Marschall et al, 2003 simplectitas comparación

Marschall et al, 2003 simplectitas...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 PAGES 227–253 2003 P–T Evolution of a Variscan Lower-Crustal Segment: a Study of Granulites from the Schwarzwald, Germany H. R. MARSCHALL∗, A. KALT† AND M. HANEL ¨ MINERALOGISCHES INSTITUT, UNIVERSITAT HEIDELBERG, IM NEUENHEIMER FELD 236, D-69120 HEIDELBERG, GERMANY RECEIVED NOVEMBER 27, 2001; REVISED TYPESCRIPT ACCEPTED AUGUST 2, 2002 Pressure–temperature–time (P–T–t) paths of orogenic granulites provide important information on the thermal and chemical structure of the lower continental crust through time, and constraints on tectonic processes. We present the first detailed petrological investigation of granulites from the Variscan Schwarzwald. Pelitic granulites from the Central Schwarzwald Gneiss Complex (CSGC) are characterized by the peak assemblage garnet + rutile + kyanite + antiperthite ± quartz. Felsic to intermediate granulites from the Southern Schwarzwald Gneiss Complex (SSGC) exhibit different peak assemblages with clinopyroxene, orthopyroxene, ternary feldspar, garnet, quartz and sillimanite, and manifold retrograde reaction textures. Peak P–T conditions were calculated by two-feldspar thermometry, garnet–orthopyroxene thermometry and various geobarometers. Minimum estimates for peak conditions are 950–1010°C and 1·4–1·8 GPa for the granulites of the CSGC, which followed a clockwise P–T path. The retrograde path is characterized by initial isothermal decompression, associated with partial melting, followed by isobaric cooling. Peak conditions for the SSGC are 1015°C and 1·5 GPa (minimum temperature, maximum pressure). No prograde relics are preserved, and isothermal decompression was less pronounced than in the CSGC. Other Variscan HP–HT granulites from Central Europe show similar lithologies, equilibration temperatures and ages (340–335 Ma). The heat for widespread high-temperature metamorphism in the Variscan lower crust could have been supplied by repeated intrusion of subduction-related basic magmas. Rapid, near-isothermal decompression of the granulites may have been facilitated by considerable volumes of partial melt and by orogenic extension. INTRODUCTION General background KEY WORDS: granulites; near-isothermal decompression; two-feldspar thermometry; HT metamorphism; Variscan Schwarzwald Granulites are typical rocks of the Earth’s middle to lower crust under high-temperature conditions. They are found as xenoliths in basaltic volcanic rocks, mainly within continental rifts, but most granulites occur as complexes or terranes in orogenic settings. Orogenic granulites display a wide compositional range (Harley, 1989, 1998) and are known from a variety of collisional belts that formed during different episodes since the Archaean (e.g. Harley, 1998; Clarke et al., 2000; Moller ¨ et al., 2000). Determination of bulk-rock and mineral compositions, calculation of peak equilibration conditions, and dating of orogenic granulites thus provide us with important constraints on the thermal and chemical structure of the Earth’s continental crust at different geological times. Harley (1989) showed that equilibration conditions deduced from natural orogenic granulites cover a wide range. In particular, pressures are highly variable, and the granulite field may be divided into a low-pressure (LP), a medium-pressure (MP) and a high-pressure (HP) facies according to Green & Ringwood (1967). Peak temperatures of many granulite terranes scatter around 800°C (Bohlen, 1987), but an increasing number of ultra-high-temperature (UHT) granulite complexes (900–1100°C and 0·7–1·3 GPa; Harley, 1998) are being recognized (e.g. Harley & Hensen, 1990; Dasgupta et al., 1995; Klemd & Brocker, 1999; Hokada, 2001). Regarding ¨ retrograde P–T paths, Harley (1989) distinguished two basic types of granulites: those that experienced isobaric ∗Corresponding author. Telephone: ++49 6221 54 6021. Fax: ++49 6221 54 4805. E-mail: hmarscha@min.uni-heidelberg.de †Present address: Institut de Geologie, Universite de Neuchatel, Rue ˆ ´ ˆ Emile Argand 11, CH-2007 Neuchatel, Switzerland. Oxford University Press 2003 JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 FEBRUARY 2003 cooling (IBC) and those that were subject to isothermal decompression (ITD). Whereas IBC granulites can form in a variety of tectonic settings, ITD granulites generally display clockwise P–T paths and high peak pressures that evidence equilibration at lower-crustal or even mantle depths. They seem to be produced by crustal thickening and subsequent thinning processes, and are typical of continental collision belts. Hence, orogenic granulites not only constrain the equilibration conditions of the continental crust; their reaction textures, the form of their P–T paths and the duration of peak metamorphism and cooling can provide important information on major tectonic processes during orogeny, on the nature and origin of the heat necessary for granulite formation, and on mechanisms that trigger exhumation. Lower-crustal granulites are of particular interest in this context, because the lower crust is the potential site of heat and material transfer between mantle and crust during orogeny. Variscan background Within the Variscan belt, Pin & Vielzeuf (1983, 1988) distinguished two groups of granulites. According to this subdivision, group I granulites are associated with eclogites and peridotites and equilibrated at HP conditions during an early Variscan episode (450–400 Ma). The granulites of all Variscan massifs in central Europe (Fig. 1), such as the Bohemian Massif, the Schwarzwald, the Vosges and the Massif Central, have been assigned to this group. Group II granulites are not associated with eclogites and peridotites and equilibrated at LP conditions during a late Variscan event (>300 Ma). According to Pin & Vielzeuf (1983, 1988) all granulites representing Variscan crust involved in Alpine tectonic processes, such as the Ivrea Zone, the Pyrenees and Southern Calabria, belong to group II. More recent, detailed petrological and geochronological investigations, particularly in the Bohemian Massif, have revealed a more complex picture (see section on comparison with granulites from the Bohemian Massif and the Vosges). The existing geochronological data on the Bohemian Massif now indicate three events of granulite formation at approximately 400, 340 and 324 Ma (e.g. von Quadt, 1993; Wendt et al., 1993; O’Brien et al., 1997; Kroner & Willner, 1998; Kalt et al., 2000b; Romer ¨ & Rotzler, 2001). Thermobarometric studies indicate ¨ that equilibration conditions within the older two groups varied between (ultra)high-pressure [(U)HP] and UHT metamorphism (e.g. Carswell & O’Brien, 1993; Kryza et al., 1996; Willner et al., 1997; Klemd & Brocker, 1999; ¨ Rotzler & Romer, 2001). Low-pressure–high-tem¨ perature (LP–HT) granulites and migmatites (Kalt et al., 1999, 2000b) appear to be confined to the youngest group. Fig. 1. Simplified geological map of the Schwarzwald. The metamorphic basement is divided into three tectonic units according to Hanel et al. (1999). The inset shows the Variscan massifs in central Europe and their subdivision into three zones (RH, ST, MO) according to Kossmat (1927). Abbreviations in the inset (after Franke, 1989): A, Alps; AM, Armorican Massif; BM, Bohemian Massif; MC, Massif Central; MO, Moldanubian zone; RH, Rhenohercynian zone; ST, Saxothuringian zone; SW, Schwarzwald; VG, Vosges. Abbreviations in the map: BBZ, Baden-Baden Zone; BLZ, Badenweiler–Lenzkirch Zone; CSGC, Central Schwarzwald Gneiss Complex; SSGC, Southern Schwarzwald Gneiss Complex. Equilibration conditions of the granulites from the Vosges have not yet been determined. However, phase assemblages in metapelitic granulites (Latouche et al., 1992; Schaltegger et al., 1999) point to high pressures and temperatures, attained at 335–337 Ma (Schaltegger et al., 1999). Like most granulites of the Bohemian Massif, the granulites of the Vosges are associated with peridotites and occur within the tectonically uppermost unit or nappe (Latouche et al., 1992). In contrast, the granulites of the Schwarzwald form part of a tectonically lower unit and contain only small lenses of eclogite and peridotite (Kalt et al., 2000a, and references therein). Preliminary petrographic and geochronological data pointed to (U)HT equilibration conditions in Early Carboniferous time (Hanel et al., 1993; Kalt et al., 2000a, and references therein). It is against this background that this paper presents the first detailed petrographic, mineral composition and 228 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES thermobarometric data on granulites from the Schwarzwald, located between the Vosges and the Bohemian Massif (Fig. 1). A P–T path is derived and a geodynamic model for the formation and exhumation of the granulites is presented. The possible relations with the granulites from the Vosges and the Bohemian Massif as well as implications for interpretative models of granulite formation and exhumation in the Variscan belt are discussed. Geochronological data for these granulites will be presented elsewhere (Kober et al., in preparation). GEOLOGICAL SETTING The Schwarzwald basement is dominated by high-grade gneisses and migmatites intruded by several post-collisional Variscan granites (Fig. 1). The metamorphic rocks can be divided into three units (Hanel et al., 1999; Fig. 1). Common to all three units is an HT–LP metamorphic overprint at >330 Ma (Kalt et al., 1994a; Lippolt et al., 1994), with P–T conditions of 730–780°C at 0·4–0·45 GPa (Kalt et al., 1994a, 2000a). Unit 1 consists mainly of migmatites and gneisses (e.g. Wimmenauer, 1984) that show no evidence of an earlier metamorphic stage. Unit 2 is more varied, with gneisses, amphibolites, calcsilicate rocks and marbles. In places, it contains relics of a medium-temperature–mediumpressure (MT–MP) stage such as kyanite + staurolite + quartz + garnet, indicating former metamorphic conditions of 550–650°C and 0·5 GPa minimum pressure (Rehfeld, 1983). The gneisses of unit 3 show relics of an earlier granulite-facies stage that comprised the assemblage garnet + rutile + kyanite + antiperthite. Units 1 and 3 contain lenses of eclogite, spinel peridotite and garnet–spinel peridotite (Klein & Wimmenauer, 1984; Hanel et al., 1993; Kalt et al., 1995; Kalt & Altherr, 1996). These lenses appear to be absent in unit 2. The metamorphic units of the Schwarzwald occur within two crystalline blocks, the Central Schwarzwald Gneiss Complex (CSGC) and the Southern Schwarzwald Gneiss Complex (SSGC; Eisbacher et al., 1989), separated by the east–west-trending fault zone of Badenweiler– Lenzkirch (BLZ; Fig. 1). In the CSGC, all three metamorphic units can be found whereas in the SSGC, unit 1 is absent (Fig. 1). The BLZ is a fault and shear zone that was active during both the compressional and the extensional stage of the Variscan orogeny (Echtler & Chauvet, 1991–1992). By 330 Ma, the BLZ was no longer active, when the southern Schwarzwald granites intruded the SSGC and sealed the BLZ (Schaltegger, 2000). The metamorphic grade of the sedimentary and volcanic rocks within the fault zone trends from amphibolite facies at the northern contact to the CSGC to non-metamorphic in the south. The geodynamic character of the BLZ is not yet clear. It may have been a syn-orogenic basin (Echtler & Altherr, 1993), which was filled by flysch sediments derived from the surrounding Variscan mountains during the early Carboniferous (Visean) and which closed during the final compressional stage of the orogeny. Other workers have proposed the BLZ to be a major suture zone separating two crustal segments, and to be the site of Variscan subduction (Loeschke et al., 1998). This model is based on the occurrence of greywackes of late Devonian and early Carboniferous age in the BLZ, which contain detritus of volcanic rocks and chromite grains thought to be derived from obducted ophiolites. The three units of the Schwarzwald with their different metamorphic histories were probably juxtaposed between the granulite-facies stage of unit 3 and the HT–LP stage. The mechanism of this juxtaposition remains unclear. Hanel & Wimmenauer (1990) found evidence for a nappe complex in the CSGC and suggested that unit 3 forms a nappe thrust over the gneisses of unit 2. Both units form a tectonic window within the uppermost nappe (unit 1). In the SSGC, a nappe structure was described by Hann & Sawatzki (2000), with the Wehra–Wiesental complex (Fig. 1) thrust over unit 3, which in turn rests on top of unit 2. SAMPLE SELECTION AND ANALYTICAL TECHNIQUES The samples investigated in our study were taken from unit 3 in the SSGC and in the CSGC. In the SSGC, felsic, garnet–orthopyroxene-bearing granulites alternate with mafic, clinopyroxene-bearing granulites. At some localities, the rocks show a foliation-parallel compositional layering, with felsic and mafic layers alternating on centimetre to decimetre scale. This texture was interpreted as a pre-metamorphic layering of rhyolitic and basaltic volcanic rocks (Lammlin, 1981). We studied granulites ¨ from five localities within the SSGC gneiss area (Fig. 1). The results presented here focus on four samples (Table 1), containing clinopyroxene + orthopyroxene (samples 14893 and 2101), garnet + orthopyroxene (sample 2100) and garnet + sillimanite (sample TM-53). The occurrence of kyanite in the SSGC was described by Metz (1980) from loose blocks near Todtmoos, but no kyanitebearing rocks were found in the course of our study. Unit 3 in the CSGC is dominated by pelitic to psammitic gneisses containing relics of a granulite-facies stage. Garnet, hercynite–garnet and rutile–kyanite–garnet granulites were sampled at about 15 localities. The investigations