Full text of DOI:10.1007/s004100000233

Contrib Mineral Petrol %2001) 141: 322±346
DOI 10.1007/s004100000233
Leander Franz á Rolf L. Romer á Reiner Klemd
È
Robert Schmid á Roland Oberhansli á Thomas Wagner
Dong Shuwen
Eclogite-facies quartz veins within metabasites of the Dabie Shan
eastern China): pressure±temperature±time±deformation path,
ꢀ
composition of the ¯uid phase and ¯uid ¯ow during exhumation
of high-pressure rocks
Received: 17 May 2000 / Accepted: 22 November 2000 / Published online: 31 March 2001
Ó Springer-Verlag 2001
Abstract Metabasites in the high-pressure unit of the 221 Ma for the vein formation, which ®ts in the geo-
southern Dabie Shan %eastern China) contain quartz tectonic framework of the orogen. This rather wide age
veins with high-pressure mineral assemblages. Two veins range arises from the heterogeneous nature of the initial
with the parageneses quartz±paragonite±ankerite±rutile lead isotopic composition, re¯ectingthe hetero ge neous
and quartz±kyanite±talc±zoisite±rutile±calcite %pseud- sources of the lead and the ¯uid phase. Part of the ¯uid
omorph after aragonite) as well as quartz±kyanite±par- was probably generated by prograde dehydration reac-
agonite±garnet±omphacite were investigated in detail. tions in the country rocks of the veins, whereas strong
Host rocks of these veins are garnet amphibolites with evidence is also given for the derivation from an older
eclogitic relics and quartz eclogites, respectively. The basement, which also underwent subduction. The veins
oldest phase of deformation %D1) is recorded in the may have formed by extensive hydraulic fracturingor by
eclogite, which displays a schistosity with shape align- volume reduction duringeclo igt isation of the crust.
ment of high-pressure minerals such as omphacite, Subsequent to the formation of the veins, great parts of
phengite and garnet. An estimate of the metamorphic the high-pressure rocks experienced an intensive,
P±T conditions yields 19±21 kbar at 570±620 °C for the syn-tectonic %D2) amphibolite facies metamorphism at
eclogites. Fluid inclusion studies reveal a primary, low- 8.5±10.5 kbar and 590±645 °C, indicatingisothermal
salinity aqueous ¯uid phase responsible for the forma- decompression duringexhumation. In the course of this
tion of the quartz veins. This is in correspondence with exhumation, pseudosecondary CO2 inclusions were
calculations of phase equilibrium curves on minerals of trapped in quartz of the veins. The last recorded stage of
the vein paragenesis, which yield P±T conditions of the retrograde overprint took place under static,
1
9.4 kbar and 591 °C in the presence of an aqueous ¯uid greenschist facies conditions %450±480 °C) in the stability
phase. Geochronology using the U±Pb system of rutile ®eld of kyanite. The shape of the metamorphic P±T path
gives evidence for a Triassic minimum age of 207± for this stage is re¯ected by the ¯uid phase, which either
re-equilibrated duringthis event or penetrated the rock
on fractures. These ®ndings show that ¯uid ¯ow during
L. Franz %&)
Institut fur Mineralogie, TU Bergakademie Freiberg,
Brennhausgasse 14, 09596 Freiberg, Germany
E-mail: lfranz@mineral.tu-freiberg
subduction-related high-pressure metamorphism and
subsequent uplift is rather complex; the ¯uid phase may
originate from di€erent sources and may interact with
country rocks in the form of a chromatographic column
transformingmineral assembla ge s, and also chan gi ngits
own character.
È
R.L. Romer
GeoForschungsZentrum Potsdam, Telegrafenberg,
1
4473 Potsdam, Germany
R. Klemd á T. Wagner
Institut fur Mineralogie, Am Hubland,
È
Universitat Wurzburg, 97070 Wurzburg, Germany
È
È
È
Introduction
È
R. Schmid á R. Oberhansli
Institut fur Geowissenschaften, Universita
È
È
t Potsdam,
Karl-Liebknechtstraûe 24, 14455 Potsdam, Germany
Quartz veins, which are a common feature in metamor-
phic rocks, are generally interpreted to form syn-meta-
morphically because of high ¯uid pressure, giving rise to
extension %e.g. Yardley 1986). This high ¯uid pressure
often results from dehydration reactions duringprograde
metamorphism, which may induce hydraulic fracturing
D. Shuwen
Institute of Geomechanics, Chinese Academy of Geological
Science, 100081 Beijing, China
Editorial responsibility: J. Hoefs

3
23
in the rock %Voll 1960; Yardley 1983). These fractures units, as well as the Luzhenguan and Foziling `groups' in
enhance and channelise the ¯uid ¯ow in the rock and the greenschist facies %e.g. Okay 1993; Hacker et al.
allow the precipitation of quartz as well as the formation 1996, 2000; Liou et al. 1996). The peak P±T conditions
of other metamorphic minerals that also occur in the host for the highest grade rocks during the Palaeozoic/Me-
rock %Etheridge 1983). Mineral assemblages of such veins sozoic orogeny have recently been constrained at 40±
are usually/often generated during prograde greenschist 48 kbar and <750 °C %Schmid et al. 2000).
facies conditions and may incorporate quartz, albite,
Presently available chronological studies on the HP
calcite and, in some cases, sheet silicates such as chlorite and UHP rocks report ages of 800±700 Ma for the pro-
and biotite %Voll 1969). In other cases, however, mineral toliths and 240±220 Ma for the HP to UHP event %Ames
assemblages in such veins indicate formation during peak et al. 1993, 1996; Li et al. 1993; Xiao et al. 1995; Rowley
pressure and temperature conditions %e.g. Austrheim et al. 1997; Hacker et al. 1998). Initial uplift to crustal
1
1
987, 1990; Manning1994; A ug e 1997; Becker et al. levels %e.g. Hacker et al. 2000) was followed by Cretaceous
999; Brunsmann et al. 2000), whereas in some instances to Cenozoic extensional processes, facilitatingexhuma-
quartz veins may also form near the end of a metamor- tion to the surface and intrusions of syenitic to tonalitic
phic loop %e.g. von Raumer 1984). As these veins always plutons at ꢀ140 Ma ago %Ratschbacher et al. 2000).
have a very speci®c relation concerningtheir textures and
The area south of Lake Hualiangting %Fig. 1) is situ-
mineral assemblages to the metamorphic host rock, they ated in the HP terrain of Dabie Shan and is largely
may serve as a tool to unravel the tectonometamorphic dominated by white mica- and garnet-bearing orthog-
evolution as well as the character of the ¯uid phase and neisses, metasediments and lenses of quartz eclogites and
its ¯ow duringmetamorphism.
garnet amphibolites. Large quartz-dominated veins are
This paper presents a combination of petrological, associated with eclogites and garnet amphibolites, which
geochronological, and ¯uid inclusion studies on high- show eclogitic relics. The oldest recognisable deformation
pressure veins and their host rocks to constrain the event %D1) occurred under eclogite facies conditions and
metamorphic path as well as the composition and spe- led to a shape alignment of high-pressure minerals in the
ci®c properties of the syn-metamorphic ¯uid phase %i.e. eclogites and the formation of the S1 schistosity, which
the P±T±t±d±x conditions) in quartz eclogites from the follows in general a NW±SE trend with a south-westward
Dabie Shan %eastern China). We demonstrate that the dip of 30±40°. Subsequent to D1, the formation of the
¯
uid phase in Dabie Shan high-pressure rocks seems to quartz veins occurred on fractures oblique to S1. The D2
be rather uniform in composition as early trapped deformation took place under amphibolite facies condi-
aqueous ¯uid inclusion in quartz and omphacite in tions creatingthe S2-foliation in the amphibolites and
quartz eclogite of the Taihu area observed by You et al. gneisses. The orientation of the S2-foliation is parallel±
%1996) show great anities to our ®ndings. The mineral subparallel to the S1-schistosity in the eclogites.
assemblage used to derive the P±T conditions of vein
formation includes up to 5-cm large rutile crystals. Two occurrences of quartz veins were investigated. The
Geochronology of this rutile, therefore provides con- ®rst vein [see HPV1 %high-pressure quartz vein) in Fig. 1]
straints on the generation of the mineral assemblage is located within ankerite-bearing ga rnet amphibolites
of these veins. Datingof this rutile, furthermore, is
and outcrops near the road on the slope south of Lake
attractive as rutile has grown during vein formation and Hualiangting [GPS location: N 30 28,1361 E 116
is not inherited or derived from the wallrock. Although 08,3745; source of samples HPV1 %qtz-vein, rutile crys-
it is debated whether rutile U±Pb ages represent cooling tal) and 98117 %wallrock garnet amphibolite)]. It has a
or formation ages %e.g. Mezger et al. 1989; Santos Zal- thickness of ꢀ1 m and formed on a steep, NE±SW
duegui et al. 1996) the U±Pb age of rutile will constrain strikingfracture that lies discordantly to the SW-dipping
the exhumation history and its initial lead isotopic S2-foliation of the amphibolites. The vein consists
composition provides a ®ngerprint to the ¯uid source.
mainly of quartz, randomly oriented white mica ¯akes,
and large ankerite crystals as well as rutile %see Fig. 2a).
The second outcrop %see HPV2 in Fig. 1) is located to
the east of the ®rst locality near a road junction [GPS
location: N30 27,6919 E 116 09,5657; source of samples
HPV2a,b,c %qtz-veins), HPV2b %rutile crystal) and 98118
Geological setting and position of the veins
The Dabie Shan forms part of the Qinlingcollisional
belt between the Yangtze and Sino-Korean cratons in %host-eclogite)]. There, several interconnected quartz
east-central China %Mattauer et al. 1985). Continental veins occur within fresh eclogites. The eclogitic host
northward subduction and subsequent exhumation of rocks of the vein often show a compositional layering
the Yangtze plate led to a complex, tectonically imbri- with alternatingquartz-rich and ga rnet±omphacite-rich
cated sequence of deformed metamorphic rocks. From sections and in part reveal a penetrative foliation with
south to north it can be subdivided into the Yangtze- aligned omphacite prisms. Part of the eclogite, however,
foreland fold-and-thrust belt, the blueschist, the has a compact, porphyroblastic texture. The size of the
high-pressure amphibolite, the high-pressure %HP) quartz veins strongly varies from <1 to >50 cm. The
quartz-bearing eclogites, the ultrahigh-pressure %UHP) distribution and orientation of mineral phases in these
coesite-bearing eclogites, and the `northern orthogneiss' veins are highly variable. Two representative sections

3
24
Fig. 1 Geographic position of
the Dabie Shan and geological
sketch map with the sample
locations. Map modi®ed after
You et al. %1996) and the geo-
logical map of Dabie mountains
area of Anhui province from
Anhui Institute of Regional
Geological Survey %AIRGS,
1
997)
within this network of veins were investigated. The ®rst Major and minor elements were determined at 15 kV acceleration
voltage and a beam current of 20 nA with counting times of 20 s for
Si, Al, Mg, Ca, Sr, Ba and K, and 30 s for Fe, Ni, Na, Cr, Mn and
vein %HPV2a) has a thickness of ꢀ10 cm and mainly
consists of quartz with up to 5-cm-longrutile prisms and
minor amounts of kyanite and talc. The second vein
Ti. The standard sets of the Smithsonian Institute %Jarosewich et al.
980) and of CAMECA were used for reference. Selected microp-
1
%
HPV2b; HPV2c) has an irregular, gnarly shape %see robe analyses of the most important minerals are listed in Tables 1,
2, 3 and 4. The entire data set is available from the ®rst author on
request.
Fig. 2b). Beside quartz, it is made up of kyanite, talc,
paragonite, pseudomorphs of calcite after aragonite,
large zoisite prisms, rutile, as well as garnet and om-
phacite. Both investigated veins strike ENE±WSW and
are steeply dippingtowards NNW. They discordantly
cut the S1-foliation in the eclogites, which is rather
variable in this location showinga south-westward to
south-eastward dip of 40±65°.
Quartz veins
The quartz veins in the two locations show similar textural features
as they all formed on fractures that crosscut the main metamorphic
foliation of their host rocks. The vein HPV1 shows a porphyrob-
lastic texture in its centre with granoblastic quartz grains, randomly
oriented white mica ¯akes, and large ankerite porphyroblasts.
Occasionally, zoisite prisms appear. Accordingto microprobe
analyses, white mica is paragonite with a prominent amount of
celadonite and minor amounts of muscovite %Pg75±88Cel11±16
Ms0±5Mrg0±2). Ankerite yields an elevated amount of mag-
Petrography and mineral chemistry of veins
and host rocks
nesite component and shows
a chemical composition of
Minerals described below were analysed usinga JEOL JXA-8900R
Cc53Mgs34Sid12Rhc<1, which is very similar to the carbonate in the
electron microprobe at the University of Freibergor with Cameca hosting ga rnet amphibolite %see below). At the immediate contact
SX-50 and SX-100 electron microprobes at the GFZ Potsdam. to the surroundingmetabasites, dynamic recrystallisation led to the

