Granitic Liquids Research Articles

The petrography and geochemistry of the Clericy Pluton: an ultrapotassic pyroxenite-syenite suite of late Archaean age from the Abitibi region, Quebec

The Archaean Clericy Pluton is a small, bimodal suite of ultramafic and syenitic rocks located in the Abitibi region of Quebec. The syenite unit consists predominantly of alkali feldspar and (aegirine)-augite and contains numerous elongated mafic xenoliths. The ultramafic component is predominantly clinopyroxenite. Using the geochemical screen proposed by Foley et al. (1987) the mafic to intermediate members of the suite (53.8% < SiO 2 < 60.9%) are shown to possess an ultrapotassic character. Crystallization, mainly of clinopyroxene and alkali feldspar, explains the transition from ultramafic to syenitic compositions. Magma mixing between a minimum melt granitic liquid and syenite is required to explain the most evolved compositions. The Clericy rocks were probably emplaced in a small release zone which developed along one of a number of major transcurrent faults which formed as a consequence of oblique subduction at the end of the Archaean era.

On the genesis of Archean granite through two-stage melting of the Quetico accretionary prism at a transpressional plate boundary

Regional geologic relationships and geochemical data support the idea that late-stage biotite granite of the Lac La Croix type may have formed by partial melting of previously migmatized metasedimentary rocks in the Quetico subprovince of the Superior craton. The proposed parent material is an intimately layered rock composed of approximately half to two-thirds biotite schist and half to one-third granitoid leucosome; the bulk chemical attributes of such a material were estimated from detailed mapping and drilling of two large outcrops near Pearl Lake, Minnesota. Previous work has demonstrated that the granitoid leucosome of the layered migmatite is the intrusively injected product of partial melting of a metasedimentary protolith. Therefore the proposed derivation of the younger Lac La Croix-type granite from layered migmatite in effect represents a second distillation of granitic liquid from meta-sedimentary rocks that had been enriched ingranitic components in the first cycle of partial melting and melt migration. It is postulated that the implied two-stage melting of metasedimentary rocks was related to thermal consequences within the Quetico accretionary complex as tectonism along the Quetico-Wabigoon boundary alternated between subduction and transform kinematics during Late Archean time.

Mid-Proterozoic lamprophyre commingled with late-stage granitic dikes of the anorogenic San Isabel batholith, Wet Mountains, Colorado

More than a dozen composite mafic-granitic dikes crop out along the northern margin of the San Isabel batholith, Wet Mountains, Colorado. The dikes crosscut and contain xenoliths of 1441 Ma granitic dikes and are in turn intruded by, but probably are coeval with, late-stage granitic dikes of the 1360 Ma San Isabel batholith. Only two other mafic representatives of the mid-Proterozoic anorogenic suite are known from Colorado. In the field, the dikes contain up to one-third granitic material by volume. Textures range from a swirling, lacy, mortarlike granite within basalt to subovoid pillows of basalt in granite. Many of the pillows contain small ovoids of the granite. The mutual inclusions and rounded to cuspate liquidlike contacts with minor mechanical mixing strongly indicate that these dikes formed through commingling of mafic and granitic liquids. The mafic dikes have the mineralogy and chemistry of lamprophyres, containing andesine, amphibole, biotite, sphene, and magnetite, with minor quartz and epidote and rare K-feldspar. Chemically, they average average 52% SiO2, and they are enriched in incompatible elements including the light rare earth elements. They are tholeiitic, compositionally similar to within-plate basalts with rift affinity. They partially mixed with the granite to produce intermediate-composition rock with 56% SiO2. The granite is a monzogranite with biotite, hornblende, and sphene. Its mineralogy and chemistry are characteristic of the San Isabel batholith. The lamprophyres may be related to mantle-derived mafic liquids that were hypothesized to have melted lower crust to produce the San Isabel batholith. The lamprophyres may have migrated up around the margin of the batholith and mixed with late-stage granite to form these composite dikes.

