Journal of Petrology | Volume 39 | Number 10 | Pages 1787-1804 | 1998
© Oxford University Press 1998
Fluid-Absent Melting Behavior of a Two-Mica Metapelite: Experimental Constraints on the Origin of Black Hills Granite
Department of Geological Sciences, University of Oregon Eugene, or 97403-1272, USA
Received May 22, 1997; Revised typescript accepted April 17, 1998
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
We have studied experimentally the vapor-absent melting behavior of a Proterozoic metapelite (HP-60-1) from the Black Hills, South Dakota, to assess whether the high-
18O, tourmaline-bearing granites of the Harney Peak suite resulted from muscovite-dominated dehydration-melting at 10 kbar. Samples were sealed with no added H2O in gold capsules, then run in a 0.5 inch (1.27 cm) piston-cylinder apparatus over the temperature interval 812–975°C. Experiments were conducted at 10 kbar to allow comparison with other metapelite melting studies. Mass balance calculations imply the melting reaction muscovite + biotite + plagioclase + quartz 
melt + alkali feldspar + aluminosilicate + garnet until muscovite is exhausted (<812°C), then the reaction biotite + plagioclase + quartz + alkali feldspar ± aluminosilicate melt + garnet. The inferred muscovite reaction is completed below 812°C and produces 1–2 wt % melt. Melt production of the continuous biotite reaction increases steadily with rising temperature, to 32 wt % in our highest temperature run. The low melt productivity of the muscovite reaction at 10 kbar suggests that melting at lower pressures may be necessary to generate mobile melt fractions by muscovite-dominated reactions in our starting material. In addition, melts of HP-60-1 are considerably more mafic than the tourmaline-bearing granites, with average combined TiO2 + MgO + FeO contents of 1.85 wt % and 0.84 wt %, respectively. In particular, melts of HP-60-1 are enriched in Ti, and would probably stabilize biotite rather than tourmaline upon cooling, arguing against HP-60-1 as a potential source rock for the tourmaline granites. KEY WORDS: Black Hills; dehydration-melting; granite; metasediments; muscovite
|
Introduction |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
Peraluminous leucogranites such as the Miocene granites of the Himalaya and tourmaline granites of the Harney Peak suite in South Dakota are widely considered to be the product of muscovite dehydration-melting; yet few experimental studies of muscovite dehydration-melting reactions have been performed, because the low temperatures of the reactions hinder the approach to equilibrium. In addition, melt fractions generated are typically low, resulting in small melt pockets which are very difficult to analyze with the electron microprobe. In this study, we assess the possible genetic relationship between the tourmaline-bearing granites of the Harney Peak suite and the aluminous metasediments these granites intrude. Although the attempt to investigate the muscovite dehydration-melting reaction was hindered by the surprisingly low melt productivity, we were able to infer a balanced reaction, illuminating the dominantly peritectic nature of fluid-absent muscovite breakdown under the conditions investigated.
Isotopic and trace element studies indicate a genetic relationship between S-type granites and aluminous metasediments in the Black Hills of South Dakota, where the Harney Peak Granite (HPG) intruded Proterozoic and Archean schists and metagreywackes undergoing regional metamorphism. The timing and mode of emplacement of the HPG, and the deformation and regional metamorphic history of the area have been described by numerous workers (e.g. Riley, 1970
; Redden et al., 1985; Duke et al., 1988, 1990; Norton & Redden, 1990; Helms & Labotka, 1991). Pertinent to this study, the HPG consists of a large pluton flanked by numerous satellite intrusions. As shown in Fig. 1, biotite is the dominant ferromagnesian mineral in the core of the main pluton, but tourmaline predominates along the perimeter and in the satellite intrusions (Nabelek et al., 1992a). The presence of tourmaline indicates enrichment of boron (B) during the magmatic stage, offering a clue to the petrogenesis of the tourmaline-bearing granites of the HPG suite. However, as we have not analyzed our run products for B we will not discuss the role of B in this paper, except in reference to the previous work that motivated this study.
|
Large-scale geochemical variation between the biotite- and tourmaline-bearing granites of the HPG indicates separate source rocks for the two suites, and possibly variable temperature conditions (Nabelek et al., 1992a
, 1992b; Krogstad et al., 1993). The average 18O values of the granites increase from 11.5
± 0.6 in the core of the pluton to 13.2 ± 0.8 in the flanks (Nabelek et al., 1992b), suggesting two isotopically distinct metasedimentary sources. The average oxygen isotopic composition of the Proterozoic schist wallrocks is identical to that of the high-18O tourmaline granites, suggesting that these schists, or their equivalent at depth, may have been the source rock (Nabelek et al., 1992b). Also supporting this inference are lead isotopic data suggesting that the high-18O granites were derived from a source with a relatively short crustal residence time (Krogstad et al., 1993). In contrast, the lower 18O values and more radiogenic lead isotopic signature of the biotite-bearing granites require a somewhat more immature metasedimentary source, possibly the Archean metasediments exposed in the core of the HPG complex, or their equivalent at depth (Nabelek et al., 1992a; Krogstad et al., 1993). In addition, trace element studies of Nabelek et al. (1992a) revealed that the ratio of B/TiO2 is significantly higher in the high-18O tourmaline-bearing granites. As the likely residence of boron in aluminous metasediments is muscovite (e.g. Eugster & Wright, 1960; Stubican & Roy, 1962; Shearer & Papike, 1986), and that of TiO2 is biotite, the high B/TiO2 ratio suggests low degrees of muscovite dehydration-melting of the schists without a significant contribution to the melt from biotite. The lower B/TiO2 ratios in the low-18O granites are consistent with contribution of TiO2 as a result of the breakdown of biotite.
To explain the variable trace element signatures and 18O values of the core and perimeter granites, Nabelek et al. (1992a) proposed two petrogenic models. In the first model, regional sillimanite grade metamorphism induced large degrees of biotite dehydration-melting of a low-18O source. The metamorphic event culminated in the ascent of the core biotite granite to higher levels in the crust, where heat from the intrusion initiated muscovite-dominated dehydration-melting of the surrounding Proterozoic schist wallrocks. In the second model, the low- and high-18O granite suites may represent melts generated in situ in a layered crust in response to a regional thermal event. In this scenario, the low-18O source at a greater depth would be hotter, inducing large extent biotite dehydration-melting. At a higher level in the crust, the high-18O source would be cooler, resulting in lower degrees of muscovite-dominated dehydration-melting. In either case, the geochemical evidence implies that the probable origin of the tourmaline-bearing granite suite is by melting of a high-18O metasedimentary source at conditions such that muscovite is the only hydrous phase contributing significantly to the melting reaction. This could either result from melting a source rock in which muscovite was the dominant hydrous phase, or be a consequence of melt extraction at a temperature below that at which biotite breaks down. As our starting material contains more biotite than muscovite, our focus was on whether the tourmaline granites of the HPG resulted from the extraction of melt formed by muscovite-dominated dehydration-melting of a Proterozoic aluminous metasedimentary rock.