concentrated on the last group, showing the greatest variety of prograde and retrograde reaction textures. The description and discussion will focus on 229 JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 FEBRUARY 2003 Table 1: Sample localities, mineral assemblages and reaction textures of investigated granulite samples Sample Complex Unit Locality Type Paragenesis 14893 SSGC 3 North of Happach Cpx–Opx Cpx, Opx, Atp 2101 SSGC 3 Todtmoos–Schwarzenbach Cpx–Opx Cpx, Opx 2100 SSGC 3 Todtmoos–Schwarzenbach Grt–Opx Grt, Opx TM-53 SSGC 3 Todtmoos–Mattle ¨ Grt–Sil Reaction textures Grt, Sil 1, 2 G-O/2 CSGC 3 Hochkopf, Gengenbach Rt–Ky–Grt Grt, Rt, Ky, Atp La-1b CSGC 3 Hohengeroldseck, Lahr Rt–Ky–Grt Grt, Rt, Ky, Atp 2, 9, 10, 11, 13, 14 2, 3, 4, 5, 6, 8, 9, 10, 14 Mue-1a CSGC 3 Muhlenbach ¨ Rt–Ky–Grt Grt, Rt, Ky 2, 5, 6, 9, 10, 11, 14 Sample names as used throughout the text. SSGC, Southern Schwarzwald Gneiss Complex; CSGC, Central Schwarzwald Gneiss Complex; Unit 3, unit with granulite-facies relics (see Fig. 1); types of granulite are named as in the text. Mineral abbreviations after Kretz (1983); Atp, antiperthite. Numbers of reaction textures refer to reactions discussed in the text: 1, Qtz-free Opx + Pl corona arround garnet; 2, Ilm + Pl + Qtz aggregates in Grt-bearing rocks; 3, high-Ca garnets with Pl + Qtz inclusions showing negative crystal shapes; 4, Hc and Ky inclusions in Grt; 5, Ky inclusions in Grt; 6, Ky and Bt inclusions in Grt; 8, Pl + Hc symplectites; 9, Crd + Hc symplectites; 10, Crd coronas around Grt and Ky; 11, Pl coronas around Grt and Ky; 13, euhedral Crn in Crd + Hc symplectites around Ky; 14, Crd + Qtz + Bt symplectites around Grt surrounded by Kfs. three typical samples (G-O/2, La-1b and Mue-1a; Table 1). The compositions of mineral phases were determined with a Cameca SX 51 electron microprobe at the Mineralogisches Institut, Heidelberg, equipped with five wavelength-dispersive spectrometers. Operating conditions were 20 nA beam current and 15 kV acceleration voltage. The electron beam was defocused to 5–10 m for feldspar analyses to avoid loss of alkalis. Counting time was 10 s on peak and 10 s on background for all elements except Ba (20 s) and Zn (30 s). PAP correction was applied to the data. Natural and synthetic oxide and silicate standards were used for calibration. PETROGRAPHY, MINERAL COMPOSITIONS AND REACTIONS Clinopyroxene–orthopyroxene granulites The investigated clinopyroxene-orthopyroxene granulites (samples 14893 and 2101) are mostly fine-grained rocks of basic composition, consisting of plagioclase, orthopyroxene and clinopyroxene with minor K-feldspar and quartz. Accessory phases are apatite, ilmenite and pyrite. Some of the granulites show foliation and isoclinal folding. They are equigranular with K-feldspar, plagioclase and quartz forming equilibrium grain boundaries, indicating recrystallization after deformation (sample 2101). Others contain large crystals of plagioclase with K-feldspar lamellae that did not recrystallize (sample 14893; Fig. 2a). The chemical compositions of these feldspar hosts (An35–42) and lamellae (Or86–92) are similar to those of the surrounding small plagioclase and K-feldspar grains, respectively (Table 2). The reintegrated bulk composition of the exsolved feldspars is Or13–18Ab52–56An31–33 (sample 14893). In all investigated samples, orthopyroxene grains are generally small (<100 m) with Mg-number [= 100 × Mg/(Fetot + Mg)] between 44 and 52 and Al2O3 contents between 0·8 and 1·6 wt % (Table 3). Clinopyroxene is a diopside–hedenbergite solid solution with Mg-number between 57 and 68, 0·3–0·5 wt % Na2O, and >2 wt % Al2O3 (Table 3). Larger grains of clinopyroxene (500 m) show exsolution lamellae of orthopyroxene. Both orthopyroxene and clinopyroxene grains are surrounded by greenish amphibole and reddish-brown biotite. The amphibole is a pargasitic hornblende with >1 wt % K2O, 2 wt % TiO2 and Mg-number of 45–50. Biotite contains >5 wt % TiO2. Its Mg-number is also between 45 and 50. These features are interpreted to indicate that during peak metamorphic conditions, the clinopyroxene–orthopyroxene granulites consisted of clinopyroxene with a considerable enstatite–ferrosilite component, orthopyroxene, ternary feldspar and quartz. During retrograde cooling and hydration, clinopyroxene exsolved orthopyroxene lamellae, feldspar exsolved K-feldspar lamellae and amphibole and biotite formed around pyroxene. Garnet–orthopyroxene granulites 230 The garnet–orthopyroxene granulites (sample 2100) are also fine-grained rocks consisting of plagioclase, K-feldspar, quartz, orthopyroxene, ilmenite and rare garnet MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES Fig. 2. Crossed-polar photomicrograph (a) and back-scattered electron images (b)–(f ) of important microtextures in granulites of the CSGC and the SSGC. Mineral abbreviations are according to Kretz (1983). (a) Lamellae of potassium feldspar in plagioclase of a granulite from the SSGC, indicating a former ternary feldspar. (b) Inclusions of kyanite and hercynite in type 2 garnet from a metapelitic granulite of the CSGC. (c) Plagioclase–hercynite symplectite inside a garnet grain from a metapelitic granulite of the CSGC, accompanied by biotite and an aluminium hydroxide (diaspore, DSp), formed by reaction (7). (d) Cordierite–hercynite symplectite around kyanite from a metapelitic granulite of the CSGC, probably formed by reactions (9)–(13). (e) Cordierite–hercynite symplectite around garnet from a metapelitic granulite of the CSGC, probably formed by reactions (9)–(13). (f ) Small irregular patches of intimately intergrown plagioclase and quartz between garnet and K-feldspar from a metapelitic granulite of the CSGC, probably formed by reaction (14). (For further explanation, see section on petrography and mineral compositions.) porphyroclasts with small inclusions of rutile. It is interpreted that no ternary feldspar is preserved because of recrystallization. In sample 2100, garnet is surrounded by coronas of plagioclase and orthopyroxene devoid of quartz (Fig. 3). In some parts of the rock, garnet is absent but coarse-grained plagioclase domains intergrown with ilmenite and orthopyroxene indicate its former presence. Orthopyroxene is a ferrosilite–enstatite solid solution with Mg-number between 38 and 52 (Table 3). The Al2O3 content varies between 08 and 2·3 wt % and generally decreases from core to rim (Fig. 4). Variations in Mg-number and Al2O3 content are observed with different textural positions. In the plagioclase–orthopyroxene coronas around garnet, orthopyroxene shows the highest Al2O3 content and Mgnumber between 45 and 52. In the domains where garnet is absent, and in the matrix, Al2O3 content is between 1·0 and 1·5 wt % and Mg-number is 38–40. The An content of matrix plagioclase is 24–27 mol %. Plagioclase grains in the coronas with orthopyroxene 231 JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 FEBRUARY 2003 Table 2: Representative analyses of plagioclase, K-feldspar and recalculated ternary feldspar in granulite samples from the SSGC Sample: 14893 Texture: host 2100 host lamella lamella ternary ternary at Grt at Grt at Opx matrix matrix matrix matrix SiO2 58·81 58·47 64·32 64·17 59·52 59·57 58·50 58·69 60·34 61·65 61·78 64·51 64·55 Al2O3 25·58 25·43 17·92 17·94 24·19 24·50 26·88 26·88 25·55 24·58 24·58 19·24 19·26 FeOT 0·12 0·11 0·04 0·00 0·08 0·11 0·25 0·04 0·10 0·13 0·06 0·06 0·01 CaO 8·20 8·07 0·03 0·08 6·80 6·92 7·98 7·94 6·31 5·54 5·35 0·05 BaO 0·00 0·02 0·49 0·53 0·10 0·07 Na2O 7·06 6·80 0·95 1·10 5·93 6·14 6·99 6·97 7·79 8·24 8·37 1·56 1·57 K2O 0·26 0·39 14·86 14·99 2·71 2·41 0·35 0·36 0·48 0·53 0·61 14·21 13·84 100·03 99·28 98·61 98·83 99·31 99·72 100·94 100·88 100·57 100·66 100·74 99·63 99·32 Total n.d. n.d. n.d. n.d. n.d. n.d. 0·08 n.d. Structural formula on the basis of 8 oxygens Si 2·632 2·633 3·008 3·002 2·696 2·685 2·595 2·602 2·673 2·722 2·726 2·970 2·975 Al 1·349 1·350 0·988 0·989 1·289 1·300 1·405 1·404 1·334 1·279 1·278 1·044 1·046 Fe3+ 0·004 0·004 0·002 0·000 0·003 0·004 0·009 0·002 0·004 0·005 0·002 0·002 0·000 Ca 0·393 0·389 0·002 0·004 0·328 0·333 0·379 0·377 0·300 0·262 0·253 0·003 Ba 0·000 0·000 0·009 0·010 0·002 0·001 Na 0·612 0·594 0·086 0·100 0·518 0·534 0·601 0·599 0·669 0·705 0·716 0·139 0·140 K 0·015 0·022 0·887 0·895 0·161 0·142 0·020 0·021 0·027 0·030 0·035 0·835 0·814 Total 5·005 4·993 4·981 5·000 4·996 4·999 5·008 5·005 5·006 5·003 5·010 4·993 4·979 n.d. n.d. n.d. n.d. n.d. n.d. 0·004 n.d. Calculation of end-members anorthite, albite, orthoclase and celsian An 38·5 38·7 0·2 0·4 32·5 33·0 37·9 37·8 30·1 26·3 25·2 0·3 0·4 Ab 60·0 59·0 8·7 9·9 51·4 52·9 60·1 60·1 67·2 70·7 71·4 14·3 14·7 Or 1·5 2·2 90·2 88·7 15·9 14·0 Csa 0·0 0·0 0·9 1·0 0·2 0·1 2·0 n.d. 2·1 n.d. 2·7 n.d. 3·0 n.d. 3·4 n.d. 85·5 84·9 n.d. n.d. Element analyses (oxide wt %). n.d., not detected. FeOT, total iron content as FeO. Textures: host, plagioclase host crystals; lamella, K-feldspar exsolution lamella; ternary, recalculated ternary feldspar compositions (see text for explanation); at Grt, plagioclase in quartz-free coronas in contact with garnet; at Opx, plagioclase in quartz-free coronas in contact with orthopyroxene; matrix, plagioclase and K-feldspar in Opx–Pl–Kfs–Qtz matrix. show a compositional gradient from An25 in the matrix to An38 at the contact with garnet (Fig. 5). Garnet shows a broad chemical plateau in its core with no indication of prograde zoning (Fig. 6). Core compositions are Alm56Prp23·5Grs19Sps1·5 (Mg-number 30), with 0·18 wt % TiO2 (Table 4). The outer 250 m of the garnet crystals are characterized by zonation with higher Fe and Mn and lower Ca, Mg and Ti contents. The typical rim composition is Alm70Prp14Grs12Sps4 (Mg-number 17; Table 4), with Ti below detection limit. Locally, biotite (Mg-number 45–50; 4 wt % TiO2) formed around orthopyroxene and garnet as a result of retrogression and hydration. The observed retrograde zoning in garnet and the quartz-free coronas of plagioclase and orthopyroxene are interpreted to have formed during decompression by the reaction Grt + Qtz = Opx + Pl. (1) Our preferred interpretation is that before decompression, the assemblage Grt + ternary feldspar + Qtz + Opx was stable, with ternary feldspar containing significant Or-component as a result of high temperatures. The zoning patterns of corona plagioclase, garnet and orthopyroxene are a result of continuous reaction under decreasing pressures and falling temperatures. Before decompression and cooling, the An content of the feldspar must have been lower than 24–27 mol % (as retained in the matrix plagioclase) as a result of a higher Or component. At this stage, garnet had 19 mol % grossular and Mg-number of 30, as recorded by the core plateau. The stability of orthopyroxene in the rock is generally limited by pressure. With rising pressure, the 232 0·09 16·38 0·62 50·24 0·07 1·45 0·01 30·23 0·63 16·22 0·61 0·02 0·00 99·49 SiO2 TiO2 Al2O3 Cr2O3 FeOT MnO MgO CaO Na2O K 2O Total 233 99·75 0·00 0·48 21·54 12·29 0·25 11·40 0·10 1·63 0·11 51·95 Cpx 99·51 0·01 0·32 21·68 12·31 0·28 11·34 0·05 1·65 0·16 51·72 Cpx 99·37 0·02 0·03 0·72 15·25 0·85 31·27 0·04 1·15 0·11 49·94 Opx 2101 99·37 0·00 0·00 0·74 15·26 0·87 31·23 0·02 1·19 0·08 49·97 Opx 0·001 1·958 68·2 65·8 0·000 0·035 0·870 0·690 0·008 0·322 0·037 0·003 0·072 0·003 1·957 67·6 65·9 0·000 0·023 0·879 0·694 0·009 0·333 0·026 0·001 0·074 0·004 1·961 47·0 46·5 0·001 0·002 0·030 0·893 0·028 1·007 0·020 0·001 0·053 0·003 1·962 46·9 46·6 0·000 0·000 0·031 0·894 0·029 1·010 0·015 0·000 0·055 0·002 63·3 61·4 0·000 0·038 0·860 0·638 0·015 0·370 0·031 0·001 0·088 0·005 1·954 99·84 0·00 0·51 21·19 11·29 0·45 12·67 0·04 1·96 0·17 51·55 Cpx 60·7 60·6 0·000 0·027 0·865 0·625 0·015 0·405 0·002 0·000 0·097 0·006 1·958 99·25 0·00 0·36 21·12 10·97 0·46 12·73 0·00 2·16 0·21 51·23 Cpx 52·0 50·5 0·001 0·001 0·029 0·984 0·025 0·907 0·058 0·001 0·046 0·002 1·946 99·45 0·02 0·01 0·70 17·00 0·76 29·73 0·05 1·00 0·08 50·12 Opx 2100 50·5 48·6 0·000 0·002 0·023 0·951 0·019 0·933 0·071 0·003 0·069 0·003 1·926 99·44 0·00 0·02 0·54 16·37 0·57 30·82 0·09 1·50 0·11 49·42 Opx 47·8 46·2 0·000 0·000 0·021 0·901 0·022 0·984 0·066 0·000 0·078 0·003 1·925 100·12 0·00 0·00 0·51 15·51 0·68 32·20 0·01 1·70 0·11 49·40 Opx 45·9 44·4 0·001 0·001 0·017 0·860 0·018 1·014 0·062 0·009 0·105 0·003 1·911 99·57 0·02 0·02 0·41 14·67 0·53 32·71 0·28 2·26 0·10 48·58 Opx 45·1 43·8 0·000 0·002 0·029 0·841 0·018 1·022 0·057 0·009 0·109 0·004 1·909 99·39 0·00 0·02 0·68 14·29 0·54 32·71 0·30 2·34 0·13 48·38 Opx 38·4 37·7 0·001 0·002 0·026 0·731 0·028 1·170 0·039 0·000 0·044 0·002 1·957 99·58 0·02 0·02 0·61 12·22 0·83 36·05 0·00 0·94 0·08 48·82 Opx Element analyses (oxide wt %); Mg-number = 100 × Mg/(Mg + Fetot); Mg value = 100 × Mg/(Mg + Fe2+); FeOT, total iron content as FeO. 49·1 49·6 0·001 48·9 Mg-no. 0·002 0·000 Na K 0·026 0·942 0·025 0·959 0·018 Mg value 49·3 0·942 0·026 0·021 Mn Mg 0·967 Fe2+ Ca 0·018 Fe3+ 0·000 0·071 0·067 0·000 Al Cr 1·954 0·003 1·956 0·002 Si Ti Structural formula on the basis of 6 oxygens, charge balanced to 4 cations 100·37 0·02 0·01 0·77 30·28 0·01 1·56 50·63 Opx Mineral: Opx 14893 Sample: Table 3: Representative analyses of clinopyroxene and orthopyroxene in granulite samples from the SSGC 42·9 42·2 0·000 0·002 0·025 0·822 0·024 1·092 0·031 0·001 0·039 0·001 1·964 99·50 0·00 0·02 0·58 13·90 0·71 33·87 0·04 0·84 0·03 49·53 Opx 44·6 44·1 0·000 0·000 0·025 0·856 0·023 1·063 0·021 0·001 0·044 0·002 1·965 99·83 0·00 0·00 0·60 14·60 0·68 32·96 0·03 0·94 0·08 49·94 Opx MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 FEBRUARY 2003 Fig. 3. Photomicrograph of an eye-shaped reaction texture in a garnet–orthopyroxene granulite from the SSGC (sample 2100). Relic garnet is surrounded by a corona of plagioclase and orthopyroxene, indicating decompression reaction (1). (For