ORIGINAL PAPER

Phase equilibria modeling of kyanite-bearing eclogitic metapelitesin the NCKFMASHTO system from the Egere terrane (CentralHoggar, South Algeria)

Amar Arab & Khadidja Ouzegane & Amar Drareni &Sidali Doukkari & Souad Zetoutou & Jean-Robert Kienast

Received: 9 December 2013 /Accepted: 31 March 2014# Saudi Society for Geosciences 2014

Abstract Kyanite-bearing metapelites from the Egere terrane(Central Hoggar, South Algeria) have a high-pressureeclogitic facies paragenesis; furthermore, they preserve evi-dence of partial melting with the development of metamorphictextures involving garnet–kyanite–biotite–white mica–plagio-clase–K-feldspar–quartz–rutile and ilmenite. Garnetporphyroblasts in these rocks display a calcium zonationformed during high-Pmetamorphism followed by decompres-sion. Moreover, the large garnet poikiloblasts reflect complexmineral–melt relationships, and their growth was stronglylinked to melt production. In order to give an account of thevarious textures and explain the evolution of the differentparageneses in relation with pressure and temperature, ther-modynamic modeling in the Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O–TiO2–Fe2O3 (NCKFMASHTO) systemis a powerful approach. The intersection of isopleths of gros-sular content and Fe⁄(Fe+Mg) of core of garnet providesconstraints on the P–T conditions of the metamorphic peak.Kyanite-bearing metapelites experienced a clockwise P–Tpath culminating at about 850 °C and 17–19 kbar. Peakconditions were followed by decompression to about 9 kbarand 750 °C.

Keywords Egere . Hoggar . High-Pmetapelites .

Pseudosection . THERMOCALC

Introduction

The Egere terrane (north of Hoggar) exposes some of thehighest grade metamorphic rocks in the Pan-African orogen.These rocks preserve a variety of textures related to partialmelting. The studied metapelites are characterized by an olderparagenesis of garnet–kyanite–phengite–quartz and rutile thatcharacterize the metamorphic peak in eclogite facies. Theretrograde P–T path is evidenced by the growth of a secondaryparagenesis involving garnet–biotite–plagioclase–sillimanite–K-feldspar–ilmenite associated to melt crystallization. Thesefeatures examined the framework of P–T pseudosections,provide constrains on the P–T conditions, and elucidate thetrajectory followed by the studied samples which can place anessential constraint on the tectonothermal evolution of theEgere terrane.

Detailed observations of the field relationship and texturalfeatures are essential for providing a framework to understandthe P–T evolution and the exhumation history of migmatizedhigh grade metapelites. Careful petrographic studies provideimportant information on the consumption or appearance ofcertain index minerals at different stages during the metamor-phic evolution. Observations and analysis of leucosomes andmelanosomes commonly show that the leucosomes representanhydrous early crystallized products formed during cooling,and being composed primarily of quartz and feldspars, wherethe melanosomes include mainly kyanite, biotite, and silli-manite; in addition, the greater part of H2O in the rock occursinside the melanosome and is incorporated in hydrous phasessuch as micas (e.g., Ashworth 1976;Waters andWhales 1984;Waters 1988; Kriegsman and Hensen 1998; Kriegsman 2001;Brown 2002).

Pseudosections approach can be used to constrain P–Tconditions of the thermal peak by comparing the calculatedisopleths of mineral compositions with the measured compo-sitions (Powell et al. 2005). For instance, the grossular

A. Arab (*) :K. Ouzegane :A. Drareni : S. Doukkari : S. ZetoutouLGGIP, FSTGAT, USTHB, BP 32 El Alia, Dar el Beida,16111 Algiers, Algeriae-mail: amarab@usthb.dz

J.<R. KienastIPGP, 1 rue Jussieu, 75238 Paris Cedex 05, France

Arab J GeosciDOI 10.1007/s12517-014-1413-z

contents z (g) [=Ca⁄(Ca+Fe+Mg)] and x (g) [=Fe⁄(Fe+Mg)].For the reason that garnet composition may have been mod-ified to some extent by diffusional Fe–Mg re-equilibrationduring cooling, P–T conditions derived from compositionisopleths should be strictly considered as minimum estimates(Štıpská and Powell 2005; Hollis et al. 2006; Indares et al.2008). Nevertheless, they are likely to represent the closestapproximation to the thermal peak and the grossular zoning isstill likely to be useful for deriving the P–T path.

The aim of this contribution is to examine the microstruc-tures and the mineral chemical variations and compare themwith thermodynamic modeling of the rocks exposed withinthe Egere terrane, in order to understand their pressure–tem-perature evolution.

Geological setting

The Touareg shield, composed of Hoggar, Aïr, and Iforasregions in Central Sahara covers approximately550,000 km2 (Fig. 1a). It was assembled during the Pan-African orogeny, at the end of the Neoproterozoic, as a resultof the collision between the Tuareg terranes and the WestAfrican Craton (Black et al. 1994).

Furthermore, it is considered as a result of the amalgam-ation, during the Pan-African orogenesis, of terranes withstrongly contrasting lithologies and tectono-metamorphic his-tories (Black et al. 1994; Liégeois et al. 1994, 2003). Thesedifferent terranes are separated from each other by subvertical

strike-slip shear zones of several hundred kilometers, or bymajor thrust faults.

The Egere area represents to the northern part of Egere–Aleksod’s terrane (Central Hoggar, South Algeria, Black et al.1994; Liégeois et al. 2003, Fig. 1a, b). This terrane togetherwith the Laouni, Azrou N’Fad, and Tefedest terranes formsthe LATEAmetacraton (Liégeois et al. 2003; Black et al. 1994,Fig. 1a, b).

The first detailed description of the lithologies of the Egerearea was given by Lelubre (1952) then by Duplan (1972).They described the structure of this region as anticlines ofgneisses separated by vast synclines of metasedimentaryrocks, so they distinguished two main formations, theArechchoum and the Egere series. The former consists ofmigmatitic gneisses and less common amphibolites; the latterconsists of amphibolites facies aluminous gneisses, marbles,quartzites, and metabasic rocks. The eclogite facies was notrecognized by these authors, but garnet amphibolites withkelyphitic textures around garnet were described. In this area,a detailed geological map was made by Duplan (1972). Thestructural and lithological setting as well as metamorphicevolution of the Egere region have been recently modifiedafter field work. Therefore, the Arechchoum series is not onlya lithological complex but also a tectonic unit. The Egere areacan thus be subdivided into two tectonic units. The infracrustalunit (Arechchoum) is dominated by granodioritic and garnet-bearing amphibolite lenses. The supracrustal unit in the stud-ied Tighsi area contains eclogites with metasedimentary andmeta-igneous parentage. The metasedimentary serie consists

Lower Palaeozoic

basic and ultra-basic massif

Thrust with ophiolite

Thrust

juvenile Terranes

LATEA terrane (mainly Palaeoproterozoic

Other terranes

(580-530 Ma interval)

(Neoproterozoic)

(Neoproterozoic)

(Neoproterozoic)