|
Characterization of the Starting Material and Experimental Procedures |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
All of our experiments utilized a natural Proterozoic metapelite from the Black Hills, SD, provided by P. I. Nabelek. The bulk composition, mineral compositions and mode of HP-60-1 are presented in Table 1. HP-60-1 contains comparable amounts of feldspar and mica, and might be considered by some to be a metagreywacke. However, the presence of aluminosilicate and muscovite, and the relatively high alumina saturation index [ASI; molar Al2O3/(CaO + Na2O + K2O)] of 1.8 compares better with experimentally studied metapelite starting materials than with metagreywacke (e.g. Patiño-Douce & Beard, 1996
). Thus, we refer to HP-60-1 as ametapelite, in keeping with other studies utilizing peraluminous yet feldspar-rich starting materials (e.g. Le Breton & Thompson, 1988; Vielzeuf & Holloway, 1988; Gardien et al., 1995). Although similar in mode to the metapelite melted by Vielzeuf & Holloway (1988), the bulk composition of HP-60-1 is comparatively richer in silica and poorer in alumina and mafic components, reflected in higher quartz and slightly lower overall mica modal abundances in HP-60-1. In contrast to the plagioclase-poor rock studied by Patiño-Douce & Johnston (1991), HP-60-1 contains abundant plagioclase, but comparatively less combined mica. Differing from both previous studies, HP-60-1 contains alkali feldspar, and lacks garnet. HP-60-1 occurs within the second-sillimanite zone along the southwestern flank of the Harney Peak granitic intrusion (P. I. Nabelek, personal communication, 1995), and contains clots of fibrolite within muscovite laths. Geothermometry work on surrounding metasedimentary rocks by Helms & Labotka (1991) indicated that the peak-metamorphic temperature achieved in the Sil + Kfs zone along the southwestern flank of the Harney Peak Granite was 
662°C. Friberg et al. (1996) estimated nearly 100°C higher maximum temperatures using garnet core analyses rather than rims, suggesting that the lower temperature estimates proposed by Helms & Labotka (1991) may record retrograde metamorphism. The higher temperature estimates are consistent with extensive migmatization within the second-sillimanite zone (Nabelek & Ternes, 1997). Helms & Labotka (1991) estimated that regional metamorphism of metasediments in the southern Black Hills occurred between 2.0 and 4.4 kbar, whereas Drops & Friberg (1996) obtained 5.3–7.6 kbar using amphibole-plagioclase geobarometry on amphibolite inliers in the core of the Harney Peak Granite. The fibrolitic clots in HP-60-1 record either incipient sub-solidus dehydration of muscovite at the lower peak metamorphic pressure estimates, or the onset of vapor-absent dehydration-melting at the higher pressure estimates (Thompson, 1974). Assuming the pressure estimate of Drops & Friberg (1996) represents regional prograde metamorphic pressure at the time of emplacement of the Harney Peak Granite, then the melting event which produced the HPG may have occurred somewhat deeper, at a pressure >7 kbar, similar to the conditions of our experiments.
|
Experiments were conducted at 10 kbar, covering the temperature interval 812–975°C. Listed in Table 2, run durations varied from 3 days (950°C) to 17 days for the lowest temperature investigated (812°C). Furnace assemblies were constructed of graphite heater-tubes with NaCl outer sleeves. All experiments were run in a 0.5 inch (1.27 cm) piston-cylinder apparatus at the University of Oregon. Samples were ground to <10 µm (some biotite laths were >10 µm) and stored in a 110°C oven for several days before being loaded into gold capsules. Loaded capsules were then reheated in the drying oven for 0.5 h before being crimped shut while warm and sealed by arc-welding. Sealed capsules were weighed before the experiments. After every run, the capsules were carefully cleaned and inspected for tears or weight loss. If the capsule was breached or weight loss occurred, the run was discarded and repeated.
|
Pressures reported are nominal Heise gauge pressures, believed to be accurate to within 0.5 kbar for the NaCl assembly used (Bohlen, 1984
). Experiments were pressurized at room temperature to 
2 kbar below the target pressure, then heated to the target temperature. In every case, thermal expansion resulted in the final pressure adjustment being a pressure release (hot, piston out). During the experiments, pressures were maintained within ±83 bars of the nominal target value.
Temperature was measured using W–Re5/W–Re26 thermocouples relative to an Omega electronic ice point (0°C). During all runs, a Eurotherm 808 digital temperature controller maintained temperatures to within ±5°C of the target value. Furnace assemblies were carefully dissected after every experiment. In each case, the crimped and folded capsules had compressed considerably in the vertical dimension, and were slightly indented by the thermocouple. Thus we infer the entire sample was close to the reported temperature (see Pickering et al., 1998).
Patiño-Douce & Beard (1994, 1995) estimated the f(O)2 of a furnace assembly like ours to be at QFM (quartz–fayalite–magnetite) or QFM –2, using orthopyroxene–hematite–quartz and biotite–ilmenite equilibria. The lack of Fe-oxides in our experiments precluded their use in estimating f(O)2. Instead, we attempted to use the equilibrium 6 Grs + 4 Alm + 3 O2 6 Adr + 2 Qtz + 10 Als, with which Patiño-Douce et al. (1993) estimated f(O)2 to be at or below QFM in an identical NaCl-based assembly to the one used in this study. However, the calculated Fe3+ contents of garnet in our experiments were very low, resulting in negligible activities of andradite component in garnet. Thus we were unable to use the above equilibrium to give a reliable estimate of f(O)2 during our experiments. The apparent paucity of Fe3+ in our experimental run products implies that the experimental conditions were relatively reducing. We therefore assume that oxygen fugacity in our experiments was at or below QFM, as determined by Patiño-Douce & Beard (1994, 1995) in a similar assembly.
Although the attainment of equilibrium cannot be rigorously demonstrated because none of the experiments was reversed, a number of observations support an acceptable approach to equilibrium. Melt compositions and modes change in a regular and consistent manner throughout the temperature interval investigated (Fig. 2 and Table 2). In a manner consistent with progressive partial melting, mg-numbers of both biotite and garnet increase systematically with increasing temperature (Fig. 3). In addition, melt and residual mineral compositions are nearly constant throughout any given experimental charge. Using average phase compositions in each run product, multiple linear regressions provide an internally consistent set of phase proportions. Small residuals on all regressions indicate good fits to the compositional data. Taken together, we feel these observations indicate that the small grain-size of the starting material, long run durations (3–17 days) and presence of an H2O-bearing melt facilitated an acceptable approach to equilibrium.
|
standard deviations on analyses.
