Time calibration of a P–T path from a Variscan high ... · Granulite-facies metamorphism and migmatite forma-tion by partial melting in mid to upper crustal levels previously thickened

Angelika Kalt á Fernando Corfu á Jan R. Wijbrans

Time calibration of a P±T path from a Variscan high-temperaturelow-pressure metamorphic complex (Bayerische Wald, Germany),and the detection of inherited monazite

Abstract A temperature±time path was constructed forhigh-temperature low-pressure (HT±LP) migmatites ofthe Bayerische Wald, internal zone of the Variscan belt,Germany. The migmatites are characterised by pro-grade biotite dehydration melting, peak metamorphicconditions of approximately 850 °C and 0.5±0.7 GPaand retrograde melt crystallisation at 800 °C. The time-calibration of the pressure±temperature path is basedon U±Pb dating of single zircon and monazite grainsand titanite separates, on 40Ar/39Ar ages obtained byincremental heating experiments on hornblende sepa-rates, single grains of biotite and K-feldspar, and on40Ar/39Ar spot fusion ages of biotite determined in situfrom sample sections. Additionally, crude estimates ofthe duration of peak metamorphism were derived fromgarnet zoning patterns, suggesting that peak tempera-tures of 850 °C cannot have prevailed much longerthan 2.5 Ma. The temperature±time paths obtained fortwo areas approximately 30 km apart do not dierfrom each other considerably. U±Pb zircon ages re¯ectcrystallisation from melt at 850±800 °C at 323 Ma(southeastern area) and 326 Ma (northwestern area).The U±Pb ages of monazite mainly coincide with thosefrom zircon but are complicated by variable degrees ofinheritance. The preservation of inherited monazite and

the presence of excess 206Pb resulting from the incor-poration of excess 230Th in monazite formed duringHT±LP metamorphism suggest that monazite ages inthe migmatites of the Bayerische Wald re¯ect crystal-lisation from melt at 850±800 °C and persistence ofolder grains at these temperatures during a compara-tively short thermal peak. The U±Pb ages of titanite(321 Ma) and 40Ar/39Ar ages of hornblende (322±316 Ma) and biotite (313±309 Ma) re¯ect coolingthrough the respective closure temperatures of ap-proximately 700, 570±500 and 345±310 °C published inthe literature. Most of the feldspars' ages (305±296 Ma)probably record cooling below 150±300 °C, while twograins most likely have higher closure temperatures.The temperature±time paths are characterised by ashort thermal peak, by moderate average cooling ratesand by a decrease in cooling rates from 100 °C/my attemperatures between 850±800 and 700 °C to 11±16 °C/my at temperatures down to 345±310 °C. Fur-ther cooling to feldspar closure for Ar was probablyeven slower. The lack of decompressional features, themoderate average cooling rates and the decline ofcooling rates with time are not easily reconciled with amodel of asthenospheric heating, rapid uplift and ex-tension due to lithospheric delamination as proposedelsewhere. Instead, the high peak temperatures atcomparatively shallow crustal levels along with theshort thermal peak require external advective heatingby hot ma®c or ultrama®c material.

Introduction

Granulite-facies metamorphism and migmatite forma-tion by partial melting in mid to upper crustal levelspreviously thickened by continental collision have beendescribed from quite a few orogenic belts (e.g. De Yoreoet al. 1991 and references therein). These ®ndings seemin apparent con¯ict with most commonly acceptedthermal models for lithosphere thickened by collision.These models predict high-temperature metamorphism

A. Kalt (&)Mineralogisches Institut, Im Neuenheimer Feld 236,D-69120 Heidelberg, Germanye-mail: [email protected]: +49-6221-544805

F. Corfu1

Royal Ontario Museum, 100 Queen's Park, Toronto,Ontario, M5S 2C6, Canada

J. R. WijbransFaculty of Earth Sciences, Vrije Universiteit,De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

Present address:1Mineralogical-Geological Museum,Sars gate 1, N-0562 Oslo, NorwayEditorial responsibility: J. Hoefs

Published in Contributions to Mineralogy and Petrology 138, issue 2, 143-163, 2000which should be used for any reference to this work

1

and partial melting due to crustal stacking, increasedradiogenic heat production and thermal relaxation onlyin lower crustal levels (Thompson and Connolly 1995and references therein). Heating of the mid and uppercrust to the temperatures observed in high-temperaturelow-pressure (HT±LP) orogenic belts would take sometens of millions of years. In collisional orogens wherethis time span is not available, additional heat is re-quired to trigger HT±LP metamorphism and melting(De Yoreo et al. 1991; Thompson and Connolly 1995;Zen 1995). In most common models the asthenosphere isinvoked as an additional heat source, brought up to thebase of the crust either by convective removal of thelithospheric mantle (e.g. England and Houseman 1989),delamination of the lithosphere (Bird 1979) or by slabbreako (Davies and von Blanckenburg 1995). Othermodels consider advective heating by rising magmas orpre-collisional lithospheric extension (De Yoreo et al.1991 and references therein).

Determining the pressure±temperature±time (P±T±t)paths of metamorphic rocks in orogenic belts can be ofconsiderable help in distinguishing the possible causes ofHT±LP metamorphism. Heating by burial and subse-quent thermal relaxation and exhumation of the crustresult in comparatively steep, sometimes counterclock-wise P±T paths of time spans of tens of millions of years.On the other hand, lithospheric or slab removal/detachment induces rapid uplift and extension of thepreviously thickened crust as recognised from quite afew mountain belts (e.g. Dewey 1988; Malavieille 1993;Ruppel 1995). In these cases, decompression (exhuma-tion) of metamorphic rocks is accompanied by heating(Platt and England 1994). Heating during decompres-sion and rapid exhumation of metamorphic rocks withcooling rates of >100 to 500 °C/my have been describedfrom a number of extensional settings (e.g. Kohn et al.1993; van der Wal and Vissers 1993; GarcõÂ a-Casco 1996;Soto and Platt 1999; Zeck and Whitehouse 1999). Animportant boundary condition for this mechanism is aconsiderable time lapse (at least 10 my; Platt et al. 1998)between lithospheric thickening and the onset of exten-sion. Heating by underplating or rise of magmas inducesP±T paths comparable to those typical of contactmetamorphism, characterised by an almost isobarictemperature increase and slow isobaric cooling.

Within the Variscan belt of Europe, remnants of anHT metamorphic belt can be traced from the BohemianMassif in the east via Schwarzwald, Vosges, MassifCentral and the Armorican Massif to the Iberian Massifin the southwest (e.g. Matte 1986). Most of the beltrepresents fairly shallow crustal levels consisting ofgneisses, migmatites and amphibolites. Granitoid plu-tons are abundant. The P±T conditions recorded bythese rocks require an external heat source for HT±LPmetamorphism (e.g. Le Me tour 1978; Latouche et al.1992; Brown and Dallmeyer 1996; Kalt et al. 1999). Lateorogenic extension has been recognised from large-scalestructures within the Variscan belt (e.g. Me nard andMolnar 1988; Costa and Rey 1995). However, within

some of the HT±LP metamorphic areas, compressionalstructures often dominate (e.g. Tanner and Behrmann1995; Behrmann and Tanner 1997) and evidence forconsiderable decompression from phase assemblagesand thermobarometry is often lacking (e.g. Latoucheet al. 1992; Kalt et al. 1999). HT±LP metamorphism hasbeen dated at approximately 330±310 Ma B.P., depend-ing on the location within the elongate belt (e.g. Grauertet al. 1974; Pin and Peucat 1986; Kalt et al. 1994a;Brown and Dallmeyer 1996). Eclogites contained in theHT±LP rocks occur as isolated exotic bodies of variablesize. In the Schwarzwald, they record a high-pressureevent at approximately 345±332 Ma (Kalt et al. 1994b;Kalt et al. 1997), suggesting that the time lapse betweenlithospheric thickening and the onset of heating may besmall. Hence, at least for some parts of the VariscanHT±LP metamorphic belt, heating may have occurredduring crustal thickening and may not have been ac-companied by extension.

The Bayerische Wald in the Bohemian Massif (Fig. 1)forms part of the Variscan HT±LP metamorphic belt.The P±T path of the migmatites is characterised byprograde dehydration melting, by granulite-facies peakmetamorphic conditions and by the absence of signi®-cant decompression during heating (Kalt et al. 1999;Fig. 1). The purpose of this investigation is to time-calibrate the P±T path of the migmatites in order toconstrain boundary conditions for, and to distinguishbetween, various causes for HT±LP metamorphism.Due to the high peak temperatures, information on theprograde part of the P±T path is not to be expected. In

Fig. 1a,b Variscan basement outcrops in central Europe, simpli®edgeological map of the Bayerische Wald with sample locations, andP±T path of the migmatites. a Simpli®ed geological map of theBayerische Wald, modi®ed after Kalt et al. (1999). Samplelocations of migmatites, gneisses and amphibolites are indicatedby numbers from 1±7 (compare Table 1). The inset shows Variscanbasement outcrops in central Europe. BM Bohemian Massif; MOMoldanubian zone; ST Saxothuringian zone; RH Rhenohercynianzone. The MO and ST de®ned by Kossmat (1927) represent theinternal part of the Variscan belt, characterised by high-grademetamorphism and plutonism. The black rectangle indicates thearea shown in the map. b P±T path for the migmatites of theBayerische Wald, modi®ed after Kalt et al. (1999). 1 Experimen-tally determined minimum temperatures and pressures for dehy-dration melting of metapelites in the absence of aqueous ¯uids bythe reaction biotite + sillimanite + plagioclase + quartz gar-net + K-feldspar + melt (Le Breton and Thompson 1988). 2aExperimentally determined minimum temperatures and pressuresfor dehydration melting of metagreywackes in the absence ofaqueous ¯uids by the reaction biotite + plagioclase + quartz =garnet + orthopyroxene + K-feldspar + melt (Vielzeuf andMontel 1994); 2b biotite-out curve at higher temperatures in thesame experiments. 2a is a minimum temperature for migmatites ofthe Bayerische Wald. The biotite-out curve marks maximumtemperatures for the migmatites as textures and biotite composi-tions indicate that biotite was not exhausted during partial melting.A, B, C and D refer to metamorphic stages as described in thesection on geological setting, petrological context and geochrono-logical background and in Kalt et al. (1999). The shaded areamarks the results of thermobarometry, indicating equilibrationdown to stage D

2

order to constrain the cooling history of the migmatitesfrom peak metamorphism to low temperatures, the fol-lowing dating techniques were applied: U±Pb dating ofzircon, monazite and titanite, 40Ar/39Ar dating ofhornblende, biotite and K-feldspar. In order to obtain

high-precision data and to be able to link them to petro-logical information, single grains were dated whereverpossible. In order to test possible regional variations inthe cooling path, various localities several kilometresapart were sampled.

