Journal of Petrology Volume 42 Number 9 Pages 1751-1772 2001
© Oxford University Press 2001
Partial Melting of Aluminous Metagreywackes in the Northern Sierra de Comechingones, Central Argentina
1DEPARTAMENTO DE GEOLOGÍA, UNIVERSIDAD NACIONAL DE RÍO CUARTO, 5800 RÍO CUARTO, ARGENTINA
2DEPARTMENT OF GEOLOGY, UNIVERSITY OF GEORGIA, ATHENS, GA 30602, USA
Received June 29, 2000; Revised typescript accepted March 5, 2001
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
We describe a suite of metamorphic and migmatitic rocks from the Sierra de Comechingones (Sierras Pampeanas of Central Argentina) that include unmelted gneisses, migmatites and refractory granulites. The gneisses are aluminous greywackes metamorphosed in the amphibolite grade and are likely to have been the protoliths for the higher-grade migmatites and granulites. Mineralogical characteristics and major and trace element compositions show that metatexite migmatites, diatexite migmatites and granulites are all melt-depleted rocks. The migmatites (both metatexites and diatexites) have undergone H2O-fluxed melting and lost
20% melt, whereas the granulites have lost as much as 60% melt, formed chiefly by dehydration-melting of biotite. The granulites are rocks from which essentially all the granitic components have been extracted, whereas the migmatites cooled and solidified while still retaining a granitic fraction. The only evidence for the presence of melt in the study area, apart from migmatitic leucosomes, are dikes and sills of leucogranite rich in cumulate K-feldspar. These leucogranite bodies may represent the pathways along which melt was drained from the exposed rocks. Our study shows that in the Sierra de Comechingones both migmatites and granulites represent source regions of anatectic magmas, which were ‘frozen in’ at different stages of development. KEY WORDS: migmatites; granulites; anatexis; metagreywackes; granites
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
The widespread occurrence of migmatites attests to the fact that crustal anatexis is a common phenomenon. There is no consensus on whether or not migmatites represent the source regions of anatectic magmas (e.g. Brown, 1994
; Sawyer, 1998), even though Sawyer (1996) showed that diatexite migmatites have rheological and chemical properties that suggest that they are parental to granite magmas. It is also known that granulites depleted in incompatible major and trace elements are present in the lower continental crust (e.g. Lambert & Heier, 1968). These rocks have been interpreted as refractory residues of partial melting complementary to granitic magmas (e.g. Clemens, 1990; Vielzeuf et al., 1990), and there is ample theoretical and experimental evidence that supports this concept (e.g. Brown & Fyfe, 1970; Thompson, 1982; Vielzeuf & Holloway, 1988; Patiño Douce & Johnston, 1991; Patiño Douce & Beard, 1995). It is thus possible that the source regions of some granitic magmas could consist of a wide spectrum of migmatites and melt-depleted granulites. In this paper we focus on a suite of metamorphic and migmatitic rocks from the Sierras Pampeanas of Argentina. The rocks in the study area include unmelted gneisses, migmatites and refractory granulites. We describe the fabrics, mineral assemblages and field relations of the various rock types and present data on major and trace element bulk-rock compositions. We use petrographic characteristics and major and trace element compositions to show that migmatites and granulites are both residual rocks, which have undergone different degrees of melt depletion. We argue that the migmatites and the granulites represent different sources of granitic magma in the same crustal section.
|
GEOLOGICAL SETTING |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
The study area is centered on the Río Santa Rosa, within the Sierra de Comechingones (Fig. 1). The Sierra de Comechingones is one of the southernmost ranges of the Sierras Pampeanas of Central Argentina (Fig. 1). It is composed of upper amphibolite to granulite facies metamorphic and migmatitic rocks that were derived from a predominantly clastic sedimentary sequence. The most abundant protoliths were aluminous greywackes, accompanied by volumetrically minor limestones (Gordillo, 1984
; Otamendi et al., 1999). If shale intercalations were present in the pre-metamorphic sequence they must have been very thin, as they cannot be recognized in the unmelted gneisses at scales of 1 m or less. Metamorphism was the result of a regional Early Cambrian (Rapela et al., 1998; Sims et al., 1998) tectono-thermal event called the Pampean Orogeny. Peak metamorphism in the study area attained temperatures ranging from 650 to 950°C at nearly uniform pressures of 7–8 kbar (Otamendi et al., 1999).
|
The Sierra de Comechingones was affected by four major deformation phases (Martino et al., 1995). D1 is poorly preserved as a gneissic foliation folded by rootless D2 folds. D2 deformation is characterized by mineral segregation and folding. It took place during prograde metamorphism, probably close to the metamorphic peak (Gordillo, 1984). This timing is suggested by the observation that metatexite leucosomes often nucleate on the quartzo-feldspathic layers that define the S2 foliation. Sigmoidal channels oblique to S2 foliation and filled with discordant leucosomes indicate that the partially molten rocks were subject to shear stress, but it is unclear whether this deformation corresponds to D2 or to an as yet unrecognized deformation episode. This is so because the dominant structural features throughout most of the Sierra de Comechingones correspond to the D3 phase (Martino et al., 1995), which began at conditions close to the metamorphic peak but continued during the cooling stage (Otamendi el al., 1999). D3 deformation is partitioned into high- and low-strain domains, and is associated in the former with retrogression of garnet + cordierite + K-feldspar assemblages to biotite + sillimanite (Otamendi et al., 1999). S3 fabric is penetrative in NNW–SSE-trending high-strain zones that nucleated along the boundaries of homogeneous lithologic blocks. The fabric is weaker and becomes harder to distinguish from S2 in the cores of homogeneous lithologic bodies, such as the Cerro Pelado Dam diatexite massif (Fig. 1). D3 kinematic indicators of shear sense are ambiguous, but most of them seem to indicate that D3 structures formed in response to dextral shearing associated with SSE-directed extensional deformation (Martino et al., 1994). Steeply dipping NNW–SSE-trending mylonitic belts that overprint D3 structures are ascribed to a D4 event (Martino et al., 1994). D4 mylonite belts are spatially associated with the granulites that record the highest metamorphic conditions (Otamendi et al., 1999) and may be responsible for juxtaposing these rocks against lower-grade migmatites. Final exhumation of the Sierra de Comechingones occurred during the Famatinian Orogeny (490–390 Ma) and was accommodated along narrow regional-scale shear belts that are affected by low-grade metamorphic recrystallization (Martino et al., 1995; Rapela et al., 1998).
|
ROCK TYPES AND FIELD RELATIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Several rock types can be distinguished in the Sierra de Comechingones on the basis of their mineral assemblages and morphological characteristics. The spatial distributions of the rock types and the sample localities are shown in Fig. 1 (regional distribution) and Fig. 2 (detail of the Río Santa Rosa area, where the highest-grade rocks occur). The lithological characteristics, mineral compositions and metamorphic evolution of these rocks were discussed by Otamendi et al. (1999)
. The most important features, including the field relations and inferred melting reactions, are also summarized here. Phase assemblages and modal compositions of representative samples are given in Table 1. Detailed mineral compositions have been given by Otamendi et al. (1999, tables 2–6).
|
|
|
|
Biotite–garnet gneisses (Type I)
These are the most abundant rocks in the Sierra de Comechingones (Gordillo, 1984). Type I gneisses have a medium- to coarse-grained granoblastic texture and are composed of the mineral assemblage Qtz + Pl + Bt + Grt + Ap + Zrn + Rt ± Ilm ± Kfs [mineral symbols after Kretz (1983)]. They have a gneissic fabric given by the alternation of quartzo-feldspathic-rich and biotite-rich layers (Fig. 3a). In places it is possible to recognize the preservation of primary centimeter-scale sedimentary bedding (Fig. 3b). This relic bedding is defined by varying biotite/plagioclase ratios. Biotite-rich beds are strongly foliated and are separated from less foliated biotite-poor beds by millimeter-thick quartz-rich layers. Garnet abundances do not vary noticeably across this relic sedimentary bedding. The gneissic foliation is overprinted by a continuous schistosity in D3 high-strain domains. Muscovite and fibrolite commonly appear in these highly strained Type I rocks, but these Al-rich phases are otherwise notoriously absent from Type I gneisses.