3
25
HPV2a, for example, is dominated by quartz and large rutile
prisms, whereas kyanite and talc are accessory phases. In thicker
parts of the quartz vein, rather complex associations of minerals
are found. Specimen HPV2b contains the mineral assemblage
quartz±kyanite±zoisite±talc ‹ calcium carbonate %former arago-
nite as indicated by the crystal shape) ‹ rutile. This section of the
vein also bears fragments of garnet xenocrysts, whose composition
2
is rather uniform %Alm49±51Pyp31±34Grs14±16Sps ) and strongly
resembles the rim composition of the eclogitic garnet. Other parts
of the vein are dominated by quartz±paragonite %Pg90±94Mrg2±3
Ms3±6)±kyanite. Eventually, small symplectitc intergrowths of
diopside-rich clinopyroxene and albite occur. Especially paragonite
and zoisite are often surrounded by small, secondary rims of albite
%
An0.6±6.4) and tremolite. In specimen HPV2c, garnet from the host
rock displays a preferential growth into the vein, which is evident
by its elongated rim section %cf. Fig. 3). The inner part of this
garnet reveals a compositional zoning with increasing Mg contents
at the expense of Ca %core: Alm54Pyp18Grs26Sps
1
; near rim:
Alm54Pyp28Grs18Sps ), which is quite similar to the garnet of the
1
host eclogite %cf. Fig. 4 and Table 4). The newly grown rim section
is characterised by distinctly higher Mg contents and lower
almandine and grossular components %Alm46±49Pyp30±35Grs15±19
1
Sps ). Garnet is accompanied by paragonite %Pg91±94Mrg0±1Ms5±8),
kyanite, and omphacite %Jd36±40), which experienced a strong
decomposition to magnesio-hornblende. Like garnet, all these
minerals show a growth perpendicular to the wall of the vein.
Representative mineral analyses from the vein assemblages are
listed in Tables 1, 2, 3 and 4.
Host rocks
The host rocks of the veins mainly consist of metabasites, namely
garnet amphibolites in case of HPV1 and eclogites in location
HPV2.
Garnet amphibolites with a remarkable variability in texture are
the host of vein HPV1. The amphibolites usually show a penetra-
tive S2-foliation, which is made up of aligned hornblende, plagio-
clase, chlorite, and in part elongated garnet. Accessories are
epidote, biotite, ilmenite and rutile. Two distinctly di€erent gen-
erations of garnet can be distinguished in sample 98117-3: garnet
porphyroblasts %see pGrt in Table 4) with diameters of up to 1 cm
often show an elongation perpendicular to the main foliation and
reveal irregular compositional zoning patterns. The likely prograde
growth of these crystals, however, is indicated by the highest XMg
values near the rims although retrograde sections may be developed
Fig. 2 a Photograph of vein HPV1 showing a large ankerite crystal
accentuated by an ink line) in a matrix of quartz and paragonite. b
Photograph of vein HPV2b displaying randomly oriented zoisite
prisms in the centre of the vein, short prismatic rutile %right of the
zoisite) as well as kyanite and talc near the contact to the host
eclogite. The diameter of the coin is 2.5 cm
%
formation of small quartz subgrains and new grains. This recrys- where an intergrowth with hornblende appears. Small elongated
tallisation probably resulted from mylonitisation because of rhe- garnet grains with a length of 0.5±1.3 mm are also aligned in the
ological contrasts between metabasite and quartz vein during the foliation and reveal a homogeneous composition in the core
exhumation. To a smaller extent, similar dynamic recrystallisation %Alm60±62Pyp15±17Grs19±21Sps2±3; see aGrt in Table 4), which cor-
features of vein quartz are found in intergranular positions. In the responds to the rim sections of the large garnet porphyroblasts with
recrystallised sections near to the contact of the host rock, frag- maximum XMg. The rim sections of these small garnet grains also
ments of strongly chloritised garnet are observed, which reveal a give evidence for a retrograde overprint with a composition of
weak chemical zoningwith decreasing gr ossular contents %Grs 20±14
and increasingalmandine and pyrope contents %Alm 58±62; Pyp20±23
)
)
65 13 17 5
Alm Pyp Grs Sps . The occurrence of these di€erent types of
garnet is interpreted in the following way: porphyroblastic, irreg-
from core to rim. In some locations, a small retrograde rim with ularly zoned types highlight a polyphase evolution with prograde
distinctly decreasingX Mg [=Mg/%Mg+Fe)] values and increasing growth and subsequent fracturing and rearrangement of small
grossular contents %Alm57Pyp18Grs23Sps
garnet, numerous small ¯akes of biotite had formed and were later Although the prograde growth may have taken place prior or
a€ected by a strongchloritisation. As the composition of these
duringeclogite facies conditions, the rupturingand recrystallisation
garnet fragments strongly resembles the garnet in the host rock, may have been the consequence of exhumation and amphibolite
they are anticipated to be xenocrysts from the wallrock.
facies overprint. The fact that growth zoning was not erased in
2
) is developed. Around pieces, stripped from the formerly broken crystals, by ductile ¯ow.
The samples from vein HPV2 in the second location also show spite of the structural and metamorphic overprint of the garnet is
granoblastic quartz grains and randomly oriented porphyroblasts probably because of its relatively large size compared with garnet
in their central parts. Near the contact to the host eclogite, how- from other amphibolites %cf. Fig. 4). Small, elongated garnet
ever, kyanite, zoisite and talc tend to grow perpendicular to the crystals may have formed as a result of syn-tectonic amphibolite
vein walls, which is especially prominent in sample HPV2b and facies crystallisation. Accordingto the scheme of Leake et al.
HPV2c. This texture may be interpreted as stretched crystals in the %1997), Ca-amphibole prisms aligned in the S2-foliation can mainly
sense of Ramsay and Huber %1983) or as ®brous veins %Passchier be classi®ed as tschermakite and magnesio-hornblende, whereas
and Trouw 1996) and thus probably developed duringthe pro-
small xenoblastic grains intergrown with the rim section of garnet
gressive opening of the fracture. Remarkably, there is no dynamic are pargasite %see Table 3). The aligned prisms show a slight
recrystallisation of vein quartz near the contact of the eclogite. The increase in Al and Ti towards the rim, whereas the small grains are
modal mineral amount of these veins is rather variable. Sample unzoned. Followingthe scheme of Hey %1954), chlorite of sample

326

3
27

328

3
29

330
Fig. 3 Backscattered electron image of a garnet, which shows a the garnet rims and as small rims around garnet. Aligned plagio-
preferential growth into the vein %i.e. towards point B) and clase shows relatively constant composition of about 14 molꢀ An,
microprobe pro®le through this garnet. Note the similarity of the whereas plagioclase intergrown with hornblende has slightly ele-
compositional zoning in the core of this garnet to the garnet in the vated Ca contents of An15±19, and the highest An contents %18±
host eclogite %sample 98118-2 in Fig. 4)
23 molꢀ) are recorded in plagioclase rimming garnet %see Table 3).
Epidote is often intergrown with plagioclase and also shows an
alignment in the foliation. The small prisms are virtually unzoned
with XPs of ꢀ0.22. Ilmenite forms small tablets in the matrix and
98117-3 is ripidolite with XFe values of ꢀ0.42. Accordingto the
method of Laird and Albee %1981), chlorite only contains minor charge balance constraints indicate elevated contents of ferric iron
amounts of ferric iron. Like chlorite, biotite also shows an align- with up to 15 molꢀ of haematite component. Carbonate in the
ment in the S2-foliation. Microprobe analyses reveal TiO
of ꢀ2 wtꢀ and XMg values of ꢀ0.54. Selected analyses of biotite contents of less than 0.03 wtꢀ %Table 2).
and chlorite are listed in Table 2. Plagioclase occurs as aligned
Garnet amphibolite in the immediate vicinity of vein HPV1, i.e.
crystals in the foliation, as euhedral intergrowth with hornblende at in a layer of about 5 cm around the vein, shows a massive, por-
2
contents matrix is ankerite with Cc Mgs Sid Rhc
<1
and SrO + BaO
5
0
35
15

3
31
Fig. 4 Comparison of microprobe pro®les through composition- of the hornblende population. They are intergrown with ®ne pla-
ally zoned garnet porphyroblasts from a quartz eclogite %sample gioclase lamellae yielding anorthite contents of 15±19 molꢀ %for
98118-2) and garnet amphibolite %sample 98117-4), and through an microprobe analyses of plagioclase and Ca-amphibole see Table 3).
elongated garnet of garnet amphibolite sample 98117-3. Thin lines In the centre of the symplectites %sym), clinopyroxene with variable
indicate zones of former rupturingand recrystallisation during
amphibolite facies ductile shearing
jadeite content of 7±32 molꢀ and XFe values of 0.21±0.28 are
present %see Cpx symp in Table 4). They are armoured by plagio-
clase and sometimes rimmed by hornblende and probably represent
the ®rst stage of omphacite and garnet breakdown following the
peak of eclogite facies metamorphism. Epidote forms large prisms
phyroblastic texture. An example of this can be observed in sample in the matrix but also ®ne grains in the symplectites with both types
8117-4, where garnet crystals up to 1 cm in diameter are rimmed showinga very similar composition, i.e. X Ps values of 0.12±0.15.
9
by green hornblende and plagioclase, and the matrix is made up of Paragonite, which is often strongly decomposed, yields a compo-
quartz, epidote, paragonite, ankerite and prismatic, ®ne grained sition of Pg88±91Cel0±2Ms6±8Mrg2±4, thus resemblingthe para go nite
symplectites. Accessories are rutile, ilmenite and rare apatite. Mi- on the adjacent quartz vein %see Table 1). Similar to sample 98117-
croprobe pro®les through large garnet porphyroblasts of sample 3, ilmenite yields an elevated amount of ferric iron which results in
98117-4 reveal compositional zonation patterns with strongly ele- a haematite component of up to 10 molꢀ.
vated spessartine and grossular contents in the centre and a com-
Eclogites hosting vein HPV2 show porphyroblastic textures
position of Alm53±54Pyp9±10Grs26±27Sps9±10. Towards the rim, an with euhedral garnet in a matrix of omphacite, zoisite and white
increase in XMg and a decrease in the spessartine and grossular mica. Alignment of zoisite and omphacite often de®nes a foliation/
component is observed and only in the outermost rim section is this lineation, whereas randomly scattered garnet porphyroblasts occur
pattern reversed, which re¯ects the prograde growth followed by a both as atoll-types enclosingwhite mica, rutile, and omphacite, and
retrograde overprint %cf. Fig. 4). Garnet composition near the rim as inclusion-free grains. A retrograde development of pale-green
is Alm61±62Pyp18±19Grs18±19Sps1±2 %cf. Table 4) and this composi- hornblende mainly took place at the rim of the garnet, whereas
tion is also found in small garnet grains that are totally surrounded clinopyroxene±plagioclase symplectites developed around ompha-
by symplectites of clinopyroxene, hornblende and plagioclase and cite. Two representative eclogite samples %98118-2 and 98118-5),
probably completely re-equilibrated duringthe peak of the am-
phibolite-facies overprint. Two distinctly di€erent generations of ent mineral assemblages. Sample 98118-2 yields the paragenesis
Ca-amphibole are observed. Amphibole formingsmall rims around
omphacite±zoisite±garnet±phengite ‹ rutile, whereas sample
which were investigated with the microprobe, show slightly di€er-
garnet as well as millimetre-sized prisms in the matrix is pargasite 98118-5 displays the same assemblage except for the white mica,
and ferropargasite according to the scheme of Leake et al. %1997). which is paragonite. Both samples reveal prominent compositional
3
+
The crystals yield XFe values of 0.45±0.57 %midpoint Fe
lation usingthe method of Papike et al. 1974) and show a certain
calcu- zoningin ga rnet with increasingalmandine %Alm 43±52) and pyrope
%Prp18±38) contents from core to rim at the expense of grossular
variability in Si, Al, and Ti contents although a common zonation %Grs34±13) and spessartine %Sps4±1). This type of zonation pattern is
pattern is not observed. Small grains within the symplectites are mainly present in large, inclusion-free garnet crystals %cf. Fig. 4 and
magnesio-hornblende with distinctly lower XFe values than the rest Table 4), whereas it is less pronounced or even completely missing