Gro§er

Zwiesel

Viechtach

5 km

Gabbros,amphibolites

Cham

Sediments

Bodenmais

Tepl-Barrandian

Czech Rep

ublic

German

y

Arber

Regen

Granitoids

Bt-Fsp gneisses, locallymylonitic

Migmatites

Deformed granites

Mylonites

Diatexites includinggranitic diatexites

Bt-Pl gneisses

sample location

Mica schists

Bt-Sil gneisses

Moldanubian s.str.

Gneiss with Fsp megacrysts

Qtz-rich gneisses with local melts

7

6

52 3

1

4

4

0.8

0.6

0.4

0.2

0.0600 700 800 900 1000

T [°C]

ky

sil

and

sil

1.0

?

?

?

A

B

CD

1

P[GPa]

a

b

2a 2b

STRH

BM

Alps

MO

3

Geological setting, petrological contextand geochronological background

Most of the Bohemian Massif forms part of theinternal Variscan HT belt, the Moldanubian and Sax-othuringian zones according to the subdivision byKossmat (1927; Fig. 1). Apart from the Tepla -Bar-randian unit, the Moldanubian zone is characterised byhigh-grade metamorphism and widespread granitoidintrusions (Fig. 1). It has a complex structure withnappes of HT±HP granulites, peridotites and eclogitesthrust over HT±LP gneisses and migmatites. Thesegneisses and migmatites proper are devoid of MP orHP relics and only contain very rare exotic lenses ofeclogite.

The Bayerische Wald (Fig. 1), located at the south-western margin of the Bohemian Massif, comprises onlythe lower unit of HT±LP migmatites and gneisses, in-truded by granitoid plutons of variable size. Most mig-matites are of semipelitic to pelitic bulk composition andare characterised by the assemblages biotite + quartz +plagioclase + K-feldspar + cordierite + ilmenite. Ad-ditional garnet, spinel, orthopyroxene and sillimaniteoccur depending on bulk rock composition. The P±Tpathis most likely clockwise (Fig. 1) and can be divided intofour stages. Apart from a ®rst stage A only preserved ingarnets of a few samples, the further P±T path of allmigmatite samples is characterised by melt-producingbiotite dehydration reactions in the absence of an aqueous¯uid phase, peak conditions of approximately 850 °C/0.5±0.7 GPa and cooling (stage B) until the partial meltsformed at stage B crystallised (stage C). Subsequent stageD, during which mineral equilibria were frozen (770±846 °C/0.44±0.51 GPa), is characterised by decompres-sion and cooling. The amphibolites are included as bodiesof variable size in the migmatites and show relic assem-blages of either plagioclase+ clinopyroxene+ titanite orplagioclase + clinopyroxene+ orthopyroxene+ quartz+ ilmenite, recording a granulite-facies stage with thesame peak temperatures as those estimated for the mig-matites.

Previous geochronological studies have shown HT±LP metamorphism to be of Carboniferous age. U±Pbages of monazite grain-size fractions from migmatites ofthe Bayerische Wald range from 317 to 321 Ma (Grauertet al. 1974), similar to those obtained on metamorphicrocks of the Moldanubian zone in the adjacent Ober-pfaÈ lzer Wald (317 3±323 3 Ma; Teufel 1988). K±Ar ages of biotite grain-size fractions from variousmetamorphic rocks of the Bayerische Wald scatter be-tween 325 5 and 309 5 Ma (Kreuzer et al. 1989).For the granites intruding the migmatites, Rb±Sr min-eral ages range from 302 7 to 322 5 Ma, and K±Ar mineral ages scatter between 296 3 and320 8 Ma (compilation in Siebel 1998).

Sample selection

The aim of the study was to establish temperature±timepaths for HT±LP migmatites and gneisses of the Bay-erische Wald by U±Pb dating of zircon, monazite andtitanite and 40Ar/39Ar dating of amphibole, biotite andK-feldspar. A temperature±time path ideally requiresthat all chronometers used should be applied to oneoutcrop or restricted area. This approach was limited inthe case of the Bayerische Wald by the scarce occurrenceof amphibolites and amphibole-bearing migmatites andby the fact that most of the latter were found to behydrothermally overprinted. Therefore, amphibole dat-ing could be performed only at a few outcrops (Table 1,Fig. 1). However, a larger number of localities (Table 1,Fig. 1) had to be selected for dating in order to detectpossible regional variations in the cooling path. Thesampling sites can be roughly grouped into a south-eastern area around Bodenmais (Fig. 1), comprisinglocations 1±5 and a northwestern area around Cham(Fig. 1), comprising location 7 (and 6).

Two rock types were sampled for dating: (1) mig-matites and one gneiss for U±Pb dating of zircon andmonazite and 40Ar/39Ar dating of biotite and K-feld-spar, and (2) amphibolites for U±Pb dating of titaniteand 40Ar/39Ar dating of amphibole and in one case also40Ar/39Ar dating of biotite. Phase assemblages, meltingreactions, melt fractions and microstructures of themigmatites are described in detail in Kalt et al. (1999)and Berger and Kalt (1999). The amphibolites are ®ne-to medium-grained rocks with gneissose textures. Foli-ation is well pronounced and de®ned by amphibole andbiotite (where present). Samples BW-112 and BW-132contain quartz, clinopyroxene, orthopyroxene, biotiteand accessory ilmenite and apatite in addition to am-phibole and plagioclase. Pyroxene inclusions in someamphiboles indicate that amphibole grew at the expenseof pyroxenes. However, the coexistence of unreactedpyroxenes with amphibole, both displaying stable grainboundaries, indicates equilibrium between the minerals.Sample BW-134 shows compositional banding parallelto foliation. Amphibole- and plagioclase-rich layers al-ternate with clinopyroxene- and plagioclase-rich layers.Sample BW-118 contains amphibole, plagioclase andquartz in textural equilibrium. Samples BW-134 andBW-118 contain accessory titanite.

Characteristics of the phases and sample preparation

The compositions of biotite, K-feldspar and amphibolefrom the samples used here for dating are listed inTable 2 (for analytical procedure see Kalt et al. 1999).

Zircon is present as accessory phase in the mig-matites. It occurs interstitially in mesosomes and le-ucosomes and may also be enclosed in garnet andbiotite. In thin section, zircons display anhedral to eu-hedral shapes and in cases zoning under crossed polars.

4

For the present study, only zircons formed during HT±LP metamorphism were of interest. Hence, zircons wereonly selected from leucosomes where the chance of beingnewly crystallised from melt during HT±LP metamor-phism is high, whereas in mesosomes the probability ofzircons containing inherited radiogenic lead or olderzircon domains is much larger due to the restitic char-acter of some mesosomes (Kalt et al. 1999). Under thebinocular microscope, zircons partly showed core-rimstructures. In order to avoid inheritance problems theanalyses were carried out only on rim components(prisms and tips).

As is the case with zircon, monazite is an accessoryphase in migmatites where it occurs interstitially andforms inclusions in garnet and biotite. Monazites areanhedral to platy crystals that partly display colourzoning under crossed polars. Though much less oftenthan zircons, monazites may contain inherited radio-genic lead (e.g. Copeland et al. 1988), possibly in theform of an inherited monazite phase (Harrison et al.1995). The case was described by Teufel (1988) forgneisses of the OberpfaÈ lzer Wald. Therefore, in mig-matites with clear leucosome±mesosome boundaries(BW-70, BW-90, Table 1), mesosomes and leucosomeswere carefully separated before crushing. In one meso-some (BW-70M), monazite inclusions in garnet could be

separated. Under the binocular microscope, monazitesdisplayed clear core-rim structures only in one case(BW-90M) where fragments were broken (Table 3).

Titanite is restricted to amphibolites. It displays ir-regular, contorted shapes where grown interstitially(BW-118) or forms ¯at sphenoidal crystals where in-cluded in amphibole and plagioclase (BW-134). Nocolour zoning was detected under the polarising micro-scope.

Amphibole in amphibolites grows mainly at the ex-pense of granulite-stage pyroxenes, whereby it maycompletely replace pyroxenes in distinct layers of therocks. XMg, Al2O3 and K2O contents vary considerablyamong the samples, with K2O being generally low (0.27±0.82 wt%, Table 2). The compositions correspond tomagnesio-hornblende (BW-112), edenitic hornblende(BW-134) and actinolitic hornblende (BW-118, BW-132)according to the nomenclature of Leake (1978).

Biotite is one of the major phases in the migmatites. Itoccurs mainly in mesosomes and melanosomes and onlyrarely in leucosomes. Biotite is partly consumed by de-hydration melting reactions. Due to the divariant char-acter of the latter reactions, a large percentage of thebiotite remains stable during partial melting. Duringsubsequent cooling, new biotite grows and all biotitecompositions are re-equilibrated on the retrograde path

Table 1 Sample characteristics

Locationa Sample Lithology Phase assemblageb Dating techniquec

1 BW-22 Migmatite crd + grt + bt + pl + qtz 40Ar/39Ar IH biotiteBW-28 Migmatite crd + grt + bt + Kfs + pl +

qtz + sil + spl

40Ar/39Ar SF biotite

BW-70L Leucosome Kfs + pl + qtz + bt U-Pb zircon, U-Pb monaziteBW-70M Mesosome crd + grt + bt + Kfs + pl +

qtz + sil + splU-Pb monazite

2 BW-30 Migmatite crd + grt + bt + pl + qtz 40Ar/39Ar IH biotiteBW-32 Migmatite crd + grt + bt + Kfs + pl + qtz 40Ar/39Ar SF biotiteBW-67 Migmatite crd + grt + bt + Kfs + pl +

qtz + sil + spl

40Ar/39Ar SF biotite, 40Ar/39Ar IHK-feldspar

3 BW-86246 Gneiss crd + grt + bt + pl + qtz + mt + spl 40Ar/39Ar IH biotite

4 BW-132 Amphibolite hbl + cpx + opx + pl + qtz + bt + ilm 40Ar/39Ar IH hornblende, 40Ar/39Ar IHbiotite

BW-134 Amphibolite hbl + cpx + pl + ttn U-Pb titanite, 40Ar/39Ar IH hornblende

5 BW-118 Amphibolite hbl + pl + qtz + ttn U-Pb titanite, 40Ar/39Ar IH hornblendeBW-120 Migmatite crd + bt + Kfs + pl + qtz U-Pb monazite, 40Ar/39Ar IH

biotite, 40Ar/39Ar IH K-feldspar

6 BW-112 Amphibolite hbl + cpx + opx + pl + qtz +bt + ilm

40Ar/39Ar IH hornblende

BW-116 Migmatite crd + bt + Kfs + pl + qtz + sil U-Pb monazite, 40Ar/39Ar IHbiotite, 40Ar/39Ar IH K-feldspar