Biotite–garnet–K-feldspar–cordierite migmatites (Type II)
These are migmatites that display a wide variety of morphologies, among which metatexites and diatexites [nomenclature after Ashworth (1985)] occur as end-member types. The migmatites are distinguished from Type I gneisses on the basis of mineral assemblages (widespread coexistence of garnet and/or cordierite with K-feldspar) and structural evidence of partial melting (e.g. Sawyer, 1999). Relative to Type I gneisses, biotite in Type II migmatites (see Otamendi et al., 1999) is distinctly richer in Ti (0·15 vs 0·25 Ti atoms p.f.u.) and poorer in octahedral Al (0·38 vs 0·22 Al6 atoms p.f.u). Type II migmatites contain accessory apatite, zircon, monazite and rutile. Monazite occurs as unzoned euhedral grains, either next to or included in garnet, but almost never as isolated crystals in the leucosomes.
Contacts between gneisses and migmatites correspond in most cases to high-strain D3 shear zones, but gradational contacts are observed in some places south of the Río Santa Rosa (Fig. 1). In such cases, migmatites with layer-parallel leucosomes are intercalated with packages of unmelted gneiss. The transition from one rock type to the other is completed over distances of the order of 100 m.
Metatexites are characterized by a well-developed textural and mineralogical layering that results from the alternation of Qtz + Grt + Bt + Pl ± Crd melanosomes with Qtz + Kfs ± Grt ± Crd leucosomes. Some metatexites preserve remnants of the sedimentary bedding observed in the gneisses but others show no vestiges of pre-migmatization fabrics. Metatexites that appear to represent incipient melting contain only layer-parallel leucosomes <1 cm thick (sample SCP24). As melt contents increase layer-parallel leucosomes become thicker (1–2 cm) and a second set of leucosomes oblique to the metamorphic layering appears (Fig. 3c, sample OCP10). Leucosomes eventually coalesce to form discordant dikes and concordant sills of leucogranite (described below). The leucogranite bodies are spatially associated with stromatic migmatites in which melanosomes are interlayered with Kfs + Qtz ± Pl leucosomes <3 mm thick (sample RSR31b). These stromatic migmatites are depleted in Qtz and Pl and enriched in Bt and Grt relative to metatexites that are not associated with leucogranite bodies.
Diatexites (sample RSR32) have a coarse-grained equigranular fabric with disseminated alkali feldspar megacrysts (Fig. 3d). Melanosome and leucosome are less distinct than in the metatexites, but it is nevertheless possible to recognize a faint foliation and to identify biotite-rich schlieren oriented roughly parallel to this foliation. The diatexites also contain metatexite rafts, refractory xenoliths of mafic and calcareous rocks and leucogranite sills. The transition from metatexite to diatexite occurs over scales of a few meters.
|
Leucogranites
Tabular leucogranite bodies generally thicker than 5 cm and ranging in length from 10 to 100 m are found emplaced within Type II migmatites, but cannot be mapped as independent units at the scale of Figs 1 and 2. They include discordant dikes, lensoidal sills and bodies with irregular borders and variable thickness (Fig. 3e and f). The contacts with the migmatites are commonly sharp. The leucogranite bodies are chiefly composed of alkali feldspar and quartz. Alkali feldspar occurs as subhedral crystals, sometimes megacrysts a few centimeters long, forming an adcumulus texture with interstitial (intercumulus) anhedral quartz. Plagioclase occurs only as inclusions in poikilitic alkali feldspar crystals. Garnet is a common accessory mineral. Its composition is indistinguishable from that of garnet in Type II migmatites. Muscovite occurs only as mats growing on corroded biotite crystals and is thought to be a retrograde phase.
Biotite–garnet–K-feldspar–cordierite–sillimanite migmatites (Type III)
Type III migmatites are homogeneously foliated rocks with a ‘reversed’ schlieren structure in which leucosome streaks occur inside the dominant melanosome (e.g. sample RSR24). These rocks thus contain a greater melanosome/leucosome proportion than Type II metatexites. Large leucogranite bodies such as those associated with Type II migmatites are never found in Type III migmatites. They differ mineralogically from the former in having greater cordierite abundances and in the fact that they always contain abundant sillimanite and some hercynitic spinel (Table 1). Monazite in unzoned euhedral grains is abundant in Type III leucosomes, in contrast to Type II metatexites in which this accessory phase is mostly restricted to the melanosomes. Type III migmatites acquire stromatic structures in D3 high-strain belts (samples RSR04 and RSR16), suggesting that this deformation stage outlasted the presence of melt. Contacts with Type II migmatites always correspond to these high-strain belts.
Garnet–cordierite granulites (Type IV)
Granulites are distinguished by the absence of prograde biotite and alkali feldspar. These rocks are homogeneous and massive at the outcrop scale, but hand specimens reveal a well-developed foliation defined by compositional segregation accompanied by a stretching lineation. Garnet, cordierite, and Fe–Ti oxides are concentrated in poorly defined melanocratic bands that alternate with plagioclase- and quartz-rich leucocratic domains. Plagioclase and garnet in these rocks display noticeable Ca enrichment towards the rims (An45 to An50 and Grs3 to Grs6, respectively). Zoned subhedral crystals of zircon are very abundant in all Type IV granulites but monazite is conspicuously absent. Anthophyllite occurs in variable modal proportions and tends to be associated with melanocratic segregations. Type IV granulites also contain a small amount of Ti-poor phlogopite that is clearly a retrograde phase (Otamendi et al., 1999).
|
PETROLOGIC EVOLUTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
The transition from Type I gneisses to Type II metatexites is marked by the first appearance of K-feldspar-bearing layer-parallel leucosomes that contain ovoid grains of cordierite and by increases in the modal proportions of garnet and ilmenite [Table 1; see also Otamendi et al. (1999)
]. The migmatites are thus inferred to have been produced by anatexis at the amphibolite-to-granulite transition, according to an incongruent melting reaction of the general form
 |
The participation of at least a small amount of free H2O in this melting reaction is inferred on the basis of the fact that Type II metatexites are enriched in K and depleted in Na and Ca relative to Type I gneisses [see below, and also Patiño Douce (1996) and Patiño Douce & Harris (1998)]. Sillimanite is absent from most Type I rocks but is required by this melting reaction, which is constrained by the observed mineral assemblage of Type II migmatites. Otamendi et al. (1999) argued that the sillimanite necessary to stabilize the peritectic assemblage garnet ± cordierite in Type II metatexites was present not as a discrete phase but rather as a dioctahedral Al component in biotite, [Al2R-3]bt [with R = Fe2+ or Mg; see Patiño Douce et al. (1993)], in Type I gneisses. Dioctahedral Al in biotite is liberated with rising temperature and/or decreasing pressure according to the following reaction (see Patiño Douce et al., 1993; Otamendi et al., 1999):
 |
The fact that biotite in the migmatites is strongly depleted in Al6 relative to biotite in the gneisses is consistent with the operation of reaction (2), and the observed growth of euhedral sillimanite and quartz at the interface between resorbed garnet and biotite in Type II metatexites provides clear petrographic evidence for it (see Otamendi et al., 1999, fig. 3f). The feasibility of this melting mechanism [i.e. reaction (2) supplying normative Al2SiO5 to reaction (1)] from the point of view of mass balance is further discussed below, in the context of partial melting models.