The geology and structural setting of the Proterozoic Ammassalik Intrusive Complex, East Greenland

The rocks of the Ammassalik area comprise reworked Archaean gneisses and a variety of mainly pelitic to psammitic Proterozoic metasediments. Proterozoic deformation caused thrust intercalation of the rocks and folded them into overturned, southerly directed nappes. Three dioritic centres, collectively named the 'Ammassalik Intrusive Complex' in this paper, were emplaced into a tract occupied by metasediments, causing extensive anatexis. The granitic liquids produced from this resulted in magma mixing with the still liquid diorite; they also formed discrete sheets of augen granite intruding the country rocks. The diorite centres were emplaced as relatively dry bodies, and together with anatexis, they caused the formation of granulite facies assemblages by dehydration reactions in the country rocks. Further away from the contact zone garnet appeared in the surrounding grey gneisses.

CO2-enhanced melting of biotite-bearing rocks at deep-crustal pressure–temperature conditions

THE role of CO2 in melting of the Earth's crust has been a topic of considerable debate. Experimental studies in quartz-feldspar systems1,2 have shown that CO2 solubility in simple granitic liquids is less than 1 wt% at pressures below 10 kbar and that dilution of H2O vapour with CO2 raises the vapour-saturated solidi3, indicating that CO2 inhibits, rather than promotes, crustal melting. On the other hand, from experiments on the system K2O–MgO–Al2O3–SiO2–H2O–CO2, Wendlandt has inferred significant solubility of CO2 in crustal melts4. Wendlandt observed that biotite–quartz–feldspar assemblages melt to orthopyroxene + liquid at temperatures as low as 740 °C in the presence of CO2-rich fluids, more than 50 °C lower than equivalent orthopyroxene-producing reactions in the pure-H2O systems5. This surprising result implies a fundamental melt interaction between MgO and CO2, because neither component alone has significant solubility in granitic liquids near the H2O-saturated solidus6. In view of the potentially important consequences of these results for deep-crustal processes, additional data are needed. Here we report the results of experimental melting studies on synthetic biotite–quartz and biotite–quartz–sanidine assemblages. In general agreement with Wendlandt's results, we find that these assemblages melt in the presence of CO2-rich vapour below 760 °C at pressures <10 kbar, conditions believed to be encountered in high-grade crustal metamorphism. This opens up the possibility that melting of the deep crust may occur in the presence of CO2-rich fluids.

Experimental determination of the fluid-absent melting relations in the pelitic system

In order to provide additional constraints on models for partial melting of common metasediments, we have studied experimentally the melting of a natural metapelite under fluid-absent conditions. The starting composition contains quartz, plagioclase, biotite, muscovite, garnet, staurolite, and kyanite. Experiments were done in a halfinch piston-cylinder apparatus at 7, 10, and 12 kbar and at temperatures ranging from 750° to 1250° C. The following reactions account for the mineralogical changes observed at 10 kbar between 750° and 1250° C: Bi+Als+Pl+Q=L+Gt+(Kf), Ky=Sill, Gt+Als=Sp+Q, Gt=L+Sp+Q, and Sp+Q=L+Als. The compositions of the phases (at T>875° C) were determined using an energy-dispersive system on a scanning electron microscope. The relative proportions of melt and crystals were calculated by mass balance and by processing images from the SEM. These constraints, together with other available experimental data, are used to propose a series of P-T, T-XH2O, and liquidus diagrams which represent a model for the fluid-present and fluid-absent melting of metapelites in the range 2–20 kbar and 600°–1250° C. We demonstrate that, even under fluid-absent conditions, a large proportion (≈40%) of S-type granitic liquid is produced within a narrow temperature range (850°–875° C), as a result of the reaction Bi+Als+Pl+Q=L+Gt(+/-Kf). Such liquids, or at least some proportion of them, are likely to segregate from the source, leaving behind a residue composed of quartz, garnet, sillimanite, plagioclase, representing a characteristic assemblage of aluminous granulites. The production of a large amount of melt at around 850° C also has the important effect of buffering the temperature of metamorphism. In a restitic, recycled, lower crust undergoing further metamorphism, temperature may reach values close to 1000° C due to the absence of this buffering effect. Partial melting is the main process leading to intracontinental differentiation. We discuss the crustal cross-section exposed in the North Pyrenean Zone in the context of our experiments and modelling.

