GeolCarp_Vol58_No5_415

GEOLOGICA CARPATHICA, OCTOBER 2007, 58, 5, 415—425

www.geologicacarpathica.sk

Introduction

High-temperature metamorphism commonly associated
with granitoid magmatism is a major feature of the crystal-
line core of the Variscan orogenic belt (O’Brien 2000, see
also Montel et al. 1992; Petrakakis 1997; Kalt et al. 1999;
Henk et al. 2000 and references therein). Within the
Moldanubian Zone, several areas with intensive and wide-
spread HT metamorphism were distinguished: for ex-
ample, cordierite-bearing migmatites with hercynite and
orthopyroxene-bearing leucosome in the Bavarian Forest,
where temperatures reach up to

~770—850 ºC (Kalt et al.

1999); cordierite-bearing migmatites of the Monotonous
Group in the envelope of the South Bohemian Pluton,
with the mineral assemblages indicating temperatures of
650—800

ºC (Linner 1996; Bűttner & Kruhl 1997; Deibl et

al. 2003).

The garnet-sillimanite-cordierite kinzigite from Petrovice,

Jihlava Pluton is characterized by a hyperaluminous bulk

High-temperature to ultrahigh-temperature metamorphism

related to multiple ultrapotassic intrusions: evidence from

garnet-sillimanite-cordierite kinzigite and garnet-orthopyroxene

migmatites in the eastern part of the Moldanubian Zone

(Bohemian Massif)

JAROMÍR LEICHMANN

, MILAN NOVÁK

, DAVID BURIÁNEK

and

DAVID BURGER

Institute of Geological Sciences, Masaryk University, Kotlářská 2, 611 37 Brno, Czech Republic;

leichman@sci.muni.cz; mnovak@sci.muni.cz; xburger@sci.muni.cz

Czech Geological Survey, Leitnerova 2, 602 00 Brno, Czech Republic; burianek@cgu.cz

(Manuscript received March 6, 2006; accepted in revised form March 15, 2007)

Abstract: The garnet-sillimanite-cordierite kinzigite from Petrovice, Jihlava Pluton, Moldanubian Zone, Czech Republic
has a hyperaluminous chemical composition. Its mineral assemblage consisting of garnet and prismatic sillimanite I,
relics of cordierite, hercynite, rutile I, ilmenite and quartz, retrograde minerals – cordierite II, fibrolitic sillimanite II
and Ti-rich biotite and leucocratic portions with abundant K-feldspar, quartz and plagioclase which closely matches a
restite. The prograde part of the reaction history was virtually obliterated and the mineral reactions: (1) Crd + Hrc = Grt + Sil;
(2) Hrc + Qtz = Grt + Sil; (3) Grt + Sil + Qtz + L = Crd + Bt; (4) Grt + L = Crd + Bt + Qtz; (5) Grt + Crd + L = Bt + Sil; (6)
Crd + L = Bt + Sil + Qtz took place during cooling. The relic assemblages cordierite+hercynite and hercynite+quartz, which
represent a peak of metamorphism, are stable for P

~ 0.5 GPa at T> ~ 900 ºC. The P-T conditions determined by

geothermobarometry yield: T = 900 ± 40

ºC (garnet—ilmenite) and P=0.73—0.49 GPa (spinel—garnet); concentrations of

Zr in rutile I enclosed in garnet and sillimanite I yielded T = 840—970

ºC. We assume, on the basis of the hyperaluminous

chemistry and mineral assembly that the garnet-sillimanite-cordierite kinzigite from Petrovice is a restite, where a large
portion of melt was lost from the rock. The sequence of mineral reactions, and garnet composition with low and constant
Ca in the profile across the grain, indicate an isobaric cooling path. A moderate heat input from the Jihlava Pluton,
manifested by garnet-orthopyroxene-cordierite migmatites developed at the direct contact of durbachites, supported by
a local but large heat input of gabbro—monzogabbro associated directly with kinzigites seem probable heat sources. The
chemical U-Pb ages of monazite 319 ± 24 Ma from kinzigite, 329.8 ± 9.5 Ma from garnet-orthopyroxene migmatites,
and 335.8 ± 6.9 Ma from gabbros are very close to U-Pb ages of zircon 335.2 ± 0.5 Ma from durbachites (Kotková et al.
2003). They suggest a Variscan age of the HT (to UHT) metamorphism and support affiliation to the durbachite intrusions.

Key words: Moldanubian Zone, Jihlava Pluton, mineral reactions, HT metamorphism, heat source, garnet-sillimanite
rock, restite.

chemical composition and mineral assemblages with abun-
dant garnet and prismatic sillimanite I with relics of
cordierite I, hercynite, ilmenite, rutile, and common retro-
grade cordierite II, fibrolitic sillimanite II and Ti-rich biotite.
The kinzigite represents a new example of HT to UHT meta-
morphism in the eastern part of the Moldanubian Zone. The
rock exhibits several features, which very closely resemble
those attributed to restites (see e.g. Vielzeuf & Holloway
1988; Pati

ño Douce & Johnston 1991; Patiño Douce &

McCarthy 1998). It is very similar to Al-rich rock with garnet,
sillimanite, hercynite, cordierite and corundum from
Sepekov near Milevsko, Central Bohemia (Fiala & Losert
1976). The hyperaluminous composition of the rock is fairly
distinct from the HT cordierite-bearing migmatites men-
tioned above and from garnet-orthopyroxene migmatites in
direct contact with the durbachite. We present mineralogical
and petrological data, discuss the formation, the P-T condi-
tions and evolutions of the garnet-sillimanite-cordierite
kinzigite and garnet-orthopyroxene-cordierite migmatites.

416

LEICHMANN, NOVÁK, BURIÁNEK and BURGER

Geological setting

The Moldanubian Zone represents a crustal stack of

allochthonous units assembled during the Variscan
orogeny. It was formed by several Variscan events of su-
perimposed deformation phases and metamorphic recrys-
tallizations (Dallmeyer et al. 1995; Franke 2000). Three
fundamentally different tectonometamorphic units can be
recognized: (i) LP/HT series; (ii) HP/L-MT series; (iii)
remnants of HP/HT series (O’Brien 2000). The rocks
characterized by LP/HT are commonly associated with the
Variscan plutons (Novák & Houzar 1996; Linner 1996;
Büttner & Kruhl 1997; Henk et al. 2000).