7 BW-44 Migmatite crd + grt + bt + Kfs + pl +qtz + sil

40Ar/39Ar IH biotite, 40Ar/39Ar SFbiotite, 40Ar/39Ar IH K-feldspar

BW-46 Migmatite crd + grt + bt + Kfs + pl +qtz + spl + sil

40Ar/39Ar IH biotite, 40Ar/39Ar IHK-feldspar

BW-90L Leucosome Kfs + pl + qtz + bt U-Pb zircon, U-Pb monaziteBW-90M Mesosome crd + grt + bt + Kfs + pl +

qtz + spl + silU-Pb monazite, 40Ar/39Ar SF biotite

a Location numbers correspond to those given in Fig. 1bMineral abbreviations according to Kretz (1982)c IH incremental heating experiments on single minerals or mineralseparates with a laser, SF spot fusion experiments on minerals in

thin sections with a laser. In addition to the minerals indicated,migmatites may contain accessory zircon, monazite, apatite, ilme-nite, graphite, pyrrhotite and pyrite. Amphibolites may bear ad-ditional zircon, ilmenite and apatite

5

Table

2Mineralcompositions.Mineralabbreviationsaccordingto

Kretz

(1982);n.d.notdetermined,n.c.notcalculated.Form

ula

calculations:bt,12O,2OH,Fe2

+/Fe tot=

0.85;

Kfs:8O,Fe3

+=Fe tot;hbl,22O,15cations±Na±K,2OH,Fe3

+per

chargebalance

Location1

Location2

Loca-

tion3

Location4

Location5

Location6

Location7

BW-22BW-28

BW-30

BW-32

BW-67

86246

BW-132

BW-132BW-134BW-118

BW-120BW-120BW-112

BW-116BW-116BW-44

BW-44

BW-46

BW-46

BW-90

bt

bt

bt

bt

Kfs

bt

hbl

bt

hbl

hbl

bt

Kfs

hbl

bt

Kfs

bt

Kfs

bt

Kfs

bt

SiO

234.96

34.96

34.22

34.02

65.57

34.88

51.09

38.19

43.77

50.49

34.74

64.47

48.07

33.80

64.50

34.42

65.21

34.45

64.75

34.53

TiO

24.41

4.28

4.08

3.95

0.00

0.04

0.52

2.68

1.87

0.78

3.82

0.00

1.51

3.82

0.00

5.12

0.00

4.15

0.00

4.96

Al 2O

318.26

18.38

18.69

18.73

19.18

19.91

5.86

15.00

10.77

5.56

18.49

18.28

7.02

18.28

19.16

17.88

19.32

17.85

19.54

17.74

Cr 2O

30.11

0.08

0.07

0.07

0.00

0.01

0.21

0.26

0.15

0.08

0.05

0.00

0.17

0.35

0.00

0.13

0.00

0.14

0.00

0.11

Fe 2O

33.33

3.37

3.85

3.91

0.02

3.87

0.13

2.21

0.04

0.26

3.59

0.00

0.00

4.04

0.04

3.48

0.01

3.69

0.05

3.47

FeO

16.99

17.18

19.63

19.96

0.00

19.74

10.89

11.29

14.82

12.56

18.30

0.00

15.00

20.62

0.00

17.74

0.00

18.83

0.00

17.68

MnO

0.09

0.09

0.08

0.08

0.00

0.00

0.36

0.15

0.26

0.40

0.17

0.00

0.23

0.09

0.00

0.07

0.00

0.13

0.00

0.08

MgO

8.45

8.18

6.00

5.91

0.00

7.59

16.22

15.25

11.44

14.84

7.01

0.00

12.79

5.33

0.00

7.71

0.00

7.38

0.00

7.63

CaO

0.07

0.00

0.00

0.00

0.23

0.02

11.55

0.00

12.09

11.78

0.01

0.04

10.95

0.00

0.08

0.00

0.12

0.00

0.12

0.02

Na2O

0.17

0.11

0.18

0.21

3.71

0.10

0.57

0.18

1.60

0.55

0.17

1.64

0.73

0.15

1.71

0.17

2.77

0.17

2.10

0.16

K2O

9.25

9.29

9.12

9.06

11.24

9.09

0.28

9.09

0.82

0.27

9.17

14.11

0.67

9.14

14.06

9.11

12.47

9.15

13.54

9.31

H2O

3.97

3.96

3.91

3.90

n.d.

3.90

2.09

4.03

2.01

2.07

3.92

n.d.

2.03

3.87

n.d.

3.94

n.d.

3.92

n.d.

3.93

Total100.06

99.89

99.82

99.78

99.95

99.15

99.76

98.33

99.64

99.63

99.43

98.54

99.17

99.47

99.55

99.77

99.90

99.84

100.10

99.61

Si

2.639

2.644

2.624

2.615

2.979

2.680

7.326

2.838

6.524

7.320

2.655

3.000

7.101

2.621

2.972

2.619

2.975

2.634

2.962

2.633

Ti

0.250

0.244

0.235

0.228

0.000

0.002

0.056

0.150

0.209

0.085

0.220

0.000

0.168

0.223

0.000

0.293

0.000

0.238

0.000

0.285

Al

1.624

1.638

1.589

1.696

1.027

1.803

0.990

1.314

1.893

0.951

1.666

1.003

1.223

1.670

1.041

1.603

1.039

1.608

1.053

1.594

Cr

0.006

0.004

0.004

0.004

0.000

0.001

0.023

0.015

0.018

0.009

0.003

0.000

0.020

0.022

0.000

0.008

0.000

0.008

0.000

0.007

Fe3

+0.189

0.192

0.222

0.226

0.001

0.224

0.014

0.124

0.004

0.028

0.206

0.000

0.000

0.236

0.002

0.199

0.000

0.212

0.002

0.199

Fe2

+1.073

1.087

1.259

1.283

0.000

1.269

1.306

0.702

1.847

1.523

1.169

0.000

1.853

1.337

0.000

1.129

0.000

1.204

0.000

1.127

Mn

0.006

0.006

0.005

0.005

0.000

0.000

0.044

0.009

0.033

0.050

0.011

0.000

0.028

0.006

0.000

0.004

0.000

0.008

0.000

0.005

Mg

0.950

0.923

0.686

0.677

0.000

0.870

3.468

1.690

2.542

3.206

0.799

0.000

2.817

0.616

0.000

0.874

0.000

0.841

0.000

0.867

Ca

0.006

0.000

0.000

0.000

0.011

0.001

1.774

0.000

1.930

1.829

0.001

0.002

1.733

0.000

0.004

0.000

0.006

0.000

0.006

0.001

Na

0.025

0.016

0.026

0.031

0.327

0.014

0.159

0.026

0.463

0.153

0.025

0.148

0.210

0.022

0.153

0.026

0.245

0.025

0.186

0.024

K0.890

0.896

0.892

0.888

0.651

0.891

0.051

0.826

0.156

0.050

0.894

0.838

0.127

0.904

0.826

0.884

0.726

0.892

0.790

0.906

OH

2.000

2.000

2.000

2.000

n.c.

2.000

2.000

2.000

2.000

2.000

2.000

n.c.

2.000

2.000

n.c.

2.000

n.c.

2.000

n.c.

2.000

Total

7.659

7.651

7.642

7.653

4.996

7.756

15.210

7.729

15.619

15.203

7.647

4.991

15.278

7.656

4.997

7.638

4.991

7.671

4.999

7.647

6

Table

3ResultsofU±Pbdating

Loca-

tion

Sample,analysisno.

and

mineralproperties

aWeight

[lg]b

U [ppm]b

Th/U

cPbcom

[pg]d

206Pb/

204Pbe

206Pb/238U

f,g

207Pb/235U

f,g

207Pb/206Pbf,g

206Pb/

238U

[Ma]f

+/)

[2r]g

207Pb/

235U

[Ma]f

+/)

[2r]g

207Pb/

206Pb[M

a]f

+/)

[2r]g

1BW-70M

1,m

euincl

ingrt

[1]

11918

16.00

12.0

701

0.06842

58

0.4867

60

0.05159

44

426.6

3.6

402.6

4.2

267

20

2,m

eu[1]

12053

13.84

1.2

5547

0.05162

25

0.3764

21

0.05288

11

324.5

1.6

324.4

1.6

323.6

4.9

3,m

eu[1]

1558

16.71

1.2

1529

0.05136

26

0.3745

31

0.05261

32

324.5

1.6

323.0

2.4

312

14

4,m

sb[1]

34932

9.19

8.7

5375

0.05143

28

0.3738

23

0.05271

10

323.3

1.8

322.4

1.6

316.3

4.3

5,m

euincl

A[1]

20

3079

17.43

76.9

2548

0.05045

25

0.3667

22

0.05271

12

317.3

1.6

317.2

1.6

316.3

5.2

BW-70L

6,m

eq[1]

35785

7.68

8.7

6458

0.05144

27

0.3740

22

0.05273

10

323.4

1.6

322.6

1.6

317.0

4.3

7,m

eu[1]

10

6278

7.64

5.6

35862

0.05148

33

0.3736

25

0.05263

08

323.6

2.0

322.3

1.8

312.7

3.3

8,m

eu[1]

47205

2.50

1.9

49912

0.05153

31

0.3731

24

0.05250

08

323.9

2.0

321.9

1.8

307.3

3.6

9,m

eu[1]

13095

5.85

1.3

7743

0.05128

24

0.3729

20

0.05274

10

322.4

1.4

321.8

1.4

317.6

4.3

10,m

sbA

[1]

17466

7.11

1.2

20059

0.05103

29

0.3710

23

0.05273

10

320.9

1.8

320.4

1.8

317.2

4.6

11,zeu

tipsA

[2]

2145

0.06

2.0

489

0.05145

26

0.3746

65

0.05281

82

323.4

1.6

323.1

4.8

321

35

4BW-134

31,tfr

lbrA

[>30]

333

21

1.47

405

73

0.05108

44

0.383

23

0.0543

32

321.2

2.6

329

17

384

129

32,teu-sbflbr[>

30]