further explanation, see section on petrography and mineral compositions.) Diameter of garnet grain is 1·8 mm. Fig. 4. Zoning pattern of an orthopyroxene within a corona around garnet as shown in Fig. 3 (sample 2100). It should be noted that whereas Al shows a symmetrical pattern, the patterns for Fe2+ and Mg are asymmetric. (For further explanation, see section on petrography and mineral compositions and section on P–T conditions.) Mg-number of garnet and orthopyroxene increases. Modal abundance of garnet increases, whereas that of orthopyroxene decreases, until orthopyroxene is entirely consumed. To ascertain if orthopyroxene was part of the paragenesis at peak pressure conditions, the Mg-numbers of garnet cores and whole rock were compared. If no other Fe–Mg-bearing phase than garnet was present at peak conditions, the Mg-number of the garnet cores should be similar to that of the whole rock. Wholerock analyses performed by wavelength-dispersive X-ray fluorescence spectrometry yielded an Mg-number of 38 for sample 2100, which is considerably higher than that 234 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES Fig. 5. Chemical composition of plagioclase within a corona around garnet as illustrated in Fig. 3 (sample 2100). The compositional change is exclusively controlled by the distance from garnet and not by plagioclase grain boundaries. (For further explanation, see section on petrography and mineral compositions and section on P–T conditions.) of garnet cores (Mg-number 30; sample 2100). As biotite and ilmenite clearly formed during retrogression, orthopyroxene must have been stable at peak metamorphic pressures, with high Al2O3 contents and high Mg-number of 52. During decompression, garnet reacted with surrounding quartz, forming new plagioclase and orthopyroxene with decreasing Mg-number of both garnet and orthopyroxene, decreasing Al2O3 contents of orthopyroxene and increasing An content of plagioclase. The decompression was probably accompanied by cooling, which could have additionally influenced the zoning patterns of garnet and orthopyroxene, leading to lower Mg-number of garnet and lower Al2O3 content of orthopyroxene. Higher Mn content of garnet rims together with corroded shape indicates resorption of garnet during retrogression. Rutile is preserved only as inclusions in garnet. Some inclusions in garnet consist of ilmenite + plagioclase + quartz. The matrix is characterized by the occurrence of ilmenite. These textures document the reaction Grt + Rt = Ilm + Pl + Qtz (2) which also is pressure dependent. Garnet–sillimanite granulites The garnet–sillimanite granulites (sample TM-53) show equigranular textures and consist of plagioclase, K-feldspar, quartz, garnet and sillimanite. Garnet is free of inclusions and reaction textures. The rocks show no indication of retrogression or hydration such as hydrous minerals or alteration rims. Sillimanite is euhedral and prismatic. There is no kyanite or evidence of former kyanite. This indicates that the rocks completely equilibrated in the stability field of sillimanite. Rutile–kyanite–garnet granulites These granulites (samples La-1b, G-O/2 and Mue-1a) are characterized by fairly large grains (1–8 mm) of garnet, antiperthite, rutile and kyanite embedded in a fine-grained matrix of quartz, plagioclase, K-feldspar and biotite. The large grains are interpreted as relics of an HP granulite-facies stage, the matrix phases as products of LP–HT metamorphic recrystallization, hydration and crystallization of granitoid melt. The antiperthites are large plagioclase grains containing lamellae of K-feldspar, formed by exsolution upon cooling. Reintegration gives primary compositions of Or18–22Ab57–61An17–23 and Or15–17 Ab63–65An19–21 for samples G-O/2 and La-1b, respectively (Table 5). Two types of garnet can be distinguished with respect to inclusions and composition. Garnet of type 1 contains inclusions of rutile, Ti-rich biotite, plagioclase and quartz. In some type 1 grains small inclusions of ilmenite occur in the cores. Garnet shows core plateaux with the compositions Alm53Prp24Grs22Sps1 (G-O/2) and Alm65Prp26Grs6·5Sps2·5 (La-1b; Table 4). In samples where garnet grains have diameters exceeding 4 mm they show prograde zoning patterns with Mn-rich cores (Alm59Prp14Grs9Sps18, Mue-1a; Fig. 7). The outer 200 m 235 JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 FEBRUARY 2003 Fig. 6. Zoning pattern of a garnet grain in the centre of a corona as illustrated in Fig. 3 (sample 2100). (For further explanation, see section on petrography and mineral compositions and section on P–T conditions.) of all grains show zonations with Mn and Fe contents increasing and Mg and Ca contents decreasing rimwards, leading to the compositions Alm77Prp7·5Grs6·5Sps9 (G-O/ 2) and Alm79Prp13Grs2·5Sps5·5 (La-1b; Table 4). Inclusions of plagioclase + quartz show negative crystal shapes, indicating that the host garnet controlled their shape during its own growth. This may suggest the presence of melts or fluids during garnet formation, facilitating the development of negative crystals. These fluids or melts could have been derived from the prograde breakdown of biotite coexisting with quartz. Because of their Ca-rich composition, type 1 garnets must have formed from a Ca-rich phase, presumably plagioclase. Gardien et al. (2000) produced garnet experimentally at high pressures by the dehydration reaction Bt + high-Ca Pl + Qtz = Grt + low-Ca Pl + liquid/ vapour (L/V). (3) Restitic biotite included in garnet in the rocks studied here contains >4 wt % TiO2, so that its breakdown in the presence of quartz could have produced the ilmenite and rutile found as inclusions within garnet. In contrast to type 1, type 2 garnet is Ca-poor throughout the grains. It includes rutile, graphite, kyanite and biotite or kyanite and hercynite (Fig. 2b), and some garnet grains are in turn included in kyanite. The core composition is Alm75Prp19Grs2·5Sps3·5 (La-1b). Rims have lower Mg and higher Mn and Fe contents (Alm79Prp12Grs2·5Sps6·5; La-1b) than cores. The hercynite inclusions contain about 4·5 wt % ZnO and 4 wt % MgO (La-1b; Table 6). The kyanite and hercynite inclusions require an Al-rich phase as the source of type 2 assemblages. In Fe-rich pelites, staurolite is a likely source of garnet and might explain the high Zn contents of the hercynite. The reaction 236 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES Table 4: Representative analyses of garnet in granulite samples from the SSGC and CSGC Sample: 2100 GO/2 core core–rim rim core La-1b rim core Mue1a rim core Mg max∗ Mn min∗ rim SiO2 38·26 37·94 36·43 38·27 36·62 38·52 37·39 37·10 38·22 38·08 TiO2 0·18 0·14 0·08 0·13 0·01 0·05 0·00 0·06 0·03 0·02 37·60 0·09 Al2O3 20·96 20·69 21·94 21·58 20·86 21·57 20·95 20·44 21·21 21·17 20·91 Cr2O3 0·02 0·00 0·00 0·03 0·00 0·02 0·00 0·05 0·01 0·02 0·07 FeOT 25·92 28·25 31·40 25·09 34·70 30·24 35·37 26·92 28·62 30·60 32·61 MnO 0·64 1·18 1·92 0·62 4·00 0·92 2·54 8·26 3·67 1·72 2·62 MgO 6·14 5·09 3·55 6·48 1·88 7·09 3·38 3·53 6·99 6·32 4·36 1·97 CaO 6·98 6·38 4·15 8·17 2·27 2·07 0·78 3·56 1·35 1·94 Na2O 0·04 0·02 0·02 0·00 0·01 0·00 0·00 0·00 0·00 0·02 0·00 K2O 0·03 0·00 0·01 0·01 0·02 0·00 0·03 0·00 0·00 0·00 0·00 Total 99·16 99·68 99·50 100·38 100·37 100·47 100·44 99·91 100·09 99·90 100·23 Structural formula on the basis of 12 oxygens Si 3·005 3·001 2·926 2·966 2·973 2·999 3·002 2·987 3·000 3·002 2·998 Ti 0·011 0·008 0·005 0·007 0·001 0·003 0·000 0·003 0·002 0·001 0·005 Al 1·941 1·929 2·077 1·971 1·996 1·979 1·983 1·940 1·962 1·967 1·965 Cr 0·001 0·000 0·000 0·002 0·000 0·001 0·000 0·003 0·000 0·001 0·004 Fe2+ 1·702 1·869 2·109 1·626 2·356 1·969 2·375 1·812 1·878 2·017 2·175 Mn 0·042 0·079 0·131 0·041 0·275 0·061 0·173 0·563 0·244 0·115 0·177 Mg 0·719 0·600 0·425 0·748 0·228 0·823 0·405 0·424 0·818 0·743 0·518 Ca 0·587 0·540 0·357 0·679 0·198 0·172 0·067 0·307 0·114 0·164 0·169 Na 0·007 0·002 0·002 0·000 0·002 0·000 0·001 0·000 0·001 0·004 0·000 K 0·003 0·000 0·001 0·001 0·002 0·000 0·003 0·000 0·000 0·000 0·000 Total 8·018 8·028 8·032 8·041 8·030 8·008 8·008 8·039 8·018 8·014 8·012 Calculation of end-members almandine, pyrope, grossular, spessartine and uvarovite Alm 55·8 60·5 69·8 52·6 77·1 65·1 78·6 58·3 61·5 66·4 71·6 Prp 23·6 19·4 14·1 24·2 7·5 27·2 13·4 13·6 26·8 24·4 17·1 Grs 19·2 17·5 11·8 21·9 6·5 5·7 2·2 9·7 3·7 5·3 5·3 Sps 1·4 2·5 4·3 1·3 9·0 2·0 5·7 18·1 8·0 3·8 5·8 Uvr 0·0 0·0 0·0 0·1 0·0 0·0 0·0 0·1 0·0 0·1 0·2 ∗See Fig. 6. Element analyses (oxide wt %). FeOT, total iron content as FeO. Mineral abbreviations after Kretz (1983). St = Grt + Ky + Hc + V (4) could have formed type 2 garnets, hercynite and kyanite in quartz-absent domains. In the presence of quartz, the reaction St + Qtz = Grt + Ky + V (5) probably formed garnets without hercynite, and the reaction St + Qtz + Ms = Grt + Ky + Bt + V (6) could have led to the formation of biotite inclusions in garnet and kyanite. Biotite inclusions could also be relics of a garnet-forming, biotite breakdown reaction. Later decomposition of garnet took place via a number of reactions. There are three kinds of breakdown assemblage associated with relict garnet grains: (1) symplectites of plagioclase and hercynite (Fig. 2c); (2) symplectites of cordierite and hercynite (Fig. 2d and e); (3) symplectites of cordierite, quartz and biotite (Fig. 2f ). Plagioclase–hercynite symplectites are always surrounded by garnet that is not in contact with quartz (Fig. 2c). In some grains, the symplectite is accompanied by biotite (Fig. 2c) and an Al hydroxide phase (86 wt % 237 100·19 99·61 100·21 0·25 7·75 100·62 0·16 7·78 238 5·013 Total 0·001 4·998 0·006 0·702 0·295 2·673 5·003 0·014 0·668 0·323 0·016 1·310 2·628 5·021 0·009 0·670 0·332 0·024 1·358 2·743 5·018 0·014 0·763 0·251 0·001 1·246 5·020 0·014 0·770 0·246 0·002 1·244 2·745 99·75 0·25 8·92 0·9 Or 29·4 0·6 70·0 32·1 1·4 66·5 32·8 0·9 66·3 24·4 1·4 74·2 23·9 1·4 74·8 89·8 9·9 0·3 4·987 0·884 0·098 0·003 0·002 0·998 3·003 99·09 1·42 86·7 13·0 0·2 4·980 0·847 0·127 0·002 0·000 0·996 3·007 99·00 14·35 18·2 60·7 21·1 4·999 0·182 0·610 0·212 0·000 1·200 2·795 99·18 3·10 7·02 4·42 15·8 64·0 20·2 4·996 0·158 0·641 0·202 0·000 1·185 2·810 99·17 2·68 7·39 4·22 1·8 73·3 24·9 4·995 0·018 0·731 0·248 0·000 1·239 2·759 100·27 0·33 8·53 5·24 0·01 1·3 74·3 24·4 5·011 0·014 0·760 0·249 0·001 1·229 2·759 100·35 0·24 8·86 5·25 0·03 85·0 14·4 0·6 4·987 0·834 0·141 0·006 0·002 1·012 2·992 99·74 14·22 1·59 0·13 0·04 18·69 65·08 83·9 15·6 0·5 4·977 0·816 0·152 0·005 0·000 0·999 3·004 100·02 13·99 1·72 0·09 0·01 18·53 65·68 16·6 63·2 20·2 4·991 0·165 0·629 0·201 0·001 1·189 2·807 100·44 2·82 7·35 4·25 0·03 22·76 63·23 15·8 64·9 19·3 5·003 0·160 0·655 0·195 0·001 1·183 2·809 100·70 2·75 7·67 4·14 0·04 22·69 63·41 ternary 0·5 40·1 59·4 5·020 0·005 0·413 0·610 0·016 1·598 2·379 100·16 0·09 4·69 12·57 0·41 29·91 52·49 at Grt 1·0 62·7 36·3 5·007 0·010 0·635 0·367 0·008 1·350 2·636 99·88 0·18 7·32 7·66 0·22 25·60 58·91 matrix Element analyses (oxide wt %). FeOT, total iron content as FeO; Textures: Pl–Ky, plagioclase in coronas surrounding kyanite; Pl–Hc, plagioclase in plagioclase–hercynite symplectites; host, plagioclase host crystals; lamella, K-feldspar exsolution lamella; ternary, recalculated ternary feldspar compositions (see text for explanation); at Grt, plagioclase in contact with garnet; matrix, plagioclase in matrix. 26·3 72·8 An Ab 1·09 14·93 0·05 0·01 23·58 62·39 lamella lamella ternary NUMBER 2 Calculation of end-members anorthite, albite, orthoclase and celsian 0·738 0·009 Na Ca 2·713 1·281 99·94 0·25 8·86 0·06 0·01 23·77 62·39 host Mue-1a VOLUME 44 K 0·003 0·267 Fe3+ 2·720 1·277 Si Al Structural formula on the basis of 8 oxygens Total 0·11 8·12 5·15 0·00 22·39 62·48 8·58 5·27 0·04 22·62 62·01 0·16 6·97 0·06 18·25 64·94 K 2O 6·77 0·02 18·26 64·72 Na2O 6·17 0·64 23·71 61·66 5·61 0·42 23·79 61·74 CaO 0·02 25·93 59·15 0·07 24·97 60·05 FeOT 24·37 60·83 24·43 ternary 61·34 lamella lamella ternary SiO2 host Al2O3 host host Pl–Hc Pl–Ky Texture: Pl–Ky Pl–Hc La-1b Sample: G-O/2 Table 5: Representative analyses of plagioclase, K-feldspar and recalculated ternary feldspar in granulite samples from the CSGC JOURNAL OF PETROLOGY FEBRUARY 2003 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES Fig. 7. Zoning pattern of a large garnet grain of type 1 from a metapelitic granulite of the CSGC (sample Mue-1a). (For further explanation, see section on petrography and mineral compositions and section on P–T conditions.) Al2O3) replacing hercynite. Hercynite is chemically slightly different from the one included in type 2 garnet in containing less ZnO (3·5 wt % in La-1b; Table 6). Plagioclase has An content of >33 mol % in contrast to >25 mol % in the matrix (sample La-1b; Table 5). The garnet shows atoll-like shapes cross-cut by more or less parallel fractures (Fig. 2c). The atolls display retrograde zoning with higher Mn and Fe contents at the rims and higher Mg content in the centres (Alm74Prp20Grs2·5Sps3·5; sample La-1b). The non-radial arrangement of the fractures and the retrograde zoning suggest that the garnet atolls resulted from the decomposition of former large garnet grains rather than representing newly formed crystals around the symplectites. The symplectites inside the garnet crystals indicate that garnet may have reacted with inclusions. Kyanite inclusions could be a potential source of Al for the reaction. However, none of the observed contacts of kyanite and garnet show plagioclase–hercynite symplectites (e.g. Fig. 2b) and none of the plagioclase–hercynite symplectites contain relics of kyanite. Therefore, it is unlikely that a reaction between kyanite and garnet formed the symplectites and that kinetic factors inhibited the reaction in some cases. It is more likely that garnet included corundum at high pressures and that the