The Jihlava Pluton is located in the eastern part of the

Moldanubian Zone, E of the South Bohemian Pluton
(Fig. 1). It is mainly composed of ultrapotassic pyroxene-
biotite syenites to quartz monzonites (durbachites) hosted
in biotite to biotite-sillimanite ± cordierite gneisses locally
strongly migmatized at the contact. Enclaves of biotite-
sillimanite gneisses also form an approximately N—S
trending discontinuous belt, several km long, located in
the center of the Jihlava Pluton (Fig. 1). A NNW—SSE ori-
ented body of shoshonitic to ultrapotassic gabbro to
monzogabbro (Leichmann & Švancara 2005), approxi-
mately 2 km long, is situated E of the gneiss belt within

the Jihlava Pluton. Monzogabbro exhibits complicated
textural relations; the primary assemblage is characterized
by clinopyroxene+orthopyroxene+K-feldspar+plagioclase,
whereas biotite, quartz and amphibole represent secondary
phases. Monzogabbro resembles mafic charnockites and a
temperature of magmatic crystallization (

~1000—1100 ºC;

Leichmann et al. 2000) was calculated from the composi-
tion of mesoperthitic K-feldspar (see Nekvasil 1992). The
durbachites (syenite to quartz monzonite) from the
Jihlava Pluton differ from gabbro to monzogabbro by
their lower amount of mafic components and higher
content of K-feldspar; nevertheless, they exhibit similar
textures such as mesoperthitic K-feldspar and exsolutions
of orthopyroxene in clinopyroxene. These indicate their
magmatic crystallization at high temperature (T

~850 ºC)

but lower relative to the gabbro—monzogabbro.

Two distinct types of rocks showing HT metamorphic

conditions were found within the Jihlava Pluton. The gar-
net-sillimanite-cordierite kinzigite forms an isolated out-
crop about 15 10 m in size located at the contact between
shoshonitic gabbros and gneiss. Both gabbro as well as
gneiss enclave are hosted in the cpx-opx-bearing syenites
to monzonites. The second type – garnet-orthopyroxene
migmatite, occurring locally at the contact of Jihlava Plu-
ton and surrounding metasedimentary rocks, forms small
layers up to several dm thick. It differs from the kinzigite
examined by small size, low amount of melt, less alumi-
nous composition and simple mineral assemblage.

Analytical methods

The electron-microprobe analyses were done on a

Cameca SX-100 instrument in the wavelength-dispersion
mode at the Slovak Geological Survey in Bratislava. Accel-
erating potential was 15 kV for all elements, spot diameter
1 µm, beam current 20 nA. The samples were analysed
using the K lines from the following standards: albite (Na),
wollastonite (Ca, Si), orthoclase (K), MgO (Mg), Fe

(Fe),

chromite (Cr), gahnite (Zn), metal Mn (Mn), Al

(Al),

TiO

(Ti), NaCl (Cl), BaF

(F). Data were reduced on-line

using the PAP routine (Pouchou & Pichoir 1985). The same
equipment was used for the chemical dating of monazite by
the method of Montel (1996). The current of 250 nA and
voltage of 15 kV were used for the determination of Th, U
and Pb contents. The error of determination ranged between
0.04—0.05 wt. % for Th, 0.025—0.030 wt. % for U and about
0.01 wt. % for Pb. Monazite samples with known
concordant U-Pb ages determined by classical radiogenic
methods were used as standards.

Zirconium in rutile was determined on a Cameca SX-100

instrument in the wavelength-dispersion mode at the Joint
Laboratory of Electron Microscopy and Microanalysis, In-
stitute of Geological Sciences, Masaryk University, Brno
and Czech Geological Survey, Prague. The accelerating
voltage was 15 kV, spot diameter 1 µm, beam current 20 nA
and counting time 300 s, the standard used was ZrSiO

Part of the electron microprobe analysis was carried out

on the JEOL Superprobe JXA-8600 instrument in the

Fig. 1. Schematic geological map of the northern part of the Jihlava
Pluton (modified from Tonika 1970).

417

HIGH-TEMPERATURE METAMORPHISM RELATED TO ULTRA POTASSIC INTRUSIONS (BOHEMIAN MASSIF)

wavelength-dispersion mode at the Paris Lodron Univer-
sity, Salzburg, Austria with beam diameter of 3 µm, accel-
erating voltage of 15 kV, and sample current of 20 nA.
Natural minerals and oxides (Si, Al, Ti, Cr, Fe, Mn, Mg,
Ba) were used as standards. Data were reduced using the
ZAF-4 corrections.

Whole-rock chemical analyses were performed at the

Acme Chemical Laboratories Ltd, Vancouver, by ICP-ES
(main elements) and ICP-MS (trace elements and REE).
The melanosome-dominant portion contains a small ad-
mixture of leucosome (about 10 %), but the leucosome
portion contains modest admixture of melanosome (about
20 vol. %). We were not able to subtract the admixture of
melanosome (or leucosome) completely.

Results

Macroscopic description

The garnet-sillimanite-cordierite kinzigite is heteroge-

neous on a macroscopic scale with volumetrically domi-
nant melanosome consisting of coarse- to medium-grained
garnet + sillimanite + cordierite + biotite. The melanosome-
bearing rocks consist of approximately 50 % garnet, up to
20 % sillimanite and cordierite, while biotite forms to-
gether 30 %. It does not exhibit any apparent preferred ori-
entation of minerals. Abundant silvery white prismatic
crystals of sillimanite I, up to 1 cm long, are randomly dis-
tributed. The melanosome locally contains minor, irregu-
larly distributed, medium-grained to locally fine-grained
quartz-feldspar portions of leucosome as nets and veins,
up to 3 cm in size. Leucosome could form up to 30 % of
the rock. The content of retrograde cordierite and biotite
increases at the expense of garnet with increasing
leucosome portion. It forms a matrix, which locally sur-
rounds early-formed minerals. Contacts between melano-
some and leucosome are commonly marked by more
abundant cordierite, biotite and fibrous sillimanite, sharp
contacts of leucosome with sillimanite + garnet dominant
rock are rare. Garnet-orthopyroxene migmatites are dis-
tinctly banded rocks, where biotite-plagioclase bands dis-
tinctly prevail over quartz-feldspar leuocosome. Small
grains of garnet are randomly distributed mainly in the
leucosome portion.

