( 2006 ). Quantification of water content and speciation in natural silicic glasses (phonolite, dacite, rhyolite) by confocal microRaman spectrometry . Geochimica et Cosmochimica Acta 70 , 2868 – 2884 . Google Scholar Crossref Search ADS WorldCat Drummond M. S. , Defant M. J. ( 1990 ). A model for trondhjemite–tonalite–dacite genesis and crustal growth via slab melting: Archean to modern comparison . Journal of Geophysical Research 95 , 21503 – 21521 . Google Scholar Crossref Search ADS WorldCat Du Bray E. A. , John D. A., Sherrod D. R., Evarts R. C., Conrey R. M., Lexa J. ( 2006 ). Geochemical database for volcanic rocks of the Western Cascades, Washington, Oregon, and California . US Geological Survey Data Series 155 , 49 . Google Scholar OpenURL Placeholder Text WorldCat Eggler D. H. ( 1972 ). Amphibole stability in H2O-undersaturated calc-alkaline melts . Earth and Planetary Science Letters 15 , 28 – 34 . Google Scholar Crossref Search ADS WorldCat Eichelberger J. C. ( 1975 ). Origin of andesite and dacite: evidence of mixing at Glass Mountain in California and at other circum-Pacific volcanoes . Geological Society of America Bulletin 86 , 1381 – 1391 . Google Scholar Crossref Search ADS WorldCat Foden J. D. , Green D. H. ( 1992 ). Possible role of amphibole in the origin of andesite: some experimental and natural evidence . Contributions to Mineralogy and Petrology 109 , 479 – 493 . Google Scholar Crossref Search ADS WorldCat Frost D. J. , Wood B. J. ( 1995 ). Experimental measurements of the graphite C–O equilibrium and CO2 fugacities at high temperatures and pressure . Contributions to Mineralogy and Petrology 121 , 303 – 308 . Glazner A. F. ( 1994 ). Foundering of mafic plutons and density stratification of continental crust . Geology 22 , 435 – 438 . Google Scholar Crossref Search ADS WorldCat Goldstein J. I. , Newbury D. E., Joy D. C., Lyman C. E., Echlin P., Lifshin E., Sawyer L., Michael J. R. ( 2003 ). Scanning Electron Microscopy and X-Ray Microanalysis . Berlin : Springer . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Green T. H. ( 1980 ). Island arc and continent building magmatism—a review of petrogenetic models based on experimental petrology and geochemistry . Tectonophysics 63 , 367 – 385 . Google Scholar Crossref Search ADS WorldCat Grove T. L. ( 1993 ). Corrections to expressions for calculating mineral components in ‘Origin of calc-alkaline series lavas at Medicine Lake Volcano by fractionation, assimilation, and mixing’ and ‘Experimental petrology of normal MORB near the Kane Fracture Zone: 22°–25°N, mid-Atlantic ridge’ . Contributions to Mineralogy and Petrology 114 , 422 – 424 . Google Scholar Crossref Search ADS WorldCat Grove T. L. , Kinzler R. J., Bryan W. B. ( 1992 ). Fractionation of mid-ocean ridge basalt (MORB). In: Morgan J. P., Blackman D. K., Sinton J. M. (eds) Mantle Flow and Melt Generation at Mid-Ocean Ridges. American Geophysical Union, Geophysical Monograph 71 , 281 – 310 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Grove T. L. , Parman S. W., Bowring S. A., Price R. C., Baker M. B. ( 2002 ). The role of an H2O-rich fluid component in the generation of primitive basaltic andesites and dacites from the Mt. Shasta region, N. California . Contributions to Mineralogy and Petrology 142 , 375 – 396 . Google Scholar Crossref Search ADS WorldCat Grove T. L. , Elkins-Tanton L. T., Parman S. W., Chatterjee N., Müntener O., Gaetani G. A. ( 2003 ). Fractional crystallization and mantle-melting controls on calc-alkaline differentiation trends . Contributions to Mineralogy and Petrology 145 , 515 – 533 . Google Scholar Crossref Search ADS WorldCat Grove T. L. , Baker M. B., Price R. C., Parman S. W., Elkins-Tanton L. T., Chatterjee N., Müntener O. ( 2005 ). Magnesian andesite and dacite lavas from Mt. Shasta, northern