This preview shows page 1. Sign up to view the full content.
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
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
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
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: firstname.lastname@example.org
†Present address: Institut de Geologie, Universite de Neuchatel, Rue
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
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
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
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
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
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. 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
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
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/
(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
St + Qtz = Grt + Ky + V (5) probably formed garnets without hercynite, and the
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
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 &
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
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
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
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
(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–
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,
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,
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
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
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–
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,
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
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,
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,
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,
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,
Menard, G. & Molnar, P. (1988). Collapse of a Hercynian Tibetian
plateau into a late Palaeozoic Basin and Range Province. Nature 334,
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
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
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
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,
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,
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,
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
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.
- Spring '12