Journal of Petrology | Volume 40 | Number 8 | Pages 1241-1269 | 1999
© Oxford University Press 1999
The Petrogenesis of Some Migmatites and Granites (Central Damara Orogen, Namibia): Evidence for Disequilibrium Melting, Wall-Rock Contamination and Crystal Fractionation
1 Institut Für Mineralogie, Kristallographie Und Petrologie, Philipps-Universität Marburg 35032 Marburg, Germany
2 Max-Planck-Institut Für Chemie, Abt. Geochemie Postfach 3060, 55020 Mainz, Germany
3 Mineralogisch–Petrologisches Institut Der Universität Bonn, Poppelsdorfer Schloss 53115 Bonn, Germany
Received June 26, 1997; Revised typescript accepted February 22, 1999
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
The Oetmoed Granite–Migmatite Complex (OGMC), central Damara orogen, Namibia, consists of restite-rich, grt- and crd-bearing S-type granites and grt–crd–sil–Kfs-bearing metasediments, stromatic migmatites and nebulites. Both types of migmatites formed by limited in situ partial melting of metapelites under H2O-saturated conditions at
700°C and 5 kbar. Melanosomes of the stromatic migmatites do not resemble true residues, instead they more probably represent reaction zones between in situ melt and the metasedimentary host rock. Leucosomes of the stromatic migmatites have LREE- and HFSE-depleted disequilibrium compositions, typical of low-melt fractions generally observed in migmatite terranes. Similar 
18O values in the melanosomes and leucosomes suggest that partial melting occurred under fluid-present conditions. Nebulites are more residual than melanosomes and metasediments, indicating that separation of melt and residue must have occurred. Cordierite- and grt-bearing xenoliths in the granites do not represent residue from the site of origin of the intrusive granites; their depleted chemical composition is best explained by extensive degrees of partial melting of incorporated country rocks. Chemical variations among the grt- and crd-bearing granites are explained by fractional crystallization processes and xenolith entrainment. Major and trace element data and high 18O values suggest that the grt- and crd-bearing granites were derived from H2O-undersaturated melting of metapelitic rocks. KEY WORDS: granites; migmatites; Namibia; partial melting; Proterozoic mobile belt
|
Introduction |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
Some models for the origin of S-type granitic rocks have emphasized fractional crystallization as the dominant process controlling elemental variations in granitoid suites (e.g. Phillips et al., 1981
; Miller & Mittlefehldt, 1982; Mittlefehldt & Miller, 1983). This contrasts with other studies which, although still recognizing the importance of fractional crystallization, emphasized the importance of source rock composition. These studies have explained the mineralogical and geochemical characteristics in terms of partial melting of metasedimentary rocks and subsequent separation of melt and refractory residues (e.g. Flood & Shaw, 1975; White & Chappell, 1977; Clemens & Wall, 1981). However, a recent reinvestigation of the Lachlan Fold Belt granitoids indicates that a three-component mixing process (mantle, lower crust, middle crust) may also be a viable process generating a wide spectrum of granitoid rocks (Collins, 1996).
Separation of melt and residues [restite unmixing process of White & Chappell (1977)] will result in linear trends in most variation diagrams. However, the process of disequilibrium melting (e.g. Barbero et al., 1995) will produce granitic melts, which have different elemental contents and hence define different positions in most elemental variation diagrams. To find out what exactly constitutes a restite and to quantify the effects of its unmixing, the restites must be geochemically and mineralogically characterized. Furthermore, the granitic suites that contain members with a large proportion of restite must also contain unfractionated rock types and so the source must be well constrained. Such granitic suites where the presence of restite and the effects of unmixing have been described are rarely documented (White & Chappell, 1988; Barbero & Villaseca, 1992; Williamson et al., 1997; Sawyer, 1998). Some peraluminous granites from the Proterozoic Damara orogen are ideal objects to study because they contain a high proportion of restite phases (garnet, cordierite, xenoliths), their source can be fairly well constrained and they can be compared with associated peraluminous granites that contain less or no restite material.
Another approach to solving the problem of determining the sources of crustally derived granites is the study of migmatites, which might represent several stages of crustal anatexis, because of their inferred role as a link between high-grade metamorphism and the generation of larger-scale granitic bodies [e.g. Brown & D'Lemos (1991); for a different view see White & Chappell (1990)]. Despite the observation that many, but not all, regional-scale migmatite complexes are spatially and temporally associated with intrusive rocks of mainly granitic composition in orogenic belts, the relation between migmatites and plutons is unresolved: (1) migmatites may represent an arrested stage of granite development in whichleucosomes did not coalesce to form a large-scale body (e.g. Obata et al., 1994); (2) migmatites may result from contact effects induced by neighbouring plutons (e.g. Pattison & Harte, 1988); or (3) migmatites and granites are unrelated (e.g. White & Chappell, 1990). Before the relationship between migmatization and granite genesis can be constrained the processes by which migmatites are generated must be evaluated.
The central zone of the Proterozoic Damara orogen (Namibia) offers a unique opportunity to study these problems because it comprises an amphibolite-facies terrane where granites and migmatites are closely associated. In this paper we (1) discuss the contrasting geochemical signatures of migmatite leucosomes and intrusive granites, (2) evaluate the main evolutionary mechanism in the genesis of these leucosomes, (3) discuss the origin of these migmatites in the context of regional metamorphism and granite genesis and (4) evaluate the effects of fractional crystallization and source rock composition for the granites. Some of the intrusive granites are contaminated with cordierite-rich xenoliths, and it is appropriate to ask whether these xenoliths represent restite from the source region or if they simply reflect metasedimentary residues that result from incorporation and extensive partial melting. To contribute to this problem we evaluate the effects of restite unmixing. This discussion is based on petrographical observations, interpretation of geochemical data and mass balance calculations.
|
Geological Setting |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
The Oetmoed Granite–Migmatite Complex (OGMC) is located at 15°E,21.5°S within the Central Zone of the Pan-African Damara orogen, Namibia (Fig. 1). This complex comprises mainly schistose cordierite–sillimanite–quartz–K-feldspar–biotite-bearing metasediments of the Kuiseb Formation of Precambrian age (Miller, 1983
). Some metasediments are migmatized garnet–cordierite gneisses with small-scale, layer-parallelleucosomes with less garnet and cordierite. Minor marble, quartzite, amphibole schists and graphitic schists are also present. The granites form predominantly metre- to decimetre-wide conformable sheets within the metasediments but additionally, a round to slightly elongated body occurs in the centre of the complex (Fig. 2).
|
sillimanite, (7) sillimanite-in according to staurolite breakdown; (8) partial melting as a result of muscovite + plagioclase + quartz + H2O
|
Haack et al. (1982)
reported Rb/Sr whole-rock ages of 514 ± 22 Ma to 465 ± 31 Ma for leucogranites from the nearby Okombahe reserve, which might be comparable with the granites studied here. These ages indicate emplacement and cooling of the granites within the F2/F3 deformation phases as defined by Haack et al. (1980). A Rb/Sr whole-rock isochron for the garnet- and cordierite-bearing granites gives an age of 508 ± 15 Ma with an initial 87Sr/86Sr value of 0.7169. New U/Pb ages on titanite and monazite (Jung et al., 1998b) indicate emplacement of some granites around 495 ± 3 Ma. All types of granitic rocks studied here show local development of foliation and phenocryst alignment, some of them defining the magmatic foliation. However, in some places these structures are folded by a later phase of deformation. Therefore, it seems likely that the granites intruded close to the F3 deformation phase.
Peak metamorphic conditions have been estimated to be around 700°C at 4–5 kbar (Jung et al., 1995); these values are in agreement with previous estimates within the Central Zone (700–750°C and 5–6 kbar; Masberg et al., 1992; Bühn et al., 1994). These high-temperature conditions lead to the formation of garnet–cordierite-bearing migmatites and to incipient melting of metapelites (Jung et al., 1998a). Microscopic evidence (exolution of mesoperthitic string-perthites and antiperthites in feldspars, orientated exsolution of rutile needles in quartz, activation of basal 
a
—and prism c—glide systems in quartz under conditions of plastic deformation) indicate low-pressure granulite-facies conditions at least for the highest-grade coastal region of the Proterozoic Damara orogen (Masberg et al., 1992).
|
Description of the Country Rocks and Granites |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
Field relationships
The OGMC shows the uniformly high metamorphic grade, complex structures and abundant evidence of anatexis typical of many migmatite terranes. Most country rocks are Al-rich high-grade metasedimentary rocks in which metapelites have higher modal biotite contents than the metagreywackes. The latter are characterized by higher contents of quartz and feldspar and less biotite. The metasedimentary rocks show continuous banding on the scale of 1 mm to 1 cm, which is defined by alternating, well-foliated biotite-rich layers and granular quartz–feldspar layers.
