The relative contributions of crustal anatexis and mantle-derived magmas in the genesis of synorogenic, Hercynian granites of the Livradois area, French Massif Central.

F. Solgadi(*,1,2), J.-F. Moyen(1,3), O. Vanderhaeghe(1), E. W. Sawyer(1,2) & L. Reisberg(4)

* Corresponding author: Fabien Solgadi Université du Québec à Chicoutimi (UQAC) Sciences de la terre Etudiant 3ème Cycle CHICOUTIMI Québec Canada G7H2B1 Tel : (418) 545 5011 poste 2505 E-mail address: [email protected]

1-

Université Henri Poincaré Nancy 1, Géologie et Gestion des Ressources Minérales et

Energétiques, BP 239 54506, Vandoeuvre-lès-Nancy Cedex, France 2-

Université du Québec a Chicoutimi, Science de la Terre, Département des Sciences

Appliquées, Chicoutimi, Québec G7H2B1, Canada 3-

Departement of Geology, University of Stellenbosch, South Africa.

4-

Laboratoire du C.R.P.G, 15 rue Notre Dame des Pauvres, BP20, 54501 Vandoeuvre-

lès-Nancy, Nancy, France

1

Abstract In the Livradois area (French Massif Central), the two characteristic types of synorogenic Hercynian granitoids (porphyritic monzogranite and two-mica leucogranite) are intruded into a migmatitic paragneiss sequence. A genetic link between the two-mica leucogranite and the migmatitic paragneiss is suggested by the major and trace element whole rock compositions, and by Rb/Sr and Sm/Nd isotopic data. A model of partial melting of the migmatitic paragneiss explains the origin of the twomica leucogranite, especially if the accessory minerals present in paragneiss (zircon, monazite or xenotime) are taken into account. The origin of the porphyritic monzogranite is more difficult to constrain. This pluton belongs to a peculiar high K-Mg suite, rich both in compatible (e.g., MgO) and incompatible (e.g., K2O) elements. The porphyritic monzogranite is heterogeneous and contains microgranular mafic enclaves (MME) that represent a mafic magma. A model of mixing between; (1) a mafic magma with a composition similar to the MME and, (2) a felsic magma similar to the two-mica leucogranite accounts for the composition of the porphyritic monzogranite, both for major and trace elements, and for Rb/Sr and Sm/Nd isotopic ratios. The MME are very enriched in incompatible elements implying an enriched mantle source. Considering the geologic context of the Variscan belt of the French Massif Central, a possible cause for the enrichment of the sub-continental mantle is the contamination by sediments during subduction between 450 and 400 Ma. These results show that the two types of granitoids plutons of the French Massif Central are derived from partial melting of a paragneiss-dominated orogenic wedge mixed to various degree to magmas generated from partial melting of an enriched mantle source. The enriched mantle source is itself contaminated by a crustal-derived component. These magmas are then emplaced within the orogenic wedge during the time period corresponding to the transition from crustal thickening to orogenic gravitational collapse of the Variscan belt between 350 to 290 Ma.

2

Keyword: Granite, magma, anatexis, mixing, high K-Mg magma, mantle, Hercynian, French Massif Central

1. Introduction

Magmatism in zones of plate convergence displays various petrologic and geochemical characteristics that are interpreted to reflect the source of the magma (e.g., mantle versus crustal origins) as well as processes such as partial melting, fractional crystallization and magma mixing (Barbarin, 1990; Barbarin, 1999; Bonin, 1990; Brown, 1994; Chappell & White, 1974; Defant & Drummond, 1990; Drummond et al., 1996; Eby, 1992; Foley & Peccerillo, 1992; Peacock et al., 1994; Pitcher, 1983; Sawyer, 1998; Thompson & Connolly, 1995) Typical magmatic rocks found in zones of active or past plate convergence encompass (1) the calc-alkaline suite interpreted to represent magmatism derived from partial melting of the mantle wedge triggered by the dehydration of the subducting plate; (2) the peraluminous suite associated with anatexis of a thermally relaxed, thickened orogenic wedge; and (3) highK magmatic rocks, exhibiting both mantle and crustal characteristics the significance of which is much debated. The goal of this paper is to present new petrological and geochemical data from the Livradois region in the core of the French Massif Central, where both peraluminous and high-K granitoids are intrusive into a sequence of metasedimentary rocks that have been affected by partial melting to various degrees. These data allow us to constrain (1) the genetic link between peraluminous leucogranite and the host migmatitic paragneisses, (2) the relative contributions of mantle and crustal derived melts in the genesis of a high-K monzogranite, and (3) the impact of various processes such as partial melting, fractional crystallization, and

3

magma mixing on the geochemical evolution of these magmas in the French Massif Central. The geodynamic significance of these results is discussed within the framework of the tectonic evolution of the Hercynian belt.

2. Geological setting 2.1 Geology of the Hercynian belt in the French Massif Central

The French Massif Central exposes a segment of the Hercynian middle crust that is characterized by a large volume of granite (senso lato), migmatites and metamorphic rock (Fig. 1) (Dupraz & Didier, 1988; Duthou et al., 1984; Pin & Peucat, 1986). The Paleozoic tectonic evolution of this region was controlled by convergence between Gondwana and Laurasia accommodated by lithospheric subduction and the tectonic accretion of crustal blocks (Franke, 1989; Matte, 1986; Matte, 1991; Pin & Duthou, 1990). This evolution is recorded, in particular, by the formation of a crustal, orogenic wedge that comprises several nappes. This wedge was affected by metamorphism under different conditions ranging from low to high geothermal gradients, between the times of burial (420-350 Ma) and exhumation during gravitational collapse (350-290 Ma) (Bard & Rambeloson, 1973; Burg et al., 1989; Burg et al., 1984; Costa et al., 1993; Dufour, 1985; Gardien, 1990; Gardien et al., 1997; Ledru et al., 2001; Ledru et al., 1989; Mercier et al., 1991; Montel et al., 1992; Nicollet, 1978; Pin & Peucat, 1986; Vanderhaeghe et al., 1999). Three major lithotectonic units are recognized (Fig. 1) (Ledru et al., 1989). From top to bottom they are: (1) the Upper Gneiss Unit (UGU), (2) the Lower Gneiss Unit (LGU); and (3) the Para-autochtonous Unit. This nappe pile was pervasively melted and converted to migmatite, and intruded by various types of magmatic rocks ranging from a high-K magmatic suite to peraluminous leucogranites in the period between 350 and 290 Ma which corresponds to the transition from crustal thickening to crustal

4

thinning (Downes et al., 1990; Duthou et al., 1984; Montel et al., 1992; Pin & Duthou, 1990; Williamson et al., 1992). Based on their petrologic and geochemical characteristics, the origin of these different types of plutonic rocks is interpreted to include crustal anatexis as well as the intrusion of mantle-derived magmas.

2.2 Geology of the souther Livradois area

The study area is located in the Livradois region, in the core of the French Massif Central (Fig. 1). In this area, the nappe pile is represented by; (1) the ophiolite-bearing Upper Gneiss Unit with at its base, the so-called leptynite-amphibolite group, and (2) the Lower Gneiss Unit (Burg et al., 1984; Ledru et al., 1989). This nappe pile was affected by pervasive partial melting in the sillimanite stability field, but contains relics of a higher pressure metamorphism (Grange et al. in prep.). At the scale of the study area, a sequence of paragneisses shows a gradation from metatexite to diatexite migmatite from south to north (Fig. 1). The distinction between metatexite and diatexite is based on the percentage and the disposition of leucosome (Brown, 1973; Menhert, 1968; Sawyer, 1996; Vanderhaeghe, 2001). These rocks are characterized by a north-dipping, roughly E-W trending penetrative foliation with a northwesttrending mineral and stretching lineation. The composite foliation is underlined by the alternation of leucosome and melanosome layers superimposed on a transposed compositional layering in the metasediments, and by the alignment of enclaves and schlieren in diatexite migmatite. This foliation is concordant with the map-scale metatexite to diatexite transition, although diatexite migmatite locally appears in the cores of kilometer-scale domes and has discordant contacts with the foliation in the metatexite migmatite (Fig. 2). This sequence is intruded by a number of leucocratic granitic veins ranging from centimeter to kilometer in

5

scale. In addition, a large 10 km long and 2-3 km wide sill of porphyritic monzogranite is intruded parallel to the foliation in the diatexite migmatite.

3. Petrologic description of migmatites and granitoids of the southern Livradois area 3.1

Metatexite migmatites

The metatexite migmatites are derived from the paragneiss and characterized by a foliation defined by the alternation of leucosome and melanosome layers. The average bulk modal composition of the metatexite migmatites is: 30% quartz, 20% plagioclase (An20-26), 15% Kfeldspar, 20% biotite, 5% muscovite and 10% aluminous silicate minerals, such as cordierite, sillimanite or almandine-rich garnet. The accessory minerals monazite, xenotime and zircon are ubiquitous and commonly occur as inclusions in biotite. The melanosome is fine-grained (0.5 to 1 mm) and comprises more than 80% of the rock. It is composed mainly of biotite, corderite, sillimanite and garnet. The leucosome is medium-grained (1 to 2 mm) and represents less than 20% of the rock. It occurs principally as small lenses, or layers. The proportion of leucosome in the paragneiss increases progressively near the contacts with the intrusive granitic rocks. The composition of the leucosomes is typically 40% quartz, 30% plagioclase (An0-7), 20% K-feldspar, 8% muscovite, 2% biotite. The most common aluminous minerals in the metatexite migmatites is cordierite, but it contains some small crystals of sillimanite and this suggests the following reaction: Biotite + Quartz + Sillimanite + Plagioclase

→

Cordierite + K-feldspar + Melt

(Hoffer, 1976)

6

This reaction constrains the peak metamorphic temperature to about 800°C and indicates low to moderate pressures (between 2 and 5 kbar). Primary muscovite may have crystallized during post-peak cooling, at temperatures of around 700 °C (Storre & Karotke, 1972) in place of K-feldspar

3.2

Diatexite migmatites

Diatexite migmatites is a medium to fine grained rock (1 to 3mm) with an overall bulk mineral composition of 35% quartz, 25% plagioclase (An20-26), 10% K-feldspar, 15% biotite, 5% muscovite and 10% aluminosilicate minerals principally cordierite and rarely sillimanite. It has a very high proportion of leucosome, between 40 and 60% of the rock, and contains many centimetre-size melanosome lenses, mostly composed of biotite and cordierite, oriented parallel to the general foliation. Enclaves of paragneiss of different sizes (5cm to 1m) are also common in the more leucocratic zones and create a schollen diatexite (Menhert, 1968). The most common aluminous mineral is cordierite, and since the mineral assemblages are very similar to those in the metatexite migmatites, the diatexite records similar metamorphic conditions.

3.3 The Two-mica leucogranite

A kilometre-scale two-mica leucogranite body (Fig. 2a & b) intrudes across both the anatectic rocks (i.e., the migmatites) and an intrusion of porphyritic monzogranite (see next section) (Fig. 2a & b). The leucogranite is fine grained (1 to 2 mm) and contains 35% quartz, 30% plagioclase (An 0-7), 20% K-feldspar, 5% biotite, 10% primary muscovite; accessory minerals (monazite and zircon) are rare. Therefore, the overall composition is a “leucomonzogranite”

7

(sensu Streckeisen, 1976). A weak fabric marked by the preferential orientation of mica flakes is ubiquitous and oriented NE-SW parallel to the contact of the intrusion. This fabric is locally parallel to the foliation in the migmatites, particularly where the leucogranite occurs as dykes. The fabric in the leucogranite was acquired in the magmatic state, as indicated by the weak internal deformation of the minerals (Paterson et al., 1989). The plagioclase crystals in the leucogranite have particularly low An contents, suggesting that the leucogranite represents a differentiated, or evolved magma; the An end-member was partitioned into the first-formed plagioclase and is, therefore, principally concentrated in the early cumulates (Bowen, 1913).