170

16

1.79

178

68

0.05107

44

0.376

23

0.0534

33

321.1

2.8

324

17

347

136

5BW-118

33,tan-frlbrA

[>30]

173

110

0.64

132.2

477

0.05099

26

0.3729

36

0.05303

39

320.6

1.6

321.8

2.6

330.2

16.5

BW-120

23,m

eueq

[1]

21701

8.02

3.3

3332

0.05116

27

0.3723

23

0.05277

16

321.7

1.6

321.3

1.8

318.9

7.0

24,m

eueq

[1]

115364

5.63

3.2

15535

0.05122

26

0.3722

21

0.05270

08

322.0

1.6

321.3

1.6

315.9

3.6

25,m

eueq

[1]

22263

9.28

5.7

2590

0.05145

24

0.3737

21

0.05268

12

323.4

1.4

322.4

1.6

314.9

5.3

26,m

eueq

incl

[1]

26956

14.74

12.0

3836

0.05142

25

0.3740

21

0.05275

10

323.2

1.6

322.6

1.6

318.0

4.5

6BW-116

27,m

eu-sbeq

[1]

1519

10.95

2.6

697

0.05377

29

0.3936

44

0.05309

49

337.6

1.8

337.0

3.2

333

21

28,m

eu-sbtip[1]

31979

7.59

2.3

8527

0.05203

25

0.3828

22

0.05337

13

327.0

1.6

329.1

1.6

344.5

5.4

29,m

eueq

[1]

11780

8.31

6.7

883

0.05188

24

0.3789

27

0.05296

23

326.1

1.6

326.2

2.0

327

10

30,m

eueq

incl

[1]

34569

11.13

33

1360

0.05141

28

0.3744

25

0.05283

17

323.2

1.8

322.9

1.8

321.3

7.4

7BW-90M

12,m

eusp

[1]

2318

6.86

3.0

734

0.05421

27

0.3950

37

0.05285

39

340.3

1.6

338.0

2.8

322

17

13,m

sbA

[1]

18915

10.11

2.0

14368

0.05216

28

0.3809

23

0.05295

10

327.8

1.8

327.7

1.8

326.7

4.3

14,m

eu[1]

35287

6.75

4.1

12613

0.05200

30

0.3792

24

0.05289

08

326.8

1.8

326.4

1.8

323.9

3.6

15,m

ancore

[1]

11

933

7.57

2.4

13806

0.05168

25

0.3768

21

0.05288

09

324.8

1.6

324.7

1.6

323.6

3.7

16,m

anrim

[1]

27234

7.01

6.3

7408

0.05164

29

0.3764

23

0.05287

09

324.6

1.8

324.4

1.8

323.1

3.7

17,m

sbA

[1]

44030

7.18

2.2

23541

0.05156

26

0.3753

21

0.05280

08

324.1

1.6

323.1

1.6

320.0

3.5

BW-90L

18,m

eu[1]

112730

2.90

3.1

13420

0.05222

32

0.3805

25

0.05284

09

328.2

2.0

327.4

1.8

321.8

3.7

19,m

euA

[1]

21558

2.74

3.0

3415

0.05197

26

0.3797

22

0.05298

13

326.6

1.6

326.8

1.6

328.1

5.7

20,m

euA

[1]

410310

2.78

8.9

15019

0.05186

31

0.3780

25

0.05286

09

326.0

2.0

325.6

1.8

322.7

3.9

21,m

eu[1]

12632

7.03

0.9

9312

0.05173

24

0.3768

20

0.05283

11

325.2

1.4

324.7

1.4

321.4

47

22,zprism

[1]

1241

0.34

0.8

952

0.05186

26

0.3795

40

0.05307

46

325.9

1.6

326.7

3.0

332

20

am

Monazite,zzircon,ttitanite,

eueuhedral,sb

subhedral,ananhedral,eq

equant,f¯at,fr

fragments,incl

inclusion,lbrlight-brown,A

abraded,[1]number

ofgrainsanalysed

bWeightsknownto

betterthan10%

when

over

10

lg,andto

about50%

when

less

than2

lg;accuracy

ofU

andThconcentrationsisroughly

proportionalto

uncertainty

ofsampleweight

cModel

Th/U

ratioestimatedfrom

208Pb/206Pbratioandageofsample

dTotalcommonPbin

sample,includes

initialandblankPb

eMeasuredratio,correctedforfractionationandspikecontribution

fCorrectedforspike,

fractionation,blankandinitialcommonPb(StaceyandKramers1975)

g2runcertainty

calculatedbyerrorpropagationprocedure

thattakes

into

accountinternalmeasurementstatisticsandexternalreproducibilityaswellasuncertainties

intheblankandcommonPb

correction;errors

are

given

asthelast

decim

alplacesoftherespectivevalues

7

at very high temperatures (Kalt et al. 1999). Biotite in-clusions in garnet may potentially record older stages ofmetamorphism or earlier metamorphic events with the40Ar/39Ar chronometer (Kelley et al. 1997) through be-ing armoured from Ar diusion. Therefore, apart fromsingle grains very thin, polished rock sections were pre-pared for dating biotite inclusions in garnet and biotitegrains in the matrix. Biotites from migmatite sampleshave rather uniform compositions characterised byfairly low XMg values and K2O contents >9 wt%. Thelargest variation is found in TiO2 (3.82±5.12 wt%,Table 2). Biotite in gneiss sample BW-132 is in equilib-rium with amphibole, plagioclase and quartz. It displayslower TiO2 contents and higher XMg values compared tobiotite from migmatites (Table 2).

K-feldspar occurs in mesosomes and in leucosomesof the migmatites. K-feldspars in leucosomes clearlycrystallised from a melt during stage C. Conversely,their role in mesosomes is not clear as in generalK-feldspar may be either consumed or formed bydehydration melting reactions depending on the K2O/H2O ratio of the melt (Carrington and Watt 1995).However, K-feldspars in all samples are perthitic andthus record equilibration above approximately 700 °C.Above this temperature and at the relevant pressures,K-feldspar is generally homogeneous monoclinic,whereas below approximately 700 °C, spinodal decom-position of domains to a triclinic albite phase starts(Parsons and Brown 1991 and references therein). Theexsolution lamellae observed in K-feldspars from mig-matites of the Bayerische Wald have several scales. Witha polarising microscope, parallel or subparallel lamellaeof 50±250 lm widths can be observed. Back-scatteredelectron microprobe images reveal further exsolutionlamellae of 1±5 lm widths. In thin sections, beginningformation of braid perthite can be recognised, resultingfrom rotation, broadening and coalescence of the albitelamellae. These transformations occur at and below400 °C (Parsons and Brown 1991 and references therein)and are generally ascribed to slow cooling of the feld-spars. While K-feldspars of samples BW-67, BW-120and BW-116 only have minor inclusions, those in sam-ples BW-44 and BW-46 contain numerous inclusions ofplagioclase. All K-feldspar grains selected for datingcome from leucosomes.

U±Pb dating

Analytical techniques

The U±Pb isotopic measurements were conducted at theJack Satterly Geochronological Laboratory, Depart-ment of Earth Sciences, Royal Ontario Museum inToronto. Minerals were separated by crushing with ajaw-crusher, pulverisation with a disk mill, heavy min-eral enrichment on a Wil¯ey table, and subsequentmagnetic separation on a Frantz isodynamic separatorand density separation using heavy liquids. The minerals

to be analysed were separated under a binocular mi-croscope and in part air-abraded (Table 3) following thetechnique of Krogh (1982). Single grains or single frag-ments of grains were dated except for zircon analysis ofsample BW-70L (two grain fragments, Table 3) andtitanite analyses (>30 grains, Table 3).

After a ®nal selection, the minerals were washed inca. 4 N HNO3 on a hotplate and rinsed with H2O andacetone. A mixed 205Pb/235U spike was used for U±Pbanalyses of zircon, titanite and monazite. The spike wasadded to the sample after weighing and transfer to thedissolution vessel. Zircon was dissolved in HF(+HNO3) in Te¯on mini-bombs at ca. 190 °C, monazitewas dissolved in 6 N HCl in Savillex vials on a hotplate,and titanite was dissolved in HF (+HNO3) using Sa-villex vials on a hotplate. The solutions were subse-quently evaporated, redissolved in 3.1 N HCl andpassed through anion exchange resin in minicolumns inHCl medium to purify U and Pb (zircon and monazite).For titanite, a more complex HCl±HBr±HNO3 proce-dure was necessary to purify U and Pb. Blanks were lessthan 2 pg Pb and 0.1 pg U for zircon and monazite and10 pg Pb and 0.5 pg U for titanite.

Pb and U were collected together from the columns,loaded on outgassed Re-®laments together with H3PO4and Si-gel, and run on a VG354 mass spectrometer usinga Daly detector. Daly±Faraday conversion was 0.04%/a.m.u. Fractionation factors for U and Pb correspond to0.1%/a.m.u.

General features of the monazite and zircon data

The results for 28 monazite, 2 zircon and 3 titanite an-alyses are given in Table 3 and presented in Concordiadiagrams (Figs. 2±4). The data patterns for the mona-zites reveal various degrees of complexity and repeatedanalyses were carried out to verify the reproducibility ofthe ages and explore possible causes for the deviation ofsome of the samples.

All the monazite data plot on or slightly above theconcordia curve, a common observation in monazite,re¯ecting on the one hand the resistance of monazite tolead loss, and on the other hand the presence of excess206Pb resulting from the incorporation of excess 230Th atthe time of formation (SchaÈ rer 1984; Parrish 1990). Al-though excess 230Th mainly aects the U±Pb systematicsof very young samples it may as well lead to severalpercent of excess 206Pb in older samples provided thefractionation factor (f (Th/U)mineral/(Th/U)reservoir) islarge enough (SchaÈ rer 1984). The excess 206Pb results in206Pb/238U ages that are too high and 207Pb/206Pb ratiosthat are too low. A correction for the initial disequilib-rium is generally possible but was not applied in this caseas the Th/U ratio of the initial reservoir open to U, Thand Pb exchange with monazite is not known. Withoutcorrection, the most accurate estimate of the ages can beobtained from 207Pb/235U ratios which are not aectedby the initial disequilibrium. Hence, all the monazite

8

ages reported in Figs. 2±4 and referred to in the text arebased on weighted averages of 207Pb/235U.

