Journal of Petrology

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (54) Request Permissions Google Scholar Articles by SCAILLET, B. Articles by MACDONALD, R. Search for Related Content GeoRef GeoRef Citation Social Bookmarking          What's this?

Journal of Petrology Volume 42 Number 4 Pages 825-845 2001
© Oxford University Press 2001

Phase Relations of Peralkaline Silicic Magmas and Petrogenetic Implications

BRUNO SCAILLET^1,* and RAY MACDONALD²

¹INSTITUT DES SCIENCES DE LA TERRE D’ORLÉANS, CNRS, 1A RUE DE LA FÉROLLERIE, 45071, ORLÉANS CEDEX 2, FRANCE
²ENVIRONMENTAL SCIENCES DIVISION, IENS, LANCASTER UNIVERSITY, LANCASTER LA1 4YQ, UK

Received November 29, 1999; Revised typescript accepted July 21, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION ROCKS STUDIED EXPERIMENTAL AND ANALYTICAL... RESULTS DISCUSSION REFERENCES

 
The phase relationships of three peralkaline rhyolites from the Kenya Rift have been established at 150 and 50 MPa, at oxygen fugacities of NNO - 1·6 and NNO + 3·6 (log fO₂ relative to the Ni–NiO solid buffer), between 800 and 660°C and for melt H₂O contents ranging between saturation and nominally anhydrous. The stability fields of fayalite, sodic amphiboles, chevkinite and fluorite in natural hydrous silicic magmas are established. Additional phases include quartz, alkali feldspar, ferrohedenbergite, biotite, aegirine, titanite, montdorite and oxides. Ferrohedenbergite crystallization is restricted to the least peralkaline rock, together with fayalite; it is replaced at low melt water contents by ferrorichterite. Riebeckite–arfvedsonite appears only in the more peralkaline rocks, at temperatures below 750°C (dry) and below 670°C at H₂O saturation. Under oxidizing conditions, it breaks down to aegirine. In the more peralkaline rocks, biotite is restricted to temperatures below 700°C and conditions close to H₂O saturation. At 50 MPa, the tectosilicate liquidus temperatures are raised by 50–60°C, and that of amphibole by 30°C. Riebeckite–arfvedsonite stability extends down nearly to atmospheric pressure, as a result of its F-rich character. The solidi of all three rocks are depressed by 40–100°C compared with the solidus of the metaluminous granite system, as a result of the abundance of F and Cl. Low fO₂ lowers solidus temperatures by at least 30°C. Comparison with studies of metaluminous and peraluminous felsic magmas shows that plagioclase crystallization is suppressed as soon as the melt becomes peralkaline, whatever its CaO or volatile contents. In contrast, at 100 MPa and H₂O saturation, the liquidus temperatures of quartz and alkali feldspar are not significantly affected by changes in rock peralkalinity, showing that the incorporation of water in peralkaline melts diminishes the depression of liquidus temperatures in dry peralkaline silicic melts compared with dry metaluminous or peraluminous varieties. At 150 MPa, pre-eruptive melt H₂O contents range from 4 wt % in the least peralkaline rock to nearly 6 wt % in the two more peralkaline compositions, in broad agreement with previous melt inclusion data. The experimental results imply magmatic fO₂ at or below the fayalite–quartz–magnetite solid buffer, temperatures between 740 and 660°C, and melt evolution under near H₂O saturation conditions.

KEY WORDS: peralkaline; rhyolite; phase equilibria

	INTRODUCTION

TOP ABSTRACT INTRODUCTION ROCKS STUDIED EXPERIMENTAL AND ANALYTICAL... RESULTS DISCUSSION REFERENCES

 
Since the field recognition of peralkaline igneous rocks by Foerstner (1881), much effort has been expended in the petrological, geochemical and experimental fields to understand their petrogenesis (e.g. Washington, 1913, 1914a, 1914b; Bowen, 1937; Zies, 1960; Carmichael, 1962; Carmichael & MacKenzie, 1963; Bailey & Schairer, 1964, 1966; Thompson & MacKenzie, 1967; Ewart et al., 1968; Noble, 1968; Nicholls & Carmichael, 1969; Bailey & Macdonald, 1970; Gibson, 1972; Macdonald & Bailey, 1973; Bailey, 1974; Barberi et al., 1974, 1975; Villari, 1974; Roux & Varet, 1975; Bailey & Cooper, 1978; Harris, 1983; Mahood, 1984; Mungall & Martin, 1995; Bohrson & Reid, 1997). For the vast majority of natural occurrences, peralkaline felsic rocks can be explained by fractional crystallization of mildly alkalic or transitional basalts at both low fH₂O and low fO₂ (e.g. Harris, 1983; Caroff et al., 1993). Perhaps the best example of such a mechanism is the Boina volcanic rock suite in the Afar Rift, Ethiopia (Barberi et al., 1975). Other studies have shown, however, that this simple petrogenetic scheme may not apply everywhere. Davies & Macdonald (1987), Macdonald et al. (1987) and Black et al. (1997), for instance, have conclusively shown, mainly on geochemical grounds, that the peralkaline rhyolites cropping out near Naivasha in the Kenyan Rift Valley (the Greater Olkaria Volcanic Complex, GOVC) cannot be the daughter magmas of the associated basalts, and they proposed an origin by partial melting of the crust triggered by alkali-bearing volatiles. Mahood et al. (1990) proposed that pantellerite rhyolites at the type locality of Pantelleria are the product of a two-step process, whereby partial melting of alkali gabbros yields trachyte melts whose subsequent fractionation ends in the strongly peralkaline rhyolites [but see Civetta et al. (1998)].

All studies up to 1988 agreed as to the dry, or at least the very H₂O-poor, character of peralkaline rhyolites, in particular when compared with their metaluminous equivalents. The reasons for believing so (Bailey & Macdonald, 1987) include the nearly anhydrous state of all peralkaline obsidians, their high Cl content, thought to be incompatible with the exsolution of a hydrous vapour phase, and the fact that phase equilibrium experiments performed under dry conditions better reproduce the crystallization sequence in the rocks than those under wet conditions which seriously contrast with the rock record (Bailey, 1974). Another rarely considered factor is that peralkalinity is accompanied by a significant fall in liquidus temperatures relative to metaluminous rhyolites. This was apparent in the pioneering studies of Schairer & Bowen (1955, 1956), who worked out the phase relationships in the dry synthetic Na₂O–Al₂O₃–SiO₂ and K₂O–Al₂O₃–SiO₂ ternaries and found eutectic temperatures as low as 740°C at 1 bar on the peralkaline sides. Addition of iron to the Na₂O–Al₂O₃–SiO₂ system does not change the trend of falling temperatures (Bailey & Schairer, 1966), with even lower eutectic temperatures, around 728°C, in the SiO₂-oversaturated field of the Na₂O–Al₂O₃–SiO₂–Fe₂O₃ system. The implication is that there are no thermal limitations to generating dry silicic peralkaline melts by direct melting of crust under dry conditions (given a suitable protolith), a conclusion in sharp contrast to metaluminous or peraluminous systems where dry melting requires temperatures of 1000°C (Bowen, 1937), i.e. temperatures unlikely to be reached in the crust, even in extensional tectonic settings.

Melt inclusion studies of peralkaline rhyolites, however, have found significant amounts of dissolved H₂O (1·4–5 wt %, Ascension: Harris, 1986; Pantelleria: Kovalenko et al., 1988; Lowenstern & Mahood, 1991; Fantale, Ethiopia: Webster et al., 1993; GOVC, Kenya: Wilding et al., 1993; Mayor Island, New Zealand: Barclay et al., 1996). Although this conflicts with previously available phase equilibrium constraints (Bailey, 1974), it is worth stressing that our knowledge of the physico-chemical conditions of the evolution of peralkaline silicic magmas is still surprisingly poor. For instance, because of the scarcity of coexisting Fe–Ti oxide minerals in peralkaline rocks, pre-eruptive temperatures are poorly constrained. The few quantitative estimates available (Carmichael, 1967, 1991; Bizouard et al., 1980; Mahood, 1981; Wolff & Wright, 1981; Conrad, 1984; Novak & Mahood, 1986; Crisp & Spera, 1987; Jørgensen, 1987; Ghiorso & Sack 1991) indicate pre-eruptive temperatures in the range 700–950°C at fO₂ generally close to the fayalite–magnetite–quartz buffer (FMQ).

Available experimental work is of little use in constraining the pre-eruptive conditions of peralkaline magmas (Webster et al., 1993). Most studies have been performed in synthetic systems under either dry (Bailey & Schairer, 1966) or water-saturated conditions (Carmichael & MacKenzie, 1963; Thompson & MacKenzie, 1967). The same problem applies to the few experimental studies on natural peralkaline rhyolites (Bailey & Cooper, 1978). In addition, in none of these studies was the fO₂ carefully controlled. Experiments at H₂O saturation have all been performed under very oxidizing conditions (Carmichael, 1962; Nicholls & Carmichael, 1969; Sutherland, 1974). Thus, the phase equilibria of peralkaline magmas are poorly known compared with their metaluminous or peraluminous silicic counterparts (e.g. Clemens & Wall, 1981; Clemens et al., 1986; Webster et al., 1987; Conrad et al., 1988; Scaillet et al., 1995; Martel et al., 1998, 1999; Dall’Agnol et al., 1999; Scaillet & Evans, 1999), which seriously hinders any attempt at constraining their evolution. This work intends to fill this gap by determining the phase relationships of three natural peralkaline rhyolites under controlled fO₂ and fH₂O conditions. Although our main goal is to provide quantitative estimates of pre-eruptive conditions, the comprehensive phase equilibrium data provided should prove useful in determining the pre-eruptive conditions of other mildly to moderately peralkaline silicic rocks.

	ROCKS STUDIED

TOP ABSTRACT INTRODUCTION ROCKS STUDIED EXPERIMENTAL AND ANALYTICAL... RESULTS DISCUSSION REFERENCES

 
The three natural rhyolites used come from the Greater Olkaria Volcanic Complex in the Naivasha area of the Kenyan Rift Valley, where the geology is well known (Clarke et al., 1990) and where there has been extensive petrological and geochemical work (Bailey & Macdonald, 1987; Davies & Macdonald, 1987; Macdonald et al., 1987; Black et al., 1997; Marshall et al., 1998). The three rocks are phenocryst-poor obsidians; bulk compositions are listed in Table 1. They are comendites in the classification scheme of Macdonald (1974), with agpaitic indices [(Na + K)/Al, in moles] ranging from 1·05 to 1·36. They represent the two extremes of the geochemical range displayed by the comendite suite at Olkaria (Macdonald et al., 1987). Sample ND 002 is only marginally peralkaline and has <5% (volume) phenocrysts of alkali feldspar, clinopyroxene, fayalite, magnetite, zircon, apatite and chevkinite. The glassy matrix has the same agpaitic index as the bulk rock (Table 1). Sample BL 575 is the most peralkaline specimen (NK/A = 1·36). It has 5% phenocrysts of alkali feldspar, quartz, riebeckite–arfvedsonite, aenigmatite, biotite, ilmenite and chevkinite. BL 575 contains some spherulites with a radial arrangement of riebeckite laths and aggregates of quartz and alkali feldspar. Microphenocrysts of riebeckite–arfvedsonite also occur in the matrix glass. Sample SMN 49 is chemically similar to BL 575, being only slightly more SiO₂ rich and less peralkaline (Table 1). Its phenocryst assemblage differs from that of BL 575 in the occurrence of fluorite (Marshall et al., 1998) and the absence of riebeckite and aenigmatite. Also, it has lower Fe₂O₃/FeO than BL 575 (Table 1). In both BL 575 and SMN 49, the matrix glass is slightly more peralkaline than the bulk rock (Table 1). For all three samples, microprobe analyses of the matrix glass yield totals close to 100%, implying water contents <1 wt %. This is confirmed by Fourier transform IR spectroscopy (FTIR) analyses carried out on sample ND 002, whose matrix glass has 1800 ppm dissolved H₂O (Wilding et al., 1993). FTIR analyses performed by Wilding et al. (1993) on samples belonging to the same suite as BL 575 and SMN 49 yielded similar results.