332
in atoll-type garnet. Steadily increasing XMg values from core to strained for the eclogites using the jadeite content of
rim highlight the prograde development of this kind of garnet,
whereas retrograde rims with slightly decreasing XMg are only re-
corded in sections with secondary amphibole overgrowth. Om-
phacite in both samples shows jadeite contents of 48±54 molꢀ
omphacite %Holland 1980, 1990), which points to 15±
16 kbar at a temperature of 600 °C. Similar minimum-
pressures are obtained with phengite barometry using
%sample 98118-2) and 42±45 molꢀ %sample 98118-5), respectively. the approach of Massonne %1991). Barometry of the
While there was no zoningin jadeite content, an increase in X Mg
from core to rim could be observed in most omphacite grains.
Charge balance constraints following Droop %1987) point to ferric
iron contents of 0±40ꢀ of total iron in the omphacite, which is
inverse tschermaks exchange among garnet, clinopy-
roxene and phengite of Waters and Martin %1993; up-
dated calibration of 1996 with activity models of
variable from grain to grain %cf. Table 4). This variability may Newton and Haselton %1981) and Holland %1990) for
result from %1) uncertainties in charge balance because of low
Fetotal contents and %2) a retrograde overprint leading to partial
decomposition of the omphacites %i.e. alteration to hornblende).
Zoisite hardly reveals any zoningexcept for a little hi gh er amount
garnet and omphacite, respectively) results in pressures
of 19±21 kbar at temperatures of 600 °C for garnet rim
sections with the highest X
values, adjacent jadeite
Mg
of ferrous iron in the core section of some specimen. Phengite rims and phengite cores with the highest Si contents %i.e.
grains of sample 98118-2 yield Si contents of 3.4±3.5 p.f.u. with 3.50±3.53 p.f.u.). An estimate of the P±T-conditions
relatively constant XMg values of 0.86±0.88, whereas paragonite of
sample 98118-5 has a composition of Prg93±95Ms2±5Mrg2±3 %cf.
Table 1). Secondary pyroxene in the symplectites is diopside with
duringexhumation is hampered by evident mineral
chemical disequilibria within the eclogites as already
indicated by the mainly coronitic textures. Temperature
minor amounts of jadeite %Jd11±30; see Table 4), which shows in-
tergrowth with plagioclase of An9±13 %sample 98118-2) and An17±25 calculations for the Fe±Mg-exchange between garnet
%sample 98118-5). Secondary amphibole displays a rather variable
rims and symplectitic hornblende usingthe calibration
of Graham and Powell %1986) result in a wide scatter of
composition, which may di€er distinctly within the same section. In
the symplectites of sample 98118-2, pargasite and edenite are ob-
served. Sample 98118-5 bears pargasite, edenite, tschermakite and
460±750 °C. Similarly, secondary clinopyroxene in the
magnesio-hornblende near garnet rims and within symplectites, symplectites show strongly variable temperatures of
whereas large, blue-green intergranular grains are barroisite %am-
phibole classi®cation of Leake et al. 1997). Most of these minerals
3
70±650 °C for the Fe±Mg-exchange with the garnet rim
usingthe Grt±Cpx thermometry of Kro gh %2000).
3
+
show Fe
contents of 20±30ꢀ usingthe midpoint calculation
method of Papike et al. %1974). Microprobe analyses of amphibole
and plagioclase are listed in Table 3.
Quartz veins
Because of a lack of critical mineral assemblages, an
estimate of the P±T conditions duringthe formation of
the quartz vein HPV1 is not possible. The lack of tex-
tural constraints furthermore precludes any ultimate
statement about the time of its formation, although the
Pressure±temperature evolution
Eclogite
As evident from clinozoisite aggregates pseudomor- presence of minerals such as paragonite and ankerite,
phous after porphyroblastic lawsonite, as well as from which are also present in the adjacent relic eclogite,
inclusions of glaucophane in garnet of eclogites south of point to its generation under HP conditions.
lake Hualiangting %Castelli et al. 1998), a clockwise
Mineral textures and assemblages in quartz veins
prograde P±T path can be inferred for these rock. This is HPV2 o€er the opportunity to directly constrain the
also indicated by typical prograde compositional zona- P±T conditions during formation. The growth of garnet
tion patterns in garnet of the investigated eclogites %cf. next to omphacite, paragonite and kyanite in sample
Fig. 4), which points to maximum temperatures of less HPV2c highlights the generation of this vein under
than 650±700 °C %Yardley 1977; Schliestedt 1980; Spear eclogite facies conditions. Minimum pressures of 12.5±
1991). Furthermore, chemical zoningin ag rnet and
13.5 kbar are estimated by the jadeite barometry of
omphacite testi®es to equilibrium conditions between Holland %1980; 1990), whereas the Grt±Cpx-geother-
these minerals in their rim sections. Garnet±clinopy- mometer of Krogh %2000) points to temperatures of 580±
roxene thermometry performed on garnet and nearby 630 °C. Maximum pressures for the formation of the
omphacite %not in direct contact with the garnet) yields veins are constrained by the reaction paragonite M
temperatures of 570±620 °C at 22 kbar for sample kyanite + jadeite + H O. A calculation of the reaction
2
98118-2 usingthe latest calibration of Krogh %2000). The
curve usingthe TWEEQ-pro gr am of Berman %1991) and
use of this calibration gave results very similar to other minerals of sample HPV2c yields pressures of 20±
calibrations performed on HP rocks %e.g. Krogh 1988; 21.5 kbar at temperatures of 580±630 °C and a water
Brey and Ko
È
hler 1990), whereas older calibrations %e.g. activity of 1.0 %cf. Fig. 6). A reduction of the water ac-
Ellis and Green 1979) overestimated these temperatures tivity would lead to a distinct decrease in pressure. For
by 70±90 °C. Application of garnet±phengite thermom- sample HPV2b, the P±T-X condition were modelled in
etry %Green and Hellman 1988) on garnet rims and core the CaCMASH system with the TWEEQ-program of
sections of phengite results in temperatures of 635± Berman %1991). The vein assemblage quartz±zoisite±
711 °C, but because of the missingestimate for ferric
kyanite±talc±aragonite as well as eclogitic garnet and the
iron in phengite, these temperatures have to be regarded ¯uid phases H O and CO %for which proof was given by
2
2
as maximum values. Minimum pressures can be con- ¯uid inclusion studies, see below) were selected for this

3
33
calculation. Values for H, V and S of aragonite given in 1987). The garnet±hornblende±plagioclase±quartz geo-
the VERTEX program-package of Connolly %1990) were barometer of Kohn and Spear %1990) indicates pressures
implemented into the TWEEQ database. A calculation of 9±10.5 kbar, which coincides with results of former
of phase equilibrium curves including17 mineral reac- calibrations for this assemblage %e.g. Kohn and Spear
tions %with three independent reactions) revealed an in- 1989). Similarly, pressures of 9.5±11 kbar were calcu-
tersection at 19.6 kbar and 591 °C at a water activity of lated usingthe GRIPS- eg obarometer of Bohlen and
0.88 and a CO activity of 0.12 %Fig. 5). Although the Liotta %1986).
two data sets used are not internally consistent, the in-
2
Garnet amphibolite sample 97117-4 only locally re-
corporation of the aragonite data did not only lead to produces these thermobarometric estimates. P±T-con-
meaningful results concerning the thermobarometry, but ditions of 8.5±10 kbar at 578±648 °C are calculated for
also ®tted well with the ¯uid inclusion studies %see garnet rims and adjacent plagioclase±hornblende±sym-
below) thus justifyingthis extension of the database.
plectites. Wider hornblende rims around garnet, how-
ever, yield unrealistically high temperatures of up to
7
50 °C for the Fe±Mg-exchange with the garnet, which
Garnet amphibolite
may be caused by local mineral disequilibria. Clinopy-
roxene±plagioclase symplectites around garnet rims
Syn-tectonic amphibolite facies metamorphism generat- highlight an earlier metamorphic stage with elevated
ed the D2 foliation, which can be well observed in the pressures. The symplectitic clinopyroxene shows jadeite
host rocks of vein HPV1. There, the strongly foliated contents of up to 32 molꢀ, which points to pressures of
garnet amphibolite sample 98117-3 reveals peak meta- up to 13.5 kbar usingthe jadeite barometry of Holland
morphic temperatures of 590±630 °C, based on garnet %1980, 1990). These crystals also developed an Fe±Mg-
3
+
rim sections and hornblende %midpoint Fe
usingthe calibration of Graham and Powell %1986).
These temperatures are also obtained by garnet±chlorite Grt±Cpx-thermometer of Krogh %2000).
thermometry %Dickenson and Hewitt 1986; Ghent et al. Figure 6 shows the tectonometamorphic evolution in
estimate) exchange equilibrium with the garnet rims as indicated
by temperatures of 560±615 °C calculated with the
the southern Dabie Shan startingwith a subduction-type
prograde P±T path of the eclogites deduced from the
observations of Castelli et al. %1998) and from compo-
sitionally zoned garnet. Peak metamorphic conditions
can be estimated at 570±620 °C and 19±21 kbar, which is
in accordance with earlier investigations and the absence
of coesite %Okay 1993; Carswell et al. 1997). The error of
this P±T-box would be ‹2.5 kbar for the garnet±om-
phacite±phengite geobarometer of Waters and Martin
%
1993) and ‹30 °C for the garnet±clinopyroxene geo-
thermometer of Krogh %2000). In the light of our new
and other published P±T estimates %e.g. Carswell et al.
1997), high temperatures %³700 °C) for the quartz
eclogites of southern Dabie Shan, as reported by Castelli
et al. %1998) have to be considered as an over-estimate.
HP-metamorphism was takingplace syn-tectonically as
evident from the penetrative D1 foliation with an
alignment of high-pressure minerals such as omphacite,
phengite and clinozoisite.
The formation of the quartz veins initiated at eclogite
facies P±T conditions and was connected with the
presence of H O-rich ¯uid phase. Their generation may
2
be a consequence of extensive hydraulic fracturing
caused by prograde metamorphic dehydration or of
particular deformation episodes as a result of the bulk
volume reduction during eclogitisation of the crust %e.g.
Austrheim 1998). Fracturingand vein formation, how-
ever, may also have been triggered by extensional tec-
tonic processes and ¯uid activity duringthe initial sta ge s
Fig. 5 Calculation of phase equilibrium curves in the CaCMASH of exhumation.
system for the HP mineral assemblage of vein HPV2b using the
TWEEQ program of Berman %1991). The calculation yields an
intersection at 19.6 kbar and 591 °C at a water activity of 0.88 and
2
a CO activity of 0.12. Mineral reactions, activity models and
thermodynamic data set see Appendix
In the course of the exhumation, symplectites of
clinopyroxene and plagioclase developed in the eclogites
and, subsequently, symplectites and coarse grained in-
tergrowths of plagioclase and hornblende were gener-

3
34
Fig. 6 P±T±t±d evolution of the
HP-rocks from the Taihu re-
gion showing a prograde
clockwise P±T evolution and an
isothermal decompression.
TWEEQ calculation for the
post-D1 extensional veiningthe
are indicated by intersecting
phase equilibrium curves. Am-
phibole out reaction after Car-
swell %1990), glaucophane out
after Maresch %1977), lawsonite
out after Heinrich and Althaus
%1981, 1988) and coesite quartz
transformation as well as par-
agonite out reaction calculated
with the TWEEQ program of
Berman %1991; see Appendix).
For comparison, the P±T path
of Okay %1993) for quartz
eclogites of the southern Dabie
Shan is outlined by grey stars
ated. Later on, the penetrative D2 deformation a€ected
the rock leadingto the decomposition of the HP-
assemblages as observed in the metabasic host rocks of
vein HPV1, where schistose garnet amphibolite formed.
Memory of the HP event was only preserved near the
quartz vein, where symplectitic plagioclase and clino-
pyroxene were protected from the amphibolite facies
overprint. There, the massive quartz vein may have
acted as a chemical barrier hamperingthe penetration of
Fluid inclusion studies
Microthermometric measurements were performed us-
ing the USGS gas ¯ow heating±freezing stage at the
È
University of Wurzburg. The estimated precision of
nal meltingtemperatures %Tm) and homo eg nisation
®
temperatures of CO2 inclusions is ꢀ‹0.2 °C; ho-
mogenisation temperatures of aqueous inclusions %Th)
have a reproducibility of about ‹3 °C. A detailed
description of the analytical and calibration proce-
dures is given in Klemd et al. %1992). Calculation of
densities and isochores was carried out usingthe
FLINCOR program package %Brown 1989). Densities
¯
uid phase duringthe amphibolite facies overprint and
thus preservingtexture and mineral relics of the eclo gi te
facies in its immediate vicinity. A mechanical shielding
of the wallrock by the quartz vein can easily be excluded
as the vein quartz should ¯ow/deform much easier at
temperatures of ꢀ600 °C than its side wall. Formation
of garnet amphibolite with the paragenesis hornblende±
plagioclase±garnet±quartz took place at pressures of
of H O±NaCl inclusions were calculated applyingthe
2
equation of state of Zhangand Frantz %1987), whereas
the equation of Holloway %1981) was used for CO2
inclusions.
A total number of 161 ¯uid inclusions were mic-
rothermometrically investigated in four doubly polished
thick sections of the high-pressure quartz vein samples
HPV1, HPV2a and HPV2b. Fluid inclusions were found
in kyanite and vein quartz. Based on textural criteria
and ¯uid composition, three principal types of ¯uid
inclusions can be distinguished:
8
.5±10.5 kbar and temperatures of 590±645 °C.
At a very late stage, a static greenschist facies
overprint led to the growth of albite and tremolite in
the veins, for which temperatures of 450±480 °C may
be estimated usingTi in Hbl thermometry of Colomby
%
1988) and Plg±Hbl geothermometry %e.g. Spear 1980).
Although the exact retrograde path from the am-
phibolite facies stage cannot be reconstructed, it is
rather probable that it remained within the stability
®
eld of kyanite as indicated by the sole occurrence of Type 1. Texturally primary two-phase H O inclusions
2
this alumino-silicate phase in the veins and the host in kyanite %HPV2b). These inclusions are elongated and
rocks.
commonly display negative crystal morphologies
These data clearly highlight an isothermal decom- %Fig. 7a). They occur either as individual inclusions or as
pression path for the eclogites in the Taihu area, which groups oriented parallel to the c-axis of the kyanite
®
ts well in the pattern of P±T paths recorded for the crystals. The size of the inclusions is in the range of 10±
Dabie Shan so far %e.g. Wang et al. 1990, 1992, 1995; 150 lm; degrees of ®ll range usually between 80±
Okay 1993; Liou et al. 1995, 1996; Carswell et al. 1997; 90 volꢀ. This type of ¯uid inclusion represents the
cf. Fig. 6).
oldest recognisable generation of ¯uid inclusions.