|
|
Analytical Procedures and Mode Calculations |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
Subsequent to quenching, each run product was mounted in epoxy, sectioned, polished and carbon-coated, in preparation for electron-beam analysis. Quantitative phase analyses were acquired with a Cameca SX-50 electron probe microanalyzer (EPMA) utilizing a ZAF X-ray intensity reduction routine. A 15 keV accelerating voltage and 10 nA beam current were used for all analyses except those of garnet, which were obtained with a 20 nA beam current. Beam diameter varied from 1 µm for all garnet and biotite analyses, to 5 µm for all feldspar and most glass analyses. Where possible, a 10 µm beam was used for glass analyses to minimize alkali loss. To assess the Na2O loss, a time series of analyses was collected from the super-liquidus glass that was used to generate the bulk composition. Sodium counts were collected every 1 s for 60 s, using a 10 µm spot size and 10 nA beam current. Sodium counts decreased linearly throughout the interval, recording an average total loss of 33% in 60 s. However, over the first 10 s, the length of time during which sodium is counted during analyses of glass in the run products, the sodium loss detected in the time series analysis is <5% relative. Accordingly, all sodium values for analyses of glass listed in Table 3 were corrected for a 5% relative loss. However, recent work by Morgan & London (1996)
shows that a 10 nA beam current yields significantly lower intensities for the first second of analysis than do lower beam currents. Thus corrections based on extrapolation to zero time of time series analyses such as we have done will underestimate the true Na concentration (Morgan & London 1996). In addition, textures of the sub-liquidus run products often necessitated the use of a 5 µm spot size, and even a 3 µm spot in the lowest temperature run. Smaller spot sizes also contribute to loss of Na intensity during analyses (Morgan & London, 1996). Hence applying a 5% relative correction for Na2O loss on analyses collected with 3–5 µm spot sizes probably results in minimum estimates of melt Na2O content.
|
For modal analysis, multiple digital back-scattered electron (BSE) images were acquired for each run product with a JEOL 6300FXV scanning electron microscope. The summed areas of the acquired images for each sample covered approximately half the surface area of each sectioned charge. Grey-scale thresholding capabilities of NIH Image software provided binary images of garnet, biotite and alkali feldspar, from which modes were calculated by area. Quartz and aluminosilicate, as well as plagioclase and glass, had grey-scales too alike to be segregated by thresholding. Thus, quartz and aluminosilicate, and also plagioclase and glass, were combined and a total area for each pair of phases was calculated. Mineral modes generated by image analysis are presented in Table 2. To further distinguish quartz from aluminosilicate and plagioclase from glass, we performed mass balance calculations using multiple linear least-squares regression techniques, yielding a consistent set of modes that behave systematically with temperature. We omitted Na2O, as well as F and the minor oxides MnO and P2O5, in the linear regression calculations because of their relatively large uncertainties. To diminish the effect of small inaccuracies in compositional analyses, we followed Patiño-Douce & Johnston (1991)
and combined image analysis data with mineral compositions in the mass balance calculations, resulting in the final modes presented in Table 2. Although we cannot quantitatively assess the errors on the mode, the sum of the squares of the residuals on all the regressions are <1%, indicating very good fits to the compositional data. The slight differences in modes calculated by mass balance (vol % entries in Table 2) and image analysis are probably due to uncertainties in both the compositional analyses and image analysis, where pits and cracks in the sample result in unclassified pixels during thresholding of the BSE images. In addition, minor cracks, scratches and grain boundaries reflect lower intensities in BSE images, probably resulting in underestimation in the amount of brighter phases (e.g. biotite, garnet) and overestimation of quartz + plagioclase. |
Melting Behavior of Hp-60-1 During Progressive Anatexis |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
Melt fractions generated in our experiments are plotted as a function of temperature in Fig. 4. Based on how the modal abundances of mineral phases and glass in the run products vary with progressive melting (Fig. 5), we see evidence for the following general dehydration-melting reactions:
 |
|
|
The muscovite-dominated reaction is inferred, as muscovite was exhausted before the lowest temperature investigated (812°C). However, very little melt was present in the 812°C experiment, and the abundance of biotite was not significantly reduced from the amount in the starting rock. Therefore, we assumed that the changes in modal amounts of phases in the 812°C experiment from their amounts in the starting rock resulted primarily from dehydration-melting of muscovite. It should be noted that this approximation assumes that the mineral assemblage and mode of the starting material is appropriate for sub-solidus conditions at 10 kbar, the pressure of the experiments reported here.
The muscovite reaction is completed below 812°C and produces 1–2 wt % melt (estimated by eye). The biotite reaction takes place throughout the entire temperature interval investigated (160°C), producing an additional 31 wt % melt. Aluminosilicate appears to be a product in reaction (1) because its abundance in the 812°C run product is more than double that of the starting rock (4.4 and 1.5 wt %, respectively). However, the role of aluminosilicate is ambiguous in reaction (2). The calculated modal abundance of aluminosilicate (Table 2) remains above 3 wt % in all but one of the run products studied, but fluctuates slightly, neither decreasing nor increasing systematically.
Muscovite dehydration-melting reaction
At the lowest temperature investigated (812°C, 17 days), the experimental charge contains about 1–2 wt % melt and muscovite is absent. The abundances of quartz, plagioclase and biotite at this temperature are reduced from their values in the starting rock. Both aluminosilicate and alkali feldspar have increased considerably in abundance in the 812°C run product, and we see garnet in the 812°C run product, although it is absent in the starting rock.
Best fit weight fractions of residual crystalline phases are plotted vs melt fraction in Fig. 5. The slope of the trend for each phase in Fig. 5 signifies how the proportion of that phase changes in the residue with increasing melt fraction, with negative slopes indicating that a phase is a reactant. Thus multiplying the slope by –1 gives the reaction coefficients. Unfortunately, the 812°C run product contained very little melt, resulting in variable melt analyses that were subject to large standard deviations (Table 3). Including the average melt composition in the multiple linear regression resulted in an unreasonable mode. Therefore, we calculated the mode for the 812°C experiment without glass, and assumed 1.0 wt % glass for the determination of the muscovite-dominated melting reaction presented here. From this analysis, we infer that below 812°C the following incongruent dehydration-melting reaction occurs:
 |
Le Breton & Thompson (1988) discussed the controls on whether muscovite and biotite melt concurrently or in two stages as first muscovite and then biotite react. The steeper dP/dT slope of the biotite dehydration-melting curve implies an intersection at some pressure with the muscovite dehydration-melting curve (e.g. Clemens, 1984; Le Breton & Thompson, 1988). Above the pressure of this intersection, biotite and muscovite will melt concurrently by a reaction similar to (3). Although mass balance calculations require biotite to be a reactant in equation (3), we cannot discern whether muscovite and biotite were melting concurrently, or whether by 812°C the onset of biotite dehydration-melting at the higher-temperature biotite solidus has overprinted the muscovite reaction, which would be consistent with the P–T diagram presented by Le Breton & Thompson (1988) in their fig. 3.
In their model two-mica pelite BPQM, Gardien et al. (1995) also observed Kfs as a product of muscovite melting below 800°C at 10 kbar, but modal data imply biotite is not involved in the reaction. Gardien et al. (1995) did not observe the formation of garnet and aluminosilicate until 825°C, when the abundance of alkali feldspar and biotite began to decline. Patiño-Douce & Johnston (1991) presented a balanced reaction determined at 10 kbar, in which biotite contributes H2O to the melt, but alkali feldspar is not produced. In contrast, Vielzeuf & Holloway (1988) and Brearly & Rubie (1990) inferred biotite to be a product of the muscovite reaction under vapor-absent conditions at 10 and 1 kbar, respectively. Similarly, Patiño-Douce & Harris (1998) presented a balanced reaction determined at 6 kbar, in which both biotite and alkali feldspar are products.
The muscovite dehydration-melting reaction determined in this study differs from previous estimations in its strongly peritectic nature. Very little melt is produced as the majority of muscovite components go to produce alkali feldspar and aluminosilicate. This behavior severely limits the melt productivity of the muscovite dehydration reaction.