Tz24°

22°

AsIs

20°

7°6°5°

Te

Az

Se

Algéria

Niger

It

Isk

Eg-Al

La

Ts

N

100 km

Central Hoggar

Tighsi Metapelite

0˚

20˚

5˚E 10˚E

Aïr

In Ouzzal

Iforas

0 500Km

Archean Paleoproterozoic Neoproterozoic

IskIt

West

Afric

an Cr

aton

Ah

Is

As Ba

AoEd

Dj

Western Hoggar Central Hoggar

Tz

LATEA

24˚N

TeEg-Al

SeAz

La

Egere

baEastern Hoggar

Fig. 1 aGeological sketch of the Touareg shield with terrane boundaries(Black et al. 1994). b Geological map of the LATEA metacraton(Liégeois et al. 2003) showing location of the studied metapelite in Tighsiarea. Ah Ahnet; Aïr is the full name; Ao Aouzegueur, As Assodé, Az

Azrou-n-Fad, Ba Barghot, DjDjanet, Ed Edembo, Eg-Al Egéré-Aleksod,Is Issalane, Isk Iskel, It In Teidini, La Laouni, Se Serouenout, Te Tefedest,Tz Tazat

Arab J Geosci

of interbedded kyanite-bearing metapelites, quartzites, andmarbles (Fig. 2a) Meter-to-decameter-long eclogite lensesare observed within the migmatitic metapelites, and they formboudins stretched in the subvertical regional foliation (Fig. 2a,Doukkari et al. 2014). The first paleoproterozoic age of thebanded gneiss from Arechchoum basement in the Aleksodarea, south of the Egere region (2200 My total rock Rb/Srdate) is reported by Bertrand and Caby (1978). The augengneisses have been dated at 1940 m.y (Bertrand 1974), andthis age is in good agreement with the paleoproterozoic age ofgranite emplacement defined elsewhere in Africa (Bonhomme1962), but isolated whole rock model ages suggest an olderage (∼2500 m.y) for the Arechchoum basement.

The studied metapelitic rocks crop out along Tighsi area(North of Egere terrane, Figs. 1b and 2a). At the outcrop scale,the Egere metapelites typically consist by alternating of mil-limeter to centimeter thick leucocratic quartzo-feldspathic do-mains with dark residual layers formed mainly of biotite,kyanite, sillimanite, garnet, and ilmenite (Fig. 2c). Foliationis defined by the preferred orientation of biotite and flatteningof feldspar and quartz grains. The foliation is generallysubvertical, and it has a stretching lineation of sillimanite thatplunges about 10° N.

Petrography

The studied sample (Eg139) is a migmatitic metapelite, char-acterized by coarse-grained textures and heterogeneous distri-bution of minerals. It is essentially composed of varyingproportions of garnet, kyanite, biotite, white mica, sillimanite,

K-feldspar, and quartz with lesser amounts of plagioclase, inaddition, ilmenite and rutile. The leucocratic layers (2 to12 mm-wide) typically contain quartz, plagioclase, and garnetporphyroblasts, alternating with dark residual layers (4 to8 mm-wide) formed mainly of garnet, biotite, plagioclase, k-feldspar, sillimanite rutile, and ilmenite.

Peak or near-peak assemblages have been inferred on thebasis of coarse grain size and the nature of the peak paragen-esis. Post peak assemblages are deduced from fine-grainedmineral in melanosome.

Quartz forms elongated ribbons (q1), several millimeterslong (2 to 9 mm-wide), which are composed of large grainswith highly lobate rims or commonly corroded (Fig. 3a), orcompletely recrystallized polygonal aggregates (q2,Fig. 3a, b).

Garnet is observed in two textural positions. Largeeuhedral grains (g1, 2–7 mm in size), frequently containinginclusions, are interpreted as a part of the P-peak assemblage.Fine-grained garnet in quartz-dominated micro domains isinterpreted as secondary (g2, Fig. 3b, c).

The garnet porphyroblasts contain inclusions of rounded tolobate quartz, kyanite, and biotite (Figs. 3d, f and 4a, b), withrutile, apatite, and zircon. Moreover, these quartz inclusions ingarnet are surrounded by a thin plagioclase pool (pl2) which isconsidered to be an indication of a former melt (Platten 1983;Vernon and Collins 1988; Sawyer 1999; Guernina and Sawyer2003; Cesare 2008; Indares et al. 2008; Ferrero et al. 2012)and were likely preserved by the shielding effect of garnetduring rock evolution. This particular texture suggests thatgarnet grew during melting reaction. Some rims of primarygarnet appear corroded by biotite-plagioclase symplectite or

EclogiteEclogiteEclogiteEclogite

Metasedimentary seriesMetasedimentary series

Metapelite + eclogite)

Metapelite + eclogite)

(Quartzite +(Quartzite +

b

MarbleMarble

MarbleMarble +

a

c

EW

Fig. 2 a Outcrop photograph from Tighsi (north of Egere terrane) in thework area (photograph by Khadidja Ouzegane), showing metapelitesinterbedded with marbles and quartzites. Decameter-long eclogite lenses

are also observed. b, c Close-up view of the studied sample (Eg139). Ethe East, W the West

Arab J Geosci

q1q1

q2q2

a0,3 mm0,3 mm

j

bibi

bibi

phnphn

0,3 mm0,3 mm

bi2bi2

q1q1 q1q1q1q1

bi2bi2

g0,5 mm0,5 mm

g2g2

bi2

bi2

bi2

bi2

bi2

bi2

bi2bi2

bi2bi2

pl2pl2

pl2pl2

q1q1 q1q1

q1q1

0,3 mm0,3 mm c

q1q1q1q1q1q1

g1g1

g1g1

g1g1

bi1bi1

pl2pl2

f0,1 mm0,1 mm

g1g1

g1g1g1g1

d

g1g1

g1g1

kyky

kyky

kykyg1g1

0,1 mm0,1 mm

bibi

bi2 + sillbi2 + sill

0,3 mm0,3 mm h

phnphnkyky

bibi0,3 mm0,3 mm i

q1q1

q1q1

ru+

ilmru

+ilm

e0,1 mm0,1 mm

pl1pl1

g1rimg1rim

g1rimg1rimg1rimg1rim

pl1pl1

k

rurururu

ilmilmilmilm

0,1 mm0,1 mm

q2q2

g2g2

g2g2

0,3 mm0,3 mm

g2+bi2+pl2g2+bi2+pl2

b

q2q2 q2q2

chlchl

q2q2

g2+bi2+pl2g2+bi2+pl2

q2q2

kspksp

q1q1q1q1

q2q2

Arab J Geosci

biotite-plagioclase and K-feldspar assemblage which mightrepresent the evidence of cooling reactions between garnetand melt during the final melt crystallization (Fig. 3c).

Fractures in garnet are typically filled with one or more ofthe following minerals: biotite, muscovite, and oxides(Fig. 3d).

Biotite occurs as (1) inclusions in garnet (bi1) (Figs. 3f and4a), (2) as 0.5–4-mm-long oriented laths (bi2) randomly radi-ating, and frequently rimmed by ilmenite (Figs. 3g and 4c, d).Otherwise, it occurs as clusters of fine-grained (bi2) associatedto sillimanite needles (Figs. 3h and 4d). In some places, biotiteclusters associated to plagioclase (pl2) surround or overgrowgarnet. Finally biotite occurs also (3) as irregular filling alonggarnet fractures (Fig. 3d). In the leucosomes, the small amountof biotite typically occurs in isolated restitic masses or asrandomly oriented or radiating crystals.