The migmatites can be subdivided into stromatic migmatites with melanosomes and leucosomes and nebulites. Stromatic migmatites represent the lower-grade part of the migmatites, broadly comparable with metatexites as defined and described by Brown (1979). The pre-migmatization fabric (e.g. layering and foliation) is well developed within the palaeosome (Fig. 3a) although the neosome–paleosome ratio may be variable (Fig. 3b). Principally, there are two types of leucosome: (1) rare in situ leucosomes with mafic selvedges and (2) leucosomes without mafic selvedges that were injected from external sources (Fig. 3a). Nebulites are comparable with diatexites (sensu Brown, 1979) and comprise the higher-grade part of the migmatites. Typically, they are located close to voluminous intruding granite sheets (Fig. 3c and d). Leucosomes occur mostly as concordant veins with diffuse and irregular borders towards the mafic rock portions. There is considerable variation in the proportion of mafic material relative to leucocratic rock portion; usually the mafic material produces abundant mafic schlieren oriented parallel to the flow (Fig. 3d).
|
The granites are white to pinkish, medium- to fine-grained garnet- and cordierite-bearing granites with numerous medium- to coarse-grained cordierite- and garnet-bearing xenoliths (Fig. 3e).
Petrography
Representative mineral compositions of the differentlithologies are given in Table 1 and representative thin sections are shown in Fig. 4. Zoning profiles of garnet and cordierite are shown in Fig. 5. The most common mineral assemblage of the metapelites is biotite–cordierite – plagioclase – K-feldspar – sillimanite – quartz(Fig. 4a) with minor muscovite and accessory tourmaline, apatite, zircon, monazite and Fe–Ti oxides. Dark greenish blue grains of cordierite, sometimes with abundant inclusions of sillimanite and large porphyroblasts of K-feldspar with subparallel inclusions of biotite are typical of these metasediments.
|
|
|
The melanosomes are garnet, cordierite and biotite rich (with Qtz + Kfs + Pl + Sil), generally not very different from the surrounding metapelites. The prominent feature of these melanosomes is the occurrence of garnet that is sometimes rimmed by cordierite (Fig. 4b). The leucosomes contain plagioclase, K-feldspar, quartz, some garnet and anhedral crystals of cordierite with inclusions of biotite, sillimanite and quartz but also minor amounts of euhedral, inclusion-free cordierite crystals. Additionally, large K-feldspar porphyroblasts with sub-parallel inclusions of biotite occur within these leucosomes. Garnet from the leucosomes is chemically different from garnets of the melanosomes, it has fewer inclusions and preserves a bell-shaped compositional pattern. In the leucosomes, plagioclase dominates over alkali feldspar. Zircon is the dominant accessory phase but it is restricted to minute crystals isolated within the leucosomes.
The mineral assemblage of the nebulites is variable,but is generally biotite–cordierite–plagioclase–K-feldspar–sillimanite–quartz together with accessory zircon, monazite, apatite and rare Fe–Ti oxides (Fig. 4c).
In the garnet- and cordierite-bearing granites quartz, perthitic alkali feldspar and plagioclase make up to 90 vol. % of the rock, but proportions of alkali feldspar to plagioclase are variable. Large K-feldspar crystals may be phenocrystic. Biotite in the garnet- and cordierite-bearing granites occurs either as mineral aggregates or as individual crystals. Cordierite is either anhedral and contains abundant inclusions of sillimanite, or is euhedral and inclusion free (Fig. 4d). Garnet occurs as embayed crystals with inclusions of sillimanite and quartz sometimes rimmed by cordierite (Fig. 4e) or as euhedral to subrounded, inclusion-free crystals (Fig. 4f). Sillimanite occurs as dense fibrolitic mats or as needle-like inclusions in cordierite and garnet (Fig. 4g). Typical accessory phases within the garnet- and cordierite-bearing granites include zircon, apatite, monazite and sometimes (?)allanite.
The xenoliths contain mostly completely pinitized, inclusion-poor cordierite with embayed crystal faces (Fig. 4h) whereas others have inclusions of quartz, biotite and sillimanite, which makes these cordierites indistinguishable from those in the metasediments and migmatites.
|
Geochemistry of the High-Grade Rocks and Granites |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
Analytical methods
Major and trace elements [except for rare earth elements (REE)] were determined on fused lithium tetraborate glass beads using standard X-ray fluorescence (XRF) techniques. Additionally, trace element contents were verified by flameless atomic absorption spectrometry using the graphite furnace technique (Cr, Ni, Co, Pb) or by flame atomic absorption spectrometry (Rb). Abundances of Ba, Sr, Zr, Nb and Y were verified by inductively coupled plasma emission spectrometry (ICP-AES) using an HF–HClO4 decomposite for Ba and Sr or an HF–H2SO4 decomposite for Zr, Nb and Y. The precision of each technique is better than 5–10% for all trace elements, except for Nb below 25 ppm (30%), and the agreement between XRF and atomic absorption spectrometry is generally better than 10%. REE have been analysed by ICP-AES following separation of the matrix elements by ion exchange (Heinrichs & Herrmann, 1990
). LOI (loss on ignition) was determined gravimetrically at 1050°C (Lechler & Desilets, 1987) and FeO was measured titrimetrically with standard techniques. Accuracy has been controlled by repeated measurements against several international and in-house standards and the results are in good agreement with the recommended values. Oxygen isotope analyses were performed at the University of Bonn on 8–10 mg aliquots of powdered whole-rock samples, using purified fluorine for oxygen extraction, followed by conversion to CO2 (Clayton & Mayeda, 1963). 18O/16O measurements were made on a SIRA-9 triple-collector mass spectrometer by VG-Isogas. Analytical uncertainties are <0.2
.
Metasediments
The most abundant metasediments are Al-, Fe-, Mg- and Ti-rich metapelites (Table 2), having a composition close to ordinary shales with K2O > Na2O and low CaO (Taylor & McLennan, 1985). Relative to the granites, the metasediments have higher contents of TiO2, Al2O3, FeOtot and MgO, but their Na2O, CaO and K2O contents are lower (Fig. 6). Because of their high modal biotite/feldspar ratio Rb, Ba, Zn and V contents are high, and Sr and Pb contents are moderate (Table 2). Co, Cu, Cr and Ni contents are moderate, and all trace element contents and the REE, with their characteristically fractionated pattern and a pronounced negative Eu anomaly (Eu/Eu*: 0.38–0.63), are well within the compositional range of various post-Archaean shale estimates (Taylor & McLennan, 1985). Relatively high contents of Zr, Nb and REE reflect significant amounts of accessory minerals. K/Rb ratios are between 100 and 160, and are significantly lower than the estimates of average post-Archaean shales quoted by Taylor & McLennan (1985). The less abundant metasediments are quartz-intermediate metagreywackes, having higher SiO2 (>65 wt % SiO2), Na2O and CaO, and lower TiO2, Al2O3, FeO, MgO and K2O than the metapelites. Because of their higher modal feldspar/biotite ratio, the elements Rb, Ba, Zn, and sometimes Co, Cu, Ni and Cr contents are lower than in the metapelites. Sc and Y are lower, but Nb and Zr contents are comparable with those in the metapelites. REE contents are variable, but broadly comparable with those of the metapelites, but the negative Eu anomaly is more pronounced (Eu/Eu* 0.28–0.35; Fig. 7a).
|
|
|
Nebulites, melanosomes and xenoliths
In comparison with the metapelites, the nebulites are enriched in TiO2, P2O5, Rb and Nb, and depleted in Na2O and Sr (Fig. 6 and Table 2). Values of Al2O3, MgO, FeO, Ni, Zn, Co, Cu, Cr, Zr and Y are close to the values of the metapelites, but Pb and Ba contents show a wide scatter in abundance. Relative to the metagreywackes the enrichment and depletion in the above-mentioned elements is more pronounced. Rare earth element contents and the fractionation between the light REE (LREE) and heavy REE (HREE) are comparable with the values obtained from the metapelites, but on average a slight enrichment in total REE content, except for Eu which shows an overall depletion, occurs. Consequently, the negative Eu anomaly is more pronounced than in the metapelites, with Eu/Eu* values between 0.12 and 0.35 (Fig. 7b).
The melanosomes of the stromatic migmatites (samples Mel 1–Mel 3) have a major and trace element geochemistry indistinguishable from that of the metapelites. Garnet-free melanosomes (Mel 2 and Mel 3) are slightly lower in FeOtot, MgO, Y, Sc and Nb. Total REE contents are slightly lower than in the metapelites and metagreywackes but the fractionation of LREE relative to HREE is more pronounced in garnet-free samples Mel 2 and Mel 3 (Fig. 7c). The Eu/Eu* values are similar to those observed in the metagreywackes (Eu/Eu* 0.28–0.41).
The leucosomes of the stromatic migmatites are enriched in SiO2, CaO and Na2O, but depleted in FeOtot, TiO2 and MgO relative to the metasedimentary rocks. In comparison with the intrusive garnet- and cordierite-bearing granites, they are enriched in Na2O and CaO but depleted in K2O at a given SiO2 content (Fig. 6). The leucosomes have higher contents of Li, Zn, Pb and Cr, and lower contents of Rb, Ba, Y, Nb and REE, with a pronounced positive Eu and moderately fractionated Lan/Ybn, Gdn/Ybn and Lan/Smn ratios (Fig. 7d). Sample 89.44 is spatially associated with the other leucosomes but is broader than the leucosomes L 1 to L 3 (decimetre instead of centimetre), it is enriched in mafic material (mainly cordierite and some biotite) and lacks mafic selvedges. Because of the close spatial association it is presented with the other leucosomes.