3
35
%1), because recrystallisation of quartz within the HP
veins took place after crystallisation of kyanite.
Type 3. Secondary two-phase H O inclusions in re-
2
crystallised vein quartz %HPV1, HPV2a, HPV2b). This
¯uid inclusion type is very abundant in all investigated
samples. In contrast to inclusions of type %2), they occur
alonghealed microfractures crosscuttingthe recrystalli-
sation texture of the vein quartz. They display very ir-
regular shapes and high degrees of ®ll %>97 volꢀ). The
size of inclusions is in the range of 10±35 lm.
Microthermometric results
Temperatures of ®nal meltingof primary aqueous two-
phase inclusions in kyanite %type 1) are in the range
between ±0.4 and ±1.2 °C %Table 5), correspondingto
very low salinities of 0.7±2.1 wtꢀ NaCl-equivalent
%
Fig. 8a). Eutectic melting was optically dicult to ob-
serve and occurred between ±25 and ±28 °C, indicating
the presence of NaCl and KCl in the aqueous phase
%
Crawford 1981; Roedder 1984). Clathrate meltingcould
not be observed. All aqueous ¯uid inclusions in kyanite
homogenised into the liquid phase. Homogenisation
temperatures range between 230 and 290 °C %Fig. 9) with
a maximum at 260‹20 °C. The resultingdensities are in
3
the range of 0.71±0.83 g/cm . The salinities of these in-
clusions are rather constant as well as the homogenisa-
tion temperatures, therefore, it can be inferred that these
inclusions were derived from a single homogeneous ¯uid
phase. You et al. %1996) also exclusively observed type 1
inclusions with similar densities and textural occurrence
in quartz and omphacite in coesite-free eclogites from
the Taihu area.
Fig. 7 Photomicrographs showing representative textures of ¯uid
inclusions from the studied high-pressure veins, Dabie Shan.
a Elongated primary two-phase H
sample HPV2b. Width of ®eld, 90 lm. b Trail of pseudosecondary
monophase CO inclusions %type 2) within recrystallised quartz;
sample HPV2b. Width of ®eld, 110 lm
2
O inclusion %type 1) in kyanite;
The pseudosecondary CO inclusions in recrystallised
2
2
vein quartz %type 2) display ®nal meltingtemperatures
very close to ±56.6 °C %±56.6 to ±56.9), indicatingthe
Type 2. Texturally pseudosecondary monophase CO2 presence of insigni®cant amounts of N or CH only.
2
4
inclusions in recrystallised vein quartz %HPV2b). Fluid Raman analysis only revealed CO . Consequently, nei-
2
inclusions of this type occur as clusters or trails that do ther temperatures of initial meltingof the CO phase nor
2
not crosscut the grain boundaries between individual additional phase transitions at low temperatures could
quartz grains. They display spheroidal/ellipsoidal or be observed. The homogenisation temperatures of the
negative crystal shapes and are usually isometric CO inclusions display a relatively wide variation rang-
2
%Fig. 7b). Inclusion sizes are in the range of 5±20 lm. ingbetween ±12.6 and 9.5 °C with no clear maximum
This ¯uid inclusion type was trapped subsequent to type detectable %Fig. 10). Homogenisation always took place
Table 5 Microthermometric results of ¯uid inclusions from the studied high-pressure veins
Sample and type of ¯uid inclusion System
Tm CO
°C)
2
Th CO
%°C)
2
Tm ice
%°C)
Th tot
%°C)
Density
%g/cm )
3
%
HPV1
Type 3, secondary in Qtz %n=21)
HPV2a
Type 3, secondary in Qtz %n=27)
HPV2b
H
H
H
2
2
2
O±NaCl±KCl
O±NaCl±KCl
O±NaCl±KCl
±0.5 to ±2.0
±0.5 to ±6.1
±0.4 to ±1.2
158 to 200
123 to 165
233 to 293
0.973±0.932
0.923±0.983
Type 1, primary in Ky %n=45)
Type 2, pseudosecondary in Qtz
0.713±0.827
0.924±0.997
CO
2
±56.6 to ±56.9 ±12.6 to 9.5
%n=58)
Type 3, secondary in Qtz %n=10)
H
2
O±NaCl±KCl
±0.4 to ±0.6
150 to 173
0.904±0.928

3
36
Fig. 9 Final melting% Tm) and homogenisation temperatures %Th)
2
of both primary and secondary two-phase H O inclusions %types 1
and 3)
2
Fig. 8 Histograms displaying the salinities of two-phase H O
inclusions %types 1 and 3). a Primary inclusions in kyanite %type 1);
b secondary inclusions in quartz %type 3)
into the liquid phase and the calculated densities are in
the range of 0.82±0.99 g/cm . No correlation between
3
inclusion size and density %Th) can be observed %Fig. 11),
although re-equilibrated ¯uid inclusion assemblages may
display such a relationship as a result of pressure dif-
ferences between internal pressure and con®ningpres-
sure %Bodnar et al. 1989; Hall et al. 1991).
Secondary aqueous two-phase inclusions in vein
quartz %type 3) display ®nal meltingtemperatures in the
range between ±0.5 and ±6.1 °C, correspondingto rela- densities of 0.87±0.98 g/cm . There is no correlation
tively low salinities of 0.7±9.3 wtꢀ NaCl equivalent between ®nal meltingtemperatures and homo ge nisation
Fig. 10 Histogram of homogenisation temperatures of pseudosec-
ondary CO inclusions in vein quartz %type 2)
2
3
%Fig. 8b). Most inclusions of this textural type have sa- temperatures detectable %Fig. 9).
linities in the range of 0.7±4.0 wtꢀ NaCl, including all
inclusions from samples HPV2a and HPV1, whereas
only 12 inclusions from sample HPV2b display some- Conditions of trappingand sources of the ¯uids
what higher salinities in the range 5.4±9.3 wtꢀ NaCl
equivalent. The resultingdistribution is, therefore, dis-
Maximum and minimum density isochores were con-
tinctly bimodal %Fig. 8b and Fig. 9). Temperatures of structed for the respective highest- and lowest-density
eutectic melting are, although not generally observable, inclusions of each principal ¯uid inclusion type de®ned
in the range between ±17 and ±29 °C, indicatingthe
presence of NaCl and KCl. The absence of clathrate clusions in kyanite %type 1) and the CO inclusions in
above. The isochores of both the low-salinity H O in-
2
2
meltingprecludes the presence of si gn i®cant amounts of
CO . Homogenisation temperatures %homogenisation retrograde P±T path during greenschist facies conditions
recrystallised quartz %type 2) overlap with the inferred
2
into the liquid phase) are in the range between 120 and in a relatively late exhumation stage of orogenic evolu-
tion %cf. Fig. 12). The calculated densities for these two
200 °C with a maximum at 160‹20 °C, resultingin

3
37
such as kyanite + H O ‹ diaspore + quartz would
2
have caused a density change in the low-salinity ¯uid.
Salinities of the inclusion ¯uid would have been in-
creased through such a back reaction. Solid inclusions
alongthe inclusion walls have not been found by Raman
analysis or petrographic means. The nature of very ®ne-
grained %<1 lm) solids can only be identi®ed by SEM
methods %Heinrich and Gottschalk 1995). Based on the
textural relationships of the H O inclusions in kyanite
2
%
type 1), which clearly reveal primary features, initial
trappingof this inclusion type can be related to the post
D1 stage of extensional high-pressure veining close to
peak-metamorphic conditions. Fluid compositions of
the H O inclusions in kyanite would, therefore, re¯ect
2
the ¯uid regime during the prograde and peak-meta-
morphic evolutionary stages. These observations, as well
as the presence of H O-bearingminerals such as zoisite,
2
paragonite and phengite in the veins and the eclogite,
inclusions re¯ect elevated water fugacity in these rocks during the
HP event %see also Holland 1979; Jamtveit et al. 1990;
Brunsmann et al. 2000).
Fig. 11 Relationship between size and density of CO
type 2) in vein quartz
2
%
The initial trappingconditions of the texturally
pseudosecondary CO ¯uid inclusions %type 2) within
inclusion types are too low to relate them to either the
static symplectite growth or the subsequent syn-tectonic
2
vein quartz are dicult to assess. These ¯uids may have
been derived from decarbonation of marbles and calc-
silicate rocks present in the tectono-metamorphic units
%
D2) amphibolite facies metamorphism. However, it
should be emphasised that back-reactions between the
inclusion ¯uid and the kyanite host may have occurred.
Back-reactions between inclusion ¯uids and mineral
hosts other than quartz are a common process %Heinrich
and Gottschalk 1995). A possible retrograde reaction
hosting the eclogites and high-pressure veins. The CO2
phase contains no other gas species such as N , which
2
are usually present within high-density CO -rich ¯uid
2
inclusions in eclogite terranes %Klemd et al. 1992; An-
dersen et al. 1993; Klemd et al. 1995). The N trapped in
such prograde to peak-metamorphic high-density ¯uid
inclusions can be derived from reactions involvingthe
2
Fig. 12 Position of the isochores of the di€erent ¯uid inclusion
types in the P±T space and schematic P±T path %black) of the Dabie
Shan high-pressure rocks