Zircon commonly incorporates very little Th at thetime of formation and hence remains essentially unaf-fected by disequilibrium. In order to date only zirconparts grown during HT±LP metamorphism, very smallzircon fragments had to be used that contained very littleU and Pb, aecting the precision of the 207Pb/235U and207Pb/206Pb ages through the common lead correction.Therefore, the 206Pb/238U ages are reported for zirconanalyses. Titanite also incorporates only very little Thand is characterised by low U and Pb abundances.Therefore, 206Pb/238U ages are also indicated for titanite.

Ages of monazite and zircon from migmatitesand gneisses

For sample BW-70, ®ve analyses of monazite from theleucosome cluster tightly on or slightly above Con-cordia, yielding an average 207Pb/235U age of

321.8 0.7 Ma (Fig. 2a). Analysis of two abradedzircon tips yields a concordant data point and an over-lapping 206Pb/238U age of 323.4 1.6 Ma (Fig. 2a).More complex relations are observed for the mesosome(Fig. 2b). Three of the monazite analyses yield over-lapping results at a mean 207Pb/235U age of323.3 1.0 Ma. One analysis (no. 5, Table 3) yields ayounger age of 317.2 1.6 Ma (Fig. 2b). This graincontained several inclusions that probably contributedto the elevated initial common Pb (Table 3), but theireect on the age is not known. Another analysis (no. 1,Table 3) provides a reversely discordant but much olderage (Fig. 2b). This grain was enclosed in garnet.

For sample BW-90, analyses of four monazite grainsfrom the leucosome yield a weighted average 207Pb/235Uage of 326.1 1.7 Ma (Fig. 3a). The scatter of thedata, however, exceeds analytical uncertainty. Analysisof a zircon prism from the leucosome yields a concor-dant data point and a 206Pb/238U age of 325.9 1.6 Ma(Fig. 3a), identical to the monazite age. Five analyses ofmonazite from the mesosome are clustered tightly on or

Fig. 2a, b U±Pb concordia diagram showing the results of U±Pbdating of monazite and zircon from migmatite sample BW-70 oflocation 1. a Leucosome; bmesosome. Errors are given at the 2r level.For further explanation see section on general features of the zirconand monazite data and section on ages of monazite and zircon frommigmatites and gneisses

Fig. 3a, b U±Pb concordia diagram showing the results of U±Pbdating of monazite and zircon from migmatite sample BW-90 oflocation 7. a Leucosome; bmesosome. Errors are given at the 2r level.For further explanation see section on general features of the zirconand monazite data and section on ages of monazite and zircon frommigmatites and gneisses

9

slightly above the Concordia curve (Fig. 3b). The aver-age weighted 207Pb/235U age is 325.3 1.9 Ma, but thescatter of the data is beyond analytical uncertainty. Asixth analysis (no. 12, Table 3) of a simple prismaticcrystal gave a signi®cantly older age of 340 Ma. Some ofthe monazite grains of BW-90 mesosome showedapparent cores and overgrowths. One of these grainswas broken into core and rim fragments (nos 15±16,Table 3) to test for a possible inherited origin of thecore. However, both parts of the grain yielded identicalages (Table 3).

In migmatites BW-120 and BW-116, mesosome andleucosome could not be separated. Four monazite grainsfrom sample BW-120 yield data points that overlap onor slightly above the Concordia curve with a mean207Pb/235U age of 321.9 0.8 Ma (Fig. 4a). The resultsfor four monazite analyses of sample BW-116 are morecomplex (Fig. 4b). Three of the analyses are concordantbut have 207Pb/235U ages ranging from 337 to 323 Ma.Another analysis (no. 28, Table 3) is slightly discordant(Fig. 4b). There is no obvious morphological or colourdierence between the four monazite grains. The

207Pb/235U age of the youngest monazite is322.9 1.8 Ma.

Ages of titanite from amphibolites

Two analyses were performed on titanite of sampleBW-134, one of them using fragments that were abra-ded, the other using ¯at spheroidal crystals. Both showlow U contents of 16±21 ppm and provide identical206Pb/238U age values of 321.2 2.7 Ma. Although thelow 206Pb/204Pb ratios of about 70 make the age cal-culation quite sensitive to the choice of initial Pb, thevariation introduced by this correction on the206Pb/238U age is minimal. For example, the use ofStacey and Kramers (1975) model Pb compositions,calculated for ages spanning almost the entire Paleozoicand Mesozoic, yields a range of 206Pb/238U valuescovered almost entirely by the quoted analytical un-certainty. The titanite population of sample BW-118consists of anhedral grains with irregular, contortedshapes. As the U content and the proportion of ra-diogenic Pb are much higher than in sample BW-134(Table 3), the 206Pb/238U age of 320.6 1.6 Ma ismuch more precise, and much less dependent on thechoice of initial Pb. In this case a Paleozoic±Mesozoicrange of model corrections produces a variation of lessthan half the analytical error.

40Ar/39Ar dating

Analytical techniques

The 40Ar/39Ar isotopic measurements were conductedusing the VULKAAN argon laserprobe at the Facultyof Earth Sciences, Vrije Universiteit Amsterdam. Forincremental heating experiments, mineral separates wereobtained by carefully crushing ®nger- to hand-sizedsamples, and sieving through 500 lm, 250 lm and125 lm mesh fractions. The mineral separates werewashed with deionised water, separated magneticallyand handpicked. For an incremental heating experi-ment, 1±3 grains of biotite, 60 grains of hornblende and1 grain of K-feldspar were used, respectively. The sep-arates were put into Al tablets containing 20, 2-mmholes. On each tablet, 16 positions were loaded withsamples and 4 with a ¯ux monitor of known age (stan-dard DRA-1, 24.99 Ma). For spot fusion experiments onbiotite and K-feldspar, polished sample sections of2 ´ 2 cm, 90 lm thick, were prepared. Their corners andedges were then broken to ®t the sections onto an Altablet. Additional tablets containing only the standardwere also prepared. All tablets were wrapped in Al foiland loaded into cylindrical Al containers, with samplesections and standard tablets alternating. The containerswere then irradiated in the Triga Reactor at the OregonState University for 24 (®rst sample charge) and 18 h(second sample charge).

Fig. 4a±c U±Pb concordia diagrams showing the results of U±Pbdating of monazite and titanite. a Monazite from migmatite sampleBW-120 (location 5); b monazite from migmatite sample BW-116(location 6); c titanite from amphibolite samples BW-134 (location 4)and BW-118 (location 5). Errors are given at the 2r level. For furtherexplanation see section on general features of the zircon and monazitedata and section on ages of monazite and zircon from migmatites andgneisses

10

The VULKAAN argon laserprobe consists of a 24 Wargon ion laser (visible light, 488±524 nm), beam optics,a low volume UHV gas inlet system, and a MAP 215-50noble gas mass spectrometer. Full technical and ana-lytical laserprobe dating procedures were described byWijbrans et al. (1995). The isotopic composition of Arwas measured in static mode. Intensities were recordedwith a secondary electron multiplier detector usingswitchable pre-ampli®er resistor settings (10, 100,1000 MW) and peak jumping at half mass intervals frommass 40 down to mass 35.5. All measured values werecorrected for mass discrimination, 37Ar and 39Ar decay,neutron-induced interferences from Ca and K, andprocedural blanks. Correction factors were as quoted inWijbrans et al. (1995). The J values (irradiation pa-rameter) were calculated by measuring replicate stan-dards from dierent locations within the tablets and thecontainers and by ®tting a function through the datapoints, respectively. Ages were calculated using thedecay constants recommended by Steiger and JaÈ ger(1977). The quoted errors on the ages (2r) includethe uncertainty on the irradiation parameter J, theuncertainty on the mass spectrometric analyses of the Arintensities, errors on the blanks and the mass interfer-ences of radiogenic Ar with Ar derived from K, Ca andCl. The uncertainty of a plateau age is calculated fromthe uncertainties of the individual steps, using the in-verse variance of the 40Ar/39Ar ratios as a weightingfactor.

Incremental heating experiments were performed byprogressively degassing the sample grains using a defo-cused laser for 60 s at successively greater laser powerincrements. System blanks were measured before the®rst experiment and every ®fth analysis thereafter.The blank correction applied to each sample usedthe blank integrated over the time of sampleanalyses. Typical system blanks were 3.92±6.63 ´ 10)17,5.18±8.91 ´ 10)19, 8.91±33.7 ´ 10)20, 2.61±3.23 ´ 10)18and 5.87±12.6 ´ 10)19 mol for masses 40, 39, 38, 37 and36, respectively. Spot fusion experiments on biotite andK-feldspar in sample sections were performed using a2 W laser beam focused at 20 lm for approximately 2 s in0.5 s pulses. System blanks were measured before eachsample and every third analysis thereafter. Typical systemblanks were 0.93±2.21 ´ 10)16, 1.03±1.60 ´ 10)18, 7.59±8.02 ´ 10)19, 1.78±3.13 ´ 10)18 and 1.61±1.78 ´ 10)18mol for masses 40, 39, 38, 37 and 36, respectively.

40Ar/39Ar ages

The results of incremental heating and spot fusion ex-periments are given in Table 4. Representative agespectra obtained by incremental heating on hornblende,biotite and K-feldspar are shown in Figs 5, 6 and 7,respectively. Age plateaus are de®ned by at least threeconsecutive steps that overlap at the 95% con®dencelevel (exclusive of uncertainty in J) and together containat least 50% of the 39ArK released by the sample. All

samples were additionally examined on inverse isochrondiagrams (36Ar/40Ar vs 39Ar/40Ar), in which the age isgiven by the intercept on the 39Ar/40Ar axis and thecomposition of any trapped non-radiogenic argon by theintercept on the 36Ar/40Ar axis. In principle, the inverseisochron approach allows to determine whether thetrapped argon is air-derived and/or excess argon.However, because the samples analysed here are fairlyold and potassium-rich (except for hornblende), most ofthe data plot on or near the 39Ar/40Ar axis. Therefore,the 36Ar/40Ar intercepts have large uncertainties andgive little information on initial ratios. The ages given bythe 39Ar/40Ar intercepts generally coincide with theplateau and total fusion ages (Table 4).

Hornblende from three dierent locations (Fig. 1,Table 4) was analysed by incremental heating experi-ments. The separates yield plateau ages between321.9 3.3 and 313.7 3.7 Ma (Table 4). Within er-ror limits, there is good agreement between the plateauages and the integrated total fusion ages (Table 4). Thedierent plateau ages are not related to dierent samplelocations as almost the entire range of ages is displayedby hornblende from location 4 (315.6 3.5 to321.9 3.3 Ma, Fig. 5). A common characteristic ofthe hornblende age spectra is that the ®rst, usually verysmall, heating step released argon with apparent agesthat are considerably higher or lower than the plateauages. The contribution of air argon in these ®rst steps ofincremental heating is comparatively large and the de-viation of the apparent ages is most likely induced bythe release of argon trapped on cracks or cleavageplanes.