Petrography

The garnet-sillimanite-cordierite kinzigite displays min-

eral assemblages with complex textural relationships in
thin sections. Melanosome consists of euhedral, prismatic
crystals of sillimanite I (Fig. 2a,b), with locally abundant
inclusions of cordierite I, hercynite, quartz, ilmenite and
short prismatic grains of rutile I (all inclusions up to
50 µm in size) randomly distributed within crystals
(Fig. 2a,b). Hercynite and quartz were not found in direct
contact, they are always separated by sillimanite, but the
distance of both minerals is commonly smaller than the size
of their grains. Hercynite and ilmenite are commonly

closely associated (Fig. 2d). Euhedral prismatic sillimanite I
is weakly affected by subsequent mineral reactions
(Fig. 2a,b). Anhedral to locally subhedral zoned grains of
garnet, up to 5 mm in diameter, contain inclusions of
cordierite I, hercynite, ilmenite, quartz and rutile I concen-
trated apparently in cores of garnet (Fig. 2a); abundance,
volume ratios and textural relations of the individual
inclusions in garnet cores are similar to those described in
prismatic sillimanite I. Small inclusions of hercynite were
scarcely found in cordierite I (Fig. 2d). Rare inclusions of
monazite occur in garnet closely associated with other in-
clusions – hercynite, rutile I, cordierite I, ilmenite and
quartz. Monazite is homogeneous in transmitted light as
well as in the BSE image. Very rare accicular rutile II oc-
curs in the outer parts of garnet grains, where the inclu-
sions of cordierite I, hercynite, quartz, rutile I and ilmenite
are not present. Garnet was partly consumed during
mineral reactions and replaced by cordierite II ± biotite
chiefly along garnet rims (Fig. 2c). Garnet cores with
common inclusions were locally replaced as well, but in-
clusions of hercynite locally survived, whereas rutile I,
cordierite I and quartz were completely consumed. Il-
menite is abundant as inclusions in early garnet and silli-
manite, in cordierite II as well as in biotite I and II. It
occurs very likely in several distinct types, which we
were not able to distinguish. Anhedral cordierite II replac-
ing garnet is commonly replaced by aggregate of
biotite + fibrolitic sillimanite II. It seems that biotite is
present in several textural and paragenetic types
(Fig. 2a,b,c). Abundant coarse flakes of biotite I are com-
monly associated with cordierite II and/or fibrolitic
sillimanite II. Small anhedral to subhedral flakes of
biotite II fill late fractures in garnet (Fig. 2a). Cordierite II,
biotite I and fibrolitic sillimanite II are commonly devel-
oped between aggregates of prismatic sillimanite I + garnet,
and quartz-feldspar rich leucosome. Additional accessory
minerals include zircon, xenotime and pyrite. During the
late retrograde stage, chlorite-dominated, fine-grained
pseudomorphs after cordierite II were formed (Fig. 2b,c).
Orthopyroxene, kyanite and other relevant minerals (e.g.
osumilite, corundum, sapphirine) were not identified.

The leucosome portions form irregular masses, veins

and veinlets consisting of anhedral quartz, K-feldspar and
commonly less abundant subhedral plagioclase. Both
feldspars are rather homogeneous. Based on the CL and
EMP study, one compositional type of K-feldspar and pla-
gioclase, respectively, were observed. Anhedral grains of
quartz associated with subhedral plagioclase occur in rela-
tively fine-grained matrix located around large grains of
sillimanite I, garnet, cordierite II and biotite.

The garnet-orthopyroxene migmatite from the

exocontact of the Jihlava Pluton is composed mainly of
quartz, plagioclase and biotite I. Orthopyroxene forms
subhedral colourless grains, up to 1 mm in size, garnet was
found as very irregular poikiloblastic grains, about
0.5 mm in diameter, with abundant inclusions of quartz.
Garnet and orthopyroxene are closely spatially associated
and are commonly replaced by flakes of biotite II. Irregu-
lar grains of cordierite, about 1 mm in size, commonly oc-

419

HIGH-TEMPERATURE METAMORPHISM RELATED TO ULTRA POTASSIC INTRUSIONS (BOHEMIAN MASSIF)

Table 2a: Representative chemical composition of minerals.

Table 2b: Representative chemical composition of minerals.

tot and MgO, whereas SiO

, CaO, Na

O, K

O, Ba, Sr

and Rb are low. The domains consisting of dominant
leucosome are characterized by elevated amounts of K

SiO

, Ba, Rb and Sr. Concentrations of HFSE are generally

high, but the melanosome is systematically enriched rela-
tively to the leucosome (Table 1). The leucosome (A.S.I.
4.71) is slightly enriched in light REE’s (Eu/Eu* = 1.33;
LREE 83.7 ppm, HREE 33.0 ppm) and exhibits a weak

positive Eu anomaly relative to the melanosome, which is
enriched in HREE’s (LREE 60.1 ppm, HREE 60.5 ppm) and
shows distinct negative Eu anomaly (Eu/Eu* = 2.7).

Chemical composition of minerals

Hercynite is a hercynite-spinel-gahnite solid solution

(Table 2). Most inclusions exhibit similar composition

420

LEICHMANN, NOVÁK, BURIÁNEK and BURGER

52—66

Spl

26—37

Gah

10—18

, with a good negative correlation

Fe/Zn and low Cr

< 1.46 wt. % in all samples. Excep-

tionally Zn-rich (16.31 wt. % ZnO) and Zn-poor
(1.23 wt. % ZnO) compositions not involved in the above
formula were found. Hercynite associated with cordierite I
has the composition Hc

Spl

Gah

Garnet from kinzigite (X

= 0.30—0.16) exhibits simple

zoning, with core slightly enriched in pyrope component,
and rim enriched in almandine and spessartine compo-
nents, whereas concentrations of Ca are low and almost
constant across the grains (core – Alm

Pyr

Sps

Grs

rim – Alm

Pyr

Sps

Grs

). A typical profile of a garnet

grain is given in Fig. 3. The chemical compositions of
garnet from garnet-orthopyroxene-cordierite migmatite
(X

= 0.28—0.22) locally exhibits higher Ca (Grs

4—13

Cordierite yielded two distinct compositions; inclu-

sions of slightly heterogeneous cordierite I show
X

= 0.73—0.79, whereas homogeneous large grains of

cordierite II have X

= 0.68—0.69. Both exhibit low Na

contents mostly below 0.10 wt. %.