symplectites were formed during decompression by the reaction Grt + Crn = Pl + Hc. (7) The presence of biotite and the An content of plagioclase (33 mol %; sample La-1b) indicate that the reaction was not isochemical, but that K+, Na+ and H2O were added. The complete reaction probably was Grt + Crn + K+ + Na+ + H2O = Pl + Hc + Bt. (8) The cordierite–hercynite symplectites occur around kyanite (Fig. 2d), garnet (Fig. 2e) and sillimanite pseudomorphs after kyanite. Thus, they seem to reflect a reaction of garnet and Al2SiO5 at silica-undersaturated conditions. However, as mentioned above, the rocks show intergrowths of garnet and kyanite as inclusions in one another without any reaction textures. The symplectites occur only where garnet as well as kyanite are in contact with the matrix. This suggests either that formation of symplectite was inhibited by kinetic factors and only possible where kyanite and garnet were in contact with fluid or melt, or that a third phase was involved in the reaction. The occurrence of biotite grains with resorbed grain boundaries in the vicinity of the symplectites makes biotite a probable candidate. The reaction Grt + Als + Bt = Crd + Hc + L (9) was reported by Kriegsmann & Hensen (1998) from migmatites. In the metapelitic granulites of the Schwarzwald, some melt probably remained in situ, and crystallized as fine-grained quartz–plagioclase–K-feldspar–biotite domains. The melt-forming reaction (9) was probably restricted to silica-undersaturated domains. It 239 JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 FEBRUARY 2003 Table 6: Representative analyses of hercynite in granulite samples from the CSGC Sample: G-O/2 Mue-1a La-1b Texture: Crd–Hc Crd–Hc Crd–Hc Crd–Hc Crd–Hc Crd–Hc Hc–Grt Hc–Grt Hc–Pl Hc–Pl SiO2 0·04 0·06 0·11 0·07 0·02 0·03 0·04 0·04 0·02 TiO2 0·00 0·00 0·04 0·08 0·06 0·06 0·05 0·05 0·00 0·01 0·05 Al2O3 59·14 57·94 59·33 59·52 58·33 58·73 58·41 57·91 60·02 58·81 Cr2O3 0·00 0·12 0·23 0·18 0·08 0·11 0·18 0·43 0·03 0·10 FeOT 34·08 37·98 29·42 31·63 36·99 36·07 32·43 31·18 31·06 32·33 MnO 0·53 0·58 0·17 0·17 0·23 0·27 0·13 0·12 0·12 0·40 ZnO 1·22 0·66 3·86 3·02 1·78 1·71 4·66 4·20 3·43 3·53 MgO 4·19 1·96 5·16 4·81 2·82 2·97 3·46 4·53 5·14 4·04 CaO 0·00 0·01 0·02 0·01 0·03 0·00 0·00 0·01 0·00 0·04 Na2O 0·01 0·02 0·09 0·08 0·01 0·04 0·09 0·09 0·12 0·07 K2O 0·08 0·00 0·01 0·05 0·02 0·01 0·04 0·00 0·00 0·01 Total 99·28 99·34 98·45 99·62 100·57 100·15 99·48 98·56 100·06 99·52 Structural formula on the basis of 4 oxygens, charge balanced to 3 cations Si 0·001 0·002 0·003 0·002 0·000 0·001 0·001 0·001 0·001 0·000 Ti 0·000 0·000 0·001 0·002 0·001 0·001 0·001 0·001 0·000 0·001 Al 1·974 1·967 1·984 1·974 1·953 1·967 1·965 1·952 1·978 1·969 Cr 0·000 0·003 0·005 0·004 0·002 0·002 0·004 0·010 0·001 0·002 Fe3+ 0·028 0·027 0·009 0·022 0·044 0·030 0·035 0·037 0·027 0·030 Fe2+ 0·779 0·888 0·689 0·722 0·835 0·827 0·739 0·709 0·700 0·738 Mn 0·013 0·014 0·004 0·004 0·006 0·006 0·003 0·003 0·003 0·010 Zn 0·026 0·014 0·081 0·063 0·037 0·036 0·098 0·089 0·071 0·074 Mg 0·177 0·084 0·218 0·202 0·120 0·126 0·147 0·193 0·214 0·171 Ca 0·000 0·000 0·001 0·000 0·001 0·000 0·000 0·000 0·000 0·001 Na 0·000 0·001 0·005 0·004 0·001 0·002 0·005 0·005 0·007 0·004 K 0·003 0·000 0·000 0·002 0·001 0·000 0·001 0·000 0·000 0·000 Calculation of end-members hercynite, magnetite, spinel and gahnite Hc 78·0 88·7 69·3 72·1 82·0 82·2 73·3 69·7 69·7 Mag 1·4 1·3 0·5 1·1 2·2 1·5 1·7 1·9 1·3 1·5 Spl 18·0 8·5 22·1 20·4 12·1 12·7 15·0 19·5 21·8 17·4 Gah 73·6 2·6 1·4 8·2 6·4 3·8 3·6 10·0 9·0 7·2 7·5 Mg-no. 18·5 8·7 24·1 21·8 12·5 13·2 16·6 21·4 23·5 18·8 Mg value 18·0 8·4 23·8 21·3 12·0 12·8 16·0 20·6 22·8 18·2 Element analyses (oxide wt %). Mg-number = 100 × Mg/(Mg + Fetot); Mg value = 100 × Mg/(Mg + Fe2+); FeOT, total iron content as FeO. Mineral abbreviations after Kretz (1983). Gah, gahnite. produced a granitoid melt and cordierite–hercynite symplectites as restite. The symplectites are always enclosed in coronas of cordierite and sometimes also of plagioclase (Fig. 2d). The cordierite coronas isolate the quartzbearing matrix from the hercynite in the symplectites. They could either be the product of a melt-forming reaction under silica-saturated conditions or they could have formed by the reaction Grt + Als + Qtz = Crd. (10) The plagioclase corona might be a result of the reaction Grt + Als + Qtz = Pl (11) or of late exchange between cordierite and surrounding melt, as postulated for example by Waters (1991): Crd + Na+ + Ca2+ ± Si4+ = Pl + Mg2+ + Fe2+. (12) The two types of symplectites can be clearly distinguished from one another, and there is no transition 240 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES Table 7: Representative analyses of cordierite in granulite samples from the CSGC Sample: G-O/2 Mue-1a Texture: Qtz–Bt Qtz–Bt Crd Crd Crd–Hc Crd–Hc Crd–Hc Crd–Hc Crd Crd SiO2 47·65 47·73 47·63 47·51 47·33 47·02 49·26 48·58 48·65 48·74 TiO2 0·02 0·00 0·03 0·00 0·05 0·02 0·00 0·03 0·00 0·00 Al2O3 32·23 32·16 32·33 32·10 32·22 32·23 33·20 32·72 32·67 32·93 Cr2O3 0·03 0·04 0·00 0·00 0·03 0·05 0·03 0·02 0·00 0·04 FeOT 11·48 11·76 11·31 11·60 9·12 11·40 6·65 7·90 8·04 7·63 MnO 0·37 0·44 0·54 0·64 0·35 0·62 0·04 0·17 0·06 0·14 MgO 6·47 6·63 6·05 6·15 7·92 6·14 9·55 8·74 8·86 9·15 CaO 0·00 0·02 0·05 0·04 0·01 0·05 0·05 0·02 0·00 0·00 Na2O 0·20 0·17 0·17 0·15 0·27 0·25 0·11 0·08 0·12 0·10 K2O 0·01 0·01 0·00 0·01 0·01 0·00 0·02 0·01 0·02 0·01 Total 98·45 98·95 98·12 98·19 97·29 97·77 98·90 98·27 98·42 98·75 Structural formula on the basis of 18 oxygens Si 4·984 4·977 4·996 4·990 4·961 4·961 5·002 4·998 5·000 4·985 Ti 0·002 0·000 0·003 0·000 0·004 0·001 0·000 0·002 0·000 0·000 Al 3·973 3·952 3·997 3·973 3·981 4·008 3·974 3·967 3·956 3·970 Cr 0·002 0·003 0·000 0·000 0·002 0·004 0·002 0·002 0·000 0·003 Fe2+ 1·004 1·025 0·992 1·019 0·799 1·006 0·565 0·680 0·691 0·653 Mn 0·033 0·039 0·048 0·057 0·031 0·056 0·004 0·015 0·005 0·012 Mg 1·008 1·030 0·946 0·964 1·238 0·966 1·446 1·341 1·357 1·395 Ca 0·000 0·002 0·006 0·005 0·001 0·005 0·005 0·002 0·000 0·000 Na 0·040 0·034 0·035 0·031 0·054 0·052 0·021 0·016 0·023 0·021 K 0·001 0·002 0·000 0·001 0·001 0·000 0·003 0·001 0·002 0·002 Total 11·047 11·064 11·021 11·039 11·071 11·058 11·021 11·024 11·035 11·040 Mg-no. 50·1 50·1 48·8 48·6 60·8 49·0 71·9 66·4 66·3 68·1 Element analyses (oxide wt %); Mg-number = 100 × Mg/(Mg + Fetot); FeOT, total iron content as FeO. Textures: Qtz–Bt, cordierite in symplectites of cordierite, quartz and biotite; Crd, cordierite in coronas surrounding kyanite; Crd–Hc, cordierite in symplectites of cordierite and hercynite. between the two. No plagioclase is observed within the cordierite–hercynite symplectites and no cordierite is observed in the plagioclase–hercynite symplectites. Further retrogression of the symplectites led to pinitization of cordierite, albitization of plagioclase and replacement of hercynite by Al hydroxide in the plagioclase–hercynite symplectites. Hercynite in the cordierite–hercynite symplectites contains less ZnO (0·7–1·8 wt %; samples La-1b and G-O/ 2) than in the plagioclase–hercynite symplectites and the hercynite inclusions in garnet (Table 6). Its Mg-number varies between 8 and 18 as a result of retrograde exchange with cordierite. The Mg-number of cordierite in the symplectites varies accordingly between 41 and 61 and between 46 and 50 in the cordierite coronas (Table 7). The Na contents of cordierite scatter around 0·046 and 0·033 cations per formula unit (c.p.f.u.) for symplectites and coronas, respectively (sample G-O/2). Within the cordierite–hercynite symplectites surrounding relic kyanite, we observed euhedral grains of corundum (Fig. 2d). In the immediate vicinity of these grains, hercynite is absent. Corundum is interpreted to be a product of the reaction of the newly formed hercynite with the relic kyanite Ky + Hc = Crd + Crn. (13) Cordierite–quartz–biotite symplectites occur around garnet grains embedded in K-feldspar (Fig. 2f ). The symplectites were probably formed by the back-reaction of a granitoid melt [see reaction (9)] with garnet and Kfeldspar: 241 Grt + Kfs + L = Crd + Qtz + Bt. (14) JOURNAL OF PETROLOGY VOLUME 44 Between the K-feldspar and garnet, small irregular patches of intergrown plagioclase and quartz can be found (Fig. 2f ). These patches are interpreted as crystallized remnants of the melt. The cordierite grains in the symplectites have Mg-number of 49–52 and Na contents of 0·036 c.p.f.u. (Table 7; sample G-O/2). Biotite has low TiO2 contents of 0·5 wt % and Mg-number of 35–40 (sample G-O/2). THERMOBAROMETRIC METHODS To evaluate peak P–T conditions and retrograde paths, we used both conventional geothermobarometers and the computer program TWQ-2.02 (Berman, 1991). Details are given in the following section. Also, large feldspar grains containing exsolution lamellae were used for determination of peak temperatures by reintegration of the original ternary composition. This was done by measuring the chemical compositions of lamellae and hosts by microprobe spot analysis. The volumetric ratios of the two phases were determined using the graphical software NIH Image on back-scattered images. The molar ratios were calculated using the measured compositions and the molar volumes of pure anorthite, albite and orthoclase assuming linear changes in molar volumes within the ternary system. The determined ternary compositions were plotted in the isobaric ternary feldspar system for 0·5 GPa containing the solvus isotherms of Elkins & Grove (1990). The location of the plotted points with respect to the solvus isotherms yields the minimum temperature necessary to stabilize a certain ternary feldspar composition at a pressure of 0·5 GPa. The influence of pressure on the stability of ternary feldspars was investigated by Seck (1971) and Green & Usdansky (1986). The Clapeyron slope of the critical feldspar curves is positive (7–8 MPa/K for binary Ab–Or feldspars; Seck, 1971) and increases with increasing An contents (Green & Usdansky, 1986). The Schwarzwald ternary feldspars have An contents between 17 and 33. Thus, temperatures were corrected with a Clapeyron slope of 10 MPa/K. P–T CONDITIONS Granulites from the SSGC The investigated clinopyroxene–orthopyroxene granulites are strongly influenced by retrograde cooling. The larger pyroxene grains show exsolution lamellae and all grains are surrounded by retrograde phases. Therefore, two-pyroxene thermometry is not the recommended tool to determine peak-metamorphic conditions. The original ternary composition of exsolved feldspars indicates minimum temperatures of 1050 ± 50°C, assuming pressures NUMBER 2 FEBRUARY 2003 of 1·5 GPa (Figs 8 and 9a). Pressure constraints come from the sillimanite–garnet granulites and the garnet–orthopyroxene granulites (see below). Sillimanite–garnet granulites do not show any reaction textures, but the prismatic, euhedral shape of sillimanite indicates equilibration in the sillimanite stability field. Sillimanite-bearing granulites are widespread within unit 3 of the Southern Schwarzwald, whereas kyanite is only known from one locality (Metz, 1980). Assuming the same temperatures as for the clinopyroxene–orthopyroxene granulites (1050 ± 50°C), maximum pressures are limited to >1·5 GPa. Orthopyroxene–garnet granulites were used for calculating the peak conditions and the various stages of decompression. We considered the zonation patterns of the minerals involved in reaction (1). The barometers and thermometers used were: Al-in-opx thermometer (Aranovich & Berman, 1997); grt–opx Fe–Mg exchange thermometer (Harley, 1984); grt–opx–pl–qtz barometry (Perkins & Chipera, 1985); grt–rt–ilm–pl–qtz (GRIPS) barometry (Bohlen & Liotta, 1986). This set of geothermobarometers was also calculated with TWQ-2.02 (Berman, 1991). This includes an Mg and Fe calculation for both Al-in-opx thermometry and grt– opx–pl–qtz barometry, it calculates the Al and Ca exchange between plagioclase, orthopyroxene and garnet, and it calculates the stability of rutile + garnet in the rock. The TWQ calculation includes six reactions of which three are independent, plus GRIPS barometry yielding minimum pressures. As stated in the previous section, reaction (1) starts at high pressures with Mg-rich garnet, Mg–Al-rich orthopyroxene and Ab-rich plagioclase. It proceeds towards increasingly Fe-rich garnet and orthopyroxene (the latter with decreasing Al contents) and An-rich plagioclase as the coronas grow. Hence, garnet cores and the outermost plagioclase and orthopyroxene grains within the coronas should reflect the P–T conditions at the start of the reaction, whereas garnet rims and adjacent plagioclase and orthopyroxene should represent the end of the decompression reaction. However, the zoning patterns of orthopyroxene (Fig. 4) clearly reveal that whereas Al shows a symmetrical profile corresponding to that of garnet (Fig. 6), Fe and Mg patterns are asymmetric, indicating later modification and hence a decoupling of Mg-number and Al in orthopyroxene. Al-rich cores of orthopyroxene grains of coronas around garnet are not in equilibrium with the garnet cores, in having Mgnumber that are too low. Therefore, the Fe and Mg calculations of the Al-in-opx thermometer by TWQ yield temperatures that differ by >300°C. Calculating reaction 242 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES Fig. 8. Ternary composition diagram for feldspar with the solvus isotherms of Elkins & Grove (1990). The location of the plotted ternary feldspar compositions with respect to the solvus isotherms yields a minimum temperature for a pressure of 0·5 GPa. (For pressure correction and further information, see section on thermobarometric methods.) (1) with the most Ab-rich plagioclase from the matrix (An25), garnet cores (Alm56) and a hypothetical orthopyroxene with the highest Al2O3 content (2·3 wt %) and