Cordierite- and garnet-bearing xenoliths have a more residual mineralogy and geochemistry than the melanosomes or nebulites. These xenoliths have similar contents of FeO, but most of them are enriched in TiO2, MgO and Al2O3 and are strongly depleted in Na2O, K2O, CaO (Fig. 6), Rb, Pb, Ba, HFSE and REE relative to the metasediments and migmatites. The most obvious geochemical feature is the strong depletion of the LREE and the less strong depletion of the HREE, which leads to low total REE contents and a low Lan/Ybn ratio. Additionally, Eu is strongly depleted, resulting in low Eu/Eu* values of 0.09–0.25 (Fig. 7c).
Granites
The granites have SiO2 contents of 68–76 wt % (Table 3). All samples are peraluminous [molar A/CNK (i.e. Al2O3/CaO + Na2O + K2O) > 1] with K2O > Na2O and a strong variation in Na2O and CaO content. The granites have low MgO contents (<1.0 wt % MgO), except for some samples that contain restitic xenoliths (samples 89.57 and 89.75). Because it is likely that some of the garnet- and cordierite-bearing granites contain some residual material we compare these granites with crustally derived granites that show no sign of incorporation of residual material on an outcrop scale. Figure 8 shows major element and Rb, Sr and Zr variations from crustally derived granites from the central Damara orogen (McDermott et al., 1996; S. Jung et al., unpublished data, 1999). Generally, TiO2, FeOtot, MgO, CaO, K2O and Al2O3 decrease with increasing SiO2 and Na2O (not shown) shows some scatter. However, K2O, CaO and FeOtot abundances are lower but Al2O3 is higher (at a given SiO2 content) in the granites from the OGMC relative to the other granites. Generally, Rb and Ba are enriched relative to Sr in the garnet- and cordierite-bearing granites from the OGMC and therefore Rb/Ba and Rb/Sr ratios are high but Sr/Ba ratios are low. On the basis of their REE composition (Fig. 9a–c), the granites can be subdivided into three groups.
|
|
|
Group I granites (Fig. 9a) have the lowest SiO2 but the highest Sr, Rb and Zr contents among the garnet- and cordierite-bearing granites, and are REE enriched with a pronounced negative Eu anomaly. These features suggest that this group comprises the least differentiated samples. Group II granites (Fig. 9b) are characterized by lower LREE contents but higher HREE contents and consequently lower Lan/Ybn, Lan/Smn and Gdn/Ybn ratios. Additionally, a sometimes strong negative Eu anomaly is observed. Samples 89.71 and 89.75 are contaminated with substantial amounts of migmatitic material and show high FeO and MgO contents, a high alumina saturation index (ASI: 1.34 and 1.93) and the smallest negative Eu anomaly within this group. Group III granites (Fig. 9c) have very low total REE contents and a marked positive Eu anomaly.
Stable isotopes
Generally, all granite samples show high 18O values, typical for crustal rocks. The variation of 18O within the individual samples of the garnet- and cordierite-bearing granites is between 13.6 and 15.2. The group I granites, which are considered to be relatively unfractionated, have constant 18O values of 14.3–14.5 and this value generally increases with differentiation; i.e. samples 9.1 and 13, which are the most fractionated samples, have 18O values of 15.1. Lower 18O values are observed only in some of the group III granites. The oxygen isotopic composition of metasediments and migmatites is fairly homogeneous and varies in the range 11.6–13.6 and 11.5–13.5, respectively. Rare 18O data of the restitic xenoliths indicate slightly higher 18O values between 13.2 and 14.2.
|
Discussion |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
Interpretation of textures and implications from mineral data
Field evidence indicates that some disintegration of metasedimentary material with subsequent integration into the intruding granite has occurred (Fig. 3e and d). The main evidence for this is that enclaves of metasedimentary rocks are often disintegrated and are randomly distributed within the host granite. There is little textural evidence that most of the biotite in these granites represents restitic material because biotite in the granites is usually coarse grained and has far fewer inclusions of zircon.
Whether cordierite in peraluminous melts is a primary magmatic mineral (Phillips et al., 1981; Allen & Barr, 1983; Maillet & Clarke, 1985; Georget & Fourcade, 1988) or represents a restite component (Hensen & Green, 1973; Birch & Gleadow, 1974; Flood & Shaw, 1975; Green, 1976; Clemens & Wall, 1981) remains controversial. Cordierite may form as a cotectic magmatic phase relatively late in peraluminous melts or as a result of peritectic reactions during either an increase in temperature or a decrease in pressure (Clarke, 1995). All cordierites from this study are Fe rich, and the cordierites from the nebulites have Na2O contents slightly higher as usually observed in metamorphic rocks (Miyashiro, 1957). The pinitized cordierite with embayed crystal faces present in some granites certainly represents entrained material from the metasediments (Fig. 4h). It is unclear whether the inclusion-free cordierite, now incorporated in the leucosomes, was incorporated as a portion of the melanosome, which was partly digested by the intruding melt, or grew as a product of the melt-producing incongruent breakdown of biotite.
The inclusion-poor cordierite from the nebulites (Fig. 4c) and from some granites is believed to have crystallized in the presence of a melt. The main criteria for this suggestion are: (1) a euhedral shape, (2) lack of strong zoning (Fig. 5f) and (3) scarcity or absence of inclusions. Another characteristic feature seen in cordierite-bearing granitic melts is micrographic cordierite–quartz intergrowths (e.g. Villaseca & Barbero, 1994).
On the basis of textural evidence, at least two different types of garnet can be distinguished. Garnets from the melanosomes and the leucosomes are inclusion rich, anhedral and sometimes embayed, which suggests that these garnets formed as a product of biotite-dehydration melting reactions followed by textural modification within the leucosomes. Some of the intruding granites contain garnet that is texturally similar to garnets in the leucosomes (Fig. 4e), whereas other garnets show no inclusions and have euhedral shapes (Fig. 4f), indicating crystallization of these garnets from a melt (Vernon & Collins, 1988). The occurrence of inclusion-free, euhedral garnet crystallizing in the peraluminous granites suggests that garnet crystallized at high temperature, high pressure and low aH2O (Clemens & Wall, 1988).
Generally, most garnets have flat profiles from core to rim without steep zonation profiles, which suggests that they became homogenized by volume diffusion athigh temperatures (Fig. 5). Garnets from the stromaticmigmatites have slightly different compositions inthe melanosomes, leucosomes and at the melanosome–leucosome interface. In the melanosomes, garnets have rather low Mn and Mg contents, suggesting equilibration with coexisting biotite and cordierite. At the melanosome–leucosome interface, garnet is notably enriched in Mg and Mn, which might be due to the decomposition of biotite during migmatization. In the leucosomes, garnet has rather low Mg and Mn contents, but shows a pattern in which Mn increases from core to rim and Mg decreases from core to rim. We interpret this pattern as being the result of reaction during garnet breakdown to produce cordierite, an interpretation consistent with the observed cordierite rims. Euhedral garnets from the granites usually display an increase in Mn from core to rim together with a decrease of Mg from core to rim. This zonation can be interpreted as a magmatic feature, in which Mg behaves compatibly and Mn behaves incompatibly during crystallization of the garnet and evolution of the granite. Exceptions are rare garnets with very low Mn but slightly higher Ca, which may be the result of partial melting of igneous source rocks, enriched in Ca and depleted in Mn relative to metasedimentary source rocks.
Chemical attributes of the nebulites, stromatic migmatites and xenoliths
The chemical data, i.e. the depletion of several incompatible elements (such as Sr, Na, Ca) in the nebulites and the xenoliths can be explained by processes linked with anatexis. Normalizing the nebulite and xenolith data against the inferred source rock (i.e. pelite) allows us to estimate which elements are residual and which have been concentrated into a melt.
The nebulite samples 89.64 and 89.5 and the melanosome M 1 of the stromatic migmatites are characterized by a strong depletion of Sr, CaO and Na2O, a moderate depletion of Eu, Pb and Ba, and a slight enrichment of K2O, La, Rb, Al2O3, MgO, Yb and FeOtot (Fig. 10). These geochemical features are qualitatively consistent with the partial melting of a metapelitic protolith, leaving a residue enriched in ferromagnesian(biotite, cordierite, garnet) and accessory phases (zircon, monazite, apatite). The conclusions drawn above are further supported from the REE data. Nebulites show a wider range in LREE and HREE abundances but generally lower Eu contents than the metasediments. This can be explained by the extraction of a feldspathic melt and selective entrainment of accessory phases within the melanocratic portion of the nebulites. In addition, strong variations in HREE patterns from the melanosomes of the stromatic migmatites are related to the formation of either garnet or cordierite during melting. The observation that the melanosomes from the stromatic migmatites are on average not depleted in LREE and HREE but show rather low Eu contents suggests the presence of an LREE- and HREE-poor partial melt with selective enrichment of Eu. Some of these features (moderate REE contents, positive Eu anomaly) are monitored by the leucosomes and are characteristics of leucosomes usually found in high-grade terranes (Barbey et al., 1989, 1990; Sawyer, 1991; Watt & Harley, 1993; Carrington & Watt, 1995; Watt et al., 1996). These leucosomes are interpreted to originate through disequilibrium melting processes of the host metasedimentary rock. It is equally possible that they represent melt plus some accumulated feldspar (e.g. Sawyer & Barnes, 1988) but petrographic evidence argues against this possibility in the case of the Damaranleucosomes. The Damaran leucosomes have higher Na2O and lower K2O contents than the intruding granites, which can be explained by higher water activities during melting, different melting conditions and/or different source rocks relative to the intrusive granites.
|
Although the melanosomes are enriched in several elements linked with residual Fe, Mg phases, e.g. biotite and garnet (TiO2, Al2O3, FeOtot, MgO, K2O, Rb, V, Cr, Sc), other elements that should also be concentrated within the residue are depleted (P2O5, Nb) or show considerable overlap with the values from the metagraywacke (Ni, Y, Zr). Thus it seems likely that these melanosomes are not strictly residues but represent a reaction zone where the intruding granite injected into the metasediments, thereby initiating local partial melting.