338
breakdown of micas and/or K-feldspar %Andersen et al. 1995). Such a correlation, however, is not observed in
993). The absence of signi®cant amounts of N in the the investigated samples. However, even small inclusions
1
2
CO ¯uid inclusions trapped in recrystallised quartz may %<10 lm) will re-equilibrate as a result of large pressure
2
be indicative of an e€ective dilution of the N generated gradients during exhumation. Furthermore, preferential
2
duringthe pro gr ade to peak-metamorphic evolution. loss of H O duringre-equilibration of the smaller in-
2
The prograde dehydration reactions generated a ¯uid clusions as described by Hall and Sterner %1993) cannot
phase dominated by H O, as demonstrated by the ¯uid be proven, as there is no correlation between inclusion
2
inclusions trapped in kyanite %type 1) close to the peak size and salinity. The fact that some inclusions of both
metamorphic conditions, whereas CO is entirely absent inclusion types seem to have decrepitated or re-equili-
2
within these inclusions. In contrast, no texturally pri- brated duringpost-amphibolite facies conditions, is
mary or pseudosecondary CO ±H O±NaCl or H O± probably the result of individual ¯uid density geometry
2
2
2
NaCl ¯uid inclusions could be found in the recrystallised as well as size of inclusion.
vein quartz. The absence of H O in the CO ¯uid in- The secondary H O inclusions %type 3) trapped along
2
2
2
clusions trapped in quartz can be explained by two en- healed fracture zones in vein quartz can be related to the
tirely di€erent processes. Either the ¯uid phase that was latest evolutionary stage of the high-pressure veins.
trapped in these inclusions had a very low H O activity Fracturingand in®ltration by these low-salinity H O
2
2
or the composition of these inclusions was changed ¯uids took place under dominantly brittle conditions
through post-entrapment modi®cations. It is well known subsequent to the latest recrystallisation of vein quartz.
that CO ±H O inclusions are subjected to preferential As minimum temperatures for the recrystallisation of
2
2
H O loss duringrecrystallisation of the host quartz
2
quartz are usually assumed to be in the range of 270±
crystals, thereby producingCO 2 ¯uid inclusions %e.g. 300 °C %Voll 1969), this ¯uid phase clearly penetrated the
Hollister 1990; Bakker and Jansen 1991; Johnson and rock at the last stage of orogenic evolution.
Hollister 1995). No evidence was observed to suspect
¯
uid±¯uid-immiscibility usingthe criteria of Ramboz
et al. %1982). Alternatively, enhanced regional devola- U±Pb systematics of rutile
tilisation through veins and fractures during conditions
of initial extension followingthe peak metamorphic
The three analysed samples of specimen HVP1 are small
stage could have most e€ectively led to the expulsion of fragments from a 4-cm-long single rutile crystal. They
the prograde H O±NaCl ¯uids and ®nally, resulted in a are characterised by little radiogenic lead, and very low
2
relative ¯uid de®ciency. Increased decarbonation of uranium contents %<0.1 ppm; Table 6). The isotopic
marbles and calc-silicate rocks duringthe initial iso-
thermal decompression stage of the P±T-evolution attributed to in-situ lead growth %Fig. 13) and patterns
variability amongthem is, however, too lar eg to be
2
06
204
207
204
206
204
208
would have also generated a relatively pure CO ¯uid on the
phase to be trapped within the recrystallised quartz.
Pb/ Pb- Pb/ Pb and
Pb are incompatible with secondary addition or re-
Alternatively, such CO ¯uids may have been supplied moval of lead. Note that, although blank lead %15‹8 pg)
Pb/ Pb± Pb/
2
2
04
2
from an external reservoir.
The CO ¯uids within vein quartz were most proba- %170±180 pg), the observed variation in the isotopic
contributes signi®cantly to total lead in samples 1±3
2
bly trapped duringthe isothermal decompression stage
followingpost D1 extensional hi gh -pressure veiningand
composition is not compatible with incorrect estimation
of blank level or composition. Thus, the initial lead
composition of rutile from specimen HVP1 was isotop-
®
nally resultingin static symplectite gr owth and syn-
tectonic %D2) amphibolite facies metamorphism. Strong ically heterogeneous, which implies that the lead in the
recrystallisation of the vein quartz duringthe P±T-evo- ¯uid was isotopically variable duringrutile crystallisa-
lution followingamphibolite facies metamorphism re-
sulted in re-equilibration of the CO ¯uid inclusions. of the in¯owing¯uid or may re¯ect local wallrock±¯uid
tion. This heterogeneity may re¯ect a genuine character
2
The calculated densities re¯ect the ®nal stage of re- interaction, notably leachingor dissolution of the wall-
equilibration during rg eenschist facies conditions, as
rock %e.g. Banks et al. 1991; Romer and Heinrich 1998).
densities of inclusions that were trapped or retrapped The low uranium contents in this specimen re¯ect low
duringrecrystallisation of quartz will monitor those P±T availability of uranium in the ¯uid or the preferential
conditions %e.g. Johnson and Hollister 1995).
The H O inclusions in kyanite as well as the CO
partitioningof uranium into the ¯uid.
Rutile in specimen HPV2b occurs in a medium-
2
2
inclusions in recrystallised quartz were extensively re- grained zone in the vein. Three analysed samples show
equilibrated duringthe relatively late gr eenschist facies variably radiogenic lead. Two samples have uranium
P±T-conditions %see Fig. 12). Such a process explains the contents at ꢀ3 ppm and low lead contents. Their lead
variation in densities for both types of inclusions. isotopic composition is distinctly radiogenic. The third
Modi®ed ¯uid inclusions of the same shape should re- sample has little uranium and a high content of lead. Its
veal a correlation between Th and inclusion size because lead isotopic composition, thus, still closely resembles
larger inclusions, as a result of pressure di€erences, re- the initial composition acquired at the time of rutile
equilibrate before smaller inclusions duringexhumation crystallisation. Usingthis isotopic composition initially
2
06
238
%
e.g. Bodnar et al. 1989; Hall et al. 1991; Klemd et al. for the other two analysed samples results in Pb/
U

3
39
Table 6 Pb±Pb and U±Pb analytical data of rutile from high-pressure veins
a
206Pb
Pb
206Pb
Pb
207Pb
Pb
208Pb
Pb
238U
Pb
Sample
Weight %mg) Concentra-
tions %ppm)
U
Pbtot
204
204
204
204
204
c
c
c
d
Measured
ratios
corr.
corr.
corr.
corr.
b
Rutile
Sample HPV1
1
2
3
0.319
0.347
0.638
ꢁ0.1
1.11
0.46
0.23
18.121
19.086
19.541
17.87
18.70
19.05
15.37
15.66
15.33
37.18
38.67
37.63
±
±
±
<0.07
<0.08
Sample HPV2b
e,f
4
5
6
0.587
0.509
0.749
3.19
3.28
<0.16
0.24
0.35
6.38
57.248
43.615
19.409
6329
46.09
19.43
17.60
16.85
15.69
37ꢀ52
1,360
820
±
e,f
38.41
38.60
a
205
206
Pb/ Pb=21.693 for isotopic tracer composi-
Rutile concentrates were obtained usingstandard mineral-sepa-
spectively, and
ration procedures and were puri®ed through separation by hand tion. Data reduction was performed by Monte Carlo modellingof
under the binocular. Care was taken to use only fracture-free and 1,000 random normally distributed data sets that ®t above un-
inclusion free clear crystals or crystal fragments. Pb± U mixed certainty limits, allowingfor error correlation when appropriate
2
05
235
d
tracer was added before sample dissolution. All samples were dis-
Lead corrected for fractionation, 15 pglead blank, tracer con-
solved with 52ꢀ Hf in autoclaves at 220 °C for 4 days, dried, and tribution, and initial lead. Uranium corrected for fractionation and
transferred overnight into chloride-form using 6 N HCl. Lead and 1 pguranium blank
e
uranium were separated using ion-exchange chromatography as
described by Tilton %1973) and Manhe
uranium were loaded on separate single Rc-®laments using silica- the e€ects of initial isotopic heterogeneity and do not necessarily
The Pb/U ratios are very dependent on the common lead corec-
Â
s et al. %1978). Lead and tion. The here calculated Pb/U and age values are for illustration of
2
06
204
gel emitter and H
3
PO
4
%Gerstenberger and Haase 1997) and ana- represent the preferred values. For common lead
Pb/ Pb=
207
235
208
204
lysed at 1,200±1,260 °C and 1,350±1,400 °C, respectively, on a 17.5‹0.1, Pb/ Pb=15.4‹0.05, Pb/ Pb=37.0‹0.2 %Fig. 13),
2
06
238
207
235
Finnigan MAT262 multicollector mass-spectrometer using ion Pb/U ratios are Pb/ U=0.0337 and 235Pb/ U= 0.223 %ana-
2
lysis 4) and Pb/ U=0.0349 and Pb/ U=0.244 %analysis 5).
06
238
207
counting
Lead isotope ratios corrected for fractionation with 0.1ꢀ/a.m.u. Usingthe constants of Ja€ey et al. %1971) recommended by IUGS
b
c
206
238
Pb/ U ages are
Lead corrected for fractionation, 15 pglead blank, and tracer
contribution. Because of very low total lead contents, blank and 214 Ma %analysis 4) and 221 Ma %analysis 5)
%Steiger and Ja
È
ger 1977), the corresponding
tracer lead represent signi®cant contributions. Uncertainties of 0.3±
.4ꢀ for isotope ratios were calculated takinginto consideration
f
206
204
Pb/ Pb=19.0‹0.1,
207
204
Pb/ Pb
As in e with common lead
=15.65‹0.05, Pb/ Pb=38.3‹0.2 %Fig. 13), the following va-
0
208
204
the followinguncertainties: measurement errors, 30ꢀ for frac-
tionation correction, 50ꢀ for blank level, 0.1, 0.05, 0.2 for
207
235
lues are obtained: analysis 4: 206Pb/238U=0.0326,
=0.198 and 207Pb/235Uage=207 Ma; and analysis 5:
=
Pb/
Pb/
U
U
206
238
206
238
2
06
204
Pb/ Pb,
207
204
Pb/ Pb, and
208 204
Pb/ Pb blank composition, re-
206 238
0.0331 and Pb/ U=0.202 and Pb/
Uage=210 Ma
2
and Pb/ U and indicates that %1) the plot to the left increase with time, such a pattern would suggest either
07
235
207
204
of the concordia, i.e. the Pb/ Pb used for the com- that uranium and thorium contents in rutile anticorrelate
mon lead correction was too high, and %2) that they or that the initial lead isotopic composition was hetero-
206
238
di€er slightly in their apparent Pb/ U age from 207± geneous. As rutile is not known to incorporate signi®cant
10 Ma %Table 6, see footnote). Given the initial lead amounts of thorium %cf. Mezger et al. 1989; Brenan et al.
2
isotopic heterogeneity of rutile in specimen HPV1, there 1994), and thorium might have been little available be-
is little reason to assume that the initial lead isotopic cause of its limited mobility in aqueous ¯uids, a hetero-
composition of rutile in specimen HVP2b would be geneous initial lead isotopic composition seems to be the
homogeneous. Thus, the rutile data were also calculated more likely explanation. %3) The slope of the regression
2
06
204
238
204
usingthe least radio eg nic lead isotopic composition
observed in the veins. Then, the two radiogenic rutile
samples yield apparent
214 Ma %Table 6) respectively.
line in the Pb/ Pb± U/ Pb diagram is too ¯at to
2
be in agreement with the Pb/ U data %not shown).
06
238
206
238
Pb/ U ages at 221 and
Constraints on the age of rutile. Because of the hetero-
Additional arguments for a heterogeneous initial geneous nature of the initial lead isotopic composition, it
lead isotopic composition in rutile specimen HVP2b is not possible to determine a precise age for rutile
206
204
207
204
208
204
come from the
Pb/ Pb±2 Pb/ Pb± Pb/ Pb crystallisation. It should be noted that the range in ap-
06 204 207 204
206 238
systematics %Table 6). %1) In the Pb/ Pb± Pb/ Pb parent Pb/ U ages from 207 to 221 Ma %1) is entirely
diagram, the three HVP2b samples fall on a line with a caused by the uncertainty in the true common lead
Pb/ Pb slope at 0.0435 %Fig. 13A), which is ¯atter composition and %2) represents a minimum range, as it is
2
07
206
than that for `instantaneous lead formation at present' possible that the initial lead of the dated fragments has
and which would correspond to a negative age if the lead even a less radiogenic lead isotopic composition than
line is interpreted as secondary lead isochron. %2) In the observed for the low-U fragments %Table 6). In this case,
2
06
204
208
204
206
238
Pb/ Pb± Pb/ Pb diagram %Fig. 13B), the three the apparent
HVP2b samples scatter about a line with negative slope. pendence of the rutile age on the common lead com-
Pb/ U age would be older. The de-
2
06
204
Pb/ Pb and
208
204
Pb/ Pb ratios only can position is also illustrated in Fig. 13C using samples 4
Because