3.4 Porphyritic monzogranite

The porphyritic monzogranite forms a large sill, (8 x 4) km wide (Fig.2). It is a coarse-grained rock (2-3 mm) with 1-3 cm poikiolitic, K-feldspar megacrysts that have inclusions of biotite, quartz and plagioclase. The bulk rock contains 30% quartz, 30% plagioclase (Ca. An30), 30% K-feldspar (mostly as megacrysts), and 10% biotite. Accessory minerals e.g., zircon, apatite and allanite are common. In its central part, the porphyritic monzogranite displays a weak fabric and a weak internal deformation of the minerals that is inferred to be magmatic in origin (Paterson et al., 1989). This fabric is marked by the preferred orientation of K-feldspar megacrysts and by biotite flakes. Quartz-filled fractures are present in some plagioclase crystals (Fig. 3a) and indicate that deformation of the solid crystal-framework was synchronous with crystallization, down to low fractions of melt, and that the residual liquids were able to percolate through the crystal framework (Bouchez et al., 1992). Locally, stretched and recrystallized quartz ribbons are developed and indicate some degree of subsolidus deformation of the porphyritic monzogranite (Paterson et al., 1989; Vernon et al., 1983). In contrast, the southern border of the intrusion is mylonitic indicating intensive sub-

8

solidus deformation (Fig. 2). Altogether, these features point to a syn-tectonic emplacement for this pluton. The porphyritic monzogranite is heterogeneous at the outcrop scale (Fig. 4a), and comprises; (1) foliation-parallel, meter-sized layers rich in K-feldspar crystals; (2) zones concordant to the foliation that are devoid of megacrysts and, (3) veins of biotite-bearing leucogranite (20-50 cm wide), that have diffuse borders discordant to the foliation of the porphyritic monzogranite. Collectively these features, specifically the megacryst-free layers, are thought to indicate the segregation of liquid from magma during crystallization (Vigneresse & Tikoff, 1999). The veins represent the liquid extracted from the regions now marked by the K-feldspar accumulations. This accumulation of K-feldspar parallel to the magmatic foliation can occur during magmatic flow by crowding of the crystals (Vernon, 1986). The porphyritic monzogranite contains layered (mm-scale) xenoliths of muscovite-rich paragneiss. At magmatic temperatures, other aluminous phases e.g., cordierite, sillimanite should be stable, as observed in the paragneiss, and their absences suggests that they have been destabilized to muscovite. Myrmekite is developed at the edges of the paragneiss xenoliths (Fig. 3b) and indicates chemical disequilibrium between the paragneiss and the host monzogranite magma. The average K2O content of the porphyritic monzogranite is 4,6 wt% which contrasts with the average K2O of 3,5 wt% in the paragneiss. The following reactions may occur at the edges of the xenoliths and lead to the formation of myrmekite:

KAlSi3O8 + Na = NaAlSi3O8 + K, Orthoclase

albite

in combination with:

2KAlSi3O8 + Ca = CaAl2Si2O8 + 4SiO2+ 2K Orthoclase

anorthite

quartz

(Phillips, 1974)

Two types of plagioclase are observed in the porphyritic monzogranite:

9

1) Plagioclase with a simple zoning pattern (Fig. 5a): The An content in these plagioclases gradually decreases from An30 at the core, to An20 at the rim. The normal plagioclase zoning records the evolution of the magma composition during fractional crystallization (Wiebe, 1968). The plagioclase grew in equilibrium with an enclosing melt that became progressively more sodic during the course of crystallization (Singer & Pearce, 1993; Wiebe, 1968). 2) Plagioclase with a complex zonation (Fig. 5b): These plagioclase crystals have composite cores with some irregularly-shaped regions that are very rich in An component (An40), and some zones with lower An content (An30).

Plagioclases with such complex cores are

interpreted as indicators of hybrid magmas (Castro, 2001; Tsuchiyama, 1985);

First, a

plagioclase with a moderate Ca content (An30) crystallizes in a felsic magma, then the intrusion of a high-temperature, more mafic magma causes resorption of the first plagioclase and new, Ca-rich (An40), plagioclase crystallizes around the resorbed plagioclase, thus forming the complex core structures.

During subsequent fractional crystallization of the host

composite magma, the complexly zoned plagioclase develops a Ca-poor rim (Fig. 6). Complex core zoning records mechanical mingling and mixing between felsic and mafic magmas (Sparks & Marshall, 1986; Vernon, 1984).

3.5 Microgranular Mafic Enclaves

Microgranular Mafic Enclaves (MME) are common in the porphyritic monzogranite (Barbarin, 1988; Didier & Barbarin, 1991). They are 5-20 cm long (Fig. 4b), elongated in the foliation, consistent with deformation under magmatic conditions (Wiebe & Collins, 1998). Typically, they are fine-grained (0.2 to 1 mm) and consist of 15% quartz, 30% plagioclase (An30), 15% K-feldspar and 40% biotite, and very rarely amphibole. Accessory minerals (zircon and apatite) are common. The grain size of biotite in the MME varies from < 1 mm in

10

the core of an enclave, to > 1 mm at the rim of an enclave (Fig. 3c) which is comparable to the size of biotite in the porphyritic monzogranite. This suggests that biotites crystallized in the monzogranite host have been mechanically assembled around the enclave during cooling of the magma, thus forming the dark rim (Fig. 3c). Large (2 mm), crystals of allanite occur in the rim zone of MME and are stabilized by the high Ca content and the low peraluminosity of the MME (Cuney & Friedrich, 1987). The stability of the allanite in the MME is commonly interpreted as evidence for hybridation (Dini et al., 2004). Rounded feldspar phenocrysts (Fig. 3d), which are macroscopically similar (size, compositional zonation and poikiolitic nature) to K-feldpar megacrysts in the porphyritic monzogranite, are found in the MME. Commonly, this type of relationship is interpreted to be the result of chemical destabilization, and subsequent corrosion, of the feldspar megacryst crystallized from the felsic porphyritic monzogranite, following its incorporation into the mafic magma from which the enclave crystallized (Barbarin, 1988; Hibbard, 1981; Pin, 1990; Pin et al., 1990; Pin & Duthou, 1990).

4. Geochemistry and origin of the paragneiss and leucogranite 4.1. Major elements

The migmatitic paragneiss is clearly peraluminous, with A/CNK ratios between 1.30 and 2.45 supporting a metasedimentary origin, but at the same time it is rich in Fe2O3 ( ~6 wt% ). The two-mica leucogranite is also clearly peraluminus with A/CNK ratio between 1.2 and 1.5, but it contains < 2 wt% Fe2O3 (Table 1). Harker diagrams (Fig. 7) for the anatectic metasedimentary rocks (metatexite migmatite, diatexite migmatite and enclaves of paragneiss) and the leucogranites show a good alignment for all samples; SiO2 is negatively correlated with Al2O3, and Fe2O3T, suggesting one of two different processes:

11

1) Initial variation in the protolith metasediment composition, between a clay dominated, pelitic (Al and Fe rich, Si poor) end-member, and a quartz+feldspar-dominated, greywacke end-member. The observed trends would then just reflect variations in the mixing ratio of the two end-members sediment compositions. 2) Variable accumulation of residual aluminous phases such as biotite and cordierite in the solid product left by partial melting. The second possibility is more likely for the following reason. The samples on Figure 7 that are relatively enriched in Al2O3 and Fe2O3 are from paragneiss enclaves collected from the diatexite migmatites which probably represents a hotter environment than the metatexite migmatite. The proportion of biotite is greater in the paragneiss enclaves than in the metatexite migmatite (10% more) suggesting that the enclaves were affected by partial melting and that biotite is a phase in excess. Harker diagrams (Fig. 7c, d)) show rough positive correlations of CaO and Na2O with SiO2. Plagioclase is the principal CaO and Na2O bearing mineral in these rocks. This positive correlation can be explained by varying proportions of leucosome composed principally of quartz and plagioclase in the rocks. This observation is consistent with a process of partial melting and segregation of the liquid, because the two-mica leucogranite contains more plagioclase than K-feldspar, like the leucosomes in the metatexite migmatite. In addition, the two-mica leucogranite has a composition very close to the experimental glass composition obtained by melting of a metapelite or greywacke (compilation of data from Gardien et al., (1995); Patino Douce & Harris (1998); Patino Douce & Johnston (1991).

12

4.2. Trace elements

The migmatitic paragneisses have relatively high trace element concentrations compared to the two-mica leucogranite (Fig. 8), e.g., La between 48 and 77 ppm for migmatitic paraneisses and between 2.3 and 26 ppm for two-mica leucogranite. Table 1 shows also a dispertion in trace element concentration for the two-mica leucogranite e.g., La between 2.3 and 26 ppm; Ce between 4.2 and 60 ppm. 4.3. Rb/Sr and Sm/Nd isotopes

Four samples of leucogranite and three of the migmatitic paragneiss (two of metatextite and one of diatexite) were chosen for the determination of their Rb/Sr and Sm/Nd isotopic ratios. The analytical procedures used are described in the appendix A; and results are given in Table 2. The goal was to constrain the origin of the two-mica leucogranite, and to test its potential genetic link to the migmatitic paragneiss. The Sr and Nd isotope compositions were recalculated to 300 My; the probable age of the granite emplacement in this region (Ledru et al., 1989). The Sr and Nd isotopic ratios were similar for both the anatectic rock and the twomica leucogranite at that time. Therefore, two-mica leucogranite has isotopic ratios that are consistent with a derivation, at 300 Ma, by partial melting of the paragneiss (Fig. 9). 4.4. Geological modeling for the petrogenesis of the two-mica leucogranite

The two-mica leucogranite can be modeled as partial melts of pelitic metasediment, using the equilibrium partial melting equation (Allegre & Minster, 1978; Shaw, 1970)

Cl = C0/ (Drs+f*(1-Drs)

(E1)

Where: 13

Cl is the concentration of an element in the liquid C0 is the concentration of this element in the parent rock F is the fraction of liquid formed, and DRS is the bulk partition coefficient for the element in the residual solid (a function of mineral proportions in the residue, and partition coefficients). Partition coefficients used in this modeling are from the compilation by Rollinson (1993).

A simple model with a biotite-rich residuum, and modal proportions similar to those observed in the residual paragneiss enclaves in the diatexite migmatite (Fig.10a) does not account for the trace element composition of the leucogranites. In particular, the predicted REE contents for the modeled liquid are considerably too high compared to the measured values in the twomica leucogranite. REE contents in the migmatitic paragneiss are higher than in the leucogranite: this implies that the REE behaved as compatible elements during melting. The REE have low partition coefficients in major phases and this implies that phases with high partition coefficients (zircon, monazite, xenotime or apatite) were stable in the residue. Such a situation is common in leucogranites, or S-type granites (Brouand, 1989; Montel, 1993). Adding 1% of zircon to the model restite causes a dramatic increase in the bulk D values for REE from e.g., 0,4 to 5,6 for DYb, producing a melt composition in good agreement with one of the leucogranites. (Fig.10 b). One percent of zircon is far too much compared to the amount of zircon present in the paragneiss enclaves, but it is taken to represent the effect of all the other accessory phases such as monazite and/or xenotime, for which the partition coefficients are even larger than those of

zircon (Bea et al., 1994). The large dispersion in REE

composition of the two mica leucogranite can, therefore, be related to accumulation, or melting, of various amounts of zircon, or monazite, during melting (Montel, 1993).