The xenolith samples display high Al2O3, MgO, FeOtot, TiO2 and Yb contents, but low CaO, K2O, Na2O, P2O5, Sr, Pb, Eu, Ba, Nb, Zr and REE abundances with very low Eu contents. These features indicate (1) a residual component in the xenoliths, (2) dissolution of accessory phases into the derived melt and (3) that most of the feldspar was used to form melt, which was lost to the enclosing magma.
Constraints from oxygen isotopic compositions
Previous studies have shown that the leucogranites of the central part of the Damara orogen have high 18O values up to 14.3 (Haack et al., 1982; Hoernes & Hoffer, 1985), typical for crustal melts. The granites studied here have 18O values up to 15.2. It is therefore reasonable to assume that the oxygen isotope composition has been modified to some extent by interaction with a fluid phase. Figure 11a and b shows the variation of the oxygen isotopic composition as a function of Fe3+/Fe2+ and LOI (loss on ignition) values. There is a broad negative correlation between the oxygen isotopic composition and the Fe3+/Fe2+ values and the LOI values, suggesting that interaction with a hydrothermal aqueous system has occurred, thereby lowering the 18O values whilst increasing the Fe3+/Fe2+ and LOI values. Despite the evidence for some alteration, the magnitude of lowering the 18O values is rather small (1 for the granites and 2 for the metasediments). Additionally, a very large amount of H2O is required to produce significant shifts in the 18O values, for which there is no field or petrographic evidence, i.e. no late crystallization of muscovite, no rehydration of garnet or cordierite, and no alteration of feldspar. Furthermore, if the high 18O ratios are the result of low-temperature hydrothermal alteration one would expect to find abnormal oxygen isotope fractionations between coexisting minerals, especially quartz and feldspar, because the rates of oxygen isotope exchange between mineral and aqueous fluids are usually different for different minerals. According to Hoernes & Hoffer (1985), 18O quartz–feldspar ratios of granites and granitic anatexite range between 1.0 and 1.5, typical values for unaltered granites world-wide (Shieh, 1985).
|
If the high
18O values of the granites were the result of isotopic exchange with the high-grade country rocks and migmatites one should expect to find lower values of 18O in the samples with a strong anatectic overprint, e.g. the nebulites. This is not observed because the nebulites and metasediments show similar 18O values, which are positively correlated with the alumina saturation index (Fig. 12). This positive correlation mainly results from the high 18O values of the primary Al-rich clay minerals in the original shale (e.g. Hoernes & van Reenen, 1992). Therefore, we conclude that the high 18O values of about 14 of the group I granites are primary and are not due to isotopic exchange with the high-grade country rocks.
|
In contrast to the metasediments and migmatites, the garnet- and cordierite-bearing granites show an obvious negative correlation between the ASI and
18O values (Fig. 12). This is usually not explained by fractional crystallization processes, which lead to an enrichment of both the ASI value and 18O. However, Barbarin (1996) reported decreasing alumina saturation during differentiation in some peraluminous granitic suites. It is therefore probable that the evolution from group I granitoids with ASI values of 1.3 and 18O values of 144 to the most fractionated granites from group III (samples 9.1 and 13) with ASI values of 1.06 and 18O values of 15.1 follows such trends. Some samples have higher ASI values coupled with lower 18O ratios. Here, it is more likely that the negative correlation of 18O with ASI results from incorporation of Al-rich minerals (garnet, cordierite, biotite, sillimanite), which usually have much lower 18O values than the major rock-forming minerals quartz, plagioclase and K-feldspar.
The 18O values of the nebulites are indistinguishable from those of the metasediments; this feature indicates that partial melting processes are not monitored by the oxygen isotopic composition.
|
Implications for the Evolution of the Grt- and Crd-Bearing Granites |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
Nature of the source for the granites
The geochemical characteristics (enrichment of Ba and Rb relative to Sr, high K2O, low CaO) of the garnet- and cordierite-bearing granites indicate a crustal source rock. Even the group I granites, which are believed to be the less fractionated group, display a pronounced negative Eu anomaly. This negative Eu anomaly either can be interpreted as a characteristic feature of the source rocks or indicates the presence of residual plagioclase during melting. Low Na2O (<3.5 wt %) and CaO (<1.2 wt %) but high K2O (>5.0 wt %) contents, the large ion lithophile element (LILE) contents (Rb 220–350 ppm, Sr 70–100 ppm, Ba 220–260 ppm), Rb/Ba > 0.25, and high 87Sr/86Sr (initial 87Sr/86Sr 0.715–0.735) and
18O (10.5–15.5) values indicate a pelitic source rock (Haack et al., 1982; Miller, 1985; Williamson et al., 1997). Rb/Sr ratios are high (>2.6) and Sr/Ba ratios are low (<0.4), suggesting that these granites were generated under water-undersaturated conditions through biotite-dehydration melting (Harris & Inger, 1992). McDermott et al. (1996) have presented a large amount of data on crustally derived Damaran granitoids, and the garnet- and cordierite-bearing granites studied here are similar in composition (K2O/Na2O, Rb, Sr, Ba, Zr) to the leucogranites of Stinkbank and Bloedkoppje. For these granites biotite-dehydration melting of a pelitic greywacke source has been suggested (McDermott et al., 1996).
Partial melting processes
Generally, the garnet- and cordierite-bearing granites (1) might represent different magma pulses of a deep-seated magma or (2) they might correspond to unrelated magma batches generated from different source rocks (e.g. Deniel et al., 1987). These two possibilities are not mutually exclusive, but because most of the garnet- and cordierite-bearing granites have rather constant K2O/Na2O ratios and K2O > Na2O, we suggest that differentiation of different magma pulses from similar sources is more likely than partial melting of very different source rocks, which might produce granitic rocks with very different K2O/Na2O ratios.
Estimates of the conditions of formation of the granites may be obtained by the application of saturation equations for Zr, LREE and P2O5 (Watson & Harrison, 1983; Rapp et al., 1987; Bea et al., 1992; Montel, 1993). This calculated temperature can be interpreted as the temperature of extraction of a granitic melt from its source, given that no fractional crystallization has occurred and that the magma does not contain accessory phases as xenocrysts. Such temperatures are probably minimum estimates because they can reflect crystallization temperatures, i.e. they are lower than the temperatures of source partial melting. Consistent results can be expected only if chemical equilibrium prevailed during melting, the relevant accessory minerals control the trace element budget and the whole-rock composition represents a frozen liquid (e.g. Montel, 1993). Estimates for a typical unfractionated (Group I) garnet and cordierite granite (sample 89.66) are between 800 and 850°C (Table 4). These temperature estimates are well within the range usually assumed to be realistic for anatectic granites (LeBreton & Thompson, 1988; Vielzeuf & Holloway, 1988; Vielzeuf & Montel, 1994). Furthermore, the study of Watson & Harrison (1983) has shown that a typical peraluminous melt will be saturated in Zr at a level of 150–250 ppm at 800–860°C. These temperatures and Zr contents are close to the values for the Group I granites, which are considered to be unfractionated, indicating that these granites contain no significant amounts of inherited zircon. Furthermore, most separated zircon crystals (from a detailed geochronological study; S. Jung, unpublished data, 1999) are long to short prismatic with well-developed crystal faces. These features probably indicate that the major proportion of zircon is magmatic and not xenocrystic. Early saturation of apatite is indicated by abundant euhedral apatite enclosed in phenocrystic K-feldspar. Experimental work (Pichavant et al., 1992) has shown that in peraluminous granites containing <0.5 wt % P2O5 most of the apatite is magmatic rather than restitic in origin. Temperatures in excess of 800°C for the unfractionated (group I) garnet- and cordierite-bearing granites are also monitored by their low Sm/Nd ratios (0.21), as has been calculated by Ayres & Harris (1997) using an extended apatite dissolution model.
|
The average MgO (0.47–0.55 wt %) and FeOtot (1.50–1.77 wt %) contents of the group I granites suggest that they might represent melts with no entrained Fe–Mg-rich crystals (assuming melting at the temperature conditions given above) because at T = 800–850°C, the solubilities of MgO and FeO in peraluminous granitic melts are 0.22–0.90 wt % and 1.27–3.10 wt %, respectively (Clemens & Wall, 1981
; Puziewicz & Johannes, 1988; Holtz & Johannes, 1991). We therefore conclude that the melting conditions for the garnet- and cordierite-bearing granites were between 800 and 900°C. Similar temperatures have been calculated by McDermott et al. (1996), on the basis of the monazite solubility model, and are also supported by experimental results (Vielzeuf & Montel, 1994), which show that the P–T domains between 800°C at 2 kbar and 900°C at 10 kbar are the regions where most crustal metasedimentary rocks can melt under fluid-absent conditions.