Biotite is an annite-phlogopite solid solution with sub-

stantial amount of eastonite-siderophylite components. In
kinzigite, biotite I is Ti-rich with 4.48—5.51 wt. % of TiO

(0.26—0.31 apfu) and characterized by X

= 0.47—0.44,

Si = 2.65—2.67 apfu,

Al = 0.51—0.54 apfu. Biotite II has

= 0.44, Si = 2.70 and it is slightly Al- (

Al = 0.43 apfu)

and Ti-depleted (up to 3.59 wt. % TiO

—0.21 apfu Ti).

Low concentrations of Na and Ba of

~0.02—0.01 apfu

and F < 0.15 apfu are typical. Biotite from migmatite is
Mg-enriched (X

= 0.61—0.55), shows slightly lower TiO

(3.33—4.85 wt. %),

Al = 0.38—0.22 apfu but higher

Si = 2.70—2.73 apfu.

Orthopyroxene from migmatite (X

= 0.49—0.48,

Table 2) contains a low amount of Al

(up to

2.18 wt. % – 0.10 apfu Al). Prismatic sillimanite I and
fibrolitic sillimanite II from kinzigite are close to the
end-member composition Al

SiO

with low concentration

of Fe

and MgO. Plagioclase An

39—36

from the

Fig. 3. A typical compositional profile of a garnet grain.

Fig. 4. Concentrations of Zr in inclusions of rutile I in garnet and
sillimanite I.

leucosome is rather homogeneous. K-feldspar is slightly
Ba-enriched, with up to 1.5 wt. % of BaO. K-feldspar from
migmatites is Ba depleted (BaO 0.55 wt. %). Ilmenite has
X

= 0.01—0.03 and low concentrations of Mn. Repre-

sentative compositions of minerals are given in Table 2.
The concentrations of Zr in rutile I inclusions, enclosed in
garnet cores or in prismatic sillimanite I, vary from 750 to
2200 ppm (see Fig. 4). Monazite enclosed in garnet from
kinzigite yielded a chemical age of 319 ± 24 Ma, while
monazite from migmatites has an age of 329.8 ± 9.5 Ma.

Discussion

Mineral reactions and petrogenetic grid

The metamorphic evolution of garnet-sillimanite-cordi-

erite kinzigite is illustrated in a petrogenetic grid in the
KFMASH system and its derivates. Abundant ilmenite and
rutile I inclusions and high Ti in biotite indicate substan-
tial participation of TiO

as a minor extra component in

the system.

The early mineral assemblage A (Table 3) is defined by

abundant inclusions of cordierite I + hercynite + ilmenite +
+ rutile I+quartz enclosed in prismatic sillimanite I and in
cores of garnet (Fig. 2a,c). Textural relations strongly sug-
gest that this mineral assemblage represents the earliest
stage recorded in the garnet-sillimanite-cordierite kinzigite.
The reactions producing cordierite I, hercynite, ilmenite,
rutile I, and quartz cannot be constrained due to absence of
early-formed minerals. The occurrence of a relatively abun-
dant assemblage of ilmenite+rutile+hercynite indicates
breakdown of a Ti-rich spinel mineral such as ulvöspinel
(see Sengupta et al. 1999); however, the textural relations
are not developed enough to specify the reaction.

The mineral assemblage B, volumetrically dominant in

the melanosome, is characterized by prismatic sillimaniteI +

421

HIGH-TEMPERATURE METAMORPHISM RELATED TO ULTRA POTASSIC INTRUSIONS (BOHEMIAN MASSIF)

+garnet and represents the second stage of the evolution.
Relics of cordierite I plus hercynite (in direct contact)
found in garnet cores suggest that the mineral assemblage
B formed via the FMAS reaction (Fig. 5):

Crd + Hrc = Grt + Sil. (1)

Relics of hercynite and quartz (not in direct contact) oc-

curring in sillimanite I and in garnet cores may suggest the
reaction:

Hrc + Qtz = Grt + Sil. (2)

The presence of melts during the reactions (1) and (2) is

unclear, but formation of the assemblage garnet + sillimanite
and garnet alone may have proceeded via more compli-
cated reactions in the KFMASH system. High Zn contents
in some hercynite may come from early staurolite.

The mineral assemblage C, characterized by the assem-

blage cordierite II ± biotite (Fig. 2b,c, Fig. 6, Table 3), rep-
resents the third stage. The reactions (3) and (4) are
portrayed in the relevant KFMASH system:

Grt + Sil + Qtz + L = Crd + Bt, (3)

Grt + L = Crd + Bt + Qtz. (4)

They include garnet, sillimanite, quartz and melt as re-

actants; however, the textural relations show that pris-
matic sillimanite I is rather stable during these reactions
(Fig. 2b). It may suggest significant participation of the re-
action (4) relative to the reaction (3). Rare hercynite inclu-
sions in cordierite II show that it was not or it was only
partly consumed during garnet decomposition. Rutile I
was likely consumed during the above reactions and Ti
was incorporated into Ti-rich biotite and/or ilmenite. Il-
menite inclusions are common in both cordierite II and
garnet, thus ilmenite did not participate in the reactions or
it was more likely newly formed along with cordierite II,
biotite and K-feldspar.

The volumetrically abundant mineral assemblage D in-

volving biotite + fibrolitic sillimanite II is produced at the
expense of garnet and particularly cordierite II via the re-
actions (Figs. 5, 6):

Grt + Crd + L = Bt + Sil, (5)

Crd + L = Bt + Sil + Qtz. (6)

As in the reactions (3) and (4), the presence of melt is re-

quired in the reactions (5) and (6). Abundance of the min-
eral assemblages C and D suggests essential participation
of melt in the back (cooling) reactions.

Absence of orthopyroxene is a typical feature of garnet-

sillimanite-cordierite kinzigite from Petrovice disregard-
ing the fact that the estimated P-T conditions are suitable
to produce this mineral. The hyperaluminous nature of the
rock seems to be a crucial factor, which controls the ab-
sence of orthopyroxene. We assume that all Fe + Mg were
consumed by garnet and the remaining Si and Al were ac-
commodated in sillimanite.