14

5. Geochemistry and origin of the porphyritic monzogranite and associated MME 5.1. Major elements The porphyritic monzogranite is weakly peraluminous, with A/CNK ratios between 1.06 and 1.07, but at the same time it is rich in MgO (~2 wt%) and Fe2O3 ( ~4 wt% ) and has high K2O contents (~4,5 wt%) similar to those of the two-mica leucogranite (Table 1). Thus the porphyritic monzogranite falls within the group of plutonic rocks described as high K-Mg by Bogaerts et al. (2003) and Laporte et al. (1991). The MME, although more mafic than the porphyritic monzogranite, also display a similar high K-Mg signature (A/CNK =1.01 to 1.04, ~3,5 wt% MgO ,~7 wt% Fe2O3 ,~3wt% K2O). The origin of high K-Mg magmas is not well understood but some authors think this type of magma has a mixed origin (Barbarin, 1999; Ferre et al., 1998; Laporte et al., 1991) While the high compatible element contents do point to a mantle-derived origin, at least in part, the high K2O and related element contents suggest, either a very peculiar (i.e., strongly enriched) mantle, and/or a crustal influence (i.e., contamination). Figures 11 a and b show two binary diagrams for major elements. The data for the porphyritic monzogranite lie between the two-mica leucogranite and the MME. As described above, the petrographic features of the porphyritic monzogranite support a hybrid origin, and the geochemical data suggest that the leucogranite and the MME are the two components of this mixture. The mixing lines traced on Fig. 11 give an idea of the relative percentage of the mafic magma required to mix with a leucocratic two-mica magma to produce the porphyritic monzogranite. The analyses for the two MME differ. We used the more extreme composition for the mixing line because we think that this enclave corresponds to a more primitive (less contaminated) mafic magma.

15

5.2 Trace elements

The porphyritic monzogranite and the MME it contains are both rich in compatible elements (e.g., Ni ~30 and ~100 ppm; Cr ~100 and ~500 ppm respectively) and in incompatible (LIL) elements (e.g., Rb ~200 ppm for monzogranite and MME) (see Table 1). Typically, they also have rather high HFSE contents (e.g., Nb between 13 and 25 ppm and Y ~20 ppm). The REE patterns for the porphyritic monzogranite and its MME are similar, but MME are more enriched. They are LREE enriched (La ~50 ppm) and moderately fractionated (La/Yb ~20), and display a moderate negative Eu anomaly. Figure 11c shows a binary diagram for trace elements (Nb vs. Rb). As for major elements the analyses for the porphyritic monzogranite, the two-mica leucogranite and the MME from an array aligned which also supports a mixing process.

5.3 Rb/Sr and Sm/Nd isotopic data

The porphyritic monzogranite and the MME have very similar age-corrected (300 Ma) Sr and Nd isotopic ratios. This is because the leucogranite component had relatively low Sr and Nd concentrations, and thus had little effect on the Sr and Nd isotopic compositions of the mixture.The isotopic compositions of these rocks are intermediate between those observed for the migmatitic paragneiss and the two-mica leucogranite, and those of the mantle array (Fig. 9.). This observation precludes derivation solely from a local crustal source and strongly implicates the mantle in their creation.

16

5.4. Geochemical modeling for the petrogenesis of the porphyritic monzogranite

The geochemical evidence suggests that the porphyritic monzogranite represents a hybrid magma, produced by mixing between a component derived from a potentially enriched mantle source, and an anatectic melt derived from crustal rocks, such as the paragneisses of the area. This conclusion is in good agreement with both the field evidences (porphyritic body intrusive into a partially molten crust), and the petrological data (gain size, complex zonation in plagioclase). Therefore, its petrogenesis can be described as a two-steps model; (1) formation of a mafic magma and, (2) mixing of this magma with the anatectic melts already in the crust. This model will now be tested using geochemical modeling.

The most likely felsic end-member of the mixing in the model is taken to be a magma similar to the nearby two-mica leucogranites. Although it is possible that the MME represent hybrids between a “true” mafic magma and the host granite, we use the composition of the MME to model the mafic magma end-member.

Magma mixing can be modeled by the simple mass balance equation: Cm = γ Ca + (1 - γ) Cb (E2) Where: Cm = concentration of an element in the mix (taken to be the porphyritic monzogranite ) Ca = concentration of an element in one end-member of the mixing (taken to be the MME) Cb = concentration of the other end-member (the two mica leucogranite) γ = mass fraction of the mafic end-member in the mixture .

17

Following the “mixing test” of Fourcade & Allegre (1981), we rewrite this equation (E2) as Cm-Cb = γ (Ca-Cb) (E3)

In a (Ca - Cb) vs (Cm - Cb) diagram, if the rock “m” is indeed a mixture of end members “a” and “b”, then representative points for all the major elements should plot along a straight line passing through the origin, the slope of which corresponds to γ.

In our case, a mixture of 70% MME and 30% leucogranite (Fig. 12a) fits the measured major element compositions. This mixture is also tested for the REE (Fig. 12b), and yields a modeled REE pattern very close to that obtained from the porphyritic monzogranite (corresponding major and trace element compositions are in Table 1). Small differences in Eu could be related to feldspar separation, evidence of which is seen at the outcrops scale. Feldspar accumulation is also suggested by the misfit for Al2O3 in the mixing diagram Fig. 14a.

6. Discussion 6.1. Types of granites in the Livradois area

Two coexisting magmas were present in the Livradois region at ca. 300 Ma: (1) The leucogranite suite, corresponding to the two mica leucogranite intrusion. This sort of magma is genetically linked to the nearby diatexitic paragneisses ; (2) The high K-Mg suite corresponding to porphyritic monzogranite, which we demonstrated was derived from mixing between a mafic component represented by the MME, and anatectic melts similar to the leucogranites.

18

The origin of both the leucogranite and the mafic component of the mixing; will now be discussed. 6.2. Crustal anatexis and formation of the leucogranitic melts

The two-mica leucogranite is, of sedimentary origin because it is very peraluminous. Such a conclusion is consistent with most experimental work on leucogranites (Patino Douce & Johnston, 1991; Vielzeuf & Holloway, 1988), which invariably proposes such genetic link. The modelling shows that the partial melting of the paragneiss found in the study area is a possible origin of the leucogranite. The interesting thing is that the accessory minerals play an important role in the geochemical composition of the magma. The fact that accessory minerals are only a very minor component of the leucogranite indicates that the solid-liquid separation was very effective, leaving the accessory phases with the solid fraction. Two possibilities may explain why these accessory minerals remained in the residue during partial melting: (1) The temperature during partial melting was relatively low and the accessory minerals were not affected by the partial melting. For example Montel (1993) show an exponential solubility of monazite in the melt as a function of temperature. Low temperature can also have an impact on the degree of partial melting. With a low degreee of melting the accessory minerals are more easily trapped in the restite, the solid part acts as a filter to trap them. (2) The second possibility is that the accessory minerals are included in other minerals and are inaccessible during partial melting (Watson et al., 1989). Indeed, in the migmatitic paragneisses, the accessory minerals are commonly included in biotite. Biotite is abundant in the paragneiss (20%) and so not much was likely consumed in the partial melting reaction. It is difficult to choose between these two hypothesis. In addition, these processes are not

19

mutually exclusive and both could contribute to the very low content of incompatible elements in the leucogranite.

6.2. Origin of the mafic components in the porphyritic monzogranite

Trace elements

Given the mafic nature of the MME and the implied mafic component of the mixing, these magmas must originate from a mafic lithology. The source of the MME and the mafic component in the porphyritic monzogranite could potentially be either the mantle, or the lower crust. The Hercynian lower crust is known from xenoliths in nearby Cenozoic volcanoes, such as the Bournac pipe, 80 km to the SE of our study area (Dostal et al., 1980; Downes et al., 1990). It consists of unknown proportions of mafic granulites (47 wt% < SiO2 < 54 wt%), felsic granulites, and metasedimentary rocks. None of these rocks are mafic enough to represent the source of the low SiO2 mafic component in the porphyritic monzogranite. Furthermore, (Dostal et al., 1980) found that the Hercynian lower crust is significantly depleted in incompatible elements, LILE, REE & HFSE. Accordingly, the low-SiO2 mafic nature and the incompatible element enrichment in MME are not consistent with the known characteristics of the Variscan lower crust. A spidergram (Fig. 13) for the incompatible elements shows the chemical similarity between the MME, and Hercynian lamprophyre dykes, of about the same age from the Thiers area, 50 km north, and the Cévennes area, 120 km south in the Massif Central (Agranier, 2001). While the exact origin of lamprophyres is still debated, there is little doubt that they are mantle-derived melts, owing to their low SiO2 values (50-55 wt%). Such a degree of enrichment for MME and lamprophyres (100 to 1000 times chondritic, i.e. 10-100 times MORB (Hart & Zindler, 1986), implies an extremely enriched mantle source.

20

Modeling of mantle melting is inconclusive, owing to the large number of possibilities for the mantle source composition. To circumvent this problem we use the inverse approach and recalculated the composition that a mantle source required to yield a mafic partial melt composition consistent with that of the MME. The bulk melting equation Cl=C0/(Drs+f(1Drs) (E1) from Allegre & Minster (1978) and Shaw (1970) used the same conventions as used for the leucogranite model, is reformulated as C0=Cl * (Drs+f(1-Drs) (E4), to facilitate calculation of the mantle source given in Fig. 16. For reasonable values of f, in the range 0.01 to 0.2, it is noticeable that this hypothetical mantle is; 1) very enriched, from 5 to 30 times chondritic and, 2) already displays prominent positive anomalies in Th (and U), and negative anomalies in K, Sr and Ti. Furthermore, the nature of the Al-rich phase present in such a mantle, garnet or spinel, is not very important (see Fig. 14a), except, to a limited extent, for the HFSE and HREE. While this model is calculated using equilibrium batch melting, the choice of equation used is not particulary critical. Using other equations (e.g., Rollison, 1993; Allègre & Minster, 1978) for different types of fractional melting will affect the resulting compositions to some degree; but the differences between the different models are minor (Rollinson, 1993), except perhaps for very low melt fractions – and in all instance, these differences are small enough to have no effect on the interpretations discussed below. The modelled composition of the source mantle can be compared with actual mantle samples (Figs. 14b to f). First, the Massif Central peridotites that occur as enclaves in Cenozoic basalts (Lenoir et al., 2000) show a depletion in most incompatible elements (Fig. 14b) that is indicated by a concave trace element pattern; they are also one order of magnitude too depleted in most trace elements to be an adequate source for the MME. This is not a surprise, because it is known from seismic tomography that Cenozoic volcanism is linked to the rise of the asthenospheric (or even lower) mantle (e.g. Goes et al., 1999; Granet et al., 1995) therefore, this mantle does not correspond to the sub-continental mantle that existed in late

21

Hercynian times. Orogenic peridotites from the Ronda massif (Frey et al., 1985), and Fig. 16c are even more depleted than the Massif Central xenoliths, and cannot represent the expected mantle source either. Maury et al. (1992) have described nodules of metasomatized peridotite found as enclave in calk-alkali lavas from the Philippines, which they consider to represent the sub-arc metasomatized mantle; such a tectonic setting could be a realistic source of mantle enrichment in a convergent setting. However, Fig. 14d shows that these rocks are hardly a realistic source either, as they reach only the lower envelope of possible mantle compositions. In fact, the only peridotites described in the literature that do have the adequate degree of enrichment are from enclaves of old sub-continental lithosphere, e.g. from kimberlite (Fig. 16e and references on the figure). The implication of this observation is that a possible mantle source is an old, enriched sub-cratonic mantle. This, however, seems an unlikely hypothesis, since evidence for a pre-existing craton, with its associated enriched mantle in Western Europe is presently lacking. Therefore, the only likely hypothesis is a composite source, with both mantle and enriched, crustal material present in the source, or at least involved in the early evolution of the magmas (Fig. 14). As an illustration, the composition of a mantle into which 1-20 % of sediments has been added was computed. This kind of source has been proposed for the compositionally similar vaugneritic magmas that are widespread in the Massif Central (Michon, 1987; Montel & Weisbrod, 1986). The composition of the sediments used is the average post-Archaean shale of Taylor & McLennan (1985), corresponding to average fine-grained silicate detritic sediments, and not very far from an average upper crust composition. For this model the initial trace elements composition of the mantle is irrelevant, since the sediments are one or two orders of magnitude enriched in incompatible trace elements over the mantle, thus the sediment component completely controls the trace elements composition of the mixture. As

22

can be seen Fig. 14-f, the composition of the mixture has degrees of enrichment comparable with the expected mantle source; furthermore, the shapes of the trace element pattern are also consistent, with the observed positive spikes in Th and negative spikes in K, Sr and Nd, together with a “plateau” for Nb to Ce and Nd to Sm. To summarize, the most likely source for the mafic magmas seems to be a composite one, combining an ultramafic (mantle) and a crustal component. The geological conditions allowing such mixing to occur will discussed below.

