GeolCarp_Vol58_No2_107

GEOLOGICA CARPATHICA, APRIL 2007, 58, 2, 107—119

www.geologicacarpathica.sk

P-T pseudosections in KFMASH, KMnFMASH,

NCKFMASH and NCKMnFMASH systems: a case study

from garnet-staurolite mica schist from the Alpine

metamorphic basement of the Pannonian Basin (Hungary)

PÉTER HORVÁTH

Institute for Geochemical Research, Hungarian Academy of Sciences, Budaörsi út 45, H-1112 Budapest, Hungary;

phorvath@geochem.hu

(Manuscript received March 28, 2006; accepted in revised form June 22, 2006)

Abstract: P-T pseudosections (quantitative phase diagrams) in the KFMASH, KMnFMASH, NCKFMASH and
NCKMnFMASH systems were calculated with PERPLEX and THERMOCALC from garnet-staurolite mica schist origi-
nating from the basement of the Pannonian Basin (Hungary) in the P-T range 0.2—1.2 GPa and 450—700 ºC with quartz
and H

O in excess. The previously published peak P-T conditions (650 ± 30 ºC and 0.9 ± 0.1 GPa) are confirmed by the

resultant pseudosections in all systems with garnet-biotite-kyanite-muscovite-plagioclase-quartz as the stable paragen-
eses. Mineral composition isopleths mostly model the mineral chemical changes. The P-T path outlined in the NCKFMASH
system shows that biotite and quartz inclusions in host staurolite with matrix muscovite and plagioclase are stable at 560—
650 ºC and 0.35—0.6 GPa. During simultaneous P-T increase garnet started to form at about 0.6 GPa in the garnet-
biotite-staurolite-muscovite-plagioclase-quartz field, and then kyanite appeared at the expense of staurolite near the P-
T peak. The cooling path passed the garnet-biotite-staurolite-muscovite-plagioclase-quartz field again and ended in
chlorite-bearing assemblages. The presented P-T pseudosections in combination with the published Ar-Ar age data on
muscovite (85.5 ± 1.2 Ma) support the Eo-Alpine age of amphibolite facies metamorphism in the basement of the Tisza
Unit (southwestern part of the Pannonian Basin), but do not exclude the possibility of a polymetamorphic origin
(Variscan-Alpine).

Key words: Pannonian Basin, Alpine metamorphism, metamorphic petrology, pseudosection calculations, PERPLEX
and THERMOCALC, mica schist.

Introduction

The model system KFMASH (K

O-FeO-MgO-Al

-SiO

O) has been studied intensively in order to develop

petrogenetic grids for the range of common mineral assem-
blages in metapelitic rocks. Early qualitative grids were
based on natural mineral parageneses (e.g. Albee 1965;
Pattison & Tracy 1991; Droop & Harte 1995); later quanti-
tative grids were based on internally-consistent thermody-
namic data sets (e.g. Spear & Cheney 1989; Powell &
Holland 1990). In recent years a number of studies have
been carried out to augment the system with additional
components such as Na

O, CaO, TiO

, Fe

and MnO

(e.g. Mahar et al. 1997; Worley & Powell 1998; White et
al. 2001). From these components MnO has the most im-
portant role in stabilizing ferromagnesian minerals, mostly
garnet. Mahar et al. (1997) calculated a quantitative grid
for the KMnFMASH system in the range 450—700 ºC and
0—2 GPa based on the Holland & Powell (1990) thermody-
namic dataset to determine the effect of Mn on mineral
stabilities. Wei et al. (2004) presented a new calculation of
the petrogenetic grid in the KMnFMASH system for low-
and medium-pressure conditions using THERMOCALC
3.1 with the internally-consistent thermodynamic data set
of Holland & Powell (1998 and upgrades) and updated

models of activity-composition relationships. Their main
conclusions are that the addition of Mn to the KFMASH
system: enhances the stability field of garnet, extends the
medium-P stability of muscovite and reduces the stability
of staurolite and cordierite. Vance & Mahar (1998),
Tinkham & Stowell (2000) and Stowell et al. (2001) used
the NCKMnFMASH system to derive garnet growth P-T
paths from medium-grade metapelites. Stowell et al.
(2001) concluded that the NCKMnFMASH is the mini-
mum system required to apply quantitatively pseudosec-
tions to natural metapelites containing garnets and Na-
and Ca-bearing phases (e.g. plagioclase, zoisite).

There are three main focuses of this paper. First, a com-

parison between pseudosections (quantitative phase dia-
grams) calculated with PERPLEX (Connolly 1990;
Connolly & Petrini 2002) and THERMOCALC (Powell et
al. 1998) are performed in the KFMASH and KMnFMASH
systems on a representative garnet-staurolite mica schist
sample from the basement of the Pannonian Basin, Hunga-
ry (Újszentiván Uszi-2 borehole). THERMOCALC phase
diagrams involving solid solutions are calculated by solv-
ing sets of non-linear equations, while PERPLEX uses the
minimization of Gibbs energy. A similar type of compari-
son was performed by Hoschek (2004) who compared the
programs THERMOCALC, PERPLEX and DOMINO on a

108

HORVÁTH

kyanite eclogite from the Tauern Window (Eastern Alps,
Austria). Then, we present NCKFMASH and NCKMnF-
MASH pseudosections with mineral compositional isop-
leths to model mineral stability and mineral compositional
changes in the studied sample. Finally, we present a P-T
path for the studied mica schist sample in the NCKF-
MASH system.