The transition from Type II to Type III rocks is characterized by increases in the modal proportions of cordierite and sillimanite (Table 1). Otamendi et al. (1999) attributed these changes solely to a rise in the peak metamorphic temperature from Type II to Type III migmatites that shifted reaction (2) to the right and increased the degree of melting, thus producing more cordierite by reaction (1). In addition to this temperature effect, it is also possible that extraction of a plagioclase-rich melt, formed by melting in the presence of a small amount of free H2O, contributed further to Al enrichment in the residue-rich Type III migmatites (see also below).
Type IV granulites are thought to represent refractory residues formed by extraction of relatively H2O-poor melts at extreme metamorphic temperatures (Otamendi et al., 1999; see also below). In addition to their refractory mineral assemblage, the fact that garnet and plagioclase in the granulites are simultaneously enriched in Ca from core to rim is consistent with these rocks having undergone melt extraction (see Otamendi et al., 1999). Further geochemical arguments that support the residual nature of the granulites are discussed below.
|
BULK-ROCK CHEMICAL COMPOSITIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Sample descriptions and analytical procedures
Bulk-rock chemical analyses (Table 2) were carried out on gneisses, leucogranite bodies (width >2 cm), migmatites and granulites. Samples of Type I gneisses were taken along profiles 1–2 m long oriented perpendicular to the compositional banding, so as to obtain bulk compositions averaged over the length scales characteristic of lithological variability in the gneisses. Migmatite samples were taken so as to be large enough to be representative of the bulk migmatite composition, including mesosome, melanosome and leucosome. This sampling technique was adopted to determine whether the migmatites as a whole are rocks enriched or depleted in melt, or rocks that partially melted and in which the melt underwent only local redistribution and solidified essentially in situ.
Each sample consisted of 5–12 kg of material collected in the field and crushed. After crushing, samples were split to obtain representative fractions of 300 g. Each sample was finally pulverized in a tungsten-carbide mill. The container was always pre-contaminated with the sample about to be pulverized. A 100 mg fraction of powder was digested for 24 h in an HNO3 and HF solution in a Teflon vessel at 120°C. After evaporation to dryness each sample was diluted in 100 ml of 2% HNO3 solution. Gneisses and migmatites were analyzed for major elements by atomic absorption spectrometry and for trace elements by inductively coupled plasma-mass spectrometry (ICP-MS), both at the University of Huelva’s Central Research Services facility (Spain). Major element analyses were duplicated. Average precision and accuracy for most elements is 5–10% relative, determined by repeated analysis of the SARM-1 international rock standard (see also Table 2). Leucogranites were analyzed for major elements and Zr by X-ray fluorescence (XRF) at Activation Laboratories Ltd, Ancaster, Ontario, Canada, and for trace elements by ICP-MS at the University of Huelva.
Major element compositions
Type I gneisses that do not show evidence of anatexis and that we assume to represent the protolith composition for the Sierra de Comechingones anatectic rocks have major element compositions that are within the range of quartz-intermediate greywackes (see Taylor & McLennan, 1985). Relative to the average of quartz-intermediate greywackes (Fig. 4a), Type I gneisses are somewhat depleted in Al2O3, CaO and Na2O, have similar contents of SiO2, and are slightly enriched in K2O, FeOtotal and MgO. Despite these differences, which may reflect lower plagioclase and higher chlorite contents in the protoliths of Type I gneisses compared with average quartz-intermediate greywackes, it is clear that the regionally dominant protolith in the Sierra de Comechingones was an immature clastic sediment.
|
The major element compositions of some Type II metatexites from the central Sierra de Comechingones (samples OCP10 and SCP24; see Fig. 1) are virtually indistinguishable from those of Type I gneisses (Table 2, Figs 4b and 5). These metatexites contain layer-parallel leucosomes and also oblique leucosomes (e.g. Fig. 3c), but are not spatially associated with larger leucogranite bodies. They are found in the central Sierra de Comechingones, several kilometers to tens of kilometers south of the exposure of the highest-grade rocks along the Río Santa Rosa (Fig. 1; see also Otamendi et al., 1999
). These metatexites with unmodified major element compositions are important in two respects. In the first place, they provide chemical evidence supporting the lithological continuity between the regional gneisses and the migmatites. Second, they suggest that some migmatites in the Sierra de Comechingones have undergone little or no melt extraction.
|
In contrast to these rocks, the compositions of other Sierra de Comechingones migmatites differ from those of Type I gneisses. This group of migmatites with modified bulk compositions includes stromatic Type II metatexites associated with leucogranite bodies, Type II diatexites and all Type III migmatites. All of these rocks are enriched in K2O and in major elements contained in ferromagnesian minerals (FeO*, MgO, TiO2 and MnO), and depleted in CaO and Na2O, relative to Type I gneisses (Table 2, Figs 4b and 5). These compositional characteristics argue against a model in which the migmatites are the products of biotite dehydration-melting of the gneisses followed by partial melt extraction, as such processes would leave behind residues enriched in Ca and depleted in K. They also argue against the migmatites being the products of injection of melts formed by dehydration-melting of other lithologies, as this model would not lead to enrichment in ferromagnesian components, but rather the opposite. We argue that the most plausible explanation for these migmatites is that they are the products of partial melting of Type I gneisses (aluminous metagreywackes) in the presence of a small amount of free H2O (H2O-fluxed melting, as opposed to H2O-saturated melting; see Patiño Douce & Harris, 1998), followed by melt extraction. Compared with dehydration-melting, H2O-fluxed melting consumes a greater proportion of plagioclase relative to micas. Extraction of such melts therefore leaves behind residues enriched in biotite components (K2O, FeO, MgO, TiO2) and depleted in plagioclase components (CaO, Na2O) relative to the residues of dehydration-melting (Figs 4b and 5). Owing to the low solubility of excess Al2O3 in H2O-rich, relatively low-temperature melts (e.g. Patiño Douce & Harris, 1998), H2O-fluxed melting is also able to explain the noticeable Al enrichment in the migmatites relative to their putative metagreywacke precursors.
Compared with Type I gneisses, the leucogranites are strongly enriched in K, have similar Na and Al contents, and are depleted in Ca (Figs 4b and 5). The relative contents of these elements are such that the leucogranites are less peraluminous than Type I gneisses, with alumina saturation indices [ASI = Al2O3/(CaO + Na2O + K2O) molar] of 1·12 and 1·2 for leucogranites and gneisses, respectively (see also Table 2). Removal of melts with the ASI of leucogranites is thus consistent with Al enrichment of the residues. It should be noted that the migmatites in which melt extraction appears to have been insignificant (i.e. samples OCP10 and SCP24) are not enriched in Al relative to Type I gneisses (Table 2, Fig. 4b). Because the leucogranites are strongly enriched in K2O and depleted in CaO relative to Type I gneisses, however, they cannot represent unmodified melts formed by H2O-fluxed melting of the gneisses (see Patiño Douce & Harris, 1998). In fact, the very high K contents of the leucogranites (7–8 wt % K2O, Table 2) strongly suggest that they do not represent melt compositions at all [compare experimental melt compositions from metagreywackes (e.g. Patiño Douce & Beard, 1995; Patiño Douce, 1996) and metapelites (e.g. Patiño Douce & Harris, 1998); see also below]. Magmas in the leucogranite bodies may have undergone K-feldspar accumulation and removal of residual melt, for which there is petrographic evidence in the adcumulus texture of K-feldspar megacrysts. An alternative explanation for K enrichment in the leucogranites is the production of peritectic Kfs, formed by melting in a system with high H2O/K2O ratio (e.g. Carrington & Watt, 1994), and subsequent incorporation of some of this Kfs into the extracted magmas. Although we cannot rule out the operation of this process, we believe that this could not have been the dominant process, because both the leucogranites and the associated migmatites are enriched in K relative to the gneisses (Fig. 5).
