Search

The P–T paths of the exhumation of Precambrian granulite complexes at craton boundaries usually include two stages: subisothermal decompression and a decompression–cooling stage with a more gently sloped P–T path. Our goal is to understand the possible causes of the change in the slope of the P–T exhumation path of the Central Zone (CZ) of the Limpopo granulite complex, South Africa, located between the Kaapvaal and Zimbabwe cratons. For this purpose, rocks (mainly, metapelites) were studied in various structural settings within the Central Zone, i.e., in dome structures, regional cross folds, and in local and regional shear zones. The metapelites are gneisses of similar bulk composition. The rocks contain various amounts of relics of leucosomes composed of quartz–feldspar aggregates with garnet and biotite, and melanocratic domains that are enriched in cordierite and usually mark shear microzones that envelope and/or break garnet porphyroblasts. Study of polymineralic (crystallized melt and fluid) inclusions in the garnet, its zoning with respect to the major (Mg, Fe, and Ca) and some trace (P, Cr, and Sc) elements, fluid inclusions in quartz, as well as phase equilibria modeling (PERPLE_X) showed that the rocks coexisted with granite melts and saline aqueous carbonic fluids ( = 0.74–0.58) at the peak of metamorphism at 800–850°C and 10–11 kbar. Partial melting of the rocks initiated their subisothermal exhumation to 7.5–8 kbar during diapirism of granitic magmas in the Neoarchean (2.65–2.62 Ga). This is reflected in the specific zoning of the garnet grains in terms of the grossular content. A change in the rheology of the rocks as a result of partial removal and crystallization of melt activated the shear zones during further exhumation to 6–5.5 kbar along a decompression–cooling P–T path at 95–100°/kbar, reflecting the slower uplift of the rocks in the middle crust. This process was resumed due to thermal effects and interaction of the rocks with aqueous fluids ( > 0.85) in the Paleoproterozoic (~2.01 Ga). Such a scenario of metamorphic evolution implies that the Limpopo granulite complex in general and its Central Zone in particular resulted from the evolution of an ultrahot orogen, in which vertical tectonic movements associated with diapirism were coupled to horizontal tectonic processes caused by the convergence of continental blocks. [ABSTRACT FROM AUTHOR]

Xenoliths in kimberlites are the most promising material for studying the composition and structure of the lower levels of the continental crust. This study is aimed at the estimation of P–T parameters and fluid regime of metamorphism for garnet–biotite–feldspar and orthopyroxene–garnet–biotite–feldspar rocks found as xenoliths in kimberlites of the Yubileynaya and Sytykanskaya pipes, Yakutian kimberlite province. The seven studied samples show inverse dependences of relative contents of garnet and orthopyroxene, orthopyroxene and biotite, garnet and plagioclase, plagioclase and potassium feldspar. This indicates a consistent series of transformations of the assemblage garnet + plagioclase + orthopyroxene ± quartz to the assemblage garnet + biotite + potassium feldspar. In this process, the replacement of plagioclase by potassium feldspar was the leading reaction. Evidence of this reaction is specific reaction textures in the rocks, negative correlations between the contents of the minerals, and petrochemical characteristics of the rocks. Modeling of the mineral assemblages of the xenoliths using the pseudosection approach (PERPLE_X) revealed two groups of rocks corresponding to different depth levels of the Siberian cratonic crust. For rocks without orthopyroxene or with this mineral as single relics, the pressure was estimated at 9.5–10 kbar, and it is 6–7 kbar for orthopyroxene-bearing samples. The xenolith rocks have close metamorphic peak temperatures of 750–800°C. They experienced 200–250°C cooling and 3–4 kbar decompression, regardless of the level of the crust at which they had initially occurred. This indicates that the metamorphic evolution of the rocks during their exhumation was probably associated with collisional processes during the amalgamation of individual terrains of the Siberian craton. Xenoliths enriched in K-feldspar might have been products of metamorphic reactions with participation of aqueous–(carbonic)–salt fluids, which were sourced from basaltic magmas in the lower crust. The most strongly metasomatized rocks were located closest to the place of accumulation of crystallizing magmas. [ABSTRACT FROM AUTHOR]