California: products of fractional crystrallization of H2O-rich mantle melts . Contributions to Mineralogy and Petrology 148 , 542 – 565 . Google Scholar Crossref Search ADS WorldCat Hall L. J. , Brodie J., Wood B. J., Carroll M. R. ( 2004 ). Iron and water losses from hydrous basalts contained in Au80Pd20 capsules at high pressure and temperature . Mineralogical Magazine 68 , 75 – 81 . Google Scholar Crossref Search ADS WorldCat Hernlund J. , Leinenweber K., Locke D., Tyburczy J. A. ( 2006 ). A numerical model for steady-state temperature distributions in solid-medium high-pressure cell assemblies . American Mineralogist 91 , 295 – 305 . Google Scholar Crossref Search ADS WorldCat Hildreth W. , Moorbath S. ( 1988 ). Crustal contributions to arc magmatism in the Andes of Central Chile . Contributions to Mineralogy and Petrology 98 , 455 – 489 . Google Scholar Crossref Search ADS WorldCat Holland T. , Blundy J. ( 1994 ). Non-ideal interactions in calcic amphiboles and their bearing on amphibole–plagioclase thermometry . Contributions to Mineralogy and Petrology 116 , 433 – 447 . Google Scholar Crossref Search ADS WorldCat Holloway J. R. , Burnham C. W. ( 1972 ). Melting relations of basalt with equilibrium water pressure less than total pressure . Journal of Petrology 13 , 1–29 . Google Scholar Crossref Search ADS WorldCat Hürlimann N. , Müntener O., Ulmer P., Nandedkar R., Chiaradia M., Ovtcharova M. ( 2016 ). Primary magmas in continental arcs and their differentiated products: petrology of a post-plutonic dyke suite in the Tertiary Adamello batholith (Alps) . Journal of Petrology 57 , 495 – 534 . Google Scholar Crossref Search ADS WorldCat Iacono-Marziano G. , Morizet Y., Le Trong E., Gaillard F. ( 2012 ). New experimental data and semi-empirical parameterization of H2O–CO2 solubility in mafic melts . Geochimica et Cosmochimica Acta 97 , 1 – 23 . Google Scholar Crossref Search ADS WorldCat Jagoutz O. , Burg J.-P., Hussain S., Dawood H., Pettke T., Iiszuka T., Maruyama S. ( 2009 ). Construction of the granitoid crust of an island arc part I: geochronological and geochemical constraints from the plutonic Kohistan (NW Pakistan) . Contributions to Mineralogy and Petrology 158 , 739 – 755 . Google Scholar Crossref Search ADS WorldCat Jagoutz O. , Müntener O., Schmidt M. W., Burg J.-P. ( 2011 ). The roles of flux- and decompression melting and their respective fractionation lines for continental crust formation: evidence from the Kohistan arc . Earth and Planetary Science Letters 303 , 25 – 36 . Google Scholar Crossref Search ADS WorldCat Jakobsson S. ( 2012 ). Oxygen fugacity control in piston-cylinder experiments . Contributions to Mineralogy and Petrology 164 , 397 – 406 . Google Scholar Crossref Search ADS WorldCat Jenner F. E. , O’Neill H. S. C., Arculus R. J., Mavrogenes J. A. ( 2010 ). The magnetite crisis in the evolution of arc-related magmas and the initial concentration of Au, Ag and Cu . Journal of Petrology 51 , 2445 – 2464 . Google Scholar Crossref Search ADS WorldCat Jull M. , Kelemen P. B. ( 2001 ). On the conditions for lower crustal convective instability . Journal of Geophysical Research 106 , 6423 – 6445 . Google Scholar Crossref Search ADS WorldCat Kägi R. , Müntener O., Ulmer P., Ottolini L. ( 2005 ). Piston-cylinder experiments on H2O undersaturated Fe-bearing systems: an experimental setup approaching fO2 conditions of natural calc-alkaline magmas . American Mineralogist 90 , 708 – 717 . Google Scholar Crossref Search ADS WorldCat Kawamoto T. ( 1996 ). Experimental constraints on differentiation and H2O abundance of calc-alkaline magmas . Earth and Planetary Science Letters 144 , 577 – 589 . Google Scholar Crossref Search ADS WorldCat Kay R. W. ( 1978 ). Aleutian magnesian andesites: melts from subducted oceanic crust . Journal