Rb/Sr and Sm/Nd isotopic data Isotopes provide an independent constraint on this discussion. The highly enriched signature shown by the porphyritic monzogranite and its MME can be interpreted in one of the following ways: (1) The isotopic signature is purely related to the mixing in the migmatitic amphibolite-facies crust between the leucogranitic magmas related to in-situ melting (see above), and the mafic magma. In this case, the MME would not display a pristine source-related signature, but would already have a hybrid signature between the true mafic magma, and the anatectic melts. Figure 15 curve (b) shows that mixing of the leucogranitic component with a mafic magma with a CHUR-like isotopic composition can indeed account for the signature of the porphyritic monzogranite and its MME, with around 50 – 60 % of the crustal component; this is in good agreement with the mixing model presented above. In this scenario, the MME would have later undergone complete isotopic reequilibration with their host, homogenizing both their signatures. Yet, we regard this as unlikely, on both geological and geochemical grounds. First, more primitive rocks are not found in our study area. Furthermore, even the mafic rocks emplaced in a shear zone within the paragneisses, south of the porphyritic body, are similar to the MMEs. Finally, the lamprophyres, which are probably nearly devoid of any

23

upper crustal influence, also display enriched (although to a lesser degree, cf. Fig. 15) isotopic characteristics (Agranier, 2001; Turpin et al., 1988) together with a similar trace elements signature. On geochemical grounds, it should be noted that the mixing we described is between mafic melts, and leucogranitic melts; we see no evidence for direct mixing of mafic melts with sediments. The leucogranitic melts, because of the residual phases present during partial melting of their paragneissic source, are actually relatively depleted in most incompatible trace elements (eg., La), and consequently cannot account for the enriched nature of the mafic melts. For instance, the leucogranitic melts have La contents of 20 ppm, while the MMEs have 65 ppm of La. Consequentely, if the MMEs themselves where the products of mixing between an unknown magma, and the leucogranitic melts, this unknown magma should be even more enriched. Therefore, we think that the enriched character of the rocks is related to the source, rather than to interactions with the migmatitic crust. (2) The isotopic signature is composite, and related to the interactions between a CHUR-like or depleted mantle, and the lower crust, in granulite facies (during magma ascent). Samples from the granulitic lower crust have been brought to the surface by Cenozoic volcanoes, at Bournac for instance (Dostal et al., 1980), about 100 km east of our study locality. Bournac granulites, however, have a composition which does not allow them to be the source of the enrichment of the mantle component; as can be seen in Fig. 16, they are far too depleted, and have incorrect signatures (e.g., La/Yb and Ba/Sr ratios), to be an adequate source component. More generally, granulites are commonly depelted in trace elements (Blein et al., 2003; Dessai et al., 2004) relative to the whole crust, and it is therefore unlikely that an enriched lower crust was present. The low incompatible element contents of the nearby Variscan lower crust (Dostal et al., 1980) implies that assimilation of lower crust will not significantly alter the trace element composition of the mafic magma; but the very radiogenic signature of the lower crust (Downes et al., 1990) could produce an isotopically enriched magma. Quantitative

24

modeling (Fig. 15, curve a) indicates that this process can indeed yield the correct isotopic signature, but assuming assimilation of very large amounts of lower crustal (60 – 80 %) material: it is unlikely that such a high degree of assimilation would have no effect on the petrography and geochemistry of the mafic magmas. Therefore, interactions with the Variscan lower crust are unable to solely account for the isotopic and trace elements signature of the magma (although its influence cannot be completely ruled out), and such interactions cannot be the principal source of enrichment of the mafic magmas. (3) The mantle source had an enriched signature to begin with. This would imply, for instance, an old subcontinental mantle with high Rb/Sr and low Sm/Nd that evolved in a closed system for a long period of time. This hypothesis is not incompatible with conclusions drawn from trace elements in the MME. It would imply the existence of an old lithosphere preserved beneath the continent involved in the Hercynian collision. The problem, however, is that old lithospheric components (Proterozoic) are not really known in the Hercynian belt (Liew & Hofmann, 1988), although omnipresent mid-Proterozoic Nd model ages (Downes & Duthou, 1988; Downes et al., 1997; Pin et al., 1990; Pin & Duthou, 1990; Simien et al., 1999; Turpin et al., 1990) suggest that such a component might well have existed. (4) Finally, the mantle source itself could be a mixture of components with contrasting origins, as proposed earlier from the trace element data. Again, as an illustration, the isotopic composition (at 300 Ma) of a CHUR-like mantle in which 1-20 % of sediments were mixed at 420 Ma, the age of the subduction in the Massif Central, has been plotted on Fig. 15 (curve and field c ). This process is able to produce a very radiogenic source with isotopic signatures compatible with the present-day signature of the MME. Altogether, the isotopic evidence alone is rather equivocal, as already mentioned (Agranier, 2001; Pin et al., 1990; Turpin et al., 1988). An old lithospheric component, mantle contaminated with sediments, the lower crust and the upper crust, can all have influenced the isotopic signature and it is difficult to identify

25

their relative contributions. However, the combination of both trace elements and isotopic informations suggest that the most likely explanation is a composite, enriched source, made of a mantle mixed with some sort of crustal material. 6.3. Nature of the high K-Mg signature

High K-Mg granitoids are a volumetrically minor, but relatively frequent, component of orogenic and post-orogenic domains (Barbarin, 1999; Laporte et al., 1991; Liegeois et al., 1998). Therefore, it is important to discuss the origin of this sort of granitoid, and the nature of its geochemical signature. In the light of this study, it appears that the high K-Mg signature is, at least partially, related to source characteristics. Subsequent interactions with either the lower crust, or the migmatitic middle crust, cannot be ruled out, but they do not suffice to explain the geochemical characteristics of the rocks – in particular, their enrichment in incompatible elements. Interactions with the migmatitic crust do demonstrably occur; but they do not generate this signature. These interactions, if anything, make the resulting magmas less enriched; the anatectic melts are poorer in most incompatible trace elements than the mafic high K-Mg melts. High K-Mg granitoids can, therefore, be related to an enriched mantle source. This is actually consistent with experimental studies on the genesis of potassic and ultra-potassic magmas from the mantle (Conceicao & Green, 2004; Niida & Green, 1999) showing that a phlogopite and/or amphibole-bearing lherzolite is a possible source for these melts. The cause of the enrichment is actually poorly known. In different geological settings, and for different members of the high K-Mg suite, it has been ascribed to slab-released fluids (Wyllie, 1987) percolation of mafic magmas (Bodinier et al., 1990; VanderWal & Bodinier, 1996), or metasomatism by felsic melts (Martin et al., 2005; Moyen et al., 2001). In this study, we show that materials with a crustal signature are likely contaminants for the mantle. However,

26

they could correspond to a wide range of components such as sediments, elements of subducted felsic basement, melts generated during continental subduction. (1) Mixing sediments into the mantle is geologically possible; for instance, in a subduction zone, a variety of crustal material is recycled back into the mantle, with the subducting plate. In the Massif Central, subduction occured at ca. 400 Ma (Ducrot et al., 1983; Lardeaux et al., 2001; Paquette et al., 1995; Pin & Lancelot, 1982; Pin & Peucat, 1986), providing ample opportunities for adding sediment to the mantle, and forming an enriched mantle wedge near, or above, the subducting slab, directly through mechanical mixing of the sediments with the peridotites. Furthermore, this suggestion is consistent with conclusions of other work on ultrapotassic, syn-convergent magmatism (Bachinski & Scott, 1980; Carr & Fardy, 1984; Michon, 1987; Montel & Weisbrod, 1986; Rock, 1984; Rogers et al., 1982; Venturelli et al., 1984). (2) With similar processes, it is also possible to mechanically inject felsic ortho-derived “basement” into the mantle. Continental subduction of felsic fragments is now well documented in orogenic wedges, e.g., in the alpine belt (Chopin, 1984; Chopin et al., 1991; Reinecke, 1991; Schertl et al., 1991), and has actually been suggested for the Hercynian belt itself (Lardeaux et al., 2001). (3) Finally, a slightly more speculative case can be made for a third possibility; the role of felsic melts derived from partial melting of the subducted continental material. Indeed, subduction of continental crust, as discussed above, might bury such rocks to conditions where they are able to melt; therefore, some felsic magmas could be generated at great depth, and below the mantle wedge, in collisional settings. Ascent of such magmas would allow their reaction with the overlying mantle wedge (Rapp et al., 1999) and, therefore, make possible the formation of an enriched mantle; this mantle, having been enriched via materials which are effectively of crustal nature, would be a possible source for the magnesio-potassic series(Moyen et al., 2001)

27

6.4. Granite genesis in a partially molten orogen

During the late-orogenic event, the Hercynian crust of the Massif Central underwent a thermal regime that allowed pervasive melting of the fertile lithologies of the lower gneiss unit. Indeed, on a regional scale, the LGU is common consists of metatexite migmatite and locally domes of diatexite migmatite of LGU material (e.g., the Velay dome, just south of our field area) form antiformal structures rising through the upper gneissic unit. “S-type” granites, formed via partial melting of this paragneiss material, occur commonly throughout the Massif Central, generally as intrusive bodies cutting across the migmatites (Margeride, Guéret), but also in the cores of domal structures (Velay) ; they correspond to regions of the local collection and mobilization of the anatectic melts, extracted from the LGU migmatites. At the same time, mafic melts formed in the mantle also intruded the partially molten crust. They probably intruded preferentially along weak zones in the crust, such as active shear zones (Brown & Solar 1998). Immediately south of our field area, similar mafic melts that intruded a shear zone are indeed observed. More generally, occurrences of dioritic material with high K-Mg affinity are common in the Massif Central, for instance the “vaugnérites”, common in all of the south-eastern Massif Central, or the late lamprophyres observed throughout the area. In some cases e.g., in the Velay dome – (Montel, 1988; Montel & Weisbrod, 1986) or in the Margeride pluton (Couturie & Caenvachette, 1980), vaugnérites are demonstrably co-magmatic with anatectic, S-type granites or diatexite complexes. When such melts reach a magma-dominated zone in the anatectic crust, they are able to interact with the S-type melts there. The addition of the mafic melts also brings heat to the crust, allowing an increase in the degree of partial melting, which locally increase the fraction of melt in the rock. Therefore, such melt-rich zones become places where major interactions, mixing and mingling between the coexisting two types of melts occur e.g., Castro et al.,

28

(2003) resulting in the formation of high K-Mg magmas. We suspect that, depending on the composition of the partially molten region at the site of the magma intrusion, a large variety of high K-Mg granitoids can be generated in this way. This could account for some of the diversity within this group. High K-Mg granitoids are observed throughout the Massif Central, but are not very common. The most typical of the high K-Mg plutons occur in the Bois Noirs and Saint-Julien-la-Vêtre plutons, in the north (Forez Fig. 1) (Barbarin, 1988) in the Livradois (studied here), in parts of the Lozère massif and in the Aigoual-Saint-Guiral pluton in the south (Cevenes Fig. 1). Only in the Livradois are they observed to intrude into the migmatitic UGU and LGU directly; elsewhere, they typically intrude low grade units found in the uppermost or lowermost levels of the metamorphic pile (Cévennes schists, Forez métavolcanics). Therefore, the Livradois area provides a window on the region where these magmas formed, allowing a better constraint of their origin. 6.5. Mantle implication during orogenic collapse of the Hercynian belt

An interesting implication of the above discussion is that, during the late-orogenic events in the Massif Central area, melting affected not only the crust, but also the whole of the lithosphere. Indeed, while melting in the crust generated the S-type and similar magmas, melting of an enriched lithospheric mantle generated the Mg-K mafic magmas, some of which would subsequently interact with the anatectic melts in the crust and generate Mg-K granitoids. In other terms, the late-orogenic melting (and melt-assisted collapse of the orogen) was actually a lithosphere-scale process, and was not restricted to just the continental crust. The influx of the mafic magmas carried some heat into the crust, amplifying the melting there. Evidence for such a heat supply is provided by P-T loops during late orogenic evolution of the

29

belt (Gardien et al., 1997), which show not only an isothermal decompression, but also heating and a high-temperature “excursion” during the decompression. What triggered melting of the enriched lithospheric mantle remains unclear. Obviously, this enriched mantle was very fertile and melting it did not require important amounts of heat. In similar situations e.g.,Western Himalaya, (Maheo et al., 2002), Mg-K granitoids and mafic rocks have been related to slab-breakoff and delamination of the mantle part of the orogenic stack. It is possible that, in the Hercynian belt too, the transition between collision and collapse was triggered by such a process. Slab-breakoff would not only remove the cold lithospheric root under the orogenic crust, but would also generate mafic magmas in the mantle (Nelson, 1992). In the hot, already partially molten, orogenic crust, this relatively minor heating from both sources (removal of thermal barrier, and influx of mafic magmas) might well have been sufficient to induce massive melting on an orogenic scale, therefore generating migmatites in most of the crust, and allowing its gravitational collapse, together with the generation of massive amounts of granite – mostly crustal melt, but also, locally, hybrid melts involving the mantle-derived component.