340

3
41
b
pic closure %Cherniak 2000; ꢀ600 °C for 100-lm large
Fig. 13 Pb±Pb diagrams for rutile samples from Dabie Shan high- grains) and, furthermore, the large size of the analysed
pressure veins. Sample numbers as in Table 6. Evolution curves for
average lead reservoirs %Zartman and Doe 1981) are for reference
rutile crystal, and the preservation of the initial isotopic
heterogeneity. If there had been signi®cant mobility of
Pb after crystallisation, initial isotopic heterogeneities
2
06
204
207
204
only. A
Pb/ Pb± Pb/ Pb diagram. Note excess scatter at
204
2
06
low Pb/ Pb and the slope of the regression line, which is ¯atter
than the reference lines for
0
and 220 Ma, respectively. probably would have been largely erased.
2
06
204
208
204
B
Pb/ Pb± Pb/ Pb diagram. Note the large variation of
204
Pb/ Pb, which for a low-Th phase is not reconcilable with in-
The demonstrated initial isotopic heterogeneity of
lead in rutile and the general sensitivity of age determi-
nations of minerals with low Pb/ Pb to the correc-
2
08
2
06
204
207
204
situ lead growth. C Pb/ Pb± Pb/ Pb diagram, %enlargement
from A). Shaded arrays for samples 4 and 5 are calculated usingthe
2
06
204
measured lead isotopic composition, the Pb and U contents, and tion of initial lead have direct implications for U±Pb
2
06
204
the ages indicated by the short vertical lines and assumingthat rutile
represents a closed system for U and Pb, i.e. %1) any assumption of
initial lead fallingon a shaded array would yield a concordant age
chronology of low- Pb/ Pb minerals from meta-
morphic rocks and ¯uid-dominated systems. %1) Scatter
in age may not re¯ect secondary disturbance, slow
2
06
238
in the concordia for that sample, and %2) the Pb/ U age is given
by the short vertical line. Note that the arrays for samples 4 and 5 cooling, or formation over long time. Instead, it may
do not intersect each other, and thus, samples 4 and 5 cannot have re¯ect heterogeneous initial lead isotopic compositions
had a common initial lead isotopic composition. Samples 1, 2, 3,
%
È
for a detailed discussion see also Romer and Rotzler
and 6 are shown as ellipses; the recalculated lead isotopic
composition for 220 Ma is shown as dashed ellipses. Asterisks 1
and 2 refer to initial lead used to calculate data in Table 6. For lead
2001). %2) In cases where the initial lead isotopic com-
position cannot be measured %phases consumed, ¯uid
of the lead composition shown by asterisk 1, all isotope data will gone), the choice of initial composition will be arbitrary.
fall to the left of the concordia. For lead of the lead composition
shown by asterisk 2, sample 5 will yield concordant data and
other chronometers may result in incorrect inferences
sample 4 still would plot to the left of the concordia. The lead
Comparison of such ages with those obtained from
about the conditions of isotopic closure. The perfect ®t
growth curves for average lead reservoirs %Zartman and Doe 1981)
are shown for orientation. The mixingarray % dotted) is shown for of the data along a straight line %Fig. 13A) has geo-
illustration and does not have age signi®cance. It is ®tted %1) to
encompass lead that is less radiogenic than the measured isotopic
compositions and %2) to intercept the arrays of concordant data of
samples 4 and 5, respectively, at the same age
chronologically no signi®cance, which is obvious from
the `future' age. Furthermore, it implies that the initial
238
204
lead isotopic composition of a low- U/ Pb mineral
not necessarily is identical with the one of cogenetic
minerals with higher
2
38
204
U/ Pb %and thus after some
and 5 from Table 6. The shaded area for each sample time show a more radiogenic lead isotopic composi-
represents those initial lead isotopic compositions that tions) and, therefore, in some cases may be unsuitable
yield concordant data in the concordia diagram, whereas to constrain the slope of lead lines in the
2
06
204
Pb/ Pb-
2
07
204
Pb/ Pb and Pb/ Pb and U/ Pb diagrams.
206
204
238
204
the numbers refer to the age obtained. Note that con-
cordant data are obtained for contrastinginitial lead in
accordance with the anomalous ¯at slope in the
Source of Pb and ¯uid. The initial lead isotopic com-
position represents a ®ngerprint for the lead source, and
possibly indirectly also the ¯uid source. The two samples
2
06
204
207
204
Pb/ Pb± Pb/ Pb diagram %Fig. 13A) and the
heterogeneous isotopic composition of the low-U rutile
samples, which is given for orientation %Fig. 13C). The
dotted array at ꢀ220 Ma %Fig. 13C) is here considered to
be the best estimate for the U±Pb age of the rutile
crystals and the formation of these veins because this
array yields a coherent age for the two fragments with
radiogenic lead %samples 4 and 5) and encompasses also
the least radiogenic lead observed. This estimate coin-
cides well with the age of collision and UHP metamor-
2
07
204
with the lowest
Table 6) also show the lowest Pb/ Pb values. Sample
%Table 6) could have the same initial lead signature, as
Pb/ Pb values %samples 1 and 3 in
2
08
204
4
2
08
204
indicated by low Pb/ Pb %Fig. 13B, C) and its curve
of possible initial lead isotope composition, if rutile was
a closed system after crystallisation. Samples 2 and 6
2
07
204
%Table 6) in comparison have high
208
Pb/ Pb and
204
Pb/ Pb values. Initial lead of sample 5 %Table 6) is
phism %Ames et al. 1993, 1996; Li et al. 1993; Xiao et al. intermediate between these two groups of samples in
1995; Rowley et al. 1997; Hacker et al. 1998). The lead both lead diagrams %Fig. 13B, C).
array %dotted, Fig. 13C) by no means includes anoma-
2
06 204
Pb/ Pb ratios seem to be characteristic for
Low
lously non-radiogenic lead, as shown by the range of the entire Dabie Shan region. For instance, leached K-
initial lead isotopic compositions encompassed by lea- feldspar from 125±140 Ma old post-metamorphic
ched K-feldspar from ꢀ130 Ma old granitoids from the granitoid intrusions of the northern Dabie Shan range in
2
06
204
north Dabie Shan %NDG, Fig. 13C). Although the their Pb/ Pb ratios from 15.8±16.7 %Romer, unpub-
2
06
204
temperature of isotopic closure of rutile is debated %e.g. lished data). Such low
Mezger et al. 1989; Santos Zalduegui et al. 1996), the lead is derived from old rocks or minerals that are
Pb/ Pb ratios indicate that
2
38
204
U±Pb systematics of rutile is here interpreted to be re- characterised by low U/ Pb and, therefore, had ex-
lated with the formation of rutile rather than with some perienced only little lead growth after their formation.
moment duringthe coolinghistory of the rock. Such an
2
38
Low U/ Pb are found in rocks that had lost signi®-
204
interpretation is in agreement with the recent experi- cant portions of uranium during high-grade metamor-
mental determination of lead di€usion in rutile, which phism, and in feldspar-dominated magmatic and
corresponds to a distinctly higher temperature of isoto- sedimentary rocks because feldspar strongly favours the

342
incorporation of lead over uranium. During high-grade near-isothermal decompression. Another potential
metamorphism, not only uranium, but to a lesser extent source of ¯uids is the basement of the Dabie Shan area,
also lead, is lost. Hence, lower-crustal rocks may have which also underwent dehydration in the course of the
2
32
204
238
204
enhanced
trastingTh/Pb and U/Pb ± with increasingtime ± result
Th/ Pb and low
U/ Pb. These con- Triassic metamorphic overprint. Duringsubsequent ex-
humation, ¯uids penetrated the rock and led to the
in distinct lead isotopic compositions. For instance, old formation of quartz veins on fractures. It is, however,
lower crustal rocks are characterised by rather high unclear to what extent early produced ¯uids still were
Pb/ Pb at low Pb/ Pb values, whereas old feld- available duringthe exhumation of the HP and UHP
2
08
204
206
204
2
08
204
spar-dominated rocks commonly have low
and low
Pb/ Pb units %gneisses and metabasites) of Dabie Shan. An in-
Pb/ Pb for dication for the origin of the ¯uid phase is given by the
2
06
204
Pb/ Pb values. The low
206
204
initial lead of these intrusions, as well as the initial lead of vein minerals ankerite, zoisite, paragonite and ompha-
the vein-bound rutile, indicates signi®cant amounts of cite, which display similar composition to those forming
concealed old basement in the Dabie Shan region.
The initial isotopic composition of rutile %projected chemistry of the vein-®llingwas rock-controlled and that
back alonglead gr owth-lines for 220 Ma) encompasses the ¯uid phase was ± at least in part ± locally derived
their host eclogites. This correlation shows that the
the entire range of isotopic composition characteristic %e.g. Scambelluri et al. 1998). This is also supported by
for old lower crust, or possibly old carbonates or old the observed heterogeneous population of ¯uid inclu-
feldspar-dominated rocks %Fig. 13C), and thus may sions ranging from low salinity brines to CO -rich
2
originate from such sources %see reference curves in inclusions. Accordingto Giaramita and Sorensen
Fig. 13C). The initial lead isotopic heterogeneity of rutile %1994), this favours a local source for the ¯uid, which
demonstrates that the ¯uid was temporally heteroge- was generated during progressive metamorphism. Fur-
neous in its lead isotopic composition and was derived thermore, this is in agreement with the ®ndings of Xiao
from contrastingsources %la re g variation in initial
Pb/ Pb).
et al. %2000) who suggested that the preservation of
primary high salinity inclusions in omphacite and kya-
nite indicate a relatively closed ¯uid system and a limited
2
08
204
¯
uid ¯ow duringUHP-metamorphism in the Bixiling
eclogites. Their model is in agreement with previous
uids inclusion studies, which indicate that large ¯uid
Fluid evolution along the P±T path
¯
Petrographic and petrologic features suggest a subduc- activity gradients could persist during eclogite facies
tion-related, prograde metamorphic path for the inves- conditions as a consequence of local ¯uid production
tigated rocks, passing through the stability ®elds of and/or bu€eringas well as limited ¯uid ¯ow at depth
lawsonite and glaucophane %cf. Fig. 6; see also Castelli >60 km in a subduction zone %e.g. Philippot and Sel-
et al. 1998). Thermobarometric data demonstrate peak verstone 1991; Klemd et al. 1992).
conditions for the HP eclogite-facies at 19±21 kbar and
Beside hydraulic fracturing, the formation of frac-
5
70±620 °C. The HP metamorphism was accompanied tures may, however, also be induced by volume reduc-
by a penetrative D1 deformation, which was well pre- tion duringeclo gi tisation %Austrheim 1998) or later by
served in the pristine eclogites. Although the exact time hydration and volume increase duringback-reaction of
of formation of quartz vein HPV1 remains unclear, eclogites to amphibolites during exhumation %e.g. Hein-
crosscuttingrelations de®nitely show that the formation rich 1982). The behaviour of all ¯uids generated during
of HPV2 post-dates the eclogite-facies D1 deformation. the metamorphic path of the HP and UHP rocks follows
Mineral assemblages within quartz vein HPV2, espe- the general sequence of dehydration with build-up of an
cially the growth of garnet and omphacite, highlight its overpressure and eventual fracturingof the rocks,
formation under eclogite facies conditions. This is also channelled escape of the ¯uids alongthese fractures to-
con®rmed by calculations of phase equilibrium curves wards lower pressure %i.e. generally lower levels). In this
on the vein assemblage quartz±kyanite±talc±zoisite± general sequence of processes, which occur at di€erent
garnet±aragonite using the TWEEQ-program of Ber- depth, the ¯uid passes through the rocks as if they were a
man %1991), which yield an intersection at 19.5 kbar and chromatographic column, exchanging with the rocks,
5
91 °C in the presence of an aqueous ¯uid phase. These and thus, changing them and becoming changed itself.
observations demonstrate that both the peak-metamor- Therefore, the ¯uid does not necessarily re¯ect its source,
phism of the host-rocks and extensional fracturingwith but may instead re¯ect the last lithological unit with
subsequent growth of hydrous vein-assemblages oc- which it had equilibrated. Such a local bu€eringof the
curred under similar eclogite-facies conditions. Fur- ¯uid has been demonstrated for ¯uid inclusion isotopes
thermore, it testi®es that a free ¯uid phase may well exist from low-grade rocks in the Pyrenees %Banks et al. 1991)
in deeply subducted crust %cf. Austrheim 1998). From and for Sr and Pb isotopes from wollastonite-mantled
this the question of the source of this ¯uid phase arises. chert layers in marbles %Romer and Heinrich 1998). The
Considerable amounts of ¯uids were released from studied rocks from HPV1 and HPV2 also show this
subducted metasediments and metavolcanics %cf. Fig. 1) general sequence, although the various stages were re-
duringthe transition of blueschist-facies rocks to
corded at di€erent times as the samples moved through
eclogite-facies rocks %e.g. Ridley 1984) and later during this chromatographic column. Our results concur with