Two generations of white mica have been recognized. Theolder ones are usually coarse-grained than the later ones.Primary white mica occurs as big euhedral laths (0.5–3 mm)long (Figs. 3i, j and 4e) with resorbed outlines partially cor-roded by fine-grained biotite–k-feldspar aggregates.Frequently, these white micas have tardy biotite exsolutionsthat grow along their cleavages (Figs. 3j and 4e). This texturesuggests that this white mica is phengite but certainly itbecomes unstable during the metamorphic evolution.Secondary fine-grained muscovite overgrows biotite (bi2).Moreover, rare grains of muscovite replace kyanite.

K-feldspar is present in the mesocratic domains associatedwith biotite, sillimanite, and ilmenite (Fig. 4c, d, f). It alsosurrounds quartz (Fig. 4f). Furthermore, plagioclase is foundin small laths both in the leucocratic and mesocratic domains(Fig. 4c) and frequently associated to biotite around garnet.Finer-grained plagioclase crystals are included in garnet rim(pl1) and correspond to the first plagioclase developed inequilibrium with garnet rim along the retrograde path(Fig. 3e).

Kyanite is found as small laths (Fig. 3i) destabilized tosillimanite at their rims. It occurs either as a relict phase, asinclusions in garnet (Fig. 3d), or as bladed crystals oriented

with bi2. Sometimes, kyanite is corroded and replaced byplagioclase-biotite aggregates. Kyanite is regarded as remainsof the high-pressure metamorphic assemblage that becamemetastable during transition to a low-pressure overprint.Sillimanite is ubiquitous, occurring as fine-grained crystalsassociated to biotite defining the main foliation (Figs. 3h and4d).

Rutile (sometimes rimmed by ilmenite) as well as ilmeniteare common in melanosome (Fig. 3k).

The compositional layering of this studied sample impliesthat the protoliths may have been heterogeneous, or evencomposite; Feldspar-rich domains are interpreted as sites ofmelt crystallization. The corroded rims of quartz ribbonsagainst these domains imply that the deformation responsiblefor the development of ribbons predated partial melting. Thetextural arrangement of composite inclusions in garnet, withrelict phases in plagioclase pool, is considered to be thecharacteristic of the former melt (Platten 1983; Vernon andCollins 1988; Sawyer 1999; Guernina and Sawyer 2003;Cesare 2008; Indares et al. 2008; Ferrero et al. 2012) andwas likely preserved by the shielding effect of garnet duringrock.

Replacement of garnet by biotite+plagioclase is consistentwith retrograde reaction during melt crystallization (Waters2001). Furthermore, the interstitial character of biotitethroughout the matrix, away from garnet replacement tex-tures, suggests a retrograde origin for most biotite.

Mineral chemistry

Mineral chemical analyses were undertaken using aCAMECA SX100 electron microprobe at the University ofParis VI with operating conditions of 15 kV and 10 nA. Therepresentative mineral analyses are shown in Tables 1, 2, 3,and 4.

Two distinct stages of garnet growth are apparent. The firstinvolved subhedral garnet porphyroblasts which reveal anumber of important textural and striking chemical zonationfeatures. This primary garnet is almandine-rich (XAlm=0.67–0.76) with significant amounts of grossular (XGrs=0.13–0.21)and pyrope (XPrp=0.07–0.11), with a minor spessartine (Sps=0.01–0.02) (Table 1, Fig. 5a). This garnet is characterized by agradual decrease of XCa (grossular) [=Ca/(Ca+Mg+Fe)] to-ward the rims relative to cores (in relation with pressure/temperature evolution), (core, XCa=0.20–0.21; rim, XCa=0.13–0.17). XFe [=Fe/(Fe+Mg)] shows only a minor increasein rim (core, XFe=0.86–0.87; rim, XFe=0.89–0.91). This pat-tern is typical of diffusional homogenization at high-T, follow-ed by further Fe–Mg exchange between garnet rims andbiotite during cooling (Pattison and Begin 1994; Spear1991; Indares et al. 2008). The X (g) [=Fe/(Fe+Mg)] of garnetcore is considered to represent a minimum value for the X (g)

�Fig. 3 Photomicrographs of mineral textures (mineral abbreviations: bi,biotite; g, garnet; ky, kyanite; ksp, K-feldspar; mu, muscovite; phn,phengite; pl, plagioclase; q, quartz; chl, chlorite; ru, rutile; ilm,ilmenite). a Elongated ribbons of quartz completely enveloped by arecrystallized polygonal quartz aggregates. b Second generation of fine-grained garnet intergrown with quartz, biotite, and plagioclase in aleucosome. c Large poikiloblast of primary garnet well preserved andsurrounded by thin films of biotite–plagioclase–quartz and K-feldspar. dInclusions of primary kyanite armored in garnet and late biotite alongcracks of garnet. e, f Close-up view of primary garnet shows biotite andplagioclase inclusions in garnet. g Secondary biotite in long-oriented lathsthat underline a part of the rock foliation. h Fine-grained biotite clustersassociated to sillimanite needles. i Kyanite laths associated to phengite. jBiotite exsolutions growing along phengite cleavages. k Rutilesurrounded by ilmenite

Arab J Geosci

acquired at the thermal peak, because, even in the case of largegarnet porphyroblasts, it may have been modified to some

extent by diffusion during cooling (Pattison and Begin 1994;Spear 1991).

bi2bi2 bi2bi2

ilmilm

c

pl2pl2

q1q1

b

a

pl2pl2

pl2pl2 q1q1

bi1bi1

q1q1

g1g1

q1q1

q1q1

q1q1q1q1

d

q1q1

e

bi2bi2

bi2bi2

sillsill

sillsill

phnphn

bi2bi2

bi2bi2

q1q1

kspksp

kspksp

bi2bi2

bi2bi2

ilmilm

g1g1

g1g1

g1g1

f

kspksp

plpl

plpl

kspksp

plpl

bi2bi2

bi2bi2

ilmilm

ilmilm

q1q1

q1q1

q1q1

q1q1q1q1

kspksp

kspksp

kspksp

kspksp

g1g1

bi2bi2

bi2bi2

500µm500µm

100µm100µm

100µm100µm200µm200µm

50µm50µm

50µm50µm

Fig. 4 Back-scattered electron (BSE) images showing a various texturalrelationship. a Quartz, plagioclase, and biotite in inclusions and aroundgarnet. b Close-up view of (Fig. 4a) which evidentiates the plagioclasecorona (pl2) around quartz. c Biotite laths associated to ilmenite, k-feldspar, and plagioclase. d Small elongated sillimanite, associated tobiotite and K-feldspar in melanosome. Note ilmenite and K-feldspar

coronas around quartz. e Phengite destabilization to biotite (in core andborders). f General representative texture of mesocratic domain showingfoliation defined by biotite and sillimanite needles. Plagioclase occurs assmall grains in small patches with biotite, sillimanite, and ilmenite. K-feldspar surrounds also quartz (drawn in a yellow color)

Arab J Geosci

The second stage involved the formation of secondarygarnet which have either a similar composition to rims ofprimary garnets (XCa=0.13) or shows a lower grossular con-tent (XCa=0.08).

Biotite shows XFe values that range between 0.67 and 0.72,along with Ti values which vary between 0.34 and 0.43 c.p.f.u(cations per formula unit). In addition, there is no systematicvariation in AlVI (0.67–0.75 c.p.f.u) and Al IV (2.67–2.78 c.p.f.u, Table 2).