Geology and previous data on metamorphism

The basement of the southwestern part of the Pannonian

Basin (Tisza or Tisia Unit) covered by several thousand
meter-thick sedimentary sequences consists of polymeta-
morphic formations. The Tisza Unit (Fig. 1a) originated
from the northern, European margin of Tethys by mostly
Meso-Alpine horizontal displacements of microplates
(Géczy 1973; Kovács 1982; Kázmér & Kovács 1985;
Kovács et al. 2000). It forms the basement of the southern
part of Hungary, and is bounded by the Mid Hungarian (or
Zagreb-Zemplín) Line to the north, while it can be fol-

lowed over the state boundary to Northern Croatia and
Serbia-Montenegro, and to Western Transylvania (Roma-
nia) in the southern and eastern directions, respectively.
Szederkényi (1984) divided the pre-Alpine (mostly
Variscan) basement complexes of the Tisza Unit into two
major parts: the Parautochthon Unit and the South Hun-
garian Nappe. Fülöp (1994) distinguished three tectonic
units on the basis of Mesozoic sedimentary facies zones:
the Mecsek, Villány-Bihar (VBU), and Békés-Codru Units
(BCU). The first two represent the Parautochthon Unit,
whereas the latter is equivalent to the South Hungarian
Nappe. These tectonic units are of Late Cretaceous age, so
their presence in pre-Alpine tectonic reconstructions is not
unambiguous. The prevailing rocks are paragneisses, mica
schists and granitoids with minor amphibolites and – in
some areas – marble intercalations.

In general, the first metamorphic event recorded in the

Tisza Unit in Hungary is characterized by Barrow-type
amphibolite facies regional metamorphism (330—350 Ma).
Árkai (1984) and Árkai et al. (1985) calculated peak con-
ditions of 500—600 ºC and 0.5—0.9 GPa for gneisses, mica

Fig. 1. a – Main tectono-stratigraphic units of the Pannonian Basin and neigh-
bouring areas. Box indicates area enlarged in Fig. 1b; b – Pre-Tertiary geological
map of the Tisza Unit. 1 – Cretaceous-Paleogene flysch deposits, 2 – Mesozoic
rocks, 3 – Permian and Carboniferous molasse-type rocks, 4 – metamorphic and
igneous basement rocks.

schists and intercalated amphibolites. This
event was overprinted by a low-pressure
Variscan event in the BCU (270—330 Ma). No
reliable isotopic ages older than Variscan are
available for the metamorphic basement of the
Hungarian part of the Tisza Unit (Lelkes-
Felvári et al. 1996; 2003). Recently, Balen et
al. (2006) reported pre-Variscan (428± 25 and
444± 19 Ma) monazite U-Pb ages for the medi-
um-grade metamorphism of garnet-bearing
mica schists from the Slavonian Mts (NE
Croatia). Eo-Alpine tectonism caused a very
low- to low-grade prograde metamorphism in
the Permian-Mesozoic rocks beneath the over-
thrusted Variscan basement which shows
strong Alpine retrogression (Árkai et al. 2000).
Some parts of the Variscan basement in the
BCU were affected by Eo-Alpine amphibolite
facies metamorphism (Algyő basement-high,
Horváth & Árkai 2002). Non-metamorphic
Late Paleozoic overstep sequences were depos-
ited on different parts of the Tisza Unit
(Kovács et al. 2000). Lower Permian and Me-
sozoic formations occur in the entire area;
however they do not form a continuous cover
above the basement. Lower-Middle Miocene
conglomerates contain the basement rocks as
pebbles. In these clastic sediments exotic rock
types such as eclogites and garnetiferous am-
phibolites occur (Horváth et al. 2003).

The metamorphic basement near the Algyő

basement-high is built up mainly by Variscan
andalusite-bearing metapelites (Ar-Ar cooling
ages of muscovite are in the range of 318—
321 Ma, Lelkes-Felvári et al. 2003). Sm-Nd
analyses of a garnet concentrate yielded a Per-
mian age of 273± 7 Ma (Lelkes-Felvári et al.
2002, 2003). According to Lelkes-Felvári et al.

109

P-T PSEUDOSECTIONS IN KFMASH, KMnFMASH, NCKFMASH AND NCKMnFMASH SYSTEMS

(2003) the garnet, together with K-feldspar and crystal-
shape relics of andalusite formed during a high-T/low-P
event. Kyanite aggregates (sometimes with staurolite) re-
placed andalusite, and the rock pile suffered intense mylo-
nitic deformation during the Eo-Alpine tectono-
metamorphic event. An Ar-Ar age dating on muscovite
(85.5±1.2 Ma) from the Uszi-2 sample proved the Eo-Al-
pine age of metamorphism (Lelkes-Felvári et al. 2003).
Slightly younger Ar-Ar plateau ages (68.4—84.3 Ma) were
published by Balogh & Pécskay (2001). Horváth & Árkai
(2002) studied the mica schist samples from the Algyő base-
ment-high in detail and found several garnet generations
distinguished by petrographic studies and compositional
features as well: 1. complex zoned garnet porphyroblasts
with 3 growth stages; 2. zoned garnets with S-shaped inclu-
sion trails; 3. small, homogeneous garnets with composi-
tions similar to the rim of the large porphyroblasts. Having
applied various thermobarometric methods and calibrations
(TWEEQU of Berman 1991 and THERMOBAROMETRY
by Kohn & Spear 1995) combined with P-T calculations
from garnet zoning profiles (GIBBS program of Spear &
Menard 1989) a P-T path with increasing P and T condi-

Fig. 2. Petrographic images from the Uszi-2 sample. a – Hypidioblastic staurolite (St) inclusions in garnet (Grt); b – Corona of small
garnets on larger garnet porphyroblasts; c – idioblastic garnet aggregate in biotite-muscovite-quartz matrix; d – Ky aggregate
(pseudomorphs after andalusite?) in muscovite-rich matrix. Note the discordant orientation of the aggregate in relation to the main foli-
ation of the rock represented by biotites (horizontal in the thin section).

tions was established. Peak conditions of 650±30 ºC and
0.9±0.1 GPa were reached and followed by subsequent
cooling (Fig. 9 in Horváth & Árkai 2002).

Petrography and mineral chemistry

Chemical analyses of minerals were carried out by a

JEOL JXCA-733 electron microprobe equipped with 3
WDS in the Institute for Geochemical Research, Hungari-
an Academy of Sciences, Budapest. The measuring condi-
tions were: 15 kV acceleration voltage; 40 nA sample
current; electron beam with a diameter of 5 m; 5 s count-
ing time. Matrix effects were corrected by using the ZAF
method. The following standards were used for quantita-
tive analysis: orthoclase (K, Al, Si), synthetic glass (Fe,
Mg, Ca), spessartine (Mn), rutile (Ti) and albite (Na).

The garnet-staurolite mica schist sample (Uszi-2) chosen

for this study contains garnet porphyroblasts that are near-
ly homogeneous and have idioblastic staurolite inclusions
(Fig. 2a). The garnets are Alm

78—80

Prp

12—15

Grs

3—4

Sps

1—3

whereas some small areas (< 10 m) have lower Alm and

110

HORVÁTH

Table 1: Representative mineral chemical analyses from the Uszi-2 garnet-staurolite mica schist.

Fig. 3. Zoning profiles of garnets from Uszi-2 mica schist: a – large
porphyroblast; b – small garnet corona (shaded) on large porphyro-
blast.

Prp, and higher Grs and Sps contents (Alm

Prp

Grs

Sps

). These areas occur inside the garnets near the

rims (Fig. 3a), but they are not related to any specific mi-
crotextural positions (e.g. cracks, grain-edges), so tenta-
tively they are interpreted as relics of an earlier lower-T
event. The absence of growth zonation in garnet either re-
sults from fast garnet growth near the metamorphic peak or
from diffusive smoothing out of an initial growth zonation
(Zeh 2001). Small garnets with compositions similar to the
rims of the large porphyroblasts form overgrowths on them
(Figs. 2b and 3b) or occur as aggregates in the matrix
(Fig. 2c). XFe [Fe

/(Fe

+Mg)] varies between 0.84 and

0.87. Staurolites occur either in garnet, or they are present
in the matrix as large (up to 5—7 mm) porphyroblasts. These
porphyroblasts are rimmed or cross-cut by kyanite aggre-
gates. Staurolites have quite uniform XFe values ranging
from 0.77 to 0.82. There is no zonation found inside the
analysed samples or differences in chemical composition re-
gardless of their textural position (i.e. inclusion in a garnet
core or relic porphyroblast in the matrix). The matrix assem-
blage consists of muscovite, biotite, kyanite, quartz, plagio-
clase and accessory minerals such as zircon, tourmaline,
ilmenite and magnetite. Fine-grained aggregates of kyanite
appear in the muscovite-rich matrix (Fig. 2d); they were in-
terpreted as pseudomorphs after andalusite by Szederkényi
(1984). The orientation of the kyanite aggregates is some-
times discordant to the main foliation of the matrix. Matrix
biotites are chemically homogeneous with XFe values of
0.44—0.52 and 1.5—1.9 wt. % TiO

. Biotite inclusions in

staurolite have lower XFe values (0.38—0.41). Muscovites
have 6.2—6.3 Si atoms p.f.u., and plagioclases are An

10—15

Retrograde chlorite replaces biotite and garnet; it occurs in
larger quantities in some of the Újszentiván samples than in

111

P-T PSEUDOSECTIONS IN KFMASH, KMnFMASH, NCKFMASH AND NCKMnFMASH SYSTEMS

Fig. 4. P-T pseudosections (a – KFMASH, b – KMnFMASH) of
Uszi-2 mica schist sample calculated with PERPLEX. Box indicates
calculated peak P-T conditions after Horváth & Árkai (2002).

the others from different part of the basement (e.g. Ferenc-
szállás or Algyő samples). Representative mineral chemical
analyses are listed in Table 1.

Pseudosections

Phase relations are best illustrated and understood using

pseudosections (quantitative phase diagrams) where the
bulk composition of a rock is incorporated into the calcu-
lations. The PERPLEX calculations (Connolly 1990; Con-
nolly & Petrini 2002) were performed with an updated
(2002) version of the internally-consistent thermodynamic
data set of Holland & Powell (1998), and the solid solu-
tion models incorporated in the software package. The
THERMOCALC pseudosection modelling (Powell et al.
1998) was undertaken with the 3.25 version of the soft-
ware and the internally-consistent thermodynamic dataset
5.5 (August 2004 upgrade). The datafile coding of the ac-
tivity-composition relationships of the minerals used in
the MnNCKFMASH calculations is that of Stowell &
Tinkham (2003). For the KFMASH and KMnFMASH sys-
tems we used the coding of Wei et al. (2004), and for
NCKFMASH the coding of White et al. (2001). All the
constructed pseudosections are in the P-T range 0.2—
1.2 GPa and 450—700 ºC with quartz and H

O in excess.

The bulk composition in the system NCKMnFMASH is

( S i O

: A l

: M g O : F e O : K

O : C a O : N a

O : M n O )

77.07:27.55:4.94:9.01:5.58:0.57:2.29:0.27 in molar amounts,
with the corresponding KFMASH, KMnFMASH and NCK-
FMASH compositions involving the omission of MnO,
CaO or Na

O if necessary. Major element composition of

the bulk rock sample was determined using a Perkin Elmer
5000 atomic absorption spectrophotometer (AAS), after
digestion with lithium metaborate. In addition to the AAS
technique, permanganometric (FeO), gravimetric (SiO

TiO

, H

O and P

) and volumetric (CO

) methods were

applied.

KFMASH and KMnFMASH pseudosections using
PERPLEX

The KFMASH pseudosection is scarcely consistent with

the 650±30 ºC and 0.9±0.1 GPa peak P-T range (Fig. 4a).
The largest field in the peak P-T range is occupied by gar-
net-biotite-staurolite-muscovite without kyanite which is
not in accordance with the petrographic observation that
the sample has abundant kyanite. Moreover, garnet-free as-
semblages such as biotite-staurolite-muscovite, biotite-kya-
nite-muscovite or biotite-staurolite-kyanite-muscovite form
parts of the P-T range. Chlorite is stable up to about 570 ºC
at 0.9 GPa. The lower stability of biotite is at 520 ºC at low
pressures. Garnet is not stable below 0.9 GPa. In the KMnF-
MASH pseudosection (Fig. 4b) only three assemblages are
stable at the above-mentioned P-T conditions: garnet-bi-
otite-staurolite-muscovite, garnet-biotite-staurolite-kyanite-
muscovite and garnet-biotite-kyanite-muscovite. The bi-
otite-staurolite-muscovite assemblage disappears when
MnO is added to the calculations.

KFMASH and KMnFMASH pseudosections using
THERMOCALC

The main feature of the KFMASH pseudosection

(Fig. 5a) is that garnet is stable only above 0.8 GPa at
570—650 ºC. The lower stability of biotite is at about
550 ºC. The lower stability of the peak assemblage of gar-
net-biotite-muscovite-quartz is at 620—650 ºC and 0.7—
1.2 GPa, partly covering the calculated peak P-T range of
650± 30 ºC and 0.9± 0.1 GPa. Staurolite is stable in the P

115

P-T PSEUDOSECTIONS IN KFMASH, KMnFMASH, NCKFMASH AND NCKMnFMASH SYSTEMS

garnet-biotite-staurolite-muscovite, the isopleths change
dramatically, they show slightly negative slope (Fig. 7a).
The same holds true for garnet and biotite isopleths in gar-
net-biotite-muscovite field. Isopleths for measured garnet
and biotite compositions intersect at around 670—690 ºC
and 0.8—0.95 GPa matching fairly well with peak condi-
tions calculated (650±30 ºC and 0.9±0.1 GPa). Biotite in-
clusions in staurolite have x(bi)=0.38—0.41 presuming
higher-P conditions. Staurolite x(st) isopleths yield P-T
conditions at around 650 ºC and 0.9 GPa.

KMnFMASH system

X(g) [=Fe

/(Fe

+Mg+Mn)] isopleths in garnet-biotite-

staurolite-muscovite field show an interesting feature
(Fig. 7b). They have a very gentle negative slope with in-
creasing x(g) until 0.8 GPa. Then the x(g)=0.73 isopleth
has a turned U-shape and x(g) starts to decrease with a
steep negative slope. The m(g) [=Mn/(Mn+Fe

+Mg)]

isopleths do not show this effect. Garnet isopleths in the
narrow garnet-biotite-staurolite-kyanite-muscovite field
are nearly vertical, while in garnet-biotite-kyanite-musco-
vite they have a gentle negative slope with decreasing
values. X(bi) [=Fe

/(Fe

+Mg+Mn)] isopleths are strongly

P-dependent aside from the garnet-biotite-staurolite-kyan-
ite-muscovite field (Fig. 7c). X(st) [=Fe

/(Fe

+Mg+Mn)]

isopleths have a steep positive slope in garnet-biotite-
staurolite-kyanite-muscovite field, while in the upper
part of the garnet-biotite-staurolite-muscovite field they
feature a steep negative slope (Fig. 7d). This steepness
becomes almost vertical then turns into a positive slope
near the area where the x(g) isopleths change their slope.
The x(g) isopleths are lower in the pseudosections than
those measured, while m(g) isopleths are higher. Calcu-
lated biotite and staurolite isopleths weakly match the
measured ones.

NCKFMASH system

The z(g) [= Ca/(Ca+Fe

+Mg)] isopleths are nearly ver-

tical and decrease with increasing T from 16 to 3 in the T
range of 550—680 ºC (Fig. 8a). They closely match the
measured ones in garnet-biotite-kyanite-muscovite (3—
4). The areas with higher Ca content in garnet (10) proba-
bly formed at lower T. On the other hand, the x(g) [= Fe

(Fe

+Mg)] isopleths are lower than the measured ones

similarly to the KMnFMASH system. This is assigned to
diffusional re-equilibration. The x(bi) isopleths are
strongly P-dependent in garnet-biotite-staurolite-musco-
vite-plagioclase and garnet-biotite-kyanite-muscovite-
plagioclase, but change drastically in biotite-staurolite-
plagioclase (Fig. 8b). They are nearly uniform with
x(bi)=0—57—0.58. The x(bi) isopleths in biotite-chlorite-
staurolite-muscovite-plagioclase are subvertical showing
strong T-dependence. The x(st) isopleths have a similar be-
haviour to biotite ones (Fig. 8c). The ca(pl) [= Ca/(Ca+Na)]
isopleths decrease intensively with increasing T (Fig. 8d).
The calculated isopleths in the peak assemblage closely
match the measured ones (10—15).

NCKMnFMASH system

The x(g) isopleths are lower than the measured ones in

the peak assemblage. The m(g) [=Mn/(Mn+Fe

+Mg+Ca]

and z(g) [= Ca/(Ca+Fe

+Mg+Mn)] isopleths fairly match

the measured ones. They yield P-T conditions at around
630—660 ºC and 0.9—1.0 GPa (Fig. 9a). The slopes of the
m(g) isopleths are positive in the pseudosection with values
decreasing with T. The z(g) isopleths have a gentle positive
slope in the garnet-biotite-kyanite-muscovite-plagioclase
and garnet-biotite-staurolite-muscovite-plagioclase fields,
but become steeper in garnet-biotite-chlorite-staurolite-
muscovite-plagioclase. The slopes turn negative in garnet-
chlorite-staurolite-muscovite-plagioclase. The x(g) isop-
leths are subvertical in this field and become gently
negative in the biotite-bearing fields. The biotite and stau-
rolite isopleths are similar to the ones in NCKFMASH in
chlorite-free assemblages (Fig. 9b and c). They turn nearly
vertical when chlorite becomes part of the mineral assem-
blage. They fairly match the measured compositions in the
peak assemblage. The ca(pl) isopleths have a similar slope
to the NCKFMASH ones but plagioclase is Ca-poorer in
this system at the same P-T conditions (Fig. 9d). Ca(pl) is
between 10 and 16 in the peak assemblage fitting well with
the measured plagioclase compositions.

Discussion

PERPLEX vs. THERMOCALC pseudosections in
KFMASH and KMnFMASH systems

Quantitative phase diagrams (pseudosections) were cal-

culated in the P-T range 0.2—1.2 GPa and 450—700 ºC in the
systems KFMASH and KMnFMASH for a representative
garnet-staurolite mica schist sample from the basement of
the southwestern part of the Pannonian Basin (Figs. 4 and 5)
with PERPLEX and THERMOCALC computer programs.
Despite the fact that the two programs use different ways of
approaching the calculation of mineral equilibria involving
solid solutions, one should expect the same results with the
two calculation methods, especially when the same mineral
solution models are used. Hoschek (2004) pointed out sev-
eral differences in the resulting P-T pseudosections when he
compared PERPLEX, DOMINO and THERMOCALC for a
kyanite eclogite. Our results have similar conclusions. The
KFMASH pseudosections show similarities in the medium-
pressure range (0.5—0.8 GPa). Both are dominated by the
assemblages biotite-staurolite-muscovite, biotite-staurolite-
sillimanite-muscovite, biotite-sillimanite-muscovite and bi-
otite-staurolite-kyanite-muscovite (all assemblages have
quartz in excess). The same holds true for the high-pres-
sure part. There are some differences in the low-pressure
region. This could be due to the different Al

SiO

triple

point calibrations used by the two software packages, but
additional, currently unknown causes can be assumed. In
the KMnFMASH system the medium- and high-pressure
parts of the P-T diagrams match each other closely. The
differences in the low-pressure regions are probably due to

117

P-T PSEUDOSECTIONS IN KFMASH, KMnFMASH, NCKFMASH AND NCKMnFMASH SYSTEMS

Fig. 10. P-T path of the Uszi-2 mica schist sample. Bold arrow
indicates the outlined P-T path, the dashed line indicates the un-
certain parts (see text for explanation). Box indicates calculated
peak P-T conditions after Horváth & Árkai (2002). Roman
numbers (I and II) are described in the text.

during their P-T history (Szederkényi 1984; Horváth & Ár-
kai 2002; Lelkes-Felvári et al. 2003). The P-T pseudosec-
tion calculated in the NCKFMASH system is used,
because it can model garnet composition and the effect of
Na- and Ca-bearing phases (plagioclase, paragonite, and
zoisite) commonly found in metapelites (Fig. 10). The
main difference between NCKFMASH and NCKMnF-
MASH systems is the presence of garnet at low-pressure
(< 0.4 GPa) in the latter. Since the Sps content in our gar-
nets are generally low (1—3) and the other rock-forming
minerals (biotite, staurolite) have MnO contents near the
detection limit of EMP, the system NCKFMASH is
enough to quantitatively model the mineralogical and
mineral compositional changes in our sample during the
P-T evolution. Within the framework of the NCKMnF-
MASH system the outlined P-T path (see below) would be
similar to the NCKFMASH system, only the garnet-free as-
semblages would be garnet-bearing.

Szederkényi (1984) described the fine-grained kyanite

aggregates as pseudomorphs after andalusite. If this holds
true then the prograde P-T path starts from biotite-stauroli-
te-andalusite-muscovite-plagioclase or biotite-andalusite-
muscovite-plagioclase (point I in Fig. 10). There are sever-
al boreholes in the BCU near to the study area with
Variscan andalusite-bearing mica schists and gneisses. Ar-
Ar muscovite age data are 305—322 Ma in these rock types
(Lelkes-Felvári et al. 2003). The pseudomorphs could also
represent former sillimanite according to Lelkes-Felvári et
al. (2003). The P-T path barely touches the biotite-stauro-

lite-sillimanite-muscovite-plagioclase field, so this assump-
tion is questionable. Horváth & Árkai (2002) described sil-
limanite-bearing mica schist from the Algyő-54 borehole
located about 3 km north of the Uszi-2 borehole. Garnet
zoning patterns and thermobarometric data indicate growth
during prograde conditions from 610 ºC and 0.4 GPa to
peak P-T conditions at 650—680 ºC and 0.5—0.6 GPa. An-
other possibility is that the pseudomorphs were staurolites
(point II). Petrographic evidence shows that staurolite por-
phyroblasts occurring in the matrix are replaced by kyanite.
In this case, the P-T path starts from chlorite-staurolite-mus-
covite-plagioclase or biotite-chlorite-staurolite-muscovite-
plagioclase and passes the same fields. No chlorite inclu-
sions were found in staurolite beside biotite and quartz. So
far there is no conclusive evidence to decide between the
two possibilities. The assemblage biotite-staurolite-quartz
with matrix muscovite and plagioclase is stable at 560—
650 ºC and 0.35—0.6 GPa in NCKFMASH. This is the earli-
est detectable mineral assemblage in the sample. The chem-
ical compositions of the inclusion biotites and the host
staurolites do not match with the calculated ones in the bi-
otite-staurolite-muscovite-plagioclase field. Instead, they
reflect compositions in the garnet-biotite-staurolite-musco-
vite-plagioclase field suggesting chemical re-equilibration
during the prograde evolution of the rock at about 620—
640 ºC and 0.7—0.8 GPa. Lelkes-Felvári et al. (2003) also
mentioned K-feldspar porphyroclasts with relic staurolite,
andalusite (now pseudomorphed by kyanite) and garnet as
a pre-tectonic assemblage with respect to their S

defined

by fine-grained biotite, muscovite, plagioclase and quartz.
We were unable to find K-feldspar in our sample, so this as-
sumption could not be confirmed by our study. In the calcu-
lated pseudosections there is no joint stability field for stau-
rolite, andalusite and K-feldspar with or without garnet
which is in accordance with other published grids (e.g.
White et al. 2001; Wei et al. 2004). The lowest T where K-
feldspar enters the parageneses is about 630 ºC at low-pres-
sure conditions, and even higher at high-pressure. The K-
feldspar is most probably a relic of an earlier high-T
metamorphic event. Continuous T increase in the garnet-bi-
otite-staurolite-muscovite-plagioclase field followed, and
then kyanite appeared at the expense of staurolite. The peak
conditions represented by garnet-biotite-kyanite-musco-
vite-plagioclase were reached at about 650—670 ºC and
0.7—0.8 GPa. The observed prograde path is similar to the
one in Horváth & Árkai (2002); they used the Gibbs meth-
od on zoned garnets (Spear & Menard 1989) on mica schist
samples from Ferencszállás (4 km east from the present
study). After the thermal peak the cooling path passed the
garnet-biotite-staurolite-muscovite-plagioclase field again
and ended in chlorite-bearing assemblages. Since no pla-
gioclase breakdown was observed, the cooling path was be-
low 0.6 GPa at 600 ºC, and was above 0.4 GPa due to the
absence of sillimanite. The P-T path presented here is out-
lined in Fig. 10, the dashed line indicates parts of the P-T
path which are not constrained by mineral chemical data.

Several authors published Eo-Alpine Ar-Ar ages on

muscovites from the Algyő basement-high (85.5±1.2 Ma
in Lelkes-Felvári et al. 2003; 68.4—84.3 Ma in Balogh &

118

HORVÁTH

Pécskay 2001). These data are interpreted as cooling ages.
The oldest ages from this tectonic unit were determined by
Sm-Nd analyses from garnet- and kyanite-bearing mylo-
nitic rocks (Lelkes-Felvári et al. 2002, 2003). The isoch-
ron calculated from a garnet concentrate yielded a Permi-
an age of 273±7 Ma. Unfortunately, no chemical data from
garnet are presented in these papers. Taking into consider-
ation these features the Permian ages obtained by the
above-mentioned authors could be mixed Variscan and
Alpine ages or represent a separate thermal event. Horváth
& Árkai (2002) showed strongly zoned garnets from the
Algyő basement-high. They evaluated the core of the
zoned garnets as possible relics of a Variscan metamorphic
event, and calculated 520—560 ºC and 0.8—1.0 GPa for the
Ca-rich outer core using inclusions of muscovite, plagio-
clase and quartz. In the calculated NCKFMASH pseudosec-
tion (and in NCKMnFMASH as well) garnet-chlorite-
chloritoid-muscovite-paragonite would be the stable as-
semblage in the studied sample which is not confirmed in
this study. A polymetamorphic evolution with garnet-
staurolite (±andalusite/sillimanite) in the first (Variscan)
cycle, and garnet-kyanite in the second (Eo-Alpine) is
evaluated here. The Variscan cooling ages from an-
dalusite-bearing mica schists outside the Algyő basement-
high support this idea. On the other hand, there is no con-
clusive evidence for the Variscan age of the andalusites in
the studied sample, and within the framework of the pre-
sented P-T pseudosections both andalusite and kyanite
(and even sillimanite) can be the products of a single (Eo-
Alpine) tectono-metamorphic event.

Conclusions

1 – P-T pseudosections (quantitative phase diagrams)

in the KFMASH and KMnFMASH systems were calculat-
ed with PERPLEX and THERMOCALC from garnet-stau-
rolite mica schist originating from the basement of the
Pannonian Basin (Hungary). The resultant pseudosections
are similar to each other highlighting the effectiveness of
both software packages. Differences only occur in the low-
pressure regions. The peak P-T conditions calculated with
thermobarometric methods (650±30 ºC and 0.9±0.1 GPa,
Horváth & Árkai 2002) are confirmed by the mineral as-
semblage garnet-biotite-kyanite-muscovite in both sys-
tems. The mineral composition isopleths in the KFMASH,
KMnFMASH, NCKFMASH and NCKMnFMASH systems
mostly model the mineral chemical changes.

2 – The P-T path of the Uszi-2 mica schist was outlined

in the NCKFMASH system. The biotite- and quartz inclu-
sions in host staurolite with matrix muscovite and plagio-
clase are stable at 560—650 ºC and 0.35—0.6 GPa. The
mineral compositions reflect chemical re-equilibration in
the garnet-biotite-staurolite-muscovite-plagioclase field
during the prograde evolution of the rock at about 620—
640 ºC and 0.7—0.8 GPa. During simultaneous P-T in-
crease garnet started to form at about 0.6 GPa in the
garnet-biotite-staurolite-muscovite-plagioclase field, and
then kyanite appeared at the expense of staurolite. The

peak conditions in garnet-biotite-kyanite-muscovite-pla-
gioclase-quartz were reached at about 650—670 ºC and
0.7—0.8 GPa. After the T peak the cooling path passed the
garnet-biotite-staurolite-muscovite-plagioclase-quartz

field

again and ended in chlorite-bearing assemblages. Since no
plagioclase breakdown was observed, the cooling path
was under 0.6 GPa at 600 ºC, and was over 0.4 GPa due to
the absence of sillimanite. The observed P-T path is simi-
lar to the one published by Horváth & Árkai (2002).

3 – From the study area andalusite and sillimanite as

possible Variscan phases were speculated by Szederkényi
(1984) and Lelkes-Felvári et al. (2003). No petrographic
evidence was found for the existence of these phases in
the studied sample. All three Al

SiO

polymorphs are

present in the study area; so the possibility of their forma-
tions during one single P-T path could not be excluded.
Together with published Ar-Ar age data on matrix musco-
vite (85.5±1.2 Ma) from the Uszi-2 sample, the presented
P-T pseudosections also support the Eo-Alpine age of am-
phibolite facies metamorphism in the basement of the Tis-
za Unit (southwestern Pannonian Basin).

Acknowledgments:

This work is financially supported by

the Hungarian National Science Fund (grant number:
OTKA F 047322). James Connolly is thanked for an earlier
review of the paper and his continuous help with PERPLEX
pseudosection calculations. The author received valuable
help from Martin Racek regarding THERMOCALC calcu-
lation procedures. Péter Árkai provided the rock sample for
this study and gave his support which is greatly acknowl-
edged here. Thanks are given here for the helpful and con-
structive reviews by Alexander Proyer and Kálmán Török,
and the editorial handling of Marian Janák.

References

Albee A.L. 1965: A petrogenetic grid for the Fe-Mg silicates of pel-

itic schists. Amer. J. Sci. 263, 512—536.

Árkai P. 1984: Polymetamorphism of the crystalline basement of

the Somogy-Dráva Basin (Southwestern Transdanubia, Hun-
gary). Acta Mineral. Petrogr. 26, 129—153.

Árkai P., Bérczi-Makk A. & Balogh K. 2000: Alpine low-T pro-

grade metamorphism in the post-Variscan basement of the
Great Plain, Tisza Unit (Pannonian Basin, Hungary). Acta
Geol. Hung. 43, 1, 43—63.

Árkai P., Nagy G. & Dobosi G. 1985: Polymetamorphic evolution

of the South-Hungarian crystalline basement, Pannonian Ba-
sin: geothermometric and geobarometric data. Acta Geol.
Hung. 28, 3—4, 165—190.

Balen D., Horváth P., Tomljenović B., Finger F., Humer B., Pamić

J. & Árkai P. 2006: A record of pre-Variscan Barrovian re-
gional metamorphism in the eastern part of the Slavonian
Mountains (NE Croatia). Mineral. Petrology 87, 143—162.

Balogh K. & Pécskay Z. 2001: K/Ar and Ar/Ar geochronological

studies in the PANCARDI region. Acta Geol. Hung. 44, 2-3,
281—299.

Berman R.G. 1991: Thermobarometry using multi-equilibrium cal-

culations. A new technique, with petrologic applications. Ca-
nad. Mineralogist 29, 835—855.

Connolly J.A.D. 1990: Multivariable phase diagrams: an algorithm

119

P-T PSEUDOSECTIONS IN KFMASH, KMnFMASH, NCKFMASH AND NCKMnFMASH SYSTEMS

based on generalized thermodynamics. Amer. J. Sci. 290,
666—718.

Connolly J.A.D. & Petrini K. 2002: An automated strategy for cal-

culation of phase diagram sections and retrieval of rock prop-
erties as a function of physical conditions. J. Metamorphic
Geology 20, 697—708.

Droop G.T.R. & Harte B. 1995: The effect of Mn on the phase rela-

tions of medium-grade pelites: constraints from natural assem-
blages on petrogenetic grid topology. J. Petrology 36,
1549—1578.

Fülöp J. 1994: Geology of Hungary. Paleozoic II. Akadémiai Ki-

adó, Budapest, 1—445 (in Hungarian).

Géczy B. 1973: The origin of Jurassic faunal provinces and the

Mediterranean plate tectonics. Ann. Univ. Sci. Budapest R.
Eötvös Nom. Sect. Geol. 16, 99—114.

Ghent E.D. & Stout M.Z. 1981: Geobarometry and geothermome-

try of plagioclase-biotite-garnet-muscovite assemblages. Contr.
Mineral. Petrology 76, 92—97.

Holland T.J.B. & Powell R. 1990: An enlarged and updated inter-

nally consistent thermodynamic dataset with uncertainties and
correlations: the system K

O-Na

O-CaO-MgO-MnO-FeO-

-Al

-TiO

-SiO

-C-H

-O

. J. Metamorphic Geology 8,

89—124.

Holland T.J.B. & Powell R. 1998: An internally consistent thermo-

dynamic data set for phases of petrological interest. J. Meta-
morphic Geology 16, 309—343.

Horváth P. & Árkai P. 2002: Pressure-temperature path of

metapelites from the Algyő-Ferencszállás area, SE Hungary:
thermobarometric constraints from coexisting mineral assem-
blages and garnet zoning. Acta Geol. Hung. 45, 1, 1—27.

Horváth P., Kovács G. & Szakmány Gy. 2003: Eclogite and garnet

amphibolite pebbles from Miocene conglomerates (Pannonian
Basin, Hungary): implications for the Variscan metamorphic
evolution of the Tisza Megaunit. Geol. Carpathica 54, 6,
355—366.

Hoschek G. 2004: Comparison of calculated P-T pseudosections for

a kyanite eclogite from the Tauern Window, Eastern Alps,
Austria. Eur. J. Mineral. 16, 59—72.

Kázmér M. & Kovács S. 1985: Permian-Paleogene paleogeography

along the Eastern part of the Insubric-Periadriatic lineament
system: evidence for the continental escape of the Bakony-
Drauzug Unit. Acta Geol. Hung. 28, 1—2, 71—84.

Kohn M.J. & Spear F.S. 1995: Thermobarometry. Dept. Geol.,

Renssealer Polytechnic Inst.

Kovács S. 1982: Problems of the “Pannonian Median Massif” and

the plate tectonic concept. Contributions based on the distribu-
tion of Late Paleozoic-Early Mesozoic isopic zones. Geol. Rd-
sch. 71, 617—640.

Kovács S., Haas J., Császár G., Szederkényi T., Buda Gy. & Nagy-

marosy A. 2000: Tectonostratigraphic terranes in the pre-Neo-
gene basement of the Hungarian part of the Pannonian area.
Acta Geol. Hung. 43, 3, 225—328.

Lelkes-Felvári Gy., Árkai P. & Sassi F.P. 1996: Main features of the

regional metamorphic events in Hungary: a review. Geol. Car-
pathica 47, 4, 257—270.

Lelkes-Felvári Gy., Frank W. & Schuster R. 2002: Basement evolu-

tion of the Great Hungarian Plain: Variscan, Permo-Triassic
and Alpine metamorphism. Földt. Közl. 132, 1, 125—127.

Lelkes-Felvári Gy., Frank W. & Schuster R. 2003: Geochronologi-

cal constraints of the Variscan, Permian-Triassic and Eo-Al-
pine (Cretaceous) evolution of the Great Hungarian Plain
basement. Geol. Carpathica 54, 5, 299—315.

Mahar E.M., Baker J.M., Powell R., Holland T.J.B. & Howell N.

1997: The effect of Mn on mineral stability in metapelites.
Contr. Mineral. Petrology 99, 226—237.

Pattison D.R.M. & Tracy R.J. 1991: Phase equilibria and ther-

mobarometry of metapelites. In: Kerrick D.M. (Ed.): Contact
metamorphism reviews in mineralogy. Vol. 26. Mineral. Soc.
Amer., Washington, DC, 105—206.

Powell R. & Holland T.J.B. 1990: Calculated mineral equilibria in

the pelite system KFMASH (K

O-FeO-MgO-Al

-SiO

-H

O).

Amer. Mineralogist 75, 367—380.

Powell R., Holland T.J.B. & Worley B. 1998: Calculating phase dia-

grams involving solid solutions via non-linear equations, with
examples using THERMOCALC. J. Metamorphic Geology 16,
577—588.

Spear F.S. & Cheney J.T. 1989: A petrogenetic grid for pelitic

schists in the system SiO

- Al

- FeO-MgO-K

O-H

O. Contr.

Mineral. Petrology 101, 149—164.

Spear F.S. & Menard T. 1989: Program GIBBS: a generalized

Gibbs method algorithm. Amer. Mineralogist 74, 942—943.

Stowell H.H. & Tinkham D.K. 2003: Integration of phase equilibria

modeling and garnet Sm-Nd chronology for construction of
P-T-t paths: examples from the Cordilleran Coast Plutonic
Complex, USA. In: Vance D., Muller W. & Villa I. (Eds.):
Geochronology: linking the isotopic record with petrology
and textures. Geol. Soc. Spec. Publ. 220, 119—145.

Stowell H.H., Taylor D.L., Tinkham D.K., Goldberg S.A. &

Ouderkirk K.A. 2001: Contact metamorphic P-T-t paths from
Sm-Nd garnet ages, phase equilibria modelling, and ther-
mobarometry. J. Metamorphic Geology 19, 645—660.

Szederkényi T. 1984: Crystalline basement of the Great Hungarian

Plain and its geological connections. D.Sc. Thesis, Budapest,
1—216 (in Hungarian).

Tinkham D.K. & Stowell H.H. 2000: Lack of evidence for loading

during garnet growth: Southern Nason terrane, Cascades Crys-
talline Core, Washington. Geol. Soc. Amer., Cordilleran Sec-
tion, 2000 Annual Meeting, 32, A-71.

Vance D. & Mahar E. 1998: Pressure-temperature paths from P-T

pseudosections and zoned garnets; potential, limitations and
examples from the Zanskar Himalaya, NW India. Contr. Min-
eral. Petrology 132, 225—245.

Wei C.J., Powell R. & Clarke G.L. 2004: Calculated phase equilibria

for low- and medium-pressure metapelites in the KFMASH and
KMnFMASH systems. J. Metamorphic Geology 22, 495—508.

White R.W., Powell R. & Holland T.J.B. 2001: Calculation of par-

tial melting equilibria in the system Na

O-CaO-K

O-FeO-

MgO-Al

-SiO

-H

O (NCKFMASH). J. Metamorphic

Geology 19, 139—153.

Worley B. & Powell R. 1998: Singularities in NCKFMASH (Na

O-

CaO-K

O-FeO-MgO-Al

-SiO

-H

O). J. Metamorphic Geol-

ogy 16, 169—188.

Zeh A. 2001: Inference of a detailed P-T path from P-T pseudosec-

tions using metapelitic rocks of variable composition from a
single outcrop, Shackleton Range, Antarctica. J. Metamorphic
Geology 19, 329—350.