of Volcanology and Geothermal Research 4 , 117 – 132 . Google Scholar Crossref Search ADS WorldCat Kay R. W. , Kay S. M. ( 1991 ). Creation and destruction of lower continental crust . Geologische Rundschau 80 , 259 – 278 . Google Scholar Crossref Search ADS WorldCat Kay S. M. , Ramos V. A. ( 2006 ). Evolution of an Andean margin: a tectonic and magmatic view from the Andes to the Neuquén Basin (35°–39°S lat.). Geological Society of America Special Papers 407 , 1 – 343 . Keller C. B. , Schoene B., Barboni M., Samperton K. M., Husson J. M. ( 2015 ). Volcanic–plutonic parity and the differentiation of the continental crust . Nature 523 , 301 – 307 . Google Scholar Crossref Search ADS WorldCat Kelley K. A. , Cottrell E. ( 2009 ). Water and the oxidation state of subduction zone magmas . Science 325 , 605 – 607 . Google Scholar Crossref Search ADS WorldCat Korolyuk Y. N. , Usova L. V., Nigmatulina E. N. ( 2009 ). Accuracy in the determination of the compositions of main rockforming silicates and oxides on a JXA-8100 microanalyer . Journal of Analytical Chemistry 64 , 1042 – 1046 . Google Scholar Crossref Search ADS WorldCat Kress V. C. , Carmichael I. S. E. ( 1991 ). The compressibility of silicate liquids containing Fe2O3 and the effect of composition, temperature, oxygen fugacity and pressure on their redox states . Contributions to Mineralogy and Petrology 108 , 82 – 92 . Google Scholar Crossref Search ADS WorldCat Kushiro I. ( 1972 ). Effects of H2O on the composition of magmas formed at high pressures . Journal of Petrology 13 , 311 – 344 . Google Scholar Crossref Search ADS WorldCat Kushiro I. ( 1974 ). Melting of hydrous upper mantle and possible generation of andesite magma. An approach from synthetic systems . Earth and Planetary Science Letters 22 , 294 – 299 . Google Scholar Crossref Search ADS WorldCat Kushiro I. ( 1975 ). On the nature of silicate melt and its significance in magma genesis: regularities in the shift of liquidus boundaries involving ol–opx and silica minerals . American Journal of Science 275 , 411 – 431 . Google Scholar Crossref Search ADS WorldCat Kushiro I. ( 1990 ). Partial melting of mantle wedge and evolution of island arc crust . Journal of Geophysical Research 95 , 15929 – 15939 . Google Scholar Crossref Search ADS WorldCat Lattard D. , Sauerzapf U., Käsemann M. ( 2005 ). New calibration data for the Fe–Ti oxide thermo-oxybarometers from experiments in the Fe–Ti–O system at 1 bar, 1000–1300°C and a large range of oxygen fugacities . Contributions to Mineralogy and Petrology 149 , 735 – 754 . Google Scholar Crossref Search ADS WorldCat Leake B. E. , Woolley A. R., Arps C. E. S., Birch W. D., Gilbert M. C., Grice J. D., Hawthorne F. C., Kato A., Kisch H. J., Krivovichev V. G., Linthout K., Laird J., Mandarino J. A., Maresch W. V., Nickel E. H., Rock N. M. S., Schumacher J. C., Smith D. C., Stephenson N. C. N., Ungaretti L., Whittaker E. J. W., Guo Y. Z. ( 1997 ). Nomenclature of amphiboles: report of the subcommittee on amphiboles of the International Mineralogical Association, commission on new minerals and mineral names . American Mineralogist 82 , 1019 – 1037 . Google Scholar OpenURL Placeholder Text WorldCat Lee C.-T. A. , Leeman W. P., Canil D., Li Z.-X. A. ( 2005 ). Similar V/Sc systematics in MORB and arc basalts: implications for the oxygen fugacities of their mantle source regions . Journal of Petrology 46 , 2313 – 2336 . Google Scholar Crossref Search ADS WorldCat Lee C. T. A. , Cheng X., Horodyskyj U. ( 2006 ). The development and refinement of continental arcs by primary basaltic magmatism, garnet pyroxenite accumulation, basaltic recharge and delamination: insights from the Sierra Nevada, California . Contributions to Mineralogy and Petrology 151 , 222 – 242 . Google Scholar Crossref Search ADS WorldCat Lee C.