Primary phengite with biotite exsolution exhibits a narrowcompositional variation in major elements with XFe (0.47–0.69) and XNa [=Na⁄(Na+K)]=(0.05–0.07); furthermore, Al#[=AlVI/(AlVI+Fe+Mg)]= (0.91–0.95) (Table 3). Thisphengite is very poor in celadonite component (Si 3.01–3.13) and appears reequilibrated to muscovite. However,

TiO2 content reaches 1.62 wt%, and texture of biotite exsolu-tion in white mica is indicator of presence at high pressure ofTi-muscovite (Nahodilová et al. 2014).

Plagioclase in this metapelitic rocks ranges from oligoclaseto andesine a XAn [=Ca/(Ca+Na+K)]=(0.18–0.35) (Table 4,Fig. 5b).

Phase equilibria modeling of the anatectic metapelite

In order to recognize the metamorphic evolution of the Egeremetapelites and understand the mineral assemblages develop-ment related to P–T evolution of the studied rocks, a quanti-tative mineral equilibria modeling was undertaken usingTHERMOCALC 3.33 (Powell and Holland 1988, updated

Table 1 Representative mineral analyses of garnet, including list of abbreviations and notations used in the text

Mineral Garnet

Sample Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139g1 (core) g1 (core) g1 (core) g1 (rim) g1 (rim) g1 (rim) g1 (rim) g2 g2 g2 g2

SiO2 37.69 37.76 37.47 37.10 36.84 37.24 36.81 36.33 36.06 35.95 37.09

TiO2 0.02 0.04 0.03 0.05 0.04 0.02 0.06 0.06 0.07 0.01 0.27

Al2O3 21.36 21.12 21.21 21.04 20.95 21.01 20.97 20.71 20.69 20.80 21.17

Cr2O3 0.00 0.03 0.02 0.00 0.00 0.00 0.00 0.03 0.00 0.01 0.02

FeOt 31.59 31.87 30.50 33.93 33.83 34.19 34.39 35.23 35.20 35.50 36.42

MnO 0.43 0.52 0.45 0.71 0.67 0.76 0.73 0.17 0.15 0.21 1.01

MgO 2.79 2.71 2.77 1.87 2.01 2.41 2.29 2.24 2.16 2.12 2.66

CaO 7.15 7.15 7.24 5.70 5.85 4.48 4.45 4.38 4.34 4.40 2.89

Na2O 0.02 0.01 0.00 0.00 0.01 0.00 0.02 0.00 0.01 0.00 0.13

K2O 0.00 0.00 0.00 0.01 0.02 0.03 0.00 0.01 0.00 0.00 0.00

Total 101.05 101.20 99.69 100.43 100.21 100.13 99.71 99.16 98.68 99.00 101.77

Oxygen 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00

Si 2.97 2.97 2.99 2.97 2.96 2.99 2.97 2.95 2.94 2.93 2.94

Al,IV 0.03 0.03 0.01 0.03 0.04 0.01 0.03 0.05 0.06 0.07 0.06

Al,VI 1.95 1.94 1.98 1.96 1.94 1.98 1.96 1.93 1.94 1.92 1.92

Ti 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02

Fe 2.03 2.04 2.02 2.24 2.21 2.27 2.29 2.33 2.34 2.34 2.35

Mn 0.03 0.03 0.03 0.05 0.05 0.05 0.05 0.01 0.01 0.01 0.07

Mg 0.33 0.32 0.33 0.22 0.24 0.29 0.28 0.27 0.26 0.26 0.31

Ca 0.60 0.60 0.62 0.49 0.50 0.38 0.38 0.38 0.38 0.38 0.25

Na 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02

K 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Cr 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Total 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00

XFe 0.86 0.87 0.86 0.91 0.90 0.89 0.89 0.90 0.90 0.90 0.88

Alm 0.68 0.68 0.67 0.75 0.74 0.76 0.76 0.78 0.78 0.78 0.79

Sps 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.00 0.00 0.00 0.02

Grs 0.20 0.20 0.21 0.16 0.17 0.13 0.13 0.13 0.13 0.13 0.08

Prp 0.11 0.11 0.11 0.07 0.08 0.10 0.09 0.09 0.09 0.09 0.11

Alm=Fe⁄(Ca+Fe+Mg+Mn); Prp=Mg⁄(Ca+Fe+Mg+Mn); Grs=Ca⁄(Ca+Fe+Mg+Mn); Sps=Mn⁄(Ca+Fe+Mg+Mn); XFe=Fe⁄(Fe+Mg)

Arab J Geosci

June 2009) and the internally consistent data set of Hollandand Powell (1998); data set (file tc-ds55.txt), created 22November 2003). Calculations were constructed in the chem-ical system Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O–TiO2–Fe2O3 (NCKFMASHTO), which is a near-comprehensive compositional analog of natural metapeliticrock systems. Manganese was not considered in calculations,due to its very small content (Mahar et al. 1997; Sajeev et al.2006; White et al. 2007). The activity–composition (a–x)models and mineral abbreviations used are as follows: garnet(g: White et al. 2007), biotite (bi: White et al. 2007), musco-vite (mu: Coggon and Holland 2002), plagioclase and K-feldspar (pl, ksp: Holland and Powell 2003), and melt (liq:White et al. 2007). Quartz (q) and aluminosilicate (ky/sill) arepure end-member phases.

A pseudosection was constructed using the measured bulk-rock composition obtained by XRF method (CRPG Nancy).Therefore, the pseudosection allows the exploration of thephase equilibria of the metamorphic history and the assess-ment of the near-peak and retrograde evolution.

Figure 6 shows a P–T pseudosection constructed for themodel aluminous metapelite composition, for a P–T range 2–25 kbar and 700–950 °C. In this P–T range, quartz andilmenite were considered in excess. Furthermore, the bulk-rock composition of this metapelite is characterized by a smallamount of H2O calculated by using the loss of ignition (LOI)method; H2O is fixed in the hydrous minerals (i.e., biotite andmuscovite). The value of Fe2O3 was likely to be low becauseno Fe3+-rich oxides were observed. The P–T pseudosectioncalculated using the measured bulk-rock composition wasused tomodel the whole metamorphic evolution of the studiedsample, that is, from peak P–T conditions to complete meltcrystallization (White et al. 2007; Indares et al. 2008; Whiteand Powell 2010, 2011).

This pseudosection (Fig. 6) is characterized by four mainfields; (1) field I contains the peak metamorphic assemblageformed by garnet–kyanite–white mica–k-feldspar–melt– andrutile which represents the eclogite facies. This field is bor-dered by the white mica breakdown limit and occurs in thehigh-pressure part of the pseudosection and (2) field II and III