View this table: [in this window] [in a new window]   Table 1: Bulk-rock and matrix glass compositions (wt %)

 

	EXPERIMENTAL AND ANALYTICAL TECHNIQUES

TOP ABSTRACT INTRODUCTION ROCKS STUDIED EXPERIMENTAL AND ANALYTICAL... RESULTS DISCUSSION REFERENCES

 
Preparation of charges
Hydrothermal experiments were performed using rock powders as starting material. Therefore, given the near-aphyric character of the natural rhyolites, most experiments in this study were of crystallization type (e.g. Clemens & Wall, 1981; Pichavant, 1987; Conrad et al., 1988; Scaillet et al., 1995; Martel et al., 1998; Dall’Agnol et al., 1999; Scaillet & Evans, 1999). Charges consisted of Au capsules (length 1·5–1·2 cm, thickness 0·02 cm, inner diameter 0·29–0·27 cm), first loaded with deionized water, then with silver oxalate as a source of CO₂ for H₂O-undersaturated experiments, and finally with the rock powder. Fluids of various H₂O/CO₂ ratio were used as a practical means of varying the water activity (aH₂O) at any given temperature (T) and pressure (P) but also because geological evidence points to a role for CO₂ in the petrogenesis of felsic peralkaline magmas in the Kenya rift (Bailey & Macdonald, 1987; Macdonald et al., 1993). The weight proportion of H₂O + CO₂ [i.e. (H₂O + CO₂)/(H₂O + CO₂ + silicate)] in most charges was maintained in the range 7–10 except for charge numbers 1–7, in which this parameter was around four (in the following, XH₂O_in refers to the initial mole fraction of H₂O in the fluid phase loaded to the capsule). A single series of experiments was also performed with only H₂O added in amounts below those required for saturation. Nominally dry charges were also prepared simply by loading the rock powder into Au capsules. These charges contain in fact 0·2–0·4 wt % water, owing to the presence of residual H₂O in the matrix glass and adsorbed water. Upon loading, capsules were sealed by arc-welding, immersed in a 120°C oil bath, and then weighed to check for leaks. After completion of an experiment, capsules were reweighed to ensure closed-system behaviour at P and T.

Equipment
All experiments were performed in standard cold-seal pressure vessels with or without semi-permeable H₂ membranes (respectively designed CSPV-H₂ and CSPV). Total pressure is known to within 2 MPa. Temperatures were continuously recorded by external unsheathed type-K thermocouples and are known to ±8°C. Experiments performed with an H₂ membrane were terminated by removing the vessel from the furnace and allowing it to cool in air. To minimize exsolution, most experiments (charges with numbers higher than 12) were terminated using a small ventilator to accelerate cooling and by maintaining isobaric conditions down to 150°C. Experiments performed without H₂ membranes at 150 MPa were drop quenched. For each composition, phase equilibria were established at two pressures, 50 and 150 MPa, and under two contrasting redox conditions, in the temperature range 660–800°C. Run durations varied according to temperature, between 7 days at around 800°C and 51 days at or below 700°C. A total of 158 phase equilibrium experiments was performed.

fH₂ control
Two types of vessels were used to perform either reduced (CSPV-H₂) or oxidized (CSPV) experiments. In the former, the hydrogen fugacity (fH₂) was read continuously via H₂ membranes connected to a transducer with an uncertainty of 0·01 MPa (Scaillet et al., 1992; Schmidt et al., 1995; Scaillet & Evans, 1999). Under oxidizing conditions the fH₂ of the vessel was determined by using the dependence of the ferric/ferrous ratio of the three bulk compositions on fH₂. This relationship was calibrated by annealing the rock powders at 150 MPa, 800°C, H₂O saturation and various fH₂, as monitored with H₂ membranes. Volumetric titration of the ferrous iron content of the run products yields the relationship between FeO content and fH₂ specific to each composition. Subsequent determination of the ferrous content of H₂O-saturated charges around 800°C thus gives the fH₂ of that run. The fH₂ varies among the vessels, between 0·011 and 0·004 MPa; these values correspond to fO₂ 3·6–4·4 log units above that of the Ni–NiO solid buffer (NNO + 3·6 to NNO + 4·4) at H₂O saturation. The fH₂ obtained at that temperature for a given vessel was considered constant for that vessel and used for subsequent fO₂ calculations at lower temperatures. The fO₂ of two vessels used without an H₂ membrane was also measured using solid NiPd sensors (Taylor et al., 1992; Pownceby & O’Neill, 1994). Values of NNO + 3 to NNO + 5·1 were obtained, straddling the haematite–magnetite equilibrium.

fO₂ determination
The fO₂ of fluid-saturated charges was calculated using the equation (see Scaillet & Evans, 1999)

For experiments performed at high fH₂, the vast majority of fO₂ values fall in the range NNO - 1 to NNO - 2, with an average at NNO - 1·67 ± 0·4. For experiments performed at low fH₂, most of the calculated fO₂ are in the range NNO + 3 to NNO + 4, with an average at NNO + 3·63 ± 0·44. Thus, for simplicity, average values of NNO - 1·6 and NNO + 3·6 are used in the following to characterize runs performed under reduced and oxidized conditions, respectively. At XH₂O values <0·3, however, the fO₂ starts to decrease significantly relative to that at H₂O saturation, by several orders of magnitude for extremely H₂O-poor charges (see Webster et al., 1987). For nominally dry charges the amount of H₂O present in the fluid phase, if any, depends on the amount of both adsorbed and dissolved H₂O of the rock powder, which can vary between charges and cannot be evaluated rigorously. For instance, the fO₂ of nominally dry charges calculated using equation (1) assuming a fluid phase composition of XH₂O = 0·01 is around NNO for dry charges at low fH₂, or 5 log units below it (NNO - 5) for dry charges at fH₂ = 20–30 bar. The same calculation with XH₂O = 0·1 yields fO₂ of NNO + 2 to NNO - 3, respectively. This shows that, if extremely dry, runs performed at low fH₂ may have fO₂ values approaching those of H₂O-saturated charges at high fH₂.

For experiments carried out with CSPV-H₂ at H₂O saturation, the calculated fO₂ values are accurate to within 0·03 log units whereas those calculated for H₂O-undersaturated charges must be seen as maxima, as equation (1) overestimates the abundance of H₂O in the fluid and neglects the departure from ideality of H₂O–CO₂ fluid. For experiments performed in CSPV, the uncertainty of fO₂ is larger. In runs for which fH₂ has been determined (runs at 800°C), we evaluate the uncertainty to be around ±0·6 unit log. For runs performed in CSPV at lower temperatures, the error on our calculated fO₂ can be much larger, possibly approaching 1 log unit. We consider, however, that all the runs performed in CSPV have their fO₂ within the range NNO + 3 to NNO + 5 (excluding dry runs).

Analytical methods
Run products were characterized by polarized light microscopy, and by electron microprobe analysis (EMPA) and scanning electron microscopy (SEM). Analytical conditions for EMPA were accelerating voltage 15 kV, sample current 6 nA, counting time 10 s on peak for all elements and a focused beam for minerals. For glasses the beam was defocused to 5 µm. Na and K were analysed first and a ZAF correction procedure was applied. Na₂O migration under the probe beam was assessed using two sets of hydrous metaluminous and peralkaline (SMN 49) rhyolitic glass standards of known H₂O contents (Scaillet & Evans, 1999). At the beginning of each analytical session the two set of standards were analysed, the metaluminous being used for retrieving the Na₂O correction factor to be applied to composition ND 002, whereas the peralkaline set was used for calculating that specific to compositions BL 575 and SMN 49.

Melt H₂O contents (H₂O_melt) of drop-quenched glasses at 800 and 750°C were estimated using the by-difference method (Devine et al., 1995) after correction for alkali migration. The precision of melt water contents is estimated to be 0·5 wt %. Experiments performed at lower temperatures display textural evidence of fluid exsolution during quench (sub-micrometric bubbles). The melt water content of these charges was estimated using the linear correlations between XH₂O_in and melt H₂O content obtained in drop-quenched charges at 800 and 750°C. These correlations were also used to calculate the melt water content of near-solidus charges in which the melt composition could not be microprobe analysed. For nominally anhydrous charges the melt H₂O content was calculated assuming perfectly incompatible behaviour for both H₂O and F during crystallization and using a bulk H₂O content for the three starting rock powders of 0·4 wt %.

Attainment of equilibrium
The successful crystallization of all the naturally occurring minerals apart from aenigmatite shows that no major nucleation problems were encountered in this study. At 800°C and 150 MPa, melting of the three H₂O-saturated samples was complete after 10 days. A few relict minerals were observed in some charges, mainly at temperatures below 750°C and at low XH₂O (Table 2), but their modal abundance is <0·5%, implying that the non-reactive part of the system was minimal. Experimental work performed in dry and H₂O-saturated peralkaline systems has shown that equilibrium conditions are attained during run durations of the order of a week (Carmichael & Mackenzie, 1963; Thompson & Mackenzie, 1967; Bailey et al., 1974; Bailey & Cooper, 1978). In the present work, run duration was at least 10 days at 800°C and up to 8 weeks at temperatures below 700°C (Table 2). Previous researchers carrying out phase equilibrium experiments in metaluminous or peraluminous silicic systems have concluded that bulk equilibrium was reached for run durations similar to those applied in the present study (e.g. Clemens & Wall, 1981; Pichavant, 1987; Webster et al., 1987; Holtz et al., 1992; Scaillet et al., 1995; Dall’Agnoll et al., 1999). These lines of evidence argue for the attainment of near-equilibrium conditions in the experimental charges. Probe analyses show that in some runs, contamination by Ni from the vessel occurred, with the Ni content of ferromagnesian phases of a single charge varying in an erratic way from below detection up to several weight percent, suggesting that Ni was not instrumental in crystal growth. The FeO_tot content of glasses of supra-liquidus charges at low fO₂ is 6% lower than the starting bulk composition, showing that iron loss toward the Au capsules was minimal.

View this table: [in this window] [in a new window]   Table 2: Experimental results on Naivasha rhyolites

 

	RESULTS

TOP ABSTRACT INTRODUCTION ROCKS STUDIED EXPERIMENTAL AND ANALYTICAL... RESULTS DISCUSSION REFERENCES

 
Phases and textures
Phases grown in our experiments are alkali feldspar, quartz, plagioclase, hedenbergite, magnetite, ilmenite, chevkinite, biotite, fayalite, zircon, fluorite, titanite, aegirine, ferrorichterite and riebeckite–arfvedsonite. Montdorite, a tetrasilicic mica (Robert & Maury, 1979), was identified in one charge from SMN 49 (Table 2). The only phase in the natural samples not identified in this study is aenigmatite.