Type IV granulites are strongly enriched in FeO*, MgO, MnO and TiO2, and to a lesser extent in Al2O3, relative to Type I gneisses (Fig. 4b). CaO is also enriched in the granulites but K2O and Na2O are strongly depleted in them (Fig. 5). The major element compositions of the granulites reflect their refractory and residual mineralogical composition. The strong K depletion, however, shows that the granulites must be the products of a melting process different from that responsible for the formation of the migmatites (see Fig. 5). Biotite dehydration-melting is the most likely explanation for the origin of the granulites.
Trace element abundances
Amphibolite-grade Type I gneisses have rare earth element (REE) contents that are similar to those of post-Archaean sediments [Table 2, Fig. 6a; see also Taylor & McLennan (1985)]. This is in good agreement with our inference that these rocks represent a metamorphosed clastic sedimentary sequence. Type I gneisses show weak negative Eu anomalies [Eu/Eu* 0·7, where Eu* is the expected Eu value estimated as described by Taylor & McLennan (1985)] and absolute abundances of La and Ce around 100x chondrite. In contrast, the heavy REE (HREE) abundances are less uniform, ranging from about 10x chondrite for gneisses in the southern sector of the study area (Río Quillinzo, see Fig. 1) to 15x for the Río Santa Rosa gneisses (Fig. 6a). This difference almost certainly reflects the greater modal proportion of garnet in the Río Santa Rosa gneisses compared with the Río Quillinzo gneisses (Table 1).
|
Relative to Type I gneisses, Type II metatexites with modified major element compositions and all Type III migmatites have slightly lower Eu concentrations (EuN = 16–18 in Type I rocks, 15 in Type II rocks and 14 in Type III rocks, where EuN denotes chondrite-normalized values) and are also depleted in Sr (Figs 6b and c and 7a). These trace element characteristics support H2O-fluxed melting, in which plagioclase is a chief reactant, followed by melt extraction. These rocks are also generally depleted in incompatible elements [large-ion lithophile elements (LILE) and light rare earth elements (LREE)] relative to Type I gneisses (Figs 6b and 7a). In contrast, Type II metatexites with major element compositions similar to those of the gneisses (samples OCP10 and SCP24) are also generally similar to the latter in their trace element abundances (Figs 6b and 7a, Table 2), in agreement with our hypothesis that melt loss from these rocks has been negligible. Type II diatexites have distinct trace element patterns that are not depleted in Eu nor Sr relative to Type I gneisses, and that are enriched in LILE and LREE relative to the latter (Fig. 7a). These characteristics would appear to contradict the interpretation that the diatexites are melt-depleted rocks, yet some degree of melt loss is required by their major element compositions (Figs 4b and 5). The explanation for this discrepancy may reside in the fact that metatexites and diatexites are the products of different melt migration and homogenization length scales.
|
Type IV granulites have very low LREE and LILE concentrations (Figs 6c and 7b), even compared with other granulites that are considered to be residual products of partial melting and large-scale melt extraction. Examples of such granulites (from the Phanerozoic Ivrea Zone, the Late Proterozoic Pan-African rocks from Yaoundé and the Archaean Limpopo Belt) are compared with the Río Santa Rosa granulites in Fig. 7b. The very low abundances of Ba and Rb in the latter rocks reflect the low abundances of host phases for these elements, such as biotite and alkali feldspar. Breakdown of feldspars during melting also liberated Eu and Sr, which were thus depleted in the residual granulitic assemblages (Fig. 7b).
The strong depletion of LREE in the granulites compared with the migmatites (Fig. 6b and c) reflects a fundamental change in the behavior of monazite with rising melting temperature. Monazite is present as unzoned euhedral crystals in Types II and III migmatite and is the dominant LREE reservoir in these rocks. In contrast, monazite is absent from Type I gneisses, in which apatite appears to be the chief LREE reservoir, and from Type IV granulites, in which there is no LREE-rich accessory phase. These petrographic relationships are consistent with crystallization of peritectic monazite (Wolf & London, 1995) during initial H2O-fluxed melting of apatite-bearing metasediments (at T 800°C, Otamendi et al., 1999), followed by dissolution of monazite as melting temperatures rosed above 850°C (Rapp et al., 1987; see also Otamendi et al., 1999).
In contrast to the LREE, the HREE are enriched in the Río Santa Rosa granulites (Fig. 6c) as in other examples of residual granulites (Fig. 7b). This enrichment is consistent with the observed accumulation of residual garnet and zircon in the Río Santa Rosa granulites.
The leucogranites are very rich in Ba and have positive Eu anomalies (Eu/Eu* 6·7, Figs 6d and 7c). In contrast, the trace elements that are partitioned by ferromagnesian minerals (HREE, V, Cr, Sc, Co and Ni) are strongly depleted in these rocks. Zr abundances in the leucogranites (<50 ppm) are much lower than those required to saturate peraluminous granitic melts (e.g. Watson & Harrison, 1983) at the granulite-facies temperatures recorded in the Río Santa Rosa migmatites (800–850°C; see Otamendi et al., 1999). Strong Zr undersaturation in melts can be explained by the fact that almost all zircon in the gneisses and migmatites occurs as inclusions in biotite and, given that biotite tends to stay in the residue during H2O-fluxed melting (Patiño Douce & Harris, 1998), zircon is effectively shielded from equilibrating with the melt. Crystallization of peritectic monazite would have prevented LREE enrichment of the melts, and the HREE would also have been trapped in the residues by peritectic and/or residual garnet. The trace elements in the melts would then chiefly derive from feldspar breakdown, explaining the strong enrichment in Ba and Eu (Watt & Harley, 1993). The leucogranites, however, are undoubtedly cumulate rocks with some amount of trapped interstitial melt, as attested by their textures and by their major element compositions. The cumulate nature of the rocks can also explain all of the trace element characteristics of the leucogranites. The observed trace element abundances in these rocks could thus be the result of two distinct causes that act in concert: the behavior of accessory phases during melting and the accumulation of K-feldspar during crystallization. Identifying the relative contributions of each process may not be feasible.
|
MODELING OF PARTIAL MELTING OF METAGREYWACKES IN THE SIERRA DE COMECHINGONES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Our purpose in this section is to estimate the degree of melting and melt extraction that is represented by the various migmatites and granulites in the Sierra de Comechingones. We model partial melting in the Sierra de Comechingones in two stages, each with a different modeling strategy. This is required because of the differences in our degree of confidence in the identity of the protolith, the melting reaction and the extent of melting represented by each stage. In the first stage we model formation of the migmatites from limited partial melting of aluminous metagreywackes (Type I gneisses). In the second stage we assess whether the granulites represent widespread melting of the same aluminous metagreywackes or of already melt-depleted migmatites.