Table 2 Representative mineral analyses of biotite

Mineral Biotite

Sample Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139

SiO2 33.44 33.60 33.03 32.34 33.07 33.32 33.30

TiO2 3.40 3.52 3.27 2.80 3.22 3.60 3.56

Al2O3 18.22 17.89 17.83 18.48 18.11 18.71 19.11

Cr2O3 0.10 0.06 0.00 0.06 0.35 0.10 0.07

FeO 23.35 23.81 24.17 23.78 24.40 22.57 22.81

MnO 0.07 0.10 0.04 0.10 0.04 0.07 0.09

MgO 5.45 5.73 5.40 5.79 5.66 6.13 6.18

CaO 0.03 0.04 0.06 0.03 0.04 0.00 0.00

Na2O 0.11 0.14 0.16 0.18 0.12 0.09 0.18

K2O 9.13 9.22 8.98 9.12 8.98 9.33 9.47

Total 93.29 94.10 92.94 92.66 93.97 93.92 94.77

Oxygen 22.00 22.00 22.00 22.00 22.00 22.00 22.00

Si 5.33 5.33 5.31 5.22 5.27 5.26 5.22

Al,IV 2.67 2.67 2.69 2.78 2.73 2.74 2.78

Al,VI 0.75 0.67 0.69 0.74 0.67 0.74 0.74

Ti 0.41 0.42 0.40 0.34 0.39 0.43 0.42

Fe 3.11 3.16 3.25 3.21 3.25 2.98 2.99

Mn 0.01 0.01 0.01 0.01 0.00 0.01 0.01

Mg 1.29 1.35 1.29 1.39 1.34 1.44 1.44

Ca 0.01 0.01 0.01 0.01 0.01 0.00 0.00

Na 0.03 0.04 0.05 0.05 0.04 0.03 0.06

K 1.86 1.86 1.84 1.88 1.82 1.88 1.89

Cr 0.01 0.01 0.00 0.01 0.04 0.01 0.01

Total 15.47 15.52 15.54 15.63 15.51 15.50 15.55

XFe 0.71 0.70 0.72 0.70 0.71 0.67 0.67

XFe=Fe⁄(Fe+Mg)

Table 3 Representative mineral analyses of white mica

Mineral White mica

Sample Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139

SiO2 43.82 44.91 44.69 44.27 47.16 46.96 46.81

TiO2 0.44 0.45 0.92 0.46 1.47 1.62 1.46

Al2O3 35.11 36.08 34.65 36.05 33.60 33.93 33.24

Cr2O3 0.00 0.00 0.00 0.00 0.00 0.01 0.06

FeO 1.29 1.24 1.44 1.13 1.47 1.49 1.44

MnO 0.04 0.00 0.04 0.01 0.06 0.00 0.00

MgO 0.37 0.31 0.43 0.35 0.93 0.86 0.90

CaO 0.00 0.00 0.04 0.05 0.00 0.00 0.01

Na2O 0.52 0.33 0.36 0.48 0.45 0.41 0.51

K2O 9.86 8.08 10.51 10.31 11.23 11.04 10.95

Total 91.46 91.40 93.08 93.10 96.37 96.32 95.38

Oxygen 11.00 11.00 11.00 11.00 11.00 11.00 11.00

Si 3.03 3.06 3.05 3.01 3.12 3.11 3.13

Al,IV 0.97 0.94 0.95 0.99 0.88 0.89 0.87

Al,VI 1.90 1.96 1.84 1.91 1.74 1.75 1.74

Ti 0.02 0.02 0.05 0.02 0.07 0.08 0.07

Fe 0.07 0.07 0.08 0.06 0.08 0.08 0.08

Mn 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Mg 0.04 0.03 0.04 0.04 0.09 0.09 0.09

Ca 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Na 0.07 0.04 0.05 0.06 0.06 0.05 0.07

K 0.87 0.70 0.92 0.90 0.95 0.93 0.93

Cr 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Total 6.98 6.84 6.99 6.99 7.00 6.98 6.99

XFe 0.66 0.69 0.65 0.64 0.47 0.49 0.47

XFe=Fe⁄(Fe+Mg)

Arab J Geosci

define the high-T boundary of the pseudosection. For field II,the paragenesis is characterized by the garnet–k-feldspar–pla-gioclase–melt–kyanite–rutile assemblage; here, plagioclase(pl1), a feature of this metapelite, is stable. Field III showsan assemblage formed by the garnet–k-feldspar–plagioclase–melt–silimanite; it is bordered by the rutile to ilmenite

transition (Fig. 6); furthermore, along the retrograde path—within this field—the kyanite/sillimanite polymorphic transi-tion is crossed. (3) In the low P–T part of this pseudosection,field IV is characterized by the garnet–k-feldspar–plagio-clase–biotite–white mica–melt–sillimanite assemblage. It isbordered by the solidus which marks the disappearance or

Table 4 Representative mineral analyses of feldspar

Mineral Pl ksp

Sample Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139 Eg139

SiO2 63.65 60.11 60.01 60.56 60.72 60.51 58.09 62.20 61.95 65.48 65.04

TiO2 0.00 0.01 0.02 0.04 0.03 0.01 0.03 0.01 0.00 0.00 0.00

Al2O3 22.96 25.38 25.56 23.76 23.30 23.20 25.21 24.28 24.69 18.48 18.29

Cr2O3 0.00 0.00 0.00 0.00 0.03 0.00 0.01 0.02 0.00 0.00 0.00

FeO 0.12 0.17 0.20 0.21 0.19 0.11 0.32 0.14 0.12 0.04 0.13

MnO 0.03 0.00 0.00 0.00 0.00 0.02 0.00 0.01 0.07 0.02 0.03

MgO 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01

CaO 3.87 6.86 6.73 5.52 5.06 5.14 7.51 5.41 5.37 0.01 0.01

Na2O 9.98 8.23 7.91 8.82 9.17 8.83 7.70 8.78 8.33 0.42 0.38

K2O 0.14 0.16 0.15 0.19 0.23 0.21 0.14 0.17 0.14 15.54 15.39

Total 100.75 100.91 100.56 99.10 98.73 98.03 99.01 101.00 100.66 99.97 99.27

Oxygen 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00

Si 2.80 2.66 2.66 2.72 2.74 2.74 2.63 2.74 2.73 3.01 3.01

Al 1.19 1.32 1.34 1.26 1.24 1.24 1.35 1.26 1.28 1.00 1.00

Ti 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Fe 0.00 0.01 0.01 0.01 0.01 0.00 0.01 0.01 0.00 0.00 0.00

Mn 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Mg 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Ca 0.18 0.33 0.32 0.27 0.24 0.25 0.36 0.25 0.25 0.00 0.00

Na 0.85 0.71 0.68 0.77 0.80 0.78 0.68 0.75 0.71 0.04 0.03

K 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.91 0.91

Cr 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Total 5.04 5.03 5.01 5.04 5.05 5.03 5.04 5.01 4.99 4.96 4.96

An 0.18 0.32 0.32 0.26 0.23 0.24 0.35 0.25 0.26 0.01 0. 01

Ab 0.82 0.68 0.67 0.74 0.76 0.75 0.64 0.74 0.73 0. 04 0. 04

Or 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.96 0.96

An=Ca⁄(Ca+Na+K); Ab=Na⁄(Ca+Na+K); Or=K⁄(Ca+Na+K)

gt1 core

gt2

gt1 rim

Gross Pyrope

Alm+Spess

Gross Pyrope

Alm+Spes s

Sanidine

Anorthoclas

e

Albite Oligoc

laseAndesine

Labrado

rite

Bytown

iteAnorthite

Ab An

Or

a b

60

70

80

90

40

30

20

10

60 70 80 90

pl1

pl2

Ksp

Fig. 5 a Garnet compositionsplotted in the (almandine+spessartine)–grossular–pyropediagram. b Plagioclasecompositions plotted in thealbite–anorthite–orthoclasediagram

Arab J Geosci

the total crystallization of the melt and by biotite appearance,which is associated to sillimanite.

Approximately, melt occurs in all the pseudosection fields;however, the low-T part of pseudosection represents thesubsolidus portion which may not be applicable to the pro-grade evolution of the system.

In our sample, the observed thin section textures are in-ferred to have developed in fields I, II, III, and IV (Fig. 6).Textural evidence for the presence of melt in these assem-blages is best preserved in the feldspar pools included ingarnet (Fig. 4a, b).