To study the formation of S-type ultra-potassic granitic melts under the conditions of the continental crust, we conducted experiments on partial melting of plagioclase-free garnet–two-mica metapelite (with accessory apatite and ilmenite) at 6, 10, and 15 kbar and 700–900°C. Rock melting is driven by a series of peritectic melting reactios as a function of pressure and starts at temperatures of 750–800, ~800, and ~850°C at 6, 10, and 15 kbar, respectively, reflecting the positive dP/dT slope of the solidus. These P–T-conditions of melting, compositions of mineral assemblages and granitic melts formed during melting of plagioclase-free metapelites are similar to those formed during melting of plagioclase-bearing two-mica assemblages. The high K2O content (7–8 wt %) at a K2O/Na2O ratio of >8–10 is the main feature of the composition of the melt, which forms upon melting of plagioclase-free metapelite. These features are similar to the compositions of S-type ultra-potassic rhyolites, which form in extension tectonic settings. [ABSTRACT FROM AUTHOR]

One of the mechanisms explaining relationships between CO2-rich fluids and granitoid magmas at high-temperature crustal metamorphism is the melting of protoliths that originally contained carbonate minerals. In order to study the coupled processes of dehydration/decarbonation and melting, experiments were conducted with carbonate–biotite gneiss from an Archean greenstone belt at pressures of 6, 10, and 15 kbar in the temperature range of 800–950°C, and phase relations in this rock were modeled using the pseudo-section method. The experiments and modeling revealed a subvertical positive dP/dT slope of the solidus of the rock. In comparison to the calculated solidus temperatures, the experiments showed higher melting temperatures (~800°C at 6 kbar and ~850°C at 10 and 15 kbar). The products of the experiments at pressures of 6 and 10 kbar and temperatures >850°C were found out to contain assemblages of clinopyroxene, orthopyroxene, and ilmenite. The products of the experiments at 15 kbar did not contain either orthopyroxene or ilmenite, but calcium garnet and rutile were stable. The first portions of the near-solidus melts at 6 and 10 kbar were poor in SiO2 (44–50 wt %) and were formed because carbonate phases were involved in the melting reactions. With a temperature increase, the melts acquired a granite composition that was close to the composition of melts formed during the melting of the carbonate-free plagioclase + biotite + quartz assemblage. An aqueous–carbonic fluid containing Ca–Mg–Fe carbonate components coexisted with the melts. The phase assemblages and compositions of the granite melts obtained in the experiments are consistent with the modeling results. Comparison of the experimental results with published data on the partial melting of the carbonate-free plagioclase + biotite + quartz assemblages led us to the preliminary conclusion that Ca–Mg–Fe carbonates are able to decrease the melting temperature. The experiments have demonstrated that granite magmas can be derived together with aqueous–carbonic fluids from a carbonate-bearing protolith during high-grade metamorphism in the middle and lower crust. The occurrence of clinopyroxene or two-pyroxene assemblages in granitoids can be considered as a mineralogical indicator of this process. [ABSTRACT FROM AUTHOR]

The results of experimental study of crystallization of chromium-bearing Ba-titanates (redlegeite, lindsleyite, and hawthorneite) in the chromite–rutile/ilmenite system in the presence of the H2O–CO2–BaCO3 fluid at pressures of 1.8, 3.5, and 5.0 GPa and temperatures of 1000–1200°C, simulating the formation of these phases through metasomatism of peridotite of the upper mantle, are reported. Experiments showed the formation of Ba–Cr-titanates in the entire pressure range studied and the possibility of joint crystallization of titanates. However, redledgeite is formed only in the Fe-poor chromite–rutile–H2O–CO2–BaCO3 system, whereas minerals of the magnetoplumbite group are preferably crystallized in the system with ilmenite. Lindsleyite was not detected at a pressure of 1.8 GPa. A positive dependence of the Cr content in titanates on pressure was shown. The Raman spectra of redlegeite, lindsleyite, and hawthorneite are presented. [ABSTRACT FROM AUTHOR]