7. Conclusion

Two types of granitoids are present in the study area; a two mica leucogranite, and a high KMg porphyritic monzogranite. Petrological study and geochemical modelling suggest that the former corresponds to partial melting and melt extraction from the nearby migmatitic paragneisses, while the latter are ascribed to mixing between such magmas, and mantle derived mafic melts, related to the vaugnérites, or lamprophyres, found in the Massif Central area. These mafic magmas were derived from patial melting of a very enriched mantle,

30

probably resulting from the mechanical emplacement of metasediments during a previous subduction event. Therefore, it appears that the late-orogenic stages in the Livradois area resulted in the formation and interaction of two types of magmas: melts derived from crustal anatexis, and magmas formed in the enriched mantle. The late-orogenic period appears to be a time of important anatexis that affected the lithosphere as a whole.

Appendix A. Analytical techniques The chemical whole rock analyses were performed at the SARM (Service d'Analyse des Roches et des Minéraux) CNRS-CRPG laboratory, Vandoeuvre (France), by ICPAES for major elements and ICP-MS for trace elements. For major elements the precision is estimated to be better than 2% for values higher than 5 wt %. For trace elements the precision is better than 15% in the range of 1 to 10 ppm For Rb-Sr and Sm-Nd isotopic analyses, approximately 100 mg of sample powder were spiked with isotopic tracers and dissolved in a concentrated HF-HNO3-HClO4 acid mixture heated (110°C) for at least 72 hours in a closed teflon beaker. Rb and Sr were separated by conventional cation exchange techniques using AGX 50W resin with 2.5 N HCl acid. After rinsing with 2.9 N HNO3, REE were extracted from the same cation exchange columns with 4.4 N HNO3. Nd and Sm were subsequently isolated from the other REE using HDEHPcoated teflon resin columns with 0.27 N and 0.5 N HCl, respectively. Rb isotopic compositions were determined by the SARM (Service d'Analyse des Roches et des Minéraux) using an ELAN 6000 ICPMS. Sr, Nd and Sm isotopic compositions were determined in static multicollection mode using a Finnigan MAT 262 mass spectrometer at CNRS-CRPG in Nancy. Concentrations were calculated by isotope dilution. Sr samples were loaded on single W filaments, while Sm and Nd were loaded on Ta filaments and ionized

31

using a Re filament. Measured

87

isotope ratios were normalized to

Sr/86Sr ratios were normalized to 146

86

Sr/88Sr = 0.1194. Nd

Nd/144Nd = 0.7219. During the period of analysis, our

value for the La Jolla Nd standard was 0.511799 ± 0.000022 (2σ), while the value for the NBS987 Sr standard was 0.710166 ± 0.000048 (2σ). These values are significantly different from the generally accepted values for these standards (La Jolla: 0.511854; NBS987: 0.71024), most probably because of aging of the Faraday cups. Nevertheless, determinations of other standards during the same time period showed equivalent shifts, demonstrating that the shift is insensitive to isotopic composition. All the results presented in Table II have been corrected by the amounts needed to bring the La Jolla and NBS987 standards into agreement with the internationally accepted values for these standards. Total chemical blanks were < 2 ng for Sr and ~ 0.4 ng for Nd, and thus have no significant effect on the isotopic compositions. Initial ratios and εNd have been calculated at 300 My (estimated age from geological relationships). The maximum age uncertainty remains limited (± 20My) and has only a minor affect on the calculated initial ratios (εNd changes by a maximum of 0.3 units). Small but systematic differences are observed between the Sm concentrations obtained by ICPMS (after fusion with LiBO4) and those obtained by isotope dilution after acid digestion, especially for the leucogranites. This probably reflects a failure to completely dissolve zircon during acid digestion. Nevertheless, in most cases the Sm/Nd ratios of the two methods agree within 10 %, and the difference is only 24 % in the worst case (sample 8). In any event, assuming the samples were in isotopic equilibrium at 300 My, this should have no effect on the calculated initial ratios.

Acknowledgment This work was partially funded by the BRGM, and was integrate in the project of 1/50 000 geological mapping of France. The authors are thankful to P. Rossi for this support. We wish

32

to thank C. Zimmermann and C. Parmentier of the laboratory of isotopic analysis of the CRPG for their help. This work is part of the M.Sc project of F. Solgadi. References:

AGRANIER, A. (2001): Un exemple de magmatisme ultrapotassique syn-orogénique : les lamprophyres du Massif Central Français. Approche géochimique et géochronologique. In D.E.A. (Master thesis), Université B.Pascal, Clermont-Ferrand, France., pp. 51pp. ALLEGRE, C. J. & MINSTER, J. F. (1978): Quantitative models of trace-element behavior in magmatic processes. Earth Planet. Sci. Lett. 38, 1-25. BACHINSKI, S. W. & SCOTT, R. B. (1980): Rare-earth and other trace-element contents and the origin of minettes - a critical comment - Reply. 44, 1389-1392. BARBARIN, B. (1988): Field evidence for successive mixing and mingling between the piolard diorite and the Saint-Julien-La-Vetre monzogranite (Nord-Forez, Massif Central, France). Can. J. Earth Sci. 25, 49-59. ________. (1990): Granitoids - main petrogenetic classifications in relation to origin and tectonic setting. Geol. J. 25, 227-238. ________. (1999): A review of the relationships between granitoid types, their origins and their geodynamic environments. Lithos 46, 605-626. BARD, J. P. & RAMBELOSON, R. (1973): Métamorphisme plurifacial et sens de variation du degré géothermique durant la tectonogenèse polyphasée hercynienne dans la partie orientale de la zone axiale de la Montagne Noire (Massif du Caroux, Sud du Massif Central français). Bull. Soc. Geol. Fr. 15, 579-586.

33

BEA, F., PEREIRA, M. D. & STROH, A. (1994): Mineral leucosome trace-element partitioning in a peraluminous migmatite (a Laser Ablation-Icp-Ms Study). Chem. Geol. 117, 291312. BLEIN, O., LAFLECHE, M. R. & CORRIVEAU, L. (2003): Geochemistry of the granulitic Bondy gneiss complex: a 1.4 Ga arc in the Central Metasedimentary Belt, Grenville Province, Canada. Precamb. Res. 120, 193-217. BODINIER, J. L., VASSEUR, G., VERNIERES, J., DUPUY, C. & FABRIES, J. (1990): Mechanisms of mantle metasomatism - geochemical evidence from the Lherz orogenic Peridotite. J. Petrol. 31, 597-628. BOGAERTS, M., SCAILLET, B., LIEGEOIS, J. P. & AUWERA, J. V. (2003): Petrology and geochemistry of the lyngdal granodiorite (Southern Norway) and the role of fractional crystallisation in the genesis of proterozoic ferro-potassic A-type granites. Precamb. Res. 124, 149-184. BONIN, B. (1990): From orogenic to anorogenic settings - evolution of granitoid suites after a major orogenesis. Geol. J. 25, 261-270. BOUCHEZ, J. L., DELAS, C., GLEIZES, G., NEDELEC, A. & CUNEY, M. (1992): Submagmatic microfractures in granites. Geology 20, 35-38. BOWEN, N. L. (1913): The melting phenomena of the plagioclase feldspars. Am. J. Sci. 35, 577-599. BROUAND, M. (1989): Pétrogenèse des migmatites de la dalle du Tibet (Himalaya du Népal). Thèse de diplôme, Institut National Polytechnique de Lorraine, Nancy, France BROWN, M. (1973): The definition of metatexis, diatexis and migmatite. Proc. Geol. Assoc. 84, 371-382.

34

________. (1994): The generation, segregation, ascent and emplacement of granite magma the migmatite-to-crustally-derived-granite connection in thickened orogens. Earth-Sci. Rev. 36, 83-130. BROWN, M. & SOLAR, G. S. (1998): Shear-zone systems and melts: feedback relations and self-organization in orogenic belts. J. Struct. Geol. 20, 211-227. BURG, J.-P., DELOR, C. P., LEYRELOUP, A. & ROMNEY, F. (1989): Inverted metamorphic zonation and Variscan thrust tectonics in the Rouergue area (Massif Central, France): P-T-t record from mineral to regional scale. In Evolution of Metamorphic Belts, Geological Society Special Publication,, Vol. 43 (ed. R. A. C. J. S. Daly, and B. W. D. Yardley, eds.,) 423-439. BURG, J.-P., LEYRELOUP, A., MARCHAND, J. & MATTE, P. (1984): Inverted metamorphic zonation and large-scale thrusting in the Variscan Belt: an example in the French Massif Central. Exhumation processes normal faulting, ductile flow and erosion Geological Society Special Publications 14, 47-61. CARR, P. F. & FARDY, J. J. (1984): Ree geochemistry of late Permian shoshonitic lavas from the Sydney Basin, New-South-Wales, Australia. Chem. Geol. 43, 187-201. CASTRO, A. (2001): Plagioclase morphologies in assimilation experiments. Implications for disequilibrium melting in the generation of granodiorite rocks. Mineral. Petrol. 71, 3149. CASTRO, A., CORRETGE, L. G., DE LA ROSA, J. D., FERNANDEZ, C., LOPEZ, S., GARCIAMORENO, O. & CHACON, H. (2003): The appinite-migmatite complex of Sanabria, NW Iberian massif, Spain. J. Petrol. 44, 1309-1344. CHAPPELL, B. W. & WHITE, A. J. R. (1974): Two contrasting granite types. Pac. Geol. 8, 173174.

35

CHOPIN, C. (1984): Coesite and pure pyrope in high-grade blueschists of the Western Alps - a first record and some consequences. Contrib. Mineral. Petrol. 86, 107-118. CHOPIN, C., HENRY, C. & MICHARD, A. (1991): Geology and petrology of the coesite-bearing terrain, Dora-Maira Massif, Western Alps. European Journal of Mineralogy 3, 263291. CONCEICAO, R. V. & GREEN, D. H. (2004): Derivation of potassic (shoshonitic) magmas by decompression melting of phlogopite plus pargasite lherzolite. Lithos 72, 209-229. COSTA, S., MALUSKI, H. & LARDEAUX, J. M. (1993): 40Ar-39Ar chronology of Variscan tectono-metamorphic events in an exhumed crustal nappe: the Monts du Lyonnais complex (Massif Central, France). Chem. Geol 105, 339-359. COUTURIE, J. P. & CAENVACHETTE, M. (1980): Westphalien age of leucogranites intruding the Margeride Granite (French Massif Central). C. R. Acad. Sci. Ser. D 291, 43-45. CUNEY, M. & FRIEDRICH, M. (1987): Physicochemical and crystal-chemical controls on accessory mineral paragenesis in granitoids - implications for uranium metallogenesis. Bull. Mineral. 110, 235-247. DEFANT, M. J. & DRUMMOND, M. S. (1990): Derivation of some modern arc magmas by melting of young subducted lithosphere. Nature 347, 662-665. DESSAI, A. G., MARKWICK, A., VASELLI, O. & DOWNES, H. (2004): Granulite and pyroxenite xenoliths from the Deccan Trap: insight into the nature and composition of the lower lithosphere beneath cratonic India. Lithos 78, 263-290. DIDIER, J. & BARBARIN, B. (1991): Enclaves and Granite Petrology, Elsevier edition. Elsevier, Amsterdam. DINI, A., ROCCHI, S. & WESTERMAN, D. S. (2004): Reaction microtextures of REE-Y-Th-U accessory minerals in the Monte Capanne pluton (Elba Island, Italy): a possible indicator of hybridization processes. Lithos 78, 101-118.