Biotite dehydration-melting reaction
Biotite is present in every run product, indicating that the biotite melting reaction spans a temperature interval of >160°C (812–975°C), with melt fraction increasing steadily with increasing temperature (Fig. 4). The abundance of alkali feldspar systematically decreases from 6.9 wt % in the 812°C run product to 2.9 wt % at 975°C. Once muscovite is exhausted, it appears that Kfs switches from product to reactant. The variation of mineral modes over this temperature interval is consistent with the continuous dehydration-melting reaction
 |
|
The role of alkali feldspar in dehydration-melting reactions has been extensively discussed (e.g. Vielzeuf & Clemens, 1992
; Carrington & Watt, 1995; Patiño-Douce & Beard, 1995). In this study, modal data suggest that alkali feldspar is a product of reaction (1), then switches to reactant in reaction (2) (Table 2, Fig. 5). Experimental studies by Carrington & Watt (1995) indicated that the H2O/K2O ratio of the biotite relative to the melt formed during H2O-undersaturated melting may control whether alkali feldspar is a reactant or a product during the melting reaction. Performing the analysis of Carrington & Watt (1995) on our experimental products suggests that alkali feldspar should be a product throughout both the muscovite and biotite melting intervals. The discrepancy between what the H2O/K2O ratios of the melt and biotite imply and what the modal data show probably arises from uncertainties in estimates of H2O content of the two phases, or uncertainties in melt alkali contents.
Results in comparison with previous experimental work
Partial melting of lower-crustal rocks with subsequent melt segregation and emplacement of high-level granitic plutons is considered a primary mechanism of crustal differentiation (Vielzeuf et al., 1990). Although large quantities of melt can be generated under H2O-saturated conditions (e.g. Piwinskii, 1973), the amount of H2O required to saturate granitic melts at high pressures (10 wt % at 10 kbar, e.g. Burnham, 1979) is not likely to be present in the lower crust (e.g. Clemens & Vielzeuf, 1987; Yardley & Valley, 1997). Furthermore, the negative dP/dT slope of the H2O-saturated solidus would prevent such melts from ascending without solidifying. Thus, mobile granitoid magmas emplaced as plutons in the mid to upper crust probably result from vapor-undersaturated dehydration-melting of assemblages containing hydrous phases.
To investigate the conditions under which crustal anatexis takes place, numerous workers have studied the vapor-absent melting behavior of a myriad of rock types. Relevant to this study, the dehydration-melting behavior of muscovite and biotite in metapelitic rocks has been investigated by many workers, including Thompson (1982), Le Breton & Thompson (1988), Vielzeuf & Holloway (1988), Peterson & Newton (1989), Holtz & Johannes (1991), Patiño-Douce & Johnston (1991), Gardien et al. (1995), and a study in the model system KFMASH performed by Carrington & Watt (1995). A discussion of our results referenced to the span of previous experimental work is beyond the scope of this paper. Therefore, in the following section, we compare our findings with a limited number of similar metapelite melting studies.
Melt productivity
The NKM pseudoternary projection in Fig. 7 includes HP-60-1 liquid compositions, as well as those reported by Vielzeuf & Holloway (1988), Patiño-Douce & Johnston (1991) and Gardien et al. (1995). Clustering together, the lowest temperature liquid compositions define a region (labeled M) comprising melt compositions generated at temperatures near the onset of biotite-dominated dehydration-melting. Patiño-Douce & Johnston (1991) noted that once melt of composition (M) has been generated, subsequent melting trends are governed by the bulk composition (hence the mode) of the protolith. Those workers further noted that the bulk composition chosen by Vielzeuf & Holloway (‘VH’) lies close to the intersection of the biotite–plagioclase join with the join between region (M) and the MgO + FeO vertex (the projected position of garnet), a geometric relationship that favors a quasi-isothermal melting step which generates a large amount of melt similar to (M). Vielzeuf & Holloway (1988) reported that their bulk composition generates >50 wt % melt over a 10–20°C temperature interval by the reaction Bt + Pl + Als + Qtz Melt + Grt. Biotite is exhausted in this reaction by 875°C, and subsequent melting of mafic phases results in a liquid trend emanating from region (M), heading for the MgO–FeO apex. In contrast, the bulk composition of Patiño-Douce & Johnston (‘PDJ’) exhausts plagioclase before biotite, which then melts over a wide temperature interval (150°C) by the reaction Bt + Qtz + Als Grt + Melt, generating liquid compositions projecting along the peritectic path R1. After exhausting biotite, liquid compositions then traverse the Grt + Qtz + Als liquidus field toward the PDJ bulk composition along path R2.
|
) and those of Vielzeuf & Holloway (1988)
), Patiño-Douce & Johnston (1991)
), after Patiño-Douce & Johnston (1991)In contrast to the above studies, the biotite melting reactions in HP-60-1 and in the bulk composition (BPQM) studied by Gardien et al. (1995)
do not exhaust any phases in the temperature interval investigated. Thus the melt compositions of HP-60-1 and BPQM cluster near the region (M) without departing on a peritectic or cotectic path, even at the highest temperatures studied. The variation in HP-60-1 melt compositions evident in Fig. 7 probably arises from the fact that at each temperature, melts approach equilibrium with residual phase compositions that have evolved slightly from the last experiment. The whole array of HP-60-1 melt compositions is offset from the region (M), defined by Patiño-Douce & Johnston (1991). This is undoubtedly due primarily to underestimates in Na2O contents of the glasses.
Differing from VH, bulk compositions HP-60-1, BPQM and PDJ display continuous melting behavior, with the biotite dehydration-melting reaction spanning a wide temperature interval (125–160°C). The refractory nature of biotite in these studies can be attributed to high concentrations of Ti and F (not analyzed in BPQM), which extend the thermal stability of biotite (Forbes & Flower, 1974; Munoz, 1984; Le Breton & Thompson, 1988; Peterson et al., 1991; Patiño-Douce, 1993; Skjerlie & Johnston, 1993; Dooley & Patiño-Douce, 1996). With a TiO2 content of 1.78 wt %, the starting composition of biotite in VH is considerably different from that of the biotite in PDJ, BPQM or HP-60-1 (2.68, 2.70 and 2.81 wt % TiO2, respectively). Furthermore, in a study of the fluid-absent melting of F-rich phlogopite, Dooley & Patiño-Douce (1996) determined that F and Ti in combination increase the thermal stability of mica to a much greater degree than either component independently. Fluorine contents of biotite in PDJ and HP-60-1 start at 0.31 and 0.28 wt % F, respectively, and increase with progressive melting. As Vielzeuf & Holloway (1988) did not analyze biotite for F, we infer that the biotite in their protolith was F poor, and the presence of Ti alone did not significantly enhance the thermal stability of biotite during their experiments.