P–Tevolution and textural evidences (decompression P–Tpath)

Some new studies on metapelites have shown how factorssuch as grain size distribution and progressive chemical iso-lation of porphyroblast cores can influence the mineralogical

development during metamorphic evolution through modifi-cation of the local effective bulk composition (EBC,e.g.,Marmo et al. 2002; Evans 2004; Verdecchia et al. 2013).The studied metapelite is characterized by coarse-grainedminerals that formed at or near metamorphic peak. To varyingdegrees, they contain inclusions of relic minerals within them,which are also rimmed by overgrowths⁄coronas of finer-grained minerals. First-order constraints on the P–T range ofthe thermal peak are provided by the location of the stabilityfield of peak assemblage: garnet–kyanite–white mica–K-feld-spar–melt–quartz and rutile on the pseudosection. This assem-blage reflects the maximum degree of dehydration and byinference the highest grade of metamorphism.

Further information on the P–T evolution within melt-present fields is constrained by comparing the observed tex-tures and the measured mineral compositions with those cal-culated (isopleths). The upper pressure limit that characterizesthe P–T evolution of the studied sample is provided by theintersection of garnet isopleths reflecting the measured

700 750 800 850 900 950

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

g wm ksp pl ky ru

g wm ksp pl ky

g bi ksp pl sill

g ks

p liq

ky ru

g ksp pl liq sill

19

21

11

13

15

17

11

13

3

5

7

15

17

19

21

86

86

87

9

3

5

7

v4v4

v5v5

v5v5

v4v4

v4v4

v5v5

v5v5

v4v4

v5v5

NCKFMASHTO + q ilm

ky

sill

a

700 750 800 850 900 950

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

g ks

p liq

ky ru

g ksp pl liq sill

v5v5

v4v4

v5v5

v5v5

v5v5

NCKFMASHTO + q ilm

63

6567

69

7173

75

6369717375

22 24 26 28 30 32 34 36 38

38 40 42 44 46 48 50

16 18 20 22 24 26 28

3234363840 42 44

18 20

6765

2826

g bi ksp pl sill

ky

sill

g wm ksp pl ky

v4v4

9

b

peak paragenesis

v5v5

P(K

bar)

T(C°)

P(K

bar)

T(C°)

g ksp pl liq ky ru

g wm ksp ky ru

g wm ksp pl ky ru

g wm ksp ky ru

I

II

III

8688

9086

8890

8

8

9

868890

v4v4

740 750 760 7707.5

8.5

9.5

1

2

5

6

7

8

9

8

7

86

88

90

6569717375

9

6367

8868890

v4v4

IV

retrograde paragenesis

36

Fig. 6 P–T pseudosections constructed with the measured bulk compo-sition of the investigated metapelite, show the general topologies. a Thepeak paragenesis is outlined with white ellipsoid and provided by theintersection of z(g) [=Ca⁄(Ca+Fe+Mg)] with those of [x(g-core)=Fe⁄(Fe+Mg)] in garnet which is in equilibrium with kyanite, white mica,melt, k-feldspar, and rutile. b The retrograde paragenesis is illustrated inpurple ellipsoid and formed by a second generation of garnet, biotite,sillimanite, k-feldspar, plagioclase, and ilmenite. It is evidentiated via theintersection of z(g), x(g), and x(bi) close to the solidus. Green line

indicates the solidus, red dotted line indicates x(g)%, yellow dotted lineindicates z(g) %, light brown dotted line indicates x(bi) %, blue dotted lineindicates ca (pl) %. The dark red dotted arrow in (Fig. 6b) shows the P–Tpath. g, garnet; wm, white micas; bi, biotite; ky, kyanite; sill, sillimante;pl, plagioclase; ksp, K-feldspar; liq, melt; ru, rutile; in addition to quartz(q) and ilmenite (ilm) which are in excess; v, variance; (I): g–ky–wm–ksp–liq–ru–q–ilm field; (II): g–pl–ksp–liq–ky–ru–q–ilm field; (III): g–ksp–pl–liq–sill–q–ilm (IV): g–ksp–pl–bi–wm–liq–sill–q–ilm fields

Arab J Geosci

grossular content in large garnet porphyroblasts; z (g)[=Ca⁄(Ca+Fe+Mg)=0.21] with those of [x (g-core)=Fe⁄(Fe+Mg)] on pseudosection (Fig. 6a), the intersection ofisopleths defines P–T conditions that provide the thermal andpressure peak in the model system. The compositional param-eters x (g) and z (g) in garnet are proved to be robust toconstrain the metamorphic conditions of this eclogiticmetapelite. This is because the diffusion of Ca in garnet isextremely sluggish (Chakraborty and Ganguly 1991). On thecontrary, relict white mica appears to reequilibrate easily asindicated by exsolutions textures and is therefore a poormetamorphic indicator of this high-pressure stage.

Retrograde reactions are promoted bymelt crystallizing; onthe other hand, the potential for substantial development ofretrograde textures is proportional to the amount of melt thatremained present at the thermal peak. The fact that garnet andkyanite are enveloped and/or partially replaced by biotite+plagioclase implies retrograde reaction during melt crystalli-zation (e.g., Waters 2001; Cenki et al. 2002). The end of thisretrograde trajectory is typically testified by the developmentof biotite, a second generation of garnet and the appearance ofsillimanite at the expense of kyanite; furthermore, it is con-firmed by the intersection of z (g) [=0.08], x (g) [=0.88], and x(bi) [=0.69] close to the solidus at 9 kbar and 750 °C in themodel (Fig. 6b).

The slope of z (g) garnet isopleths suggests that depletion inz (g) values in the rim of garnet could be achieved by apressure decrease. The z (g) garnet in this pseudosection issimilar to the observed garnet zoning in the studied sample [z(g) core=0.20–0.21, rim 0.13]. Plagioclase is not stable atpeak pressure conditions, but becomes stable upon decom-pression, with a composition An18–An35, which is consistentwith the observed mineral compositions. This metapelite has aclockwise metamorphic P–T path involving decompressionprior to cooling. In the pseudosection, the range in isoplethsvalues matches the measured secondary garnet x (g)=0.88, thegrossular content of garnet z (g)=0.08, and the x (Fe) in thesecondary biotite x (Fe)=0.69 at P=8.9 Kbar and T=752 °C.Thus, the H2O-under saturated pseudosections explains ade-quately the textural evolution of the garnet metapelite and isreasonably consistent with measured mineral composition.

General P–Tevolution

Based on modeling and textural observations, the Egereeclogitic metapelites followed a clockwise P–T path to ametamorphic peak, around 18 Kbar and 850 °C followed bya decompression retrograde path to around 8 Kbar and 750 °Cassociated to melt crystallization.

Considering the petrographic texture observations, andthermodynamic modeling, the retrograde path followed bythe studied sample is underlined by two main observations:

a → The development of the peak assemblage formed es-sentially by kyanite, phengitic white mica, melt, rutile, andanatectic garnet through a prograde trajectory is evidencedby the various inclusions in garnets.