36

DOSTAL, J., DUPUY, C. & LEYRELOUP, A. (1980): Geochemistry and petrology of metaigneous granulitic xenoliths in Neogene volcanic-rocks of the Massif Central, France Implications for the Lower Crust. Earth Planet. Sci. Lett. 50, 31-40. DOWNES, H., DUPUY, C. & LEYRELOUP, A. F. (1990): Crustal evolution of the Hercynian Belt of Western-Europe - evidence from lower-crustal granulitic xenoliths (French MassifCentral). Chem. Geol. 83, 209-231. DOWNES, H. & DUTHOU, J. L. (1988): Isotopic and trace-element arguments for the lowercrustal origin of Hercynian granitoids and Pre-Hercynian orthogneisses, Massif Central (France). Chem. Geol. 68, 291-308. DOWNES, H., SHAW, A., WILLIAMSON, B. J. & THIRLWALL, M. F. (1997): Sr, Nd and Pb isotopic evidence for the lower crustal origin of Hercynian granodiorites and monzogranites, Massif Central, France (vol 136, pg 99, 1997). Chem. Geol. 140, 289289. DRUMMOND, M. S., DEFANT, M. J. & KEPEZHINSKAS, P. K. (1996): Petrogenesis of slabderived trondhjemite-tonalite-dacite/adakite magmas. Trans. R. Soc. Edinb.-Earth Sci. 87, 205-215. DUCROT, J., LANCELOT, J. R. & MARCHAND, J. (1983): U-Pb Dating on zircons on La-Borie eclogite (Haut-Allier, France) and consequences on the ante-Hercynian evolution of occidental Europe. Earth Planet. Sci. Lett. 62, 385-394. DUFOUR, E. (1985): Granulite facies metamorphism and retrogressive evolution of the Monts du Lyonnais metabasites (Massif Central, France). Lithos 18, 97-113. DUPRAZ, J. & DIDIER, J. (1988): Le complexe anatectique du Velay (Massif Central francais): structure d'ensemble et évolution géologique:. Geol. Fr., 73-88.

37

DUTHOU, J. L., CANTAGREL, J. M., DIDIER, J. & VIALETTE, Y. (1984): Paleozoic granitoids from the French Massif Central - age and origin studied by Rb-87-Sr-87 system. Phys. Earth Planet. Inter. 35, 131-144. EBY, G. N. (1992): Chemical subdivision of the A-type granitoids - petrogenetic and tectonic implications. Geology 20, 641-644. ERLANK, A. J., WATERS, F. G., HAWKESWORTH, C. J., HAGGERTY, S. E., ALLSOPP, H. L., RICKARD, R. S. & MENZIES, M. A. (1987): Evidence for mantle metasomatism in peridotite nodules from the Kimberley Pipes, South Africa. In Mantle metasomatism (ed. M. a. Hawkesworth), pp. pp. 221-311., London. FERRE, E. C., CABY, R., PEUCAT, J. J., CAPDEVILA, R. & MONIE, P. (1998): Pan-African, post-collisional, ferro-potassic granite and quartz-monzonite plutons of Eastern Nigeria. Lithos 45, 255-279. FOLEY, S. & PECCERILLO, A. (1992): Potassic and ultrapotassic magmas and their origin. Lithos 28, 181-185. FOURCADE, S. & ALLEGRE, C. J. (1981): Trace-elements behavior in granite genesis - a casestudy the calc-alkaline plutonic association from the Querigut Complex (Pyrenees, France). Contrib. Mineral. Petrol. 76, 177-195. FRANKE, W. (1989): Variscan plate-tectonics in central-Europe - current ideas and open questions. Tectonophys. 169, 221-228. FREY, F. A., SUEN, C. J. & STOCKMAN, H. W. (1985): The Ronda high-temperature peridotite - geochemistry and petrogenesis. Geochim. Cosmochim. Acta 49, 2469-2491. GARDIEN, V. (1990): Evolutions P-T et structures associées dans l'est du Massif Central Français : Un exemple de l'évolution thermomécanique de la chaîne paléozoique, Thèse de diplôme, Université Grenoble I, Grenoble , France.

38

GARDIEN, V., LARDEAUX, J. M., LEDRU, P., ALLEMAND, P. & GUILLOT, S. (1997): Metamorphism during late orogenic extension: Insights from the French Variscan belt. Bull. Soc. Geol. Fr. 168, 271-286. GARDIEN, V., THOMPSON, A. B., GRUJIC, D. & ULMER, P. (1995): Experimental melting of biotite + plagioclase + quartz + or - muscovite assemblages and implications for crustal melting. J. Geophys. Res. B Solid Earth Planets 100, 15,581-15,591. GOES, S., SPAKMAN, W. & BIJWAARD, H. (1999): A lower mantle source for central European volcanism. Science 286, 1928-1931. GRANET, M., WILSON, M. & ACHAUER, U. (1995): Imaging a mantle plume beneath the French Massif Central. Earth Planet. Sci. Lett. 136, 281-296. HART, S. R. & ZINDLER, A. (1986): In search of a bulk-earth composition. Chem. Geol. 57, 247-267. HIBBARD, M. J. (1981): The magma mixing origin of mantled feldspars. Contrib. Mineral. Petrol. 76, 158-170. HOFFER, E. (1976): Reaction Sillimanite + Biotite + Quartz reversible Cordierite + KFeldspar + H-2o and Partial Melting in System K-2O-FeO-MgO-Al-2O-3-Si0-2-H-20. Contrib. Mineral Petrol. 55, 127-130. LAPORTE, D., ORSINI, J. B. & FERNANDEZ, A. (1991): Le complexe d'Ile-Rousse, Balagne, Corse du Nord-Ouest: pétrologie et cadre de mise en place des granitoïdes magnésiopotassiques. Geol. Fr. 4, 15-30. LARDEAUX, J. M., LEDRU, P., DANIEL, I. & DUCHENE, S. (2001): The Variscan French Massif Central - a new addition to the ultrahigh pressure metamorphic 'club': exhumation processes and geodynamic consequences. Tectonophys. 332, 143-167. LEDRU, P., COURRIOUX, G., DALLAIN, C., LARDEAUX, J. M., MONTEL, J. M., VANDERHAEGHE, O. & VITEL, G. (2001): The Velay dome (French Massif Central):

39

melt generation and granite emplacement during orogenic evolution. Tectonophys. 342, 207-237. LEDRU, P., LARDEAUX, J. M., SANTALLIER, D., AUTRAN, A., QUENARDEL, J. M., FLOCH, J. P., LEROUGE, G., MAILLET, N., MARCHAND, J. & PLOQUIN, A. (1989): Où sont les nappes dans le Massif Central Francais? Bull. Soc. Geol. Fr. 5, 605-618. LENOIR, X., GARRIDO, C. J., BODINIER, J. L. & DAUTRIA, J. M. (2000): Contrasting lithospheric mantle domains beneath the Massif Central (France) revealed by geochemistry of peridotite xenoliths. Earth Planet. Sci. Lett. 181, 359-375. LIEGEOIS, J. P., NAVEZ, J., HERTOGEN, J. & BLACK, R. (1998): Contrasting origin of postcollisional high-K calc-alkaline and shoshonitic versus alkaline and peralkaline granitoids. The use of sliding normalization. Lithos 45, 1-28. LIEW, T. C. & HOFMANN, A. W. (1988): Precambrian crustal components, plutonic associations, plate environment of the Hercynian fold belt of central-Europe indications from a Nd and Sr isotopic study. Contrib. Mineral. Petrol. 98, 129-138. MAHEO, G., GUILLOT, S., BLICHERT-TOFT, J., ROLLAND, Y. & PECHER, A. (2002): A slab breakoff model for the Neogene thermal evolution of South Karakorum and South Tibet. Earth Planet. Sci. Lett. 195, 45-58. MARTIN, H., SMITHIES, R. H., RAPP, R., MOYEN, J. F. & CHAMPION, D. (2005): An overview of adakite, tonalite-trondhjemite-granodiorite (TTG), and sanukitoid: relationships and some implications for crustal evolution. Lithos 79, 1-24. MATTE, P. (1986): Tectonics and plate-tectonics model for the Variscan belt of Europe. Tectonophys. 126, 329-374. ________. (1991): Accretionary history and crustal evolution of the Variscan belt in WesternEurope. Tectonophys. 196, 309-337.

40

MAURY, R. C., DEFANT, M. J. & JORON, J. L. (1992): Metasomatism of the sub-arc mantle inferred from trace-elements in philippine xenoliths. Nature 360, 661-663. MENHERT, K. R. (1968): Migmatites and the origin of granitic rocks. Elsevier Publ. Co., Amsterdam - Oxford - New York. MENZIES, M. A., ROGERS, N. W., TINDLE, A. & HAWKESWORTH, C. J. (1987): Metasomatic and enrichment processes in lithospheric peridotites, an effect of asthenospherelithosphere interaction. In Mantle metasomatism (ed. M. a. Hawkesworth), Acad. Press, London, United Kingdom. 313-361. MERCIER, L., VAN ROERMUND, H. L. M. & LARDEAUX, J. M. (1991): Comparison of P-T-t paths in allochthonous high-pressure metamorphic terrains from the Scandinavian Caledonides and the French Massif-Central: contrasted thermal structures during uplift. Geol. Rundsch. 80, 333-348. MICHON, G. (1987): Les Vaugnerites de l'Est du Massif Central Français : apport de l'analyse multivariée à l'étude géochimique des éléments majeurs. Bull. Soc. Geol. Fr. 3, 591600. MONTEL, J. M. (1988): First discovery of an orthopyroxene-bearing Vaugnerite - petrography, geochemistry, and implications on the genesis of Vaugnerites. C. R. Acad. Sci. Ser 2 306, 985-990. ________. (1993): A model for monazite/melt equilibrium and application to the generation of granitic magmas. Chem. Geol. 110, 127-146. MONTEL, J. M., MARIGNAC, C., BARBEY, P. & PICHAVANT, M. (1992): Thermobarometry and granite genesis - the Hercynian low-P, high-T Velay anatectic dome (French MassifCentral). J. Metamorph. Geol. 10, 1-15.

41

MONTEL, J. M. & WEISBROD, A. (1986): Characteristics and evolution of Vaugneritic magmas - an analytical and experimental approach, on the example of the Cevennes Medianes (French Massif-Central). Bull. Mineral. 109, 575-587. MOYEN, J. F., MARTIN, H. & JAYANANDA, M. (2001): Multi-element geochemical modelling of crust-mantle interactions during late-Archaean crustal growth: the Closepet granite (South India). Precamb. Res. 112, 87-105. NELSON, K. D. (1992): Are crustal thickness variations in Old Mountain Belts like the Appalachians a consequence of lithospheric delamination. Geology 20, 498-502. NICOLLET, C. (1978): Pétrologie et tectonique des terrains cristallins anté-permiens du versant du dôme du Lévézou (Rouergue, Massif Central). Geol. Fr. I, 225-263. NIIDA, K. & GREEN, D. H. (1999): Stability and chemical composition of pargasitic amphibole in MORB pyrolite under upper mantle conditions. Contrib. Mineral. Petrol. 135, 18-40. NIXON, P. H., ROGERS, N. W., GIBSON, I. L. & GREY, A. (1981): Depleted and fertile mantle xenoliths from Southern African Kimberlites. Annu. Rev. Earth Planet. Sci. 9, 285309. PAQUETTE, J. L., MONCHOUX, P. & COUTURIER, M. (1995): Geochemical and isotopic study of a norite-eclogite transition in the European Variscan Belt - implications for U-Pb zircon systematics in metabasic rocks. Geochim. Cosmochim. Acta 59, 1611-1622. PATERSON, S. R., VERNON, R. H. & TOBISCH, O. T. (1989): A review of criteria for the identification of magmatic and tectonic foliations in granitoids. J. Struct. Geol. 11, 349-363. PATINO DOUCE, A. E. & HARRIS, N. (1998): Experimental constraints on Himalayan anatexis. J. Petrol. 39, 689-710.