The modeling focuses on the systematics of three distinct groups of trace elements. The first group consists of Rb, Sr, Ba and Eu. The bulk of these elements is contained in essential mineral phases (feldspars and biotite) which are crucial participants in the melting reactions. Because experiments have shown that equilibrium between these phases and melt is attained in very short times (of the order of days, e.g. Patiño Douce & Beard, 1995) we argue that it is reasonable to treat the distribution of Rb, Sr, Ba and Eu as an equilibrium process (e.g. Harris & Inger, 1992). The second group of elements consists of the HREE, which are chiefly contained in garnet and zircon. The behavior of these elements is likely to be more complex than that of the first group, because garnet is an essential participant of the melting reactions and zircon is not, and also because HREE in garnet have high KD values and low chemical diffusivities (e.g. Hickmott & Shimizu, 1990; Schwandt et al., 1996; Otamendi et al., 2001). As a first approximation, however, we also model distribution of the HREE as an equilibrium process, and then discuss the consequences of this assumption. Finally, the third group consists of the LREE, for which the principal host phases in the Sierra de Comechingones rocks are either apatite (in Type I gneisses) or monazite (in Types II and III migmatites). Modeling of LREE distribution is problematic regardless of whether this is done as an equilibrium or disequilibrium process, chiefly because the LREE are essential structural components of monazite and thus do not follow Henry’s law, and also because estimates of the modal abundances of accessory phases carry large uncertainties. We have modeled the behavior of the LREE assuming equilibrium between melt and essential mineral phases only. This assumption predictably leads to significant discrepancies between measured and modeled LREE abundances. We show, however, that these discrepancies can be explained in a straightforward way on the basis of the observed petrographic characteristics of monazite and apatite and of the reaction relationship between these two phases during melting of aluminous metasediments, which was demonstrated experimentally by Wolf & London (1995).
Formation of the migmatites
We used a generalized mixing model (Le Maitre, 1979) to estimate the stoichiometry of the partial melting reaction on the basis of mass balance of major elements. The procedure solves for the coefficients of the melting reaction by multiple linear least-squares regression. Mineral compositions used for this calculation are shown in Table 3 (data from Otamendi et al., 1999), and were taken from Type I gneisses for reactant minerals (quartz, plagioclase and biotite) and from the migmatites for peritectic minerals (garnet, ilmenite and cordierite). The data were fitted to the incongruent melting reaction
 |
|
Two different melt compositions were used to perform the calculations (Table 4). Model I uses the average of all glass compositions produced by dehydration-melting of a model metagreywacke (SBG) at 7 kbar by Patiño Douce & Beard (1995). Another set of calculations (Model II) was performed using the average glass compositions obtained by Patiño Douce (1996) by melting of the same model metagreywacke as used by Patiño Douce & Beard (1995) at 7–10 kbar but with 1–2 wt % added H2O. The melts produced by Patiño Douce (1996) by H2O-fluxed melting have lower K2O and higher Na2O and CaO contents than the melts produced by Patiño Douce & Beard (1995) by dehydration-melting of the same starting material (Table 3). These differences reflect the increased participation of plagioclase relative to biotite in the H2O-fluxed melting reaction, compared with dehydration-melting.
|
Cordierite in the Río Santa Rosa migmatites and granulites could have formed both as a peritectic phase and as a product of the breakdown of the assemblage garnet + sillimanite + quartz during decompression (see Otamendi et al., 1999). The appearance of cordierite in the rocks cannot be unambiguously related to the melting process and we therefore computed two sets of models (Table 4), one which does not include cordierite (model A), and another in which cordierite is included as a peritectic phase (model B). The combination of cordierite-present and cordierite-absent models with two different melt compositions leads to four sets of stoichiometric coefficients. Table 5 shows the stoichiometric coefficients obtained for each of these models. The predicted melt compositions are compared with the target melt compositions in Table 4.
|
Regardless of the melt composition used in the calculations, models that include cordierite as a peritectic phase yield much better statistical fits than those in which cordierite is absent (Table 4). Among cordierite-bearing models, the sum of the squares of the residuals is 0·11 for model IB [experimental melt compositions from Patiño Douce & Beard (1995), dehydration-melting] and 0·08 for model IIB [experimental melt compositions from Patiño Douce (1996), H2O-fluxed melting]. The difference between the two models may not be statistically significant, and a choice between them is best made on petrological and geochemical grounds. Given that the migmatites are enriched in K2O and depleted in CaO, Na2O, Eu and Sr relative to their putative Type I precursors, we choose the stoichiometry of model IIB (H2O-fluxed melting with production of peritectic cordierite) as representative of the initial melting reaction in the Sierra de Comechingones (Table 5). Our hypothesis is that this reaction was responsible for partial melting of (Type I) aluminous metagreywackes, and that the migmatites (Type II and Type III) are the products of varying degrees of extraction of these melts.
The major element compositions of the leucogranites do not match those of the melts formed by either dehydration-melting or H2O-fluxed melting of metagreywackes (Fig. 8). The leucogranites are thus unlikely to represent melts extracted from the migmatites, but they could be rocks rich in cumulate K-feldspar derived from parental melts formed by H2O-fluxed melting of the migmatites (Fig. 8).
|
The stoichiometric coefficients for the preferred reaction (IIB, Table 5) show that sillimanite must be consumed by the melting reaction, in a proportion approximately half that of biotite. Comparison of the modal proportion of biotite in the purported protolith (Type I gneisses) with that in the migmatites indicates that melting in the transition from gneisses to migmatites consumed 6% biotite. Because sillimanite is very scarce, or is altogether absent, in the gneisses, there must be some mechanism that can supply the 3% of sillimanite that is required by the melting reaction. The net-transfer reaction (see Patiño Douce et al., 1993; Otamendi et al., 1999)
 |
is a mechanism that can supply normative sillimanite. On average, Type I gneisses have 25% modal biotite. The transition from Type I gneisses to Type II migmatites is accompanied by a decrease in the Al2O3 content of biotite of 2·5 wt % (Otamendi et al., 1999). According to reaction (4), this change in biotite composition can supply 2 wt % of normative sillimanite, even if no biotite breaks down. Breakdown of biotite did take place, however (see Table 1), and could have supplied an additional amount of normative sillimanite. We conclude that the proposed cordierite-forming melting reaction (model IIB, Table 5) is possible in aluminous metagreywackes with low or negligible modal proportions of aluminous minerals (muscovite, sillimanite, etc.), and in which the excess Al is contained in biotite.
Equilibrium trace element distributions were modeled using the batch melting equation (Hanson, 1978)
 |
 |
The bulk distribution coefficient (D) is calculated from D = 
iXiKdi, where Kdi is the partition coefficient between mineral i and felsic melt (values given in Table 6) and Xi is the weight fraction of mineral i in the mineral assemblage that coexists with melt at a given value of F. The weight fractions of the residual minerals for any given value of F (shown in Table 7) were computed as follows. The weight percent composition of the initial mineral assemblage was estimated from the mode of Type I rocks and mineral densities taken from Deer et al. (1966). These initial weight fractions of Qtz, Pl, Bt, Grt and Ilm were then varied according to the preferred stoichiometry for the melting reaction (model IIB in Table 5), yielding the values shown in Table 7. We assumed that sillimanite was consumed by melting reaction (3) at the same rate as it was produced by reaction (4), resulting in sillimanite-free residues. Given the very low partition coefficients of most trace elements in sillimanite, the effect of this assumption on the model results is likely to be negligible. The trace element abundances in the source were obtained by averaging the trace element abundances in Type I gneisses (data taken from Table 2).
|
The Río Santa Rosa migmatites contain 15% biotite (Table 1), which corresponds to residues formed by somewhat more than 20% melting (Table 7). Complete consumption of biotite requires >40% melting (Table 7). Trace element abundances in the different varieties of migmatites are compared in Fig. 9a with modeled abundances in residues formed by extraction of 20% and 40% melt. The measured LREE abundances in all migmatite types are significantly higher than the modeled abundances. This discrepancy obviously arises from neglecting monazite in the melting models, but it is consistent with the migmatites being residual rocks in which peritectic monazite accumulated during melt extraction. There is ample petrographic evidence for this process, in the form of abundant euhedral and unzoned monazite crystals. Agreement between measured and modeled abundances is generally good for HREE and LILE, except for Ba (for which abundances are higher in the migmatites than in the model residues, perhaps reflecting a poorly constrained mineral–melt partition coefficient for Ba). Although the resolution of the models is rather poor, they suggest that Type II metatexites associated with leucogranite bodies (sample RSR31b) and Type III migmatites (samples RSR16 and RSR24) may represent residues of 20–40% melt extraction. The model results are also consistent with our qualitative conclusions that these rocks have undergone greater melt extraction than Type II diatexites and than Type II metatexites from the southern end of the study area, which are not associated with leucogranite bodies (sample OCP10). In any event, it is clear that the trace element signatures of the migmatites support the inference derived from major element compositions that many of the Sierra de Comechingones migmatites (both metatexites and diatexites) are melt-depleted residue-rich rocks, rather than melt-dominated rocks.
|
The model trace element abundances in melts generated by 20–40 % melting of Type I gneisses are compared in Fig. 9b with measured abundances in the leucogranites. As with the migmatites, the greatest discrepancy occurs in the LREE, which are strongly depleted in the leucogranites compared with the model results. This discrepancy is consistent both with accumulation of peritectic monazite in the residue of partial melting and with accumulation of K-feldspar in the leucogranite magmas.