The segregation and emplacement of granitic magmas

The segregation of granitic magma from residual crystals at low melt-fraction is strongly dependent on the viscosity of the melt. Theoretical considerations imply that for the typical range of granitic melt viscosities (10 4 Pas to 10 11 Pas) only very limited separation will be possible by a compaction mechanism over the typical duration of a crustal melting event ( c . 10 6 years). Small-scale segregations (millimetre to metre) of the type observed in migmatite terranes may be generated by compaction (possibly assisted by continuous deformation), or by flow of melt into extensional fractures, but low melt-fraction liquids are unlikely to be extracted to form large (kilometre-size) granitic plutons because of the limited separation efficiency. At higher melt-fractions (&gt;30%) the rapid decrease in strength and effective viscosity during partial melting allows other segregation processes to operate. Calculations and experiments indicate that in granitic systems the effective viscosity of partially melted rocks, having a very narrow melt fraction range of 30–50% will fall rapidly to levels at which convective overturn of kilometre-thick zones can occur. Convective motion within anatectic regions is capable of generating large (kilometre-size) homogeneous, high crystal-fraction, crustally-derived magma bodies, which are orders of magnitude greater in size than low melt-fraction segregates. Before convective instability is reached, small (centimetre- to metre-sized) pods of granitic liquid may rise buoyantly through, and pond at the top of such partly molten zones; such a process is consistent with the observation that some granulites appear to be residue rocks, chemically depleted in a minimum melt component. The effective viscosity (and hence the susceptibility to convection) of a partially melted zone within the crust, is strongly dependent on the water content of the system at a given pressure and temperature, because this controls both the quantity of melt generated and also the viscosity of the melt. The intrinsic water content of most crustal lithologies is incapable of promoting the high percentages of partial melting, or the low liquid viscosities, required to form large kilometre-sized granitic plutons by convective homogenization, at typical crustal temperatures. This suggests that the anatexis involved in the generation of large crustally-derived magma bodies has in many cases been promoted by an influx of externally derived aqueous fluid. These magma segregation processes are illustrated with respect to the petrogenesis of three different types of granitoid pluton from a Hercynian low-pressure, metamorphic-anatectic terrane in the Pyrenees.

Monzo-anorthosite from the Tague� ring-complex, A�r, Niger: a hybrid rock with cumulus plagioclase and an infiltrated granitic intercumulus liquid?

An unusual hybrid rock composed of cumulus plagioclase with the interstices occupied by abundant quartzo-feldspathic material occurs in the Taguei ring complex, Niger. The monzo-anorthosite consists of about 75% plagioclase in large tabular to elongate crystals with interstices occupied in two stages, firstly by clinopyroxene, titanomagnetite and apatite and then by an intergrowth of quartz and alkali feldspar associated with brown-green amphibole and zircon. Secondary green amphibole, chlorite, epidote and calcite may occur. Four stages in the crystallization history were identified: (1) Cumulus stage represented by the cores of the plagioclase laths (An56-An66 with reversed oscillatory zoning) and rare clinopyroxene (T ∼1,150° C), (2) Early intercumulus stage with a wide overgrowth zone on plagioclase (An62-An15 with oscillatory zoning and increase in Or), clinopyroxene, apatite, titanomagnetite (T ∼1,150-1,050° C), (3) Late intercumulus stage with alkali feldspar, quartz, opaque oxide, brown-green amphibole, apatite and zircon (T ∼750-700° C). Alkali feldspar gave cryptoperthite on further cooling. (4) Deuteric stage with development of turbidity in the alkali feldspar and plagioclase rims, and formation of patch perthite with microcline and secondary minerals (T ∼400° C).