Biotite from all locations was analysed by incremen-tal heating experiments on one or two grains, respec-tively. The obtained plateau ages range between309.6 3.5 and 313.2 3.2 Ma (Fig. 6) and coincidewith the integrated total fusion ages within error limits.Most spectra are characterised by very even plateauswith minor age deviations only in the ®rst heating steps.As for hornblende, there are no signi®cant age dier-ences between the locations as the values overlap withinerror limits. Additionally, total fusion ages of singlebiotite grains from sample sections were obtained forsome of the locations (Table 4). Garnet zoning patternsrevealed that the inner cores of large garnets record theoldest metamorphic stage (Kalt et al. 1999). Therefore,biotite inclusions in garnet were analysed by spot fusionexperiments and their mean ages were calculated sepa-rately from those obtained on matrix biotites (Table 4).However, no signi®cant age dierence between biotitefrom these two dierent textural positions is recognis-able in any sample, and the mean spot fusion agescoincide with those obtained by incremental heating inevery sample (Table 4). The spread in ages obtained byspot fusion experiments on single grains from samplesections is inmost cases large, expressed by large errors onthe mean ages (Table 4). This is probably because biotitegrains were often rather small and some of the gas mayhave come from adjacent minerals or grain boundaries.

11

Table

4Resultsof40Ar/39Ardating

Location

Sample

Mineralproperties

aIncrem

entalheating

Inverse

isochron

Spotfusion

J

nb

MSWD

39Ar

[%]c

PLdage

[Ma]

+/)

[2r]e

TFfage

[Ma]

+/)

[2r]e

nb

MSWD

40Ar/36Ar

intercept

+/)

[2r]e

Age

[Ma]

+/)

[2r]e

ng

Ageh

[Ma]

+/)

[2r]i

1BW-22

bt,250±500

lm[2]

10

1.8

100.0

311.7

3.2

311.0

5.4

10

0.7

170.7

48.9

313.0

3.0

0.006161

BW-28

bt,in

matrix

6306.2

6.1

0.005569

2BW-30

bt,250±500

lm[2]

11

2.4

96.8

311.6

3.5

310.3

5.7

11

2.4

173.0

174.5

312.4

5.5

0.006160

BW-32

bt,incl

ingrt

5313.6

1.8

0.005427

BW-32

bt,in

matrix

3314.8

4.8

0.005427

BW-67

bt,incl

ingrt

8312.2

7.9

0.005259

BW-67

bt,incl

ingrt

1311.4

1.6

0.005259

BW-67

Kfs,250±500

lm[1]

10

109.2

100.0

299.3

j2.9

292.2

2.9

10

109.2

)1099.5

1785.5

302.6

10.25

0.005259

BW-67

Kfs,250±500

lm[1]

10

203.2

299.9

300.0

j2.9

291.7

2.9

9271.6

949.4

*k

285.9

224.91

0.005259

386246

bt,250±500

lm[1]

50.5

99.7

312.0

3.1

312.1

3.9

50.4

330.0

145.6

311.9

3.2

0.006097

86246

bt,250±500

lm[2]

70.3

96.0

312.6

6.2

308.2

24.8

11

0.5

181.3

346.1

313.7

4.9

0.006160

4BW-132

bt,250±500

lm[3]

81.3

99.1

309.6

3.9

311.0

8.9

90.03

40771.1

*267.2

311.4

0.006161

BW-132

hbl,125±250

lm[60]

10

3.1

97.7

321.9

3.3

323.6

4.1

11

5.6

453.2

124.5

314.6

9.4

0.004425

BW-134

hbl,125±250

lm[60]

81.0

100.0

319.1

3.0

318.8

3.6

80.9

271.7

29.5

319.5

3.0

0.004425

BW-134

hbl,125±250

lm[60]

811.8

99.1

315.6

3.5

315.4

3.6

12

8.0

267.3

199.8

316.4

7.5

0.004425

5BW-118

hbl,125±250

lm[60]

91.8

98.4

313.7

3.7

316.3

7.2

10

3.3

358.6

75.9

308.8

9.6

0.004425

BW-120

bt,250±500

lm[2]

50.3

92.5

311.8

3.0

310.8

3.7

61.7

159.2

45.7

312.4

3.1

0.006097

BW-120

Kfs,250±500

lm[1]

945.8

82.8

296.4

3.0

296.3

3.2

10

14.5

878.3

1816.2

291.5

20.2

0.006136

6BW-112

hbl,125±250

lm[60]

95.4

98.8

320.3

3.2

318.9

3.5

10

18.2

119.3

190.4

322.6

8.3

0.004425

BW-116

bt,250±500

lm[3]

40.7

69.6

309.7

4.3

303.6

11.4

81.7

138.3

93.1

311.2

4.6

0.006161

BW-116

Kfs,250±500

lm[1]

720.3

87.7

305.3

3.8

306.9

3.3

82.2

2424.7

2390.5

296.9

13.0

0.006136

7BW-44

bt,in

matrix

8315.8

10.8

0.005364

BW-44

bt,incl.in

Kfs

2306.2

2.9

0.005364

BW-44

Kfs,250±500

lm[1]

11

24.1

84.3

305.1

3.6

307.2

3.2

12

14.0

518.8

358.9

302.8

11.2

0.005364

BW-46

bt,250±500

lm[2]

80.8

100.0

313.2

3.2

313.0

5.3

80.7

172.6

92.8

314.5

3.3

0.006097

BW-46

bt,250±500

lm[2]

12

1.7

100.0

311.5

3.7

312.1

8.7

12

1.6

108.2

119.7

314.4

3.5

0.006161

BW-46

Kfs,250±500

lm[1]

92.8

96.5

319.7

3.1

318.4

3.4

12

0.2

20010.0

*269.5

70.1

0.006136

BW-46

Kfs,250±500

lm[1]

32.3

84.0

317.8

3.1

318.0

3.3

95.7

3047.0

*297.6

39

0.006136

BW-90

bt,in

matrix

4316.6

2.8

0.005154

BW-90

bt,incl.in

grt

7302.0

13.9

0.005154

aMineralabbreviationsaccordingto

Kretz

(1982);incl.,inclusion;[1],number

ofgrains

bnnumber

ofheatingstepsincluded

intheplateauagecalculation

cPercentageof39Arincluded

intheplateauagecalculation

dPL

Plateau

eQuotederrors

includetheuncertainty

ontheirradiationparameter

J,theuncertainty

onthemass

spectrometricanalysesoftheArintensities

(includingerrors

ontheblanks),themass

interferencesofradiogenic

ArwithArderived

from

K,CaandClandtheerrorontheweightedmeanoftheindividualsteps(D

alrymple

andLanphere1971)

fTFTotalfusion

gnNumber

ofsingle

grainsanalysed

hMeanagecalculatedfrom

thetotalfusionages

ofsingle

grains

iStandard

deviationofthemeanage

jApparentageofthe®nalheatingstep

(noplateauagecould

becalculated)

kAsteriskindicateserrortoolargeto

beadequately

displayed

12

Spot fusion ages of K-feldspar from sample sectionscould not be obtained. It was not possible to fuse thespots by hitting them with the focused laser beam forseveral seconds. As longer laser treatment would havethermally activated other parts of the section, the spotfusion experiments on K-feldspar were not continued.K-feldspar was broken from the sections and fragmentsof grains were then analysed by incremental heatingexperiments. Incremental heating experiments wereperformed on single K-feldspar fragments from loca-tions 1, 5, 6 and 7 (Table 4). In general, the age spectraobtained from K-feldspar are not as even as those ofbiotite and hornblende (Fig. 7). Plateau ages that in-clude more than 80% of the released Ar and that coin-cide with the integrated total fusion ages could becalculated for samples BW-44, BW-46 and BW-116. Fora K-feldspar of sample BW-120, a plateau age includingonly 60% of the released Ar, but coinciding with theintegrated total fusion age, could be calculated. The twoK-feldspars of sample BW-67 are characterised by a

gentle staircase-like increase in apparent ages and by thelack of plateaus. The integrated ages are in goodagreement with the total fusion ages for each sample,whereas the apparent ages of the ®nal degassing step areslightly higher (Table 4). The K-feldspar plateau ages(ages of the ®nal degassing step for sample BW-67) formtwo distinct groups: K-feldspars from locations 2, 5 and6 have ages of 299.3 2.9, 300.0 2.9, 296.4 3.0and 305.3 3.8 Ma, the latter being identical to an ageof 305.1 3.6 Ma of one K-feldspar from locality 7.Two other K-feldspars from location 7 have consider-ably higher ages of 319.7 3.1 and 317.8 3.1 Ma.

Signi®cance of mineral ages and link to the P±T path

U±Pb zircon ages

Microstructural data and modes show the leucosomes ofmigmatites from the Bayerische Wald to have either been

Fig. 5a±c Results of incremental heating experiments of horn-blende separates from amphibolite samples BW-132 (a), BW-134(b) and BW-112 (c). Errors are given at the 2r level. For furtherexplanation see section on 40Ar/39Ar dating

Fig. 6a±c Results of incremental heating experiments of biotitefrom migmatite samples BW-22 (a), BW-120 (b), and BW-46 (c).Errors are given at the 2r level. For further explanation see sectionon 40Ar/39Ar dating

13

pure melt (Qtz±Kfs leucosomes) or melt and residualcrystals (Plg±Qtz±Kfs leucosomes, Berger and Kalt1999). Therefore, the analysed zircon tips from the leu-cosomes must have coexisted with melt at some stage.The euhedral shapes of the zircons strongly suggest thatthe tips grew during HT±LP metamorphism by precipi-tation from a melt. This interpretation is supported bythe fact that most crustal lithologies contain a sucientamount of zirconium (Zr) and light rare earth elements(LREE) to saturate a melt in these components and allowprecipitation of accessory phases such as zircon andmonazite (see Watt and Harley 1993 for a discussion).The case may be dierent in small-volume disequilibriummelts (Sawyer 1991), but estimates of former melt volumefractions in migmatites of the Bayerische Wald yieldvalues of 20±40% (Berger and Kalt 1999). The core-rimstructures of zircon grains visible under the binocularmicroscope probably represent residual zircons that wereentrained in the partial melt formed during HT±LP

metamorphism, partly dissolved and later overgrownduring zircon precipitation. Hence, the concordant U±Pbages of the zircon tips must be crystallisation ages asopposed to resetting ages. The ages do not record cool-ing. This interpretation is generally accepted for zircons,based on low diusivities for U, Pb and Th (Cherniaket al. 1997; Lee et al. 1997) and very high closure tem-peratures for Pb in zircon (Lee et al. 1997), implying thatvolume diusion is not the dominant process for Pb lossin zircon (Mezger and Krogstad 1997).