-T. A. , Luffi P., Le Roux V., Dasgupta R., Albarade F., Leeman W. P. ( 2010 ). The Redox state of arc mantle using Zn/Fe systematics . Nature 468 , 681 – 685 . Google Scholar Crossref Search ADS WorldCat Le Maitre R. W. , Bateman P., Dudek A., Keller J., Lameyre J., Le Bas M. J., Sabine P. A., Schmid R., Sorensen H., Streckeisen A., Woolley A. R., Zanettin B. ( 1989 ). A Classification of Igneous Rocks and Glossary of Terms: Recommendations of the International Union of Geological Sciences Subcommission on the Systematics of Igneous Rocks . Oxford : Blackwell Scientific . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Le Voyer M. , Rose-Koga E. F., Shimizu N., Grove T. L., Schiano P. ( 2010 ). Two contrasting H2O-rich components in primary melt inclusions from Mount Shasta . Journal of Petrology 51 , 1571 – 1595 . Google Scholar Crossref Search ADS WorldCat Lindsley D. H. ( 1983 ). Pyroxene thermometry . American Mineralogist 68 , 477 – 493 . Google Scholar OpenURL Placeholder Text WorldCat Luth R. W. ( 1989 ). Natural versus experimental control of oxidation state—effects on the composition and speciation of C–O–H-fluids . American Mineralogist 74 , 50 – 57 . Google Scholar OpenURL Placeholder Text WorldCat Macera P. , Ferrara G., Pescia A., Callegari E. ( 1983 ). A geochemical study on the acid and basic rocks of the Adamello Batholith . Memorie della Società Geologica Italiana 26 , 223 – 259 . Google Scholar OpenURL Placeholder Text WorldCat Meen J. K. ( 1987 ). Formation of shoshonites from calcalkaline basalt magmas: geochemical and experimental constraints from the type locality . Contributions to Mineralogy and Petrology 97 , 333 – 351 . Google Scholar Crossref Search ADS WorldCat Melekhova E. , Annen C., Blundy J. ( 2013 ). Compositional gaps in igneous rock suites controlled by magma system heat and water content . Nature Geoscience 6 , 385 – 390 . Google Scholar Crossref Search ADS WorldCat Mercier M. , Di Muro A., Giordano D., Metrich N., Lesne P., Pichavant M., Scaillet B., Clocchiatti R., Montagnac G. ( 2009 ). Influence of glass polymerisation and oxidation on micro-Raman water analysis in alumino-silicate glasses . Geochimica et Cosmochimica Acta 73 , 197 – 217 . Google Scholar Crossref Search ADS WorldCat Miyashiro A. ( 1974 ). Volcanic rock series in island arcs and active continental margins . American Journal of Science 274 , 321 – 355 . Google Scholar Crossref Search ADS WorldCat Morton D. M. , Miller F. K., Kistler R. W., Premo W. R., Lee C.-T. A., Langenheim V. E., Wooden J. L., Snee L. W., Clausen B. L., Cossette P. ( 2014 ). Framework and petrogenesis of the northern Penninsular Ranges Batholith, Southern California. In: Morton D. M., Miller F. K. (eds) Peninsular Ranges Batholith, Baja and Southern California. Geological Society of America, Memoirs 211 , 61 – 143 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Munoz-Espadas M.-J. , Lunar R., Martinez-Frias J. ( 2000 ). The garnet placer deposit from SE Spain: industrial recovery and geochemical features . Episodes 23 , 266 – 269 . Google Scholar OpenURL Placeholder Text WorldCat Müntener O. , Ulmer P. ( 2006 ). Experimentally derived high-pressure cumulates from hydrous arc magmas and consequences for the seismic velocity structure of lower arc crust . Geophysical Research Letters 33 , L21308 . Google Scholar Crossref Search ADS WorldCat Müntener O. , Ulmer P. ( 2018 ). Arc crust formation and differentiation constrained by experimental petrology . American Journal of Science 318 , 64 – 89 . Google Scholar Crossref Search ADS WorldCat Müntener O. , Kelemen P. B., Grove T. L. ( 2001 ). The role of H2O during crystallization of primitive arc magmas under uppermost mantle conditions and genesis of igneous pyroxenites: an experimental study . Contributions to Mineralogy and Petrology 141 , 643 – 658 . Google Scholar Crossref Search ADS WorldCat Nandedkar R. H. ( 2013 ). Evolution of Hydrous Mantle-Derived Calc-Alkaline Liquids by Fractional Crystallization at 0·7 and 0·4 GPa—An Experimental Study . PhD thesis, ETH Zurich, 232 pp. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Nandedkar R. H. , Ulmer P., Müntener O. ( 2014 ). Fractional crystallization of primitive, hydrous arc magmas: an experimental study at 0·7 GPa . Contributions to Mineralogy and Petrology 167 , 101 . Google Scholar Crossref Search ADS WorldCat Nandedkar R. H. , Hürlimann N., Ulmer P., Müntener O. ( 2016 ). Experimental amphibole–melt trace element partitioning in andesitic to rhyolitic systems at mid- to lower crustal conditions . Contributions to Mineralogy and Petrology 171 , 71 . Google Scholar Crossref Search ADS WorldCat Newman S. , Lowenstern J. B. ( 2002 ). VolatileCalc: a silicate melt–H2O–CO2 solution model written in Visual Basic for Excel . Computers and Geosciences 28 , 597 – 604 . Google Scholar Crossref Search ADS WorldCat Nimis P. , Ulmer P. ( 1998 ). Clinopyroxene geobarometry of magmatic rocks Part 1: an expanded structural geobarometer for anhydrous and hydrous, basic and ultrabasic systems . Contributions to Mineralogy and Petrology 133 , 122 – 135 . Google Scholar Crossref Search ADS WorldCat Parkinson I. J. , Arculus R. J. ( 1999 ). The redox state of subduction zones: insights from arc-peridotites . Chemical Geology 160 , 409 – 423 . Google Scholar Crossref Search ADS WorldCat Pichavant M. , Macdonald R. ( 2007 ). Crystallization of primitive basaltic magmas at crustal pressures and genesis of the calc-alkaline igneous suite: experimental evidence from St Vincent, Lesser Antilles arc . Contributions to Mineralogy and Petrology 154 , 535 – 558 . Google Scholar Crossref Search ADS WorldCat Pichavant M. , Martel C., Bourdier J.-L., Scaillet B. ( 2002 ). Physical conditions, structure, and dynamics of a zoned magma chamber: Mount Pelée (Martinique, Lesser Antilles Arc) . Journal of Geophysical Research 107 , doi:10.1029/2001JB000315. Google Scholar OpenURL Placeholder Text WorldCat Pistone M. , Caricchi L., Ulmer P., Reusser E., Ardia P. ( 2013 ). Rheology of volatile-bearing crystal mushes: mobilization vs. viscous death . Chemical Geology 345 , 16 – 39 . Google Scholar Crossref Search ADS WorldCat Rapp R. P. , Watson E. B. ( 1995 ). Dehydration melting of metabasalt at 8–32 kbar: implications for continental growth and crust–mantle recycling . Journal of Petrology 36 , 891 – 931 . Google Scholar Crossref Search ADS WorldCat Ratajeski K. , Sisson T. W. ( 1999 ). Loss of iron to gold capsules in rock-melting experiments . American Mineralogist 84 , 1521 – 1527 . Google Scholar Crossref Search ADS WorldCat Ridolfi F. , Renzulli A. ( 2012 ). Calcic amphiboles in calc-alkaline and alkaline magmas: thermobarometric and chemometric empirical equations valid up to 1,130°C and 2.2 GPa . Contributions to Mineralogy and Petrology 163 , 877 . Google Scholar Crossref Search ADS WorldCat Rudnick R. L. ( 1995 ). Making continental crust . Nature 378 , 571 – 578 . Google Scholar Crossref Search ADS WorldCat Schwandt C. S. , McKay G. A. ( 2006 ). Minor- and trace-element sector zoning in synthetic enstatite . American Mineralogist 91 , 1607 . Google Scholar Crossref Search ADS WorldCat Sisson T. W. , Bronto S. ( 1998 ). Evidence for pressure-release melting beneath magmatic arcs from basalt at Galunggung, Indonesia . Nature 391 , 883 – 888 . Google Scholar Crossref Search ADS WorldCat Sisson T. W. , Grove T. L. ( 1993 ). Experimental investigations of the role of H2O in calc-alkaline differentiation and subduction zone magmatism . Contributions to Mineralogy and Petrology 113 , 143 – 166 . Google Scholar Crossref Search ADS WorldCat Sisson