Textural relations in garnet-orthopyroxene-cordierite

migmatite are also complicated. Orthopyroxene probably
originated by melting of biotite by the reaction:

Bt + Pl + Qtz = Opx + Kfs + L + Grt (7)

Table 3: Mineral assemblages A—D.

Fig. 5. P-T diagram for the KFMASH system showing the locations
of selected melting and dehydratation reactions (modified from
Spear et al. 1999). Dashed field (reaction A—B) calculated from
garnet-ilmenite thermometer (Pownceby et al. 1991) and spinel-
garnet geobarometer (Nichols et al. 1992). The arrow indicates the
path of cooling. The X

= 0.33 in hercynite relative to X

= 0.30

in associated garnet shifts position of the reaction Alm + Sill = Hc + Qtz
to the higher-pressure side of the reaction Alm + Sill = Spl + Crd (Spear
et al. 1999 – see Fig. 3).

Fig. 6. Schematic P-T diagram for KFMASH system showing se-
lected reactions involving cordierite (modified from Vielzeuf &
Holloway 1988). The arrows mark reaction directions in the
garnet-sillimanite rock.

422

LEICHMANN, NOVÁK, BURIÁNEK and BURGER

(see Vielzeuf & Montel 1994). Other textural relations
indicate a similar cooling path as in the kinzigites as
documented by the reactions (4), (5) and (6) However, no
relics of spinel, rutile, ilmenite or Mg-rich cordierite,
common in the garnet and sillimanite I from kinzigite,
were found, hence the prograde evolution of both rocks
was probably different. The position of garnet in
leucosome indicates that the garnet represents a restite
phase remaining after melt producing breakdown of
biotite I.

Estimation of metamorphic conditions

Kinzigite

The mineral assemblages, textural relations and chemi-

cal composition of minerals indicate a complex meta-
morphic evolution of the garnet-sillimanite-cordierite
kinzigite. The diagnostic mineral assemblages and min-
eral reactions define the individual metamorphic stages
A to D. However, the resetting of the assemblages during
cooling, documented by the retrograde zoning of the gar-
net, exclude the application of most Fe-Mg cation-ex-
change thermometers for the estimation of the peak
metamorphic conditions. The high disequilibrium in
mineral assemblages, presence of melt in the reactions
followed by widespread back reactions, and absence of
orthopyroxene and plagioclase in the metamorphic min-
eral assemblages A + B excludes the application of
multielement geothermobarometers, like GASP (garnet-
alumosilicate-plagioclase) or GRAIL (garnet-rutile-alumo-
silicate-ilmenite). The assemblage A represented by the
assemblage cordierite I+hercynite+rutile I+ilmenite+quartz
may record peak conditions in garnet-sillimanite-cordier-
ite kinzigite. We have several lines of evidence, which
suggest HT to perhaps UHT metamorphic conditions:

a) The assemblage cordierite I + hercynite is stable in

the LP/HT field in the P-T diagram (see Fig. 5); high
X

= 0.73—0.79 in cordierite I suggests high temperature

at T>

~900 ºC at P~0.5—0.7 GPa (see Fig. 5).

b) The assemblage hercynite+quartz shows a similar

stability field (Fig. 5), but high concentrations of Zn in
hercynite indicate higher pressure (Shulters & Bohlen
1989; Dasgupta et al. 1995) and elevated spinel compo-
nent in hercynite supports higher temperature (Waters
1991) relative to the end-member hercynite. Neverthe-
less, hercynite and quartz, although closely associated,
were not found in direct contact, so this evidence is not
sufficient for UHT conditions. Moreover, we do not
know the parental mineral reaction producing hercynite.
It may have formed by breakdown of sillimanite, kyanite
and/or staurolite (Spear et al. 1999), which may be a
source of Zn in hercynite. Nevertheless, crystallization of
garnet at the expense of hercynite + quartz according to
the reaction (2) may also significantly increase the
amount of Zn in the relic hercynite.

c) The potential existence of spinel with elevated of

TiO

component as a precursor of the assemblage

hercynite + ilmenite + rutile, shown by common close as-

sociation of hercynite and ilmenite (Fig. 2d), may indicate
the presence of primary Ti-rich spinels a good indicator
of UHT conditions (Sack & Ghiorso 1991; Sengupta et
al. 1999); however, no relics of such a Ti-rich spinel were
observed and the textural relations of hercynite + ilmenite
are not convincing enough to provide evidence for the
existence of Ti-rich spinel.

d) The concentrations of Zr in inclusions of rutile I en-

closed in garnet cores and in prismatic sillimanite I is a
novel method to determine T in UHT metamorphic rocks
(Zack et al. 2004). The obtained data yielded T

~840—

910

ºC in inclusions from garnet and slightly higher

~890—970 ºC in inclusions enclosed in prismatic

sillimanite I (see Fig. 4).

e) We applied geothermobarometry, disregarding the

high degree of disequilibrium in kinzigite, to the garnet
cores, where the enclosed minerals may be in equilibrium
with the surrounding garnet. The garnet-ilmenite ther-
mometer (Pownceby et al. 1991) yielded T = 900 ± 40

ºC,

the spinel-garnet geobarometer (Nichols et al. 1992)
yielded P = 0.68 GPa for temperature 900

ºC, and 0.8 GPa

for 950

ºC. Disregarding the fact that formation of the as-

semblage A at T>

~900 ºC and P~0.5—0.7 GPa (or slightly

higher) is supported by the lines of evidence discussed
above; none of them is sufficiently convincing to con-
firm unequivocally the UHT metamorphic conditions in
the Petrovice kinzigite.

The assemblage B garnet (X

= 0.30) + sillimanite re-

placed by cordierite II (X

= 0.68—0.69) ± biotite is stable

at T>

~720 ºC (Fig. 5) and high T~750 ºC and P~0.5 GPa

is also indicated by high X

in cordierite II (e.g. Spear

et al. 1999). The survival of hercynite during replace-
ment of garnet by cordierite II in the stability field of
hercynite + cordierite II suggests T>

~800 ºC for P~0.3 GPa

and generally low pressure P<

~0.5 GPa (Fig. 5) for the

assemblage B. The temperature T>

~720—800 ºC and

~0.5 GPa is suggested for the assemblage B.