3
43
those from ¯uid inclusion studies on UHP rocks from the
Bixilingcomplex in Dabie Shan, by Xiao et al. %2000).
These investigators showed that the ¯uid phase
underwent distinct changes in composition during the
metamorphic evolution and ± to a great extent ±
re-equilibrated duringthe last sta ge s of the metamor-
phism. Especially the in®ltration of a low salinity aque-
ous phase on the retrograde metamorphic path is in
concurrence with our observations from the HP rocks.
Evidence for the character of the ¯uid phase during
1. 5 Aragonite+3 a-quartz+2 zoisite=3 grossular+
H O+5 CO
2. 4 Aragonite+3 kyanite+3 a-quartz+H O=2 zoi-
2
2
2
site+4 CO2
3. 9 a-Quartz+3 pyrope+4 aragonite+4 H O=3
talc+2 zoisite+4 CO2
2
4. 3 Aragonite+kyanite+2 a-quartz=grossular+3
CO2
5. 3 Kyanite+3 grossular+2 H O+CO2=arago-
2
nite+4 zoisite
6. a-Quartz+5 kyanite+4 grossular+3 H O=6 zoi-
the HP event is given by primary, two-phase H O in-
2
2
clusions trapped in vein kyanite. The main source for
this hydrous ¯uid phase could be dehydration reactions
that occurred in the course of the eclogite facies meta-
morphism in the ma®c host rock and neighbouring
metasediments or from destabilisation of hydrous
phases duringdecompression. This conclusion is in ac-
cordance with the observation that eclogite-facies
site
7. 8 Aragonite+pyrope+7 a-quartz+2 zoisite=
talc+ 4 grossular+8 CO2
8. 11 Aragonite+3 talc+8 zoisite=3 pyrope+9
grossular+7 H O+11 CO
2
2
9. 4 Grossular+5 pyrope+11 a-quartz+8 H O=6
2
zoisite+5 talc
metamorphism is most typically accompanied by a ¯uid 10. 3 Aragonite+pyrope+4 a-quartz+H O=talc‡
2
regime dominated by H O but, however, in places also
2
grossular+3 CO2
by N ‹CO %Andersen et al. 1989; Klemd 1989; Touret 11. Talc+a-quartz+4 kyanite+4 aragonite=pyrope‡
2
2
1
992; Andersen et al. 1993; Klemd and Bro
The initial lead of rutile, however, demonstrates that the 12. 8 Aragonite+9 kyanite+3 talc=4 zoisite+3 py-
ma®c host rocks and the associated metasediments were rope+H O+8 CO
È
cker 1999).
2 zoisite+4 CO
2
2
2
not the only source of the ¯uid, but part of the ¯uid 13. Pyrope+2 a-quartz+H O=talc+kyanite
2
phase may have been released from old basement rocks 14. 5 Aragonite+grossular+7 kyanite+2 talc=4 zoi-
of the Dabie Shan.
The eclogite-facies stage was followed by a static 15. 3 Talc+8 kyanite+4 grossular=3 pyrope+5 a-
growth of symplectites, and in turn followed by a syn- quartz+6 zoisite
tectonic %D2) amphibolite-facies metamorphism. In the 16. Talc+11 kyanite+8 grossular+5 H O=pyrope+
site+2 pyrope+5 CO2
2
course of this retrogression, a CO -rich ¯uid penetrated
12 zoisite
the rocks. Large amounts of the eclogite facies rocks 17. Talc+2 kyanite+3 aragonite=grossular+py-
2
were totally retrogressed and deformation a€ected even
part of the high-pressure quartz veins. The last overprint
took place under static greenschist facies conditions and
rope+3 CO2+H O
2
These calculations were performed with the TWEEQ
led to a re-equilibration of all captured ¯uid inclusions program of Berman %1991) using the thermodynamic
up to this stage. Furthermore, low-salinity H O ¯uids dataset BA96.dat %see also Berman 1988). Aragonite
2
penetrated the rocks via fracture zones. The low amount data have been extracted from the VERTEX %Conolly
of this ¯uid phase and the lack of any deformation 1990) database sup92ver.dat. The followingactivity
during gr eenschist facies metamorphism are responsible
for the modest extent of this overprint.
models were applied: grossular, pyrope: Berman %1990);
jadeite: Holland %1980, 1990); zoisite, diopside, talc,
paragonite: ideal; ¯uid phase %estimated): water: 0.88;
Acknowledgements Dieter Rhede, Oona Appelt and Klaus Becker carbon dioxide: 0.12; kyanite, aragonite, quartz, coesite:
are thanked for assistance with the microprobe analyses at the GFZ
Potsdam and at the University of Freiberg. Klaus-Peter Kelber at
a=1.
È
University of Wurzburgkindly provided the black-and-white
photographs of ¯uid inclusions. We are grateful to Lothar Rats-
chbacher and Jens C. Schmid who supplied structural data and
eclogite samples from location HPV2 and we enjoyed the company
and helpful comments of Dr X. Liu and Mr H. Xue duringour
References
Ague JJ %1997) Crustal mass transfer and index mineral growth
in Barrow's garnet zone, northeast Scotland. Geology 25:
®
reviews by Hakon Austrheim and Bruce Yardley. Fundingwas
eldwork in China. The paper pro®ted considerably from thorough
Ê
7
3±76
AIRGS %Anhui Institute of Regional Geological Survey) %1997) 1:
50,000 Geological map of Dabie mountains area of Anhui
provided by the grants OB 80/17 and KL692/9 from the Deutsche
Forschungsgemeinschaft %DFG).
2
Province. Anhui Institute of Regional Geological Survey
Ames L, Tilton GR, Zhou G %1993) Timingof collision of the Sino-
Korean and the Yangtse cratons: U±Pb zircon dating of coesite-
bearing eclogites. Geology 21:339±342
Appendix
Ames L, Zhou G, Xiong B %1996) Geochronology and geochem-
istry of ultrahigh-pressure metamorphism with implications for
collision of the Sino-Korean and the Yangtze cratons, central
China. Tectonics 15:472±489
Compilation of the metamorphic mineral reactions for
the invariant point of the high-pressure mineral assem-
blage in HPV2b.

3
44
Andersen T, Burke EAJ, Austrheim H %1989) Nitrogen-bearing, Connolly JAD %1990) Multivariable phase diagrams: an algo-
aqueous ¯uid inclusions in some eclogites from the Western
Gneiss Region of the Norwegian Caledonides. Contrib Mineral
Petrol 103:153±156
rithm based on generalized thermodynamics. Am J Sci 290:
666±718
Crawford ML %1981) Phase equilibria in aqueous ¯uid inclusions.
In: Hollister LS, Crawford ML %eds) Fluid inclusions: appli-
cations to petrology, vol 6. Short course handbook. Mineral
Association of Canada, Calgary, pp 75±100
Andersen T, Burke EAJ, Austrheim H, Elvevold S %1993) N
CO in deep crustal ¯uids: evidence from the Caledonides of
Norway. Chem Geol 103:153±165
2
and
2
Austrheim H %1987) Eclogitization of lower crustal granulites by Dickenson MP, Hewitt D %1986) A garnet±chlorite geothermome-
ter. Geol Soc Am Abstr Prog18:584
Droop GTB %1987) A general equation for estimating Fe
¯
8
uid migration through shear zones. Earth Planet Sci Lett
1:221±232
3
+
con-
Austrheim H %1990) The granulite±eclogite facies transition: a
comparison of experimental work and a natural occurrence in
the Bergen Arcs, western Norway. Lithos 25:163±169
centrations in ferromagnesian silicates and oxides from mi-
croprobe analyses, usingstoichiometric criteria. Mineral Mag
51:431±435
Austrheim H %1998) In¯uence of ¯uid and deformation on meta- Ellis DH, Green EH %1979) An experimental study of the e€ect of
morphism of the deep crust and consequences for the geody- Ca upon garnet±clinopyroxene Fe±Mg exchange equilibria.
namics of collision zones. In: Hacker BR, Liou JG %eds) When Contrib Mineral Petrol 71:13±22
continents collide: geodynamics and geochemistry of ultrahigh- Etheridge MA %1983) Di€erential stress magnitudes during regional
pressure rocks. Kluwer, Dordrecht, pp 297±323
Bakker RJ, Jansen JBH %1991) Experimental post-entrapment
deformation and metamorphism: upper bound imposed by
tensile fracturing. Geology 11:231±234
2 2
water loss from synthetic CO ±H O inclusions in natural Gerstenberger H, Haase G %1997) A highly e€ective emitter sub-
quartz. Geochim Cosmochim Acta 55:2215±2230
Banks DA, Davies GR, Yardley BWD, McCaigAM, Grant NT
stance for mass spectrometric Pb isotope ratio determinations.
Chem Geol 136:309±312
1991) The chemistry of brines from an Alpine thrust system in Ghent ED, Stout MZ, Black PM, Brothers RN %1987) Chloritoid
%
the central Pyrenees; an application of ¯uid inclusion analysis
to the study of ¯uid behaviour in orogenesis. Geochim Cos-
mochim Acta 55:1021±1030
bearingrocks with blueschists and eclo igt es, northern New
Caledonia. J Metamorph Geol 5:239±254
Giaramita MJ, Sorensen SS %1994) Primary ¯uids in low-temper-
ature eclogites: evidence from two subduction complexes %Do-
minican Republic and California, USA). Contrib Mineral
Petrol 117:279±292
Banno S %1959) Aegirinaugites from crystalline schists in Sikoku.
J Geol Soc Jpn 65:652±657
Becker H, Jochum KP, Carlson RW %1999) Constraints from high-
pressure veins in eclogites on the composition of hydrous ¯uids Graham CM, Powell R %1986) A garnet±hornblende geother-
mometer: calibration, testing, and application to the Pelona
schists, southern California. J Metamorph Geol 2:13±31
in subduction zones. Chem Geol 160:291±308
Berman RG %1988) Internally-consistent thermodynamic data for
stoichiometric minerals in the system Na
FeO±Fe ±Al ±SiO ±TiO ±H O±CO
Berman RG %1990) Mixing properties of Ca±Mg±Fe±Mn garnets.
Am Mineral 75:328±344
Berman RG %1991) Thermobarometry usingmultiequilibrium cal-
culations: a new technique with petrologic applications. Can
Mineral 29:833±855
2
O±K
2
O±CaO±MgO± Green TH, Hellman PL %1988) Fe±Mg-partitioning between coex-
2
O
3
2
O
3
2
2
2
2
. J Petrol 29:445±522
isting garnet and phengite at high-pressure, and comments on a
garnet±phengite geothermometer. Lithos 15:253±266
Hacker BR, WangX, Eide EA, Ratschbacher L %1996) Qinling±
Dabie ultrahigh-pressure collisional orogen: a critical review of
existing data and suggestions for future research. In: Yin A,
Harrison TM %eds) The tectonic evolution of Asia, Rubey
volume. Cambridge Univ Press, Cambridge, pp 345±370
Bodnar RJ, Binns PR, Hall DL %1989) Synthetic ¯uid inclusions.
VI. Quantitative evaluation of the description behaviour of Hacker BR, Ratschbacher L, Webb L, Ireland T, Walker D, Dong
¯uid inclusions in quartz at one atmosphere con®ningpressure.
J Metamorph Geol 7:229±242
Bohlen SR, Liotta JJ %1986) A barometer for garnet amphibolites
and garnet granulites. J Petrol 27:1025±1034
Brenan JM, Shaw HF, Phinney DL, Reyerson FJ %1994) Rutile-
aqueous ¯uid partitioningof Nb, Ta, Hf, Zr, U, and Th: im-
plications for high ®eld strength element depletions in island-arc
basalts. Earth Planet Sci Lett 128:327±339
S %1998) U/Pb zircon ages constrain the architecture of the ul-
trahigh-pressure Qinling±Dabie Orogen, China. Earth Planet
Sci Lett 161:215±230
Hacker BR, Ratschbacher L, Webb LE, McWilliams M, Ireland T,
DongS, Calvert A, Wenk HR %2000) Exhumation of the ul-
trahigh-pressure continental crust in east-central China: Late
Triassic±Early Jurassic extension. J Geophys Res 105:13339±
13364
È
Brey GP, Kohler T %1990) Geothermobarometry in four-phase Hall DL, Sterner SM %1993) Preferential water loss from synthetic
lherzolites II. New thermobarometers, and practical assessment
of existingthermobarometers. J Petrol 31:1353±1378
Brown PE %1989) FLINCOR: a microcomputer program for the
reduction and investigation of ¯uid inclusion data. Am Mineral
¯uid inclusions. Contrib Mineral Petrol 114:489±500
Hall DL, Bodnar RJ, CraigJR %1991) Fluid inclusion constraints
on the uplift history of the metamorphosed massive sul®de
deposit at Ducktown, Tennessee. J Metamorph Geol 9:551±565
Heinrich CH %1982) Kyanite±eclogite to amphibolite facies evolu-
tion of hydrous ma®c and pelitic rocks, Adula nappe, central
Alps. Contrib Mineral Petrol 81:30±38
74:1390±1393
Brunsmann A, Franz G, Erzinger J, Landwehr D %2000) Zoisite-
and clinozoisite-segregations in metabasites %Tauern window,
Austria) as evidence for high-pressure ¯uid±rock interaction. Heinrich W, Althaus E %1981) Die obere Stabilita
J Metamorph Geol 18:1±21 Lawsonit plus Albit bzw. Jadeit. Fortschr Mineral 58:49±50
Carswell DA %1990) Eclogites and the eclogite facies. In: Carswell Heinrich W, Althaus E %1988) Experimental determination of the
È
tsgrenze von
DA %ed) Eclogite facies rocks. Blackie, Glasgow, pp 1±13
Carswell DA, O'Brien PJ, Wilson RN, Zhai M %1997) Thermoba-
rometry of phengite-bearing eclogites in the Dabie Mountains
of central China. J Metamorph Geol 15:239±252
Castelli D, Rolfo F, Companioni R, Xu S %1998) Metamorphic
veins with kyanite, zoisite and quartz in the Zhu±Jia±Chong
eclogite, Dabie Shan, China. Island Arc 7:159±173
reactions 4 Lawsonite+1 Albite=1 Paragonite+2 Zoisite+2
Quartz+6 H O and 4 Lawsonite+1 Jadeite=1 Paragonite+2
Zoisite+1 Quartz+6 H O. N Jahrb Mineral 56:Beiheft 1, 38
2
2
Heinrich W, Gottschalk M %1995) Metamorphic reactions between
¯uid inclusions and mineral hosts. I. Progress of the reaction
2
calcite+quartz=wollastonite+CO in natural wollastonite-
hosted ¯uid inclusions. Contrib Mineral Petrol 122:51±61
Cherniak DJ %2000) Pb di€usion in rutile. Contrib Mineral Petrol Hey MT %1957) A new review of the chlorites. Mineral Mag
1
39:198±207
Colomby A %1988) Me
iques des Alpes ouest-centrales %ge
Locarno). PhD Thesis, University Lausanne
30:245±253
Â
tamorphisme et ge
Â
ochemie des roches maf- Holland TJB %1979) High water activities in the generation of high-
Â
opro®l Vie
Á
ge±Domodossola±
pressure kyanite eclogites of the Tauern window. Austria.
J Geol 87:1±27