Mg-number (52) found in the sample, we obtain equilibrium and conditions of 1·5 GPa and 1015°C (Fig. 9a). Using corona orthopyroxene (Mg-number 45, 1·0–1·5 wt % Al2O3) and plagioclase (An30) and garnet near-rim compositions (Alm60) yields similar conditions of 1·3 GPa and 1010°C. The GRIPS barometer using matrix plagioclase and garnet core compositions yields minimum peak pressures of 1·5 GPa. The end of the decompression reaction was calculated using reaction (1) with garnet rims, corona plagioclase (An38) and orthopyroxene with low Mg-number (42) from the corona. The reaction curves determined intersect at 0·6 GPa and 700°C (Fig. 9a). Calculation of this late stage has a large uncertainty, as garnet is resorbed at rims, Mg–Fe diffusion probably ceased at higher temperatures, and the formation of ilmenite and biotite might have influenced Mg-number of garnet and orthopyroxene rims. In summary, no relics of prograde metamorphism are preserved in the granulites of the SSGC. The P–T path is characterized by peak metamorphic conditions of approximately 1·5 GPa and 1015°C, an initial decompression to 1·3 GPa and 1010°C and further decompression to 0·6 GPa and 700°C. Granulites from the CSGC In the metapelites, the widespread occurrence of the assemblage antiperthitic feldspar + garnet + rutile + kyanite allows peak conditions to be estimated. The intersection of the temperature field, determined by reintegration of feldspar compositions, with the kyanite– sillimanite boundary indicates peak metamorphic temperatures of 950–1010°C at minimum pressures of 1·4–1·8 GPa (Fig. 9b). Peak pressures were also determined by the stability of garnet cores together with kyanite, rutile and feldspar using the GRIPS (garnet–rutile–ilmenite–plagioclase–quartz) and GASP (garnet–aluminosilicate–quartz–plagioclase) barometers. GRIPS barometry yields minimum pressures only, because ilmenite was not stable at peak pressures. The 243 JOURNAL OF PETROLOGY VOLUME 44 NUMBER 2 FEBRUARY 2003 Fig. 9. Pressure–temperature plot showing calculated equilibria of peak and retrograde metamorphism as well as P–T paths for granulites from the Schwarzwald. (a) SSGC. The GRIPS barometer (Bohlen & Liotta, 1986) was applied to determine minimum pressures. Reactions are: a, Grs + 2 Alm + 2 Rt = 2 Ilm + An + Qtz; b, 3 An + 6 En = Grs + 2 Prp + 3 Qtz; c, Ok + 3 Fs = Alm; d, 3 Fs + Prp = 3 En + Alm. Reactions b–d were calculated for sample 2100 with TWQ-2.02 (Berman, 1991) for peak conditions using hypothetical garnet and orthopyroxene compositions (referred to with suffix 1) and for two stages of decompression using measured compositions (referred to with suffixes 2 and 3). For clarity, reactions c1 and d1 are not shown. Ok is orthocorundum, the hypothetical Al end-member of orthopyroxene. The dark shaded area marks the results of ternary feldspar thermometry for sample 14893. (b) CSGC. The GASP barometer (Koziol & Newton, 1988) was applied using the measured plagioclase compositions of sample La-1B. The GASP ternary was calculated with the reintegrated ternary feldspar composition. The GRIPS barometer (Bohlen & Liotta, 1986) was applied to determine minimum pressures. The shaded area marks the results of ternary feldspar thermometry for samples G-0/2 and La-1B. The partial P–T grid shows retrograde reactions in the system KFMASH after Waters (1991) and Bucher-Nurminen & Ohta (1993), with grey areas indicating divariant fields. Dashed lines show univariant reactions in the Mg-free system KFASH. Continuous lines indicate reactions in a bulk system of Mg-number 30. [Hc] and [Bt] are invariant points, characterized by the absence of hercynite and biotite, respectively. [For reactions (10), (13) and (14), see section on petrography and mineral compositions and section on P–T conditions.] Aluminosilicate stability fields were calculated with TWQ-2.02 (Berman, 1991). pressures determined from GASP for 1000°C are between 1·4 and 1·6 GPa using measured plagioclase compositions and between 1·6 and 1·8 GPa using the reintegrated ternary feldspar compositions (Fig. 9b). The large number of symplectites and reaction textures described above allows a very detailed determination of the decompression and cooling history of the granulites from the CSGC. It is not possible to calculate equilibria for all reactions in one petrogenetic grid using a single whole-rock composition, because most of the coronas and symplectites form chemical sub-systems. In these sub-systems, the effective chemical compositions can be distinctly different from the average whole-rock composition. On the one hand, this complicates the determination of the P–T grid. On the other hand, the number of reactions that can be used for determining the P–T history is increased. In the case of the pelitic granulites, products of retrograde reactions are located 244 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES in coronas around garnet and kyanite grains. These assemblages are commonly silica undersaturated and include hercynite or corundum, isolated from the quartzbearing matrix by cordierite coronas. Equilibration temperatures of cordierite in the presence of plagioclase can be determined on the basis of its Na content (Mirwald, 1986). This thermometer is independent of pressure and has an uncertainty of ±35°C (Mirwald, 1986). In sample G-O/2, the composition of cordierite could be measured in all three assemblages. In the cordierite coronas, Na contents are 0·033 c.p.f.u., corresponding to a temperature of 780°C. In the cordierite–hercynite symplectites, cordierite has Na contents of 0·046 c.p.f.u., indicating 750°C. In the cordierite– quartz–biotite symplectites, cordierite contains 0·036 Na p.f.u. (770°C). The retrograde P–T path was further determined by combining the P–T grids of Waters (1991) and BucherNurminen & Ohta (1993) for the system KFMASH (Fig. 9b). Garnet cores in all samples have a maximum Mgnumber of 30. As garnet was the only Fe- and Mgbearing phase that was stable at peak metamorphic conditions (apart from minor biotite included in garnet and kyanite) the Mg-number of the whole rock can be inferred from the Mg-number of garnet cores. All reactions in the FAS system are univariant and become divariant in KFMASH. Hence, the invariant point [Bt] of Waters (1991) moves to higher pressures and temperatures with the addition of Mg, stabilizing cordierite to higher pressures. The incorporation of H2O into cordierite also enlarges its stability field (Aines & Rossman, 1984). Hercynite in the symplectites contains 2% gahnite and 1% magnetite component, which shifts the invariant point [Bt] to lower temperatures by >20°C (Nichols et al., 1992). The low-temperature part of the grid shown, based on Bucher-Nurminen & Ohta (1993), displays the reactions involving biotite. These are dependent on parameters such as H2O activity and melt composition, which cannot be determined precisely. However, the invariant point [Hc] is located at the intersection of reactions (10) and (14). Thus, the pressure of [Hc] can be determined via the cordierite-forming decompression reaction (10) (e.g. Nichols et al., 1992; calculation by TWQ ), and its temperature by the Na-in-Crd thermometer (Mirwald, 1986), using the cordierite formed by reaction (14). The shaded areas in the grid show divariant reaction fields. Within these fields, the coexistence of different Fe–Mg-bearing phases is possible, leading to a couple of further reactions. Kriegsman & Hensen (1998) described melt-producing reactions, e.g. (9), that can occur during heating or decompression. In the Schwarzwald granulites, all decompression reactions that produced cordierite must have occurred in the temperature range given by the two invariant points [Bt] and [Hc], which are connected by reaction (10). This indicates temperatures between 700 and 900°C, which can be further defined by Na-in-Crd thermometry. Another limit for the P–T path is given by the occurrence of corundum formed by reaction (13). This reaction is restricted to low pressures at high temperatures. The back-reaction of melt in reaction (14) produced biotite, quartz and cordierite. This is evidence for pressures below the invariant point [Hc] at >0·4 GPa. In summary, the granulites from the CSGC probably traversed the kyanite stability field during the prograde part of the P–T path as garnet and kyanite form inclusions within each other (see section on petrography and mineral chemistry). Breakdown reactions of staurolite and biotite probably formed kyanite, garnet and rutile. The peak metamorphic assemblages require minimum temperatures between 950 and 1010°C at minimum pressures of 1·4–1·8 GPa. After their equilibration at HT–HP conditions, the rocks were strongly decompressed to pressures below 0·4 GPa associated with cooling to 750°C. During this exhumation of >30 km, as well as during prograde metamorphism, partial granitoid melts formed by the breakdown of biotite [reactions (3) and (9)]. At least some of the melt was not segregated from the rocks but remained in situ, crystallizing to fine-grained blebs of K-feldspar, plagioclase, quartz and biotite. Part of the melt reacted with garnet, replacing it with symplectites of cordierite, quartz and biotite. GEOLOGICAL IMPLICATIONS For unit 3, the ages of granulite-facies metamorphism and of pre-metamorphic sedimentation–magmatism are now well defined. SHRIMP zircon dating of CSGC metasedimentary granulites (Kalt et al., 2000a, and references therein; B. Kober et al., in preparation) yielded an age for the granulite-facies metamorphism of 340–335 Ma. In the SSGC, ages of >342 Ma were obtained on rocks from unit 3 (Pb–Pb evaporation method and conventional U–Pb dating on zircons, Sm–Nd model ages on whole-rock samples; Hegner et al., 2001). Concordant U–Pb ages of 332–329 Ma of monazites from unit 3 rocks were inferred to date the peak of HT–LP metamorphism (Kalt et al., 1994a). These results indicate a fairly rapid succession of the HT–HP and the HT–LP stage within a single orogenic event and metamorphic cycle, and argue against an early or pre-Variscan granulite-facies event in the Schwarzwald at >400 Ma, as postulated by Pin & Vielzeuf (1983, 1988) and Vielzeuf & Pin (1989). The pressure calculations for granulites from the SSGC and the CSGC along with geochronological data indicate that these rocks represent Carboniferous lower crust. Petrographic observations show that probably all of the gneisses and migmatites of unit 3 in the Schwarzwald represent former granulites that are strongly retrogressed. Therefore, 245 JOURNAL OF PETROLOGY VOLUME 44 the entire unit 3 represents Variscan (Carboniferous) lower crust. The temperature calculations for the granulites reveal that this lower-crustal segment was hotter than would correspond to a normal geothermal gradient. The geodynamic processes potentially responsible for these high temperatures are discussed below. According to the definition of Harley (1998), the granulites of the Schwarzwald are just at the upper pressure limit for UHT metamorphism (900–1100°C and 0·7–1·3 GPa). Several workers pointed out the lithological differences between the metamorphic basement of the CSGC and the SSCG (e.g. Stenger et al., 1989; Wimmenauer et al., 1989) and others even interpreted both as representing two terranes or microcontinents (Loschke et al., 1998; see ¨ section on geological setting). This raises the question of whether the granulites of the CSGC and the SSGC represent the lower crust of two distinct terranes. Criteria that can be used to assess this problem are granulite lithology, age of metamorphism, peak metamorphic conditions and association with other rocks. In the CSGC, pelitic and psammitic compositions dominate over basic, intermediate and felsic (igneous) compositions in well-preserved granulites. However, taking into account the gneisses that represent retrogressed granulites, igneous felsic and intermediate compositions are more abundant, as is the case in the SSGC. The ages of granulite-facies metamorphism in CSGC and SSGC are indistinguishable (see above). The metamorphic conditions calculated for granulites from both areas are very similar (see section on P–T conditions). Although the temperatures are virtually identical, granulites of the CSGC equilibrated in the kyanite stability field whereas granulites of the SSGC were formed in the sillimanite stability field. However, the calculated minimum and maximum pressures, respectively, coincide at 1·5 GPa. Ultramafic rocks occur as bodies of various sizes in unit 3 of the CSGC and the SSGC. In the CSGC, garnet–spinel peridotites and spinel peridotites can be found that record various equilibration conditions, but in general lower temperatures and higher pressures than their granulite hosts (Kalt & Altherr 1996). Also, basic granulites from the CSGC show evidence of an eclogitefacies stage before granulite-facies metamorphism, at lower temperatures and higher pressures (Hanel et al., 1993). Within unit 3 of the SSGC, eclogite-facies relicts are absent. Lens-shaped bodies of garnet–spinel and spinel peridotites can be found that equilibrated at approximately the same P–T conditions as the granulites (R. Altherr et al., unpublished data, 1996). In conclusion, the granulites of the CSGC may represent a slightly deeper level of Carboniferous lower crust than the granulites of the SSGC, but there are no convincing arguments to assign them to two separate terranes. Three tectonometamorphic units have been defined in the Schwarzwald. Apart from a contemporaneous NUMBER 2 FEBRUARY 2003 HT–LP overprint, the three units are thought to differ in their early metamorphic evolution (Fig. 1). The peak P–T conditions and the P–T path obtained here for the granulites of unit 3 confirm this hypothesis. The obtained values of approximately 1000°C and 1·5 GPa are distinctly