The structures of two K-richterite crystals with the crystal chemical formulas (K0.44Na0.320.24)Σ=1(Ca1.18Na0.82)Σ=2Mg5Si8O22OH2 and (K0.83Na0.020.15)Σ=1(Ca1.11Na0.89)Σ=2Mg5Si8O22OH2 synthesized at a pressure of 3 GPa and a temperature of 1000°С in the MgSiO3 + CaMgSi2O6 + K2CO3 + Na2CO3 + CO2 + H2O system were studied by single-crystal X-ray analysis and Raman spectroscopy. The monoclinic cell parameters were obtained: a = 10.0256(5) and 9.9748(11) Å, b = 17.9874(7) and 17.9879(16) Å, c = 5.2687(3) and 5.2746(6) Å, Ve.c. = 916.17(18) and 918.52(8) Å3, β = 104.520(12)° and 104.821(5)°, sp. gr. С2/m (12), and Z = 2. The sites M(1), M(2), and M(3) are inhabited by Mg2+, while site M(4) is occupied simultaneously by Ca2+ and Na+. The Na+ cations not included in M(4) are located in position A, which also accommodates K+ cations. Raman spectroscopy made it possible to reveal vacancies in position A in both samples. The structure corresponds to the "ideal" structure of richterite group minerals. The unit cell volume of the measured crystals is directly proportional to the K content in position A. Based on the generalization of new and published data, an equation of the dependence of Vu.c. for amphiboles of the richterite Na(NaCa)Mg5Si8O22(OH)2–K-richterite K(NaCa)Mg5Si8O22(OH)2 series with a low tremolite component on the K content in position A is proposed. [ABSTRACT FROM AUTHOR]

The paper summarizes authors' earlier and newly obtained experimental data on the formation of phlogopite and chromium-bearing potassium titanates of the crichtonite, magnetoplumbite, and hollandite groups, which are indicator minerals that characterize various stages of modal metasomatism in the upper mantle. Phlogopite-forming reactions were studied in the garnet–orthopyroxene system in the presence of H2O–KCl fluid at 3 and 5 GPa and 900–1000°C as simulating phlogopite formation in garnet peridotites and pyroxenites. Experiments have demonstrated regular variations in Ca and Cr contents in the garnet and Al in the pyroxenes, as well as the composition of the newly formed phlogopite, depending on the concentration of the potassium component (KCl or K2CO3) of the fluid. Experiments on the formation of potassium titanates (yimengite, mathiasite, and priderite) in the chromite–rutile/ilmenite–K2CO3–H2O–CO2 system at 3.5 and 5 GPa and 1200°C have proved that these minerals can be formed when chromite reacts with potassic aqueous–carbonic fluid, but these reactions require additional sources of titanium. These sources may be rutile and/or ilmenite, which are usually also produced by modal metasomatism of peridotites. The experiments thus confirm that the formation of titanates marks the most advanced stages of metasomatism in mantle peridotites. The experiments have shown that the formation of potassium titanates follows the formation of phlogopite, whereas assemblages of the titanates with phlogopite are produced at higher concentrations of the potassium component of the fluid than phlogopite alone. The relationships between the titanates are also a function of the activity of the potassium component in the fluid and, probably, pressure. The relationships and trends reproduced in the experiments clearly illustrate features of mineral assemblages and variations in compositions of minerals from metasomatized peridotites in the lithospheric mantle. [ABSTRACT FROM AUTHOR]