The formation of Ti-rich biotite (up to 0.31 apfu Ti) in

the assemblages C and D by the reactions including
rutile and ilmenite (not specified in detail) also suggests
a high temperature up to T

~800 ºC (cf. Sengupta et al.

1999). High Fe in biotite (X

= 0.47—0.44) relative to

Mg-rich biotite from UHT metamorphic rocks (e.g.
Carrington & Harley 1995; Sengupta et al. 1999) and the
biotite obtained during experimental studies (e.g. Pati

ño

Douce & Johnston 1991; Pati

ño Douce & McCarthy

1998; Kriegsmann & Hensen 1998; Sengupta et al. 1999)
may reflect rather low pressure. The garnet-biotite ther-
mometer calibrated according to Bhattacharya et al.
(1992) yielded temperatures of 663—671

ºC for the pres-

sure 0.45 GPa. The abundance of K-feldspar in quartz-
feldspar nests and absence of muscovite indicate
P<

~0.4 GPa during the late stage at T<~650 ºC in the

lower pressure side of the invariant point I (Fig. 5). Nev-
ertheless, its position is controlled by the activity of H

and its lower activity shifts the invariant point I to a
higher pressure (Holtz & Johannes 1996).

The sequence of the mineral assemblages in the garnet-

sillimanite-cordierite kinzigite (A to D): cordierite I+

423

HIGH-TEMPERATURE METAMORPHISM RELATED TO ULTRA POTASSIC INTRUSIONS (BOHEMIAN MASSIF)

+ hercynite + rutile I + ilmenite + monazite + quartz

prismatic

sillimanite I +garnet + ilmenite+ rutile II

cordierite II ± biotite

( ± hercynite, ilmenite)

biotite + fibrolitic sillimanite II +

+ ilmenite, their textural relations (resorbed garnet), chemi-
cal composition (low Ca contents in garnet) and sequence
of metamorphic reactions producing cordierite II (reac-
tions 3—4) and consuming cordierite II (reactions 5—6) do
not allow us to specify the P-T path in greater detail.
However, some lines of evidence, such as low Ca con-
tents in garnet and presence of the assemblage
cordierite + hercynite, indicate low P<

~0.5—0.7 GPa and

high T>

~800—950 ºC. Consequently, an isobaric cooling

path seems more likely relative to isothermal decompres-
sion (see Figs. 5, 6). High disequilibrium observed in the
kinzigite as well as granophyric textures (Shelley 1993)
developed locally in the leucosome indicates rapid cool-
ing at relatively low water pressure, during the reactions
producing the assemblages C and D.

Garnet-orthopyroxene migmatites

The peak conditions of garnet-orthopyroxene

migmatites could be estimated by comparison of the sys-
tem published by Vielzeuf & Montel (1994). The authors
estimated, for rocks of a similar composition (Qtz-Pl-Bt-
bearing metagreywackes), that the Opx – in temperature
range from 810

ºC for the pressure 0.2 GPa, and 885 ºC

for 1 GPa. The Bt – out temperature from 860

ºC for

0.2 GPa and 980

ºC for 1 GPa. Garnet is stable by pres-

sures higher than 0.5 GPa, whereas cordierite at lower
pressure, spinel appears instead of garnet at temperatures
higher than 850

ºC. Therefore, the garnet-orthopyroxene-

cordierite migmatites originated certainly at temperatures
higher than 800

ºC, and pressure higher than 0.5 GPa.

The skeletal form of garnet and high portion of

unreacted biotite found in the biotite-plagioclase bands
indicate that the temperature was not high enough to de-
compose the biotite completely, or the heat input was
relatively short-lived, producing only a small amount of
melt. The melt was just segregated, but not fully ex-
tracted from the parental gneiss. Kinzigite is, on the
other hand, a restite after almost complete extraction of
the melt. The absence of spinel and other minerals com-
mon as inclusions in garnet from kinzigite could be in-
terpreted as evidence for relatively lower temperature
origin of garnet-orthopyroxene-cordierite migmatites rela-
tive to kinzigite. The garnet-biotite thermometry yielded
temperatures of 658—686

ºC for the pressure 0.5 GPa.

These values are almost identical with those, calculated
from Grt-Bt pair in the kinzigite.

Potential heat source

The peak conditions of metamorphism at T up to

~800—950 ºC and P~0.5—0.7 GPa estimated for the gar-
net-sillimanite-cordierite kinzigite from Petrovice and
feasible isobaric cooling path (see discussion above) re-
quire an external heat source (cf. Spear 1993; Thompson
1999). The chemical U-Pb age (319 ± 24 Ma) of monazite

closely associated with the assemblage A suggests the
Variscan age of the HT to UHT metamorphism at
T>

~900 ºC. In the region of the Petrovice kinzigite the

following potential heat sources are discussed.

a) A large body of two-pyroxene ultrapotassic syenite

to monzonite (Jihlava Pluton) located around the
kinzigite body seems to be a dominant heat source in
this region. Moreover, the radiometric age of the
Jihlava Pluton’s crystallization (U-Pb zircon, TIMS
335.2 ± 0.5 Ma, Kotková et al. 2003) is close to the
chemical U-Pb age of the monazite 329.8 ± 9.5 Ma from
the garnet-orthopyroxene-cordierite migmatites. The
mineral assemblage in the garnet-orthopyroxene-cordier-
ite migmatites from the exocontact of the Jihlava Pluton
does not exhibit temperatures high enough to produce a
high-grade melting and thus such a hyperaluminous
restite—kinzigite with abundant sillimanite and hercynite.
However, the retrograde (cooling) stage seems to be fairly
similar to those from the kinzigite.

b) The gabbro-monzogabbro body occurring in close

vicinity to the garnet-sillimanite-cordierite kinzigite
with the calculated temperature of crystallization of
monzogabbro at T

~1000—1100 ºC (Leichmann et al.

2000) represents the possible heat source for the
kinzigite. The U-Pb monazite age of the gabbro,
335.8 ± 6.9 (Leichmann & Švancara 2005) lies within the
error of the kinzigite (319 ± 24 Ma) and Grt-Opx
migmatite (329.8 ± 9.5 Ma) age estimation. The gabbroic
intrusion may provide enough heat to produce high-de-
gree melting of a metapelitic or metapsammitic protolith
at temperatures

~900 ºC (cf. e.g. Patiño Douce &

Johnston 1991; Spear et al. 1999).