different from those of unit 1 (730–780°C and 0·4–0·45 GPa) and unit 2 (550–650°C and 0·5 GPa) and there is no metamorphic gradient within unit 3, with decreasing P–T conditions towards the contacts with units 1 and 2. Therefore, the three units probably do not represent a coherent crustal section. This conclusion is further supported by the age spectra of detrital zircon. The age of the youngest detrital zircon population in unit 3 is significantly younger (Devonian) than in the other two units (Ordovician and Neoproterozoic), suggesting a different palaeogeographical position in the Palaeozoic (Kober et al., in preparation). The P–T paths obtained in this study on granulites of the Schwarzwald are characterized by initial decompression, followed by cooling (Figs 9 and 10). In the CSGC, decompression is nearly isothermal and cooling is more or less isobaric. These trends are less pronounced in the SSGC. The decompressional parts of the paths are fairly well constrained by the petrological and geochronological data. The peak conditions of approximately 1000°C and 1·4–1·8 GPa were dated at 342–335 Ma (see above). The HT–LP stage common to all three units (730–780°C and 0·4–0·45 GPa) was dated at 332–329 Ma by the U–Pb method on monazite (see above). However, the applied dating methods are only mineral- and/or temperaturesensitive and cannot date pressure or record decompression. Temperatures of 730–780°C, prevailing during monazite growth, may have been attained at different pressures by granulites and other metamorphic rocks of the Schwarzwald, as suggested by the obtained spread in pressures (Fig. 10). Therefore, the time span of >3–10 Myr may include a decompression of 1·1 GPa but perhaps as little as 0·4 GPa (Fig. 10). The resulting vertical exhumation rates of 0·9–5 mm/a are on average fairly rapid and comparable with those for other granulite terranes (e.g. Harley & Hensen, 1990; Willner et al., 1997). Cooling after the HT–LP stage varied regionally in the Schwarzwald as documented, for example, by the large spread in 40Ar/39Ar ages of metamorphic hornblende (320–350 Ma; Lippolt et al., 1994). COMPARISON WITH GRANULITES FROM THE BOHEMIAN MASSIF AND THE VOSGES 246 In the Bohemian Massif (BM), several areas with granulites have been recognized. As in the Schwarzwald, the granulites of the BM are partly retrogressed and occur as MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES Fig. 10. Pressure–temperature diagram showing the P–T paths of granulites from the CSGC and the SSGC. Also shown are equilibration conditions of gneisses and migmatites from unit 1 (HT–LP stage; Kalt et al., 2000a) and the P–T path of spinel peridotites associated with granulites of unit 3 of the SSGC (R. Altherr, unpublished data, 1997). distinct tectonic units. Most of the granulites of the southern BM (Austria, Czech Republic) form klippen (Gfohl nappe; ¨ Fuchs, 1971) on top of HT–LP gneisses and migmatites. Felsic and intermediate lithologies dominate over mafic and pelitic compositions. The Austrian and the Czech granulites are associated with eclogites and peridotites. The granulites in the Austrian part of the southern BM experienced (U)HT conditions of 1·6 GPa and 1000°C (Carswell & O’Brien, 1993; Cooke, 2000; Cooke & O’Brien, 2001). The granulites from the Czech part of the Gfohl ¨ nappe record very similar P–T conditions at 340 ± 3 Ma (e.g. Kroner et al., 2000, and references therein). ¨ In the northern BM, granulites form large coherent areas in the Erzgebirge and in the Granulitgebirge (Czech Republic and Germany). The Erzgebirge is interpreted as a mega-antiform (e.g. Willner et al., 1997). Within this antiform, granulites occur in the so-called Gneiss– Eclogite Unit together with eclogites and peridotites. Intermediate and felsic granulites dominate over pelitic lithologies. The granulites equilibrated at 830°C and pressures between 1·2 and 2·4 GPa (Willner et al., 1997) at 340–341 Ma (e.g. Kotkova et al., 1996; Kroner & ´ ¨ Willner, 1998). Tectonic contacts separate the granulites recording different pressures. The Granulitgebirge is interpreted as a dome (e.g. Rotzler & Romer, 2001) with ¨ the granulites forming the core of the structure. Felsic rocks are by far the dominant lithology. The granulites were metamorphosed at the same time as those of the Erzgebirge (340–342 Ma, von Quadt, 1993; Kroner et ¨ al., 1998; Romer & Rotzler, 2001) with similar peak ¨ conditions (967°C and 2·2 GPa; Rotzler & Romer, 2001). ¨ In the northeastern BM, granulites are known mainly from two localities in the Polish Sudetes: the Gory Sowie ´ block and the Snieznik area. The Gory Sowie block is ´ bounded on all sides by faults and is anisofacial with its country rocks. Felsic and intermediate granulites are most abundant and associated with garnet peridotites and eclogites. The granulites experienced peak metamorphic conditions of 1·5–2·0 GPa and 900–1000°C (Kryza et al., 1996; O’Brien et al., 1997). SHRIMP dating on zircon shows that these equilibration conditions prevailed at 402 ± 1 Ma (O’Brien et al., 1997). Granulites from the nearby Snieznik area partly record UHP equilibration conditions (800–1000°C and 2·1–2·8 GPa) and are younger than 360 Ma (Klemd & Brocker, 1999). ¨ In the Vosges, granulites are associated with peridotites occuring within the tectonically uppermost unit or nappe (Rey et al., 1989) and equilibrated at 335–337 Ma (Schaltegger et al., 1999). They are of felsic, intermediate, mafic and pelitic compositions. Equilibration conditions of these granulites have not yet been determined. However, phase assemblages in metapelitic granulites point to high pressures and temperatures. In summary, the granulite occurrences in the BM and 247 JOURNAL OF PETROLOGY VOLUME 44 the Schwarzwald (and the Vosges) seem to be characterized by the following features: (1) they form separate tectonic units; (2) their lithology is heterogeneous; (3) equilibration temperatures are fairly uniform at >1000°C (except for the Erzgebirge); (4) pressures vary from 1·2 to 2·8 GPa, which seems to reflect the composite ´ nature of some granulite units; (5) except for the Gory Sowie block, all dated granulite massifs equilibrated at 335–340 Ma. These features were already summarized for granulites of the BM by O’Brien (2001). GEODYNAMIC IMPLICATIONS Heat sources The high equilibration temperatures attained at the same time within most of the granulite occurrences in the Schwarzwald, the BM, and presumably also the Vosges suggest that the now isolated granulite units may once have formed a coherent segment of Variscan lower crust that was affected by a large-scale thermal event in the Carboniferous. The lowest obtained pressures of 1·2 GPa indicate a minimum crustal thickness of >40 km. Pressures in excess of 2 GPa imply that at least some parts of the Variscan continental crust were even considerably thicker (>70 km). In any case, the range in pressures recorded by most of the granulites (1·2–1·6 GPa) indicates that different levels of the lower crust attained very similar temperatures during this thermal event. An explanation for the extremely high temperatures must be found within the Variscan context. Internal heating by increased radiogenic heat production as a result of crustal stacking is probably not sufficient to produce temperatures of >1000°C in the lower crust. Coupled thermal–mechanical models of convergent orogens show that Moho temperatures of 700–800°C can be attained only if heat-producing material is subducted or if a thick orogenic wedge is built up, and time scales of 10–30 Myr would be necessary to reach these temperatures (e.g. Jamieson et al., 1998, 2002). The preservation of growth zoning in Variscan granulites that experienced >1000°C (e.g. Cooke et al., 2000; this study) strongly argues for short-lived heating. Moreover, the uniform temperature distribution throughout the lower Variscan crust is more in favour of an external heat source and advective rather than conductive heating. Several models of heating by external sources have been discussed in the context of Variscan orogeny (e.g. Henk et al., 2001). For the HP–HT granulites of the BM, penetration and detachment of a subducting slab by ascending, hot asthenosphere (slab breakoff; e.g. Davies & Blanckenburg, 1995) has been inferred as a major mechanism (e.g. Bruckner & Medaris, 2000; O’Brien, ¨ 2001). However, this process has not been used so much as a possible heat source but has been invoked mainly NUMBER 2 FEBRUARY 2003 to explain the presence of peridotite bodies with different origins (asthenospheric, continental-lithospheric and metasomatic) and varying equilibration conditions in the HP–HT granulite nappes (Bruckner & Medaris, 2000; ¨ O’Brien, 2001). In the invoked slab-breakoff setting, the granulites are viewed as a part of subducting continental crust (O’Brien, 2001), which is in line with the broadly granitic upper-crustal composition of the dominant felsic granulites and with the high pressures some of them record. However, subduction usually implies a relatively high P/T gradient in the downgoing plate, even in ‘hot’ subduction settings (e.g. high initial temperature of the subducted slab, low convergence rates, low subduction angle; Peacock 1996; Stein & Stein, 1996). The only heat source in the slab is internal radioactive decay. It is questionable if conditions of >1000°C and 1·2–1·6 GPa, recorded by most Variscan granulites, can be attained in a subducting plate. A more evident mechanism for heating the lower crust to extremely high temperatures seems to be detachment of thickened mantle lithosphere and upwelling of hot asthenosphere (e.g. Bird, 1979; England & Houseman, 1989). The heat source for the lower crust could be either the asthenosphere or large volumes of basaltic melt underplated at the crust–mantle boundary. This process usually leads to extension–gravitational collapse of the overlying crust and exhumation of high-grade rocks. There is convincing structural evidence for crustal extension and collapse in the Variscan belt (e.g. Menard ´ & Molnar, 1988; Rey et al., 1989; Costa & Rey, 1995), but these structures are thought to postdate high-grade metamorphism. In the Schwarzwald, for example, granulite-facies metamorphism is 5–10 Myr earlier than the ubiquitous HT–LP metamorphism. The latter seems to be older than or contemporaneous with the juxtaposition of CSGC, BLZ and SSGC in a compressional setting (Echtler & Chauvet, 1991–1992). The onset of extension in the Schwarzwald is thought to be marked by large volumes of granites at 330–325 Ma that cut through the nappe structures and shear zones. Therefore, lithospheric detachment and asthenospheric upwelling are probably processes important for the exhumation of the granulites, but not their initial heat source. Our preferred model is that the granulites of the Schwarzwald, and probably also other Variscan granulites, equilibrated in the lower part of a thickened crust of a hanging continental plate in a convergent setting. The presumably clockwise P–T path of the metapelites of the CSGC suggests that at least a part of the granulites was originally upper crust, buried in a compressional setting. The initial heat source of these granulites could have been large volumes of basaltic magma, intruded below and in the lower crust of the hanging plate in the context of subduction-related melting. The presence of active subduction zones at 360–333 Ma is evident from 248 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES HP–LT eclogites that equilibrated at this time (e.g. Kalt et al., 1994b; Schmadicke et al., 1995). Subduction-related ¨ melting at this time can be inferred from magmatic rocks. Although of minor volume compared with the postcollisional peraluminous granites, there are calc-alkaline granitoids (e.g. Altherr et al., 1999a, 2000) and lamprophyres (Hegner et al., 1998) that indicate an enriched mantle. The proposed setting would also be in line with the presence of some HP–LT eclogites and various peridotites in the granulite nappes. It is often argued that large volumes of basic magma are required for such a process and that only few Carboniferous basic rocks (magmatic or metamorphic) are found in the Variscan belt, at the surface or at depth (e.g. O’Brien, 2001). In Alpine areas representing fossil crust–mantle transitions, this percentage is usually much larger (e.g. Voshage et al., 1987; Hermann et al., 1997). However, experimental and modelling work has shown that heat transfer into the crust and melting of the crust by basaltic intrusions is a rapid process (e.g. Huppert & Sparks, 1988; Barboza & Bergantz, 1996) and that repetitive intrusion of smaller volumes on short time scales can be even more efficient in heating the crust than one large single body (Petford & Gallagher, 2001; Annen & Sparks, in preparation). The absence of Carboniferous basic rocks at the surface of the Variscan belt may be explained by the formation of crustal melts, the mixture of mantle and crustal melts, and magma differentiation during ascent through a thickened crust (e.g. Altherr et al., 1999b). The absence of a thick mafic lower crust today may be simply due to later modification of the entire European lithosphere during Mesozoic rifting, the Alpine orogeny and Cenozoic mantle plume activity. Moreover, the Carboniferous crust–mantle boundary is not exposed anywhere in the Variscan belt. The lower-crustal granulite nappes found today were obviously detached from the mantle and tectonically dismembered. Exhumation conditions, metagreywackes yield similar melt fractions with residues rich in garnet, orthopyroxene and plagioclase (e.g. Montel & Vielzeuf, 1997; Stevens et al., 1997). Partial melting of amphibolites and amphibole-bearing tonalites at conditions of 1000°C and 1–1·4 GPa yields lower melt fractions