Formation of the granulites
The stoichiometry of the melting reaction estimated for the migmatites (Table 5), which represents H2O-fluxed melting, is not necessarily applicable to formation of the granulites. This is so because, at the high temperatures recorded by the granulites (
900°C, Otamendi et al., 1999), dehydration-melting of biotite is likely to have been the overriding melting mechanism, and also because it is not clear whether the protoliths for the granulites were Type I gneisses or already melt-depleted migmatites. In these circumstances, tying to base the melting model on the stoichiometry of a specific melting reaction would be unconstrained. We thus adopted an alternative strategy, which consisted of finding the degree of melting required to minimize the difference between the abundances of LILE and REE calculated in a residual assemblage with the modal composition of Type IV granulites and the abundances actually measured in the granulites.
Trace element abundances were modeled with equations (5) and (6), using the observed modal composition of Type IV granulites to calculate the value of the bulk distribution coefficient, D (mineral partition coefficients were those listed in Table 6). We generated two sets of models using different source materials (Type I gneisses or Type III migmatites) to set the initial concentration values, C0. Equations (5) and (6) were then solved for the value of F that would simultaneously minimize the difference between each of the calculated values of CS and the measured abundance of the same element in Type IV gneisses.
The best fit models were obtained for melt fractions of 50% and using Type III migmatites as the source material (Fig. 10a). The granulites contain a small amount (3%) of Mg-rich biotite which, on the basis of mineral chemistry, we consider to be a retrograde phase [see discussion by Otamendi et al. (1999)]. Inclusion of even this small amount of biotite in the trace element models results in calculated residues with Rb and Ba contents significantly higher than those measured in the granulites. This is consistent with biotite in the granulites being a low-temperature retrograde phase rather than a high-temperature residual phase, and for this reason biotite was excluded from the assemblage used to model trace elements in the granulites.
|
The greatest discrepancy between measured and calculated trace element abundances is in the LREE. These elements are noticeably less abundant in the granulites than in the model calculations (Fig. 10a). This discrepancy reflects complete dissolution of monazite, which is absent from Type IV granulites. Using the monazite solubility model of Montel (1993)
, we estimate that complete dissolution of monazite in the Sierra de Comechingones rocks required temperatures in the range 830–860°C (assuming H2O contents between 1 and 4 wt %). We have estimated peak temperatures of 800°C for the migmatites and 900°C for the granulites (Otamendi et al., 1999). The strong contrast in LREE contents between migmatites and granulites is thus consistent with accumulation of peritectic monazite in the residues during the relatively low-temperature H2O-fluxed melting stage that gave rise to the migmatites, followed by dissolution of monazite during the higher-temperature melting stage that gave rise to the granulites.
The complementary anatectic melts extracted from the granulites must have been rich in Rb, Ba and LREE and poor in Sr and HREE (Fig. 10b). The predicted LREE enrichment does not match the composition of the leucogranites from the study area. However, peraluminous granitoids with trace element signatures that resemble closely those that could be expected of anatectic melts that have left behind residues such as the Río Santa Rosa granulites are known throughout the southern Sierras Pampeanas (Fig. 10b, data after Rapela et al., 1998; J. E. Otamendi, unpublished data, 2000). Many of these granitoids are intruded in metamorphic sequences that are similar to those of the Sierra de Comechingones. These relationships suggest that aluminous greywackes may have been an important crustal source in the generation of peraluminous granitoid magmas during the Early Cambrian Pampean Orogeny.
|
DISCUSSION AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
The geochemical characteristics of the migmatites in the Sierra de Comechingones show that they are residue-rich rocks. Because pelitic lithologies are uncommon in the region, the most likely protoliths for the migmatites are the aluminous metagreywackes (Type I gneisses), which are the only regionally abundant quartzo-feldspathic rocks in the Sierra de Comechingones. It is important to emphasize that, whereas migmatitic lithologies in the Sierra de Comechingones extend over distances of the order of kilometers (e.g. Figs 1 and 2), pelitic intercalations in the sub-solidus Type I gneisses are not recognized at scales of 1 m or less. The generation of residual assemblages that contain aluminous phases (cordierite and sillimanite) from protoliths that contain Al-rich biotite but no muscovite or sillimanite is seen in this context as the result of two contributing factors: the decrease of the Al solubility in biotite with rising temperature (Patiño Douce et al., 1993
; Otamendi et al., 1999), which can lead to saturation of aluminous phases, and the removal of melts that are less peraluminous than their protoliths. Because the solubility of normative corundum in anatectic melts is small (commonly <3 wt % Al2O3, e.g. Patiño Douce & Johnston, 1991; Patiño Douce & Harris, 1998) partial melting of even slightly peraluminous sources can lead to strongly peraluminous residues. This effect may be particularly important during H2O-fluxed melting, as these melts are rich in normative plagioclase (Patiño Douce, 1996; Patiño Douce & Harris, 1998), which has ASI = 1.
Significantly, the migmatites in the Río Santa Rosa area are enriched in K and depleted in Na, Ca, Sr and Eu relative to the metagreywackes. These geochemical characteristics indicate that the migmatites could not have formed by dehydration-melting of the aluminous metagreywackes and subsequent extraction of those melts. Rather, they suggest H2O-fluxed melting, with subordinate participation of biotite relative to plagioclase in the melting reaction (see Patiño Douce & Harris, 1998). This initial melting stage took place at temperatures of the order of 800°C (Otamendi et al., 1999). Breakdown of residual biotite in the migmatites became the dominant melting mechanism at a later stage, when temperature rose to 900°C (Otamendi et al., 1999) and free aqueous fluids were perhaps no longer available. Incongruent dehydration-melting of biotite was thus the dominant process in the generation of the cordierite granulites. The dramatic drop in LREE contents from migmatites to granulites is consistent with destabilization of monazite resulting from the substantial rise in melting temperature that we estimated independently on the basis of phase equilibrium (Otamendi et al., 1999).
Geochemical modeling suggests that the Río Santa Rosa migmatites are the source regions of anatectic granites after extraction of at least 20% melt and perhaps more. Similar conclusions have been reached for other migmatitic terranes (e.g. Sawyer, 1996, 1998; Carson et al., 1997), suggesting that at least in some cases there is a link between migmatites and granitic magmas. Geochemical modeling also suggests that the refractory cordierite granulites from the Río Santa Rosa are the products of extraction of an additional 50% of melt, starting from already melt-depleted migmatites. Assuming 20% melt extraction from gneisses to migmatites followed by 50% melt extraction from migmatites to granulites we can conclude that the granulites are the residues of a total melt extraction of the order of 60%. It is significant that many experiments have shown that this is approximately the maximum amount of granitic melt that can be obtained by melting a wide range of crustal rocks (e.g. Vielzeuf & Holloway, 1988; Patiño Douce & Johnston, 1991; Patiño Douce & Beard, 1995, 1996). The Río Santa Rosa granulites are thus likely to embody the ultimate residue of crustal anatexis, explaining their extreme trace element patterns compared with other residual granulites.
