ULTRAHIGH-PRESSURE METAMORPHISM OF GARNET PERIDOTITES FROM POHORJE MTS. (EASTERN ALPS, SLOVENIA)

A detailed study of the trace-element composition of olivine from a variety of mantle lithologies has been undertaken to (1) constrain the variability within and between various lithologies, (2) to obtain a better understanding of trace element partitioning into olivine and (3) to explore the potential of olivine for thermobarometry of mantle rocks. Results from mantle xenoliths from three localities will be presented, two from kimberlite pipes (Kaalvallei, South- Africa, and Kirkland Lake, Canada) and one from a basaltic volcanic centre (Ray Pic, France). Xenoliths from the latter are all spinel peridotites (ol, opx, cpx, sp), whereas xenoliths from the kimberlites are garnet peridotites (ol, opx, gt, ±cpx, ±sp). Analyses were performed either in situ or on olivine separates by laser ablation ICP-MS at our lab in Gothenburg. Typically five olivines per sample were ablated using 300- µm craters ablated at 10Hz for 120 s, resulting in detection limits between 1 and 10 ppb for most elements. Accuracy was monitored using an in-house standard prepared from olivines separated from a peridotite xenolith from the Kimberley pipe, which has a variability of all trace elements analyzed of less than 3%. Major elements were analyzed by SEM-EDS recording 400s spectra. Accuracy was monitored against San Carlos olivine, of which repeat analysis gave a value of Fo90.13 ± 0.07. Forsterite contents of all olivines lie between 90.1 and 93.3, with the lowest values in the spinel peridotites from France, indicating a lower degree of depletion compared to cratonic xenoliths. The variation in Li, Mn, Co, Ni and Zn is small (less than 25%), in accordance with olivine being the major host for these elements. Manganese and Li correlate positively within each locality, and are higher in spinel than in garnet peridotites. The largest variations are observed in Na, Al, Cr, REE, Zr and are highest for Ti, which varies from 0.4 to 240 ppm, although the variation in much more limited in the spinel peridotites. Similar large variations are observed in Ti contents of minerals co-existing with olivine, which indicates that variation is dominated by bulk-rock concentrations. The extreme Ti variation in cratonic xenoliths reflects the enormous diversity in the compositions of mantle rocks created by multiple events of melt depletion and metasomatic refertilization, and olivine is a sensitive indicator of these processes. Similar considerations may explain variations in Zr and possibly REE. On the other hand, Na, Al and Cr are relatively constant in co-existing minerals, requiring another explanation for the variation of these elements. Sodium and Cr co-variation defines a narrow band of compositions with Na and Cr concentrations being almost equal, independent of lithology. Al and Cr, on the other hand, show significant difference between garnet and spinel peridotites, where spinel peridotites are offset to higher Al values. This is probably related to differences in partitioning behavior between garnet-olivine and spinel-olivine, although currently we cannot rule out a pressure effect on partitioning. A chart of all monovalent cations (Li, Na) against trivalent cations (Al, Cr, Sc) also shows significant differences. Whereas garnet peridotites show a close to 1:1 correlation, which can be expected because of coupled substitution of a tri- and a monovalent cation for two divalent cations (Mg2SiO4 + Na+ + Cr3+ = NaCrSiO4 + 2Mg2+), spinel peridotites show a 1:3 correlation. The excess of trivalent cations can possibly be explained by a Tschermak-type substitution, i.e., Mg2SiO4 +2Al3+ = MgAlVIAlIVO4. + Mg2+ + Si4+. As the size of the Al ion is in-between the size of the octahedral and tetrahedral sites, such partitioning behavior can be expected based on lattice-strain theory [1]. However, in contrast to most other rock-forming silicates, silica tetrahedrons in olivines do not share oxygens, and replacement of Si by Al would therefore significantly distort the crystal lattice, which seems energetically unfavorable [2]. However, the 1:3 correlation in the spinel peridotites presents the interesting possibility that substitution occurs through double coupled substitution, i.e, 2Mg2SiO4 + 3Al3+ + Na+ = NaAlVISiO4 + MgAlVIAlIVO4 + 3Mg2+ + Si4+, in which the two Al-bearing molecules may have a mutually stabilizing effect. Admittedly the proposition is speculative and requires further study. Alternative explanations could be vacancy substitution or the presence of significant amounts of H+, which has not been analyzed for. Why trivalent cation excess occurs in spinel peridotites only is unclear, and the analysis of similar xenolith suites should tell whether it is a general trend. Irrespective of the substitution mechanism, the concentration of Al in olivine from mantle peridotites is strongly temperature dependent, and can therefore be used as a geothermometer. Different calibrations for spinel and garnet peridotite are necessary. Peridotite xenoliths from the Kaalvallei kimberlite (South-Africa) were used for calibration of the garnet-bearing variety of the thermometer, and xenoliths from Ray Pic (France) for the spinel-bearing one. Aluminum contents of the olivines range from 8-140 ppm, with a few anomalous values up to 310 ppm, for the Kaalvallei olivines, and from 49-191 ppm for Ray Pic olivines. Mg# lie between 91.5 and 93.3, and between 90.1 and 91.4, respectively. P-T conditions of the peridotites were estimated with the Al-in-opx geobarometer and cpx-opx geothermometer [3], and plot on conductive geotherms of ca. 40 and 60 mW/m2, respectively. The temperature ranges are 900- 1380°C and 830-1060°C, respectively. Only samples plotting on the local geotherms were considered, as the other samples showed disequilibrium features. The expression for the thermometers are: TGt Al-in-ol(°C) = 11390 / [ 11.88 - ln(ppm Al) ] – 273 (garnet peridotites) TSp Al-in-ol(°C) = 10871 / [ 7.46 - ln(ppm Al) ] – 273 (spinel peridotites) with average residuals of 12°C and 29°C, respectively. As the compositional range of mantle olivine is very small, no correction for major chemical components is necessary. In addition, no correction for Al activity of the system is necessary, as long as an Al-saturated phase such as garnet or spinel is present. Combined with the Ca-in-olivine barometer [4], the new thermometer has the potential to determine P-T conditions of single olivines. Whether olivine is derived from garnet of spinel peridotite can easily be determined from the Cr/Al ratio of the olivine (<0.7 for spinel peridotites). As olivine is an abundant component of heavy mineral separates from kimberlites and glacial till, it could serve as a new tool for diamond exploration. Vanadium and Cr show similar temperature-dependent variations as Al, but to a lesser degree, and would therefore yield less accurate geothermometers. In addition, partitioning of these elements is sensitive to variations in oxidation state. The pressure dependence of the thermometer is the subject of future research and it is recommended that the Alin- olivine thermometer in its current form is applied to rocks derived from comparable geotherms only.

In the Ulten Zone (Upper Austroalpine), small bodies of mantle peridotites are incorporated within high-grade basement rocks (gneisses and migmatites) which represent remnants of lower crust subducted and reequilibrated at eclogite- facies conditions during the Variscan orogenic cycle. The Ulten peridotites record a complex metamorphic and deformative evolution, which is testified by the transition from coarse-grained protogranular spinel-bearing peridotites, to fine-grained garnet and amphibole (Ca-hornblende) -bearing peridotites with porphyroclastic to mosaic granoblastic textures. Thermometric estimates on the coarse-type spinel lherzolites have yielded high temperatures of equilibration, in the range 1100-1300°C (Obata and Morten, 1987). In the porphyroclastic peridotites, the metamorphic recrystallization to (garnet + amphibole)-facies conditions is evidenced by the development of: i) garnet coronas around spinel, ii) fine-grained granoblastic aggregates made by olivine + garnet + Ca-hornblende + new pyroxenes, iii) garnet and Ca-hornblende exsolutions within primary spinel-facies clino- and ortho-pyroxenes. The P-T conditions of the high-pressure eclogitic recrystallization which produced the spinel- to garnet-facies transition have been recently estimated to 850°C and 27 kbar (Nimis and Morten, 1999). The peculiar thermobarometric reequilibration recorded by the Ulten peridotites has been interpreted as the result of a wedge to slab evolution (Nimis and Morten, 1999; Godard et al., 1996). In this scenario, the spinel peridotites represent portions of a mantle wedge which were incorporated (by convection ) in a downgoing slab of cold continental crust, and were then subducted together with the slab to depths of about 90 km. Entrainment in the cold slab and subduction caused the reequilibration of the peridotites at 850°C and 27 kbar. The metamorphic transition from spinel- to garnetbearing assemblage occurred therefore in a dynamic regime, and was accompanied by significant input of metasomatic fluids, as testified by the crystallization of abundant amphibole in the garnet-bearing high-pressure assemblage. Petrologic investigations on the host gneissic basement rocks have evidenced that they also experienced high-pressure recrystallization, which was accompanied by in-situ partial melting and migmatization (Godard et al., 1996). The particular geodynamic evolution of the Ulten peridotites thus offer the unique opportunity to investigate the effects of crustal-derived metasomatism on mantle rocks involved in a subducting environment. Previous whole-rock chemical and isotopic investigations on the Ulten peridotites have evidenced that the fine-type garnet-facies ultramafics are enriched in LREE, K, Sr and that the alkalis enrichment is positively correlated with the 87Sr/86Sr ratios. In this study, we present the results of detailed in situ investigations (performed by the ion microprobe operating at CSCC, Pavia, Italy) on the trace element chemistry of the main mineral phases (clinopyroxene, amphibole and garnets) from seven selected samples representative of the various stages of the tectonometamorphic evolution recorded by the peridotites. Major aims have been to investigate the geochemical signature of fluids responsible of the amphibole crystallization, and provide further constraints on the nature of the metasomatic processes. The data obtained are potentially usefull to characterize, by direct evidence, the chemical changes induced in mantle rocks by crustal metasomatism. The coarse-type spinel peridotites, which are relics of the “pre-subduction”, mantle-wedge equilibration stage, display modest metasomatic effects. In these samples, modal metasomatism is only recorded by the incipient crystallization of amphibole as rims around clinopyroxene. Clinopyroxenes have almost flat REE spectra (CeN/SmN = 0.76-0.87) at 4-8 x C1 values, or display concave shape with selective LREE enrichment (CeN/SmN = 2.50-4.50, SmN/YbN = 0.53-0.97). The REE concentrations of amphiboles are very similar to those of clinopyroxenes. Both amphiboles and clinopyroxenes in sample MK5D, a coarse-type garnet-bearing peridotite, exhibit a convex-upward REE pattern characterized by LREE and HREE depletion (CeN/SmN = 0.17-0.19; SmN/YbN = 3.07-6.13). Their low HREE abundances are due to the equilibration with garnet which, as expected, show severely fractionated patterns (CeN/YbN < 0.001; HREE at about 20-30 x C1). Amphiboles, in the coarse-type rocks, also show low Sr (18-35 ppm) and K (171-964 ppm) abundances. The most evident metasomatic effects are recorded by the eclogite-facies recrystallized fine-type peridotites. In these rocks, modal metasomatism is documented by abundant crystallization of amphibole (Ca-hornblende, Mg values: 90- 92) in equilibrium with garnet. In some samples, the gnt+cpx+amph+opx+ol assemblage is replaced by amph+opx+ol assemblages, this feature indicating progressive degrees of hydration. Amphiboles display significant LREE enrichment (CeN/YbN = 3.90-11.50; LREE in the range 20-50 x C1) and high Sr (150-250 ppm), K (1910- 7280 ppm) and Ba (280-800 ppm) contents. By contrast, they have relatively low concentrations in HFSE (e.g., Zr = 14-25 ppm, Y = 6.7-16 ppm, Ti = 1150-2500 ppm, Nb = 2-7 ppm). The geochemical signature recorded by amphibole in the fine-type peridotites, i.e. the strong enrichment in LILE relative to HFSE, is a peculiar feature of crustal-derived metasomatic agents. The lack of evidence of major element modifications in mantle minerals (e.g. Mg-value decrese, crystallization of orthopyroxene around olivine) strongly suggest that the metasomatic agent was an hydrous fluid rather than a silica-rich melt. Moreover, experimental studies have demonstrated that aqueous fluids preferentially partition elements like alkalies, Ba, Sr and Pb, whereas they have scarce affinity for HFSE. The results of our study therefore indicate that the chemical modifications occurred in the Ulten peridotites during the high-pressure reequilibration were most likely produced by the input of hydrous, LILE-enriched, fluids, which caused crystallization of abundant amphibole. Such H2O-rich fluids could represent the residual fluids left after the crystallization of leucosomes, starting from water-undersaturated melts produced during migmatization of the host gneisses.

Mantle peridotite xenoliths from Archean cratons generally have high molar Mg/(Mg+Fe), or Mg#. The best known suites, from the Kapvaal and Siberian cratons, have high modal orthopyroxene (Opx). These high Opx compositions are probably not residues of partial melting (Kelemen et al., 1998). Less well known cratonic xenolith suites from Greenland and North America include high Mg# peridotites with much lower modal Opx. Such low Opx compositions could be residual from high degrees of polybaric, decompression melting, ending in the spinel lherzolite stability field at pressures of 30 to 20 kbar (Bernstein et al., 1998). Similarly, based on Mg# and trace element evidence, we infer that the great majority of both spinel- and garnet-bearing xenoliths are also residues of polybaric melting that ended at pressures £ 30 kbar. Where xenoliths record equilibration pressures > 30 kbar, this must result from tectonic transport of peridotites to greater depth after melting (Kelemen et al., 1998). Proposed mechanisms for producing the high Mg#, high Opx compositions include metamorphic differentiation of high pressure residues, mixtures of residual peridotites and high pressure igneous cumulates from ultramafic magmas, and addition of SiO2 to low Opx peridotites via melt/rock reaction. A positive correlation between Ni contents of olivine and modal proportions of Opx in mantle xenoliths (Kelemen et al., 1998) is probably not produced by partial melting, metamorphic differentation, or formation of igneous cumulates. It can be produced by reaction between SiO2-rich liquids (e.g., small degree melts of subducted eclogite) and previously depleted, low Opx peridotites. Kelemen et al. (1998) proposed a two step process. First, high Mg#, low Opx peridotites were created by large degrees of polybaric melting ending at pressures < 30 kbar. Later, these depleted residues were enriched in Opx by interaction with SiO2-rich melts generated mainly by partial melting of eclogitic basalt and sediment in a subduction zone. Magmas modified by such a process could have formed a major component of the continental crust. Thus, this hypothesis provides a genetic link between cratonic upper mantle and continental crust. We are now conducting isotopic and trace element investigations of several suites of cratonic peridotites, with the following goals: (1)The hypothesis that high Opx peridotites with high Ni in olivine (e.g., xenoliths from the Premier kimberlite, Boyd, pers. comm. 1998) gained SiO2 via reaction with SiO2-rich partial melts of eclogite implies that high Opx peridotites should have mineral compositions in equilibrium with light rare earth element (REE) enriched melts, and probably should be depleted in Nb and enriched in Sr relative to light REE. Kelley and Kelemen are attempting to measure REE contents of Opx via ion probe and laser ICP-MS to determine whether this is the case. (2)The hypothesis that low Opx peridotites (e.g., xenoliths from East Greenland, Bernstein et al., 1998) represent true residual compositions, which have not been modified by subsequent major element metasomatism, would be strengthened if it could be shown that some or all of these also have trace element contents indicative of a residual origin. Hanghoj and Kelemen are using microbeam techniques to evaluate this. (3)The inference that highly depleted, low Opx peridotites, such as those from East Greenland, are Archean must be evaluated. Hanghoj, Blustajn and Frei have completed preliminary Os isotope measurements which give Archean “Re depletion ages”. Efforts to produce a true isochron have met with failure, and we have documented substantial, recent Re addition to most or all samples. (4)Modeling by Kelemen et al. (1998) presented a single melt/rock reaction scenario, involving a high SiO2 melt composition produced by experimental melting of eclogite at 30 kbar by Rapp and co-workers. This forward model fit the observed variation of Ni in olivine vs. modal Opx in cratonic peridotites. However, we were uncertain about the uniqueness of this model. Kelemen is conducting similar modeling using a variety of different initial melt compositions to evaluate alternative possibilities. (5)Jordan (e.g., 1988) proposed that the cratonic upper mantle is neutrally buoyant with respect to the convecting, oceanic upper mantle, suggesting that the lower temperature of the cratonic upper mantle is offset by compositional differences (mainly, lower Fe). We have noticed that the data set for the Kapvaal craton (e.g., Boyd, pers. comm. 1998) suggests instead that “low temperature” peridotites of the cratonic upper mantle are compositionally buoyant compared to the convecting upper mantle.This supports the idea that even diamond-bearing low temperature peridotite xenoliths were residues of partial melting which ended at pressures less than 30 kbar. Kelley and Kelemen are making calculations to support this alternative idea, for Kapvaal data from Boyd (pers. comm. 1998) and also for an extensive data set from Eggler (pers. comm. 1998). We will report on the progress of most or all of these efforts at the Lherzolite Conference.

Mantle heterogeneity is very important for magma genesis within the upper mantle. Melting of the heterogeneous mantle consisting of peridotite and mafic (or pyroxenite/ eclogite) rocks will yield voluminous and compositionally diverse magmas upon a melting process because of selective fusion of the mafic layers (e.g., Yasuda et al., 1994; Hauri, 1996; Takahashi et al., 1998). Melting experiments using heterogeneous mixtures of peridotite and MORB as a starting material, for example, have just started to examine this problem (Yaxley and Green, 1998; Kogiso et al., 1998). Mantle heterogeneity demonstrated by layered structure of peridotites and mafic (or pyroxenite/eclogite) rocks is conspicuous in many orogenic lherzolite massifs. We need to know how and when the inhomogeneous mantle materials have been produced. The Horoman Peridotite Complex, Hokkaido, northern Japan, has two specific features; one is the presence of symplectite, possibly of garnet origin, and the other is a symmetric layered structure characterized by an arrangement of cumulus peridotite and mafic rock in the middle of a series of residual peridotite with increasing melt component outward. The symmetrical layered structure repeats several times with intervals of several meters to a few hundred meters in the complex (Niida, 1984; Obata and Nagahara, 1987; Takahashi, 1992). Symplectite-bearing rocks in the Horoman complex are divided into three types based on petrography. The first, and the most abundant, is a kind of mantle restite, that is clinopyroxene- rich spinel lherzolite and plagioclase lherzolite. The second is a pyroxenite alternating with cumulus peridotites. The third is pyroxenites, which occur as thin layers in the mantle restite. Mineral assemblages and chemical compositions of the symplectites suggest that they were generally formed by the decompression reaction between pyrope-rich garnet and olivine. The presence of symplectite in cumulus peridotite and pyroxenite suggests the garnet was involved in the formation of cumulates at about 2 GPa or more. The mafic rocks in the Horoman complex have been divided into several types. One of these mafic layers, which is called Type II layer of Takazawa et al. (in press) or GB II of Niida (1984) and Shiotani and Niida (1984), restrictedly occurs in the cumulus peridotite which is located in the middle of residual peridotite with the symmetric layered structure. Some textural characteristics of the Type II mafic rocks are similar to those of the symplectite-bearing pyroxenites in the cumulus peridotite. The Type II mafic rocks have the same origin as symplectite-bearing pyroxenites, that is the subsolidus breakdown product of garnet-bearing pyroxenites of high-pressure origin to the gabbroic rock at lowerpressure conditions. On the other hand, their geochemical signatures indicated that the Type II mafic rocks were originally formed at lower-pressure conditions (Shiotani and Niida, 1997; Takazawa et al., in press). Textural characteristics of a corundum-bearing Type II mafic rock (Morishita and Kodera, 1998) show that corundum was not stable at the latest P-T conditions of the Horoman complex and require that it had experienced heating and/or decompression. A possible P-T history for the Type II mafic rock is as follows. (1) Type II mafic rock was formed as a cumulate at lower-pressure conditions from the melts responsible for the formation of the cumulus peridotite. (2) The protolith of Type II mafic rocks had been metamorphosed to garnetbearing pyroxenite at high P-T conditions during compression due to subduction or convection within the mantle. (3) The complex ascended from the garnet stability field to the plagioclase peridotite stability filed as a diapir. The Type II mafic rocks as a member of the diapir were formed from garnet-bearing pyroxenite through symplectite-bearing rock due to breakdown of garnet and corundum at low pressures. The Type II mafic rocks have a complex P-T trajectory after it was formed as a member of the layered structure. We favor the possibility that the symmetrical layered structure in the Horoman complex have been repeated by deformation processes (Toramaru, 1997), not by the melting process along multiple parallel cracks (Takahashi, 1992). The melting process, however, had an important role in formation of a stratified lithological unit composed of cumulate rock, residual peridotite and primitive peridotite.

UHP GARNET ORTHOPYROXENITES (DABIE SHAN, CHINA) AS MONITORS OF PERCOLATION OF HYDROUS SI-RICH MELTS IN DEEP SUBDUCTION ENVIRONMENTS

The Erro-Tobbio (ET) ophiolitic peridotite (Voltri Massif – Liguria, Italy) represents a sector of subcontinental lithospheric mantle that has been emplaced at crustal, sub-oceanic levels during rifting and opening of the Jurassic Ligurian Tethys (Ernst and Piccardo, 1979). Structural and petrologic works (e.g. Drury et al., 1990; Vissers et al., 1991; Hoogerduijn Strating et al., 1993) have demonstrated that the Erro-Tobbio peridotites were uplifted along a subsolidus P-T trajectory, characterized by progressively decreasing temperature. Pristine granular mantle protoliths, completely equilibrated at subcontinental lithospheric mantle depths (i.e. 1000-1100°C and spinel-facies conditions), were deformed along km-scale extensional shear zones, where they were transformed to spinel peridotite tectonites, spinel- and plagioclase-bearing mylonites, hornblende/ chlorite peridotite mylonites and, finally, serpentine mylonites. This composite P-T evolution has been interpreted as the exhumation trajectory of mantle sections evolving as the footwall of an asymmetric extensional system dominated by simple shear mechanisms. Recent contributions have revealed the presence of: i) reactive spinel peridotites, that were formed by melt-peridotite interaction under spinel-facies conditions (Piccardo et al., 2004; Rampone et al., 2004); ii) large areas of melt impregnated, plagioclase-rich peridotites, that are cut by a network of replacive spinel dunite channels and gabbroic dikelets (Piccardo et al., 2004). Our ongoing field, microstructural and geochemical investigations allow: i) to evidence and document some main steps in the composite evolution of the Erro-Tobbio peridotite, which are characterized by significant melt-peridotite interaction, and ii) to reconstruct the time-space relationships between these melt-related events and the subsolidus tectonicmetamorphic evolution of the Erro-Tobbio peridotite. Pristine mantle protoliths are sporadically preserved in the Erro-Tobbio massif: they are moderately depleted lherzolites, showing complete recrystallization under spinel-facies conditions (T below 1100°C; P below 2.5 GPa, according to Hoogerduijn Strating et al., 1993), which has been related to the accretion of pristine asthenospheric masses to the mantle lithosphere. They have relatively Na-Al-rich clinopyroxenes and preserve structural relics of their previous evolution, i.e. rounded opx+sp clusters, suggesting spinel-facies breakdown of a precursor garnet, i.e. a pristine garnet-bearing protolith. These spinel-facies lithospheric peridotites were subsequently affected by extensional deformation which formed km-scale shear zones, consisting of spinel peridotite tectonites and mylonites (Hoogerduijn Strating et al., 1993). In the field, spinel tectonite peridotites are replaced by coarse granular spinel peridotites, which: i) show microstructural (i.e. pyroxene dissolution and olivine precipitation) and compositional features indicating their reactive origin, due to interaction with percolating pyroxene-undersaturated melts; ii) preserve microstructural features (i.e. the opx+sp clusters), which indicate that they were formed by almost complete recovering of pristine lithospheric peridotites. Available clinopyroxene trace element compositions suggest complete trace element equilibration with the percolating melts, consisting of depleted melt increments formed by 6% of fractional melting of a DMM asthenospheric mantle source. Both tectonite and coarse granular spinel peridotites are replaced by plagioclase-rich peridotites, showing rather sharp contacts with the spinel peridotites. Plagioclase peridotites show microtextural and compositional characteristics [i.e.: i) opx replacement on mantle olivine, ii) mm-size noritic pods and veins, iii) opx+plg coronas replacing mantle cpx] which indicate melt/peridotite interaction and interstitial crystallization of pervasively percolating melts, having orthopyroxene(-silica)-saturated, clinopyroxene-undersaturated characteristics. Plagioclase peridotites frequently preserve microstructural features (i.e. olivine coronas replacing pyroxenes) which indicate that the pre-impregnation spinel peridotites frequently were represented by reactive peridotites. Geochemical modeling indicates that melts which percolated and impregnated the Erro-Tobbio spinel peridotites had a strongly depleted signature: they, most probably, were formed as depleted melt increments by fractional melting and attained orthopyroxene(-silica)-saturation during reactive migration in the lithospheric mantle column. Presence of channels of replacive spinel dunites, cutting both spinel and plagioclase peridotites, indicates that: i) further upwelling melts were forced to migrate within focused channels where both ortho- and clinopyroxenes were completely dissolved by reaction with pyroxene-undersaturated melts, and ii) these high permeability channels allowed more “rapid” migration of melts. Frequently coarse granular dunites replace spinel peridotite mylonite bands, suggesting the close relationships between active deformation and focused melt migration. A subsequent compaction of the dunite channels squeezed out the migrating melts, forming cm-size gabbronoritic dikelets. Geochemical modeling indicates that melts which crystallized in the gabbroic dikelets have a strongly LREE depleted signature, and have been formed as depleted melt increments by fractional melting of a DMM asthenospheric mantle source. Subsequently, the Erro-Tobbio peridotite underwent a composite tectonic-metamorphic evolution under lowered pressure conditions, were intruded by shallow MORB gabbroic bodies and basaltic dikes and, finally, were exposed at the sea-floor of the Jurassic Ligurian Tethys basin. This composite scenario of tectonic-metamorphic and melt-related events suggests the the Erro-Tobbio peridotites, after their accretion to the thermal lithosphere, were progressively exhumed to shallower levels, during lithospheric extension. As a whole, field relationships between rocks recording melt-peridotite interaction and subsolidus tectonic- metamorphic evolution indicate that the melt percolation processes, forming reactive spinel peridotites, impregnated plagioclase peridotites and replacive dunites, postdate the early stages of extension, under spinel-facies conditions, of the lithospheric mantle. During extension, when the rising asthenosphere began to melt, the overlain extending mantle lithosphere were percolated by depleted melt fractions. Early reactive percolation formed spinel-facies reactive peridotites, subsequent interstitial crystallization at shallower conductive levels formed impregnate plagioclase peridotites, whereas further upward migration of depleted melt fractions was forced within high permeability channels of replacive dunites. Abundance of reactive spinel peridotites and impregnated plagioclase peridotites suggest that asthenosphere/lithosphere interaction by melt percolation modified the compositional and rheological characteristics of large sectors of the lithospheric mantle during exhumation related to pre-oceanic lithosphere extension. The thermal softening of the mantle lithosphere could be an important factor in the dynamics of the extensional system during transition from passive lithosphere stretching to active oceanic rifting (Ranalli et al., 2004).

The compositional and mineralogical distinction between mantle lithosphere underlying old continental cratons and that beneath orogenic belts and ocean basins is well established (Boyd, 1989). The origin of many orogenic and oceanic peridotites as residues of partial melting at pressures less than 3 GPa in the spinel stability field is generally accepted. At such pressures, residues will be Ol-rich. In contrast, the highly depleted character, yet relatively Olpoor (and Opx-rich) mineralogy recognized in xenoliths from many cratons require that such lithosphere is not a simple residue. Hypotheses for the formation of cratonic lithosphere were reviewed by Kelemen et al. (1998) who provided compelling evidence that most of the mantle lithosphere beneath cratons originates by partial melting in the spinel stability field, and that cratonic garnet peridotites with higher equilibration pressures have been tectonically transported to such depths. In this way, ‘on-craton’ lithosphere could be considered oceanic peridotite that has underplated to form Archean continents with the deep lithospheric keels. Nonetheless, of the hypotheses reviewed by Kelemen et al. (1998), none have convincingly tied the petrological origin of cratonic mantle lithosphere to a specific tectonic setting. It has been speculated that Opx-enrichment in cratonic mantle resulting from melt/rock reaction would proceed in a subduction zone, whereas the same feature would occur from high degrees of melting and related cumulate enrichment in the root of a mantle plume. Distinct chemical or mineralogical lines of evidence to link the mantle samples to these tectonic environments is so far lacking. When interpreted in the light of a growing experimental data base for V partitioning (Canil, 1997; 1999) the systematics for V in cratonic and oceanic mantle help to further distinguish the conditions and the tectonic setting during melt depletion to form these two contrasting types of lithosphere. Examination of a data base of 200 spinel- and garnet peridotite analyses of on- and off-craton mantle xenoliths, and orogenic massifs from various settings show that the V abundances in mantle lithosphere lies along two distinct coherent trends (see Figure) when plotted against a depletion index such as wt% MgO. For a given degree of depletion, the V abundances of all cratonic peridotites, whether spinel- or garnet-bearing, are at lower levels than those of oceanic peridotites. The covariation of V with depletion for oceanic peridotites is explained (and modelled) as partial melting in the spinel stability field. Spinel has a DV that is ~ 10 times higher than any other peridotite mineral at a given oxygen fugacity (fO2), and as a residual phase during partial melting at low/moderate pressures, and fO2’s near that of MORB generation (NNO- 2 or -3), will dominate the budget for V. On the other hand, the spectrum of cratonic mantle compositions in the Figure requires either of the following: (1) Oceanic and cratonic peridotites represent residues from two completely different primitive mantle (PUM) sources; the source for cratonic mantle having a lower overall V abundance relative to PUM. (2) Unlike oceanic peridotites, cratonic peridotites are residues of melting at high pressures in the garnet stability field. Because DV Gt is less than DV Sp, less V is retained in the residue for a given degree of depletion. Quite possibly, neither garnet nor spinel are residual in the original melt depletion events related to cratonic peridotites, and both pyroxene and/or olivine, having low DV are responsible for the smaller amount of V in the residues. (3) Cratonic mantle originates by depletion in the Sp stability field at low pressure where Sp is a residual phase, but at much higher fO2 than oceanic peridotites related to MORB generation. The DV for Sp is very fO2 sensitive and varies almost an order of magnitude over the range of terrestrial fO2’s. The lower DV Sp during melting at higher fO2 (near NNO) would result in less V being retained in the residue. Hypothesis (3) would constrain the formation of cratonic peridotites to a region of melt generation in the mantle of relatively high FO2, most likely related to subduction. The distinctly lower Ti/V ratios recognized in many arc-related tholeiites (Shervais, 1982) are a complimentary signature related to this process. The above hypotheses will be quantitatively described and explored using appropriate peridotite melting models and partition coefficients. If hypothesis (3) is accepted then partial melting in the Archean was more oxidized than hitherto realized, some form of subduction was operative, and the mantle lithosphere beneath Archean cratons is intimately related to the subduction process.

Relic majoritic (or super-silicic) garnets were recently discovered within orogenic garnet peridotites at the island of Otroy, Western Gneiss Region, W. Norway (Van Roermund and Drury, 1998; Van Roermund et al., 1999). However within the Otroy garnet peridotites relic majoritic garnet microstructures are rare and the petrogenetic relationship( s) between relic majoritic- and “normal” garnet is unknown. For this reason we have made a detailed SEM study of microstructures and solid-phase inclusions present exclusively within all garnets from the Otroy garnet peridotites. MAJORITIC GARNET MICROSTRUCTURES AND CHEMISTRY The super-silicic garnet microstructure is characterised by the following microstructural elements: a) Majoritic garnet nodules consist of polycrystalline garnets ranging in size from 2 to 8 mm. Garnet-garnet grain boundaries are straight or gently curved, triple point junctions are common, interstitial orthopyroxenes decorate some garnet-grain boundaries. b) In the larger garnet grain cores two pyroxene exsolution needles occur 5-10 mm thick and oriented // grt. The relative vol% between exsolved cpx and opx is 1: 9. c) Two-pyroxene-bearing garnet-cores are surrounded by 2 mm thick precipitation-free rims. Backscattered scanning electron microscopy was used to estimate the amount of majorite component present in garnet nodules before exsolution. Large interstitial orthopyroxenes at garnet grain boundaries comprise up to 3.6 vol.%. In the garnet grain cores the maximum pyroxene content is 1 vol.%. Including both types of pyroxene-exsolution the maximum amount of exsolved pyroxene is 3.6-4.0 vol.%. This corresponds to pressure estimates around 6-6.5 GPa and a minimum depth of origin for the Otroy garnet peridotites in the range of 185-200 km. Microprobe analyses of exsolved pyroxenes and garnethost reveal homogeneous mineral compositions, except for Al2O3 wt% in opx. The most dominant Al2O3 wt% value in opx is 0.7-0.8. Estimated PT conditions for mineral-chemical equilibration, using standard geothermobarometric methods, indicate upper mantle conditions around 805 ± 40°C and 3.2 ± 0.2 GPa. These PT estimates correspond to cold continental lithosphere conditions at depths of around 105 km. From a combination of both depth estimates it can be concluded that the microstructural memory extends backwards to twice as great a depth-range than can be obtained by thermobarometric methods. In addition to the exsolved two pyroxenes we have found native nickel-particles as solid inclusions in relic majoritic garnets. This is indicative for relatively reduced oxygen fugacity conditions. The nickel particles are surrounded by complex symplectitic coronas involving Ni-Al oxide, Orthopyroxene, SiO2 and Ni-bearing garnet. NON-MAJORITIC GARNET MICROSTRUCTURES AND CHEMISTRIES The following solid-phases occur as inclusions in “normal” pyrope-rich garnets: a) Ni-Fe-Cu sulfides. Two types occur: 1) co-linear trails of rounded particles less than 1 mm in size and interpreted to represent precipitation along healed cracks. 2) Coarsened irregular-shaped inclusions up to 50 mm in size. The latter are often present in clusters. b) (Fe-) Ti (oxide??) needles (oriented // grt). c) (Fe-) Ti -oxides randomly precipitated or decorating dislocations. d) Isolated spinel needles and precipitates. With respect to the garnet host two types occur: oriented and randomly distributed crystals. e) Recrystallised garnet 2 - spinel - cpx/amphibole - orthopyroxene (± Ni-Fe-Cu sulfides) mineral assemblages. In addition to the various solid-garnet-inclusions garnetdeformation microstructures are present including naturally decorated dislocations. In thin-section garnet is often heavily fractured, and a penetrative crack-spacing down to the micron-scale is not uncommon. More widely-spaced or more heterogeneously fractured garnets exhibit the following microstructural domains: i) in between the cracks garnet is geometrically strained and deformation-induced microstructures such as undulous extinction and cross-hatched deformation bands are visible. This is recognised by the local non-isotropic nature of garnet. Within some of the strained domains dislocations are identified due to the fact that micron-sized particles have been precipitated along the dislocation cores. TEM and attached analytical facilities demonstrate that the solid inclusions are titanium-oxides, most likely rutile. ii) In between the cracks garnet is fully isotropic and strainfree. Within such domains randomly oriented solid inclusions are recognised one order of magnitude larger than the precipitates that decorate dislocations in strained garnet domains. All solid garnet inclusions and deformation microstructures predate the well-known two-pyroxene-spinel corona’s that surround garnet and form due to the instability of the mineral assemblage garnet-olivine. All garnet microstructures are consistent with a polycyclic exhumation model involving diapiric upwelling of deep-mantle peridotites to lithospheric depths followed by isobaric-cooling down to normal continental lithosphere conditions. Crustal incorporation occurs during renewedsubduction of continental crust during continental collision subsequently followed by final exhumation.

In the search for a reliable indicator of chemical processes in the Earth’s mantle, we have undertaken a study of the Lithium contents in mantle minerals. The motivation behind this investigation is that Li is a relatively mobile alkali metal on the one hand, but its small ionic radius makes it only a moderately incompatible element (e.g., Blundy & Wood, 1994). Such characteristics could make Li a sensitive indicator of metasomatic or magmatic processes. To test this hypothesis, we have measured Li concentrations in orthopyroxene, clinopyroxene, olivine, garnet and spinel from two suites of spinel peridotites, garnet peridotites and garnet pyroxenites; one suite comprises xenoliths that are very well equilibrated in terms of major and a number or minor and trace elements, and the other suite is composed of samples for which there is clear evidence of metasomatism. The metasomatised samples are from several orogenic lherzolite massifs in the French Pyrenees and the Ivrea Zone (Italy) as well as being xenoliths from Victoria, Australia. The Li concentrations were measured using an ion microprobe (SIMS). The suite of equilibrated samples indicates that Li is preferentially incorporated into olivine, (1-2 ppm) with lesser concentrations (100s of ppb) in both clinopyroxene and orthopyroxene. The Li contents of garnet and spinel are at or near to the detection level (» 10 ppb). The following partitioning relationships are valid for peridotite and pyroxenite bulk compositions respectively: ol>cpx³opx>>gt,sp and cpx³opx>>gt. Of particular importance is the observation that the intercrystalline partitioning of Li between these phases is basically independent of temperature and pressure and chemistry, within the limits of such mantle bulk compositions. Combining our results with modal abundances for the samples, the bulk Li content for fertile to moderately depleted lithospheric mantle is 1-1.5 ppm, which is in excellent agreement with other previous estimates for primitive mantle (Jagoutz et al., 1979; Frey et al., 1985; Ryan and Langmuir, 1987; O’Neill, 1991; McDonough and Sun, 1995). Both the peridotites and pyroxenites display variations in absolute Li abundances, although the partitioning relationships between olivine, clinopyroxene and orthopyroxene are maintained. Lower overall Li abundances in peridotites are consistent with depletion from partial melting and subsequent melt extraction. The pyroxenes in the pyroxenites have variable Li contents with values typically of 1-3 ppm. This is consistent with their origin as cumulates. Some samples have elevated Li contents, which could be due to a higher degree of trapped melt present during crystallisation. Thus the greater variation in Li contents in the pyroxenites can be explained by variable degrees of trapped melt associated with the cumulate crystals, which themselves have relatively low Li concentrations due to the expected melt/solid partitioning behaviour of Li. The Li contents of the metasomatised samples are generally higher than those observed in the equilibrated samples and usually display much more variability within a single sample. Olivine and/or clinopyroxene can contain in excess of 5 ppm Li in these metasomatised samples. Modal metasomatic phases, such as amphibole and phlogopite also contain Li, but often in concentrations less than in the coexisting olivine or clinopyroxene. Analysis of a hornblendite dike from the Lherz massif indicates that Li is preferentially partitioned into phlogopite compared with amphibole. The most notable feature of the metasomatised samples is that most record a disequilibrium distribution of Li between olivine and the coexisting pyroxenes. Preferential enrichment of Li can occur in either olivine or clinopyroxene apparently depending on the type of metasomatising agent; olivine is preferentially enriched in samples from Victoria, which have been affected by carbonatitic melts and clinopyroxenes are preferntially enriched in samples that have interacted with mafic silicate melts. Samples from Finero (Italy), that apparently interacted with fluids, also Li distributions indicative of moderate disequilibrium. The partitioning of Li between olivine and the pyroxenes provides an additional indicator for recognising cryptic metasomatism. Given the correct conditions (a high enough temperature, long enough time for reaction etc..), an equilibrium distribution can be achieved in metasomatised samples, however, these will still display an overall enrichment in Li compared to ‘normal’ mantle values.

The easternmost end of the Periadriatic Lineament (PAL) is situated at the complex Alps-Dinarides-Pannonian Basin triple junction in Slovenia and Croatia. Tonalitic plutonism ceased along the PAL as a whole at the end of Oligocene time but not at its easternmost segment, where it continued as volcanism on both sides of the PAL in Early-Middle Miocene times. The paper deals with geochemical modeling of PAL magmatism, in order to put forward a viable petrogenetic model to explain the continuation, compatible with recent geotectonic interpretations in the area. Geochemical modeling of major and trace elements recognizes two groups of volcanic rocks, the Northern and the Southern one, geographically separated by the PAL. Geochemical trends broadly resemble those of the back-arc-trench- arc duality. The magma source was chemically enriched lithospheric mantle, metasomatized during an earlier subduction- related episode. The Northern group was formed by lower degree melting of garnet peridotite, whereas primary magmas of the Southern group were formed by extensive melting of garnet and/or spinel peridotite, developing a tholeiitic-calc- alkaline character. In Miocene times, melting of the mechanical enriched lithospheric mantle was driven by still unobstructed rise of the hot, convective astenosphere, after detachment of the subducted slab and steady lithospheric attenuation and extrusion of ALCAPA. In contrast, in the central portion of the PAL, continuing collision restrained magmatic activity. Different depth of melting arises from different thickness of the lithospheric blocks, south and north of the PAL.