The melt fraction produced with increasing temperature of HP-60-1 and similar pelite starting materials is shown in Fig. 4. The melt productivity of the starting materials shown directly reflects the control that H2O content of a protolith has on the amount of melt formed. The protolith studied by Vielzeuf & Holloway (1988) has the highest H2O content (2.15 wt %) and produces the most melt at any given temperature. Schist HP-60-1 contains the least amount of H2O (0.92 wt %) and consequently produces the least amount of melt at any given temperature. Also shown in Fig. 4 is the Clemens & Vielzeuf (1987) model melt productivity of fluid absent melting at 10 kbar of pelite containing 0.9 wt % H2O. This model curve represents the predicted amount of melt that would be formed if all the mica present in HP-60-1 were to completely break down at any given temperature along the curve. The experimentally determined melt productivity of HP-60-1 is lower than the model curve, undoubtedly because of the refractory nature of the biotite, which results in its continuous breakdown throughout the temperature interval investigated.
|
Discussion |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
Melt segregation
Segregation of melt from its source region is a necessary precursor to the emplacement of a granitic pluton. Separation of melt from residual phases is impeded by the high viscosities of granitic melts, and thus many proposed segregation mechanisms such as compaction (McKenzie, 1987
), and diapirism and stoping (e.g. Marsh, 1982; Wickham, 1987; Miller et al., 1988) were argued to require large melt fractions, above a rheological critical melt fraction (CMF) that marks a strength transition from a crystal–melt system acting more like a solid to one acting more like a liquid. Estimates for such a rheological CMF have ranged from 20 ± 10% melt (Arzi, 1978), to 30–35% melt (van der Molen & Paterson, 1979), and to 50% melt (Marsh, 1981; Miller et al., 1988). The requirement of such large melt fractions would preclude separation by these mechanisms of the low melt-fraction melts formed by muscovite dehydration-melting.
Recent experimental rock deformation studies have shown that extraction of viscous melt can occur at considerably lower melt fractions than the CMF. These studies indicate that melt segregation is controlled by several variables, including the type of mineral melting reaction, melt viscosity, and deformation rate (Rushmer, 1996). At very low melt fractions, dehydration-melting reactions associated with a positive 
V enhance cataclastic behavior by contributing to elevated pore pressures and the subsequent lowering of effective normal stresses (e.g. Hibbard & Watters, 1985; Clemens and Mawer, 1992; Connolly et al., 1997). Indeed, dynamic deformation experiments involving dehydration-melting reactions produce melt in cracks and localized shear zones, enhancing permeability (Hacker & Christie, 1990; Rushmer, 1995; Rutter & Neumann, 1995). The relatively low dihedral angles observed in hydrostatic partial melt studies of a variety of rock types (e.g. Vicenzi et al., 1988; Laporte 1994; Laporte & Watson, 1995; Lupulescu & Watson, 1995) indicate that melt interconnectivity can be achieved at low melt fractions in crustal rocks. Once permeability and melt interconnectivity are established, melt migration may be limited by high melt viscosities (Laporte & Watson, 1995). However, deformation produces pressure gradients which can enhance melt movement along a developing fracture network (Rushmer, 1996). Dynamic deformation experiments have shown that differential stress can enhance melt migration under a variety of strain rate conditions (Dell'Angelo & Tullis, 1988; Rutter & Neumann, 1995).
Theoretically, it seems possible that the melt produced by muscovite dehydration-melting of HP-60-1 could segregate to form a granitic intrusion. The positive V of the muscovite dehydration-melting reaction would probably initiate the development of melt-filled microcracks (e.g. Connolly et al., 1997). The demonstrated effect of H2O and F on reducing viscosities of peraluminous granitoid melts (e.g. Baker & Vaillancourt, 1995) would enhance melt mobility, reducing estimated time scales for melt segregation through a permeable fracture network by an order of magnitude (Rutter & Neumann, 1995). However, the estimated melt fraction of 1–2 wt % produced in the inferred muscovite reaction in this study is at the lower limit of melt fractions which could conceivably separate from its source region. Although such a low melt fraction is sufficient to establish melt interconnectivity (e.g. Wolf & Wyllie, 1995), separation of this small amount of melt from the source rock would require the source to drain to zero melt. Regardless of the segregation mechanism, the ability of the fraction of melt providing interconnectivity to drain completely from the source has not been demonstrated, and seems unlikely.
Depth of melting
We have tested the possibility that the high-18O granites represent melts formed by muscovite dehydration-melting at 10 kbar, which then intruded into shallower levels of the crust. The very low melt productivity of the muscovite reaction in HP-60-1 at 10 kbar suggests the above scenario is improbable, as the melts formed would not be likely to separate from the restite. Considering the range in estimated regional metamorphism of metasediments surrounding the HPG from 2.0–4.4 kbar (Helms & Labotka, 1991) to 5.3–7.6 kbar (Drops & Friberg, 1996), a pressure of 10 kbar is at least 3 kbar, and possibly as much as 7 kbar, greater than the pressure would have been had the tourmaline granites resulted from in situ melting at the level of intrusion. At lower pressures, the melt H2O content at any given temperature would be lower than in the 10 kbar melts presented here (e.g. Holtz & Johannes, 1994). Lower melt H2O content would result in more melt if an equal amount of muscovite breakdown occurred (Clemens & Vielzeuf, 1987). Given the typically low melt productivity of muscovite dehydration-melting reactions (<20%, Harris et al., 1995), melts generated at shallower depths are more likely to segregate to form granitic plutons, because of enhanced melt productivity. Helms & Labotka (1991) showed that the schist wallrocks evolved to above the Sil + Kfs isograd along the southwestern flank of the HPG complex. The fibrolitic clots in HP-60-1 may record incipient dehydration-melting, suggesting that the schist wallrocks underwent partial melting at the level of emplacement of the HPG. This inference supports a hypothesis that the high-18O granites are the product of in situ melting of the Proterozoic schist wallrocks.
Melt compositions compared with HPG suite: biotite vs tourmaline granites?
Peraluminous leucogranites frequently contain biotite or tourmaline as the dominant ferromagnesian mineral. Although these two minerals occasionally occur together in the same sample, they are more commonly mutually exclusive. Biotite- and tourmaline-bearing granites often form distinct coeval or genetically related plutons of a larger granite suite (e.g. Scaillet et al., 1990; Nabelek et al., 1992a, 1992b). In the latter case, cooling magmas would typically crystallize biotite first, and after extended fractionation, the B content of the melt would increase to a sufficient value for tourmaline saturation. This is consistent with observations in the Badrinath-Gangotri plutons of the high Himalayan leucogranites, in which biotite- and tourmaline-bearing granites appear to be related by crystal fractionation (Scaillet et al., 1990). In contrast, magmas forming peraluminous granitic rocks containing tourmaline only (no biotite) must crystallize tourmaline first. This seems to have been the case with the high-18O granites of the HPG, as these granites do not appear to be related to the biotite-bearing granites solely through crystal fractionation (Nabelek et al., 1992a; Krogstad et al., 1993).