The Río Santa Rosa area illustrates the roles of migmatites and granulites in the origin of anatectic granites. Geochemically depleted granulites are the ultimate residues left by almost complete extraction of anatectic melts. Migmatites (both metatexites and diatexites) are also residual rocks that have undergone partial melt extraction, but that were caught in the process of becoming depleted granulites. They represent intermediate, lower-temperature, steps in the process of differentiation of the crust into a fusible granitic fraction and a refractory aluminous and ferromagnesian fraction.
In the Sierra de Comechingones some Type II metatexites are more strongly depleted in melt than Type II diatexites, even though both have crystallized at comparable P–T conditions (see Otamendi et al., 1999). This emphasizes that the generation of distinct migmatitic fabrics is critically dependent on the relationship between rate of melt generation and rate of melt extraction, which is in turn a function of strain rate (see Sawyer, 1994). Metatexites may thus represent partially molten domains from which melt is actively being removed perhaps as fast as it forms, so that melt proportions large enough to allow a melt-dominated rheology never build up (see also Sawyer, 1994). These partially molten rocks thus retain a metamorphic appearance. Diatexites, in contrast, may represent domains from which melt extraction is more sluggish than in metatexites (perhaps as a consequence of lower strain rates), so that melt fraction builds up, allowing the transition to a melt-dominated rheology to take place. The partially molten rock can then undergo a significant degree of homogenization, losing much of its metamorphic appearance (e.g. Brown, 1994; Sawyer, 1996). The difference between metatexites and diatexites is thus not necessarily one of temperature nor of degree of melting, but is rather determined by how fast melt is removed from the partially molten rock relative to how fast it forms.
A puzzling aspect of the Sierra de Comechingones high-grade terrane is that, although the migmatites and granulites are melt-depleted rocks, the melts themselves cannot be positively identified in the study area. The leucogranites could be cumulate-rich rocks formed from parental melts such as those extracted from the migmatites and granulites, but their volume is too small to account for the amount of ‘missing melt’. The leucogranite dikes and sills could conceivably represent melt pathways through which the partially molten rocks were being drained. The absence of unmodified melt products in the neighborhood of melt depleted rocks is thus an indication of the fact that anatectic melts are capable of migrating over long distances and undergoing fractional crystallization in the process. Whether such melts can give rise to large granitic intrusions in the absence of mass influx from the mantle is still an open question (see Castro et al., 1999; Patiño Douce, 1999).
|
ACKNOWLEDGEMENTS |
|---|
Geochemical data were obtained at the Servicios Centrales de Investigación y Desarrollo of the University of Huelva (Spain). We thank J. de la Rosa for performing, and often repeating, whole-rock geochemistry, and A. Demichelis for helping in the field and for drafting Figs 1 and 2. Financial support was provided by NSF Grants EAR-9316304 and EAR-9725190 to A.E.P.D. and FONCYT Grant PICT99 to J.O. This work was initiated while J.O. was at the University of Georgia supported by a postdoctoral fellowship from CONICET (Argentina), and was completed with the support of a research fellowship from the same agency. Comments from journal reviewers Gordon Watt, John Kaszuba and Fernando Bea, and editorial suggestions from George Bergantz, helped us to greatly improve the paper, and are most appreciated.
|
FOOTNOTES |
|---|
*Corresponding author. Telephone: 001-706-542-2394. Fax: 001-706-542-2425. E-mail: klingon{at}3rdrock.gly.uga.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGROCK TYPES AND FIELD...PETROLOGIC EVOLUTIONBULK-ROCK CHEMICAL COMPOSITIONSMODELING OF PARTIAL MELTING...DISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Arth, J. G. (1976). Behaviour of trace elements during magmatic processes—a summary of theoretical models and their applications. Journal of Reseach of the US Geological Survey 4, 41–47.
Ashworth, J. R. (1985). Introduction. In: Ashworth, J. R. (ed.) Migmatites. Glasgow: Blackie, pp. 1–35.
Barbey, P., Macaudiere, J. & Nzenti, J. P. (1990). High-pressure dehydration melting of metapelites: evidence from the migmatites of Yaoundé, Cameroon. Journal of Petrology 31, 401–427.
Bea, F., Pereira, M. D. & Stroh, A. (1994). Mineral/leucosome trace-element partitioning in a peraluminous migmatite (a laser ablation-ICP-MS study). Chemical Geology 117, 291–312.[Web of Science]
Bonalumi, A. A. & Gigena, A. A. (1987). Relacion entre las metamorfitas de alto grado y las rocas basicas y ultrabasicas en el departamento Calamuchita, Cordoba. Revista de la Asociación Geológica Argentina 42, 73–81.
Brown, G. C. & Fyfe, W. S. (1970). The production of granite melts during ultrametamorphism. Contributions to Mineralogy and Petrology 28, 310–318.
Brown, M. (1994). The generation, segregation, ascent and emplacement of granite magma: the migmatite-to-crustally-derived granite connection in thickened orogens. Earth-Science Reviews 36, 83–130.
Carrington, D. P. & Watt, G. R. (1995). A geochemical and experimental study of the role of K-feldspar during water-undersaturated melting of metapelites. Chemical Geology 122, 59–76.
Carson, C. J., Powell, R., Wilson, C. J. L. & Dirks, P. H. G. M. (1997). Partial melting during tectonic exhumation of a granulite terrane: an example from the Larsemann Hills, East Antarctica. Journal of Metamorphic Geology 15, 105–126.[Web of Science]
Castro, A., Patiño Douce, A. E., Corretgé, L. G., de la Rosa, J., El-Biad, M. & El-Hmidi, H. (1999). Origin of peraluminous granites and granodiorites, Iberian Massif, Spain. An experimental test of granite petrogenesis. Contributions to Mineralogy and Petrology, 135, 255–276.
Clemens, J. D. (1990). The granulite–granite connection. In: Vielzeuf, D. & Vidal, Ph. (eds) Granulites and Crustal Evolution. Dordrecht: Kluwer, pp. 25–36.
Deer, W. A., Howie, R. A. & Zussman, J. (1966). An Introduction to the Rock-forming Minerals. Harlow: Longman, 528 pp.
Gordillo, C. E. (1984). Migmatitas cordieríticas de la Sierras de Córdoba, condiciones físicas de la migmatización. Miscelánea de la Academia Nacional de Ciencias 68, 1–40.
Hanson, G. N. (1978). The application of trace elements to the petrogenesis of igneous rocks of granite composition. Earth and Planetary Science Letters 38, 26–43.[Web of Science]
Harris, N. B. W. & Inger, S. (1992). Trace element modelling of pelite-derived granites. Contributions to Mineralogy and Petrology 110, 46–56.
Hickmott, D. D. & Shimizu, N. (1990). Trace element zoning in garnet from the Kwoiek area, British Columbia: disequilibrium partitioning during garnet growth? Contributions to Mineralogy and Petrology 104, 619–630.
Kretz, R. (1983). Symbols for rock-forming minerals. American Mineralogist 68, 277–279.[Abstract]
Lambert, I. B. & Heier, K. S. (1968). Geochemical investigations of deep-seated rocks in the Australian Shield. Lithos 1, 30–53.
Le Maitre, R. W. (1979). A new generalized petrological mixing model. Contributions to Mineralogy and Petrology 71, 133–137.
Martino, R., Escayola, M. & Saal, A. (1994). Estructura del cuerpo de ‘kinzigita’ del río Santa Rosa, Departamento Calamuchita, Provincia de Córdoba. Revista de la Asociación Geológica Argentina 49, 3–10.
Martino, R., Kraemer, P., Escayola, M., Giambastiani, M. & Arnosio, M. (1995). Transecta de las Sierras Pampeanas de Córdoba a los 32° S. Revista de la Asociación Geológica Argentina 50, 60–77.
Montel, J.-M. (1993). A model for monazite/melt equilibrium and application to the generation of granitic magmas. Chemical Geology 110, 127–146.
Nash, W. P. & Crecraft, H. R. (1985). Partition coefficients for trace elements in silicic magmas. Geochimica et Cosmochimica Acta 49, 2309–2322.[Web of Science]
Otamendi, J. E., Patiño Douce, A. E. & Demichelis, A. H. (1999). Amphibolite to granulite transition in aluminous greywackes from the Sierra de Comechingones, Córdoba, Argentina. Journal of Metamorphic Geology 17, 415–434.
Otamendi, J. E., de la Rosa, J., Patiño Douce, A. E. & Castro, A. (2001). Rare-earth and yttrium zoning in metamorphic and peritectic garnet. Geology submitted.
Patiño Douce, A. E. (1996). Effects of pressure and H2O content on the compositions of primary crustal melts. Transactions of the Royal Society of Edinburgh: Earth Sciences 87, 11–21.[Web of Science]
Patiño Douce, A. E. (1999). What do experiments tell us about the relative contributions of crust and mantle to the origin of granitic magmas? In: Castro, A., Fernandez, C. & Vigneresse, J. L. (eds) Understanding Granites. Integrating New and Classical Techniques. Geological Society, London, Special Publications 158, 55–75.
Patiño Douce, A. E. & Beard, J. S. (1995). Dehydration-melting of biotite gneiss and quartz amphibolite from 3 to 15 kbar. Journal of Petrology 36, 707–738.
Patiño Douce, A. E. & Beard, J. S. (1996). Effects of P, f(O2) and Mg/Fe ratio on dehydration-melting of model metagreywackes. Journal of Petrology 37, 999–1024.
Patiño Douce, A. E. & Harris, N. (1998). Experimental constraints on Himalayan anatexis. Journal of Petrology 39, 689–710.
Patiño Douce, A. E. & Johnston, A. D. (1991). Phase equilibria and melt productivity in the pelitic system: implications for the origin of peraluminous granitoids and aluminous granulites. Contributions to Mineralogy and Petrology 107, 202–218.[Web of Science]
Patiño Douce, A. E., Johnston, A. D. & Rice, J. (1993). Octahedral excess mixing properties in biotite, a working model with applications to geobarometry and geothermometry. American Mineralogist 78, 113–131.[Abstract]
Rapela, C. W., Pankhurst, R. J., Casquet, C., Baldo, E., Saavedra, J., Galindo, C. & Fanning, C. M. (1998). The Pampean Orogeny of the southern proto-Andes: Cambrian continental collision in the Sierras de Córdoba. In: Pankhurst, R. J. & Rapela, C. W. (eds) The Proto-Andean Margin of Gondwana. Geological Society, London, Special Publications 142, 181–217.
Rapp, R. P. Ryerson, F. J. & Miller, C. F. (1987). Experimental evidence bearing on the stability of monazite during crustal anatexis. Geophysical Research Letters 14, 307–310.
Sawyer, E. W. (1994). Melt segregation in the continental crust. Geology 22, 1019–1022.
Sawyer, E. W. (1996). Melt segregation and magma flow in migmatites: implications for the generation of granite magmas. Transactions of the Royal Society of Edinburgh: Earth Sciences 87, 85–94.[Web of Science]
Sawyer, E. W. (1998). Formation and evolution of granite magmas during crustal reworking: the significance of diatexites. Journal of Petrology 39, 1147–1167.
Sawyer, E. (1999). Criteria for the recognition of partial melting. Physics and Chemistry of the Earth (A) 24, 269–279.
Schnetger, B. (1994). Partial melting during the evolution of the amphibolite- to granulite-facies gneisses of the Ivrea Zone, northern Italy. Chemical Geology 113, 71–101.
Schwandt, C. S., Papike, J. J. & Shearer, C. K. (1996). Trace element zoning in pelitic garnet of the Black Hills, South Dakota. American Mineralogist 81, 1195–1207.[Abstract]
Sims, J. P., Ireland, T. R., Camacho, A., Lyons, P., Pieters, P. E., Skirrow, R. G., Stuart-Smith, P. G. & Miró, R. (1998). U–Pb, Th–U and Ar–Ar geochronology from the southern Sierras Pampeanas, Argentina: implications for the Palaeozoic tectonic evolution of the western Gondwana margin. In: Pankhurst, R. J. & Rapela, C. W. (eds) The Proto-Andean Margin of Gondwana. Geological Society, London, Special Publications 142, 259–281.
Taylor, S. R. & McLennan, S. M. (1985). The Continental Crust: its Composition and Evolution. Oxford: Blackwell Scientific, 312 pp.
Taylor, S. R., Rudnick, R. L., McLennan, S. M. & Eriksson, K. A. (1986). Rare earth element patterns in Archean high-grade metasediments and their tectonic significance. Geochimica et Cosmochimica Acta 50, 2267–2279.[Web of Science]
Thompson, A. B. (1982). Dehydration melting of pelitic rocks and the generation of H2O-undersaturated granite liquids. American Journal of Science 282, 1567–1595.
Vielzeuf, D. & Holloway, J. R. (1988). Experimental determination of the fluid-absent melting reactions in the pelitic system: consequences for crustal differentiation. Contributions to Mineralogy and Petrology 98, 257–276.[Web of Science]
Vielzeuf, D., Clemens, J. D., Pin, C. & Moinet, E. (1990). Granites, granulites and crustal differentiation. In: Vielzeuf, D. & Vidal, Ph. (eds) Granulites and Crustal Evolution. Dordrecht: Kluwer, pp. 59–85.
Watson, E. B. & Harrison, T. M. (1983). Zircon saturation revisited: temperature and composition effects in a variety of crustal magma types. Earth and Planetary Science Letters 64, 295–304.[Web of Science]
Watt, G. R. & Harley, S. L. (1993). Accessory phase controls on the geochemistry of crustal melts and restites produced during water-undersaturated partial melting. Contributions to Mineralogy and Petrology 114, 550–566.
Wolf, M. B. & London, D. (1995). Incongruent dissolution of REE- and Sr-rich apatite in peraluminous granitic liquids: differential apatite, monazite, and xenotime solubilities during anatexis. American Mineralogist 80, 765–775.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  N. HARRIS, A. McMILLAN, M. HOLNESS, R. UKEN, M. WATKEYS, N. ROGERS, and A. FALLICK Melt Generation and Fluid Flow in the Thermal Aureole of the Bushveld Complex J. Petrology, June 1, 2003; 44(6): 1031 - 1054. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. E. Johnson, T.E. Johnson, N.F.C. Hudson, and G.T.R. Droop Evidence for a genetic granite-migmatite link in the Dalradian of NE Scotland Journal of the Geological Society, May 1, 2003; 160(3): 447 - 457. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. FORNELLI, G. PICCARRETA, A. DEL MORO, and P. ACQUAFREDDA Multi-stage Melting in the Lower Crust of the Serre (Southern Italy) J. Petrology, December 1, 2002; 43(12): 2191 - 2217. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Otamendi, J. D. de la Rosa, A. E. P. Douce, and A. Castro Rayleigh fractionation of heavy rare earths and yttrium during metamorphic garnet growth Geology, February 1, 2002; 30(2): 159 - 162. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||