The metamorphic peak is constrained by the kyanite-bearing assemblage; in combination with garnet isoplethsz(g) and x(g), the peak metamorphic P–T conditions areestimated to have been around 18 Kb and 850 °C. Themodeled peak assemblage is associated to the productionof a significant amount of melt and its partial extractionduring the P–T evolution of the rock (White et al. 2001).

b → The development of a retrograde paragenesis wheregarnet, kyanite, and melt are required to be reactant inorder to produce new garnet, plagioclase and K-feldspar,and biotite and sillimanite. Along this path, rutile istransformed to ilmenite. Furthermore, the retrogradepath which models the exhumation from high- to mid-crustal depth is testified by (1) partial replacementtextures, for instance, plagioclase-biotite which re-places garnet, and white mica destabilization to biotite,as well as kyanite transformation to sillimanite. Theretrograde path is associated to melt crystallization atlower pressure (Fig. 6b).

Discussion and conclusions

Rocks investigated from the Egere region show progressivechanges in mineral assemblages, mineral chemistry, and meta-morphic textures consistent with a significant change in meta-morphic grade. The studied garnet–kyanite–biotite–white mi-ca–plagioclase–K-feldspar–rutile–ilmenite and quartzeclogitic metapelites from Egere exhibit microstructural evi-dences for widespread peritectic growth of garnet. As a con-sequence, melt and peritectic reactions were focused aroundthe growing garnet.

Films or pools that enclose garnet inclusions are formedduring peritectic phase growth and are representative of theformer melt (Cesare et al. 2009, 2011; Vernon 2011). Theseinclusions are sheltered by the host phase and are likely tosurvive the retrograde metamorphic events (Holness et al.2011).

The use of quantitative phase diagrams has greatly in-creased our ability to interpret the complex reaction texturesand hence use such textures to constrain the P–T evolution ofmetamorphic anatectic rocks.

Pseudosections constructed in the NCKFMASHTO systemusing the measured bulk-rock composition provide criticalinformation on the stable mineral assemblages and the chang-es that have happened.

Microstructural relationships, mineral chemical data, andthermodynamic modeling have allowed us to constrain the P–

Arab J Geosci

Tevolution of a representative sample from Tighsi area (Egereterrane). The resulting P–T path shows a clockwise shape,characterized by a retrograde decompression trajectory fromthe metamorphic peak down to pressure at which meltcompletely crystallized. The overall clockwise shape of theP–T path and the decompression-dominated P–T path reflectthe progressive exhumation of the Egere rocks.

Peak metamorphic conditions in the range of 18 Kbar and850 °C are estimated from the coarse-grained assemblagepreserved in this studied sample, garnet–kyanite–phengite–melt–plagioclase–k-feldspar–rutile and quartz, which are per-fectly in agreement with those derived from the model data. Inaddition, the inferred maximum pressure temperature condi-tions are calculated using isopleths of mineral compositions.

The retrograde evolution of the aluminous Egeremetapelites involves substantial development of garnet, bio-tite, and K-feldspar during melt crystallization, in addition tosillimanite which replaces kyanite. The preservation of high-pressure mineral paragenesis requires that a substantial pro-portion of melt produced during the prograde and peak meta-morphism to be lost from the source rock (e.g., Powell andDownes 1990; Sawyer 1991; Brown 1994; White et al. 2002;Brown 2007).

In metapelites, late metamorphic events often obliteratemost of the prograde assemblage. The preservation of high-pressure assemblages during exhumation has been regarded asa significant problem. Thermal models predict that rocksburied to 60–70 km by underthrusting should undergo heatingduring exhumation and pass through the granulite or amphib-olites facies (Thompson and England 1984). Rapid exhuma-tion could favor the preservation of high-pressure assem-blages, provided they remain preserved by influx of H20 alongshear zones.

Acknowledgments We gratefully acknowledge the OPNA (Office duParc National de l’Ahaggar), ORGM (Office National de RechercheGéologique et Minière), and COMENA (Commissariat à l’EnergieAtomique) for logistic support during fieldwork. This paper is a contri-bution to CNEPRU Project entitled “Modélisation thermodynamique etimplication géodynamique des zones de suture de haute pression duHoggar”. We are deeply grateful to anonymous reviewers for their de-tailed reviews and constructive criticism which improved both qualityand clarity of the manuscript.

References

Ashworth JR (1976) Petrogenesis of migmatites in the Huntly–Portsoyarea, northeast Scotland. Mineral Mag 40:661–682

Bertrand JML (1974) Evolution polycylique des gneiss précambriens del’Aleksod (Hoggar centrale, Sahara Algérien). Aspects structuraux,pétrologiques, géochimiques et géochronologiques. Thèse EtatUniversité de Montpellier, Ed. C.N.R.S. (C.R.Z.A), sér. Geol. no.19, p 350

Bertrand JM, Caby R (1978) Geodynamic evolution of the Pan-Africainorogenic belt: a new interpretation of the Hoggar shield (AlgerianSahara). Geol Rundschau 67:357–383

Black R, Liégeois JP, Latouche L, Caby R, Bertrand JM (1994) Pan-African displaced terranes in the Touareg shield (Central Sahara). JGeol 22:641–644

Bonhomme M (1962) Contribution à l’étude geochronologique de laplate forme de l’Ouest africain. Thesis Clermont Ferrand, 69p

Brown M (1994) The generation, segregation, ascent and em-placement of granite magma: the migmatite-to-crustally de-rived granite connection in thickened orogens. Earth-Sci Rev36:83–130

Brown M (2002) Retrograde processes in migmatites and granulitesrevisited. J Metamorph Geol 20:24–40

Brown M (2007) Crustal melting and melt extraction, ascent and em-placement in orogens: mechanism and consequences. J Geol SocLon 164:709–730

Cenki B, Kriegsman LM, Braun I (2002) Melt-producing and melt-consuming reactions in the Achankovil cordierite gneisses, SouthIndia. J Metamorph Geol 20:543–561

Cesare B (2008) Crustal melting: working with enclaves. In: Workingwith Migmatites (eds Sawyer E.W. and Brown M), pp 37–55.Mineralogical Association of Canada

Cesare B, Ferrero S, Salvioli-Mariani E, Pedron D, Cavallo A (2009)Nanogranite and glassy inclusions: the anatectic melt in migmatitesand granulites. Geology 37:627–630

Cesare B, Acosta-Vigil A, Ferrero S, Bartoli O (2011) Melt inclusions inmigmatites and granulites in the Science of Microstructure—Part II,Journal of the Virtual Explorer (Forster M.A and Fitz Gerald, J.D.),Electronic Edition, ISSN 1441-8142, 38, paper 2

Chakraborty S, Ganguly G (1991) Compositional zoning and cationdiffusion in aluminosilicate garnets. In: Ganguly J (ed) Diffusion,ordering and mass transport in physical geochemistry, vol 8.Springer, Heidelberg, pp 120–175

Coggon R, Holland TJB (2002)Mixing properties of phengitic micas andrevised garnet-phengite thermobarometers. J Metamorph Geol 20:683–696

Doukkari SA, Ouzegane K, Arab A, Kienast J-R, Godard G, Drareni A,Zetoutou S, Liégeois J-P (2014) Phase relationships and P–T path inNCFMASHTO system of the eclogite from the Tighsi area (Egereterrane, Central Hoggar, Algeria). J Afr Earth Sci. doi:10.1016/j.jafrearsci.2014.02.016

Duplan L (1972) La chaine de l’Egéré (Hoggar septentrional). Bulletin duService géologique de l’Algerie Tome I and II. Tome I, 200 p. TomeII, p 157

Evans TP (2004) A method for calculating effective bulk compositionmodification due to crystal fractionation in garnet-bearing schist:implications for isopleth thermobarometry. J Metamorph Geol 22:547–557