We compare our melt compositions with those of the low- and high-18O HPG suites using the same ATF diagram (Fig. 8) as used by Nabelek et al. (1992a), who noted that this diagram effectively separates the low-18O biotite-bearing granites from the high-18O granites containing tourmaline. The melts generated in this study do not overlap with the tourmaline-bearing high-18O granites. Rather, melts of HP-60-1 fall within the field defined by the low-18O biotite-bearing granite suite, with higher-temperature melts plotting farther towards the T apex. Even the lowest fraction melt, inferred to be generated by muscovite dehydration-melting with only a small contribution from biotite [equation (1)], is enriched in TiO2 compared with the tourmaline-bearing granite suite. Correcting the known underestimates of the Na2O contents of HP-60-1 melts would not affect this interpretation as higher Na2O contents would reduce the calculated amount of excess alumina, shifting the HP-60-1 melts further from the A apex in Fig. 8.
|
, high-
, melts of HP-60-1. Star indicates lowest fraction melt of HP-60-1, inferred to be generated by muscovite dehydration-melting (812°C experiment). For clarity, we omitted a biotite-bearing granite sample which plotted directly beneath the star. Analyses of HPG granites are from Nabelek et al. (1992aThe ATF diagram shows that tourmaline rather than biotite crystallized in the granites with lower Ti content. It is well demonstrated that biotite is stabilized by high Ti concentrations (e.g. Forbes & Flower, 1974
; Munoz, 1984; Le Breton & Thompson, 1988). As the solubility of Ti is low in peraluminous granitic melts (Gwinn & Hess, 1989), elevated Ti contents in melts will have a strong effect on biotite stability. Alternatively, biotite may be less stable than other ferromagnesian minerals in rocks with low Ti concentrations (Scaillet et al., 1990; Nabelek et al., 1992a). Recent experimental work has shown that tourmaline is stable in peraluminous (ASI > 1.2) melts containing in excess of 2 wt % B2O3 (Wolf & London, 1997). We cannot assess the melts of HP-60-1 for this criterion because we did not analyze the melt for B. However, the experiments of Wolf & London (1997) were performed on relatively Ti-poor compositions. Concentration of Ti in leucogranitic melts may be as important as B in controlling the relative stability of biotite and tourmaline (Nabelek et al., 1992a). On the basis of this analysis, it is likely that upon cooling, the Ti-rich melts of HP-60-1 would crystallize biotite rather than tourmaline.
Melt compositions generated in muscovite and biotite dehydration-melting reactions in HP-60-1 compare well in some major oxides with published compositions of tourmaline-bearing granites of the Harney Peak suite (Nabelek et al., 1992a). On a normalized anhydrous basis, the average Al2O3 and SiO2 contents for HP-60-1 melts are 15.15 and 74.15 wt %, respectively, compared with average Al2O3 and SiO2 contents of 15.46 and 74.85 wt % for the tourmaline granites. However, melts of HP-60-1 are considerably more mafic than the tourmaline-bearing granites, with average combined TiO2 + MgO + FeO contents of 1.85 and 0.84 wt %, respectively. The remarkably low mafic component contents of rocks from the tourmaline-bearing granite suite argue that these rocks represent melts that segregated very efficiently from their source region, with little or no entrained restite. Even if they segregated from the solid residue with 100% efficiency, melts from HP-60-1 would be more mafic than rocks of the tourmaline granite suite.
|
Conclusions |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
We have tested the possibility that the high-
18O, tourmaline-bearing granites of the Harney Peak suite were generated by muscovite-dominated dehydration-melting at 10 kbar of Proterozoic schist HP-60-1. The very low melt productivity of the muscovite dehydration-melting reaction of HP-60-1, coupled with the relatively high Ti content of the melts produced, indicates it is unlikely that any of the tourmaline-bearing granites of the Harney Peak suite resulted from partial melting of a metasedimentary source like HP-60-1. Although the positive V of the fluid-absent melting reaction and the possibility of reduced melt viscosity because of the presence of F would both contribute to the likelihood of melt segregation, the very low melt fraction of 1–2 wt % in our lowest temperature (812°C) experiment is on the order of the amount of melt needed just to establish interconnectivity. Melting HP-60-1 at lower pressures would probably result in greater melt productivity, but given the small amount of melt generated at 10 kbar, even a considerable increase in melt fraction would not boost the melt productivity of HP-60-1 to the level where segregation would be expected.
Irrespective of melt productivity, the schist HP-60-1 is too enriched in Ti to be a good candidate for the source rock for the tourmaline granites of the HPG. The presence of Ti increases the thermal stability of biotite, rendering the schist very refractory. In addition, the melts generated are correspondingly enriched in Ti, and upon cooling would probably stabilize biotite rather than tourmaline (Nabelek et al., 1992a). Furthermore, melts of HP-60-1 are in general more mafic than the high-18O granites.
The diminishing melt productivity at higher pressures resulting from increased H2O solubility (e.g. Clemens & Vielzeuf, 1987) renders it increasingly difficult for melts generated at greater depths to segregate. This suggests that the first hypothesis of Nabelek et al. (1992a), that heat from the intrusion of the core biotite granite induced muscovite-dominated dehydration-melting of the surrounding schist wallrocks at the level of emplacement of the HPG, is more probable. Based on our observations, we infer that a good candidate for a source rock for the tourmaline-bearing granites would be a high-18O schist, rich in muscovite, with a bulk composition rich in alkali elements and B, but poor in Ti.
|
Appendix |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
|
|
Acknowledgements |
|---|
We thank Peter Nabelek for supplying the starting material. We gratefully acknowledge Michael Shaffer's considerable help with the microprobe analyses and SEM BSE image acquisition. T. Rushmer, M. Wolf, A. Patiño-Douce and G. Stevens provided constructive reviews which greatly improved the manuscript. The Electron Microprobe Facility was funded by NSF Grant EAR-8803960 with a matching grant from the W. M. Keck Foundation. This work was supported by NSF Grant EAR-9506045.
* Corresponding author. Fax: 541-346-4692. e-mail: pick{at}darkwing.uoregon.edu
|
REFERENCES |
|---|
|
TOPABSTRACTIntroductionCharacterization of the Starting...Analytical Procedures and Mode...Melting Behavior of Hp-60-1...DiscussionConclusionsAppendixREFERENCES |
|---|
Arzi A. A. Critical phenomena in the rheology of partially melted rocks. Tectonophysics (1978) 44:173–184.[Web of Science]
Baker D. R., Vaillancourt J. The low viscosities of F + H2O-bearing granitic melts and implications for melt extraction and transport. Earth and Planetary Science Letters (1995) 132:199–211.[Web of Science]
Bohlen S. R. Equilibria for precise pressure calibration and a frictionless furnace assembly for the piston-cylinder apparatus. Neues Jahrbuch für Mineralogie, Monatshefte (1984) H9:404–412.
Brearly A. J., Rubie D. C. Effects of H2O on the disequilibrium breakdown of muscovite + quartz. Journal of Petrology (1990) 31:925–956.
Burnham C. W. The importance of volatile constituents. In: The Evolution of the Igneous Rocks: Fiftieth Anniversary Perspectives—Yoder H. S., ed. (1979) Princeton, NJ: Princeton University Press. 439–482.
Carrington D. P., Watt G. R. A geochemical and experimental study of the role of K-feldspar during water-undersaturated melting of metapelites. Chemical Geology (1995) 122:59–76.[Web of Science]
Clemens J. D. Water contents of silicic to intermediate magmas. Lithos (1984) 17:273–287.[Web of Science]
Clemens J. D., Mawer C. K. Granitic magma transport by fracture propagation. Tectonophysics (1992) 204:339–360.[Web of Science]
Clemens J. D., Vielzeuf D. Constraints on melting and magma production in the crust. Earth and Planetary Science Letters (1987) 86:287–306.[Web of Science]
Connolly J. A. D., Holness M. B., Rubie D. C., Rushmer T. Reaction-induced microcracking: an experimental investigation of a mechanism for enhancing anatectic melt extraction. Geology (1997) 25:591–594.