42

PATINO 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. Contrib. Mineral. Petrol. 107, 202-218. PEACOCK, S. M., RUSHMER, T. & THOMPSON, A. B. (1994): Partial melting of subducting oceanic-crust. Earth Planet. Sci. Lett. 121, 227-244. PHILLIPS, E. R. (1974): Myrmekite - one hundred years later. Lithos 7, 181-194. PIN, C. (1990): Variscan oceans - ages, origins and geodynamic implications inferred from geochemical and radiometric data. Tectonophys. 177, 215-227. PIN, C., BINON, M., BELIN, J. M., BARBARIN, B. & CLEMENS, J. D. (1990): Origin of microgranular enclaves in granitoids - equivocal Sr-Nd evidence from Hercynian rocks in the Massif-Central (France). J. Geophys. Res. Solid Earth Planets 95, 1782117828. PIN, C. & DUTHOU, J. L. (1990): Sources of Hercynian granitoids from the French MassifCentral - inferences from Nd isotopes and consequences for crustal evolution. Chem. Geol. 83, 281-296. PIN, C. & LANCELOT, J. (1982): U - Pb Dating of an early Paleozoic bimodal magmatism in the French Massif Central and of its further metamorphic evolution. Contrib. Mineral. Petrol. 79, 1-12. PIN, C. & PEUCAT, J. J. (1986): Age des épisodes de métamorphisme Paléozoïque dans le Massif Central et le Massif Armoricain. Bull. Soc. Geol. Fr. 2, 461-469. PITCHER, W. S. (1983): Granite type and tectonic environment. In Mountain building processes (ed. A. Press), London, 19-40. RAPP, R. P., SHIMIZU, N., NORMAN, M. D. & APPLEGATE, G. S. (1999): Reaction between slab-derived melts and peridotite in the mantle wedge: experimental constraints at 3.8 GPa. Chem. Geol. 160, 335-356.

43

REINECKE, T. (1991): Very-high-pressure metamorphism and uplift of coesite-bearing metasediments from the Zermatt-Saas Zone, Western Alps. European Journal of Mineralogy 3, 7-17. RHODES, J. M. & DAWSON, J. B. (1975): Major and trace element chemistry of peridotite inclusions from the Lashaine Volcano, Tanzania. Phys. Chem Earth 9, 545-557. ROCK, N. M. S. (1984): Nature and origin of calk-alkali lamprophyres : minettes, vogesites, kersantites and spessartites. Trans. R. Soc. Edinb.-Earth Sci. 74, 193-227. ROGERS, N. W., BACHINSKI, S. W., HENDERSON, P. & PARRY, S. J. (1982): Origin of potashrich basic lamprophyres - trace-element data from Arizona Minettes. Earth Planet. Sci. Lett. 57, 305-312. ROLLINSON, H. R. (1993): Using Geochemical Data: Evaluation, Presentation, Interpretation. Longman scientific & technical, Singapore. RUDNICK, R. L., MCDONOUGH, W. F. & CHAPPELL, B. W. (1993): Carbonatite metasomatism in the Northern Tanzanian Mantle - petrographic and geochemical characteristics. Earth Planet. Sci. Lett. 114, 463-475. SAWYER, E. W. (1996): Melt segregation and magma flow in migmatites: Implications for the generation of granite magmas. Trans. R. Soc. Edinb.-Earth Sci. 87, 85-94. ________. (1998): Formation and evolution of granite magmas during crustal reworking; the significance of diatexites. J. Petrol. 39, 1147-1167. SCHERTL, H. P., SCHREYER, W. & CHOPIN, C. (1991): The pyrope-coesite rocks and their country rocks at Parigi, Dora Maira Massif, Western Alps - detailed petrography, mineral chemistry and Pt-path. Contrib. Mineral. Petrol. 108, 1-21. SHAW, D. M. (1970): Trace element fractionation during anatexis. Geochim. Cosmochim. Acta 34, 237-243.

44

SIMIEN, F., MATTAUER, M. & ALLEGRE, C. J. (1999): Nd isotopes in the stratigraphic record of the Montagne Noire (French Massif Central): No significant Paleozoic juvenile inputs, and pre-Hercynian paleogeography. J. Geol. 107, 87-97. SINGER, B. S. & PEARCE, T. H. (1993): Plagioclase zonation in a basalt to rhyodacite eruptive suite, Seguam Island, Alaska - observations by Nomarski contrast interference. Can. Mineral. 31, 459-466. SPARKS, R. S. J. & MARSHALL, L. A. (1986): Thermal and mechanical constraints on mixing between mafic and silicic magmas. J. Volcanol. Geotherm. Res. 29, 99-124. STORRE, B. & KAROTKE, E. (1972): Experimental-data on melting reactions of Muscovite + Quartz in Systems K2O-Al2O3-SiO2-H2 to 20 Kb water pressure. Contrib. Mineral. Petrol. 36, 343-345. STRECKEISEN, A. (1976): To each plutonic rock its proper name. Earth-Sci. Rev. 12, 1-33. SUN, S. S., MCDONOUGH, W. F., SAUNDERS, A. D. & NORRY, M. J. (1989): Chemical and isotopic systematics of oceanic basalts; implications for mantle composition and processes, vol. 42, pp. 313-345. Geological Society of London, London, United Kingdom. TAYLOR, S. R. & MCLENNAN, S. M. (1985): The continental crust :its composition and evolution., Oxford. 312. THOMPSON, A. B. & CONNOLLY, J. A. D. (1995): Melting of the continental-crust - some thermal and petrological constraints on anatexis in continental collision zones and other tectonic settings. J. Geophys. Res. Solid Earth 100, 15565-15579. TSUCHIYAMA, A. (1985): Dissolution kinetics of plagioclase in the melt of the system Diopside-Albite-Anorthite, and origin of dusty plagioclase in andesites. Contrib. Mineral. Petrol. 89, 1-16.

45

TURPIN, L., CUNEY, M., FRIEDRICH, M., BOUCHEZ, J. L. & AUBERTIN, M. (1990): Metaigneous origin of Hercynian peraluminous granites in NW French Massif Central implications for crustal history reconstructions. Contrib. Mineral. Petrol. 104, 163172. TURPIN, L., VELDE, D. & PINTE, G. (1988): Geochemical comparison between Minettes and Kersantites from the Western European Hercynian orogen - trace-element and Pb-SrNd isotope constraints on their origin. Earth Planet. Sci. Lett. 87, 73-86. VANDERHAEGHE, O. (2001): Melt segregation, pervasive melt migration and magma mobility in the continental crust; the structural record from pores to orogens. Phys. Chem. Earth 26, 213-223. VANDERHAEGHE, O., BURG, J. & TEYSSIER, C. (1999): Exhumation of migmatites in two collapsed orogens Canadian Cordillera and French Variscides. Exhumation processes normal faulting, ductile flow and erosion Geological Society Special Publications 154, 181-204. VANDERWAL, D. & BODINIER, J. L. (1996): Origin of the recrystallisation front in the Ronda peridotite by km-scale pervasive porous melt flow. Contrib. Mineral. Petrol. 122, 387405. VENTURELLI, G., CAPEDRI, S., DIBATTISTINI, G., CRAWFORD, A., KOGARKO, L. N. & CELESTINI, S. (1984): The ultrapotassic rocks from Southeastern Spain. Lithos 17, 3754. VERNON, R. H. (1984): Microgranitoid enclaves in granites - globules of hybrid magma quenched in a plutonic environment. Nature 309, 438-439. ________. (1986): K-feldspar megacrysts in granites - phenocrysts, not porphyroblasts. Earth-Sci. Rev. 23, 1-63.

46

VERNON, R. H., WILLIAMS, V. A. & DARCY, W. F. (1983): Grain-size reduction and foliation development in a deformed granitoid batholith. Tectonophys. 92, 123-145. VIELZEUF, D. & HOLLOWAY, J. R. (1988): Experimental-determination of the fluid-absent melting relations in the pelitic system - consequences for crustal differentiation. Contrib. Mineral. Petrol. 98, 257-276. VIGNERESSE, J. L. & TIKOFF, B. (1999): Strain partitioning during partial melting and crystallizing felsic magmas. Tectonophys. 312, 117-132. WATSON, E. B., VICENZI, E. P. & RAPP, R. P. (1989): Inclusion host relations involving accessory minerals in high-grade metamorphic and anatectic rocks. Contrib. Mineral. Petrol. 101, 220-231. WIEBE, R. A. (1968): Plagioclase stratigraphy - a record of magmatic conditions and events in a granite stock. Am. J. Sci. 266, 690-703 WIEBE, R. A. & COLLINS, W. J. (1998): Depositional features and stratigraphic sections in granitic plutons: implications for the emplacement and crystallization of granitic magma. J. Struct. Geol. 20, 1273-1289. WILLIAMSON, B. J., DOWNES, H. & THIRLWALL, M. F. (1992): The Relationship between crustal magmatic underplating and granite genesis - an example from the Velay Granite Complex, Massif-Central, France. Trans. R. Soc. Edinb.-Earth Sci. 83, 235245. WINTERBURN, P. A., HARTE, B. & GURNEY, J. J. (1990): Peridotite xenoliths from the Jagersfontein Kimberlite pipe .1. primary and primary-metasomatic mineralogy. Geochim. Cosmochim. Acta 54, 329-341. WYLLIE, P. J. (1987): Metasomatism and fluid generation in mantle xenoliths. In Mantle Xenoliths (ed. P. H. Nixon), Wiley, New York 609-621. Wiley, New York.

47

Figure captions: Fig. 1. Simplified geologic map of the French Massif Central with location of study area.

Fig. 2. Geological map (a) and cross (b) section of the study area showing the relationship between the different rock types

Fig. 3. Photomicrographs. (a) Plagioclase (Plg) fractured in porphyritic monzogranite. The fracture is filled by quartz (Q). (b) Myrmekitic structure (Myr) due to the chemical destabilization of K-feldspar (FK) next to the paragneiss (Pgn). (c) Entire thin sections ( ~24 by 35 mm) showing the edges of an MME. Note the larger biotite crystals at the border. (d) rounded plagioclase in an MME (Plg).

Fig. 4. (a) Photograph showing heterogeneity in the porphyritic monzogranite. (b) Mafic Microgranular Enclave (MME) in monzogranite, pen is 15 Cm long.

Fig. 5. Back-scattered electron (BSE) images (Z contrast) of plagioclase from a porphyritic monzogranite. (a) Plagioclase with simple zonation due to factional crystallisation. (b) Plagioclase with complex zoning (see explanation on Fig. 6)

Fig. 6. Formation of plagioclase with complex core structure. (a) When a plagioclase that crystallized from the hydrid magma comes to contact with a mafic magma, the plagioclase is resorbed. (b) A plagioclase rich in An crystallizes in the mafic magma around the resorbed plagioclase. (c) The plagioclase now has a complex core structure. New plagioclase of An30

48

crystallizes from the hybrid magma. By the process of fractional crystallization later plagioclase growth is progressively less calcic down to An20.