Zircon incorporates large amounts of incompatibleelements (Zr, Hf, Y, U, LREE, e.g. Hanchar and Miller1993). In the metapelitic migmatites of the BayerischeWald, these incompatible elements became available assoon as in situ melting started during prograde stage Bdehydration reactions through the dissolution of existingzircon (Watson 1996). The numerical simulations ofWatson (1996) show that diusion-controlled zircondissolution in melts of granitic composition at increasingtemperatures is a continuous process whereby dissolu-tion rates increase exponentially with temperature in therange 650±850 °C. Diusion-controlled zircon growth ingranitic melts is also favoured by increasing tempera-

Fig. 7a±f Results of incremental heating experiments of K-feldsparfrom migmatite samples BW-67 (a), BW-120 (b), BW-116 (c), BW-44(d) and BW-46 (e±f). Errors are given at the 2r level. For furtherexplanation see section on 40Ar/39Ar dating

14

tures (Watson 1996). Therefore, the zircon tips used fordating here should re¯ect growth beginning at peakmetamorphic conditions, estimated to be about 850 °C(Kalt et al. 1999), down to temperatures at which allmelt crystallised. This was at the end of stage B andduring stage C (Fig. 1). Experiments in metapelitic sys-tems indicate 800 °C as approximate minimum temper-ature for biotite dehydration melting (e.g. Le Breton andThompson 1988; Vielzeuf and Montel 1994) and ther-mobarometry on subsolidus stage D phase assemblagesyields 770±846 °C as a minimum temperature for meltcrystallisation in migmatites of the Bayerische Wald(Kalt et al. 1999). Therefore, the U±Pb ages of zircontips date a minimum temperature of approximately800 °C. This temperature is still below common esti-mates of Pb closure temperature and thus supports theinterpretation as crystallisation ages (Lee et al. 1997).

U±Pb monazite ages

In general, the interpretation of U±Pb ages is not asstraightforward for monazites as it is for zircons becausemonazite closure temperatures for U, Th and Pb are stilla matter of debate. Comparatively high temperatures of³750 °C are suggested by experimental diusion data(Smith and Giletti 1997) and crystal-chemical consider-ations (Dahl 1997). Slightly lower minimum estimates of680±720 °C emerge from the presence of inherited mo-nazites in leucogranites (Copeland et al. 1988; Kings-bury et al. 1993; Harrison et al. 1995), from partialresetting of monazites at amphibolite-facies conditions(Parrish et al. 1990) and from electron microprobe ana-lyses of zoned monazites (Suzuki et al. 1994). However,slow diusion velocities, the preservation of radiogeniclead in inherited monazite grains from magmatic andmetamorphic rocks and postmagmatic monazite re-crystallisation (Hawkins and Bowring 1997) argue thatrecrystallisation rather than volume diusion controlsthe U±Pb systematics of monazite. This view is alsosupported by the presence of very sharp boundaries inline scans of Th, U and Pb across granulite-facies mo-nazites (Zhu and O'Nions 1999). Peak metamorphictemperatures in migmatites of the Bayerische Wald areabove some of the assumed U±Pb closure temperaturesfor monazite. However, a number of observations sug-gests that the obtained U±Pb monazite ages re¯ectcrystallisation and inheritance rather than monazite re-setting and/or cooling through a closure temperature.

The euhedral shape of the monazites from leuco-somes suggest that at least their rims record growth inthe presence of a melt and not resetting of older crystals.The presence of excess 206Pb (see section on U±Pb dat-ing) in monazites from leucosomes and mesosomes is thestrongest argument for the monazite ages re¯ectingcrystallisation. If resetting of older monazites by diu-sion had occurred during HT±LP metamorphism, thisdisequilibrium eect would not have been preserved.Both mesosome samples (BW-70 and BW-90) contain

simple, euhedral to subhedral inherited monazite grainswith distinctly higher ages (Figs 2b, 3b). In mesosomeBW-70, the inherited monazite was included in earlygrown garnet that shielded it from recrystallising ordissolving in a partial melt during HT±LP metamor-phism. This eect was described by DeWolf et al. (1993)for monazites from the high-grade core of the WindRiver Range, Wyoming. In mesosome BW-90, the mo-nazite carrier phase is not known. The rest of themonazites from the mesosome of sample BW-70 clusteron concordia due to the absence of inherited monazitedomains, while those from the mesosome of sampleBW-90 show a scatter that can be best explained byinheritance of older concordant monazites.

The U±Pb systematics of monazites from migmatiteBW-116 (leucosome and mesosome could not be sepa-rated) most clearly reveal the eects of inheritance(Fig. 4). Three concordant grains with ages between 323and 337 Ma as well as one slightly discordant graindemonstrate that monazite crystallised at 322.9 1.8 Ma during HT±LP metamorphism may containvarying proportions of older monazite with either undis-turbed or partially reset U±Pb systems. Monazites ofsample BW-120 reveal a simpler picture (Fig. 4a) and arebest interpreted as re¯ecting crystallisation withoutinheritance. Future study on monazite composition mayhelp to distinguish inherited and newly crystallised parts.

In summary, the monazite U±Pb data reveal crys-tallisation at the same time as the precipitation of zirconfrom partial melts and must hence record virtually thesame minimum temperature of approximately 800 °C.Monazites in all samples except BW-120 show inheri-tance of older monazite to varying degrees. The fact thatthese older monazites survived metamorphic conditionsand partial melting at temperatures of approximately850 °C places new constraints on the U±Pb closuretemperature of monazite. There is a slight trend towardshigher U±Pb ages of zircon and monazite from thenorthwestern area around Cham. Considering the factthat the monazites from all samples are eected byvariable degrees of inheritance of older grains, it ispossible that the slightly higher U±Pb monazite agesfrom the northwestern area re¯ect inheritance. However,zircon tips that should be free of inherited componentsare also older in the northwestern area, suggesting thatthe age dierence is geologically meaningful. The veryyoung U±Pb age of 317 Ma for one monazite grain ofsample BW-70 cannot be reasonably explained at themoment. Perhaps it is due to local late-stage hydro-thermal features, although a U±Pb isotope study onhydrothermally treated monazites suggests that mona-zite becomes discordant due to lead loss under theseconditions (Teufel and Heinrich 1997).

U±Pb titanite ages

Several U±Pb dating studies document that titanite mayremain closed to Pb diusion at elevated temperatures,

15

may be reset during later metamorphism or may re-crystallise even at low temperatures (Corfu and Grunsky1985; Corfu et al. 1994; Corfu 1996). Other studiessuggest that while at temperatures above 700 °C titaniteis aected by Pb diusion, recrystallisation is the dom-inant process below approximately 700 °C (Scott and St-Onge 1995; Verts et al. 1996). Calculations of U±Pbclosure temperatures from magmatic titanite with in-herited Pb components yield minimum estimates of712 °C (Zhang and SchaÈ rer 1996). These estimates co-incide quite well with calculations from diusion ex-periments (Cherniak 1993). As titanite in theamphibolites of the Bayerische Wald is not only in-cluded in amphibole but also in plagioclase that wasformed during earlier granulite-facies metamorphism, itmost likely crystallised at temperatures (850 °C) con-siderably exceeding any estimates of closure tempera-ture. Thus, the U±Pb ages of titanites most likely do notrecord crystallisation but cooling to temperatures ofapproximately 700 °C.

40Ar/39Ar ages

Diusion of Ar in amphibole, biotite and K-feldspar iscomparatively fast, even at fairly low temperatures (e.g.Baldwin et al. 1990; Lovera et al. 1991; Dahl 1996 andreferences therein). As in many metamorphic terranesthe attained peak temperatures are considerably higherthan the closure temperatures calculated from the Ardiusion velocities, K-Ar and 40Ar/39Ar ages of am-phibole, biotite and K-feldspar can be interpreted ascooling ages, provided that thermally induced volumediusion according to Fick's Law is the dominatingprocess and that the minerals did not form below theirclosure temperatures. There are also other mechanismsof Ar loss or exchange such as ¯uid- or deformation-assisted recrystallisation (e.g. Wijbrans and McDougall1986; Villa 1998; Parsons et al. 1999), but these mainlyapply to Ar-bearing minerals that were substantiallyaected by hydrothermal overprint (metamorphism)and/or deformation after their crystallisation. Thesecomplex histories usually result in disturbed Ar spectra.In the Bayerische Wald, migmatites and amphibolitesexperienced one single HT-metamorphic event, peakmetamorphic temperatures approximated 850 °C, theinvestigated rocks and minerals were not subject to late-stage ¯uid- or deformation-assisted recrystallisation (seesections on geological setting and on characteristics ofthe phases), and apart from two feldspars, all analysedAr-bearing minerals yielded good plateaus (Figs. 5±7).Therefore, it seems reasonable to interpret the 40Ar/39Arages as recording cooling below closure temperatures.

For amphiboles it has been discussed whether the Arretentivity and thus the closure temperatures depend oncomposition. Whereas several geochronological studiesindicated the possibility of a compositional dependence(e.g. O'Nions et al. 1969; Berry and McDougall 1986;Onstott and Peacock 1987), experimental determination

of Ar diusion velocities in metamorphic hornblendeshowed no correlation between XMg and activation en-ergy (Baldwin et al. 1990). On the basis of the ionicporosity model of Fortier and Giletti (1989), Dahl(1996) argued that there should be dierences in closuretemperature due to compositional eects, but that thesecompositional eects would probably become lostwithin analytical uncertainty once the cooling rate ex-ceeded 10 °C/my. The amphiboles examined here byincremental heating experiments are magnesio-horn-blende (BW-112), edenitic hornblende (BW-134) andactinolitic hornblende (BW-118, BW-132). The calcula-tions of Dahl (1996) suggest that pure edenite shouldhave higher closure temperatures (600±520 °C) thanpure ferro-actinolite (440±470 °C) at cooling rates of10±100 °C/my, which are realistic for the BayerischeWald (see section on cooling rates). This compositionaleect cannot be seen in the 40Ar/39Ar hornblende totalfusion ages obtained here (Tables 2 and 4), because theinvestigated amphiboles have intermediate, but notendmember compositions. Magnesio-hornblende andedenitic hornblende (320.3 3.2, 319.1 3.0,315.6 3.5 Ma) do not yield signi®cantly higher agesthan actinolitic hornblende (321.9 3.3, 313.7 3.7 Ma). Therefore, combining the hornblende compo-sitions of the samples with the ionic porosity model ofDahl (1996), it can be assumed that the hornblende agesobtained record cooling of the amphibolites belowapproximately 500±570 °C.