Data on the carbon and oxygen isotopic composition of carbonates from the rocks of the Giyani greenstone belt, Kaapvaal craton, South Africa are presented. This belt is immediately adjacent to the Hout River Shear Zone, along which the Southern Marginal Zone (SMZ) of the Limpopo granulite complex is overthrusted onto the Kaapvaal craton. The δ13C values are –2.1 and –2.2‰ for carbonates from two metapelite samples, –5.1‰ for a meta ultrabasite sample, and –7.1 and –7.7‰ for two metabasite samples, respectively. The δ18О values are 17.1 and 17.5‰ for carbonates from the metapelites, 14.3‰ for carbonates from the metaultrabasite, and 12.9 and 13.0‰ for carbonates from the metabasites. The comparison of the data obtained to the published data on the isotopic composition of carbonates from the rocks of the different greenstone belts indicates both hydrothermal (in the metabasites and the metaultrabasite) and sedimentary (in the metapelites) origins of the carbonate material in the rocks studied. Based on the comparison of these data to δ13C of magnesite that formed in ultrabasic granulites during the interaction of the Limpopo complex with the craton, the graphite and fluid inclusions in leucocratic garnet-bearing granitoids that transported fluids, and the carbon isotopic composition of graphite from metapelites of the SMZ, it is concluded that the carbon source was located outside the SMZ and represented a heterogeneous carbonate-bearing lithologies. It could be carbonate-bearing volcanogenic–sedimentary sequences of the greenstone belts of the Kaapvaal craton, which underwent prograde metamorphism during the interaction with the SMZ granulites. [ABSTRACT FROM AUTHOR]

The salt components of aqueous and aqueous-carbonic fluids are very important agents of metasomatism and partial melting of crustal and mantle rocks. The paper presents examples and synthesized data on mineral associations in granulite- and amphibolite-facies rocks of various composition in the middle and lower crust and in upper-mantle eclogites and peridotites that provide evidence of reactions involving salt components of fluids. These data are analyzed together with results of model experiments that reproduce some of these associations and make it possible to more accurately determine their crystallization parameters. [ABSTRACT FROM AUTHOR]

Abstract: Data on the carbon isotope composition of graphite and CO2 from inclusions in quartz of granitoids of the Southern Marginal Zone (SMZ) of the Limpopo granulite belt, South Africa, are presented. The average values of δ13C for graphite and СО2 from fluid inclusions in quartz are found within the range that is typical of deep (mantle) carbon sources. The data are evidence of an external source of granitoid magmas. They are likely to be hybrid rocks that appeared during assimilation of host metapelites by trondhjemite magmas, while graphite in trondhjemites and granodiorites is a product of high-temperature (~900-1000°С) evolution of the CO2-rich fluid phase of granitoid magmas when they intrude the rocks of the granulite complex and was formed during assimilation that led to reduction of CO2 when sulfide-bearing metapelite material was contaminated by the magma. [ABSTRACT FROM AUTHOR]

The paper presents data on the formation of K–Na richterite in the enstatite + diopside association with K2CO3–Na2CO3–CO2–H2O fluid at 3 GPa and 1000°C as a model for the formation of this mineral in peridotites of the upper mantle. Richterite formation depends on the (H2O + CO2)/(K2CO3 + Na2CO3) and K2CO3/Na2CO3 ratios in the starting material. A high concentration of alkaline components in the fluid leads to the decomposition of clinopyroxene, the formation of olivine, and a change in the component composition of the pyroxene and amphibole. Fluids with a high potassium concentration are favorable for the formation of K-richterite similar in composition to that formed in metasomatized peridotites of the upper mantle. In some cases, such a fluid leads to the decomposition of amphibole and stabilization of alkaline melt. An increase in the activity of the sodium component results in richterite similar in composition to richterite from lamproites. The clarified relations can be used to assess the activities of fluid components and conditions for the formation of K-richterite. To update the data bank of the Raman spectra of minerals, the largest and most homogeneous amphibole crystals of different compositions were studied. [ABSTRACT FROM AUTHOR]