Minerals display euhedral shapes with sizes of 5–30 µm, except for the Na-amphibole and aegirine, which crystallize as stubby prisms over 100 µm long, and for oxides whose size rarely exceeds 5 µm. Close to the liquidus, crystallization is marked by the appearance of a few large individuals (Fig. 1a), whereas at lower temperatures, nucleation is more efficient than growth, giving rise to a large number of smaller crystals (Fig. 1b). Although relict phases are very rare (<0·5% by volume of the crystal), they are clearly distinguished on the basis of their larger size (>100 µm), their rounded shape (alkali feldspar, quartz), dissolution textures at their borders (alkali feldspar), a reaction zone surrounding the crystal (hedenbergite) or the presence of a rim compositionally different from the core. In the last texture, the rim in all cases has the same composition as small euhedral crystals that grew during the experiment. Rounded relicts of tectosilicates are observed only in cases where the charge is outside, yet close to, the stability field of the corresponding phase. All charges, including those nominally anhydrous, display evidence of fluid saturation in the form of evenly dispersed bubbles with sizes varying between 10 and 100 µm, some containing cubic minerals. This population of equilibrium bubbles is clearly distinguishable from the micrometric to sub-micrometric bubbles present in charges that experienced slow quench rates (Fig. 1b). The reason for nominally anhydrous melts to be fluid saturated may be the presence of Cl and F, elements that promote early fluid saturation compared with the purely hydrous condition (e.g. Webster, 1992, 1997).

View larger version (127K): [in this window] [in a new window]   Fig. 1. Scanning electron micrographs of run products. In both pictures quartz has a dark grey to black shade compared with alkali feldspar. (a) Coexisting quartz and alkali feldspar crystals and glass in a near-liquidus charge of BL 575. Bright minerals are oxides. (b) Quartz, alkali feldspar and riebeckite–arfvedsonite (bright equant grains) and glass assemblage. (Note the presence of micron to sub-micron sized bubbles scattered throughout the charge, interpreted to occur during the quench.) The 10 µm large void in the upper left corner represents a former equilibrium fluid bubble.

 
Phase relationships
Phase relationships for the three compositions are shown in Figs 2–4. At 150 MPa, phase relations are shown in T–melt H₂O content projections. This type of projection could not be established at 50 MPa and only the T–XH₂O_in projection is shown. No attempt was made to establish the stability fields of zircon or apatite. The small size prevented in some cases clear identification of the type of Fe–Ti oxide. As a rule, magnetite grew in ND 002 whereas ilmenite or a Ti-bearing haematite crystallized in both BL 575 and SMN 49. Ilmenite is present, however, in some run products of ND 002, especially those whose fO₂ is significantly below NNO - 1. The nominally anhydrous runs performed under high and low fH₂ are also plotted. Despite the fact that their fO₂ may have been significantly below the fO₂ of H₂O–CO₂-bearing charges held under similar fH₂ conditions, both sets of charges display consistent phase assemblages, except for charge 54 in composition SMN 49 whose results conflict with the remainder. This point is discussed further below.

View larger version (20K): [in this window] [in a new window]   Fig. 2. Phase relationships of ND 002. Mineral abbreviations are as recommended by Kretz (1983). The curves are labelled with minerals lying inside their respective stability fields, and are shown dashed when inferred. (a) 150 MPa and NNO - 1·6; (b) 150 MPa and NNO + 3·6; (c) 50 MPa and NNO - 1·6. Solidus HPG is the solidus of the subaluminous haplogranite system at 150 MPa (Johannes & Holtz, 1996).

 

View larger version (19K): [in this window] [in a new window]   Fig. 4. Phase relationships of SMN 49. Mineral abbreviations are as recommended by Kretz (1983). The curves are labelled with minerals lying inside their respective stability fields, and are shown dashed when inferred. (a) 150 MPa and NNO - 1·6; (b) 150 MPa and NNO + 3·6; (c) 50 MPa and NNO - 1·6.

 

ND 002
At 150 MPa under reduced conditions (NNO - 1·6), magnetite and ferrohedenbergite are the liquidus phases, closely followed by alkali feldspar, quartz and then fayalite (Fig. 2a). At H₂O saturation, the crystallization temperatures of oxide and ferrohedenbergite are estimated to be close to 760°C, whereas alkali feldspar precipitates near 700°C, followed by quartz near 660°C. At melt H₂O content lower than 2–3 wt % and at temperatures higher than 700°C, hedenbergite is no longer stable and the main stable ferromagnesian silicate is fayalite, coexisting with ilmenite. At low H₂O_melt (<2 wt %) and temperatures above 730°C, a ferro-richteritic amphibole is present. Below 700°C amphibole is present only in a subsolidus charge (charge 31, Table 2), and the amphibole-out curve has a rather steep slope. In the dry charge at 729°C (charge 44, Table 2), a former hedenbergite phenocryst was found rimmed by fayalite, ilmenite and fluorite, suggesting that these phases replace pyroxene at low H₂O_melt. Finally, the scarce occurrence or absence of hedenbergite in charges at temperatures below 700°C suggests that its stability field may not extend down to the solidus in ND 002 in the water-rich part of the diagram. Thus, we have drawn a hedenbergite-out curve that is crossed either as temperature falls or as melt H₂O decreases. Biotite was identified only below 700°C at H₂O saturation, or in dry charges at higher temperatures, which suggests a restricted stability for fO₂ at or below NNO - 1·6, close to the solidus. Similarly, fluorite is present only in strongly crystallized charges or below 700°C at H₂O saturation and its stability field lies close to the solidus. Although not looked for systematically, chevkinite was not found. No attempt was made to locate the solidus precisely, but it is below 660°C at H₂O saturation (Table 2), which is 40°C lower than the solidus of the haplogranite system at 150 MPa (Johannes & Holtz, 1996). Melt H₂O contents at saturation are slightly lower than 5 wt %, being similar to those found in the metaluminous haplogranite system (Johannes & Holtz, 1996) at similar P and T.

Increased fO₂ does not alter the order of crystallization of the major phases (Fig. 2b) but it yields, at H₂O saturation, magnetite crystallization above 800°C, whereas clinopyroxene saturation occurs slightly below 750°C. Although the liquidus of the alkali feldspar at H₂O saturation may be higher by 10°C, over most of the investigated domain and within the error in the melt H₂O determination, the crystallization temperatures of the tectosilicates remain unchanged relative to that at low fO₂. Biotite, fayalite and amphibole were not identified. In contrast, titanite is present in most of the 750°C charges and possibly in all the 700°C charges, and the titanite-in curve is drawn flat at temperatures close to 750°C. As at low fO₂, fluorite is present in some crystal-rich charges close to the solidus. Although data are lacking, its stability domain extends to higher temperatures, as seen under low fO₂. Chevkinite was identified in only one charge at H₂O saturation and 752°C, and its stability field is shown dashed with a shape similar to that seen in BL 575 and SMN 49 at high fO₂. The solidus seems to be at significantly higher temperature, by 30°C, than under reduced conditions. This is apparent when comparison is made between charge 33, which contains at least 10% residual glass by volume as estimated from scanning electron micrographs, and charge 27, which has no clear textural evidence of residual glass, despite the fact that both charges have similar XH₂O_in values and were run at almost identical temperatures. Finally, plagioclase was identified (by EMPA) in charge 15 run at 790°C under relatively dry conditions. This is the only instance of plagioclase crystallization in this study. Charges bracketing charge 15 do not contain plagioclase and it does not appear at lower temperature, implying a closing of the plagioclase stability field, if any, at low temperature.

Decreasing the pressure to 50 MPa under low fO₂ does not affect the relative order of crystallization, although fayalite has not been identified (Fig. 2c). The liquidus temperatures of all phases at H₂O saturation are significantly raised, by 60°C, for the tectosilicates. Alkali feldspar-in and quartz-in curves are more clearly separated from each other, suggesting that these two phases are not equally sensitive to changes in pressure. Biotite and Na-amphibole have been identified only in some subsolidus charges at 690°C (Table 2), but their stability in the presence of liquid cannot be ruled out. The solidus at H₂O saturation is probably close to 700°C, that is, 60°C lower than that of the haplogranite system at 50 MPa. Melt H₂O contents at saturation range from 2·6 to 3·7 wt %, although the lower figure is probably correct (Table 2).

BL 575
This composition is a typical comendite with higher FeO_total and Na₂O contents than ND 002. At 150 MPa under reduced conditions, an oxide is the liquidus phase, crystallizing at 750°C and H₂O saturation (Fig. 3a). It is followed by alkali feldspar and quartz, crystallizing simultaneously over the T–H₂O_melt range explored. Their temperature of appearance at H₂O saturation is slightly higher than 650°C. Fluorite was found mainly in low-temperature charges, where it usually crystallizes at lower temperature than amphibole, except at H₂O saturation. Under dry conditions, the maximum thermal stability of fluorite does not exceed 775°C. At H₂O_melt > 3 wt %, Na-amphibole crystallization is restricted to temperatures below 700°C. Its crystallization at H₂O saturation occurs slightly above 660°C. Under nominally dry conditions its thermal stability extends to near 740°C. Biotite is present only at the lowest temperature investigated and at H₂O saturation, where it is a near-liquidus phase together with amphibole and fluorite. Its liquidus curve displays a positive slope in the T–H₂O_melt projection, but the exact shape of this boundary is not yet well constrained. The chevkinite stability field seems to be restricted to low melt H₂O contents (Table 2), but is also poorly constrained. Melt H₂O contents at saturation are slightly lower than 6 wt %, being 1 wt % higher than those of ND 002. The solidus has not been located but is definitely below 650°C at H₂O saturation, possibly close to, or even below, 600°C.

View larger version (19K): [in this window] [in a new window]   Fig. 3. Phase relationships of BL 575. Mineral abbreviations are as recommended by Kretz (1983). The curves are labelled with minerals lying inside their respective stability fields, and are shown dashed when inferred. (a) 150 MPa and NNO - 1·6; (b) 150 MPa and NNO + 3·6; (c) 50 MPa and NNO - 1·6.

 
Increasing fO₂ by 5 log units increases the thermal stability of Fe–Ti oxide (Fig. 3b). Below 750°C, quartz and alkali feldspar appear to be insensitive to variations in fO₂. Fluorite crystallizes at 680°C at water saturation, and its liquidus roughly parallels that of the tectosilicates over the temperature range investigated. In contrast to low fO₂, its thermal stability seems largely to exceed 800°C. Chevkinite may have a slightly wider stability field relative to low-fO₂ conditions. Under H₂O saturation it is the second phase to precipitate after oxide at 690°C. Its thermal stability when H₂O_melt decreases is again not well constrained. Both Na-amphibole and biotite are absent at high fO₂. The former is replaced by aegirine whose thermal stability also increases as H₂O_melt decreases. Solidus conditions were approached at 678°C in charge 40 (XH₂O_in = 0·46); no glass is seen on SEM images. Comparison with results at low fO₂ (charge 35 at 676°C and XH₂O_in 0·51 has >20% glass in volume) again suggests that a rise in fO₂ produces a rise in solidus temperatures, as in ND 002. Melt H₂O contents required for quartz and alkali feldspar saturation at any given temperature are identical to those at low fO₂.

Decreasing the pressure to 50 MPa raises the crystallization temperatures by 50–60°C (Fig. 3c), as for ND 002. Quartz appears before alkali feldspar but the less extensive coverage of T–H₂O_melt data prevents accurate location of the phase boundaries. Melt H₂O contents at saturation are in the range 3·2–3·5 wt % (Table 2). At H₂O saturation, the Na-amphibole stability field is higher by 30°C compared with that at 150 MPa. Biotite was not identified; if stable, it must crystallize below 680°C, as tentatively suggested in Fig. 3c.