Fractional crystallization
The main evolutionary process for the garnet- and cordierite-bearing granites, especially the linear variation in CaO, Na2O, FeO and TiO2 (Fig. 6) as well as the decreasing REE content with increasing SiO2 content (Fig. 9), indicates igneous fractionation involving plagioclase, biotite, Ti–Fe oxides and some REE-rich accessory phases. Fractionation of plagioclase is also indicated by the positive correlation between Sr contents and the size of the negative Eu anomaly. Within the group I and group II granites Sr decreases from 100 ppm to 25 ppm whereas the size of the negative Eu anomaly increases from 0.66 to 0.14. The Ba–Sr variation (Fig. 13) indicates that the fractionating assemblage consists mainly of plagioclase and K-feldspar with less biotite. A fractionation process including REE-rich accessory phases is also monitored by the group II granites, which have lower Lan/Ybn, Lan/Smn and Gdn/Ybn ratios than group I granites. The group III granites (Fig. 7c) have very low total REE, a pronounced positive Eu anomaly, moderate Lan/Smn but low Gdn/Ybn ratios. Together with their major and trace element characteristics (high K2O, Ba, Pb and Rb contents, low Zr contents) these samples may represent strongly differentiated feldspar-rich granitic rocks with low amounts of monazite and zircon (Miller & Mittlefehldt, 1982). There is also a positive correlation between Sr contents and the size of the positive Eu anomaly, suggesting accumulation of plagioclase within these samples.
|
Restite entrainment and wall-rock contamination
White & Chappell (1977)
and Chappell et al. (1987) among others, have argued that linear chemical variations for all elements in granitoid suites reflect progressive separation of melt from restite. However, any restite unmixing process should create near-linear relationships between the melts, the residue and the source rocks, a correlation not exhibited by the different rocks that crop out in the OGMC. The samples are plotted in the Al-(K + Na + 2Ca) vs Mg + Fe + Ti diagram (Fig. 14) according to Debon & Le Fort (1983). The metasediments, migmatites and garnet- and cordierite-bearing granites show a positive correlation between the two variables. Some granites (samples 89.71 and 89.75) have higher FeO and MgO contents, and plot along the trend towards the xenoliths, metasediments and migmatites. The xenoliths plot off the trend given by the country rocks and granites, suggesting that they are not sufficiently mafic to represent residues. It is therefore appropriate to ask if the presence of these components reflects unmixing of restite from the source region or wall-rock contamination coupled with more advanced stages of melting. In some Harker diagrams (FeO, TiO2, MgO, Al2O3) the trend between the granites, metasediments, migmatites and xenoliths is roughly linear but for other elements (Na2O, CaO, K2O) it is not. The steep trend for Na2O and CaO of the garnet- and cordierite-bearing granites has been explained by plagioclase fractionation and only the sample with the highest CaO content lies on the trend of the migmatites and metasediments. It is interesting to note that some group II and group III granites (samples 89.75 and 89.57) have lower Sr, Zr and REE abundances, implying that fractional crystallization involving plagioclase and REE-rich accessory phases must have occurred before entrainment of Fe/Mg-rich material. This observation suggests that the xenoliths that are now incorporated into the granites do not correspond to the residue of the melting event that produced the intrusive granites, but are more probably residues resulting from extensive degrees of partial melting of incorporated metasedimentary material. It seems that the restite unmixing model is not applicable, at least for these rocks, because the process of fractional crystallization would have also removed the dense xenolith material from the host granite during differentiation.
|
To quantify the inferred contamination process we have used simple mass balance calculations (e.g. Ugidos & Recio, 1993
) that take the composition of the group I granites, the most contaminated granite sample 89.57 and the average composition of the xenoliths into account. A simple mixing calculation of the form
 |
|
|
Implications for the Petrogenesis of Crustally Derived Granites |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
The generation of garnet- and cordierite-bearing granites is usually interpreted in terms of partial melting of metasediments. The occurrence of different types of intrusive granites as isolated sheets and dykes implies that different melt fractions can exist, not only in the form of large single intrusions, but also by the successive expulsion of numerous magma batches, which can collect and ascend into the upper crust. Numerical models (e.g. Clemens & Mawer, 1992
; Petford et al., 1993) suggest that ascent through dykes is the most efficient way to transfer significant volumes of granitic magma through the crust. A consequence of ascent via dykes is that granitic bodies can grow incrementally via the addition of pulses of magma from chemically distinct sources. In the absence of efficient mixing, the aggregation of different batches of magma may produce heterogeneous plutons. On the basis of field evidence it is obvious that these different magma batches observed in the OGMC do not mix to form a hybrid magma. |
Formation of the Leucosomes |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
The formation of the migmatites can be attributed primarily to heating close to the regional metamorphic climax, but additional heat input from intruding granitic melts was probably also important. Several lines of evidence indicate that the leucosomes of the stromatic migmatites formed more or less in situ: (1) the textural relationships between leucosomes and migmatites are often transitional and predominantly concordant with the metasediments; (2) migmatization is more extensive close to broader, clearly intrusive granitic sheets, where the migmatites are sometimes characterized by irregular patches of leucosome material that grade into country rock (nebulites); (3) most leucosomes have not migrated far from their source; (4) the often irregular, diffuse to cross-cutting relationships observed between the leucocratic material and the mafic selvages of the migmatites are inconsistent with an origin by metamorphic differentiation but consistent with anatexis. According to McLellan (1983)
, the development of distinct melanosomes and the disruption of the metasedimentary banding are also consistent with melting, and (5) in comparison with the adjacentmelanosome the fabric of the leucosomes is rather granoblastic and isotropic. Limited fluid-present breakdown of biotite produced the leucosomes as a result of the reaction
 |
According to the restite-unmixing model, the composition of the source should fall on a straight line between the melt and the residue. It is evident from the Harker diagrams that the leucosomes plot at the upper end on all major element plots (Fig. 6). Therefore, the leucosomes are good candidates to represent initial granitic melts from partial melting of pelitic metasediments. The non-minimum composition of the leucosomes might be explained by rapid extraction of some melt, which prevented rehydration of garnet as the melt freezes (e.g. Stüwe & Powell, 1989). Such an extraction process can be expected to happen if crystallization is synchronous with deformation. If the water-saturated leucosomes were frozen in a closed system, most of the cordierite should have back-reacted to form some biotite and aluminosilicate. This is not the case, indicating that some portion of the leucosomes did not crystallize in situ, and suggests that some hydrous melt has escaped from the rock volume (e.g. Ellis & Obata, 1992).
An important property of leucosomes is that their chemistry should match that of liquids formed during partial fusion, as they are assumed to be almost pure melts having limited opportunity to fractionate. Low melt-fraction granitic melts from migmatitic complexes, when their source can be clearly identified, provide invaluable examples for studying chemical fractionation during anatexis. The leucosomes from the stromatic migmatites have high Na2O (>4.2 wt %) and CaO (>1.7 wt %) contents, but low K2O (<2.4 wt %) contents. Rb/Sr ratios are low (0.59–0.73) and Sr/Ba ratios are high (0.53–0.95), indicating that these leucosomes were generated under water-saturated conditions (Harris & Inger, 1992). The view that fluid-present melting was operative is also supported by the nearly identical 18O values of the leucosomes and the melanosomes.
Different major and trace element contents of the leucosomes relative to the intrusive granites imply that the leucosomes have had a different source rock and/or had suffered different melting conditions. Commonly, anatectic leucosomes are depleted in LREE and HREE except for Eu, whereas the corresponding melanosomes are enriched in these elements (Sawyer & Barnes, 1988; Watt & Harley, 1993; Whitney & Irving, 1994; Watt et al., 1996). This behaviour is usually explained by the formation of the leucosomes through segregation of quartz and feldspar, explaining the positive Eu anomaly, whereas the mafic components (e.g. garnet) and accessory phases remained in the restitic melanosome. Low abundances of REE in anatectic leucosomes are attributed to the observation that monazite, apatite and zircon, the main hosts for REE, fail to equilibrate with the melts, because (1) melt extraction rate exceeded accessory phase dissolution, (2) these minerals were shielded during melt segregation and extraction, or (3) undersaturation with respect to H2O leads to limited dissolution of accessory minerals (e.g. Barbey et al., 1989, 1990; Watt & Harley, 1993; Whitney & Irving, 1994; Carrington & Watt, 1995). In the Damaran case, we suggest that the fast segregation rates of the newly formed melts and some shielding of accessory phases by major rock-foming minerals preserved the disequilibrium features observed in the Damaran leucosomes.