Abstract: The paper summarizes data on the petrology of granitoid intrusions in the Southern Marginal Zone (SMZ) of the Neoarchean Limpopo granulite complex, South Africa, and discusses the thermal and fluid effects of these intrusions on the granulites. The intrusions were emplaced in SMZ at 2680-2640 Ma, when the granulite complex was overthrust on the rocks of the adjacent Kaapvaal Craton. The mineral assemblages of the granitoids reflect temperatures above 900°C for the magmas that crystallized under pressures of 6-9 kbar. The granitoid magmas assimilated host rocks and were thereby enriched in MgO, FeO, and Al2O3, which was favorable for the crystallization of garnet, spinel, sillimanite, and corundum in the granitoids. Fluid inclusions in the granitoids and estimates of the fluid composition based on mineral equilibria indicate that the CO2/H2O ratio of the fluids broadly varied. Along with H2O-CO2 fluid, the magmas carried H2O-salt fluids, which penetrated the host rocks and triggered various metasomatic reactions in them. The thermal effects of the intrusions on the host granulites resulted in the development of partial melting zones that host orthopyroxene with >7 wt % Al2O3. Depending on the fluid regime and temperature, the orthopyroxene is found in equilibrium with either garnet and potassic feldspar or with biotite. The ages of the partial melting zones are comparable with those of the intrusions. [ABSTRACT FROM AUTHOR]

The paper presents experimental data on the interaction of amphibole with NaCl–H2O and (K, Na)Cl–H2O solutions at varying salt content. When interacting with H2O–NaCl fluid, amphibole remains the predominant mineral in all experiments, and the newly formed minerals are Na-phlogopite, plagioclase, and nepheline/sodalite. At > 0.6, the amphibole melts. When amphibole interacts with H2O–NaCl–KCl fluid at < 0.40 and ХKCl/(ХKCl + ХNaCl) in the fluid, defined as ХNaCl = 0.506 – 0.84ХKCl, the amphibole is replaced by the association of nepheline with sodic plagioclase, sodalite, and biotite. At ХKCl/(ХKCl + ХNaCl) > 0.3, nepheline, sodalite, and plagioclase become unstable, K-feldspar is formed, and biotite, clinopyroxene, and amphibole remain stable. At ХKCl/(ХKCl + ХNaCl) > 0.5, the association Cpx + Bt + Kfs + Grt (grossular–andradite) is stable. Thus, grossular–andradite garnet is an indicator of a high potassium activity in the fluid, whereas nepheline testifies that the sodium activity was high. Na → K exchange is typical of the amphibole and biotite, and Ca → Na exchange occurs in the clinopyroxene, and all of these minerals (but neither nepheline nor garnet) remain generally stable within a wide range of the K/Na ratio in the fluid. Clinopyroxene in the experiments spans Ca–Fe–Mg compositions with a varying, sometimes high, Al content, and the amphiboles belong to the pargasite–hastingsite series. With an increase in ( > 0.57), i.e., a decrease in the gross salinity of the fluid, melts are generated, and their composition varies from trachyte to phonolite. An increase in the ХKCl/(ХKCl + ХNaCl) ratio in the fluids leads to a decrease in alumina content of the melts. An increase in the total salinity of the fluid leads to an increase in the content of potassium in the melt and a decrease in the content of chlorine in it. The experiments have shown that interaction between amphibole and fluids containing high NaCl and KCl concentrations results in mineral associations typically produced by alkaline metasomatism of amphibole-bearing rocks and concomitant HCl enrichment in the fluid phase. The substitution of highly saline fluids for highly acidic ones leads to the leaching of Ca, Mg, Fe from the metamorphic rocks, and the transport and redeposition of these components. It follows that significant removal of FeO, MgO, CaO from rocks is sometimes a consequence of the interaction of the host rocks with saline aqueous solutions. [ABSTRACT FROM AUTHOR]