of 10–40% and restites with clinopyroxene, orthopyroxene, plagioclase and garnet (e.g. Rutter & Wyllie, 1988; Skjerlie & Johnston, 1993; Rapp & Watson, 1995). In all experiments, the partial melts are silicic and leucocratic. The granulite nappes in the Variscan massifs usually consist of large volumes of leucocratic rocks, most of them retrogressed under amphibolite-facies conditions, accompanied by minor intermediate, basic and pelitic granulites. These lithologies could in principle represent large volumes of partial melts with some of their entrained residues. Indeed, Variscan granulites of felsic composition have been considered as representing partial melts formed from continental crust at deep crustal or even mantle pressures (e.g. Fiala et al., 1987; O’Brien et al., 1997; ´ Kotkova & Harley, 1999). Clearly, this model needs to be verified from a geochemical viewpoint and by further experiments. Nevertheless, it would offer a viable explanation for the isothermal and commonly rapid initial decompression of the Variscan granulites. Considerable percentages of felsic to intermediate melt would lower bulk viscosity and density significantly and hence facilitate the ascent or exhumation of heterogeneous lower-crustal material into the middle and upper crust. Ascent, exhumation or decompression of the buoyant lower crust requires removal of upper-crustal material. There is structural evidence for crustal extension and collapse in the Variscan belt (see above). It is very likely that crustal extension was induced by detachment or delamination of the mantle lithosphere and upwelling of hot asthenosphere, because enormous volumes of granitic magma were produced in the Variscan belt starting from 330–325 Ma (e.g. Henk et al., 2001, and references therein). In the Schwarzwald, the chemical and isotopic compositions of many late- or post-collisional S-type granites can only be explained if a considerable mantle component is invoked (Altherr et al., 1999b). Regardless of the heat source, heating of the Variscan lower crust led to partial melting, evident from the microtextures of the pelitic granulites from the Schwarzwald. Although there is evidence for in situ melt crystallization in these rocks, parts of the melt may also have escaped. In fact, dehydration experiments with different rock compositions reveal that considerable melt fractions may be formed at the P–T conditions of interest (Patino˜ Douce & McCarthy, 1998). Experiments with metapelites at conditions of 1000°C and 1–1·2 GPa reveal melt percentages of 50–60 vol. % and residues rich in garnet, aluminosilicate and spinel (e.g. Vielzeuf & Holloway, 1988; Patino-Douce & Johnston, 1991). At the same P–T ˜ A petrographic and petrological study, including the determination of mineral compositions and P–T calculations, was carried out on granulites from the Schwarzwald. Samples were taken from the CSGC and the SSGC, two high-grade crystalline blocks separated by a zone of non-metamorphic to medium-grade metamorphic Palaeozoic rocks. The following results were obtained: SUMMARY AND CONCLUSIONS 249 JOURNAL OF PETROLOGY VOLUME 44 (1) no relics of prograde metamorphism are preserved in the granulites of the SSGC. The P–T path is characterized by peak metamorphic conditions of approximately 1·5 GPa and 1015°C within the sillimanite stability field, followed by an initial decompression to 1·3 GPa and 1010°C and further decompression and cooling to 0·6 GPa and 700°C. (2) Granulites from the CSGC probably traversed the kyanite stability field during their prograde part of the P–T path. The peak metamorphic assemblages require minimum temperatures between 950 and 1010°C at pressures of 1·4–1·8 GPa. Their retrograde P–T path is characterized by initial isothermal decompression, including partial melting, followed by isobaric cooling. The major conclusions from these results are as follows: (1) an HT event affected the lower crust of the Schwarzwald at 340–335 Ma. (2) There are no convincing arguments for distinguishing the granulites of the SSGC and the CSGC. However, the latter may represent a slightly deeper level of Carboniferous lower crust than the former. (3) Given the similarities in lithology, equilibration conditions and age, the granulites of the Schwarzwald, most of the granulites of the Bohemian Massif and probably also those of the Vosges experienced the same, large-scale thermal event in Early Carboniferous time. (4) The heat source for this Early Carboniferous thermal event may have been multiple basic, mantle-derived intrusions into and below the lower crust in a subduction setting. (5) The initial isothermal decompression and rapid exhumation (0·9–5 mm/a) of the granulites may be due to the presence of considerable amounts of partial melts and to orogenic extension. ACKNOWLEDGEMENTS The authors would like to thank Hans-Peter Meyer for microprobe maintainance and for providing mineral formula calculation programs. We are grateful to Wolfhard Wimmenauer for guiding us in the field. Thanks go to Thomas Ludwig for help with image processing. H.M. is also grateful to Dominik Hezel and Stefan Prowatke for fruitful discussion. Rainer Altherr is thanked for critical comments on the manuscript. Michael Williams, Giles Droop, Andreas Moller and Patrick O’Brien ¨ are thanked for constructive reviews. Financial support by the Deutsche Forschungsgemeinschaft (Ka 1023/4-1) is gratefully acknowledged. REFERENCES Aines, R. D. & Rossman, R. D. (1984). The high temperature behaviour of water and carbon dioxide in cordierite and beryl. American Mineralogist 69, 319–327. 250 NUMBER 2 FEBRUARY 2003 Altherr, R., Henes-Klaiber, U., Hegner, E., Satir, M. & Langer, C. (1999a). Plutonism in the Variscan Odenwald (Germany): from subduction to collision. International Journal of Earth Sciences 88, 422– 443. Altherr, R., Henjes-Kunst, F., Langer, C. & Otto, J. (1999b). Interaction between crustal-derived felsic and mantle-derived mafic magmas in the Oberkirch Pluton (European Variscides, Schwarzwald, Germany). Contributions to Mineralogy and Petrology 137, 304–322. Altherr, R., Holl, A., Hegner, E., Langer, C. & Kreuzer, H. (2000). High-potassium, calc-alkaline I-type plutonism in the European Variscides: northern Vosges (France) and northern Schwarzwald (Germany). Lithos 50, 51–73. Aranovich, L. Y. & Berman, R. G. (1997). A new garnet–orthopyroxene thermometer based on reversed Al2O3 solubility in FeO–Al2O3–SiO2 orthopyroxene. American Mineralogist 82, 345–353. Barboza, S. A. & Bergantz, G.W. (1996). Dynamic model of dehydration melting motivated by a natural analogue: applications to the Ivrea– Verbano zone, northern Italy. Transactions of the Royal Society of Edinburgh 87, 23–31. Berman, R. G. (1991). Thermobarometry using multiequilibrium calculations: a new technique with petrologic applications. Canadian Mineralogist 29, 833–855. Bird, P. (1979). Continental delamination and the Colorado Plateau. Journal of Geophysical Research 84, 7561–7571. Bohlen, S. R. (1987). Pressure–temperature–time paths and a tectonic model for the evolution of granulites. Journal of Geology 95, 617–632. Bohlen, S. R. & Liotta, J. J. (1986). A barometer for garnet amphibolites and garnet granulites. Journal of Petrology 5, 1025–1034. Brueckner, H. K. & Medaris, L. G. (2000). A general model for the intrusion and evolution of ‘mantle’ garnet peridotites in high-pressure and ultra-high-pressure metamorphic terranes. Journal of Metamorphic Geology 18, 123–133. Bucher-Nurminen, K. & Ohta, Y. (1993). Granulites and garnet– cordierite gneisses from Maud Land, Antarctica. Journal of Metamorphic Geology 11, 691–703. Carswell, D. A. & O’Brien, P. J. (1993). Thermobarometry and geotectonic significance of high-pressure granulites: examples from the Moldanubian Zone of Bohemian Massif in Lower Austria. Journal of Petrology 34, 427–459. Clarke, G. L., Klepeis, K. A. & Daczko, N. R. (2000). Cretaceous high-P granulites at Milford Sound, New Zealand: metamorphic history and emplacement in a convergent margin setting. Journal of Metamorphic Geology 18, 359–374. Cooke, R. A. (2000). High-pressure–temperature metamorphism in the St. Leonhard Granulite Massif, Austria: evidence from intermediate pyroxene-bearing granulites. International Journal of Earth Science 89, 631–651. Cooke, R. A. & O’Brien, P. J. (2001). Resolving the relationship between high P–T rocks and gneisses in collisional terranes: an example from the Gfohl gneiss–granulite association in the Mol¨ danubian Zone, Austria. Lithos 58, 33–54. Cooke, R. A., O’Brien, P. J. & Carswell, D. A. (2000). Garnet zoning and the identification of equilibrium mineral compositions in highpressure–temperature granulites from the Moldanubian Zone, Austria. Journal of Metamorphic Geology 18, 551–569. Costa, S. & Rey, P. (1995). Lower crustal rejuvenation and crustal growth during post-thickening collapse: insights from a crustal cross section through a Variscan metamorphic core complex. Geology 23, 905–908. Dasgupta, S., Sengupta, P., Ehl, J. & Raith, M. (1995). Reaction textures in a suite of spinel granulites from the Eastern Ghats Belt: evidence for polymetamorphism, a partial petrogenetic grid in the MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES system KFMASH and the roles of ZnO and Fe2O3. Journal of Petrology 36, 435–464. Davies, J. H. & von Blanckenburg, F. (1995). Slab breakoff: a model of lithospheric detachment and its test in the magmatism and deformation of collisional orogens. Earth and Planetary Science Letters 129, 85–102. Echtler, H. P. & Altherr, R. (1993). Variscan crustal evolution in the Vosges Mountains and in the Schwarzwald: guide to the excursion of the Swiss Geological Society and the Swiss Society of Mineralogy and Petrology (3–5 October, 1992). Schweizerische Mineralogische und Petrographische Mitteilungen 73, 113–128. Echtler, H. P. & Chauvet, A. (1991–1992). Carboniferous convergence and subsequent crustal extension in the southern Schwarzwald (SW Germany). Geodynamica Acta 5, 37–49. Eisbacher, G. H., Luschen, E. & Wickert, F. (1989). Crustal-scale ¨ thrusting and extension in the Hercynian Schwarzwald and Vosges, central Europe. Tectonics 8, 1–21. Elkins, L. T. & Grove, T. L. (1990). Ternary feldspar experiments and thermodynamic models. American Mineralogist 75, 544–559. England, P. C. & Houseman, G. A. (1989). Extension during continental convergence, with application to the Tibetian plateau. Journal of Geophysical Research 94, 17561–17579. Fiala, J., Matejovska, O. & Vankova, V. B. (1987). Moldanubian granulites: source material and petrogenetic considerations. Neues Jahrbuch fur Mineralogie, Abhandlungen 157, 133–165. ¨ Franke, W. (1989). Tectonostratigraphic units in the Variscan Belt of Central Europe. Geological Society of America, Special Papers 230, 67–90. ¨ Fuchs, G. (1971). Zur Tektonik des ostlichen Waldviertels (Niederosterreich). Verhandlungen der Geologischen Bundesanstalt Wien 3, 424– ¨ 440. Gardien, V., Thompson, A. B. & Ulmer, P. (2000). Melting of biotite + plagioclase + quartz gneisses: the role of H2O in the stability of amphibole. Journal of Petrology 41, 651–666. Green, D. H. & Ringwood, A. E. (1967). An experimental investigation of the gabbro to eclogite transformation and its petrological applications. Geochimica et Cosmochimica Acta 31, 767–833. Green, N. L. & Usdansky, S. I. (1986). Ternary-feldspar mixing relations and thermobarometry. American Mineralogist 71, 1100–1108. Hanel, M. & Wimmenauer, W. (1990). Petrographische Indizien fur ¨ einen Deckenbau im Kristallin des Schwarzwaldes. European Journal of Mineralogy 2(Beiheft 1), 89. Hanel, M., Lippolt, H. J., Kober, B. & Wimmenauer, W. (1993). Lower Carboniferous granulites in the Schwarzwald basement near Hohengeroldseck (SW-Germany). Naturwissenschaften 80, 25–28. Hanel, M., Montenari, M. & Kalt, A. (1999). Determining sedimentation ages of high-grade metamorphic gneisses by their palynological record: a case study in the northern Schwarzwald (Variscan Belt, Germany). International Journal of Earth Sciences 88, 49–59. Hann, H. P. & Sawatzki, G. (2000). Neue Daten zur Tektonik des Sudschwarzwaldes. Jahresbericht und Mitteilungen der Oberrheinischen Geo¨ logischen Vereinigung, Neue Folge 82, 363–376. Harley, S. L. (1984). An experimental study of the partitioning of Fe and Mg between garnet and orthopyroxene. Contributions to Mineralogy and Petrology 86, 359–373. Harley, S. L. (1989). The origins of granulites: a metamorphic perspective. Geological Magazine 126(3), 215–247. Harley, S. L. (1998). On the occurrence and characterization of ultrahigh-temperature crustal metamorphism. In: Treloar, P. J. & O’Brien, P. J. (eds) What Drives Metamorphism and Metamorphic Reactions? Geological Society, London, Special Publications 138, 81–107. Harley, S. L. & Hensen, B. J. (1990). Archean and Proterozoic high grade terranes of East Antarctica (40–80°E): a case study of diversity in granulite facies metamorphism. In: Ashworth, J. R. & Brown, M. (eds) High-Temperature Metamorphism and Crustal Anatexis. London: Unwin Hyman, 320–370. Hegner, E., Kolbl-Ebert, M. & Loeschke, J. (1998). Post-collisional ¨ Variscan lamprophyres (Black Forest, Germany): 40Ar/39Ar phlogopite dating, Nd, Sr, Pb and trace element characteristics. Lithos 45, 395–411. Hegner, E., Chen, F. & Hann, H. P. (2001). Chronology of basin closure and thrusting in the internal zone of the Variscan belt in the Schwarzwald, Germany: evidence from zircon ages, trace element geochemistry, and Nd isotopic data. Tectonophysics 332, 169–184. Henk A., von Blanckenburg, F., Finger, F., Schaltegger, U. & Zulauf, G. (2001). Syn-convergent high-temperature metamorphism and magmatism in the Variscides. In: Franke, W., Haak, V., Oncken, O. & Tanner, D. (eds) Orogenic Processes: Quantification and Modelling in the Variscan Belt. Geological