Modelling of leucosome-forming process and estimation of melt fraction
To place contraints on the melting process a disequilibrium melting model (e.g. Whitney & Irving, 1994) was applied to the leucosomes:
 |
|
The degree of partial melting (F) to produce a leucosome can be estimated if the source (Co), leucosome (CL) and residue (Cr) compositions are known. An alternative approach is to use the elements that are strongly concentrated into the residue (i.e. TiO2, FeOtot, MgO, Sc, Cr, V, Co) such that the mass balance calculation Co = FCL + (1 – F)Cr reduces to F = (Cr – Co)/Cr (Sawyer, 1991
). Melting is then estimated using source-rock and residuum compositions only. The results are presented in Table 6, and it can be seen that the degree of partial melting is variable but is around 22% in the case of partial melting of metapelite and of the order of 30% in the case of partial melting of metagreywacke sources.
|
From field observations it is not entirely clear whether metapelites or metagreywackes are the sources of the migmatites. SiO2-poor metapelites are probably less fertile sources because of an excess of refractory material rich in MgO, FeOtot and Al2O3, although they contain significant amounts of H2O, usually located in mica. On the other hand, metagreywackes have higher amounts of the easily fusible components quartz and feldspar, despite the observation that they usually have lower content of micas and hence, lower contents of H2O (e.g. Conrad et al., 1988
; LeBreton & Thompson, 1988; Vielzeuf & Holloway, 1988; Patiño Douce & Johnston, 1991; Vielzeuf & Montel, 1994). Patiño Douce & Johnston (1991) have shown that the optimum melting mode to generate peraluminous felsic melts is 36–41% biotite, 21–23% plagioclase, 31–34% quartz and 6–9% aluminosilicate. Table 7 gives the calculated modal abundances of the major rock-forming minerals of the metapelites, metagreywackes and migmatites using the computer program MONA (Metzner & Grimmeisen, 1990). It is evident that the abundances of biotite, plagioclase, quartz and sillimanite of the metapelites are close to the values given by Patiño Douce & Johnston (1991) and, therefore, these metapelites are good candidates for the sources of the migmatites. Siliceous metagreywackes are too poor in biotite and too rich in quartz and plagioclase, whereas the nebulites are too rich in biotite and too poor in plagioclase and quartz. The latter observation suggests that the nebulites represent residual rocks that have lost a peraluminous melt.
|
|
The Migmatite–Granite Link |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
In the study area the occurrence of small layer-parallel leucosomes (i.e. leucosomes L 1 to L 3 from the stromatic migmatites) seems to represent an earlier stage of migmatization in contrast to the nebulitic migmatites. The deformation history of the central part of the Pan-African Damara orogen has not been investigated in detail but it seems reasonable to assume that separation of melt and residue in this case was a deformation-assisted process. Within the stromatic migmatites the melt fraction was obviously not entirely removed as it formed, suggesting that the melt generation rate exceeded the melt segregation rate. We propose that such a process might occur as a result of intrusion of hot melts and the influx of external water as the magmatic body cooled. In contrast, the nebulites have a residual mineralogy and geochemical composition, and it seems likely that they lost a melt fraction as a consequence of increasing melt production. The inferred migration of melt probably increased the local permeability of the country rocks and as more melt is produced disaggregation of incorporated wall rocks as xenoliths and schlieren-like material might have occurred (Fig. 3c–e). Such a magma with a high content of mafic components (minerals, xenoliths and wall-rock material) may be, in principle, the source for intrusive granites at higher structural levels, which are usually more mafic than theleucosomes (Sawyer, 1998
). Sawyer (1998) pointed out that once the melt fraction was high enough for magma flow, dispersed residual components might form aggregated xenoliths because of their non-uniform physical behaviour in a melt. This behaviour might explain the occurrence of dispersed xenoliths or their orientation in trails probably as a consequence of different strain gradients during magma flow. As these magmas, with their large content of mafic residual components, move away from their site of separation from the residue they must cool and consequently they begin to crystallize. At this stage, fractional crystallization becomes the dominant process in changing the mineralogical and chemical composition of granitic magmas. |
Conclusions |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
Granite that invades and crystallizes in a metasedimentary pile can be the likely source of the fluid and heat that promoted partial melting and migmatization of the surrounding country rocks. We propose that significant melting is possible in the upper amphibolite facies provided that there is an excess of H2O (e.g. Yardley & Barber, 1991
) as well as heat at the peak of metamorphism. It is possible that the stromatic migmatites [metatexites sensu Brown (1979)] represent lower-temperature migmatites and the nebulites are higher-temperature migmatites (diatexites). As more melt is produced during rising temperature, a pervasive melt fraction is produced in the nebulites and hence more effective melt migration is possible (e.g. Tracy & Robinson, 1983).
The leucosomes of the stromatic migmatites are too felsic and too depleted in FeO, MgO and TiO2, and have Zr and REE contents too low to be representative of primary intrusive granitic magmas. Their chemical compositions suggest that they are not the precursor of granitic melts of sizeable masses found in the complex. These intrusive granitic melts, which are derived from metasedimentary sources, usually have higher FeO, TiO2, MgO, LREE, Th, U and Zr contents (e.g. Sawyer, 1996). This study suggests that (1) melts generated in most migmatite terranes are unlikely to represent the precursors of S-type granitic plutons (e.g. Sawyer, 1996) and (2) granitic melts generated in the deeper crust must undergo considerable modification during ascent to achieve the chemical characteristics of pluton-sized granitic bodies. This study also emphasizes that significant volumes of crustally derived melts intrude amphibolite-facies terranes rather than being derived from them. Therefore, because volatile phase-present melting is common in most amphibolite-facies migmatite terranes, it seems unlikely that these terranes are the source for large melt fractions, which constitute larger-sized granite bodies. However, there has been good evidence that melting with a volatile phase present may produce sizeable granitic bodies in some cases (e.g. Brown, 1979; Wickham, 1987; Montel et al., 1992). The nebulites contain some melt and mafic material. Field evidence suggests that mobilization of melt is restricted; however, the dominant restitic character suggests that some separation of melt and residue has occurred. The gradual change between the concentration of mafic material and schlieren-like xenoliths on one side and increasing concentration of melt on the other side (Fig. 3d) might further indicate some separation of melt and residue. Only these rare residual nebulites can be considered to be the sources of intrusive granitic melts, but the limited areal extent makes them also unlikely to be the precursor of the intrusive granites, at least in the case of the Damaran granites.
The garnet- and cordierite-bearing granites are derived by fluid-absent partial melting of Al-rich metapelites in the deeper crust. They carry significant amounts of residual xenoliths and enclaves of country rock metasediments and migmatites. The xenoliths do not represent unmixed restite from the site of the origin of the granites; instead, they are more likely to be the result of high degrees of melting of incorporated wall-rock material. This study has also shown that some, but not all, intrusive granitic melts may represent a mixture between melt and entrained solid phases rather than pure melts (Holtz & Barbey, 1991).
Migmatite terranes may form in part in response to contact effects of plutonism (e.g. Finger & Clemens, 1995); although they are a substantial part of many high-grade orogenic belts, our considerations on migmatite formation confirm that in situ leucosomes do not represent the original source for volumetrically significant granite bodies.
|
|
|
|
Acknowledgements |
|---|
This work forms part of the Ph.D. thesis of S.J., which was partly funded by the National Science Foundation (Ho-1078/1-2). Special thanks go to M. Raith for making microprobe facilities possible, to B. Spiering (and colleagues) for assistance during microprobe work at the Mineralogical and Petrological Institute, Bonn (Germany), and to I. Klink-Bakri and D. Dohle from the Bonn fluorine-laboratory crew for careful analytical work. W. Johannes, F. Holtz, N. B. W. Harris, F. McDermott and an anonymous reviewer made many helpful comments, which substantially improved the paper. Iris Bambach (Max-Planck-Institut, Mainz) did a superb job in managing the line drawings.
* Corresponding author. Fax: +49 6131-371-051. e-mail: sjung{at}mpch-mainz.mpg.de
|
References |
|---|
|
TOPABSTRACTIntroductionGeological SettingDescription of the Country...Geochemistry of the High-Grade...DiscussionImplications for the Evolution...Implications for the...Formation of the LeucosomesThe Migmatite-Granite LinkConclusionsReferences |
|---|
Allen P. L., Barr S. M. The Ellison Lake pluton: a cordierite-bearing monzogranitic intrusive body in southwestern Nova Scotia. Canadian Mineralogist (1983) 21:583–590.[Web of Science]
Arth J. G. Behaviour of trace elements during magmatic processes—a summary of theoretical models and their applications. Journal of Research of the US Geological Survey (1976) 4:41–47.[Web of Science]
Ayres M., Harris N. REE fractionation and Nd-isotope disequilibrium during crustal anatexis: constraints from Himalayan leucogranites. Chemical Geology (1997) 139:249–269.[Web of Science]
Barbarin B. Genesis of the two main types of peraluminous granitoids. Geology (1996) 24:295–298.
Barbero L., Villaseca C. The Layos Granite, Hercynian Complex of Toledo (Spain): an example of parautochthonous restite-rich granite in a granulitic area. Transactions of the Royal Society of Edinburgh, Earth Sciences (1992) 83:127–138.[Web of Science]
Barbero L., Villaseca C., Rogers G., Brown P. E. Geochemical and isotopic disequilibrium in crustal melting: an insight from the anatectic granitoids from Toledo, Spain. Journal of Geophysical Research, Solid Earth (1995) 100:15745–15765.[Web of Science]
Barbey P., Bertrand J.-M., Angoua S., Dautel D. Petrology and U/Pb geochronology of the Telohat migmatites, Aleksod, Central Hoggar, Algeria. Contributions to Mineralogy and Petrology (1989) 101:207–219.[Web of Science]
Barbey P., Macaudiere J., Nzenti J. P. High-pressure dehydration melting of metapelites: evidence from the migmatites of Yaoundé (Cameroon). Journal of Petrology (1990) 31:401–427.