Comportements différenciés de W, Sn, U, Ta, Nb dans quelques complexes granitiques du Massif Central français

New data are presented on the W-, Sn-, Ta-, Nb-, and U-geochemistry of some granitic complexes from the French Massif Central ; the data are interpreted in terms of fractional crystallization ; the behaviour of the elements appears to be similar to that of the hygromagmaphile or incompatible elements in basaltic volcanic rocks, with straight correlations with quite identical slopes in bilogarithmic diagrams. These observations are interpreted as suggesting the existence of element-stabilizing complexes in silicate melts. A model is proposed for the extraction of metals from such magmatic ore-preconcentrations ; as a consequence of this model, ore-generation should be favoured by an important phase of magmatic evolution, up to a certain level, which is strongly dependent upon the degree of incompatibility or hygromagmaphility of the element ; the model gives a possible explanation of W/Ta relationships recorded by the most evolved granitic liquids.

Geochemical aspects of granite and related rocks

Report of a meeting held in Burlington House on 28 November 1985. The meeting was convened by K. Coe, and chaired by W. S. Pitcher and M. Brown, morning and afternoon sessions respectively. A special scientific meeting of the Society was held on Thursday 28 November 1985, with the title Geochemical aspects aspects of granite and related rocks. From the papers offered, ten were selected which dealt with data and concepts derived from British and Irish granitic bodies. One exception to this was the first contribution, presented by R. Macdonald. In this it was proposed that because of late crystallization and subsolidus processes, plutonic granites may not record magmatic chemistry. Macdonald and his co-workers suggested that silicic obsidians, rhyolites and acid ash-flow sheets may be better indicators of original composition, implying that these extruded rocks are the chemical equivalents of unmodified granitic liquids. This precept was used in interpreting data derived mainly from the Western Cordillera to show that compositions of source rocks were in some cases more important than the tectonic setting of the magma. The second paper, by D. A. C. Manning , emphasized the importance of the volatiles B, F, Li and P in granites, manifest in familiar ‘accessory’ minerals and exerting controls on various aspects of the melt, including its emplacement mechanism. Other trace and some major element abundances correlated with these indicating a genetic relationship. A lower crust source for the volatiles was suggested. Two papers followed on the younger and older granites of the

The role of contamination in the evolution of the Platreef of the Bushveld Complex

The Platreef is a thick (greater than 100 m) coarse-grained feldspathic orthopyroxenite which transgresses northward across metasediments of the Transvaal Sequence and onto a granitic floor. It is overlain by main zone gabbronorites. Vertical profiles for Rb and Sr isotope ratios through the Platreef and overlying main zone show differences depending upon footwall type. Against a granitic floor the Platreef has initial 87 Sr/ 86 Sr ratios ranging from 0.7107 to 0.7226 (calculated at 2.05 b.y.) suggesting heterogeneous contamination by a melt derived from the granite. Overlying a dolomitic footwall, the Platreef shows lower initial ratios (0.7054 to 0.7147), but high Rb contents again suggest addition of a granitic component. The extent of contamination is least at the base. The role of dolomite assimilation is difficult to quantify. The lowermost rocks of the main zone show high Rb contents, due to contamination, but their initial Sr isotope ratios have not been changed significantly, due to the high proportion of cumulus plagioclase. Orthopyroxene has a much lower initial ratio than its whole rock, indicating that the main cumulus phase formed prior to contamination and that the granitic liquid reacted with the interstitial basic magma to produce the contaminated intercumulus phases.The tenor of platinum-group elements in the Platreef precludes the formation of immiscible sulfide liquid from an interstitial liquid; rather, it must be an early formed phase. Hence the separation of sulfide liquid may predate the main siliceous contamination process.