Dell'Angelo L. N., Tullis J. Experimental deformation of partially melted granitic aggregates. Journal of Metamorphic Geology (1988) 6:495–515.[Web of Science]
Dooley D. F., Patiño-Douce A. E. Fluid-absent melting of F-rich phlogopite + rutile + quartz. American Mineralogist (1996) 81:202–212.[Abstract]
Drops M. T., Friberg L. M. Comparative geobarometry of amphibolites from the west-central section of the Black Hills, South Dakota. Geological Society of America, Abstracts with Programs (1996) 28(4):6.
Duke E. F., Redden J. A., Papike J. J. Calamity Peak layered granite-pegmatite complex, Black Hills, South Dakota: Part I. Structure and emplacement. Geological Society of America Bulletin (1988) 100:825–840.
Duke E. F., Shearer C. K., Redden J. A., Papike J. J. Proterozoic granite-pegmatite magmatism Black Hills South Dakota: structure and geochemical zonation. In: The Early Proterozoic Trans-Hudson Orogen of North America. Geologic Association of Canada Special Paper—Lewry J. F., Stauffer M. R., eds. (1990) 37:253–269.
Eugster H. P., Wright T. L. Synthetic hydrous boron micas. US Geological Survey Professional Paper (1960) 400B:B441–B442.
Forbes W. C., Flower M. F. J. Phase relations of titan-phlogopite, K2Mg4TiAl2Si6O20(OH)4: a refractory phase in the upper mantle? Earth and Planetary Science Letters (1974) 22:60–66.[Web of Science]
Friberg L. M., Dahl P. S., Terry M. P. Thermometric evolution of early Proterozoic metamorphic rocks from the southern Black Hills, South Dakota. Guidebook Geology of the Black Hills, South Dakota. South Dakota School of Mines Technical Bulletin (1996) 19:191–199.
Gardien V., Thompson A. B., Grujic D., Ulmer P. Experimental melting of biotite + plagioclase + quartz ± muscovite assemblages and implications for crustal melting. Journal of Geophysical Research (1995) 100:15581–15591.
Gwinn R., Hess P. C. Iron and titanium solution properties in peraluminous and peralkaline rhyolitic liquids. Contributions to Mineralogy and Petrology (1989) 101:326–338.[Web of Science]
Hacker B., Christie J. M. Brittle/ductile and plastic/cataclastic transitions in experimentally deformed and metamorphosed amphibolite. In: Brittle–Ductile Transitions. The Heard Volume. Geophysical Monograph, American Geophysical Union—Duba A., Durham W., Handin J., Wang H., eds. (1990) 56:127–147.
Harris N., Ayres M., Massey J. Geochemistry of granitic melts produced during the incongruent melting of muscovite: implications for the extraction of Himalayan leucogranitic magmas. Journal of Geophysical Research (1995) 100:15767–15777.
Helms T. S., Labotka T. C. Petrogenesis of early Proterozoic pelitic schists of the southern Black Hills, South Dakota: constraints on regional low-pressure metamorphism. Geological Society of America Bulletin (1991) 103:1324–1334.
Hibbard M. J., Watters R. J. Fracturing and diking in incompletely crystallized granitic plutons. Lithos (1985) 18:1–12.[Web of Science]
Holtz F., Johannes W. Genesis of peraluminous granites. I. Experimental investigation of melt compositions at 3 and 5 kb and various H2O activities. Journal of Petrology (1991) 32:935–958.
Holtz F., Johannes W. Maximum and minimum water contents of granitic melts: implications for chemical and physical properties of ascending magmas. Lithos (1994) 32:149–159.[Web of Science]
Kretz R. Symbols for rock forming minerals. American Mineralogist (1983) 68:277–279.[Abstract]
Krogstad E. J., Walker R. J., Nabelek P. I., Russ-Nabelek C. Lead isotopic evidence for mixed sources of Proterozoic granites and pegmatites, Black Hills, South Dakota, USA. Geochimica et Cosmochimica Acta (1993) 57:4677–4685.[Web of Science]
Laporte D. Wetting behavior of partial melts during crustal anatexis: the distribution of hydrous silicic melts in polycrystalline aggregates of quartz. Contributions to Mineralogy and Petrology (1994) 116:486–499.[Web of Science]
Laporte D., Watson E. B. Experimental and theoretical constraints on melt distribution in crustal sources: the effect of crystalline anisotropy on melt interconnectivity. Chemical Geology (1995) 124:161–184.[Web of Science]
Le Breton N., Thompson A. B. Fluid-absent (dehydration) melting of biotite in metapelites in the early stages of crustal anatexis. Contributions to Mineralogy and Petrology (1988) 99:226–237.[Web of Science]
Lupulescu A., Watson E. B. Tonalitic melt connectivity at low melt fraction in a mafic crustal protolith. Eos Transactions, American Geophysical Union (1995) 76:299–300.
Marsh B. D. On the crystallinity, probability of occurrence and rheology of lava and magma. Contributions to Mineralogy and Petrology (1981) 78:85–98.[Web of Science]
Marsh B. D. On the mechanics of igneous diapirism, stoping, and zone melting. American Journal of Science (1982) 282:808–855.
McKenzie D. P. The compaction of igneous and sedimentary rocks. Journal of the Geological Society, London (1987) 144:299–307.
Miller C. F., Watson E. B., Harrison T. M. Perspectives on the source, segregation and transport of granitoid magmas. Transactions of the Royal Society of Edinburgh: Earth Sciences (1988) 79:135–156.
Morgan G. B. VI, London D. Optimizing the electron microprobe analysis of hydrous alkali aluminosilicate glasses. American Mineralogist (1996) 81:1176–1185.[Abstract]
Munoz J. L. F–OH and Cl–OH exchange in micas with applications to hydrothermal ore deposits. Mineralogical Society of America, Reviews in Mineralogy (1984) 13:469–491.[Abstract]
Nabelek P. I., Ternes K. Fluid inclusions in the Harney Peak Granite, Black Hills, South Dakota, USA: implications for solubility and evolution of magmatic volatiles and crystallization of leucogranitic magmas. Geochimica et Cosmochimica Acta (1997) 61:1447–1465.[Web of Science]
Nabelek P. I., Russ-Nabelek C., Denison J. R. The generation and crystallization conditions of the Proterozoic Harney Peak Leucogranite, Black Hills, South Dakota, USA: petrologic and geochemical constraints. Contributions to Mineralogy and Petrology (1992a) 110:173–191.[Web of Science]
Nabelek P. I., Russ-Nabelek C., Haeussler G. T. Stable isotope evidence for the petrogenesis and fluid evolution in the Proterozoic Harney Peak leucogranite, Black Hills, South Dakota. Geochimica et Cosmochimica Acta (1992b) 56:403–417.[Web of Science]
Norton J. J., Redden J. A. Relations of zoned pegmatites to other pegmatites, granite, and metamorphic rocks in the southern Black Hills, South Dakota. American Mineralogist (1990) 75:631–655.[Abstract]
Patiño-Douce A. E. Titanium substitution in biotite: an empirical model with applications to thermometry, O2 and H2O barometries, and consequences for biotite stability. Chemical Geology (1993) 108:133–162.[Web of Science]
Patiño-Douce A. E., Beard J. S. H2O loss from hydrous melts during fluid-absent piston cylinder experiments. American Mineralogist (1994) 79:585–588.[Web of Science]
Patiño-Douce A. E., Beard J. S. Dehydration-melting of biotite gneiss and quartz amphibolite from 3 to 15 kbar. Journal of Petrology (1995) 36:707–738.