Fig. 7. Schematic P-T diagram modified after O’Brien (2000).
Note that the P-T evolution of the garnet-sillimanite rock from
Petrovice remarkably deviates from the three major Variscan P-T
evolution paths known in the Bohemian Massif.

424

LEICHMANN, NOVÁK, BURIÁNEK and BURGER

c) High heat flow to shallow crustal levels of 15 km

(

~0.5 GPa) subsequent to the Variscan collision (see

Fig. 7) and crustal thickening is typical of the
Moldanubian Zone (e.g. Montel et al. 1992; Kalt et al.
1999; Henk et al. 2000). Such a style of melting rather
produced large-scale melting then small kinzigite body
with the mineral assemblages showing a high degree of
disequilibrium indicating rapid cooling.

In our scenario, the Jihlava Pluton may have served

as a regional heat source with a modest heat contribu-
tion on a large scale. The intrusion of gabbro-
monzogabbro represents a feasible heat source, which
may have caused local overheating high enough to pro-
duce hyperaluminous kinzigite with the observed HT to
UHT mineral assemblages. The close relation of the
durbachite emplacement and tectonometamorphic evo-
lution of the host rock is also pointed out by Verner et
al. (2005).

Conclusions and summary

The garnet-sillimanite-cordierite kinzigite from

Petrovice is characterized by hyperaluminous bulk
chemical composition and mineral assemblages involv-
ing abundant garnet and prismatic sillimanite I with rel-
ics of cordierite I, hercynite, rutile I, ilmenite and quartz,
and retrograde cordierite II, fibrolitic sillimanite II and Ti-
rich biotite. The relic assemblages – cordierite + hercynite
and hercynite + quartz may suggest peak metamorphism
conditions at T

~900 ºC and P~0.5—0.7 GPa. We assume

the kinzigite to be a restite, where a large portion of melt
was extracted from the rock (see e.g. White & Powell
2002). Modest heat input from the Jihlava Pluton pro-
ducing garnet-orthopyroxene migmatites supported by
local but high heat input from a gabbro-monzogabbro
body seems to be the important heat source. Neverthe-
less, the prograde history of the P-T path is uncertain, al-
though the sequence of mineral reactions and garnet
composition with low Ca suggest an isobaric cooling
path for both kinzigites and Grt-Opx migmatites (see
Figs. 5, 6). The chemical U-Pb age of monazite
(319 ± 24 Ma) from the kinzigite related to assemblage A
suggests a Variscan age of the HT to UHT metamor-
phism. The P-T evolution of the garnet-sillimanite rock
from Petrovice remarkably deviates from three major
Variscan P-T evolution paths known in the Bohemian
Massif (Fig. 7) and interpretation of the potential tectonic
scenario remains unclear. However, the cooling path re-
semble the cooling part of the Carboniferous HT-LP path
(O’Brien 2000), but shifted to the higher temperature
range.

Acknowledgments: We appreciate J. Kotková for her com-
ments on the manuscript, M. Taylor for improving the lan-
guage. This work was supported by the Grant Agency of
the Czech Republic (Grant No. 205/03/0400) and the re-
search project MSM. 0021622412. The authors thank P.
Tropper and S. Vrána for their comments and suggestions.

References

Bűttner S. & Kruhl J.H. 1997: The evolution of a late-Variscan

high-T/low-P region: the southeastern margin of the
Bohemian massif. Geol. Rdsch. 86, 22—38.

Carrington D.P. & Harley S.L. 1995: Partial melting and phase re-

lations in high-grade metapelites: an experimental petroge-
netic grid in KFMASH system. Contr. Mineral. Petrology
120, 270—291.

Dallmeyer R.D., Franke W. & Weber K. (Eds.) 1995: Pre-Permian

geology of the Central and Eastern Europe. Springer, Berlin,
1—213.

Dasgupta S., Sengupta P., Ehl J., Raith M. & Barchan S. 1995: Re-

action textures in a suite of spinel granulites from the Eastern
Ghats Belt, India: Evidence from polymetamorphism, a partial
petrogenetic grid in the system KFMASH and the role of ZnO
and Fe

. J. Petrology 36, 435—461.

Deibl I., Tropper P. & Mirwald P.W. 2003: The petrological evolu-

tion of high-grade metapelites from the Sauwald, southern Bo-
hemian Massif. Mitt. Österr. Miner. Gesell. 148, 116—117.

Fiala F. & Losert J. 1976: Discovery of corundum, cordierite, gar-

net, sillimanite and tourmaline in herynite-rich rock at Chlum
near Sepekov. Čas. Mineral. Geol. 21, 326 (in Czech).

Franke W. 2000: The mid-European segment of the Variscides:

tectonostratigraphic units, terrane boundaries and plate tec-
tonic evolution. In: Franke W., Haak V., Oncken O. & Tan-
ner D. (Eds.): Orogenic processes: quantification and
modelling in the Variscan Belt. Geol. Soc. London, Spec.
Publ. 179, 35—61.

Henk A., von Blanckenburg F., Finger F., Schaltegger U. & Zulauf

G. 2000: Syn-convergent high-temperature metamorphism and
magmatism in the Variscides: a discussion of potential heat
sources. In: Franke W., Haak V., Oncken O. & Tanner D.
(Eds.): Orogenic processes: quantification and modelling in the
Variscan Belt. Geol. Soc. London, Spec. Publ. 179, 387—400.

Holtz F. & Johannes W. 1996: Petrogenesis and experimental pe-

trology of granitic rocks. Springer, Berlin, 1—333.

Kalt A., Berger A. & Blümel P. 1999: Metamorphic evolution of

cordierite-bearing migmatites from the Bayerische Wald
(Variscan Belt, Germany). J. Petrology 40, 601—627.

Kotková J., Schaltegger U. & Leichmann J. 2003: 338—335 Ma old

intrusion in the E Bohemian massif – a relic of the orogen-
wide durbachitic magmatism in European Variscides. J. Czech.
Geol. Soc. 48, 1—2, 80—81.

Kretz R. 1983: Symbols for rock-forming minerals. Amer. Mineral-

ogy 68, 277—279.