Genesis of chrysoberyl in the pegmatites of southern Kerala, India

AbstractIn southern Kerala, India, chrysoberyl occurs in granitic pegmatites in association with quartz in alkali feldspar and is thought to crystallize earlier than beryl and sillimanite. The pegmatites are thought to have derived from the residual melts of granitic liquids formed by partial melting of the khondalites. In the absence of pyrometasomatic, desilication or aluminium-contamination processes, the genesis of chrysoberyl is explained by the CO2 activity in residual melts. This model is at variance with the known concepts and may also explain the genesis of chrysoberyl in Sri Lanka where it is still controversial.

New geochronological information by Rb–Sr and U–Pb investigations from the pre-Alpine basement of Grande Kabylie (Algeria)

This study presents the first geochronological results, apart from previously obtained mica ages, from metamorphic basement complex occurring within the internal zone of the Alpine chain of Algeria. The Rb–Sr whole-rock and U–Pb zircon data reveal the existence of an episode of anatectic magmatism around 510 Ma ago that corresponds to late Pan-African orogenic events.A deep crustal fault is thought to represent a major phase of shearing associated with Hercynian blastomylonitic metamorphism and the production of granitic remelts at 271 ± 12 Ma. Reworking of an ancient basement is demonstrated by the presence of inherited radiogenic lead in zircons from lower Palaeozoic orthogneisses and metarhyolites: upper intercepts on concordia plots yield ages of 1950 ± 100 and 1321 ± 30 Ma, respectively. Apart from the dating results, which may provide new insight into the geological history of the region, this study also provides general implications for the interpretation of geochronological data in metamorphic and igneous formations.Firstly, the effect of the transformation of granite to orthogneiss appears to produce a dispersion of Rb–Sr analytical points between two apparent isochrons of 565 ± 50 and 465 ± 52 Ma. Such ages may also correspond to two separate groups of granitic protoliths that are difficult to distinguish in the field because of the superimposition of a gneissose fabric. In fact, the true emplacement age of these metagranites is given by the U–Pb zircon ages around 510 Ma. The Rb–Sr whole-rock systems of these orthogneisses have been opened during subsequent metamorphism, since model initial 87Sr/86Sr ratios calculated at 510 Ma are sometimes too low (&lt; 0.700) to be explained by heterogeneity of the source materials.Secondly, the influence of fractional crystallization of a granitic melt on the behaviour of inherited zircon xenocrysts has been investigated. This involves the U–Pb study of further zircon separates from rocks having the same emplacement age but differing degrees of differentiation. An orthogneiss showing relative little magmatic differentiation (SiO2 = 65%) contains zircons that have a component of inherited radiogenic lead and that yield a lower intercept age of 506 ± 3 Ma. Two other orthogneisses showing far greater differentiation (SiO2 = 70 and 78%) contain zircons that lack any inherited radiogenic lead and that yield an upper intercept on concordia plots of 514 ± 18 Ma. This behaviour can be explained in the light of published experimental data on the dissolution of zircons in granitic liquids.