3
45
Holland TJB %1980) The reaction albite=jadeite+quartz deter-
mined experimentally in the range of 600±1200 °C. Am Mineral
Dabie Mountains, central China: a ®eld trip guide to the 3rd
international eclogite ®eld symposium. Chin Sci Bull Suppl
40:1±66
6
5:129±134
Holland TJB %1990) Activities of components in omphacitic solid Liou JG, ZhangRY, Eide EA, Maruyama S, WangX, Ernst WG
solution. Contrib Mineral Petrol 105:446±453
Hollister LS %1990) Enrichment of CO in ¯uid inclusions in quartz
by removal of H O duringcrystal-plastic deformation. J Struct
Geol 12:895±901
Holloway JR %1981) Compositions and volumes of supercritical Manhe
uids in the Earth's crust. In: Hollister LS, Crawford ML %eds) thorium±lead and rubidium±strontium study of the Saint
Severin amphoterite: consequences for early solar system
%1996) Metamorphism and tectonics of high-P and ultrahigh-P
belts in Dabie±Sulu regions, eastern central China. In: Yin A,
Harrison TM %eds) The tectonic evolution of Asia, Rubey vol-
ume IX. Cambridge University Press, Cambridge, pp 300±343
2
2
Á
s G, Minster JF, Allegre CJ %1978) Comparative uranium±
Á
¯
Fluid inclusions: applications to petrology. Short course
handbook, vol 6. Mineral Association of Canada, pp 13±38
Ja€ey AH, Flynn KF, Glendenin LE, Bentley WC, EsslingAM
Â
chronology. Earth Planet Sci Lett 39:14±24
ManningCE %1994) Fractal clusteringof metamorphic veins.
Geology 22:335±338
Maresch W %1977) Experimental studies on glaucophane: an
analysis of present knowledge. Tectonophysics 43:109±125
%1971) Precision measurement of half-lives and speci®c activities
of U and U. Phys Rev C4:1889±1906
Jamtveit B, Bucher-Nurminen K, Austrheim H %1990) Fluid con-
235 238
trolled eclogitization of granulites in deep crustal shear zones, Massonne H-J %1991) High pressure low temperature metamor-
Bergen arcs, western Norway. Contrib Mineral Petrol 104:184± phism of pelites and other protoliths based on experiments in
2 2 3 2 2
93 O±MgO±Al O ±SiO ±H O. Habil Thesis, Univ
1
the system K
Bochum
Jarosewich E, Nelen JA, NorbergJA %1980) Reference samples for
electron microprobe analysis. Geostandards Newslett 4:43±47
Johnson EL, Hollister LS %1995) Syndeformational ¯uid trapping
in quartz: determiningthe pressure±temperature conditions of
Mattauer M, Matte P, Malavieille J, Tapponnier P, Maluski H,
Qin X, Lun L, Qin T %1985) Tectonics of the Qinlingbelt: build
up and evolution of eastern Asia. Nature 317:496±500
deformation from ¯uid inclusions and the formation of pure Mezger H, Hanson GN, Bohlen SR %1989) High-precision U±Pb
ages of metamorphic rutile: application to the cooling history of
high-grade terranes. Earth Planet Sci Lett 96:106±118
CO
2
¯uid inclusions during grain boundary migration.
J Metamorph Geol 13:239±249
Klemd R %1989) P±T evolution and ¯uid inclusion characteristics of Newton RC, Haselton HT %1981) Thermodynamics of the garnet±
retrograded eclogites, MunchbergGneiss Complex, Germany. plagioclase±Al SiO ±quartz geobarometer. In: Newton RC,
Contrib Mineral Petrol 102:221±229
Klemd R, Brocker M %1999) Fluid in¯uence on mineral reactions in
È
2
5
Navrotzky A, Wood BJ %eds) Thermodynamics of minerals and
melts. Springer, Berlin Heidelberg New York, pp 129±145
È
ultrahigh-pressure granulites: a case study in the Snieznik Mts Okay AI %1993) Petrology of a diamond and coesite bearing
%
West Sudetes, Poland). Contrib Mineral Petrol 136:358±373
Klemd R, van den Kerkhof AM, Horn EE %1992) High-density
CO ±N
MunchbergGneiss complex, SE Germany. Contrib Mineral
Petrol 111:409±441
Klemd R, Brocker M, Schramm J %1995) Characterisation of am-
metamorphic terrain: Dabie Shan, China. Eur J Mineral 5:659±
667
2
inclusions in eclogite-facies metasediments of the Papike JJ, Ross M, Baldwin K %1974) Amphiboles and pyroxenes:
2
È
characterization of other than quadrilateral components and
estimates of ferric iron from microprobe data. Geol Soc Am
Abstr Prog6:1053±1054
È
phibolite-facies ¯uids of Variscan eclogites from the Orlica± Passchier CW, Trouw RAJ %1996) Microtectonics. Springer, Berlin
Snieznik dome %Sudetes, SW Poland). Chem Geol 119:101±113 HeidelbergNew York
Kohn MJ, Spear FS %1989) Empirical calibration of geobarometers Philippot P, Selverstone J %1991) Trace-element-rich brines in
for the assemblage garnet + hornblende + plagioclase ‡
eclogitic veins; implications for ¯uid composition and transport
duringsubduction. Contrib Mineral Petrol 106:417±430
Kohn MJ, Spear FS %1990) Two new geobarometers for garnet Ramboz C, Pichavaut M, Weisbrod A %1982) Fluid immiscibility in
quartz. Am Mineral 74:77±84
amphibolites, with applications to southeastern Vermont. Am
Mineral 75:89±96
Kretz R %1983) Symbols for rock-formingminerals. Am Mineral
natural processes: use and misuse of ¯uid inclusion data. II.
Interpretation of ¯uid inclusion data in terms of immiscibility.
Chem Geol 37:29±48
6
8:277±279
Ramsay JG, Huber MI %1983) The techniques of modern structural
geology, vol 1. Strain analysis. Academic Press, London
Ratschbacher L, Hacker BR, Calvert A, Webb LE, Ireland T,
McWilliams M, DongS %2000) Exhumation of the ultrahi gh -
pressure continental crust in east-central China: Cretaceous and
Cenozoic unroo®ngand the Tan±Lu Fault. J Geophys Res
105:13303±13338
Krogh RE %1988) The garnet±clinopyroxene Fe±Mg geother-
mometer ± a reinterpretation of existingexperimental data.
Contrib Mineral Petrol 99:44±48
2
+
Krogh RE %2000) The garnet±clinopyroxene Fe ±Mg eg other-
mometer: an updated calibration. Metamorph Geol
8%2):211±219
J
1
Laird J, Albee AL %1981) Pressure, temperature, and time indi- Ridley J %1984) Evidence of a temperature-dependent blueschist to
eclogite transformation in high pressure metamorphism of
metabasic rocks. J Petrol 25:852±870
cators in ma®c schists: their application to reconstructing
the polymetamorphic history of Vermont. Am J Sci 281:127±
1
Leake BW, Wooley AR, Arps CES, Birch WD, Gilbert MC, Grice
75
Roedder E %1984) Fluid inclusions. Reviews in mineralogy, vol 12.
Mineral Soc Am
JD, Hawthorne FC, Kato A, Kisch HJ, Krivovichev VG, Romer RL, Heinrich W %1998) Transport of Pb and Sr in leaky
Linthout K, Laird J, Mandarino J, Maresch WV, Nickel EH,
Rock NMS, Schumacher JC, Smith DC, Stephenson NCN,
aquifers of the Bufa Del Diente contact metamorphic aureole,
north-east Mexico. Contrib Mineral Petrol 131:155±170
Ungaretti L, Whittaker EJW, Youzhi G %1997) Nomencla- Romer RL, Ro
È
tzler J %2001) P±T±t evolution of ultrahigh-tem-
perature granulites from the Saxon Granulite Massif, Germany.
Part II. Geochronology. J Petrol %in press)
ture of amphiboles. Report of the subcommittee on amphi-
boles of the International Mineralogical Association
Commission on new minerals and mineral names. Eur J Rowley DB, Xue F, Tucker RD, PengZX, Baker J, Davis A %1997)
Mineral 9:623±651
Li SG, Xiao YL, Liu DL, Chen YZ, Ge N, ZhangZ, Sun SS, Cong
B, ZhangR, Hart SR, WangS %1993) Collision of the North
Ages of ultrahigh pressure metamorphism and protolith or-
thogneisses from the eastern Dabie Shan: U/Pb zircon geo-
chronology. Earth Planet Sci Lett 151:191±203
China and Yangtse blocks and formation of coesite-bearing Santos Zalduegui JF, Scha
È
rer U, Ibarguchi JIG, Girardeau J %1996)
Origin and evolution of the Paleozoic Cabo Ortegal ultrama®c±
ma®c complex %NW Spain): U±Pb, Rb±Sr and Pb±Pb isotope
data. Chem Geol 129:281±304
eclogites; timing and processes. Chem Geol 109:89±111
Liou JG, WangQ, ZhangRY, Zhai M, CongBL %1995) Ultrahi gh -
P metamorphic rocks and their associated lithologies from the

3
46
Scambelluri M, Pennachioni G, Phillipot P %1998) Salt-rich aque- P±T paths, and implications for regional tectonics. J Geol
ous ¯uids formed duringeclo gi tization of metabasites in the 100:231±250
Alpine continental crust %Austroalpine Mt. Emilius unit, Italian WangX, ZhangR, Liou JG %1995) UHPM terrane in east central
Western Alps). Lithos 43:151±167
Schliestedt M %1980) Phasengleichgewichte in Hochdruckgesteinen
von Sifnos, Griechenland. PhD Thesis, TU Braunschweig
China. In: Coleman RG, WangX %eds) Ultrahi hg pressure
metamorphism. Cambridge University Press, Cambridge,
pp 356±390
Schmid R, Franz L, Oberha
È
nsli R, DongS %2000) Hi hg Si-
Waters DJ, Martin HN %1993) Geobarometry in phengite-bearing
eclogites. Terra Abstr 5:410±411; updated calibration of 1996
at: http://www.earth.ox.ac.uk/ꢀdavewa/ecbar.html
Xiao YL, Li SG, Jagoutz E, Cheng W %1995) P±T±t path for
coesite-bearing peridotite±eclogite association in the Bixiling,
Dabie Mountains. Chin Sci Bull 40:156±158
phengite, mineral chemistry and P±T evolution of ultra-high-
pressure eclogites and calc-silicates from the Dabie Shan,
eastern PR China. Geol J vol 35:185±207
Spear FS %1980) The NaSi=CaAl exchange equilibria between
plagioclase and amphibole. Contrib Mineral Petrol 72:33±41
Spear FS %1991) On the interpretation of peak metamorphic tem- Xiao YL, Hoefs J, van den Kerkhof AM, FiebigJ, ZhengY %2000)
peratures in the light of garnet di€usion during cooling.
J Metamorph Geol 9:379±388
Steiger RH, Jager E %1977) Subcommission on geochronology:
È
Fluid history of UHP metamorphism in Dabie Shan, China: a
¯uid inclusion and oxygen isotope study on the coesite-bearing
eclogite from Bixiling. Contrib Mineral Petrol 139:1±16
convention on the use of decay constants in geo- and cosmo- Yardley BWD %1977) An empirical study of di€usion in garnet. Am
chronology. Earth Planet Sci Lett 36:359±362 Mineral 62:793±800
Tilton GR %1973) Isotopic lead ages of chondritic meteorites. Earth Yardley BWD %1983) Quartz veins and devolatilisation during
Planet Sci Lett 19:321±329 metamorphism. J Geol Soc Lond 140:657±663
Touret JLR %1992) Fluid inclusions in subducted rocks. Proc K Yardley BWD %1986) Fluid migration and veining in the Conne-
Ned Akad Wet 95:385±403
Voll G %1960) New work on petrofabrics. Geol J 213:503±567
Voll G %1969) Klastische Minerale aus den Sedimentserien der
mara Schists, Ireland. In: Walther JV, Wood BJ %eds) Advances
in physical geochemistry, vol 5. Springer, Berlin Heidelberg
New York, pp 109±131
Schottischen Highlands und ihr Schicksal bei aufsteigender You Z, Han Y, YangW, ZhangZ, Wie B, Liu R %1996) The hi gh -
Regional- und Kontaktmetamorphose. Habil Thesis, TU
Berlin
von Raumer JF %1984) The External Massifs, relics of the Variscan
Basement in the Alps. Geol Rundsch 73:1±31
WangX, JingY, Liou JG, Pan G, LiangW, Xia M, Maruyama S
pressure and ultra-high-pressure metamorphic belt in the East
Qinlingand Dabie Mountains, China. China University of
Geoscience Press, Wuhan
Zartman RE, Doe BR %1981) Pumbotectonics ± the model. Tec-
tonophysics 75:135±162
%1990) Field occurrences and petrology of eclogites from the ZhangY, Frantz JD %1987) Determination of the homo ge nization
Dabie Mountains, Anhui, central China. Lithos 25:119±131
WangX, Liou JG, Maruyama S %1992) Coesite-bearingeclo gi tes
from the Dabie Mountains, central China: petrogenesis,
temperatures and densities of supercritical ¯uids in the system
NaCl±KCl±CaCl
Geol 64:335±350
2 2
±H O usingsynthetic ¯uid inclusions. Chem