Bea F., Fershtater G., Corretgé L. G. The geochemistry of phosphorus in granite rocks and the effect of aluminium. Lithos (1992) 29:43–56.[Web of Science]
Bea F., Pereira M. D., Stroh A. Mineral/leucosome trace-element partitioning in a peraluminous migmatite (a laser ablation–ICP-MS study). Chemical Geology (1994) 117:291–312.[Web of Science]
Birch W. D., Gleadow A. J. W. The genesis of garnet and cordierite in acid volcanic rocks: evidence from the Ceberean Cauldron, Central Victoria, Australia. Contributions to Mineralogy and Petrology (1974) 45:1–13.[Web of Science]
Boynton W. V. Geochemistry of the rare earth elements: meteorite studies. In: Rare Earth Element Geochemistry—Henderson P., ed. (1984) New York: Elsevier. 63–114.
Brown M. The petrogenesis of the St. Malo migmatite belt, Armorican Massif, France, with particular reference to the diatexites. Neues Jahrbuch für Mineralogie, Abhandlungen (1979) 135:48–74.
Brown M., D'Lemos R. S. The Cadomian granites of Mancellia, north-east Armorican Massif of France: relationship to the St. Malo migmatite belt, petrogenesis and tectonic setting. Precambrian Research (1991) 51:393–427.[Web of Science]
Bühn B., Häussinger H., Kramm U., Kukla C., Kukla P. A., Stanistreet I. G. Tectonometamorphic patterns developed during Pan-African continental collision in the Damara inland belt. Chemie der Erde (1994) 54:329–354.[Web of Science]
Carrington D. P., Watt G. R. A geochemical and experimental study of the role of K-feldspar during water-undersaturated melting of metapelites. Chemical Geology (1995) 122:59–76.[Web of Science]
Chappell B. W., White A. J. R., Wyborn D. The importance of residual source material (restite) in granite petrogenesis. Journal of Petrology (1987) 28:1111–1138.
Clarke D. B. Cordierite in felsic igneous rocks: a synthesis. Mineralogical Magazine (1995) 59:311–325.[Abstract]
Clayton R. N., Mayeda T. D. The use of bromine pentafluoride in the extraction of oxygen from oxides and silicates for isotope analysis. Geochimica et Cosmochimica Acta (1963) 27:43–52.[Web of Science]
Clemens J. D., Mawer C. K. Granite magma transport by fracture propagation. Tectonophysics (1992) 204:339–360.[Web of Science]
Clemens J. D., Wall V. J. Origin and crystallisation of some peraluminous (s-type) granitic magmas. Canadian Mineralogist (1981) 19:111–131.
Clemens J. D., Wall V. J. Controls on the mineralogy of S-type volcanic and plutonic rocks. Lithos (1988) 21:53–66.[Web of Science]
Collins W. J. Lachlan Fold Belts granitoids: products of three-component mixing. Transactions of the Royal Society of Edinburgh, Earth Sciences (1996) 87:171–181.[Web of Science]
Conrad W. K., Nicholls I. A., Wall V. J. Water-saturated and -undersaturated melting of metaluminous and peraluminous crustal compositions at 10 kbar: evidence for the origin of silicic magmas in the Taupo volcanic zone, New Zealand, and other occurrences. Journal of Petrology (1988) 29:765–803.
Debon F., Le Fort P. A chemical–mineralogical classification of common plutonic rocks and associations. Transactions of the Royal Society of Edinburgh, Earth Sciences (1983) 73:135–149.
Deniel C., Vidal P., Fernandez A., LeFort P., Peucat J.-J. Isotopic study of the Manaslu granite (Himalaya, Nepal): inferences on the age and source of Himalayan leucogranites. Contributions to Mineralogy and Petrology (1987) 96:78–92.[Web of Science]
Ellis D. J., Obata M. Migmatite and melt segregation at Cooma, New South Wales. Transactions of the Royal Society of Edinburgh (1992) 83:95–106.
Finger F., Clemens J. D. Migmatization and ’secondary’ granitic magmas: effects of emplacement and crystallization of ‘primary’ granitoids in Southern Bohemia, Austria. Contributions to Mineralogy and Petrology (1995) 120:311–326.[Web of Science]
Flood R. H., Shaw S. E. Cordierite-bearing granite suite from New England batholith, N.S.W. Australia. Contributions to Mineralogy and Petrology (1975) 52:157–164.[Web of Science]
Fujimaki H. Partition coefficient of Hf, Zr, and REE between zircon, apatite, and liquid. Contributions to Mineralogy and Petrology (1986) 94:42–45.[Web of Science]
Georget Y., Fourcade S. REE partitioning in magmatic cordierite: implications for cordierite-bearing granitoids as exemplified by the Huelgoat intrusion (Brittany, France). Neues Jahrbuch für Mineralogie, Abhandlungen (1988) 158:225–240.[Web of Science]
Green T. H. Experimental generation of cordierite- and garnet-bearing granitic liquids from a pelitic composition. Geology (1976) 4:85–88.
Haack U., Gohn E., Klein J. A. Rb/Sr ages of granitic rocks along the middle reaches of the Omaruru River and the timing of orogenetic events in the Damara Belt (Namibia). Contributions to Mineralogy and Petrology (1980) 74:349–360.[Web of Science]
Haack U., Hoefs J., Gohn E. Constraints on the origin of Damaran granites by Rb/Sr and 18O data. Contributions to Mineralogy and Petrology (1982) 79:279–289.[Web of Science]
Haack U., Heinrich H., Boneß M., Schneider A. Loss of metals from pelites during regional metamorphism. Contributions to Mineralogy and Petrology (1984) 85:116–132.[Web of Science]
Harris N. B. W., Inger S. Trace element modelling of pelite derived granites. Contributions to Mineralogy and Petrology (1992) 110:46–56.[Web of Science]
Hartmann O., Hoffer E., Haack U. Regional metamorphism in the Damara orogen: interaction of crustal motion and heat transfer. Geological Society of South Africa, Special Publication (1983) 11:233–241.
Häussinger H., Okrusch M., Scheepers D. Geochemistry of premetamorphic hydrothermal alteration of metasedimentary rocks associated with the Gorob massive sulfide prospect, Damara orogen, Namibia. Economic Geology (1993) 88:72–90.
Heinrichs H., Herrmann A. G. Praktikum der analytischen Geochemie (1990) Berlin: Springer.
Hensen B. J., Green D. H. Experimental study of the stability of cordierite and garnet in pelitic compositions at high temperatures and pressures. Synthesis of experimental data and geological applications. Contributions to Mineralogy and Petrology (1973) 38:151–166.[Web of Science]
Hoernes S., Hoffer E. Stable isotope evidence for fluid-present and fluid-absent metamorphism in metapelites from the Damara Orogen, Namibia. Contributions to Mineralogy and Petrology (1985) 90:322–330.[Web of Science]
Hoernes S., van Reenen D. D. The oxygen isotopic composition of granulites and retrogressed granulites from the Limpopo Belt as a monitor of fluid–rock interaction. Precambrian Research (1992) 55:353–364.[Web of Science]
Holtz F., Barbey P. Genesis of peraluminous granites. II. Mineralogy and chemistry of the Tourem complex (north Portugal)—sequential melting vs. restite unmixing. Journal of Petrology (1991) 32:959–978.
Holtz F., Johannes W. Genesis of peraluminous granites. I. Experimental investigation of melt composition at 3 and 5 kbar and various H2O activities. Journal of Petrology (1991) 32:935–958.
Jung S., Hoffer E., Masberg P., Hoernes S. Geochemistry of granitic in-situ low-melt fractions—an example from the Central Damara orogen, Namibia. Special Publications of the Geological Survey of Namibia (1995) 10:21–31.
Jung S., Mezger K., Masberg P., Hoffer E., Hoernes S. Petrology of an intrusion-related high-grade migmatite—implications for partial melting of metasedimentary rocks and leucosome-forming processes. Journal of Metamorphic Geology (1998a) 16:425–445.[Web of Science]
Jung S., Mezger K., Hoernes S. Petrology and geochemistry of post-collisional metaluminous A-type granites—a major and trace element and Nd–Sr–Pb–O-isotope study from the Proterozoic Damara Belt, Namibia. Lithos (1998b) 45:147–175.[Web of Science]
Kretz R. Symbols for rock forming minerals. American Mineralogist (1983) 68:277–279.[Abstract]
LeBreton N., Thompson A. B. Fluid-absent (dehydration melting) of biotite in metapelites in the early stages of crustal anatexis. Contributions to Mineralogy and Petrology (1988) 99:226–237.[Web of Science]
Lechler P. J., Desilets M. O. A review of the use of loss on ignition as a measurement of total volatiles in whole rock analysis. Chemical Geology (1987) 63:341–344.[Web of Science]
Mahood G., Hildreth W. Large partition coefficient for trace elements in high-silica rhyolites. Geochimica et Cosmochimica Acta (1983) 47:11–30.[Web of Science]
Maillet L. A., Clarke D. B. Cordierite in the peraluminous granites of the Meguma Zone, Nova Scotia, Canada. Mineralogical Magazine (1985) 49:695–702.[Web of Science]
Masberg H. P., Hoffer E., Hoernes S. Microfabrics indicating granulite-facies metamorphism in the low-pressure central Damara Orogen, Namibia. Precambrian Research (1992) 55:243–257.[Web of Science]
McDermott F., Harris N. B. W., Hawkesworth C. J. Geochemical constraints on crustal anatexis: a case study from the Pan-African Damara granitoids of Namibia. Contributions to Mineralogy and Petrology (1996) 123:406–423.[Web of Science]
McLellan E. L. Contrasting textures in metamorphic and anatectic migmatites: an example from the Scottish Caledonides. Journal of Metamorphic Geology (1983) 1:241–262.[Web of Science]
Metzner C., Grimmeisen W. MONA: a user-friendly computer program for calculating the modal mineralogy of rocks from chemical analyses. European Journal of Mineralogy (1990) 2:735–738.