SMN 49
This composition is close to BL 575, differing only in a slightly higher SiO₂ content (Table 1). The phase relations of SMN 49 are thus essentially the same as those of BL 575 (Fig. 4), except that at low fO₂ no stability field for Fe–Ti oxides could be established. The liquidus curves of quartz and alkali feldspar are also more clearly separated from each other. Montdorite was recognized (by EMPA) in only one low-fO₂ and near-solidus charge (33, Table 2).

At high fO₂ under H₂O-saturated conditions, Fe–Ti oxide crystallizes at slightly over 730°C, followed by chevkinite at 720°C. The thermal stability of chevkinite is significantly larger than that in BL 575. The dry charge at 731°C (54) yielded Na-amphibole and no pyroxene (Table 2). This difference is interpreted as due to a significantly lower fO₂ relative to the H₂O–CO₂-bearing charges, possibly resulting from less adsorbed water in the powder loaded to charge 54. This charge has been ignored in establishing the phase relations at high fO₂. As for BL 575, a rise in fO₂ produces a marked rise in the solidus temperature, best seen when comparing charges 38 and 46, which have the same T–H₂O_melt location but were run under different fO₂ conditions (Table 2). Whereas charge 38 (high fO₂) shows a minimal amount of residual glass in scanning electron micrographs, charge 46 (low fO₂) displays a much larger fraction of residual glass. A further feature of interest is that the crystallization sequence produced at 730°C with the H₂O-undersaturated but CO₂-free charges (charges 50–53) is virtually identical, both in terms of phase assemblages and T–melt H₂O coordinates, to that obtained in CO₂-bearing charges (charges 17–21), or within the analytical error of our melt H₂O content determination technique.

At 50 MPa, at H₂O saturation, the crystallization of both quartz and alkali feldspar occurs around 725°C (Fig. 4c), or 60°C higher than at 150 MPa. The fluorite-in curve possibly has the same relative position as at 150 MPa. Melt H₂O contents at saturation are identical to those determined for BL 575 (Table 2). Amphibole stability is slightly expanded towards higher temperature. Biotite is not present under the experimental conditions investigated at 50 MPa. If this mineral crystallizes in SMN 49 it must do so below 680°C.

	DISCUSSION

TOP ABSTRACT INTRODUCTION ROCKS STUDIED EXPERIMENTAL AND ANALYTICAL... RESULTS DISCUSSION REFERENCES

 
Aenigmatite stability
Our phase equilibrium experiments successfully reproduced the phase assemblages of the rocks, apart from aenigmatite. Several factors may have contributed to this. The low modal abundance of aenigmatite in the rock suggests that this phase may also have crystallized in such low amounts in the experiments that it was overlooked in the run products. However, phases present in equally low amounts were identified, e.g. biotite and fayalite. Previous experimental studies do not mention difficulties in growing aenigmatite under laboratory conditions similar to ours (Ernst, 1962; Thompson & Chisholm, 1969; Lindsley 1971). Rather, an inappropriate choice of one or more of the intensive parameters may have prevented aenigmatite crystallization. The experimental work of Ernst (1962) on a synthetic Na–Fe end-member amphibole has shown that Ti-free aenigmatite is unstable at pressures >90 MPa. Thompson & Chisholm (1969) synthesized Ti-bearing stoichiometric aenigmatite at 100 MPa, which suggests that Ti enhances the pressure stability of this phase. Although there are currently no data on aenigmatite stability at pressures >100 MPa, its stability may be pressure dependent, possibly explaining its absence in our experiments. On the other hand, Ernst (1962) and Lindsley (1971) showed that the occurrence of aenigmatite also depends on redox conditions, being enhanced at low fO₂. Ti-free aenigmatite is unstable at fO₂ above that of the wüstite–magnetite solid buffer whereas the Ti-bearing variety is stable to an fO₂ between NNO and FMQ. Although our lowest fO₂ is in the correct range for stoichiometric Ti-bearing aenigmatite to crystallize, a small departure of the natural aenigmatite composition from the theoretical one of Lindsley (1971) may affect the fO₂ stability of that phase in the rock. It is therefore possible that our imposed fO₂ may not have been sufficiently low to stabilize aenigmatite, even at the lowest pressure investigated here.

Comparison between ND 002, BL 575 and SMN 49
At 150 MPa, liquidus temperatures are higher in ND 002, consistent with the fact that it represents the least differentiated member of the GOVC rhyolite suite. However, comparing only the tectosilicates, which are >90% of the crystallizing assemblage, the liquidus temperature of ND 002 is only 30°C higher than those of BL 575 and SMN 49. The phase relations established at 50 MPa suggest that the positions of tectosilicate liquidi are more sensitive to composition at decreasing pressures, this being most apparent for the liquidus temperature of alkali feldspar in ND 002, which is significantly higher than in the two more peralkaline compositions. As regards crystallization sequence, the liquidus tectosilicate in ND 002 is alkali feldspar, cotectic quartz–alkali feldspar crystallization is observed in BL 575, and quartz crystallizes before alkali feldspar in SMN 49. Although this may partly be due to anhydrous compositional variations, the influence of other volatiles is possible. The fact that CO₂-free and CO₂-bearing charges yield essentially the same results in SMN 49 shows that this volatile has a limited solubility in peralkaline silicic melts and that it mainly acts as a diluting agent in the fluid phase, as documented for other silicic magmas at low pressures (e.g. Blank et al., 1993). Differences in CO₂ solubility cannot therefore explain the observed differences in liquidus minerals. In contrast, fluorine expands the primary volume of quartz in the haplogranite system (Manning, 1981). The combined effects of the slightly more silicic composition of SMN 49 compared with BL 575 and the fluorine complexing effect in alumino-silicate melts may explain why the quartz-in curve is more clearly separated from that of alkali feldspar in SMN 49 than in BL 575.

Montdorite has been identified in SMN 49 only but may be present in BL 575 as well. The stability domain of this phase needs refinement but our data agree with published occurrences of montdorite in magmatic rocks (e.g. Robert & Maury, 1979). Montdorite was first described in a comendite sill (Robert & Maury, 1979), where it occurs in the groundmass along with manganoan arfvedsonite, quartz, anorthoclase, oxides and fluorite among other accessory phases. T–fO₂ conditions of crystallization obtained from two-oxide geothermometry were 760°C and NNO, pointing to low temperature and moderately reduced fO₂ during montdorite crystallization. This estimate, together with the textural evidence and our experimental data, all suggest that montdorite occurs only late in the crystallization sequence of comendite magmas. On the basis of the relationships in SMN 49, we suggest that montdorite may replace biotite as the stable ferromagnesian potassic phase, when the residual liquid becomes strongly peralkaline.

It is well known that substitution of fluorine for hydroxyl groups in hydrous minerals such as amphibole can substantially raise their thermal stability, as shown, for instance, by Holloway & Ford (1975) for pargasite. The Na-amphibole produced in our work has a F content of 2 wt % at H₂O saturation. Despite this, the thermal stability of this phase in comendite magma at H₂O saturation does not exceed 670°C, which is lower than the upper stability limit of pure OH-riebeckite in H₂O fluid (Ernst, 1962).

The fluxing effect of the volatiles is well illustrated using the phase relationships obtained on ND 002, which most closely approached solidus conditions in this study. In Fig. 2, the solidus of the haplogranite system at 150 MPa at H₂O saturation (Johannes & Holtz, 1996) is shown in the ND 002 phase diagram obtained at 150 MPa. The difference between the two solidi is at least 40°C. The reason for such a low solidus temperature in ND 002 is the presence of additional components relative to the metaluminous base composition of the haplogranite system, most notably Fe, F and Cl. Previous experimental studies have shown that these components depress solidus temperatures to some extent in granitic systems (Wyllie & Tuttle, 1961; Abbott, 1981; Manning, 1981; Pichavant et al., 1987; Webster et al., 1987; Weidner & Martin, 1987).

Comparison with previous work
Only a few experimental studies have been performed on natural peralkaline silicic rocks, notably by Bailey and co-workers, who determined the liquidus relations, under both hydrous and anhydrous conditions, of several pantelleritic obsidians (Bailey & Cooper, 1978). The redox conditions of their H₂O-saturated experiments were imposed by the vessel, and were presumably relatively oxidizing, above NNO + 1. The more peralkaline nature of their starting product, in combination with our three starting materials, allows us to establish the effect of peralkalinity on the near-liquidus phase relations of peralkaline magmas at H₂O saturation and high fO₂. In pantellerites at 150 MPa, quartz crystallizes at 650°C, whereas alkali feldspar is the liquidus tectosilicate appearing at 690°C. Comparison with BL 575 and SMN 49 at high fO₂ (Figs 3b and 4b) shows that the alkali feldspar stability field is slightly expanded in pantellerites relative to comendites, whereas quartz preserves its position. In ND 002, the least peralkaline composition, quartz and alkali feldspar have liquidus temperatures equal to, or only marginally higher than, those in the other comendites and pantellerites. In pantellerites, the liquidus phase at high fO₂ is an acmitic pyroxene. It crystallizes at 760°C or 100°C higher than in comendites at H₂O saturation (Figs 3 and 4).

Bailey & Cooper (1978) did not perform H₂O-saturated and low-fO₂ experiments, which prevents direct comparison with our results obtained at low fO₂. Their low-fO₂ experiments (thought to be around FMQ) were nominally anhydrous. However, a qualitative agreement in the phase assemblage is observed between the two datasets. In pantellerites, aegirine is absent (but a sodic hedenbergite is reported) whereas a sodic amphibole crystallizes at 750°C, i.e. 10–20°C higher than in the comendites under the driest conditions investigated in our study. Thus, it appears that the increase of agpaicity in peralkaline silicic magmas does not produce any appreciable effect on tectosilicate liquidus temperatures. In contrast, mineral phases typifying the peralkaline condition, such as aegirine and Na-amphibole, have their stability fields expanded when peralkalinity increases, by 100°C from comendites to pantellerites in the case of aegirine.

Temperatures of the liquidus minima in synthetic peralkaline acid systems at 100 MPa (Carmichael & Mackenzie, 1963; Thompson & Mackenzie, 1967), both iron free and iron bearing, are in the range 715–690°C (Carmichael & Mackenzie, 1963), slightly higher than those obtained in the present study. At 100 MPa and H₂O saturation, composition ND 002 falls on the quartz–alkali feldspar cotectic at 700°C, whereas BL 575 and SMN 49 do so at 680°C (the values at 100 MPa for our three compositions have been obtained by linear interpolation of the data obtained at 50 and 150 MPa). Minimum cotectic temperatures in the haplogranite system are higher by 20–30°C, i.e. 720°C at 100 MPa (Tuttle & Bowen, 1958). The tight grouping of quartz–feldspar cotectic temperatures (in the range 680–720°C) in synthetic and natural systems encompassing a large range of agpaitic indices shows that peralkalinity does not induce a marked depression in liquidus temperatures of tectosilicates in H₂O-saturated silicic magmas, which contrasts with the situation under dry conditions (Bailey & Schairer, 1966). The fact that the cotectic temperatures are slighty lower in this work than in the related synthetic systems may rather be attributed to the role of other volatiles (F, Cl). The addition of 1 wt % F to the H₂O-saturated haplogranite system at 100 MPa produces a fall in the temperature of the minimum of 35°C (Pichavant & Manning, 1984). Such an effect is similar in magnitude to that observed in the present study for the two comendites with 1 wt % F, whose cotectic temperature is lower by 20°C than their synthetic F-free counterparts.