Kriegsman L.M. & Hensen B.J. 1998: Back reaction between restite

and melt: Implications for geothermobarometry and pressure-
temperature paths. Geology 26, 1111—1114.

Leichmann J. & Švancara J. 2005: Shoshonitic to ultrapotassic gab-

bros from the Jihlava pluton. 2. meeting of the Czech Geol.
Soc., Slavonice, 151—153 (in Czech).

Leichmann J., Novák M., Zachovalová K. & Topa D. 2000: Ultra-

high temperature event in the Moldanubian Zone; its signifi-
cance for the origin of durbachites. Geolines 10, 47.

Linner M. 1996: Metamorphism and partial melting of paragneisses

of the Monotonous Group, SE Moldanubicum (Austria). Min-
eral. Petrology 58, 215—234.

Montel J.M., Marignac C., Barbey P. & Pichavant M. 1992:

Thermobarometry and granite genesis: Hercynian low-P,
high-T Velay anatectic dome (French Massif Central). J. Meta-
morphic Geology 10, 1—15.

Nekvasil H. 1992: Feldspar crystallization in felsic magmas: a re-

view. Trans. Roy. Soc. Edinburgh, Earth Sci. 83, 399—407.

Nichols G.T., Berry R.F. & Green D.H. 1992: Internally consistent

425

HIGH-TEMPERATURE METAMORPHISM RELATED TO ULTRA POTASSIC INTRUSIONS (BOHEMIAN MASSIF)

gahnitic spinel-cordierite-garnet equilibria in the FMASHZn
system: geothermobarometry and applications. Contr. Mineral.
Petrology 111, 362—377.

Novák M. & Houzar S. 1996: The HT/LP metamorphism of dolo-

mite marbles in the eastern part of the Moldanubicum; a mani-
festation of heat flow related to the Třebíč Durbachite Massif.
J. Czech. Geol. Soc. 41, 139—146.

O’Brien P.J. 2000: The fundamental Variscan problem: high-tem-

perature metamorphism in different depths and high-pressure
metamorphism at different temperatures. In: Franke W., Haak
V., Oncken O. & Tanner D. (Eds.): Orogenic processes: quan-
tification and modelling in the Variscan Belt. Geol. Soc. Lon-
don, Spec. Publ. 179, 369—386.

Pati

ño Douce A.E. & Johnston A.D. 1991: Phase equilibria and

melt productivity in the pelitic system: implications for the ori-
gin of peraluminous granitoids and aluminous granulites.
Contr. Mineral. Petrology 107, 202—218.

Pati

ño Douce A.E. & McCarthy T.C. 1998: Melting of crustal rocks

during continental collision and subduction. In: Hacker B.R. &
Liou J.G. (Eds.): When continents collide: geodynamics and
geochemistry of ultrahigh-pressure rocks. Kluwer Academic
Publishers, 27—55.

Petrakakis K. 1997: Evolution of Moldanubian rocks in Austria: re-

view and synthesis. J. Metamorphic Geology 15, 203—222.

Pouchou J.L. & Pichoir F. 1985: “PAP” procedure for improved

quantitative microanalysis. Microbeam. Anal. 20, 104—105.

Pownceby M.L., Wall V.J. & O’Neill H.St.C. 1991: An experimen-

tal study of effect of Ca upon garnet-ilmenite Fe-Mn exchange
equilibria. Amer. Mineralogist 76, 1580—1588.

Sack R.O. & Ghiorso M.S. 1991: An internally consistent model for

the thermodynamic properties of Fe-Mg-titano-magnetite-alu-
minate spinels. Contr. Mineral. Petrology 106, 474—505.

Sengupta P., Se J., Dasgupta S., Raith M., Bhui U.K. & Ehl J. 1999:

Ultra-high temperature metamorphism of metapelitic granu-
lites from Khondapalle, Eastern Ghats belt: Implications for
the Indo-Antarctic correlation. J. Petrology 40, 1065—1087.

Shelley D. 1993: Igneous and metamorphic rocks under the micro-

scope. Chapman & Hall, London, 1—445.

Shulters J.C. & Bohlen S.R. 1989: The stability of hercynite and

hercynite-gahnite spinels in corundum- or quartz-bearing as-
semblages. J. Petrology 30, 1017—1031.

Spear F.S. 1993: Metamorphic phase equilibria and pressure-tem-

perature-time paths. Miner. Soc. Amer. Monograph, Washing-
ton, 1—799.

Spear F.S., Kohn M.J. & Cheney J.T. 1999: P-T paths from

anatectic pelites. Contr. Mineral. Petrology 134, 17—32.

Thompson A.B. 1999: Some time-space relationships for crustal

melting and granitic intrusions at various depths. In: Castro A.,
Fernandez C. & Vigneresse L. (Eds.): Understanding granites:
Integrating new and classical techniques. Geol. Soc. London,
Spec. Publ. 168, 7—25.

Tonika J. 1970: Geology and petrology of the Jihlava massif.

Sbor. Geol. Věd, Ř. G. 17, 105—119 (in Czech).

Verner K., Žák J., Holub F.V., Týcová P. & Nahodilová R. 2005:

Textures and fabric of the durbachite plutons as an indicator
of the tectonometamorphic evolution of the Moldanubicum
(Czech Republic). 2. meeting of the Czech Geol. Soc.,
Slavonice, 101 (in Czech).

Vielzeuf D. & Holloway J.R. 1988: Experimental determination of

the fluid-absent melting relations in the pelitic system: Conse-
quences for crustal differentiation. Contr. Mineral. Petrology
98, 257—276.

Vielzeuf D. & Montel J.M. 1994: Partial melting of Al-

metagreywackes. Part I. Fluid-absent experiments and phase
relationships. Contr. Mineral. Petrology 117, 375—393.

Waters D.J. 1991: Hercynite-quartz granulite phase relations and

implication for crustal processes. Eur. J. Mineral. 3, 567—586.

White R.W. & Powell R. 2002: Melt loss and the preservation of

granulite facies mineral assemblages. J. Metamorphic Geology
20, 621—632.

Zack T., Moraes R. & Kronz A. 2004: Empirical calibration of a

rutile thermometer and its application to UHT granulites. Inter-
national workshop on petrogenesis of granulites and related
rocks. Moravian Museum, Brno, 1—71.