Society, London, Special Publications 179, 378–399. Hermann, J., Muntener, O., Trommsdorff, V. & Hansmann, W. (1997). ¨ Fossil crust-to-mantle transition, Val Malenco (Italian Alps). Journal of Geophysical Research 102, 20123–20132. Hokada, T. (2001). Feldspar thermometry in ultrahigh-temperature metamorphic rocks: evidence of crustal metamorphism attaining >1100°C in the Archean Napier Complex, East Antarctica. American Mineralogist 86, 932–938. Huppert, H. E. & Sparks, R. S. J. (1988). The generation of granitic magmas by intrusion of basalt into continental crust. Journal of Petrology 29, 599–624. Jamieson, R. A., Beaumont, C., Fullsack, P. & Lee, B. (1998). Barrovian metamorphism: Where’s the heat? In: Treloar, P. J. & O’Brien, P. J. (eds) What Drives Metamorphism and Metamorphic Reactions? Geological Society, London, Special Publications 138, 23–45. Jamieson, R. A., Beaumont, C., Nguyen, H. M. & Lee, B. (2002). Interaction of metamorphism, deformation and exhumation in large convergent orogens. Journal of Metamorphic Geology 20, 9–24. Kalt, A. & Altherr, R. (1996). Metamorphic evolution of garnet–spinel peridotites from the Variscan Schwarzwald (FRG). Geologische Rundschau 85, 211–224. Kalt, A., Grauert, B. & Baumann, A. (1994a). Rb–Sr and U–Pb isotope studies on migmatites from the Schwarzwald (Germany): constraints on isotopic resetting during Variscan high-temperature metamorphism. Journal of Metamorphic Geology 12, 667–780. Kalt, A., Hanel, M., Schleicher, H. & Kramm, U. (1994b). Petrology and geochronology of eclogites from the Variscan Schwarzwald (FRG). Contributions to Mineralogy and Petrology 115, 287–302. Kalt, A., Altherr, R. & Hanel, M. (1995). Contrasting P–T conditions recorded in ultramafic high-pressure rocks from the Variscan Schwarzwald (F.R.G.). Contributions to Mineralogy and Petrology 121, 45–60. Kalt, A., Berger, A. & Blumel, P. (1999). Metamorphic evolution of ¨ cordierite-bearing migmatites from the Bayerische Wald (Variscan Belt, Germany). Journal of Petrology 40, 601–627. Kalt, A., Altherr, R. & Hanel, M. (2000a). The Variscan basement of the Schwarzwald. European Journal of Mineralogy 12(Beiheft 2), 1–43. Kalt, A., Corfu, F. & Wijbrans, J. (2000b). Time calibration of a P–T path from a Variscan high-temperature low-pressure migmatite–gneiss complex (Bayerische Wald, Germany) and the detection of inherited monazite. Contributions to Mineralogy and Petrology 138, 143–163. Klein, H. & Wimmenauer, W. (1984). Eclogites and their retrograde transformation in the Schwarzwald (Fed. Rep. Germany). Neues Jahrbuch fur Mineralogie, Monatshefte 1984, 25–38. ¨ Klemd, R. & Brocker, M. (1999). Fluid influence on mineral reactions ¨ in ultrahigh-pressure granulites: a case study in the Szieznik Mts. 251 JOURNAL OF PETROLOGY VOLUME 44 (West Sudetes, Poland). Contributions to Mineralogy and Petrology 136, 358–373. Kossmat, F. (1927). Gliederung des variszischen Gebirgsbaues. Abhandlungen des Sachsischen Geologischen Landesamtes 1, 1–39. ¨ Kotkova, J. & Harley, S. L. (1999). Formation and evolution of high´ pressure leucogranulites: experimental constraints and unresolved issues. Physics and Chemistry of the Earth A 24, 299–304. Kotkova, J., Kroner, A., Todt, W. & Fiala, J. (1996). Zircon dating of ´ ¨ North Bohemian granulites, Czech Republic: further evidence for the Lower Carboniferous high-pressure event in the Bohemian Massif. Geologische Rundschau 85, 154–161. Koziol, A. M. & Newton, R. C. (1988). Redetermination of the anorthite breakdown and improvement of the plagioclase–garnet–Al2SiO5– quartz geobarometer. American Mineralogist 73, 216–223. Kretz, R. (1983). Symbols for rock-forming minerals. American Mineralogist 68, 277–279. Kriegsmann, L. M. & Hensen, B. J. (1998). Back reaction between restite and melt: implications for geothermobarometry and pressure– temperature paths. Geology 12, 1111–1114. Kroner, A. & Willner, A. P. (1998). Time of formation and peak of ¨ Variscan HP–HT metamorphism of quartz–feldspar rocks in the Central Erzgebirge, Saxony, Germany. Contributions to Mineralogy and Petrology 132, 1–20. Kroner, A., Jaeckel, P., Reischmann, T. & Kroner, U. (1998). Further ¨ evidence for an early Carboniferous (>340 Ma) age of highgrade metamorphism in the Saxonian granulite complex. Geologische Rundschau 86, 751–766. Kroner, A., O’Brien, P. J., Nemchin, A. A. & Pidgeon, R. T. (2000). ¨ Zircon ages for high-pressure granulites from South Bohemia, Czech Republic, and their connection to Carboniferous high temperature processes. Contributions to Mineralogy and Petrology 138, 127–142. Kryza, R., Pin, C. & Vielzeuf, D. (1996). High-pressure granulites from the Sudetes (south-west Poland): evidence of crustal subduction and collisional thickening in the Variscan belt. Journal of Metamorphic Geology 14, 531–546. Lammlin, I. (1981). Petrographische und geochemische Un¨ tersuchungen im Gneisgebiet des Sudschwarzwaldes. Ph.D. thesis, ¨ University of Freiburg im Breisgau. Latouche, L., Fabries, J. & Guiraud, M. (1992). Retrograde evolution ` in the Central Vosges Mountains (northeastern France): implications for the metamorphic history of high-grade rocks during the Variscan orogeny. Tectonophysics 205, 387–407. Lippolt, H. J., Hradetzky, H. & Hautmann, S. (1994). K–Ar dating of amphibole-bearing rocks in the Schwarzwald, SW Germany: I. 40Ar/39Ar age constraints to Hercynian HT-metamorphism. Neues Jahrbuch fur Mineralogie Monatshefte 10, 433–448. ¨ Loeschke, J., Guldenpfennig, M., Hann, H. P. & Sawatzki, G. (1998). ¨ Die Zone von Badenweiler–Lenzkirch (Schwarzwald): Eine variskische Suturzone. Zeitschrift der deutschen geologischen Gesellschaft 149, 197–212. Menard, G. & Molnar, P. (1988). Collapse of a Hercynian Tibetian ´ plateau into a late Palaeozoic Basin and Range Province. Nature 334, 235–237. Metz, R. (1980). Geologische Landeskunde des Hotzenwaldes. Moritz Schauenburg, Lahr/Schwarzwald, 1116 pp. Mirwald, P. (1986). Ist Cordierit ein Geothermometer? Fortschritte der Mineralogie 64(Beiheft 1), 119. Moller, A., Mezger, K. & Schenk, V. (2000). U–Pb dating of meta¨ morphic minerals: Pan-African metamorphism and prolonged slow cooling of high-pressure granulites in Tanzania, East Africa. Precambrian Research 104, 123–146. Montel, J. M. & Vielzeuf, D. (1997). Partial melting of metagreywackes. Part 2: Composition of minerals and melts. Contributions to Mineralogy and Petrology 128, 176–196. NUMBER 2 FEBRUARY 2003 Nichols, G. T., Berry, R. F. & Green, D. H. (1992). Internally consistent gahnitic spinel–cordierite–garnet equilibria in the FMASHZn system: geothermobarometry and applications. Contributions to Mineralogy and Petrology 111, 362–377. O’Brien, P. J. (2001). The fundamental Variscan problem: hightemperature metamorphism at different depths and high-pressure metamorphism at different temperatures. In: Franke, W., Haak, V., Oncken, O. & Tanner, D. (eds) Orogenic Processes: Quantification and Modelling in the Variscan Belt. Geological Society, London, Special Publications 179, 369–386. O’Brien, P. J., Kroner, A., Jackel, P., Hegner, E., Zelazniewicz, A. & ¨ ¨ Kryza, R. (1997). Petrological and isotopic studies on Palaeozoic high-pressure granulites, Gory Sowie Mts, Polish Sudetes. Journal of Petrology 38, 433–456. Patino Douce, A. E. & Johnston, A. D. (1991). Phase equilibria and ˜ melt productivity in the pelitic system: implications for the origin of peraluminous granitoids and aluminous granulites. Contributions to Mineralogy and Petrology 107, 202–218. Patino-Douce, A. E. & McCarthy, T. C. (1998). Melting of crustal ˜ rocks during continental collision and subduction. In: Hacker, B. R. & Liou, J. G. (eds) When Continents Collide: Geodynamics and Geochemistry of Ultrahigh-pressure Rocks. Dordrecht: Kluwer Academic, pp. 27–55. Peacock, S. M. (1996). Thermal and petrologic structure of subduction zones. In: Bebout, G. E., Scholl, D. W., Kirby, S. H. & Platt, J. P. (eds) Subduction Top to Bottom. Geophysical Monograph, American Geophysical Union 96, 119–134. Perkins, D., III & Chipera, S. J. (1985). Garnet–orthopyroxene–plagioclase–quartz barometry: refinement and application to the English River subprovince and the Minnesota River valley. Contributions to Mineralogy and Petrology 89, 69–80. Petford, N. & Gallagher, K. (2001). Partial melting of mafic (amphibolitic) lower crust by periodic influx of basaltic magma. Earth and Planetary Science Letters 193, 483–499. Pin, C. & Vielzeuf, D. (1983). Granulites and related rocks in Variscan median Europe: a dualistic interpretation. Tectonophysics 93, 47–74. Pin, C. & Vielzeuf, D. (1988). Les granulites de haute-pression d’Europe moyenne: temoins d’une subduction eo-hercynienne. Implications ´ ´ sur l’origine des groupes leptyno-amphiboliques. Bulletin de la Societe ´´ Geologique de France 8(IV), 13–20. ´ Rapp, R. P. & Watson, E. B. (1995). Dehydration melting of metabasalt at 8–32 kbar: implications for continental growth and crust–mantle recycling. Journal of Petrology 36, 891–931. Rehfeld, U. (1983). A relictic staurolite–kyanite mineral paragenesis in paragneisses of the Central Schwarzwald. Neues Jahrbuch fur Min¨ eralogie, Abhandlungen 146, 170–181. Rey, P., Burg, J.-P., Lardeaux, J.-M. & Fluck, P. (1989). Evolutions ´ ´ ´ metamorphiques contrastees dans les Vosges orientales: temoins d’un charriage dans la chaıne varisque. Comptes Rendues de l’Academie des ˆ ´ Sciences, Serie II 309, 815–821. ´ Romer, R. L. & Rotzler, J. (2001). P–T–t evolution of ultrahigh¨ temperature granulites from the Saxon Granulite Massif, Germany. Part II: Geochronology. Journal of Petrology 42, 2015–2032. ¨ Rotzler, J. & Romer, R. L. (2001). P–T–t evolution of ultrahightemperature granulites from the Saxon Granulite Massif, Germany. Part I: Petrology. Journal of Petrology 42, 1995–2013. Rutter, M. J. & Wyllie, P. J. (1988). Melting of vapour-absent tonalite at 10 kbar to simulate dehydration-melting in the deep crust. Nature 331, 159–161. Schaltegger, U. (2000). U–Pb geochronology of the Southern Black Forest Batholith (Central Variscan Belt): timing of exhumation and granite emplacement. International Journal of Earth Sciences 88, 814–828. Schaltegger, U., Fanning, C. M., Gunther, D., Maurin, J. C., Schul¨ mann, K. & Gebauer, D. (1999). Growth, annealing and re- 252 MARSCHALL et al. P–T EVOLUTION OF SCHWARZWALD GRANULITES crystallization of zircon and preservation of monazite in high-grade metamorphism: conventional and in-situ U–Pb isotope, cathodoluminescence and microchemical evidence. Contributions to Mineralogy and Petrology 134, 186–201. Schmadicke, E., Mezger, K., Cosca, M. A. & Okrusch, M. (1995). ¨ Variscan Sm–Nd and Ar–Ar ages of eclogite-facies rocks from the Erzgebirge, Bohemian Massif. Journal of Metamorphic Geology 13, 537–552. Seck, H. A. (1971). Der Einfluss des Drucks auf die Zusammensetzung koexistierender Alkalifeldspate und Plagioklase im System NaAl¨ Si3O8–KAlSi3O8–CaAl2Si2O8–H2O. Contributions to Mineralogy and Petrology 31, 67–86. Skjerlie, K. P. & Johnston, A. D. (1993). Fluid-absent melting behaviour of an F-rich tonalitic gneiss at mid-crustal pressures: implications for the generation of anorogenic granites. Journal of Petrology 34, 785–815. Stein, S. & Stein, C. A. (1996). Thermo-mechanical evolution of oceanic lithosphere: implications for subduction process and deep earthquakes. In: Bebout, G. E., Scholl, D. W., Kirby, S. H. & Platt, J. P. (eds) Subduction Top to Bottom. Geophysical Monograph, American Geophysical Union 96, 1–18. Stenger, R., Baatz, K., Klein, H. & Wimmenauer, W. (1989). Metamorphic evolution of the pre-Hercynian basement of the Schwarzwald (Federal Republic of Germany). Tectonophysics 157, 117–121. Stevens, G., Clemens, J. D. & Droop, G. T. R. (1997). Melt production during granulite-facies anatexis: experimental data from ‘primitive’ metasedimentary protoliths. Contributions to Mineralogy and Petrology 128, 352–370. Vielzeuf, D. & Holloway, J. (1988). Experimental determination of the vapor-absent melting relations in the pelitic system. Contributions to Mineralogy and Petrology 98, 257–276. Vielzeuf, D. & Pin, C. (1989). Geodynamic implications of granulitic rocks in the Hercynian belt. In: Daly, J. S., Cliff, R. A. & Yardley, B. W. D. (eds) Evolution of Metamorphic Belts. Geological Society, London, Special Publications 43, 343–348. von Quadt, A. (1993). The Saxonian Granulite Massif: new aspects from geochronological studies. Geologische Rundschau 82, 516–530. Voshage, H., Hunziker, J. C., Hofmann, A. W. & Zingg, A. (1987). A Nd and Sr isotopic study of the Ivrea zone, Southern Alps, N-Italy. Contributions to Mineralogy and Petrology 97, 31–42. Waters, D. J. (1991). Hercynite–quartz granulites: phase relations, and implications for crustal processes. European Journal of Mineralogy 3, 367–386. Wendt, J. I., Kroner, A., Fiala, J. & Todt, W. (1993). Evidence from ¨ zircon dating for existence of approximately 2·1 Ga old crystalline basement in southern Bohemia, Czech Republic. Geologische Rundschau 82, 42–50. Willner, A. P., Rotzler, K. & Maresch, W.V. (1997). Pressure– ¨ temperature and fluid evolution of quartzo-feldspathic metamorphic rocks with a relic high-pressure, granulite-facies history from the Erzgebirge (Saxony, Germany). Journal of Petrology 38, 307–336. Wimmenauer, W. (1984). Das pravariskische Kristallin im Schwarz¨ wald. Fortschritte der Mineralogie 62(Beiheft 2), 69–86. Wimmenauer, W., Klein, H., Muller, H. & Stenger, R. (1989). Petro¨ graphy and petrology of the KTB location Schwarzwald. In: Emmermann, R. & Wohlenberg, J. (eds) The German Continental Deep Drilling Program (KTB). Heidelberg: Springer, pp. 244–263. 253 ...
View Full Document

This note was uploaded on 03/19/2012 for the course GEOLOGY 6 taught by Professor Rm during the Spring '12 term at UNAM MX.

Ask a homework question - tutors are online