Plagioclase stability in silicic magmas
Phase relationships have been established for both metaluminous (Webster et al., 1987) and peraluminous (Bénard et al., 1985; Pichavant et al., 1987; London et al., 1989) silicic rock compositions. Considering first the metaluminous rocks, Webster et al. (1987) worked on an F-enriched vitrophyre (1·2 wt % F). Apart from its higher Al₂O₃ and lower FeO_tot contents, the vitrophyre composition is similar to that of ND 002. At 100 MPa under H₂O saturation, both compositions crystallize the tectosilicates at 700°C. However, in the vitrophyre, two feldspars crystallize, unlike the case in ND 002 (Fig. 2). This is despite the fact that both rock compositions have almost identical bulk CaO content (0·40 wt % for the vitrophyre), and the vitrophyre is slightly more potassic than ND 002 (5·1 and 4·7 wt %, respectively). This suggests that even a slight deficit in Al₂O₃, such as in ND 002, inhibits plagioclase crystallization at the expense of alkali feldspar and quartz in silicic magmas. This is in accord with general petrographic evidence; in a compilation of 104 analyses of comendites and pantellerites, only one rock (from Oraefajokull, Iceland) clearly contained plagioclase phenocrysts (Macdonald & Bailey, 1973). At all water contents quartz is the liquidus phase in the vitrophyre, closely followed by sanidine and then plagioclase, the reverse of that in ND 002 (Fig. 2). This relative crystallization order is opposite to that predicted from the normative quartz content of the bulk rocks (higher in ND 002 than in the vitrophyre). As proposed by Webster et al. (1987), such a crystallization sequence in the vitrophyre probably results from fluorine expanding significantly the primary volume of quartz relative to coexisting feldspars in the haplogranite system (Manning, 1981). With respect to the minor phases, the vitrophyre lacks clinopyroxene and oxides but biotite displays a larger stability field than in ND 002. Both topaz and fluorite crystallize in the vitrophyre, yet neither phase approaches liquidus conditions (<50–75°C below), unlike fluorite in the two comendites studied here. Overall, it appears that plagioclase crystallization in Ca-poor silicic melts is not suppressed by the addition of these levels of fluorine, as it has crystallized in the F-rich vitrophyre and not in the F-poorer ND 002 sample.

For F-poor peraluminous silicic magmas, comparison of our results with those obtained by Bénard et al. (1985) shows that, in peraluminous melts, quartz and alkali feldspar crystallize at similar temperatures (710°C at H₂O saturation and 100 MPa) to ND 002, our F-poor, least peralkaline composition. Unlike in ND 002, however, plagioclase precipitates at 770°C at 100 MPa in the peraluminous system despite bulk calcium contents on average similar to, or even lower than, those of the rocks used in our study [between 0·38 and 0·54 wt % for the three natural compositions used by Bénard et al. (1985)]. Experimental work on various F, Li, B-enriched peraluminous granites (Pichavant et al., 1987; London et al., 1989) showed also that all three major tectosilicates are stable in these magmas. Again, despite bulk CaO contents of 0·2 wt %, plagioclase is the liquidus tectosilicate, closely followed by quartz and alkali feldspar. All peraluminous silicic rocks lack fluorite, despite bulk F contents in the range 1·2–2·4 wt %. The above observations suggest therefore that lack of plagioclase crystallization in peralkaline magmas is directly linked to the intrinsic peralkaline character of the melt and not to an unusual deficit in CaO.

Role of fO₂ on the phase relations of peralkaline magmas
The phase relations displayed in Figs 2–4 help in understanding the main effects of fO₂ variations in peralkaline silicic magmas. For the three rock compositions, variations in fO₂ do not significantly affect the liquidus curves of quartz and alkali feldspar. The main exception is plagioclase crystallization in ND 002, which apparently occurs under oxidizing conditions only. The alkali feldspar liquidus in ND 002 may also be slightly enhanced under high fO₂. Overall, this indicates that feldspar- and quartz-producing species in the melt are not affected by fO₂. As variations in fO₂ primarily affect the Fe²⁺/Fe³⁺ ratio of silicate melts, this in turn suggests that quartz and alkali feldspar melt components are not directly linked to iron in the melt structure. This contrasts with the experimental observations of Dall’Agnol et al. (1999) on a metaluminous silicic composition. By establishing the phase relationships under two widely different sets of redox conditions (NNO - 1 and NNO + 2·5), they showed that a rise in fO₂ produces a significant increase in the thermal stability of the tectosilicates, the effect being most dramatic for the feldspars, especially plagioclase. This strong dependence of plagioclase stability on fO₂ suggests an intimate association between the feldspar melt components and the iron species in iron-bearing metaluminous silicic melts. Such a relationship appears not to exist in relatively F-rich peralkaline silicic melts.

In contrast to the tectosilicates, all other phases crystallizing in the peralkaline rocks are sensitive to fO₂ to various degrees. In all three compositions, an increase in fO₂ is marked by an increase in the thermal stability of oxides, an effect similar to that seen in almost all Fe-bearing magmas. The accessory phase chevkinite also displays an expansion of its stability field as fO₂ increases (Fig. 4), especially in SMN 49, where its stability rises by 50°C when fO₂ increases. Differences in the stability fields of chevkinite in BL 575 and SMN 49 are possibly related to different trace element contents of the bulk rocks but could be due also to difficulties in distinguishing chevkinite from oxide in BL 575, unlike in SMN 49. Fayalite in ND 002 and biotite in all three rocks apparently are stable only at low fO₂. The Ca-rich pyroxene is present in ND 002 whatever fO₂ prevails but its thermal stability is lowered when fO₂ increases. Titanite is present under oxidizing conditions only, as in metaluminous silicic magmas (e.g. Wones, 1989; Dall’Agnol et al., 1999) and its crystallization may in part explain the diminished stability field for the Ca-pyroxene at high fO₂, as both phases compete for Ca in liquid. One of the most significant changes produced by the increase in fO₂ is the replacement in the low-temperature range of Na-amphibole by aegirine. The fact that both phases display similar stability fields suggests that they are related to each other through a redox reaction (Deer et al., 1978). Given that the amphibole is strongly enriched in F, and that aegirine accommodates no F in its structure, amphibole breakdown must release F. This F release may in part explain the slighty enhanced stability field of fluorite under oxidizing conditions.

The last important feature with respect to fO₂ concerns its influence upon the solidus. Although non-systematic, the data reported here clearly show that peralkaline silicic magmas at low fO₂ can have solidus temperatures significantly lower than at high fO₂. We estimate that the lowering of solidus temperatures as a result of decreasing fO₂ from NNO + 3·6 to NNO - 1·6 is at least 30°C and could be as high as 50°C, the magnitude of the effect increasing with the bulk iron content. Clearly, this relies on the solution properties of iron in silicate melts and their dependence on fO₂. One implication is that crystal fractionation in peralkaline silicic magmas will be enhanced by low fO₂. Another is that a peralkaline protolith will start melting at a lower temperature under low fO₂ rather than high fO₂. This redox melting mechanism has been already proposed by Bailey (1974) for mafic systems.

Role of pressure
Variation in pressure between 50 and 150 MPa in our study does not produce any fundamental change in the phase relationships, apart perhaps for aenigmatite stability. In particular, the relative order of crystallization of alkali feldspar and quartz remains unaltered. Anhydrous pantellerites, on the other hand, exhibit a change in their liquidus phase with pressure, quartz replacing alkali feldspar at pressures >100 MPa (Bailey et al., 1974). In pantellerites, the stability of Na-amphibole extends down to near-atmospheric pressure (Bailey & Cooper, 1978) and is weakly sensitive to pressure. Such behaviour is here confirmed for the riebeckite–arfvedsonite variety, showing furthermore that, in comendite magma, not only is amphibole stable at low pressure but its thermal stability increases with decreasing pressure, albeit slightly in the pressure range investigated. The thermal stability of amphibole at H₂O saturation at 50 MPa, where the melt H₂O content is 3·3 wt % H₂O, is similar to that at 150 MPa for a melt H₂O content of 3 wt %. Thus, it can be expected that at P < 50 MPa, the thermal stability of amphibole may still increase to reach the maximum observed under anhydrous conditions at 150 MPa, 730–740°C. In the least peralkaline composition, amphibole replaces clinopyroxene at low water activity, the opposite behaviour to the amphibole–pyroxene relationship observed in all metaluminous melts so far worked on (e.g. Martel et al., 1999; Scaillet & Evans, 1999). Amphibole stability in peralkaline magmas is thus fundamentally different from that in calc-alkaline, silicic to intermediate magmas, where amphibole crystallization demands melt water contents >4 wt % (e.g. Scaillet & Evans, 1999), and where it is hardly stable at pressures below 50–100 MPa (e.g. Sato et al., 1999). The contrast is probably partly related to the fact that amphiboles in peralkaline magmas are more F rich. Thus, the crystallization of amphibole in peralkaline silicic magmas does not appear to be a reliable indicator of water pressure. Rather, its occurrence in mildly peralkaline rhyolites mainly indicates that both low-fO₂ and low-temperature conditions prevailed in the magma chamber. Biotite does not follow the same pattern, as this phase is not stable at low pressure, at least under near-liquidus conditions. Biotite crystallization in comendite magmas at near-liquidus conditions implies pressures of at least 150 MPa.

Our experimental data help constrain the mineralogical evolution, mostly the behaviour of quartz, alkali feldspar and amphibole, of an ascending peralkaline magma under various T–H₂O_in melt conditions. For this purpose, the stability curves of some critical phases in a P–T projection are shown in Figs 5 and 6. In ND 002 at H₂O saturation (Fig. 5), the appearance curves of all major crystallizing phases (quartz, feldspar, clinopyroxene and magnetite) display negative slopes, which implies that, if ascent takes place at constant temperature under equilibrium conditions, extensive crystallization of the magma should occur. In contrast, if the magma starting conditions are below H₂O saturation, the appearance curves display a positive slope, down to the pressure at which the magma reaches H₂O saturation (Fig. 5). As illustrated for quartz in ND 002, the slope of the liquidus increases progressively as the bulk H₂O content of the magma decreases. A magma starting at 150 MPa, 740°C, with 4 wt % H₂O, will soon dissolve quartz upon rising, but it will do so only until 100 MPa, where it reaches H₂O saturation. Beyond, quartz crystallization will take place again. In contrast, a magma ascending from 150 MPa, 790°C and 2·5 wt % H₂O will intersect the quartz saturation boundary only at very low pressures, around 10 MPa. The alkali feldspar liquidus is reached at slightly higher pressures, at around 25 MPa. A similar line of reasoning shows that compositions BL 575 and SMN 49, both starting at 150 MPa, 760°C and 3 wt % H₂O, will reach the cotectic condition at 25 MPa (Fig. 6). The relatively weak dependence of amphibole thermal stability on pressure implies that this phase will persist upon rising, unlike in calc-alkaline magmas. Instead, amphibole crystallization may take place at very shallow levels where dehydration of the magma occurs, thus raising amphibole thermal stability. This probably explains the common occurrence of amphibole microlites in the groundmass of volcanic rocks. For instance, an H₂O-saturated magma residing at 150 MPa and 700°C, i.e. outside the amphibole stability field, will upon ascent and concomitant degassing cross the amphibole liquidus at 40 MPa and end its ascent with abundant microlites, but no phenocrysts, of amphibole (Fig. 6).