Geochemical comparison between Himalayan and Hercynian leucogranites

Isotopic (Sr, Pb, Nd, O), REE and trace element data from three Himalayan (Nepal) and six Hercynian (Brittany, France) leucogranites are compared. For the Himalayan granites—Makulu, Mustang and especially Manaslu—the 87Sr/ 86Sr versus 87Rb/ 86Sr scatter diagrams, and hence variability of the initial 87Sr/ 86Sr ratios, reflect heterogeneity of the source materials. These ratios are very high (0.730- > 0.770). For Manaslu, the very high 207Pb/ 204Pb ratios (∼ 15.8) at the corresponding 206Pb/ 204Pb ratio of ∼ 18.8, the very low 143Nd/ 144Nd ratios (ϵ i Nd ∼ −13 to −17), and the very high 18O/ 16O ratios (∼ 12‰) confirm the crustal origin with the paragneisses of the Tibetan Slab Formation I as the probable parent source. For the six Hercynian granites with an individual RbSr isochron established for many of them, the initial 87Sr/ 86Sr ratios range from 0.706 to 0.717. The lead, neodymium and oxygen isotope compositions confirm their crustal source, of presumably Palaeozoic age but not specifically identified. The rare earth patterns and thorium contents are comparable to those for Himalayan granites. The similarities of the geochemical signatures (isotopes and trace elements) of the Himalayan and Hercynian leucogranites suggest that the petrogenetic processes were comparable and that the Himalayan continental collision model can be applied to the Hercynian. The depletion in REE and Th implies either interaction of the granitic liquids with fluid or, more probably, early precipitation of monazite. The increase in Sm/Nd in the granites compared to their parent materials implies that the neodymium model ages of primary mantle extraction are maximum rather than minimum ages. In contrast to the Hercynian leucogranites, the extremely radiogenic Sr of the Himalayan leucogranites reflects the very long crustal residence time before the melting event. The apparent uniformity of the initial 87Sr/ 86Sr ratios of the individual Hercynian granites may be related to an “aging effect” where the straight line fitting improves with increase in age of the pluton.

Constraints imposed by experimental petrology on possible and impossible magma sources and products

Experimental petrology can be used in forward and inverse approaches. The forward approach defines the compositions of liquids generated by partial melting of possible source rocks at various depths. The inverse approach determines conditions for multiple-mineral saturation at the liquidus of primitive magmas, correlates them with residual minerals of possible source rocks, and thus provides estimates of depths and temperatures required for their derivation. Review of a selection of forward and inverse results is followed by evaluation of petrological and geophysical processes in layered mantle and in subduction zones. Physical constraints imposed by solidus curves and geotherms present problems for models that derive basalts from deep mantle reservoirs, separated from overlying convecting layers. Magmas from mantle are limited to compositions less siliceous than basaltic andesite, with rare exceptions. Granite liquids cannot be generated from normal peridotite, nor from oceanic crust at mantle pressures in subduction zones. In continental crust, hydrous granite liquid is generated at depths of less than 30 km. Basaltic andesite and picritic basalt are parental magmas for the calc-alkaline series. Andesite is not primary from subcontinental depths, and can be generated as liquid in continental crust only if temperatures exceed about 1100°C. Calc-alkaline magmas may contain components from mantle peridotite, subducted oceanic crust, and continental crust.

Experimental Simulation of Mantle Hybridization in Subduction Zones

Experiments conducted at 30 kbar, 850-1050°C, in piston-cylinder apparatus simulate hybridization between hydrous siliceous magma rising from subducted oceanic crust into overlying peridotite. Gold capsules containing granite and peridotite powders separated by sharp boundaries, and , were run in vertical and horizontal positions. The aqueous vapor caused minor metasomatic changes in the peridotite until it was withdrawn into the melting granite, producing -undersaturated granite liquid adjacent sintered, anhydrous peridotite. Minerals developed in the hybrid reaction zones between liquid and peridotite are orthopyroxene and jadeitic clinopyroxene in all runs, and one or more of garnet, phlogopite, and quartz in some runs. (The granite liquid moves upwards relative to the peridotite even in runs of 27 hours.) In the longest runs of 120 hours the hybridization zone remains narrow, the body of granite liquid remains crystal-free and only slightly changed in composition (lower , MgO increased from 0.1% to about 1.5%). The products are consistent with phase equilibria in synthetic model systems, and the system peridotite-granite- (determined using mixtures). The hybridization process in subduction zones would produce discrete rock bodies dominated by pyroxenites without olivine. Na is fixed in jadeitic clinopyroxene. K is fixed in phlogopite dispersed through the pyroxenite, but there is a prospect that it may become concentrated into phlogopite-rich rocks by crystal fractionation. Partial melting of these source rocks would generate magmas different from those generated in peridotite or subducted oceanic crust.

