Appendix
Geochronology: analytical procedure
Nine samples were collected across the Central Vosges metamorphic units. They were crushed and sieved, and the fraction smaller than 500 µm was processed using a Frantz magnetic separator and heavy liquids. Zircon grains were finally hand picked under a binocular microscope and mounted in araldite.
U−Pb zircon analyses were performed using a LASER ablation system coupled with a Neptune multicollector (MC) ICP-MS (TermoFisherScientific). A new CETAC 213 nm UV laser was used allowing the U/Pb fractionation to be significantly reduced. The MC-ICP-MS was equipped with a multi-ion counting system operating in static mode; all masses of interest were recorded simultaneously (202Hg, 204Pb, 206Pb, 207Pb, 208Pb on ion-counters, and 232Th, 238U on Faraday cup collectors). Isotope ratios were normalized using the zircon standard 91500 from Ontario (1065±1 Ma; Wiedenbeck et al. 1995). Details on analytical conditions are given in Cocherie and Robert (2008) and Cocherie et al. (2009). The results of U−Pb zircon analyses are available as supplementary material (supplementary Tables S3−S5).
Considering that the precision and accuracy on 206Pb/204Pb determination is satisfactory for zircon grains older than ca. 600 Ma, the associated analyses are discussed using a conventional Concordia diagram (Wetherill 1956). By contrast, younger analyses are concurrently examined in a conventional Concordia plot as well as in a 238U/206Pb−207Pb/206Pb inverse Concordia diagram (Tera and Wasserburg 1972). This dual approach was adopted because the conventional representation can easily reveal possible radiogenic Pb (Pb*) loss or Discordia trends, whereas the inverse diagram does not require initial-Pb correction and gives relatively precise average ages. In the present case, Variscan ages calculated using both methods are compatible within errors, but the inverse Concordia approach systematically involves more spot analyses. Therefore, the results obtained using this latter method have been favoured in this work.
Geochronology: sample description
Monotonous gneiss samples (EV67M, EV69M, EV372R)
In thin section, the monotonous paragneiss samples display a matrix formed by thick (up to 5 mm) coarse-grained quartz−plagioclase−K-feldspar layers alternating with biotite-rich layers (1−2 mm) where muscovite is only rarely present. They also show stripes of fibrous sillimanite and yellowish pinitized cordierite parallel to the dominant foliation. In samples containing garnet (EV69M & EV372R), sillimanite pseudomorphs after garnet are observed. Additional minerals involve chlorite or chloritized biotite, ilmenite, secondary hematite, apatite, zircon and monazite.
Varied gneiss samples (EV83M, EV90M, EV343M)
Sample EV83M is clearly migmatitic and preserves an alternation of thick (2−3 cm) leucosomes and biotite−garnet-bearing restitic layers. Sample EV90M was collected at the contact with the monotonous gneiss unit and corresponds to a light-coloured, quartz-rich and fine-grained gneiss containing few garnet. Sample EV343M shows numerous rounded K-feldspar blasts lying in a foliated biotite−garnet-bearing matrix. In thin section, all varied gneiss samples show a marked differentiation between quartz−K-feldspar-rich and biotite-rich layers, and commonly host a significant amount of garnet porphyroblasts (1−5 mm). Additional minerals are chlorite, ilmenite, secondary hematite and zircon.
Felsic granulite samples (EV136N, EV357N, EV389N)
All felsic granulite samples correspond to pink or white quartz−K-feldspar rocks showing a fine mylonitic texture and hosting tiny (< 1 mm) garnet. In thin section elongated quartz stripes alternate with large (500 µm) perthitic K-feldspar and fine-grained plagioclase-rich layers. Garnet is always present and hosts rounded inclusions of perthite or rutile. It is variably replaced by biotite or chlorite. In sample EV357N, kyanite is frequently observed in the matrix or in textural equilibrium with garnet and rutile, but sillimanite overgrowths around kyanite crystals are common. Accessory minerals are rutile, apatite and zircon.
Petrology: mineral abbreviations
Mineral abbreviations used in the present work are defined as follows: and=andalusite; bi=biotite; chl=chlorite; cd=cordierite; g=garnet with end members alm=almandine [Fe/(Fe+Mg+Ca+Mn)], grs=grossular [Ca/(Fe+Mg+Ca+Mn)], prp=pyrope [Mg/(Fe+Mg+Ca+Mn)] and sps=spessartine [Mn/(Fe+Mg+Ca+Mn)]; ilm=ilmenite; ksp=K-feldspar; ky=kyanite; liq=granitic melt; mz=monazite; mu=muscovite; pa=paragonite; pl=plagioclase with end members ab=albite [Na/(Na+Ca+K)], an=anorthite [Ca/(Na+Ca+K)], or=orthoclase [K/(Na+Ca+K)]; q=quartz; ru=rutile; sill=sillimanite; zrc=zircon.
References
Cocherie A, Fanning CM, Jezequel P, Robert M (2009) LA-MC-ICPMS and multi-ion counting system, and SHRIMP U-Pb dating of complex zircons from quaternary tephras from the French Massif Central: magma residence time and geochemical implications. Geochimica et Cosmochimica Acta 73:1095-1108
Cocherie A, Robert M (2008) Laser ablation coupled with ICP-MS applied to U–Pb zircon geochronology: A review of recent advances. Gondwana Research 14:597-608
Tera F, Wasserburg GJ (1972) U–Th–Pb systematics in three Apollo 14 basalts and the problem of initial Pb in lunar rocks. Earth and Planetary Science Letters 14:281-304
Wetherill GW (1956) Discordant uranium-lead ages 1. Transactions of the American Geophysical Union 37:320-326
Wiedenbeck M, Allé P, Corfu F, Griffin WL, Meier M, Oberli F, von Quadt A, Roddick JC, Spiegel W (1995) Three natural zircon standards for U-Th-Pb, Lu-Hf, trace element and REE analyses. Geostandards Newsletter 19:1-23
1