View larger version (16K): [in this window] [in a new window]   Fig. 5. P–T stability curves for quartz and alkali feldspar in ND 002 at H2O saturation. The stability curves of quartz for a magma containing 1, 2·5 and 4 wt % H2O are also shown. The curves are labelled with minerals lying inside their respective stability fields.

 

View larger version (15K): [in this window] [in a new window]   Fig. 6. P–T stability curves for quartz–alkali feldspar and riebeckite–arfvedsonite in BL 575 at H2O saturation. The stability curves of quartz–alkali feldspar for a magma containing 1 and 3 wt % H2O, and for riebeckite–arfvedsonite for 3 wt % H2O, are also shown. The curves are labelled with minerals lying inside their respective stability fields.

 

This general picture is similar overall to that for other types of silicic magmas (Johannes & Holtz, 1996). It indicates that, to preserve their near-liquidus character up to near-surface conditions, peralkaline magmas must be H₂O undersaturated, unless ascent rates are fast enough to prevent chemical reaction. The fact that many peralkaline rocks display dissolution textures, such as rounded or embayed quartz phenocrysts (Sutherland, 1974; Barclay et al., 1996), may potentially be related to the H₂O-undersaturated character of the ascending magma. According to our data, ascent durations of <2 days are implied for a magma stored at 150 MPa, 830°C, with 3 wt % H₂O, if quartz phenocrysts are to be preserved.

Pre-eruptive conditions at Naivasha
First-order constraints on pre-eruptive conditions can be made by direct comparison of the phase equilibria established here with the phase assemblages observed in the rocks. This assumes that, in the latter, bulk crystal–melt equilibrium prevailed shortly before eruption and was not subsequently altered during extrusion. The lack of conspicuous zoning of the major phases and of evidence for xenocrysts suggests that this is the case (Macdonald et al., 1987).

Let us consider first ND 002 (Fig. 2). The absence of titanite indicates a pre-eruptive fO₂ close to or below FMQ. The stable coexistence of ferrohedenbergite and fayalite implies that pre-eruptive conditions were at the overlap in their stability fields. If the magma was at water saturation, then temperature is constrained by the liquidus of alkali feldspar, at 700°C. Significantly lower temperatures are precluded by the hedenbergite-out reaction and by the appearance of quartz in the experiments. At 50 MPa coprecipitation of alkali feldspar, hedenbergite and magnetite (+ fayalite) is achieved at around 780°C at H₂O saturation (–2·5 wt % H₂O). Temperatures lower than 740°C are ruled out by quartz crystallization. On the other hand, lower melt water contents would require still higher temperatures. Melt water contents of melt inclusions in some Olkaria rocks reach 3·4 wt % (Wilding et al., 1993), similar to those obtained here, whatever the pressure.

The presence of Na-amphibole at near-liquidus conditions in the natural BL 575 constrains the pre-eruptive temperature to have been lower than 700°C and, along with the absence of aegirine, suggests that the fO₂ was close to, or below, FMQ. Ignoring aenigmatite, the phase assemblage is restricted to a narrow T–melt water domain, with melt water contents of 5·5 wt % and temperature below the intersection point of the biotite-in and Na-amphibole-in curves, which occurs slightly above 670°C. The estimate of melt water content is significantly higher than the maximum so far recorded in melt inclusions in quartz phenocrysts at Olkaria (Wilding et al., 1993), which might indicate that 150 MPa for the pre-eruptive pressure of the storage zone is too high. A decrease in pressure would, nevertheless, not raise significantly the pre-eruption temperature estimate, as this parameter is constrained by the occurrence of Na-amphibole whose stability is rather independent of pressure. However, a decrease in pressure removes amphibole from the liquidus (Fig. 6) and biotite has not been found at 50 MPa, both facts arguing against a pressure significantly lower than 150 MPa.

Sample SMN 49 lacks both aenigmatite and Na-amphibole as phenocrysts but has biotite, in addition to quartz and alkali feldspar. This implies a very restricted set of pre-eruptive temperatures and melt water contents, basically in the coldest and wettest part of the investigated domain, corresponding to the overlap between the stability fields of biotite, quartz and alkali feldspar but above that of Na-amphibole. At 150 MPa, a pre-eruptive temperature of 660°C and H₂O-saturated conditions are implied. Again, as for BL 575, considering the highest melt H₂O content found in melt inclusions (Wilding et al., 1993), a pressure of 150 MPa would seem to be too high. At 50 MPa, coprecipitation of quartz and alkali feldspar is reached at 720°C at H₂O saturation (3·3 wt % H₂O), but at this pressure and temperature biotite is not a liquidus phase.

Our experiments confirm the idea that peralkaline magmas may contain significant amounts of dissolved water, yet our estimates are nearly twice those from melt inclusion analyses (Wilding et al., 1993). The phase equilibrium constraints detailed above are robust and we believe that this mismatch arises from a problem in interpreting the melt inclusion data. Such a difference could arise from a variety of factors, such as post-eruptive H₂O diffusive loss in the trapped melt inclusions or inclusion formation during ascent, i.e. at pressures lower than the pre-eruptive one, either of these processes being driven by, or occurring during, open-system degassing during eruption. Evidence for open-system degassing is in fact suggested by H/D isotope data (Wilding et al., 1993). Restored pre-eruptive melt H₂O contents using measured D and the matrix H₂O values of glassy volcanic rocks at Olkaria have yielded values up to 5·7 wt % (Wilding et al., 1993), i.e. identical to the H₂O contents at saturation at 150 MPa and in agreement with our phase equilibrium constraints. We conclude that melt inclusion data at Olkaria provide only a minimum estimate of the amount of water present in the magma chamber before eruption.

In summary, the pre-eruptive intensive parameters derived in this study are characterized by last equilibration at an fO₂ at, or possibly below, FMQ, temperatures well below 800°C, and water-rich conditions. For the two more peralkaline samples, the pressure of the storage zone is probably >50 MPa, as indicated by the occurrence of near-liquidus biotite in the rocks. It cannot be much higher than 150 MPa, however, as increasing pressures will further depress the liquidus of tectosilicates, whereas that of amphibole will be correspondingly less affected, making it difficult to preserve the liquidus phase assemblage displayed by the rocks. Such a pressure for magma storage is also indicated by independent geochemical evidence (Davies & Macdonald, 1987).

	ACKNOWLEDGEMENTS

 
Thorough and constructive reviews were provided by Mike Carroll, Gail Mahood and Jim Webster. Bernard W. Evans and Michel Pichavant provided useful comments. The very careful and constructive editorial handling of Dennis Geist is gratefully acknowledged.

	FOOTNOTES

 
^*Corresponding author. Telephone: 0033-2-38-25-53-40. Fax: 0033-2-38-63-64-88. E-mail: bscaille{at}cnrs-orleans.fr

	REFERENCES

TOP ABSTRACT INTRODUCTION ROCKS STUDIED EXPERIMENTAL AND ANALYTICAL... RESULTS DISCUSSION REFERENCES

 
Abbott, R. N. (1981). AFM liquidus projections for granitic magmas, with special reference to hornblende, biotite and garnet. Canadian Mineralogist 19, 103– 110.

Bailey, D. K. (1974). Experimental petrology relating to oversaturated peralkaline volcanics: a review. Bulletin of Volcanology 38, 635– 652.

Bailey, D. K. & Cooper, J. P. (1978). Comparison of the crystallisation of pantelleritic obsidians under hydrous and anhydrous conditions. In: Progress in Experimental Petrology, NERC Publications Series 11, 230– 233.

Bailey, D. K. & Macdonald, R. (1970). Petrochemical variations among mildly peralkaline (comendite) obsidians from the ocean and continents. Contributions to Mineralogy and Petrology 28, 340– 351.

Bailey, D. K. & Macdonald, R. (1987). Dry peralkaline felsic liquids and carbon dioxide flux through the Kenya rift zone. In: Mysen, B. (ed.) Magmatic Processes: Physicochemical Principles. Geochemical Society, Special Publication 1, 91– 105.

Bailey, D. K. & Schairer, J. F. (1964). Feldspar–liquid equilibria in peralkaline liquids—the orthoclase effect. American Journal of Science 262, 1198– 1206.

Bailey, D. K. & Schairer, J. F. (1966). The system Na₂O–Al₂O₃–Fe₂O₃–SiO₂ at 1 atmosphere, and the petrogenesis of alkaline rocks. Journal of Petrology 7, 114– 170.[Abstract/Free Full Text]

Bailey, D. K., Cooper, J. P. & Knight, J. L. (1974). Anhydrous melting and crystallisation of peralkaline obsidians. Bulletin of Volcanology 38, 653– 665.

Barberi, F., Santacroce, R. & Varet, J. (1974). Peralkaline rocks of the Afar depression. Bulletin of Volcanology 38, 755– 790.

Barberi, F., Ferrara, G., Santacroce, R., Treuil, M. & Varet, J. (1975). A transitional basalt–pantellerite sequence of fractional crystallisation, the Boina Centre (Afar, Ethiopia). Journal of Petrology 16, 22– 56.[Abstract/Free Full Text]

Barclay, J., Carroll, M. R., Houghton, B. F. & Wilson, C. J. N. (1996). Pre-eruptive volatile content and degassing history of an evolving peralkaline volcano. Journal of Volcanology and Geothermal Research 74, 75– 87.

Bénard, F., Moutou, P. & Pichavant, M. (1985). Phase relations of tourmaline leucogranites and the significance of tourmaline in silicic magma. Journal of Geology 93, 271– 291.

Bizouard, H., Barberi, F. & Varet, J. (1980). Mineralogy and petrology of Erta Ale and Boina volcanic series, Afar rift, Ethiopia. Journal of Petrology 21, 401– 436.[Abstract/Free Full Text]

Black, S., Macdonald, R. & Kelly, M. R. (1997). Crustal origin for peralkaline rhyolites from Kenya: evidence from U-series disequilibria and Th-isotopes. Journal of Petrology 38, 277– 297.

Blank, J., Stolper, E. & Carroll, M. R. (1993). Solubilities of carbon dioxide and water in rhyolitic melt at 850°C and 750 bars. Earth and Planetary Science Letters 119, 27– 36.

Bohrson, W. A. & Reid, M. R. (1997). Genesis of silicic peralkaline volcanic rocks in an ocean island setting by crustal melting and open-system processes: Socorro Island, Mexico. Journal of Petrology 38, 1137– 1166.

Bowen, N. L. (1937). Recent high temperature research on silicates and its significance in igneous geology. American Journal of Science 33, 1– 21.

Carmichael, I. S. E. (1962). Pantelleritic liquids and their phenocrysts. Mineralogical Magazine 33, 86– 113.

Carmichael, I. S. E. (1967). The iron–titanium oxides of salic volcanic rocks and their associated ferromagnesian silicates. Contributions to Mineralogy and Petrology 14, 36– 64.

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.[Web of Science]

Carmichael, I. S. E. & MacKenzie, W. S. (1963). Feldspar–liquid equilibria in pantellerites: an experimental study. American Journal of Science 261, 382– 396.[Abstract]

Caroff, M., Maury, R. C., Leterrier, J., Joron, J. L., Cotten, J. & Guille, G. (1993). Trace element behavior in the alkali basalt–comenditic trachyte series from Mururoa atoll, French Polynesia. Lithos 30, 1– 22.[Web of Science]

Civetta, L., D’Antonio, M., Orsi, G. & Tilton, G. R. (1998). The geochemistry of volcanic rocks from Pantelleria Island, Sicily channel: petrogenesis and characteristics of the mantle source region. Journal of Petrology 39, 1453– 1491.