An occurrence of rare-earth-rich eudialyte from Ascension Island, South Atlantic

AbstractA mineral closely resembling eudialyte in single crystal X-ray pattern and chemistry has been found in one granite block from the 5 Mile Post area on Ascension Island. To our knowledge, eudialyte has not previously been recognized in this suite of granites although they contain two other rare zirconium silicates, dalyite and vlasovite. Textural evidence suggests that the eudialyte, which is remarkable for its high yttrium and REE content, crystallized from very late-stage magmatic fluids. Rare-earth patterns are only slightly fractionated (CeN/YN = 1.8–2.5) compared with those of allanites and sphenes from Skye granites which presumably have similar parageneses. In contrast, the whole-rock REE pattern shows greater enrichment of light REE (CeN/YN ∼ 10). This difference in REE pattern is considered to be due to the effects of REE and other element complexing related to the peralkalinity of the granite liquid.

Dehydration melting of pelitic rocks and the generation of H 2 O-undersaturated granitic liquids

Dehydration melting of pelitic rocks and the generation of H 2 O-undersaturated granitic liquids

Phase relationships of I‐type granite with H2O to 35 kilobars: The Dinkey Lakes biotite‐granite from the Sierra Nevada Batholith

The Dinkey Lakes biotite‐granite from the Sierra Nevada batholith was reacted (with varying percentages of H2O in sealed platinum capsules) in a piston‐cylinder apparatus between 10 and 35 kbar. The results were combined with the results from previously published experiments to provide comprehensive phase relationships for an I‐type granite: a P‐T diagram with excess H2O; isobaric T‐XH2O diagrams at 25, 30, and 35 kbar showing H2O‐undersaturated relations; the H2O‐undersaturated liquidus surface mapped with contours for constant H2O contents and fields for near‐liquidus minerals; and the solubility of H2O in granite liquids to 35 kbar. Results and their implications show: (1) The solidus temperature decreases from 680°C at 2 kbar to 620°C at 10 kbar, then increases to 700°C at 35 kbar because of changes from less dense to more dense subsolidus mineral assemblages. (2) The melting interval with excess H2O, which is only 35°C at 2 kbar, increases to 105°C at 10 kbar and 150°C at 35 kbar because the liquidus minimum in the complex rock system departs from granite composition with increasing pressure. (3) The solubility Of H2O in granite liquid is 27 ± 2.5 weight percent at 35 kbar and 850°C, indicating that a miscibility gap persists between H2O‐saturated silicate magmas and aqueous vapor phase, at least to pressures corresponding to 120‐km depth in the mantle. Dissolution of alkali feldspar (20% of rock) in the subsolidus aqueous vapor phase indicates that deep‐seated aqueous fluids are concentrated solutions. (4) Quartz and coesite are the liquidus minerals at mantle pressures for all H2O contents, indicating that granites and rhyolites cannot be primary magmas from mantle peridotite or subducted oceanic gabbroic crust. (5) The liquidus surface at crustal pressures, with plagioclase and quartz as primary minerals, indicates that primary liquids of granite composition with moderate H2O contents can be generated in the crust at reasonable temperatures; these liquids could rise to near surface levels without vesiculation. Granite liquid together with residual crustal minerals could constitute plutonic magmas of intermediate composition.

Mineral-chemical and isotopic studies of Namaqualand granulites, South Africa: A grenville analogue

The northwestern part of South Africa and southern South-West Africa/Namibia is amongst the most extensive granulite terranes in Africa. This work reports the results of electron microprobe studies of minerals from two-pyroxene, cordieriteorthopyroxene (-gedrite) (-sapphirine) and garnet and/or cordierite parageneses from Namaqualand, in the N.W. Cape Province of South Africa. Determined PT conditions of prograde metamorphism based on thermodynamic calculations are 800°–900° C and ca. 6–7 Kb; and it is argued that rocks of unusual composition, notably cordierite-orthopyroxene rocks, are restites after the extraction of granitic liquid from former argillites. This interpretation is consistent with previously published data on similar rocks, and with McCarthy's (1976) suggestion of extensive partial melting in the quartzofeldspathic rocks in the area.