As for amphibole, it has been discussed whether theclosure temperatures of biotite may vary with compo-sition. Whereas Harrison et al. (1985) found a negativecorrelation between XMg and activation energy for thediusion of 40Ar* in their experiments, hydrothermaldegassing experiments (Grove and Harrison 1996)yielded nearly identical results for Fe-rich biotite(XMg 0.26) and biotite of intermediate composition(XMg 0.47). Instead it was suggested that high halogencontents may increase Ar retentivity. Applying single-domain volume diusion models for argon transport inbiotite of intermediate composition and assuming geo-logically reasonable cooling rates yields closure tem-peratures between 310 and 345 °C (Harrison et al. 1985;Grove and Harrison 1996 and references therein).

The halogen contents of the biotite grains investi-gated in this study are not known. The XMg values rangefrom 0.32 (BW-116) to 0.70 (BW-132). No systematicrelationship between XMg value and age can be recog-nised (Tables 2 and 4). Moreover, all biotite ages coin-cide within error limits. Therefore, the obtained Ar±Arages are interpreted to record cooling of the samplesbelow approximately 310±345 °C.

The Ar diusion systematics of K-feldspar are gen-erally more complicated than those of micas and horn-blende, and many 40Ar/39Ar spectra obtained byincremental heating are rather complex. While earlierstudies suggested that many K-feldspars may not bereliable geochronometers due to the continuous loss of40Ar* at low temperatures and the incorporation of

16

excess argon (e.g. Faure 1977), increasing knowledge onthe transport behaviour of argon in feldspars has con-siderably increased the ability to extract geologicallymeaningful ages (see Harrison and McDougall 1982;Lovera et al. 1989; Foster et al. 1990 and referencestherein). Further studies indicated that the apparentlyanomalous 40Ar/39Ar spectra of K-feldspar re¯ect thepresence of diusion domains of variable length scales(e.g. Zeitler 1987; Lovera et al. 1989; Foster et al. 1990;Fitz Gerald and Harrison 1993; Lovera et al. 1993) andthat this behaviour can be modelled by assuming a dis-crete distribution of non-interacting domains of variablesize (Lovera et al. 1989). The resulting closure temper-atures for K-feldspar span up a range between 150 and300 °C. Recent studies of alkali feldspar microtextures(Parsons et al. 1999) provide a critical view of the mul-tiple diusion domain theory. They suggest that deut-erically altered zones formed by ¯uid±feldspar reactionsat temperatures below 500 °C provide fast pathways forAr that will be metastable in step-heating experiments.However, the study being performed on very slowlycooling, strongly zoned crystals that reacted with ¯uid,its implications are not necessarily applicable to otherrocks.

Most of the above studies and considerations applyto 40Ar/39Ar spectra of K-feldspar that do not showplateaus but saddle- or U-shaped spectra, gradients orspectra with two plateau segments (Foster et al. 1990).Most of the K-feldspars analysed in this study showbroad plateaus with the exception of the ®rst few steps ofincremental heating that mainly have lower, but some-times also higher apparent ages (Fig. 7). The fact thatthe domains with anomalous behaviour are degassedduring the low-temperature steps suggests that thesedomains are either located near crystal surfaces, cleav-age planes or cracks and may correspond to alteredmaterial. Considering a closure temperature between150 and 300 °C for K-feldspar, the plateau ages of296.4 3.0 Ma (BW-120), 305.3 3.8 Ma (BW-116)and 305.1 3.6 Ma (BW-44) ®t very well to the coolingages obtained from hornblende and biotite (see aboveand Table 4). K-feldspars from sample BW-67 do notyield plateaus, but show increasing apparent ages fromapproximately 280 to 300 Ma, the ®nal degassing stepsyielding ages (299.3 2.9, 300.0 2.9 Ma, Table 4) inagreement with the plateau ages of samples BW-116,BW-120 and BW-44. The rimward decrease in age maybe explained by loss of 40Ar* or by progressive closureof the Ar system down to very low temperatures duringslow cooling.

In contrast, the two K-feldspar ages of 319.7 3.1and 317.8 3.1 Ma (BW-46) from location 7 are sig-ni®cantly higher than the biotite ages from the samelocation. Moreover, another K-feldspar grain from thesame location yields a plateau age of 305.1 3.6 Ma(BW-44). Therefore, the age dierence between thegrains cannot be due to dierent cooling rates at thevarious locations but must re¯ect either loss of 40Ar*from the apparently younger grains, or incorporation of

excess argon by the apparently older grains, or dierentAr retentivities of the K-feldspars. Loss of 40Ar* is notapparent from the spectrum of the feldspar from BW-44and, moreover, its plateau age agrees with those fromsamples BW-116 and BW-120 (see above). All horn-blende and biotite data and the other K-feldspar datarepresent an internally consistent set, suggesting thatincorporation of excess argon is not a problem in theserocks.

Dierent Ar retentivities are the best explanation forthe high ages. They could result from dierent structuralstates of K-feldspars. However, as all the K-feldspars arechemically and structurally very similar, we expect thediusion parameters to be in the same range. Therefore,it is suggested that the eective diusion radii of thedominant subgrains in the two K-feldspars of sampleBW-46 are substantially larger than those of the otherK-feldspars. Another possible reason for the elevatedAr±Ar ages of the two K-feldspars is that they containnumerous inclusions of plagioclase. Whereas K-feldsparmost likely crystallised from melt produced during de-hydration melting reactions, plagioclase could partlyrepresent an older stage of metamorphism as indicatedby residual plagioclase crystals in some leucosomes(Berger and Kalt 1999). This hypothesis is supported bypartly low K/Ca ratios (9±87) obtained during the in-cremental heating experiments on feldspars of sampleBW-46 compared to signi®cantly higher K/Ca ratios inall other analysed K-feldspars.

In conclusion, most of the analysed K-feldspar grainsyield plateau ages or apparent ages of the ®nal stepsbetween 296 and 305 Ma. These ages ®t well with thoseobtained from biotite and hornblende at the same oradjacent localities and probably record cooling below150±300 °C. Staircase patterns may record closure of theAr system down to lower temperatures, and older pla-teau ages of 318±320 Ma most likely re¯ect diusiondomains with higher closure temperatures.

Cooling rates, duration of metamorphismand possible heat sources

As discussed in the previous sections, U±Pb zircon andmonazite ages are interpreted to re¯ect crystallisationfrom a melt at approximately 800 °C. For the north-western area, mean leucosome and mesosome monaziteages of 325.3 and 326.1 Ma are considered to date thistemperature, although uncertainty remains about in-heritance problems. U±Pb titanite ages are interpreted ascooling below 700 °C and 40Ar/39Ar ages of hornblendeand biotite are considered to re¯ect cooling below 500±570 and 310±345 °C respectively. Most K-feldsparsprobably record cooling below 150±300 °C. Two feld-spar grains from location 7 (BW-46) most likely havehigher closure temperatures and are excluded in thefollowing. Taking the mean ages for the minerals andaverage cooling ages, cooling paths can be constructedfor the southeastern area around Bodenmais (locations

17

1±5) and for the northeastern area around Cham (lo-cations 6±7; Fig. 8).

The two cooling paths are quite similar, although theclosure temperatures of the K-feldspars and hence thelow-temperature part of the paths are not well-con-strained. Both paths are characterised by rapid cooling inthe high-temperature range and a non-linear decrease inthe cooling rate towards lower temperatures. Consideringan age dierence of approximately 1 million years be-tween zircon andmonazite ages and titanite ages from thesoutheastern area, cooling rates must have been in theorder of 100 °C/my. Further cooling down to the horn-blende closure temperature is characterised by a rate ofapproximately 58 °C/my. For the northwestern area, notitanite ages are available. An average cooling rate ofabout 46 °C/Ma can be calculated for the temperaturespan between 800 °C and hornblende closure tempera-ture. The samples further cooled down to the biotiteclosure temperature at rates of approximately 16 °C/my(southeastern area) and 11 °C/my (northwestern area).Considering the average K-feldspar age of 300 Ma(excluding sample BW-46) to be geologically meaningful,the further cooling rate is 1±17 °C/my. In summary,cooling from 800 to 150±300 °C lasted approximately20±30 my, provided that exhumation was a continuousprocess.

Ideally, in order to constrain possible external heatsources and mechanisms of heat input, not only thecooling rates but also the timing of the peak and pro-grade parts of the P±T path should be known. The onlymineral that potentially records all three parts of theP±T path is garnet. Biotite included in garnets does notyield ages signi®cantly older than those of biotite in thematrix. The monazite that is preserved in a garnet corerecords older ages than those in the matrix. However,

this places no time constraints on prograde or peakmetamorphism as the age is not known precisely enoughand as it may record either metamorphism of the hostrock or detrital inheritance or a combination of the two.More data are needed to test whether the obtained age isrepresentative and reproducible.

The zoning patterns of some very large garnets(5 mm diameter), particularly their bell-shaped Mnpro®les, and their inclusion relations indicate thatgrowth of these garnets started at prograde conditions(Kalt et al. 1999, their Fig. 6a). Chakraborty andGanguly (1991) have examined the extent of relaxationof initial garnet zoning patterns in relation to grain-size,time and diusion coecients. Their model calculationsfor Mn (their Fig. 13) predict that an initially growth-zoned garnet of 5 mm diameter would have beenhomogenised with respect to Mn by diusion if tem-peratures of 850 °C (the estimated peak temperature formigmatites of the Bayerische Wald) prevailed longerthan approximately 2.5 million years. As there are manyuncertainties in this calculation (the initial zoning pat-tern is not known, processes other than diusion mayhave contributed to the zoning pattern, the ®xed edgecomposition assumed in the calculations may not bevalid), the result may just be taken to indicate the orderof magnitude for the maximum duration of peak meta-morphic conditions. A short heating period ®ts well withthe survival of older monazite grains and with thecore-rim structures of zircon grains that also suggestinheritance. No time information can be gained on theprograde part of the P±T path.

As reasoned above, the T±t path of the BayerischeWald is characterised by a short thermal peak and ex-ponentially decreasing cooling rates. As stated in