Clarke, M. C. G., Woodhall, D. G., Allen, D. & Darling, G. (1990). Geological, volcanological and hydrogeological controls on the occurrence of geothermal activity in the area surrounding Lake Naivasha, Kenya. Report. Nairobi: Ministry of Energy, 138 pp.

Clemens, J. D. & Wall, V. J. (1981). Origin and crystallization of some peraluminous (S-type) granitic magmas. Canadian Mineralogist 19, 111– 131.

Clemens, J. D., Holloway, J. R. & White, A. J. R. (1986). Origin of an A-type granite: experimental constraints. American Mineralogist 71, 317– 324.[Abstract]

Conrad, W. K. (1984). The mineralogy and petrology of compositionally zoned ash flow tuffs, and related silicic volcanic rocks, from the McDermitt caldera complex, Nevada–Oregon. Journal of Geophysical Research 89, 8639– 8664.

Conrad, W. K., Nicholls, I. A. & Wall, V. J. (1988). Water-saturated and undersaturated melting of metaluminous and peraluminous crustal compositions at 10 kb: evidence for the origin of silicic magmas in the Taupo Volcanic Zone, New Zealand, and other occurrences. Journal of Petrology 29, 765– 803.[Abstract/Free Full Text]

Crisp, J. A. & Spera, F. J. (1987). Pyroclastic flows and lavas of the Mogan and Fataga formations, Tejeda Volcano, Gran Canaria, Canary Islands: mineral chemistry, intensive parameters, and magma chamber evolution. Contributions to Mineralogy and Petrology 96, 503– 518.

Dall’Agnol, R., Scaillet, B. & Pichavant, M. (1999). An experimental study of a lower Proterozoic A-type granite from the eastern Amazonian craton, Brazil. Journal of Petrology 11, 1673– 1698.

Davies, G. R. & Macdonald, R. (1987). Crustal influences in the petrogenesis of the Naivasha basalt–comendite complex: combined trace element and Sr–Nd–Pb isotope constraints. Journal of Petrology 28, 1009– 1031.[Abstract/Free Full Text]

Deer, W. A., Howie, R. A. & Zussman, J. (1978). Rock Forming Minerals. Single-chain Silicates. Harlow, UK: Longman, 668 pp.

Devine, J. D., Gardner, J. E., Brack, H. P., Layne, G. D. & Rutherford, M. J. (1995). Comparison of microanalytical methods for estimating H₂O contents of silicic volcanic glasses. American Mineralogist 80, 319– 328.[Abstract]

Ernst, W. G. (1962). Synthesis, stability relations and occurrence of riebeckite and riebeckite–arfvedsonite solid solutions. Journal of Geology 70, 689– 736.

Ewart, A., Taylor, S. R. & Capp, A. C. (1968). Geochemistry of the pantellerites of Mayor Island, New Zealand. Contributions to Mineralogy and Petrology 17, 116– 140.

Foerstner, H. (1881). Nota preliminare sulla geologia dell’isola di Pantelleria secondo gli studi fatti negli anni 1874 e 1881. Bolletino di Rendi Conti Geologia Italia 12, 523– 556.

Ghiorso, M. S. & Sack, R. O. (1991). Fe–Ti oxide geothermometry: thermodynamic formulation and the estimation of intensive variables in silicic magmas. Contributions to Mineralogy and Petrology 108, 485– 510.[Web of Science]

Gibson, I. L. (1972). The chemistry and petrogenesis of a suite of pantellerites from the Ethiopian rift. Journal of Petrology 13, 31– 44.[Abstract/Free Full Text]

Harris, C. (1983). The petrology of lavas and associated plutonic inclusions of Ascension island. Journal of Petrology 24, 424– 470.

Harris, C. (1986). A quantitative study of magmatic inclusions in the plutonic ejecta of Ascension island. Journal of Petrology 27, 251– 276.[Abstract/Free Full Text]

Holloway, J. R. & Ford, C. E. (1975). Fluid absent melting of the fluorohydroxy-amphibole pargasite. Earth and Planetary Science Letters 25, 44– 48.

Holtz, F., Pichavant, M., Barbey, P. & Johannes, W. (1992). Effects of H₂O on liquidus phase relations in the haplogranite system at 2 and 5 kbar. American Mineralogist 77, 1223– 1241.[Abstract]

Johannes, W. & Holtz, F. (1996). Petrogenesis and Experimental Petrology of Granitic Rocks. Berlin: Springer, 335 pp.

Jørgensen, K. A. (1987). Mineralogy and petrology of alkaline granophyric xenoliths from the Thorsmork ignimbrite, southern Iceland. Lithos 20, 153– 168.

Kovalenko, V. I., Hervig, R. L. & Sheridan, M. F. (1988). Ion-microprobe analyses of trace elements in anorthoclase, hedenbergite, aenigmatite, quartz, apatite and glass in pantellerite: evidence for high water contents in pantellerite melts. American Mineralogist 73, 1038– 1045.[Abstract]

Kretz, R. (1983). Symbols for rock-forming minerals. American Mineralogist 68, 277– 279.[Abstract]

Lindsley, D. (1971). Synthesis and preliminary results on the stability of aenigmatite (Na₂Fe₅TiSi₆O₂₀). Carnegie Institution of Washington, Annual Report Geophysical Laboratory 1969–1970, 188– 190.

London, D., Morgan, G. B., VI & Hervig, R. V. (1989). Vapor-undersaturated experiments with Macusani glass + H₂O at 200 MPa, and the internal differentiation of granitic pegmatites. Contributions to Mineralogy and Petrology 102, 1– 17.

Lowenstern, J. B. & Mahood, G. A. (1991). New data on magmatic H₂O contents of pantellerites, with implications for petrogenesis and eruptive dynamics at Pantelleria. Bulletin of Volcanology 54, 78– 83.[Web of Science]

Macdonald, R. (1974). Nomenclature and petrochemistry of the peralkaline oversaturated extrusive rocks. Bulletin of Volcanology 38, 498– 505.

Macdonald, R. & Bailey, D. K. (1973). The chemistry of the peralkaline oversaturated obsidians. US Geological Survey, Professional Paper 440-N(Part 1), 1– 37.

Macdonald, R., Davies, G. R., Bliss, C. M., Leat, P. T., Bailey, D. K. & Smith, R. L. (1987). Geochemistry of high-silicic peralkaline rhyolites, Naivasha, Kenya rift valley. Journal of Petrology 28, 979– 1008.[Abstract/Free Full Text]

Macdonald, R., Kjarsgaard, B. A., Skilling, I. P., Davies, G. R., Hamilton, D. L. & Black, S. (1993). Liquid immiscibility between trachyte and carbonate in ash flow tuffs from Kenya. Contributions to Mineralogy and Petrology 114, 276– 287.

Mahood, G. A. (1981). Chemical evolution of a Pleistocene rhyolitic center: Sierra la Primavera, Jalisco, Mexico. Contributions to Mineralogy and Petrology 77, 129– 149.

Mahood, G. A. (1984). Pyroclastic rocks and calderas associated with strongly peralkaline magmatism. Journal of Geophysical Research 89, 8540– 8552.

Mahood, G. A., Halliday, A. N. & Hildreth, W. (1990). Isotopic evidence for the origin of pantellerites in a rift-related alkalic suite: Pantelleria, Italy. IAVCEI Abstracts, International Volcanological Congress, Mainz.

Manning, D. A. C. (1981). The effect of fluorine on liquidus phase relationships in the system Qz–Ab–Or with excess water at 1 kbar. Contributions to Mineralogy and Petrology 76, 206– 215.

Marshall, A. S., Hinton, R. W. & Macdonald, R. (1998). Phenocrystic fluorite in peralkaline rhyolites, Olkaria, Kenya rift valley. Mineralogical Magazine 62, 477– 486.[Abstract]

Martel, C., Pichavant, M., Bourdier, J. L., Traineau, H., Holtz, F. & Scaillet, B. (1998). Magma storage conditions and control of eruption regime in silicic volcanoes: experimental evidence from Mt. Pelée. Earth and Planetary Science Letters 156, 89– 99.

Martel, C., Pichavant, M., Holtz, F., Scaillet, B., Bourdier, J. L. & Traineau, H. (1999). Effects of fO₂ and H₂O on andesite phase relations between 2 and 4 kbar. Journal of Geophysical Research 104, 29453– 29470.

Mungall, J. E. & Martin, R. F. (1995). Petrogenesis of basalt–comendite and basalt–pantellerite suite, Terceira Azores, and some implications for the origin of ocean-island rhyolites. Contributions to Mineralogy and Petrology 119, 43– 55.[Web of Science]

Nicholls, I. & Carmichael, I. S. E. (1969). Peralkaline acid liquids: a petrological study. Contributions to Mineralogy and Petrology 20, 268– 294.

Noble, D. C. (1968). Systematic variation of major element in comendite and pantellerite glasses. Earth and Planetary Science Letters 4, 167– 172.

Novak, S. W. & Mahood, G. A. (1986). Rise and fall of a basalt–trachyte–rhyolite magma system at the Kane Spring Wash caldera, Nevada. Contributions to Mineralogy and Petrology 94, 352– 373.

Pichavant, M. (1987). Effects of B and H₂O on liquidus phase relations in the haplogranite system at 1 kbar. American Mineralogist 72, 1056– 1070.[Abstract]

Pichavant, M. & Manning, D. A. C. (1984). Petrogenesis of tourmaline granites and topaz granites: the contribution of experimental data. Physics of the Earth and Planetary Interiors 35, 31– 50.

Pichavant, M., Boher, M., Stenger, J. F., Aissa, M. & Charoy, B. (1987). Relations de phases des granites de Beauvoir entre 1 et 3 kbar en conditions de saturation en H₂O. Géologie France 2–3, 77– 86.

Pownceby, M. I. & O’Neill, H. St C. (1994). Thermodynamic data from redox reactions at high temperatures. III. Activity–composition relations in Ni–Pd alloys from EMF measurements at 850–1250 K, and calibration of the NiO + Ni–Pd assemblage as a redox sensor. Contributions to Mineralogy and Petrology 116, 327– 339.

Robert, J. L. & Maury, R. (1979). Natural occurrence of an (Fe, Mn, Mg) tetrasilicic potassium mica. Contributions to Mineralogy and Petrology 68, 117– 123.

Roux, J. & Varet, J. (1975). Alkali–feldspar liquid equilibrium relationships in peralkaline oversaturated systems and volcanic rocks. Contributions to Mineralogy and Petrology 49, 67– 81.

Sato, H., Nakada, S., Fujii, T., Nakamura, M. & Suzuki-Kamata, K. (1999). Groundmass pargasite in the 1991–1995 dacite of Unzen volcano: phase stability experiments and volcanological implications. Journal of Volcanology and Geothermal Research 89, 197– 212.

Scaillet, B. & Evans, B. W. (1999). The June 15, 1991 eruption of Mount Pinatubo. I. Phase equilibria and pre-eruption P–T–fO₂–fH₂O conditions of the dacite magma. Journal of Petrology 40, 381– 411.

Scaillet, B., Pichavant, M., Roux, J., Humbert, G. & Lefèvre, A. (1992). Improvements of the Shaw membrane technique for measurement and control of fH₂ at high temperatures and pressures. American Mineralogist 77, 647– 655.[Abstract]

Scaillet, B., Pichavant, M. & Roux, J. (1995). Experimental crystallization of leucogranite magmas. Journal of Petrology 36, 663– 705.[Abstract/Free Full Text]

Schairer, J. F. & Bowen, N. L. (1955). The system K₂O–Al₂O₃–SiO₂. American Journal of Science 253, 681– 746.