Miller C. F. Are strongly peraluminous magmas derived from pelitic sedimentary sources. Journal of Geology (1985) 93:673–689.[Web of Science]
Miller C. F., Mittlefehldt D. W. Depletion of light rare-earth elements in felsic magmas. Geolog (1982) 10:129–133.
Miller R. Mc. G. The Pan-African Damara Orogen of South West Africa/Namibia. Geological Society of South Africa, Special Publication (1983) 11:431–515.
Mittlefehldt D. W., Miller C. F. Geochemistry of the Seawater Wash pluton, California: implications for ‘anomalous’ trace element behaviour during differentiation of felsic magmas. Geochimica etCosmochimica Acta (1983) 47:109–124.
Miyashiro A. Cordierite–indialite relations. American Journal of Science (1957) 255:43–62.
Montel J. M. A model for monazite/melt equilibrium and application to the generation of granitic magmas. Chemical Geology (1993) 110:127–146.[Web of Science]
Montel J. M., Marignac C., Barbey P., Pichavant M. Thermobarometry and granite genesis: the Hercynian low-P, high-T Velay anatectic dome (French Massif Central). Journal of Metamorphic Geology (1992) 10:1–15.[Web of Science]
Obata M., Yoshimura Y., Nagakawa K., Odawara S., Osanai Y. Crustal anatexis and melt migrations in the Higo metamorphic terrane, west–central Kyushu, Kumamoto, Japan. Lithos (1994) 32:135–147.[Web of Science]
Patiño Douce A. E., Johnston A. D. Phase equilibria and melt productivity in the pelitic system: implications for the origin of peraluminous granitoids and aluminous granulites. Contributions to Mineralogy and Petrology (1991) 107:202–218.[Web of Science]
Pattison D. R. M., Harte B. Evolution of structurally contrasting migmatites in the 3-kbar Ballachulish aureole, Scotland. Journal of Metamorphic Geology (1988) 6:475–494.[Web of Science]
Petford N., Kerr K. C., Lister R. C. Dike transport of granitoid magmas. Geology (1993) 21:845–848.
Phillips G. N., Wall V. J., Clemens J. D. Petrology of the Strathbogie batholith—a cordierite bearing granite. Canadian Mineralogist (1981) 19:47–63.
Phillips G. N., Groves D. I., Reed K. Geochemistry of the Kuiseb metasediments around Windhoek, Namibia. Communications of the Geological Survey of Namibia (1989) 5:19–30.
Pichavant M., Montel J. M., Richard L. R. Apatite solubility in peraluminous liquids: experimental data and an extension of the Watson–Harrison model. Geochimica et Cosmochimica Acta (1992) 56:3855–3861.[Web of Science]
Puziewicz J., Johannes W. Phase equilibria and compositions of Fe–Mg–Al minerals and melts in water-saturated peraluminous granitic systems. Contributions to Mineralogy and Petrology (1988) 100:156–168.[Web of Science]
Rapp R. P., Ryerson F. J., Miller C. F. Experimental evidence bearing on the stability of monazite during crustal anatexis. Geophysical Research Letters (1987) 14:307–310.[Web of Science]
Sawyer E. W. Disequilibrium melting and the rate of melt–residuum separation during migmatization of mafic rocks from the Grenville Front, Quebec. Journal of Petrology (1991) 32:701–738.
Sawyer E. W. Melt segregation and magma flow in migmatites: implications for the generation of granite magmas. Transactions of the Royal Society of Edinburgh, Earth Sciences (1996) 87:85–94.[Web of Science]
Sawyer E. W. Formation and evolution of granite magmas during crustal reworking: the significance of diatexites. Journal of Petrology (1998) 39:1147–1167.
Sawyer E. W., Barnes S.-J. Temporal and compositional differences between subsolidus and anatectic migmatite leucosomes from the Quetico metasedimentary belt, Canada. Journal of Metamorphic Geology (1988) 6:437–450.[Web of Science]
Schneider A. The chemical composition of common metamorphic metasediments in the Damara orogen. In: Intracontinental Fold Belts—Martin H., Eder F., eds. (1983) Berlin: Springer. 655–678.
Schnetzler C. C., Philpotts J. A. Partition coefficient of rare-earth elements between igneous matrix mineral and rock-forming mineral phenocrysts. Geochimica et Cosmochimica Acta (1970) 34:331–340.[Web of Science]
Shieh Y.-N. High-18O granitic plutons from the Frontenac Axis, Grenville Province of Ontario, Canada. Geochimica et Cosmochimica Acta (1985) 49:117–123.[Web of Science]
Stüwe K., Powell R. Metamorphic segregations associated with garnet and orthopyroxene porphyroblast growth: two examples from the Larseman Hills, East Antarctica. Contributions to Mineralogy and Petrology (1989) 103:523–530.[Web of Science]
Taylor S. R., McLennan S. M. The Continental Crust: its Composition and Evolution (1985) London: Blackwell Science.
Tracy R. J., Robinson P. Acadian migmatite types in central Massachusetts. In: Migmatites, Melting and Metamorphism—Atherton M. P., Gribble C. D., eds. (1983) Nantwich, UK: Shiva. 163–173.
Ugidos J. M., Recio C. Origin of cordierite-bearing granites by assimilation in the Central Iberian Massif (CIM), Spain. Chemical Geology (1993) 103:27–43.[Web of Science]
Vernon R. H., Collins W. J. Igneous microstructures in migmatites. Geology (1988) 16:1126–1129.
Vielzeuf D., Holloway J. R. Experimental determination of the fluid-absent melting relations in the pelitic system. Consequences for crustal differentiation. Contributions to Mineralogy and Petrology (1988) 98:257–276.[Web of Science]
Vielzeuf D., Montel J. M. Partial melting of metagreywackes. Part I. Fluid-absent experiments and phase relations. Contributions to Mineralogy and Petrology (1994) 17:375–393.
Villaseca C., Barbero L. Chemical variability of Al–Ti–Fe–Mg minerals in peraluminous granitoid rocks from Central Spain. European Journal of Mineralogy (1994) 6:691–710.
Watson E. B., Harrison T. M. Zircon saturation revisited: temperature and composition effects in a variety of crustal magma types. Earth and Planetary Science Letters (1983) 64:295–304.[Web of Science]
Watt G. R., Harley S. L. Accessory phase controls on the geochemistry of crustal melts and restites produced during water-undersaturated partial melting. Contributions to Mineralogy and Petrology (1993) 114:550–566.[Web of Science]
Watt G. R., Burns I. M., Graham G. A. Chemical characteristics of migmatites: accessory phase distribution and evidence for fast melt segregation. Contributions to Mineralogy and Petrology (1996) 125:100–111.[Web of Science]
White A. J. R., Chappell B. W. Ultrametamorphism and granite genesis. Tectonophysics (1977) 43:7–22.[Web of Science]
White A. J. R., Chappell B. W. Some supracrustal (S-type) granites of the Lachlan Fold Belt. Transactions of the Royal Society of Edinburgh, Earth Sciences (1988) 79:169–181.
White A. J. R., Chappell B. W. Per migma ad magma downunder. Geological Journal (1990) 25:221–225.[Web of Science]
Whitney D. L., Irving A. J. Origin of K-poor leucosomes in a metasedimentary migmatite complex by ultrametamorphism, syn-metamorphic magmatism and subsolidus processes. Lithos (1994) 32:173–192.[Web of Science]
Wickham S. M. Crustal anatexis and granite petrogenesis during low pressure metamorphism: the Trois Seigneurs Massif, Pyrenees, France. Journal of Petrology (1987) 28:127–169.
Williamson B. J., Downes H., Thirlwall M. F., Beard A. Geochemical constraints on restite composition and unmixing in the Velay anatectic granite, French Massif Central. Lithos (1997) 40:295–319.[Web of Science]
Yardley B. W. D., Barber J. P. Melting reactions in the Connemara schists: the role of water infiltration in the formation of amphibolite facies migmatites. American Mineralogist (1991) 76:848–856.[Abstract]
Yurimoto H., Duke E. F., Papike J. J., Shearer C. K. Are discontinuous chondrite-normalized REE patterns in pegmatititc systems the result of monazite fractionation. Geochimica et Cosmochimica Acta (1990) 54:2141–2145.[Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||