Ferrero S, Bartoli O, Cesare B, Salvioli Mariani E, Cavallo A,Groppo C, Battiston S (2012) Microstructures of melt inclu-sions in anatectic metasedimentary rocks. J Metamorph Geol30:303–322

Guernina S, Sawyer EW (2003) Large-scale melt depletion in granuliteterranes: an example from the Archean Ashuanipi Subprovince ofQuebec. J Metamorph Geol 21:181–201

Holland TJB, Powell R (1998) An internally-consistent thermodynamicdataset for phases of petrological interest. J Metamorph Geol 16:309–344

Holland TJB, Powell R (2003) Activity–composition relations for phasesin petrological calculations: an asymmetric multicomponent formu-lation. Contrib Mineral Petrol 145:492–501

Hollis JA, Harley SL, White RW, Clarke GL (2006) Preservation ofevidence for prograde metamorphism in ultrahigh-temperature,high-pressure kyanite-bearing granulites, South Harris, Scotland. JMetamorph Geol 24:263–279

Arab J Geosci

Holness MB, Cesare B, Sawyer EW (2011) Melted rocks under themicroscope: microstructures and their interpretation. Elements 7:245–250

Indares A, White RW, Powell R (2008) Phase equilibria modelling ofkyanite-bearing anatectic paragneiss from the central GrenvilleProvince. J Metamorph Geol 26:815–836

Kriegsman LM (2001) Partial melting, partial melt extraction and partialback reaction in anatectic migmatites. Lithos 56:75–96

Kriegsman LM, Hensen BJ (1998) Back reaction between restite andmelt: implications for geothermobarometry and pressure-temperature paths. Geology 26:1111–1114

Lelubre M (1952) Recherche sur la geologie de l’Ahaggar central etoccidental (Sahara central). Bulletin Service géologique Algerie22, tome 1, 354 p, tome 2, p 387

Liégeois JP, Black R, Navez J, Latouche L (1994) Early and late Pan-African orogenies in the Aïr assembly of terranes (Tuareg shield,Niger). Precambrian Res 67:59–88

Liégeois JP, Latouche L, Boughrara M, Navez J, Guiraud M (2003) TheLATEA metacraton (Central Hoggar, Touareg Shield, Algeria):behaviour of an old passive margin during the Pan-African orogeny.J Afr Earth Sci 37:161–190

Mahar EM, Baker JM, Powell R, Holland TJB, Howell N (1997) Theeffect ofMn onmineral stability in metapelites. J Metamorphic Geol15:223–238

Marmo BA, Clarke GL, Powell R (2002) Fractionation of bulk compo-sition due to porphyroblast growth: effects on eclogite facies mineralequilibria, Pam Peninsula, New Caledonia. J Metamorph Geol 20:151–165

Nahodilová R, Štípská P, Powell R, Košler J, Racek M (2014) High-Timuscovite as a prograde relict in high pressure granulites withmetamorphic Devonian zircon ages (Běstvina granulite body,Bohemian Massif): consequences for the relamination model ofsubducted crust. Gonwana Res 25:630–648

Pattison DRM, Begin NJ (1994) Zoning patterns in orthopyroxene andgarnet in granulites: implications for geothermometry. J MetamorphGeol 12:387–410

Platten IM (1983) Partial melting of semipelite and the development ofmarginal breccias around a late Caledonian minor intrusion in theGrempian highlands of Scotland. Geol Mag 120:37–49

Powell R andDownes J (1990) Garnet porphyroblast-bearing leucosomesin metapelites: mechanisms, phase diagrams, and an example fromBroken Hill, Australia. In: High-temperature Metamorphism andCrustal Anatexis (eds Ashworth J. R. & Brown M.), pp 105–123.The Mineralogical Society Series 2, Unwin Hyman

Powell R, Holland TJB (1988) An internally consistent thermodynamicdataset with uncertainties and correlations: 3. Application, methods,worked examples and a computer program. J Metamorph Geol 6:173–204

Powell R, Guiraud M, White RW (2005) Truth and beauty in metamor-phic phase-equilibria: conjugate variables and phase diagrams. CanMineral 43:21–33

Sajeev K, Santosh M, Kim HS (2006) Partial melting and P–T evolutionof the Kodaikanal Metapelite Belt, southern India. Lithos 92:465–483

Sawyer EW (1991) Disequilibrium melting and the rate of melt-residuumseparation during migmatisation of mafic rocks from the GrenvilleFront, Quebec. J Petrol 32:701–738

Sawyer EW (1999) Criteria for the recognition of partial melting. PhysChem Earth 24:269–279

Spear F (1991) On the interpretation of peak metamorphic temperaturesin light of garnet diffusion during cooling. J Metamorph Geol 9:379–388

Štıpská P, Powell R (2005) Constraining the P–T path of a MORB-typeeclogite using pseudosections, garnet zoning and garnet-clinopyroxene thermometry: an example from the BohemianMassif. J Metamorph Geol 23:725–743

Thompson AB, England PC (1984) Pressure-temperature-time paths ofregional metamorphism. II. Their inference and interpretation usingmineral assemblages in metamorphic rocks. J Petrol 25:929–955

Verdecchia SO, Reche JE, Baldo G, Segovia-diaz E, Martinez FJ (2013)Staurolite porphyroblast controls on local bulk compositional andmicrostructural changes during decompression of a St–Bt–Grt–Crd–And schist (Ancasti metamorphic complex, Sierras Pampeanas, WArgentina). J Metamorphic Geol 31:131–146

Vernon RH (2011)Microstructures of melt-bearing regional metamorphicrocks. In: Origin and Evolution of Precambrian High-Grade GneissTerranes, with Special Emphasis on the Limpopo Complex ofSouthern Africa (eds Van Reenen DD, Kramers JD, McCourt S &Perchuk L.L), pp 1–11. Geological Society of AmericaMemoir, 207

Vernon RH, Collins WJ (1988) Igneous microstructures in migmatites.Geology 16:1126–1129

Waters DJ (1988) Partial melting and formation of granulite faciesassemblages in Namaqualand, South Africa. J Metamorph Geol 6:387–404

Waters DJ (2001) The significance of prograde and retrograde quartz-bearing intergrowth microstructures in partially melted granulitefacies rocks. Lithos 56:97–110

Waters DJ, Whales CJ (1984) Dehydration melting and the granulitetransition in metapelites from southern Namaqualand, S. Africa.Contrib Mineral Petrol 88:269–275

White RW, Powell R (2010) Retrograde melt–residue interaction and theformation of near-anhydrous leucosomes in migmatites. JMetamorph Geol 28:579–597

White RW, Powell R (2011) On the interpretation of retrograde reactiontextures in granulite facies rocks. J Metamorph Geol 29:131–149

White RW, Powell R, Holland JB (2001) Calculation of partial meltingequilibria in the system Na2O–CaO–K2O– FeO–MgO–Al2O3–SiO2–H2O (NCKFMASH). J Metamorph Geol 19:139–153

White RW, Powell R, Clarke GL (2002) The interpretation of reactiontextures in Fe-rich metapelitic granulites of the Musgrave Block,central Australia: constraints from mineral equilibria calculations inthe system K2O–FeO–MgO– Al2O3–SiO2–H2O–TiO2–Fe2O3. JMetamorph Geol 20:621–632

White RW, Powell R, Holland TJB (2007) Progress relating to calculationof partial melting equilibria for metapelites. J Metamorph Geol 25:511–527

Arab J Geosci