T. W. , Ratajeski K., Hankins W. B., Glazner A. F. ( 2005 ). Voluminous granitic magmas from common basaltic sources . Contributions to Mineralogy and Petrology 148 , 635 – 661 . Google Scholar Crossref Search ADS WorldCat Taylor S. R. , White A. J. R. ( 1965 ). Geochemistry of andesites and the growth of continents . Nature 208 , 271 – 273 . Google Scholar Crossref Search ADS WorldCat Thompson A. B. , Matile L., Ulmer P. ( 2002 ). Some thermal constraints on crustal assimilation during fractionation of hydrous, mantle-derived magmas with examples from central Alpine batholiths . Journal of Petrology 43 , 403 – 422 . Google Scholar Crossref Search ADS WorldCat Ulmer P. ( 1986 ). Basische und ultrabasische Gesteine des Adamello (Provinzen Brescia und Trento, Norditalien). Dissertation, ETH Zürich, No. 8105. Ulmer P. ( 1988 ). High-pressure phase equilibria of calc-alkaline picrobasalt: implications for the genesis of calc-alkaline magmas . Annual Report of the Director, Geophysical Laboratory, Carnegie Institution of Washington 1987–1988 28 – 35 . Google Scholar OpenURL Placeholder Text WorldCat Ulmer P. ( 1989 ). The dependence of Fe2+–Mg cation-partitioning between olivine and basaltic liquid on pressure, temperature and composition; an experimental study . Contributions to Mineralogy and Petrology 101 , 261 – 273 . Google Scholar Crossref Search ADS WorldCat Ulmer P. , Luth R. W. ( 1991 ). The graphite–COH fluid equilibrium in P, T, fO2 space: an experimental determination to 30 kbar and 1600°C . Contributions to Mineralogy and Petrology 106 , 265 – 272 . Google Scholar Crossref Search ADS WorldCat Ulmer P. , Callegari E., Sonderegger U. ( 1983 ). Genesis of the mafic and ultramafic rocks and their genetical relations to the tonalitic–trondhjemitic granitoids of the southern part of the Adamello batholith (Northern Italy) . Memorie della Società Geologica Italiana 26 , 171 – 222 . Google Scholar OpenURL Placeholder Text WorldCat Vigneresse J. L. , Barbey P., Cuney M. ( 1996 ). Rheological transitions during partial melting and crystallization with application to felsic magma segregation and transfer . Journal of Petrology 37 , 1579 – 1600 . Google Scholar Crossref Search ADS WorldCat Villiger S. , Ulmer P., Müntener O., Thompson A. B. ( 2004 ). The liquid line of descent of anhydrous, mantle-derived, tholeiitic liquids by fractional and equilibrium crystallization—an experimental study at 1·0 GPa . Journal of Petrology 45 , 2369 – 2388 . Google Scholar Crossref Search ADS WorldCat Villiger S. , Ulmer P., Müntener O. ( 2007 ). Equilibrium and fractional crystallization experiments at 0.7 GPa; the effect of pressure on phase relations and liquid compositions . Journal of Petrology 48 , 159 – 184 . Google Scholar Crossref Search ADS WorldCat Waters L. E. , Lange R. A. ( 2015 ). An updated calibration of the plagioclase–liquid hygrometer–thermometer applicable to basalts through rhyolites . American Mineralogist 100 , 2172 – 2184 . Google Scholar Crossref Search ADS WorldCat Zhang Y. ( 1999 ). H2O in rhyolitic glasses and melts: measurement, speciation, solubility, and diffusion . Reviews of Geophysics 37 , 493 – 516 . Google Scholar Crossref Search ADS WorldCat © The Author(s) 2018. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. © The Author(s) 2018. Published by Oxford University Press.
TI - Experimentally Derived Intermediate to Silica-rich Arc Magmas by Fractional and Equilibrium Crystallization at 1·0 GPa: an Evaluation of Phase Relationships, Compositions, Liquid Lines of Descent and Oxygen Fugacity
JF - Journal of Petrology
DO - 10.1093/petrology/egy017
DA - 2018-01-01
UR - https://www.deepdyve.com/lp/oxford-university-press/experimentally-derived-intermediate-to-silica-rich-arc-magmas-by-60tv97RV5N
SP - 11
EP - 58
VL - 59
IS - 1
DP - DeepDyve
ER -