Patiño-Douce A. E., Beard J. S. Effects of Pf (O2) and Mg/Fe ratio on dehydration melting of model metagreywackes. Journal of Petrology (1996) 37:999–1024.
Patiño-Douce A. E., Harris N. Experimental constraints on Himalayan anatexis. Journal of Petrology (1998) 39:689–710.[Web of Science]
Patiño-Douce A. E., Johnston A. D. Phase equilibria and melt productivity in the pelitic system: implications for the origin of peraluminous granitoids and aluminous granulites. Contributions to Mineralogy and Petrology (1991) 107:202–218.[Web of Science]
Patiño-Douce A. E., Johnston A. D., Rice J. M. Octahedral excess mixing properties in biotite: a working model with applications to geobarometry and geothermometry. American Mineralogist (1993) 78:113–131.[Abstract]
Peterson J. W., Newton R. C. Reversed experiments on biotite–quartz–feldspar melting in the system KMASH: implications for crustal anatexis. Journal of Geology (1989) 97:465–486.[Web of Science]
Peterson J. W., Chacko T., Kuehner S. M. The effects of fluorine on the vapor-absent melting of phlogopite + quartz: implications for deep-crustal processes. American Mineralogist (1991) 76:470–476.[Abstract]
Pickering J. M., Schwab B. E., Johnston A. D. Off-center hot spots: double thermocouple determination of the thermal gradient in a 1.27 cm (1/2 in.) CaF2 piston-cylinder furnace assembly. American Mineralogist (1998) 83:228–235.[Abstract]
Piwinskii A. J. Experimental studies of igneous rocks series, Central Sierra Nevada Batholith, California, Part II. Neues Jahrbuch für Mineralogie, Monatshefte (1973) H5:193–215.
Redden J. A., Norton J. J., McLaughlin R. J. Geology of the Harney Peak Granite, Black Hills, South Dakota. In: Geology of the Black Hills, South Dakota and Wyoming—Rich F. J., ed. (1985) 2nd edn. Reston, VA: American Geological Institute. 225–240.
Riley G. H. Isotopic discrepancies in zoned pegmatites, Black Hills, South Dakota. Geochimica et Cosmochimica Acta (1970) 34:713–725.[Web of Science]
Rushmer T. An experimental deformation study of partially molten amphibolite: application to low-fraction melt segregation. Journal of Geophysical Research (1995) 100:15681–15695.
Rushmer T. Melt segregation in the lower crust: how have experiments helped us? In: The Third Hutton Symposium on the Origin of Granites and Related Rocks. Geological Society of America, Special Paper—Brown M., Candela P. A., Peck D. L., Stephens W. E., Walker R. J., Zen E-an, eds. (1996) 315:73–83.
Rutter E. H., Neumann D. H. K. Experimental deformation of partially molten Westerly granite under fluid-absent conditions, with implications for the extraction of granitic magmas. Journal of Geophysical Research (1995) 100:15697–15715.
Scaillet B., France-Lanord C., Le Fort P. Badrinath–Gangotri plutons (Garhwal, India): petrological and geochemical evidence for fractionation processes in a high Himalayan leucogranite. Journal of Volcanolgy and Geothermal Research (1990) 44:163–188.
Shearer C. K., Papike J. J. Distribution of boron in the Tip Top pegmatite, Black Hills, South Dakota. Geology (1986) 14:119–123.
Skjerlie K. P., Johnston A. D. Fluid-absent melting behavior of an F-rich tonalitic gneiss at mid-crustal pressures: implications for the generation of anorogenic granites. Journal of Petrology (1993) 34:785–815.
Stubican V., Roy R. Boron substitution in synthetic micas and clays. American Mineralogist (1962) 47:1166–1173.[Web of Science]
Thompson A. B. Calculation of muscovite–paragonite–alkali feldspar phase relations. Contributions to Mineralogy and Petrology (1974) 44:173–194.[Web of Science]
Thompson A. B. Dehydration melting of pelitic rocks and the generation of H2O-undersaturated granitic liquids. American Journal of Science (1982) 282:1567–1595.
van der Molen I., Paterson M. S. Experimental deformation of partially-melted granite. Contributions to Mineralogy and Petrology (1979) 70:299–318.[Web of Science]
Vicenzi E. P., Rapp R. P., Watson E. B. Crystal/melt wetting characteristics in partially molten amphibolite. Eos Transactions, American Geophysical Union (1988) 69:482.
Vielzeuf D., Clemens J. D. The fluid-absent melting of phlogopite + quartz: experiments and models. American Mineralogist (1992) 77:1206–1222.[Abstract]
Vielzeuf D., Clemens J. D., Pin C., Moinet E. Granites granulites and crustal differentiation. In: Granulites and Crustal Evolution—Vielzeuf D., Vidal P., eds. (1990) Dordrecht: Kluwer. 59–85.
Vielzeuf D., Holloway J. R. Experimental determination of the fluid-absent melting relations in the pelitic system. Contributions to Mineralogy and Petrology (1988) 98:257–276.[Web of Science]
Wickham S. M. The segregation and emplacement of granitic magmas. Journal of the Geological Society, London (1987) 144:281–297.
Wolf M. B., London D. Boron in granitic magmas: stability of tourmaline in equilibrium with biotite and cordierite. Contributions to Mineralogy and Petrology (1997) 130:12–30.[Web of Science]
Wolf M. B., Wyllie P. J. Liquid segregation parameters from amphibolite dehydration melting experiments. Journal of Geophysical Research (1995) 100:15611–15621.
Yardley B. W. D., Valley J. W. The petrologic case for a dry lower crust. Journal of Geophysical Research (1997) 102:12173–12185.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Acosta-Vigil, I. Buick, J. Hermann, B. Cesare, D. Rubatto, D. London, and G. B. Morgan VI Mechanisms of Crustal Anatexis: a Geochemical Study of Partially Melted Metapelitic Enclaves and Host Dacite, SE Spain J. Petrology, February 6, 2010; (2010) egp095v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.D. Clemens The message in the bottle: "Melt" inclusions in migmatitic garnets Geology, July 1, 2009; 37(7): 671 - 672. [Full Text] [PDF]
|
|
|
|
|
|
F. Ferri, S. Poli, and D. Vielzeuf An Experimental Determination of the Effect of Bulk Composition on Phase Relationships in Metasediments at Near-solidus Conditions J. Petrology, May 8, 2009; (2009) egp026v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Stevens, A. Villaros, and J.-F. Moyen Selective peritectic garnet entrainment as the origin of geochemical diversity in S-type granites Geology, January 1, 2007; 35(1): 9 - 12. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||