Schairer, J. F. & Bowen, N. L. (1956). The system Na₂O–Al₂O₃–SiO₂. American Journal of Science 254, 129– 195.

Schmidt, B. C., Scaillet, B. & Holtz, F. (1995). Accurate control of fH₂ in cold seal pressure vessels with the Shaw membrane technique. European Journal of Mineralogy 7, 893– 903.

Sutherland, D. S. (1974). Petrography and mineralogy of peralkaline silicic rocks. Bulletin of Volcanology 38, 517– 547.

Taylor, J. R., Wall, V. J. & Pownceby, M. I. (1992). The calibration and application of accurate redox sensors. American Mineralogist 77, 284– 295.[Abstract]

Thompson, R. N. & Chisholm, J. E. (1969). Synthesis of aenigmatite. Mineralogical Magazine 37, 253– 255.

Thompson, R. N. & MacKenzie, W. S. (1967). Feldspar–liquid equilibria in peralkaline acid liquids: an experimental study. American Journal of Science 265, 714– 734.

Tuttle, O. F. & Bowen, N. L. (1958). Origin of granite in the light of experimental studies in the system NaAlSi₃O₈–KAlSi₃O₈–SiO₂–H₂O. Geological Society of America Memoir, 74, 153 pp.

Villari, L. (1974). The island of Pantelleria. Bulletin of Volcanology 38, 680– 724.

Washington, H. S. (1913). The volcanoes and rocks of Pantelleria: I. Journal of Geology 21, 16– 27.

Washington, H. S. (1914a). The volcanoes and rocks of Pantelleria: II. Journal of Geology 22, 653– 670.

Washington, H. S. (1914b). The volcanoes and rocks of Pantelleria: III. Journal of Geology 22, 683– 713.

Webster, J. D. (1992). Water solubility and chlorine partitioning in Cl-rich granitic systems. Effect of melt composition at 2 kbar and 800°C. Geochimica et Cosmochimica Acta 56, 679– 687.

Webster, J. D. (1997). Chloride solubility in felsic melts and the role of chloride in magmatic degassing. Journal of Petrology 38, 1793– 1807.

Webster, J. D., Holloway, J. R. & Hervig, R. L. (1987). Phase equilibria of a Be, U and F-enriched vitrophyre from Spor Mountain, Utah. Geochimica et Cosmochimica Acta 51, 389– 402.

Webster, J. D., Taylor, R. P. & Bean, C. (1993). Pre-eruptive melt composition and constraints on degassing of a water-rich pantellerite magma, Fantale volcano, Ethiopia. Contributions to Mineralogy and Petrology 114, 53– 62.

Weidner, J. R. & Martin, R. F. (1987). Phase equilibria of a fluorine rich leucogranite from the St. Austell pluton, Cornwall. Geochimica et Cosmochimica Acta. 51, 1591– 1597.

Wilding, M. C., Macdonald, R., Davies, J. R. & Fallick, A. E. (1993). Volatile characteristics of peralkaline rhyolites from Kenya: an ion microprobe, infrared spectroscopic and hydrogen isotope study. Contributions to Mineralogy and Petrology 144, 264– 275.

Wolff, J. A. & Wright, J. V. (1981). Formation of the Green Tuff, Pantelleria. Bulletin of Volcanology 44, 681– 690.

Wones, D. R. (1989). Significance of the assemblage titanite + magnetite + quartz in granitic rocks. American Mineralogist 74, 744– 749.[Abstract]

Wyllie, P. J. & Tuttle, O. F. (1961). Experimental investigation of silicate systems containing two volatile components: II—the effects of NH₃ and HF, in addition to H₂O, on the melting temperatures of albite and granite. American Journal of Science 259, 128– 143.

Zies, E. G. (1960). Chemical analyses of two pantellerites. Journal of Petrology 1, 304– 308.[Free Full Text]

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				R. Macdonald and B. Baginski The central Kenya peralkaline province: a unique assemblage of magmatic systems Mineralogical Magazine, June 15, 2009; 73(1): 1 - 16. [Abstract] [Full Text] [PDF]

				R. Macdonald, B. Baginski, H.E. Belkin, P. Dzierzanowski, and L. Jezak REE partitioning between apatite and melt in a peralkaline volcanic suite, Kenya Rift Valley Mineralogical Magazine, April 17, 2009; 72(6): 1147 - 1161. [Abstract] [Full Text] [PDF]

				A. S. Marshall, R. Macdonald, N. W. Rogers, J. G. Fitton, A. G. Tindle, K. Nejbert, and R. W. Hinton Fractionation of Peralkaline Silicic Magmas: the Greater Olkaria Volcanic Complex, Kenya Rift Valley J. Petrology, February 1, 2009; 50(2): 323 - 359. [Abstract] [Full Text] [PDF]

				M. S. Ghiorso and B. W. Evans Thermodynamics of Rhombohedral Oxide Solid Solutions and a Revision of the FE-TI Two-Oxide Geothermometer and Oxygen-Barometer Am J Sci, November 1, 2008; 308(9): 957 - 1039. [Abstract] [Full Text] [PDF]

				J. R. Vidal Solano, H. Lapierre, J. M. Stock, A. Demant, F. A. Paz Moreno, D. Bosch, P. Brunet, and A. Amortegui Isotope geochemistry and petrogenesis of peralkaline Middle Miocene ignimbrites from central Sonora: relationship with continental break-up and the birth of the Gulf of California Bulletin de la Societe Geologique de France, September 1, 2008; 179(5): 453 - 464. [Abstract] [Full Text] [PDF]

				R. Macdonald, N. W. Rogers, and A. G. Tindle Distribution of germanium between phenocrysts and melt in peralkaline rhyolites from the Kenya Rift Valley Mineralogical Magazine, December 1, 2007; 71(6): 703 - 713. [Abstract] [Full Text] [PDF]

				A. C. Simon, M. R. Frank, T. Pettke, P. A. Candela, P. M. Piccoli, C. A. Heinrich, and M. Glascock An evaluation of synthetic fluid inclusions for the purpose of trapping equilibrated, coexisting, immiscible fluid phases at magmatic conditions American Mineralogist, January 1, 2007; 92(1): 124 - 138. [Abstract] [Full Text] [PDF]

				B. SCAILLET and R. MACDONALD Experimental and Thermodynamic Constraints on the Sulphur Yield of Peralkaline and Metaluminous Silicic Flood Eruptions J. Petrology, July 1, 2006; 47(7): 1413 - 1437. [Abstract] [Full Text] [PDF]

				E. V. Badanina, R. B. Trumbull, P. Dulski, M. Wiedenbeck, I. V. Veksler, and L. F. Syritso THE BEHAVIOR OF RARE-EARTH AND LITHOPHILE TRACE ELEMENTS IN RARE-METAL GRANITES: A STUDY OF FLUORITE, MELT INCLUSIONS AND HOST ROCKS FROM THE KHANGILAY COMPLEX, TRANSBAIKALIA, RUSSIA Can Mineral, June 1, 2006; 44(3): 667 - 692. [Abstract] [Full Text] [PDF]

				A.I.S. Kemp, R.J. Wormald, M.J. Whitehouse, and R.C. Price Hf isotopes in zircon reveal contrasting sources and crystallization histories for alkaline to peralkaline granites of Temora, southeastern Australia Geology, October 1, 2005; 33(10): 797 - 800. [Abstract] [Full Text] [PDF]

				J. C. White, M. Ren, and D. F. Parker VARIATION IN MINERALOGY, TEMPERATURE, AND OXYGEN FUGACITY IN A SUITE OF STRONGLY PERALKALINE LAVAS AND TUFFS, PANTELLERIA, ITALY Can Mineral, August 1, 2005; 43(4): 1331 - 1347. [Abstract] [Full Text] [PDF]

				G. D. Ventura, G. Iezzi, G. J. Redhammer, F. C. Hawthorne, B. Scaillet, and D. Novembre Synthesis and crystal-chemistry of alkali amphiboles in the system Na2O-MgO-FeO-Fe2O3-SiO2-H2O as a function of fO2 American Mineralogist, August 1, 2005; 90(8-9): 1375 - 1383. [Abstract] [Full Text] [PDF]

				A. PECCERILLO, M. R. BARBERIO, G. YIRGU, D. AYALEW, M. BARBIERI, and T. W. WU Relationships between Mafic and Peralkaline Silicic Magmatism in Continental Rift Settings: a Petrological, Geochemical and Isotopic Study of the Gedemsa Volcano, Central Ethiopian Rift J. Petrology, November 1, 2003; 44(11): 2003 - 2032. [Abstract] [Full Text] [PDF]

				M. Marks, M. Marks, and G. Markl Ilimaussaq 'en miniature': closed-system fractionation in an agpaitic dyke rock from the Gardar Province, South Greenland (contribution to the mineralogy of Ilimaussaq no. 117) Mineralogical Magazine, October 1, 2003; 67(5): 893 - 919. [Abstract] [Full Text] [PDF]

				B. SCAILLET and R. MACDONALD Experimental Constraints on the Relationships between Peralkaline Rhyolites of the Kenya Rift Valley J. Petrology, October 1, 2003; 44(10): 1867 - 1894. [Abstract] [Full Text] [PDF]

				M. MARKS, T. VENNEMANN, W. SIEBEL, and G. MARKL Quantification of Magmatic and Hydrothermal Processes in a Peralkaline Syenite-Alkali Granite Complex Based on Textures, Phase Equilibria, and Stable and Radiogenic Isotopes J. Petrology, July 1, 2003; 44(7): 1247 - 1280. [Abstract] [Full Text] [PDF]

				W. Matthews, W. Matthews, R. L. Linnen, and Q. Guo A filler-rod technique for controlling redox conditions in cold-seal pressure vessels American Mineralogist, April 1, 2003; 88(4): 701 - 707. [Abstract] [Full Text] [PDF]

				J. Blundy, J. Blundy, and B. Wood Mineral-Melt Partitioning of Uranium, Thorium and Their Daughters Reviews in Mineralogy and Geochemistry, January 1, 2003; 52(1): 59 - 123. [Full Text] [PDF]

				B. Scaillet and M. Pichavant Experimental constraints on volatile abundances in arc magmas and their implications for degassing processes Geological Society, London, Special Publications, January 1, 2003; 213(1): 23 - 52. [Abstract] [PDF]

				R. Macdonald and H. E. Belkin Compositional variation in minerals of the chevkinite group Mineralogical Magazine, December 1, 2002; 66(6): 1075 - 1098. [Abstract] [Full Text] [PDF]

				R. Macdonald, A. S. Marshall, J. B. Dawson, R. W. Hinton, and P. G. Hill Chevkinite-group minerals from salic volcanic rocks of the East African Rift Mineralogical Magazine, April 1, 2002; 66(2): 287 - 299. [Abstract] [Full Text] [PDF]

				V. R. TROLL and H.-U. SCHMINCKE Magma Mixing and Crustal Recycling Recorded in Ternary Feldspar from Compositionally Zoned Peralkaline Ignimbrite A', Gran Canaria, Canary Islands J. Petrology, February 1, 2002; 43(2): 243 - 270. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (54) Request Permissions Google Scholar Articles by SCAILLET, B. Articles by MACDONALD, R. Search for Related Content GeoRef GeoRef Citation Social Bookmarking          What's this?

Online ISSN 1460-2415 - Print ISSN 0022-3530

Copyright © 2009 Oxford University Press

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics