Lithos 116 (2010) 300–309

Contents lists available at ScienceDirect

Lithos
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / l i t h o s

Chemical, petrological and mass balance constraints on the textural evolution of

pelitic enclaves
Antonio M. Álvarez-Valero a,b,⁎, Leo M. Kriegsman b,c
a
Instituto Andaluz de Ciencias de la Tierra (CSIC - University of Granada), Campus Fuentenueva, Faculty of Sciences, 18002 Granada, Spain
b
NCB Naturalis, Darwinweg 2, 2333 CR, Leiden, The Netherlands
c
Department of Geology, University of Turku, FIN-20014 Turun Yliopisto, Finland

a r t i c l e i n f o a b s t r a c t

Article history: Reaction histories in xenoliths are commonly based on textural analysis alone and are often highly disputed.
Received 14 February 2009 In this paper we propose a more rigorous approach that combines textural analysis, reaction balance
Accepted 25 September 2009 calculations and quantitative P–T pseudosections using model reactions. Examples highlighting the power of
Available online 12 October 2009
this approach are: disequilibrium behaviour of plagioclase in a partial melting reaction; and the relation
between biotite–plagioclase intergrowths replacing garnet and biotite patches overgrowing ilmenite.
Keywords:
Enclaves
An important message is that textural analysis in xenoliths, and probably also in other rock types, must be
Xenoliths reaction balance complemented by chemical (compositions and reaction balance) and petrological (coherent P–T projections
Melt production or pseudosections) constraints to distinguish between alternative generic possibilities. This method sheds
Textural analysis new light on a discussion on the melt volume produced during biotite dehydration melting in migmatites.
Pseudosections © 2009 Elsevier B.V. All rights reserved.

1. Introduction
Textural analysis is a powerful technique to distinguish reactants
and products in many metamorphic reactions (e.g. Carmichael, 1969;
Spry, 1969; Kriegsman and Hensen, 1998; Vernon, 2004, 2007) and is
critical in defining the sense of P–T paths (e.g. Corbett and Phillips,
1981; Hand et al., 1994; Hensen et al., 1995; Vernon, 1996; Cenki
et al., 2002; Álvarez-Valero and Kriegsman, 2007). It does, however,
suffer from a number of limitations: (i) whereas the role of major
phases is commonly straightforward, the role of minor phases may be
more difficult to discern and this is vitally important for some
accessory phases used for direct dating of the textures (see papers in
Vance et al., 2003); (ii) phases that have been completely consumed
are often not considered; (iii) the number of phases required to
balance the reaction can only be assessed through a qualitative or
(semi)quantitative chemical approach; and (iv) phases that have not
been totally consumed but no longer record their initial compositions.
Hence, textural analysis can provide constraints on metamorphic
processes, but is generally insufficient for a complete description.
On the other hand, chemical analysis alone can give first-order
estimates of thermodynamic properties such as pressure and tempera-
⁎ Corresponding author. Instituto Andaluz de Ciencias de la Tierra (CSIC -
University of Granada), Campus Fuentenueva, Faculty of Sciences, 18002 Granada,
Spain. Tel.: +34 958246286; fax: +34 958243384.
E-mail addresses: antonioa@ugr.es (A.M. Álvarez-Valero),
Kriegsman@naturalis.nnm.nl (L.M. Kriegsman).
ture, but cannot define (except through local zoning in minerals such as
garnet) the sense of reaction progress along the P–T path (e.g. isobaric
heating versus cooling, isothermal decompression versus burial). In most
cases, the determination of P–T vectors requires textural information,
such as defining reactants and products, or defining successive mineral
assemblages. In addition, many key parameters of rock histories, e.g. trace
element distributions or isotopic compositions (including age patterns)
require contextual information (see papers in Vance et al., 2003).
In this paper we provide examples from pelitic enclaves in fossil
volcanic suites within the Neogene Volcanic Province (NVP) of the Betic
Cordillera (SE Spain). After Zeck's (1970) seminal work, the area has
recently been reinvestigated in detail (e.g. López-Ruiz and Rodríguez-
Badiola, 1980; Fernández-Soler, 1996; Benito et al., 1999; Cesare et al.,
1997, 2003a,b, 2009; Cesare and Gómez-Pugnaire, 2001; Álvarez-Valero
et al., 2005, 2007; Álvarez-Valero and Kriegsman, 2007, 2008). Here, we
focus on problematic textures that were not or insufficiently discussed
in previous papers. We will pursue a multidisciplinary approach of
combining petrography (textures), petrology (P–T pseudosections) and
chemistry (mineral chemistry, reaction balance).
Vernon (2007) discussed the textural evolution of pelitic xenoliths/
enclaves using examples from, amongst others, the NVP. Some of his
conclusions are that garnets in these xenoliths/enclaves are not peritectic,
and that the high-Al pelitic enclaves are not restites to the host dacitic
magma. We emphasize here that such conclusions cannot be arrived at
with confidence by textural analysis alone, and that much can be gained
by systematically combining textural and chemical analysis, including
reaction balancing and calculating volume ratios of product phases.

0024-4937/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.lithos.2009.09.009
 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309
2. Geological setting
The Alborán Basin, located between the Betic Cordillera and the
Moroccan Rif, formed by syn- to late orogenic extension on the site of a
former mountain belt (Balanyá and García-Dueñas, 1987). Currently, it
is widely recognized as a key area for the understanding of the origin
and geodynamic evolution of the Betic Cordillera (e.g. Benito et al., 1999;
López-Ruiz et al., 2002; Platt et al., 2003). It is thus a focus for
petrological, geochemical and geophysical investigations. Although
many conflicting models have been advanced for the origin of the
Alborán Basin (see Doblas et al., 2007), the current tectonomagmatic
models favour extensional hypotheses, as originally proposed by Doblas
and Oyarzun (1989), and upwelling of plume-contaminated sub-
lithospheric material replacing the delaminating subcontinental litho-
spheric mantle during the post-collisional transition from a subduction-
related to an intraplate setting (Duggen et al., 2005).
The NVP is the most important volcanic domain of continental Spain
(Fig. 1), related to the Betic Cordillera and the Alborán Sea (Zeck, 1970;
Fernández-Soler, 1996). Scattered volcanic outcrops occur within a
250 km long, NE-trending zone from Cabo de Gata (southwestern limit)
to Mar Menor (northeastern limit). Some lavas include up to 15 vol.%
crustal material (Zeck, 1970). The high-K calc-alkaline volcanic series of
El Hoyazo, Mazarrón and Mar Menor contains the most abundant
xenoliths of residual partially melted crust (Zeck, 1970; Cesare et al.,
1997, Álvarez-Valero and Kriegsman, 2007). The crustal xenoliths were
transported to the surface from depth by volcanic processes and were
rapidly quenched, thus presenting a frozen image of deep-crustal
processes that operated in SE Spain during the Neogene.
Fig. 1. Geographical location of Miocene, high-K calc-alkaline-shoshonitic volcanics of the Neogene Volcanic Province within the Alborán Domain. Sketches of the tectonic setting of
the Betic–Rif Cordilleras on the upper left and lower right sides.
301
Geothermobarometry on the xenoliths of El Hoyazo indicates
equilibration at 850 ± 50 °C and 6 ± 1 kbar (Cesare et al., 1997, 2003a,
b; Álvarez-Valero and Kriegsman, 2007). The same T range applies at
lower P (c. 4 kbar) at Mazarrón (Cesare and Gómez-Pugnaire, 2001;
Álvarez-Valero et al., 2007; Álvarez-Valero and Kriegsman, 2007),
whereas Mar Menor samples record even lower P at slightly higher T
(Álvarez-Valero and Kriegsman, 2007, 2008). Zircon and monazite
SHRIMP ages indicate that anatexis occurred at c. 9.7–8.1 Ma at El
Hoyazo, c. 9.1 Ma at Mazarrón and c. 10 Ma at Mar Menor, while the
eruption has been dated at c. 6.3 Ma at El Hoyazo, 9.1 Ma at Mazarrón
and 9.5 Ma at Mar Menor (Zeck and Williams, 2002; Cesare et al.,
2003b, 2009).
The NVP mineral equilibria are related to three successive KFMASH
univariant reactions: (i) Bt + Sil + Qtz = Grt + Crd + Kfs + M; (ii) Bt +
Sil + Grt = Spl + Crd + Kfs + M; and (iii) Bt + Crd + Grt = Opx + Spl +
Kfs + M (mineral abbreviations after Kretz, 1983; M = melt). These
and related divariant reactions have operated along a P–T segment
with increasing T and slightly decreasing P (Álvarez-Valero and
Kriegsman, 2007, 2008; Álvarez-Valero et al., 2007).
3. Methodology
Complementary to the above studies, we here focus on problem-
atic and ambiguous textures that were insufficiently discussed
because they were felt not to be of consequence to the overall
shape of the P–T paths. We will pursue a multidisciplinary approach of
combining petrography (textures), petrology (P–T pseudosections)
and chemistry (mineral chemistry, reaction balance).
302 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309

3.1. Textures, mineral chemistry and reaction balancing Table 1

(a) NCKFMASHT bulk composition (wt.% normalised to 100) used for modelling the
Perple_X pseudosections of Figs. 4 and 6 for textures 1 and 2, respectively; (b) Comparison
Our method is similar to the one employed earlier to migmatites between observed and calculated mineral data at best-fit pressures and temperatures for
by, among others, Kriegsman and Hensen (1998) and Cenki et al. textures 1 and 2.
(2002). The first authors showed, in one sample, that some crystals of
(a) (b)
spinel were rimmed by cordierite, and others by garnet and
sillimanite. Choosing at first quartz (not observed, and possibly Pl1–Pl2 Bt–TiBt Melt-bearing Mineral composition Best P–T fit
texture texture modelled comparisons — (kbar — °C)
totally consumed) as a reactant, led to inconsistent P–T vectors and P–
texture Fe2+/(Fe2+ + Mg2+)
T estimates. By contrast, involving the quartz component of melt as a
reactant solved the inconsistencies and, in addition, led to an Observed Calculated

explanation of the previously neglected biotite growth. Reaction SiO2 36.45 45.87
balance involving a melt-consuming reaction is likely to produce a TiO2 0.86 8.87 Grt 0.87 0.81 4.3–730
Al2O3 33.84 17.29 Pl1–Pl2 Crd 0.54 0.50 4.3–730
hydrous alkaline phase like biotite. Hence, textural analysis coupled
FeO 20.07 13.74 Spl 0.80 0.81 4.3–730
with petrological analysis and mass balance led to an interpretation MgO 3.54 3.14 Bt 0.66 0.63 4.3–730
capable of explaining more of the observations. CaO 1.41 1.72
Cenki et al. (2002) observed coarse-grained orthopyroxene and Na2O 1.00 2.11 Bt 0.63 0.64 4.4–800
plagioclase at the edges of large corroded crystals of garnet in partially K2O 2.22 5.15 Bt–TiBt Crd 0.49 0.49 4.4–800
H2O 0.61 2.01 Grt 0.85 0.85 4.4–800
melted metapelites. As a first interpretation, this could be due to the
classic solid–solid reaction: Grt + Qtz = Opx + Pl. However, the
observed volume ratio of Pl/Opx was twice as high as predicted by a multi-equilibrium method; (ii) it ensures that the relative P–T
from reaction balancing. In addition, the unzoned garnet crystals had conditions of different local assemblages are better constrained than by a
very low Ca contents, normal for metapelites, whereas the grossular disparate set of independent thermobarometers; and (iii) it indicates the
component required for the inferred reaction is high, which explains reaction progress expected along any P–T trajectory, and modal changes
why the reaction is common in metabasites. The solution was to can be calculated. However, as this paper aims to explain ambiguous or
involve a Ca-bearing reactant and the only candidate present was the complicated textures, it should not be considered as the single critical
melt phase, consistent with the presence of extensive leucosome method for resolving petrological problems, but instead as a complemen-
nearby. It was concluded that the melt had provided the extra tary tool to textural analysis and reaction balance calculations.
plagioclase component during partial crystallisation and reaction with Because, in principle, phase diagrammes show equilibrium states
the solid matrix upon cooling. In this case, textural analysis coupled that reacting systems attempt to approach, they are of value even for
with reaction balance and volume calculations, proved more reliable interpreting disequilibrium processes in the enclaves. They can be
than an interpretation based merely on textures. used to constrain P–T conditions of earlier assemblages when they
In summary, the method we propose is as follows: (i) textural have been overprinted by subsequent cation exchange or the partial
analysis, using the wide array of methods described in many past progress of net transfer reactions.
publications (e.g. Vernon, 2004); (ii) defining a simplified chemical Modelling was done in the system Na2O–CaO–K2O–FeO–MgO–
system for each reaction texture and using the phase rule to see if the Al2O3–SiO2–H2O–TiO2 (NCKFMASHT), and involved the following
number of phases fits; (iii) proper reaction balancing, including the phases: (i) for texture 1 (see next section): garnet, cordierite, aluminous
calculation of volume ratios of product phases; and (iv) rigorous spinel, biotite, sillimanite, plagioclase1 (Ab-rich), plagioclase2 (An-rich),
quantification using thermodynamic software, whenever high-quality K-feldspar, ilmenite and silicate liquid; (ii) for texture 2 (see next
thermodynamic data exist. section): biotite, cordierite, plagioclase, K-feldspar, ilmenite, silicate
Reaction balancing can be done via specially designed software liquid, garnet, sillimanite and aluminous spinel.
(e.g., C-Space: Torres-Roldán et al., 2000) or on a computer The model for silicate melt, as well as activity–composition models
spreadsheet. We have chosen the second option, as it allows one to for Grt and Bt, is taken from White et al. (2007); for Spl and hydrous
quickly see what happens when a compositional variable is changed. Crd from Holland and Powell (1998); and for feldspars from Fuhrman
A balance can only be achieved in a pseudo-univariant system, which and Lindsley (1988). The activity–composition model for Bt in texture
is a consequence of the mathematical fact that, in normal cases, one 2 is from Tajcmanová et al. (2009). The model biotite is saturated in Ti
can only solve for n variables by having n independent equations. by incorporating enough TiO2 to allow excess ilmenite.
Being interested only in ratios and not absolute quantities, one X-ray fluorescence bulk analyses for these kinds of anatectic rocks
reaction coefficient is set to 1, reducing it to n − 1 equations for n have limited use for phase diagramme modelling studies, because each
variables (Thompson, 1982). In order to balance reactions with microdomain may represent a different melting “equilibrium” stage
variance >1, solid solution phases can be split into end-members (see (different equilibrium volumes sensu Stüwe, 1997) within the same
the example of a divariant reaction in Section 4.1) or into an end- rock (or thin section). Therefore, the melting evolution is investigated
member with exchange vectors (see the example of a quadrivariant using discrete modal bulk compositions to avoid the construction of an
reaction in Section 4.2). After the calculations, the phases can be unreliable or even meaningless pseudosection. Each microdomain can
reconstituted from the end-members or from the end-members and be treated as a single chemical and microstructural system (Álvarez-
exchange vectors. Valero and Kriegsman, 2007, 2008; Álvarez-Valero and Waters, 2010).
Although the domainal bulk composition approach works well in a
3.2. Metamorphic modelling procedure general sense, White et al. (2008) describe the related potential
uncertainties in the final diagramme, e.g., the shape, size and P–T
The pressure–temperature (P–T) estimates were made with the location of fields, mineral proportions and compositions. The precision
thermodynamic database of Holland and Powell (1998, with updates) to define the equilibrium volume and effective bulk rock composition
and the Perple_X computer program (Connolly, 2005) using the bulk may be influenced by diffusive interaction with more distal parts of the
compositions shown in Table 1a. Modal abundances of all minerals as rock, a process that may change the effective composition of the defined
a function of T (Table 1b) and molar proportions were calculated from domain during reaction progress (see also Stüwe, 1997; Álvarez-Valero
the werami.exe routine of the Perple_X software set. and Waters, 2010). Ignoring these complications, bulk compositions of
Phase diagramme modelling has a number of advantages in microdomains were determined by a combination of point counting
metamorphic petrology: (i) it allows P–T conditions to be determined using Imager software and visual estimation (from a summation of
 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309 303

phase compositions and modal proportions in thin sections, Table 1a).

The sources for electron microprobe mineral analyses are from earlier
studies (Álvarez-Valero, 2005; Álvarez-Valero et al., 2007).

4. Ambiguous textures and their interpretation

4.1. Texture 1: role of plagioclase in melting processes

(disequilibrium behaviour?)

The role of plagioclase in the NVP melting processes has been

difficult to assess due to ambiguous growth and breakdown textures.
Álvarez-Valero et al. (2007), and Álvarez-Valero and Kriegsman
(2007) considered that continuous reactions in the NCKFMASH system
involve two different plagioclase generations: plagioclase1 as reactant
and plagioclase2 as product. Plagioclase1 is a major component of the
matrix, in contact with or enclosing fibrolite, biotite, cordierite, garnet
and glass (Fig. 2A,B). Plagioclase2 is present within coronas around
garnet crystals (Fig. 3A), and rims hercynite crystals within the same
coronas (Fig. 3B). In addition, it shows well-defined crystal faces and
growth zoning where it occurs next to pockets of devitrified interstitial
glass. Locally, at El Hoyazo, plagioclase occurs as both euhedral crystals
next to glass pockets replacing garnet and within the matrix (Fig. 3C).
Hence, the question arises whether plagioclase2 is a prograde product
in all cases, or may, in some cases, have crystallized from the melt
(cooling/retrograde event). Textural information is insufficient here,
because the morphology of both plagioclase types ranges from

Fig. 3. Plane-polarized light microscopy views of texture 1: (A) Pl2 within coronas
around Grt crystals; (B) Pl2 rimming Spl crystals within the same previous corona;
(C) Euhedral Pl2 crystals next to pockets of glass within the matrix.

anhedral to euhedral. Chemical analyses, however, reveal clear

Fig. 2. Plane-polarized light microscopy views of xenolith matrix: (A) Spl–Crd xenolith
type matrix with Pl1 and minor Fib, Bt, Crd, Grt and glass; (B) Skeletal high-Ti Bt
crystals, within the matrix of Sil and Pl1. G.I.: Glass inclusions.
differences in Ca, Na and K contents as a function of the textural
setting: (i) plagioclase1 associated with fibrolite and skeletal biotite in
the matrix ranges from An33 to An49; (ii) plagioclase2 rimming
hercynite crystals within coronas around garnet ranges from An50 to
An54; (iii) plagioclase2 in symplectitic intergrowths with biotite shows
An55 to An58; and (iv) plagioclase2 growing next to pockets of
devitrified interstitial glass range from An58 to An81. There are no
systematic differences between core and outer rim of individual
plagioclase crystals.
304 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309

4.1.1. Analysis and interpretation
The thin plagioclase rims around euhedral hercynite crystals form
coronas around garnet (Fig. 3A,B) and have recently been interpreted in
terms of a melt-consuming process, resulting either from rapid cooling
of the xenoliths after eruption of the host dacite, or from minor cooling
before uprise and eruption (Álvarez-Valero et al., 2007).
For this particular texture, in which Spl and glass are separated by
An-rich Pl2, local mass balance does not work. The only way to balance
Spl + melt as reactants and Pl2 as a product is to significantly vary the
local melt composition, for which we have no supporting chemical
observations. It has been reported from other areas, however, that
glass may have different colours and compositions related to local
melting reactions or local rock types or layers (e.g., Sawyer, 2001;
Braun and Kriegsman, 2001; Holness and Sawyer, 2008). In addition,
the textures do not suggest Spl breakdown, because both Spl and Pl2
are euhedral.
A more logical explanation is related to melt chemistry and
sluggish diffusion in plagioclase. The melt norm composition is An18,
whereas all plagioclases in these samples have higher An values.
Hence, taking out an An18 component must have shifted the An
content of relict plagioclase to higher An values. If diffusion in
plagioclase were fast enough, one would expect a single plagioclase
with An >18%. However, slow diffusion in plagioclase (see, e.g.
melting experiments by Johannes and Holtz, 1996) may lead to partial
dissolution of Pl1 and precipitation of Pl2 from the melt, where Pl2 has
higher An values than Pl1 to compensate for low An in melt. Pl2
nucleates on Spl but does not consume Spl, consistent with the
euhedral shape of both minerals.
To test this hypothesis, we performed reaction balance calcula-
tions in the NCKFMASH system using two different methods with
phase compositions as close as possible to the measured chemical
compositions. In an 8-component chemical system, 9 phases or phase
components are needed to balance a reaction. There are 8 phases
present in this divariant equilibrium: Bt, Sil, Grt, Crd, Spl, Pl, Kfs, and
M. Replacing Pl and Kfs by the three end-members Ab, An and Or gives
the correct number of phase components. After balancing, Pl and Kfs
compositions can be reconstituted from the three end-member
components. This leads to a balance of the form Bt + Sil + Grt + Pl
(= Ab + An) = Spl + Crd + Kfs (= Or) + M (Table 2; reaction 1a). In
this reaction, the consumed Pl has a composition An18, identical to the
melt An content, but slightly lower if Kfs is also given an Ab
component. This first balance predicts the observed Spl/Crd volume
ratio (~1), but cannot explain two types of Pl textures.
In the second method, we used the measured Pl1 and Pl2
compositions (on average An41 and An52, respectively) with Or. This
leads to a reaction of the form Bt + Sil + Grt + Pl1 = Spl + Crd + Kfs
(= Or) + Pl2 + M (Table 2; reaction 1b). The second reaction again
predicts the observed Spl/Crd volume ratio (~1), but in addition yields
the observed Spl/Pl2 volume ratio (~5). Hence, we conclude that this
may be the correct procedure. Adding Ti to the system, incorporated
in Bt and the additional phase Ilm, has a very limited influence on the
mass balance.
The above interpretation has a bearing on how to consider two
generations of the same mineral. In view of diffusive barriers, we have
considered Pl1 and Pl2 as thermodynamically autonomous phases. Hence,
the balanced reaction (reaction 1b), using 9 phases in an 8-component
chemical system, should be considered univariant.
Two different generations (and compositions) of the same mineral,
i.e. Pl1 and Pl2, were therefore used as input data for thermodynamic
modelling with Perple_X. The microdomain composition was derived by
estimating phase proportions in a volume including both plagioclase
compositions (±Crd, ±Bt, ±Grt), thus extending into the surrounding
matrix (entire domain shown on Fig. 3C). In this case, Ti is an essential
component. The P–T pseudosection of Fig. 4A shows a narrow field at
T > 750 °C and P < 5 kbar in which the commonly observed assemblage
Grt–Crd–Sil–Kfs–Spl–Melt ± Bt–Ilm coexists with a small amount (up to
10 mol%) of plagioclase, enough to account for the Pl1 within the reactant
matrix and for Pl2 rimming the Spl porphyroblasts. The predicted
changes in modal abundances (Fig. 4B, Table 1b) match both the reaction
balance calculations of reacting Grt + Pl1 + Sil ±Bt and product phases
Spl + Crd + Kfs and melt, and the observed volume ratio Spl/Pl2 around
5, and Spl/Crd around 2 (Table 2). The Pl2 modal abundance increases
from c. 3.6 to 5.0% along an appropriate P–T vector constrained by the
mass balance calculations and the textural observations. The data
suggest that, in this model, Pl1 is indeed consumed, and there is some Pl2
produced, consistent with the observations. The consistency of the P–T
vector with the overall P–T path (Álvarez-Valero et al., 2007; Álvarez-
Valero and Kriegsman, 2007) strengthens this interpretation.
4.2. Texture 2: role of melt and two types of biotite in garnet–ilmenite
breakdown
Different types of biotite have been described from the NVP
enclaves (e.g. Cesare et al., 1997, 2003a,b; Álvarez-Valero et al., 2007;
Álvarez-Valero and Kriegsman, 2007): (i) corroded high-Ti (3.7–
8.5 oxide wt.%) biotite crystals, locally skeletal, within the matrix and
enclosed by garnet (Fig. 2B), are attributed to the peritectic melting
reaction Bt + Sil + Pl = Spl + Crd + Kfs + M + Ilm; and (ii) biotite–
plagioclase symplectitic intergrowths replacing both garnet and,
locally, garnet–spinel (Fig. 5A,B) at Mazarrón may attest to an event
of crystallization from and reaction with cooling glass (Álvarez-Valero
et al., 2007; Álvarez-Valero and Kriegsman, 2007) according to the

Table 2
Phase compositions used in reaction balance calculations of texture 1: Spl with Pl rim and Crd near corroded Grt (no Ti, so no Ilm).

Input phases (molar basis); 9 phases/components per reaction (see coefficients) Reaction 1a Reaction 1b

Bt Sil Spl Crd Grt Or Melt An Ab Pl1 Pl2

SiO2 2.70 1.00 0.00 5.00 3.00 3.00 10.07 2.00 3.00 2.59 2.48
AlO1.5 1.60 2.00 2.00 4.00 2.00 1.00 2.27 2.00 1.00 1.41 1.52
FeO 1.78 0.00 0.79 1.08 2.61 0.00 0.19 0.00 0.00 0.00 0.00
MgO 0.92 0.00 0.21 0.92 0.39 0.00 0.02 0.00 0.00 0.00 0.00
CaO 0.00 0.00 0.00 0.00 0.00 0.00 0.09 1.00 0.00 0.41 0.52
NaO0.5 0.00 0.00 0.00 0.00 0.00 0.00 0.43 0.00 1.00 0.59 0.48
KO0.5 1.00 0.00 0.00 0.00 0.00 1.00 0.76 0.00 0.00 0.00 0.00
H2O 1.00 0.00 0.00 0.50 0.00 0.00 2.90 0.00 0.00 0.00 0.00
Reaction 1a coefficients 3.91 12.82 − 11.10 − 2.00 1.60 − 3.15 − 1.00 0.09 0.43 – –
Reaction 1b coefficients 3.91 12.82 − 11.10 − 2.02 1.60 − 3.15 − 1.00 – – 1.60 − 1.08
1a Bt + Sil + Grt + Plg = Spl + Crd + Kfs + Liq
Pl consumed has 18% An (neglecting Ca in Grt), but lower when Kfs has an Ab component Equal Spl and Crd produced (volume ratio 0.94)
1b Bt + Sil + Grt + Pl1 = Spl + Pl2 + Crd + Kfs + Liq
Equal Spl and Crd produced (volume ratio 0.94) Calculated Spl/Pl2 volume ratio is 5.2, close to observed values

Pl1, Pl2 and M compositions as analysed; idealized composition for other phases. Each of the two balanced reactions uses 9 phase components, adding up to 8 phases in reaction 1a
(divariant) and 9 phases in reaction 1b (univariant if Pl1 and Pl2 are considered thermodynamically autonomous phases: see text). Key ratios and interpretations are added.
 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309 305

Fig. 4. (A) P–T pseudosection for texture 1 highlighting the area corresponding to the best estimate (dashed circle) between modelled and observed phase proportions and
compositions (see also Table 1b), and part of the predicted P–T path from the textures and reaction balance. Ilm in excess (NCKFMASHT system). Darker fields shading symbolizes
higher variance; (B) modal abundance evolution diagramme of all phases as a function of T.

reaction Grt (±Spl ± Kfs) + Crd + M = Bt + Pl. However, the textur-
ally ambiguous biotite generations also have different chemical
compositions: (iii) low-Ti (<2 oxide wt.%) biotite crystals intergrown
with plagioclase (An50–80) and cordierite crystals, and apparently
replacing garnet (Fig. 5C); and (iv) biotite-rich patches (around 5 wt.%
Ti oxide) typically including a core of anhedral ilmenite within the
matrix and associated with garnet breakdown (Fig. 5A,C). These two
textures are commonly associated and may possibly be explained by a
single reaction. Breakdown of Ilm and Grt and growth of Bt–Pl are
texturally unambiguous (Fig. 5A,C). However, the role of melt (melt
production or consumption), and the role of Crd and, where present,
Kfs are texturally ambiguous.
4.2.1. Analysis and interpretation
In the Bt–Pl texture, there are two textural elements that may or
may not be combined: (a) Bt–Pl symplectitic intergrowths in which
the observed Bt/Pl volume ratio is about 0.5 (Fig. 5A); and (b) large
patches of Bt that overgrow Ilm (Fig. 5A,C). The Ti content of Bt in
(b) is significantly higher than in (a). In view of relatively fast
diffusion in biotite (e.g., Spear, 1993), this implies non-equilibrium at
the scale of the patch, probably promoted by the relative isolation of
small biotites in the symplectites. If both textural elements are
combined, the observed Bt/Pl volume ratio is about 1.0–1.5.
We performed a variety of reaction balance calculations in the NCK
(FM)ASHT system, reducing Fe and Mg to one component, with the
306 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309
Fig. 5. Plane-polarized light microscopy views of texture 2: (A), (B) Low-Ti Bt–Pl
symplectitic intergrowths replacing both Grt and, locally, Grt–Spl; (C) High-Ti Bt-rich
patches typically enclosing anhedral Ilm within the matrix.
phases Grt, Bt, Pl, Kfs, Ilm, Spl, Crd and melt. First, a Bt composition close
to the measured composition was used, recalculated either on the basis
of full hydrogenation (H saturation) in the case of Bt (H-sat) or on the
basis of H-deficiency (see Cesare et al., 2003a) in the case Bt (H-def).
Feldspars were separated into their components Ab, An and Or (see
Section 4.1). The observed textures where Spl is present (e.g. Fig. 5B) are
best explained when either Spl or Sil are treated as reactant phases,
giving the reaction Grt + M + Spl/Sil + Or + Ilm = Bt + Pl + Crd
(Table 3, reaction balance 2a and 2b). H-saturated and H-deficient Bt
gave similar results.
To explain the far more common reaction domains lacking Spl, this
phase was left out of the balance. Biotite was separated into phase
components (see previous section) and was thus allowed to vary its
composition. We used Al-Tschermak, Ti-Tschermak and Ti-oxy
substitutions as vectors modifying the measured Bt composition
(Table 3, balance 2c–h). The first two vectors correspond to fully
hydrogenated Bt, the third to H-deficient Bt (see Cesare et al., 2003a).
After balancing, the composition of Bt produced or consumed in the
reaction is assembled from the components, in proportion to their
reaction coefficients. If Bt composition varies as a function of P and T
(i.e. reaction progress), the composition used for balancing is a com-
ponent or exchange vector of biotite. Hence, it should correspond to
the compositional difference between the biotites at starting and final
conditions in thermodynamic calculations.
In all cases, the calculated Bt composition lies far outside the
natural range, indicating that we are dealing with an exchange vector
in biotite. In balances 2c–f (Table 3) it is oversaturated in Ti, Si, FeMg
and undersaturated in Al, which is only possible in solid solution with
preexisting Bt that shifted its composition during reaction progress.
This is consistent with thermodynamic modelling (see below). By
implication, the estimated Bt/Pl volume ratio is not representative of
the reaction products and cannot, therefore, be used to assess the
reaction balance, contrary to Section 4.1. There is little difference
between using H-saturated or H-deficient Bt, all calculations produce
an overall reaction of the form Grt + M + Or + Ilm = Bt + Pl + Crd. The
conclusion that melt has been consumed together with Ilm is an
important result to be used in a companion paper (Kriegsman and
Álvarez-Valero, 2010-this issue) where we compare migmatite and
xenolith textures.
Results most consistent with textural observations are reached
with three feldspar components (An, Ab, and Or). By contrast, using
measured feldspar compositions (Pl and Kfs) and allowing Bt to vary
through multiple components (Table 3, balance 2g–h) gives a balance
with Bt and Ilm on the wrong side of the reaction, inconsistent with
the observations. In addition, the calculated Bt components or
exchange vectors deviate strongly from those in balance 2c–f.
For pseudosection calculations, the microdomain composition was
derived from an estimate of phase proportions in a volume including
both biotite types (i.e., low-Ti and high-Ti), enlarging into the
adjacent matrix (entire domain shown on Fig. 5C). Thermodynamic
modelling of the observed solid assemblage Bt–Pl–Ilm–Melt ± Grt–
Crd–Kfs–Sil–Spl with Perple_X cannot constrain the field of coexisting
Bt types (as opposed to the texture 1 case with two different Pl) that
have very different Ti contents in a plausible range of bulk com-
positions and is, therefore, inconsistent with textures and reaction
balance. This could partly reflect the poorly known properties of Ti–Bt,
but may also imply that the two biotite varieties cannot be regarded as
thermodynamically isolated phases, in contrast with the two
plagioclases in Section 4.1. It could also mean that thermodynamic
equilibrium is maintained only at a local level, as testified by the
different Ti contents of the two different types of biotite.
Attempts to model the domain using low-Ti and high-Ti Bt
separately with Perple_X show practically the same equilibrium phase
diagramme (slightly higher T for high-Ti Bt). The P–T pseudosection
of Fig. 6A shows a broad reaction field, which corresponds to reaction
Grt + Kfs + Ilm + M = Bt + Pl + Crd (Table 3, balance 2c–f), assuming
that none of the reactants are entirely consumed. Changes in modal
abundance (Fig. 6B, Table 1b) along a cooling path from 820 to 780 °C
at 4.4 kbar are compatible with the textural observations and mass
balance. This estimate is consistent with the P–T path for the NVP
suites as described by Álvarez-Valero et al. (2007) and Álvarez-Valero
and Kriegsman (2007).
The pseudosection (Fig. 6A) and modal abundance diagramme
(Fig. 6B) also suggest that Bt was preexistent prior to the melt-
 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309 307

Table 3
Phase compositions used in reaction balance calculations of texture 2: Bt patches rimming Ilm, close to Bt–Pl intergrowths and corroded Grt (e.g., Fig. 5C), with Crd present nearby.

Input phases (molar basis); 9 phases and phase components per reaction (see reaction coefficients)

Bt Bt Tsch Ti–Tsch Ti–oxy An Or Ab Melt Ilm Grt Crd Spl Pl Kfs

(H-sat) (H-def)

SiO2 2.65 2.65 − 0.50 0.00 0.00 2.00 3.00 3.00 10.07 0.00 3.00 5.00 0.00 2.68 3.00
AlO1.5 1.70 1.70 1.00 0.00 0.00 2.00 1.00 1.00 2.27 0.00 2.00 4.00 2.00 1.32 1.00
(FeMg)O 2.05 2.35 − 0.50 − 2.00 − 1.00 0.00 0.00 0.00 0.22 1.00 3.00 2.00 1.00 0.00 0.00
CaO 0.00 0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.09 0.00 0.00 0.00 0.00 0.32 0.00
NaO0.5 0.09 0.09 0.00 0.00 0.00 0.00 0.00 1.00 0.43 0.00 0.00 0.00 0.00 0.68 0.18
KO0.5 0.91 0.91 0.00 0.00 0.00 0.00 1.00 0.00 0.76 0.00 0.00 0.00 0.00 0.00 0.82
H2O 1.00 0.70 0.00 0.00 − 1.00 0.00 0.00 0.00 2.90 0.00 0.00 0.50 0.00 0.00 0.00
TiO2 0.30 0.30 0.00 1.00 1.00 0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00
Reaction 2a coefficients 1.31 – – – – 0.07 − 0.64 0.19 − 0.73 − 0.39 − 1.00 1.61 − 2.36 – –
Reaction 2b coefficients – 1.11 – – – 0.05 − 0.64 0.11 − 0.49 − 0.33 − 1.00 1.30 − 1.77 – –
Reaction 2c coefficients 0.39 – − 1.39 – – 0.03 − 0.14 0.08 − 0.28 − 0.12 − 1.00 0.84 – – –
Reaction 2d coefficients – 0.40 − 1.26 – – 0.02 − 0.19 0.06 − 0.23 − 0.12 − 1.00 0.80 – – –
Reaction 2e* coefficients 0.83 – − 2.24 0.50 – 0.04 − 0.40 0.12 − 0.46 − 0.75 − 1.00 1.02 – – –
Reaction 2f* coefficients – 0.48 − 1.29 – 0.08 0.02 − 0.26 0.05 − 0.22 −0.22 − 1.00 0.79 – – –
Reaction 2g coefficients − 0.05 – − 0.54 − 0.50 – – – – − 0.10 0.52 − 1.00 0.66 – 0.03 0.15
Reaction 2h coefficients – − 0.16 − 1.08 – − 0.57 – – – − 0.30 0.62 − 1.00 0.87 – 0.09 0.46
(*10 phases and phase components, with Tschermak
in Bt varied manually in the matrix)

2a,b Real Bt composition M + Grt + Spl + Or + Ilm = Bt + Pl + Crd Combined Bt composition (component/exchange vector)
Ilm on reactant side and Bt on product side of the reaction, fitting textural
2c 2d 2e 2f 2g 2h
observations; Similar for H-saturated and H-deficient Bt

2c,d Variable Al in Bt M + Grt + Or + Ilm = Bt + Pl + Crd 4.43 4.23 4.00 4.00 − 2.50 − 0.63 SiO2
Ilm on reactant side and Bt on product side of the reaction, fitting textural − 1.87 − 1.45 − 1.00 − 1.00 12.00 8.26 AlO1.5
observations; Similar for H-saturated and H-deficient Bt 3.83 3.93 2.19 3.54 − 22.46 − 4.37 (FeMg)O

2e,f Variable Al and Ti in Bt M + Grt + Or + Ilm = Bt + Pl + Crd 0.09 0.09 0.09 0.09 0.09 0.09 NaO0.5
Ilm on reactant side and Bt on product side of the reaction, fitting textural 0.91 0.91 0.91 0.91 0.91 0.91 KO0.5
observations; Reasonable Bt component/vector; approximate Bt/Pl volume ratio 1.00 0.70 1.00 0.54 1.00 − 2.74 H2O
reasonable for high Na and H2O in melt

2g,h Real feldspars, variable Bt M + Grt + Bt = Crd + Kfs + Pl + Ilm 0.30 0.30 0.91 0.46 9.98 3.74 TiO2
Ilm on product side and Bt on reactant side of the reaction, not fitting textural
observations; Unreasonable Bt composition

All balanced reactions use 9 phase components (including exchange vectors), except 2e–f where the Al-Tschermak exchange vector in Bt was manipulated manually. Components
add up to 8 phases in reaction 2a–b (divariant); 7 phases in 2c–h (trivariant). In reality, Fe–Mg exchange adds one degree of variance to all reactions, becoming trivariant and
quadrivariant, respectively. Key ratios and interpretations are added.

consuming reaction, in agreement with our mass balance calculations.
However, several other aspects deviate from textural observations
and mass balance: (i) melt is still present at the low-T end of the
diagramme, although there is no textural evidence for quenched glass,
suggesting that melt was consumed; and (ii) the modal abundance of
Kfs is very high and even increases during cooling, but only a sparse
amount of Kfs is observed in thin sections. The role of Kfs depends
strongly on the H2O content of melt (e.g., Carrington and Watt, 1995):
on the cooling path, Kfs may be consumed if the melt is relatively
water-rich, but will be produced if it is water-undersaturated. The
output file of Perple_X shows that melt is strongly increasing its water
content during cooling, implying that a water-undersaturated
component has been consumed. A constant water content during
the cooling reaction would inhibit Kfs growth or result in no change in
Kfs modal abundance.
5. General discussion and conclusions
Vernon (2007) discussed metamorphic reactions in xenoliths/
enclaves using only textural observations. His conclusions are, in
some cases, quite different from the conclusions obtained from the
combination of microstructures, chemistry and thermodynamics.
Vernon (2007) is correctly quoting Cesare's (2000) observation that
garnet in El Hoyazo is corroded and that melt is produced in the
process. In our view, however, his conclusion that garnet is not a
peritectic phase is incorrect. On the contrary, the detailed analysis of
Álvarez-Valero et al. (2005, 2007) and Álvarez-Valero and Kriegsman
(2007), using textural observations and P–T pseudosections, shows
that most garnet grains had formed during earlier biotite dehydration
melting. Hence, early peritectic garnet is broken down during a
subsequent melt-producing reaction. This strengthens our conclusion
that textural observations alone are insufficient to derive P–T paths or
to fully constrain partial melting processes.
Cesare et al. (2003a) used a variety of chemical and spectroscopic
techniques to convincingly demonstrate that at least some Ti-rich
biotites from El Hoyazo xenoliths are H-deficient. We accept the
validity of his data, but do not share his conclusion that melt volumes
calculated by Kriegsman (2001) using H-saturated biotite were
overestimated by a factor 2. First of all, it remains to be demonstrated
that the mineralogical model by Cesare et al. (2003a) has general
validity and that the biotites from Finland are also H-deficient.
Secondly, biotites involved in dehydration melting generally change
composition (see Section 4.2). Hence, accepting the H-deficiency of
biotites remain after melting does not mean that the biotite
component that was consumed during the reaction had that same
composition. It seems much more likely to us that the biotite
component consumed was H-(over)saturated to allow the shift
from a low-Ti, high-H biotite in the starting material to a high-Ti,
low-H relict biotite. For example, the reaction balance of the first
texture (Section 4.1) was done with an H-saturated biotite
308 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309
Fig. 6. (A) P–T pseudosection for texture 2 highlighting the area with the best match
between modelled and observed phase proportions and compositions (dashed circle,
see also Table 1b), and part of the P–T path consistent with the textures and reaction
balance. Ilm in excess (NCKFMASHT system). Darker shading symbolizes fields with
higher variance; (B) modal abundance evolution diagramme of all phases as a function
of T. (*) = FeO: 23.8 wt.%, H2O: 3.7 wt.%; (**) = FeO: 20.5 wt.%, H2O: 3.9 wt.%.
component, but may, in reality, produce a relict biotite that is H-
deficient. In this scenario, biotite dehydration melting in SW Finland
did produce the melt volumes calculated by Kriegsman (2001).
We conclude that the combination of textural analysis with chemical
(compositions, reaction balance) and petrological (coherent P–T projec-
tions or pseudosections) constraints, allows alternative hypotheses to be
tested when interpreting ambiguous textures in metapelitic xenoliths.
Acknowledgements
We thank Mike Brown for the invitation to contribute to this
special issue after the 33rd IGC, and Tim Johnson for his challenging
review. We also acknowledge the constructive comments of Fawna
Korhonen and an anonymous reviewer. We very much appreciate the
useful comments, discussions and dedicated time of James Connolly
and Sumit Chakraborty on some monitoring aspects of Perple_X. This
research received support from the SYNTHESYS Project http://www.
synthesys.info/ which is financed by the European Community
Research Infrastructure Action under the FP6 “Structuring the
European Research Area” Programme. Research project: Textures in
crustal xenoliths versus textures in migmatites, at the National Museum
of Natural History/Naturalis of Leiden; analytical results from a PhD
thesis grant from the Department of Mineralogy and Petrology of the
University of Padova (within the WESTMED project by the ESF under
the EUROCORES Programme EUROMARGINS, through contract no.
ERAS-CT-2003-980409 of the European Commission, DG). A-V also
wishes to thank the JAE-Doc research program (CSIC-2007).
References
Álvarez-Valero, A.M., 2005. Petrographic and thermodynamic study of the partial
melting of restitic xenoliths from the Neogene Volcanic Province of SE Spain.
Unpublished Ph.D. Thesis, 223 pp., University of Padova, Italy.
Álvarez-Valero, A.M., Kriegsman, L.M., 2007. Crustal thinning and mafic underplating
beneath the Neogene Volcanic Province (Betic Cordillera, SE Spain): evidence from
crustal xenoliths. Terra Nova 19, 266–271.
Álvarez-Valero, A.M., Kriegsman, L.M., 2008. Partial crustal melting beneath the Betic
Cordillera (SE Spain): the case study of Mar Menor volcanic suite. Lithos 101,
379–396.
Álvarez-Valero, A.M., Waters, D.J., 2010. Partially melted xenoliths as a window into
sub-volcanic processes: evidence from the Neogene magmatic province of the Betic
Cordillera, SE Spain. Journal of Petrology. doi:10.1093/petrology/egq007.
Álvarez-Valero, A.M., Cesare, B., Kriegsman, L.M., 2005. Formation of elliptical garnets in
a metapelitic enclave by melt-assisted dissolution and reprecipitation. Journal of
Metamorphic Geology 23, 65–74.
Álvarez-Valero, A.M., Cesare, B., Kriegsman, L.M., 2007. Formation of melt-bearing spinel–
cordierite–feldspars coronas after garnet in metapelitic xenoliths. Reaction modelling
and geodynamic implications. Journal of Metamorphic Geology 25, 305–320.
Balanyá, J.C., García-Dueñas, V., 1987. Les directions structurales dans le Domaine
d'Alboran de part et d'outre du Détroit de Gibraltar. Comptes Rendus — Academie
des Sciences. Paris 304, 929–933.
Benito, R., López-Ruiz, J., Cebriá, J.M., Hertogen, J., Doblas, M., Oyarzun, R., Demaiffe, D.,
1999. Sr and O isotope constraints on source and crustal contamination in the high-
K calc-alkaline and shoshonitic Neogene volcanic rocks of SE Spain. Lithos 46,
773–802.
Carmichael, D.M., 1969. On the mechanism of prograde metamorphic reactions in
quartz-bearing pelitic rocks. Contributions to Mineralogy and Petrology 20,
244–267.
Carrington, D.P., Watt, G.R., 1995. A geochemical and experimental study of the role of
K-feldspar during water-undersaturated melting of metapelites. Chemical Geology
122, 59–76.
Cenki, B., Kriegsman, L.M., Braun, I., 2002. Melt-producing and melt-consuming
reactions in anatectic granulites: P–T evolution of the Achankovil cordierite
gneisses, South India. Journal of Metamorphic Geology 20, 543–561.
Cesare, B., 2000. Incongruent melting of biotite to spinel in a quartz-free restite at El
Joyazo (SE Spain): textures and reaction characterization. Contributions to
Mineralogy and Petrology 139, 273–284.
Cesare, B., Gómez-Pugnaire, M.T., 2001. Crustal melting in the Alboran Domain:
constraints from enclaves of the Neogene Volcanic Province. Physics and Chemistry
of the Earth 26, 255–260.
Cesare, B., Salvioli Mariani, E., Venturelli, G., 1997. Crustal anatexis and melt extraction
during deformation in the restitic enclaves at El Joyazo (SE Spain). Mineralogical
Magazine 67, 15–27.
Cesare, B., Cruciani, G., Russo, U., 2003a. Hydrogen deficiency in Ti-rich biotite from
anatectic metapelites (El Joyazo — SE Spain): crystal–chemical aspects and
implications for high-temperature petrogenesis. American Mineralogist 88,
583–595.
Cesare, B., Gómez-Pugnaire, M.T., Rubatto, D., 2003b. Residence time of S-type anatectic
magmas beneath the Neogene Volcanic Province of SE Spain: a zircon and monazite
SHRIMP study. Contributions to Mineralogy and Petrology 146, 28–43.
Cesare, B., Rubatto, D., Gomez-Pugnaire, M.T., 2009. Do extrusion ages reflect magma
generation processes at depth? An example from the Neogene Volcanic Province of
SE Spain. Contributions to Mineralogy and Petrology 157, 267–279.
Connolly, J.A.D., 2005. Computation of phase equilibria by linear programming: a tool
for geodynamic modeling and its application to subduction zone decarbonation.
Earth and Planetary Science Letters 236, 524–541.
Corbett, G.J., Phillips, G.N., 1981. Regional retrograde metamorphism of a high grade
terrain: the Willyama Complex, More Broken Hill, Australia. Lithos 14, 59–73.
Doblas, M., Oyarzun, R., 1989. Neogene extensional collapse in the Western
Mediterranean (Betic–Rif alpine orogenic belt): implications for the genesis of
the Gibraltar arc and magmatic activity. Geology 17, 430–433.
Doblas, M., López-Ruiz, J., Cebriá, J.M., 2007. Cenozoic evolution of the Alboran Domain:
a review of the tectonomagmatic models. In: Beccaluva, L., Bianchini, G., Wilson, M.
(Eds.), Cenozoic Volcanism in the Mediterranean Area: Geological Society of
America Special Paper, vol. 418, pp. 303–320.
 A.M. Álvarez-Valero, L.M. Kriegsman / Lithos 116 (2010) 300–309
Duggen, S., Hoernle, K., Van Den Bogaard, P., Garbe-Schönberg, D., 2005. Post-collisional
transition from subduction- to intraplate-type magmatism in the westernmost
Mediterranean: evidence for continental-edge delamination of subcontinental
lithosphere. Journal of Petrology 46, 1155–1201.
Fernández-Soler, J.M., 1996. El volcanismo calco-alcalino en el parque natural de Cabo
de Gata-Níjar (Almería). Estudio Volcanológico y Petrológico. Sociedad Almeriense
de Historia Natural. Monografías del Medio Natural 2, 1–295.
Fuhrman, M.L., Lindsley, D.H., 1988. Ternary-feldspar modeling and thermometry.
American Mineralogist 73, 201–215.
Hand, M., Scrimgeour, I., Powell, R., Stüwe, K., Wilson, C.J.L., 1994. Metapelitic granulites
from Jetty Peninsula, east Antarctica: formation during a single event or by
polymetamorphism? Journal of Metamorphic Geology 12, 557–573.
Hensen, B.J., Zhou, B., Thost, D., 1995. Are reaction textures reliable guides to
metamorphic histories? Timing constraints from garnet Sm–Nd chronology for
‘decompression’ textures in granulites from Søstrene Island, Prydz Bay, Antarctica.
Geological Journal 30, 261–271.
Holland, T.J.B., Powell, R., 1998. An internally consistent thermodynamic data set for
phases of petrological interest. Journal of Metamorphic Geology 16, 309–343.
Holness, M.B., Sawyer, E.W., 2008. On the pseudomorphing of melt-filled pores during
the crystallization of migmatites. Journal of Petrology 49, 1343–1363.
Johannes, W., Holtz, F., 1996. Petrogenesis and Experimental Petrology of Granitic
Rocks. Springer, Berlin. 335 pp.
Kretz, R., 1983. Symbols for rock-forming minerals. American Mineralogist 68,
277–279.
Kriegsman, L.M., 2001. Partial melting, partial melt extraction and partial back reaction
in anatectic migmatites. Lithos 56, 75–96.
Kriegsman, L.M., Álvarez-Valero, A.M., 2010. Melt-producing versus melt-consuming
reactions in pelitic enclaves and migmatites. Lithos 116, 310–320 (this issue).
Kriegsman, L.M., Hensen, B.J., 1998. Back reaction between restite and melt:
implications for geothermobarometry and pressure–temperature paths. Geology
26, 1111–1114.
López-Ruiz, J., Rodríguez-Badiola, E., 1980. La región volcánica Neógena del sureste de
España. Estudios de Geología 36, 5–63.
López-Ruiz, J., Cebriá, J.M., Doblas, M., 2002. Cenozoic volcanism. I: the Iberian
Peninsula. In: Gibbons, W., Moreno, T. (Eds.), The Geology of Spain. Geological
Society of London, London, pp. 417–438.
Platt, J.P., Whitehouse, M.J., Kelley, S.P., Carter, A., Hollick, L., 2003. Simultaneous
extensional exhumation across the Alborán Basin: implications for the causes of
late orogenic extension. Geology 31, 251–254.
309
Sawyer, E.W., 2001. Melt segregation in the continental crust: distribution and
movement of melt in anatectic rocks. Journal of Metamorphic Geology 19, 291–309.
Spear, F.S., 1993. Metamorphic phase equilibria and pressure, temperature–time paths:
Mineralogical Society of America, Monographs. Serie, vol. 1. Washington, DC, USA,
799 pp.
Spry, A., 1969. Metamorphic Textures. Pergamon, Oxford. 350 pp.
Stüwe, K., 1997. Effective bulk composition changes due to cooling: a model predicting
complexities in retrograde reaction textures. Contributions to Mineralogy and
Petrology 129, 43–52.
Tajcmanová, L., Connolly, J.A.D., Cesare, B., 2009. A thermodynamic model for titanium
and ferric iron solution in biotite. Journal of Metamorphic Geology 27, 153–165.
Thompson, J.B., 1982. Composition space: an algebraic and geometric approach. In:
Ferry J.M. (Ed.), Characterization of Metamorphism through Mineral Equilibria.
Mineralogical Society of America, Reviews in Mineralogy 10, 1–32.
Torres-Roldán, R.L., García-Casco, A., García-Sánchez, P.A., 2000. CSpace: an integrated
workplace for the graphical and algebraic analysis of phase assemblages on 32-bit
Wintel platforms. Computers and Geosciences 26, 779–793.
Vance, D., Müller, W., Villa, I.M. (Eds.), 2003. Geochronology: linking the isotope record
with petrology and textures: Geological Society Special Publication, vol. 220. 266 pp.
Vernon, R.H., 1996. Problems with inferring P–T–t paths in low-P granulite facies rocks.
Journal of Metamorphic Geology 14, 143–153.
Vernon, R.H., 2004. A Practical Guide to Rock Microstructure. Cambridge University
Press, Cambridge, p. 594.
Vernon, R.H., 2007. Problems in identifying restite in s-type granites of Southeastern
Australia, with speculations on sources of magma and enclaves. Canadian Mineralogist
45, 147–178.
White, R.W., Powell, R., Holland, T.J.B., 2007. Progress relating to calculation of partial
melting equilibria for metapelites. Journal of Metamorphic Geology 25, 511–527.
White, R.W., Powell, R., Baldwin, J.A., 2008. Calculated phase equilibria involving
chemical potentials to investigate the textural evolution of metamorphic rocks.
Journal of Metamorphic Geology 26, 181–198.
Zeck, H.P., 1970. An erupted migmatite from Cerro de Hoyazo, SE Spain. Contributions
to Mineralogy and Petrology 26, 225–246.
Zeck, H.P., Williams, I.S., 2002. Inherited and magmatic zircon from Neogene Hoyazo
cordierite dacite, SE Spain—anatectic source rock provenance and magmatic
evolution. Journal of Petrology 43, 1089–1104.