Fig. 7. Selected Harker diagrams comparing the composition of anatectic metasedimentary rocks : metatexite paragneiss and diatexite migmatite (‘), enclaves of paragneiss (c) and two-micas leucogranite (¯). The hachured zone correspond to a compilation of melt compositions obtained from the partial fusion of pelite (Patino Douce & Johnston, 1991 ; Gardien et al., ; 1995 Patino Douce & Harris, 1998).

Fig. 8. Spider diagrams for anatectic rocks (dashed line) and two-mica leucogranite. The concentrations of the majority of the trace elements are higher in the anatectic rocks.

Fig. 9. eNd vs. 87Sr/86Sr diagram. All values recalculated to 300 Ma.

Fig. 10. Result of the modeling of partial melting of paragneiss (a) Restite : 65% biotite, 15% plagioclase 10% quartz, 5% K-feldspar (b) Restite : 64% biotite, 15% plagioclase 10% quartz, 5% K-feldspar, 1% zircon

Fig. 11.

Selected major and trace element diagrams showing the composition of the

porphyritic monzogranites (…) in relation to the composition of the MME (z) and the leucogranites (¯). a) K2O vs. MgO; b) SiO2 vs. CaO c) Rb vs. Nd; (+) represent a mixing curve at 10% intervals with mafic magma represent by the geochemistry of the enclave.

Fig. 12. Results of mixing between a mafic (MEE-like z) and a felsic (two-mica leucogranite-like +) magma for the formation of the porphyritic monzogranite (…). The mafic

49

component corresponds to sample 155c (see Table 1) and the felsic to sample SGL2K25a (table 1). (a) “Mixing test” of Fourcade & Allègre (1981) for major elements, showing that the major element composition of the porphyritic monzogranite is consistent with the mixing of 70 % of mafic end-member and (30) % of the felsic end-member. The R2 coefficient in this case is 0.95. (b) REE composition of a mixture of 70 % of the mafic component 155c and 30% of a felsic component SGL2K25, the dashed line (c) correspond to the model, and is compared to the real composition of the porphyritic monzogranite (…) 155b (Table 1).

Fig. 13. Spider diagrams normalized to upper crust (Taylor & McLennan, 1981) for MME (z) compared to lamprophyres from Cevennes (+) and Thiers ( c ) (Agranier, 2001)

Fig. 14. Chondrite-normalized (Sun et al., 1989) multi-element diagrams for possible compositions of the mantle source. In all diagrams, the shaded area is the field of recalculated mantle compositions (see text for details) and is compared with various peridotite samples. (a) Comparison between the recalculated mantle compositions and a mantle with residual garnet (shaded), or spinel (hachured). (b) Massif Central xenoliths in recent (Tertiary and Quaternary) lavas (Lenoir et al., 2000). (c) Orogenic peridotites, Ronda (Frey et al., 1985). (d) Sub-arc mantle, xenoliths in recent Philipinos lavas (Maury et al., 1992). (e) Peridotitic xenoliths in Kimberlites (hachured field) (Erlank et al., 1987; Menzies et al., 1987; Nixon et al., 1981; Rhodes & Dawson, 1975; Rudnick et al., 1993; Winterburn et al., 1990) ; thick lines correspond to 3 samples from Bultfontein pipe, samples RS4 and RS6 from Menzies et al. (1987) and sample JJG360A from Erlank et al. (1987). (f) Mixing of Massif Central xenolithic mantle, with 1-20 % of sediments (average Post-Archaean shale of Taylor & McLennan, 1985).

50

Fig. 15. εNd vs.

87

Sr/86Sr

diagram, all values recalculated at 300 Ma unless specified

otherwise. Field of lamprophyres (all Massif Central) after Turpin et al. (1988) and Agranier (2001) (recalculated at the time of emplacement, different for each sample but between 310 and 320 Ma); Field of granulitic metasediments (Bournac, 50 km SE of study area) after Downes et al. (1990). (a) Assimilation of granulitic lower crust by a mafic magma with CHUR-like isotopic characteristics (ticks correspond to 10 % increments); (b) Mixing between Livradois metasediments and a mafic magma with CHUR-like isotopic characteristics (ticks correspond to 10 % increments). (c) shaded field is the isotopic composition (recalculated at 315 Ma) of a CHUR-like mantle into which sediments have been mixed at 420 Ma. Isotopic composition of the sediments are taken from the pre-Devonian sediments from the Montagne Noire area, 250 km S of study area (Simien et al., 1999). Dashed line: sediments with eNd (420 Ma) = -6 ; ticks correspond to increments of 1, 2, 5, 10 and 20 %. Limits of the shaded field correspond to models using sediments with eNd (420 Ma) = -4 to –8.

Fig. 16. Influence of possible components in the source of the mafic component of the porphyritic monzogranite. Field of peridotites: Bournac granulites (Dostal et al. 1980); Sediments and paragneisses: this work and Williamson et al. (1992). In La vs Yb diagram (a), the Mg-K series is actually more enriched than the granulite with higher La/Yb ratios. Interactions of a mantle-derived melt with lower crust granulites, are therefore unable to generate the observed melt composition. In Sr vs. Ba diagram (b), granulites show Ba/Sr ratios that are too high to be the source of enrichment such that addition of a small proportion of them to mantle peridotite ( see text and Fig. 14 ) is unable to generate the expected signature.

51

Fig. 1. Simplified geologic map of the French Massif Central with location of study area.

52

Fig. 2. Geological map (a) and cross (b) section of the study area showing the relationship between the different rock types

53

Pgn

Plg Q Q My

FK

b

a

Plg

c

d

Fig. 3. Photomicrographs. (a) Plagioclase (Plg) fractured in porphyritic monzogranite. The fracture is filled by quartz (Q). (b) Myrmekitic structure (Myr) due to the chemical destabilization of K-feldspar (FK) next to the paragneiss (Pgn). (c) Entire thin sections ( ~24 by 35 mm) showing the edges of an MME. Note the larger biotite crystals at the border. (d) rounded plagioclase in an MME (Plg).

54

a)

b)

Fig. 4. (a) Photograph showing heterogeneity in the porphyritic monzogranite. (b) Mafic Microgranular Enclave (MME) in monzogranite, pen is 15 Cm long.

a

b An 39 An 33

An 26 An 20 An 24

Fig. 5. Back-scattered electron (BSE) images (Z contrast) of plagioclase from a porphyritic monzogranite. (a) Plagioclase with simple zonation due to factional crystallisation. (b) Plagioclase with complex zoning (see explanation on Fig. 6)

55

Fig. 6. Formation of plagioclase with complex core structure. (a) When a plagioclase that crystallized from the hydrid magma comes to contact with a mafic magma, the plagioclase is resorbed. (b) A plagioclase rich in An crystallizes in the mafic magma around the resorbed plagioclase. (c) The plagioclase now has a complex core structure. New plagioclase of An30 crystallizes from the hybrid magma. By the process of fractional crystallization later plagioclase growth is progressively less calcic down to An20.

56

Fig. 7. Selected Harker diagrams comparing the composition of anatectic metasedimentary rocks : metatexite paragneiss and diatexite migmatite (‘), enclaves of paragneiss (c) and two-micas leucogranite (¯). The hachured zone correspond to a compilation of melt compositions obtained from the partial fusion of pelite (Patino Douce & Johnston, 1991 ; Gardien et al., ; 1995 Patino Douce & Harris, 1998).

57

Fig. 8. Spider diagrams for anatectic rocks (dashed line) and two-mica leucogranite. The concentrations of the majority of the trace elements are higher in the anatectic rocks.

Fig. 9. eNd vs. 87Sr/86Sr diagram. All values recalculated to 300 Ma.

58

Fig. 10. Result of the modeling of partial melting of paragneiss (a) Restite : 65% biotite, 15% plagioclase 10% quartz, 5% K-feldspar (b) Restite : 64% biotite, 15% plagioclase 10% quartz, 5% K-feldspar, 1% zircon

59

Fig. 11.

Selected major and trace element diagrams showing the composition of the

porphyritic monzogranites (…) in relation to the composition of the MME (z) and the leucogranites (¯). a) K2O vs. MgO; b) SiO2 vs. CaO c) Rb vs. Nd; (+) represent a mixing curve at 10% intervals with mafic magma represent by the geochemistry of the enclave.

60

Fig. 12. Results of mixing between a mafic (MEE-like z) and a felsic (two-mica leucogranite-like +) magma for the formation of the porphyritic monzogranite (…). The mafic component corresponds to sample 155c (see Table 1) and the felsic to sample SGL2K25a (table 1). (a) “Mixing test” of Fourcade & Allègre (1981) for major elements, showing that

61

the major element composition of the porphyritic monzogranite is consistent with the mixing of 70 % of mafic end-member and (30) % of the felsic end-member. The R2 coefficient in this case is 0.95. (b) REE composition of a mixture of 70 % of the mafic component 155c and 30% of a felsic component SGL2K25, the dashed line (c) correspond to the model, and is compared to the real composition of the porphyritic monzogranite (…) 155b (Table 1).

Fig. 13. Spider diagrams normalized to upper crust (Taylor & McLennan, 1981) for MME (z) compared to lamprophyres from Cevennes (+) and Thiers ( c ) (Agranier, 2001)

62

Fig. 14. Chondrite-normalized (Sun et al., 1989) multi-element diagrams for possible compositions of the mantle source. In all diagrams, the shaded area is the field of recalculated mantle compositions (see text for details) and is compared with various peridotite samples. (a) Comparison between the recalculated mantle compositions and a mantle with residual garnet (shaded), or spinel (hachured). (b) Massif Central xenoliths in recent (Tertiary and Quaternary) lavas (Lenoir et al., 2000). (c) Orogenic peridotites, Ronda (Frey et al., 1985). (d) Sub-arc mantle, xenoliths in recent Philipinos lavas (Maury et al., 1992). (e) Peridotitic xenoliths in Kimberlites (hachured field) (Erlank et al., 1987; Menzies et al., 1987; Nixon et al., 1981; Rhodes & Dawson, 1975; Rudnick et al., 1993; Winterburn et al., 1990) ; thick lines correspond to 3 samples from Bultfontein pipe, samples RS4 and RS6 from Menzies et

63

al. (1987) and sample JJG360A from Erlank et al. (1987). (f) Mixing of Massif Central xenolithic mantle, with 1-20 % of sediments (average Post-Archaean shale of Taylor & McLennan, 1985).

64

Fig. 15. εNd vs.

87

Sr/86Sr

diagram, all values recalculated at 300 Ma unless specified

otherwise. Field of lamprophyres (all Massif Central) after Turpin et al. (1988) and Agranier (2001) (recalculated at the time of emplacement, different for each sample but between 310 and 320 Ma); Field of granulitic metasediments (Bournac, 50 km SE of study area) after Downes et al. (1990). (a) Assimilation of granulitic lower crust by a mafic magma with CHUR-like isotopic characteristics (ticks correspond to 10 % increments); (b) Mixing between Livradois metasediments and a mafic magma with CHUR-like isotopic characteristics (ticks correspond to 10 % increments). (c) shaded field is the isotopic composition (recalculated at 315 Ma) of a CHUR-like mantle into which sediments have been mixed at 420 Ma. Isotopic composition of the sediments are taken from the pre-Devonian sediments from the Montagne Noire area, 250 km S of study area (Simien et al., 1999). Dashed line: sediments with eNd (420 Ma) = -6 ; ticks correspond to increments of 1, 2, 5, 10 and 20 %. Limits of the shaded field correspond to models using sediments with eNd (420 Ma) = -4 to –8.

65

Fig. 16. Influence of possible components in the source of the mafic component of the porphyritic monzogranite. Field of peridotites: Bournac granulites (Dostal et al. 1980); Sediments and paragneisses: this work and Williamson et al. (1992). In La vs Yb diagram (a), the Mg-K series is actually more enriched than the granulite with higher La/Yb ratios. Interactions of a mantle-derived melt with lower crust granulites, are therefore unable to generate the observed melt composition. In Sr vs. Ba diagram (b), granulites show Ba/Sr ratios that are too high to be the source of enrichment such that addition of a small proportion of them to mantle peridotite ( see text and Fig. 14 ) is unable to generate the expected signature.

66