The results of experimental studies are presented for reactions in the orthopyroxene–garnet–phlogopite system in the presence of H2O–KCl fluid at 3–5 GPa and 900–1000°C, which model the processes of phlogopite formation in garnet peridotites and pyroxenites during alkaline metasomatism of the upper mantle. The experiments demonstrated regular variations in the composition of garnet, pyroxenes, and phlogopite depending on the KCl content of the fluid. With increasing KCl content of the fluid, enstatite and garnet become unstable, the Al2O3 content of enstatite decreases, and the amount of grossular and knorringite components in garnet are maximum at a KCl content of ~10 mol %. Our results illustrate well the regular variations in the compositions of the coexisting minerals and their zoning in phlogopite-bearing peridotites of the lithospheric mantle. [ABSTRACT FROM AUTHOR]

The results of experiments on the synthesis of exotic titanates (priderite and yimengite) simulating metasomatic conditions of alteration of the mantle minerals (chromite and ilmenite) are reported. Ba-free Cr-bearing priderite was synthesized for the first time. Experiments showed the possibility of crystallization of this mineral as a product of the reaction of high-Cr spinel and rutile with hydrous–carbonate fluid (melt) under the conditions of the upper mantle. In particular, the experimental data obtained provide an interpretation of the relationships between K‒Cr priderite and carbonate–silicate inclusions in chromites from garnet peridotite of the Bohemian massif. Experimental study of the reaction of chromite and ilmenite with potassic hydrous–carbonate fluid (melt) shows the presence of both titanate phases (priderite and yimengite), the mineral indicators of mantle metasomatism. This provides direct evidence for the formation of yimengite and K‒Cr priderite, as well as other titanates, due to mantle metasomatism of the upper mantle peridotite under the conditions of the highest activities of potassium. [ABSTRACT FROM AUTHOR]

The structure of a potassium compound with the chemical formula K0.90Ti5.16Cr2.94Fe2.54Mg0.87Al0.22Mn0.30O19 has been investigated by single-crystal X-ray diffraction (using synchrotron radiation of the KSRS-Kurchatov source) and Raman spectroscopy. The studied crystal was extracted from the products of an experiment performed in the chromite–ilmenite–K2CO3–H2O–CO2 system at 5 GPa and 1200°С. The crystal has a hexagonal cell with the following parameters: a = 5.8920(2) Å, c = 23.0113(8) Å, sp. gr. P63/mmc, Z = 4. The compound is isostructural to yimengite (potassium mineral of the magnetoplumbite group). In contrast to yimengite, Fe3+ is absent in the compound under consideration. The cation Fe2+ is distributed between tetrahedral М3 and octahedral М5 sites. The absence of Fe3+ in the compound under study conditioned incorporation of Cr and small amount of Al into the M1 (octahedron) and M2 (trigonal bipyramid) sites. In contrast to yimengite, chromium in the synthetic potassium phase is distributed over three sites: M1, M2, and M5. The data obtained can be used to specify the nomenclature of magnetoplumbite-group compounds. [ABSTRACT FROM AUTHOR]

The aegirine end-member (NaFe3+Si2O6) in clinopyroxenes resulted from incorporation of Fe3+ into the mineral structure. Its presence affects the accuracy of reconstruction of the P-T conditions in the high-grade metamorphic rocks and allows the evaluation of the redox conditions of their formation. The content of this end-member in clinopyroxenes is usually estimated using crystal chemical recalculations of microprobe analyses. However, in some publications on eclogites, the comparison of microprobe-based recalculations with Mössbauer data revealed significant difference between the measured and calculated Fe3+/ΣFe ratios, which can significantly affect the results of geothermometry. This paper presents the results of the Mössbauer spectroscopy measurements of clinopyroxene fractions separated from three samples of garnet–clinopyroxene granulites from the Udachnaya kimberlite pipe. The ratios Fe3+/ΣFe = 0.22–0.26 measured in the clinopyroxenes correspond to 6–10 mol % aegirine. These estimates are in good agreement with the values obtained for the same clinopyroxenes by the recalculation of microprobe analyses using the charge balance method. Following this conclusion, we believe that crystal chemical recalculations of microprobe analyzes of clinopyroxenes from non-eclogitic rocks make it possible to correctly estimate the Fe3+ content in them. Similar recalculation of microprobe analyses of clinopyroxenes in crustal xenoliths from other localities, as well as from ferrobasalts of the continental flood basalts provinces, ferrodolerite dikes, and gabbroid xenoliths (similar in bulk chemical composition to many lower–middle crustal xenoliths) revealed significant amounts of previously unaccounted aegirine (up to 13 and 4–9 mol %, respectively), which holds the potential for deciphering redox conditions in many rocks. [ABSTRACT FROM AUTHOR]

A clear case study of local-scale, fluid-induced dehydration of the Paleoarchean Sand River biotite–hornblende gneiss from the Central Zone of the Limpopo Complex is presented here. Field and petrographic examination of three adjacent zones—darker Sand River orthogneiss with local occurrence of orthopyroxene and clinopyroxene, a lighter intermediate gneissic zone with more orthopyroxene than the Sand River orthogneiss, and tonalitic veins containing large orthopyroxene-bearing patches—indicates the local transformation of a light grey, fine- to medium-grained, hornblende–biotite gneiss into a greenish brown, medium- to coarse-grained orthopyroxene-bearing dehydration zone. Field evidence indicates that the tonalitic veins were emplaced in discrete ductile shear zones, with development of large orthopyroxene-bearing patches in a sigmoidally transposed foliation bounded by shear planes. Orthopyroxene-forming reaction textures after biotite and amphibole together with the occurrence of microveins of K-feldspar along quartz–plagioclase grain boundaries in the three adjacent zones, and the higher modal abundance of orthopyroxene and K-feldspar with lesser biotite and amphibole from the Sand River orthogneiss to the intermediate gneissic zone to the orthopyroxene-bearing patches, indicate that the three adjacent zones represent progressive stages of the dehydration process. Such K-feldspar microveins along quartz–plagioclase grain boundaries have been proposed as evidence for the presence and passage of a low H2O activity fluid. Further, the occurrence of monazite inclusions in fluorapatite in orthopyroxene-bearing zones suggests dissolution and reprecipitation involving a free fluid phase. Fluid inclusion studies indicate the presence of a fluid with CO2, NaCl and H2O components, with higher salinity of the fluid (up to 29% NaCl) in the orthopyroxene-bearing patches relative to the intermediate gneissic zone. The increase in Cl content in amphibole, biotite and fluorapatite from the Sand River orthogneiss to the orthopyroxene-bearing patches supports the presence of a Cl-rich brine fraction in the fluid responsible for the dehydration process. Further, the increase in An content of plagioclase at the contact with the K-feldspar rims on quartz reflects an increase in potassium activity in the fluid. The whole-rock major, trace and rare-earth element enrichment or depletion patterns of the orthopyroxene-bearing zones relative to the precursor support the dehydration process. The diffuse contact relationship of a granite pegmatite occurring in the vicinity of the dehydration zones, together with fluid inclusion and whole-rock major, trace and rare element characteristics of samples collected along a traverse from the granite pegmatite to the Sand River orthogneiss, suggests a scenario in which the dehydrating fluids derived from an external source utilized lithological contrasts, such as the gneiss–pegmatite boundaries, as fluid conduits. Dehydration of the gneissic wall-rock occurred where permeability was sufficient for fluid penetration. The occurrence of orthopyroxene-bearing tonalitic veins along deformation-transposed foliation planes further attests to a structural control to the channeling of the dehydrating fluids. [ABSTRACT FROM PUBLISHER]