ENRICHMENT PROCESSES IN GARNET-BEARING MANTLE XENOLITHS FROM KIMBERLEY PIPES (SOUTH AFRICA)

The Erro-Tobbio (ET) peridotites (Voltri Massif, Ligurian Alps) represent subcontinental lithospheric mantle tectonically exhumed during Permo-Mesozoic extension of the Europe-Adria lithosphere. Previous studies have shown that exhumation started during Permian times, and occurred along km-scale lithospheric shear zones which enhanced progressive deformation and recrystallization from spinelto plagioclase-facies conditions (Hoogerduijn Strating et al., 1993; Rampone et al., 2005a). Ongoing field and petrologic investigations have revealed that the peridotites experienced, during uplift, a composite history of diffuse melt migration and multiple episodes of ultramafic-mafic intrusions (Piccardo et al., 2004; Borghini et al., 2005), which record their progressive exhumation from deep lithospheric depths to the sea floor. In this paper we present the results of field, structural and petrologic-geochemical investigations into a sector of the Erro-Tobbio peridotite unit which preserves this multiple intrusion history. In the investigated area, peridotites are mostly constituted by low-strain spinel tectonites. The oldest intrusion event recorded is represented by the diffuse occurrence of centimeter- scale pyroxenite bands. They often display tight isoclinal folds which are crosscut at medium-high angle by the mantle tectonite foliation. This indicates that their primary intrusion relationships have been almost completely transposed by old deformation events. As already documented by Hoogerduijn Strating et al. (1993), these features suggest that pyroxenite intrusion represents an old deep-seated magmatic event, which preceded the lithospheric exhumationrelated evolution of the ET mantle. Pyroxenites consist of a coarse-grained primary mineral assemblage, made by spinel and variable amounts of clinoand ortho-pyroxene, their modal compositions thus ranging from clinopyroxenites to spinel-websterites. The spinelbearing mineral association is severely recrystallized to a plagioclase-bearing assemblage. This is recorded by the crystallization of plagioclase + olivine aggregates around relict Cr-rich spinels, and by the development of large opx+plag exsolutions in spinel-facies pyroxenes. Preliminar geochemical investigations on two pyroxenite samples have revealed that clinopyroxenes hold an unusual trace element signature, being characterized by very high Sc (120-155 ppm) and V (765-913 ppm) contents, and strongly fractionated REE spectra, with very low MREE-HREE ratios (GdN/YbN = 0.39-0.52) and HREE concentrations up to 20xC1. Similar compositional features were documented by Vannucci et al. (1993) in clinopyroxenes from plagioclasereequilibrated spinel pyroxenites of Zabargad (Red Sea), and they were interpreted to be inherited from a precursor garnet-bearing primary magmatic assemblage. The investigated ET pyroxenites also show significantly fractionated bulk-rock REE spectra, with marked LREE depletion (CeN/YbN = 0.049), absent EuN anomaly, and HREE abundances at 7xC1. All these features point that pyroxenites originated as high-pressure cumulates. Microstructural and geochemical characteristics suggest that plagioclase formation in the studied pyroxenites was mostly driven by subsolidus reaction. The likely occurrence of garnet in the primary magmatic assemblage constrains the depth of intrusion and crystallization to P > 15-20 kbar (Hirschmann and Stolper, 1996; and quoted references). Subsequent decompression of the ET pyroxenite-peridotite association is documented by the recrystallization of the pyroxenite bands to spinel- and plagioclase-bearing assemblages. At shallower lithospheric levels, the ET peridotites were diffusely migrated and impregnated by melts. Melt impregnation is documented by significant enrichment of interstitial plagioclase between mantle minerals, and by the crystallization of unstrained poikilitic orthopyroxene replacing deformed tectonitic mantle olivine and exsolved clinopyroxene: this indicates that the impregnating melts were opx-saturated. Melt-rock interaction caused chemical changes in mantle minerals (e.g. Al decrease and REE increase in cpx; Ti and Cr# enrichment in spinel). Reacted clinopyroxenes, in spite of the overall increase in the REE contents, still exhibit strong LREE depletion (CeN/SmN =0.006-0.011), indicating a depleted signature for the percolating melts. Melt impregnation was thus related to diffuse porous flow migration of opx-saturated depleted MORB-type melt fractions, as inferred by Piccardo et al. (2004). The impregnated peridotites are intruded by a hectometre- scale stratified cumulate ultramafic body, mostly consisting of troctolites and wehrlites, showing gradational, interfingered contacts with the host mantle rocks. Subsequent intrusion events are revealed by the occurrence of olivine gabbros as decameter-wide lenses, variably thick (cm- to mscale) dykes and thin dykelets, which crosscut both the peridotite foliation and the magmatic layering in the cumulates. Overall, major and trace element compositions of minerals in the intrusives indicate that they represent variably differentiated cumulus products crystallized from rather primitive N-MORB-type aggregated melts. Slightly more evolved compositions are shown by olivine gabbros, relative to the troctolites and wehrlites of the ultramafic body. Peculiar mineral chemistry features (e.g. the Fo-An correlation and high Na, Ti, Mg# in cpx) indicate that the studied intrusive rocks crystallized at moderate pressure conditions (3-5 kbar, i.e. 9-15 km depth). Sm-Nd isotope data on two olivine gabbros have yielded a magmatic crystallization age of 180 + 14 Ma (Rampone et al., 2005b). Chemical and petrologic characteristics of ultramafic and gabbroic rocks point to compositional differences between their parental melts (olivine-saturated, N-MORB-type aggregated melts) and melts which impregnated the host peridotites (orthopyroxene- saturated single depleted melt increments). Field and geochemical evidence thus indicate that peridotite impregnation and cumulate intrusion represent unrelated events. Peridotite impregnation was caused by diffuse migration of single depleted melt fractions and occurred before the intrusion of MORB-type aggregated melts. The transition from porous flow melt migration to emplacement of magmas in fractures, most likely reflects progressive change of the lithospheric mantle rheology during extension-related uplift and cooling of the ET mantle.

Highly refractory peridotites are defined here as peridotites more depleted in melt components than abyssal peridotite, i.e. peridotites with spinel showing Cr# (Cr/(Cr+Al) atomic ratio) higher than 0.6. These correspond to the type III peridotite defined by Dick and Bullen (1984) and are reported from forearc regions, such as Izu-Ogasawara, Mariana and Tonga trenches (e.g. Bloomer and Hawkins, 1983; Bloomer and Fisher, 1987; Ishii et al., 1992) and from the Kamuikotan belt, the central axial zone of Hokkaido, northern Japan (e.g. Katoh and Nakagawa, 1986; Makita and Arai, 1997). These peridotites are, however, not common in the upper mantle and their genesis has not been explained thoroughly. We discuss the genesis of highly refractory peridotite from three complexes, the Takadomari, Iwanaidake and Nukabira, in the Kamuikotan belt, Japan, and its bearing on the process of high-Mg andesite magma genesis. The Kamuikotan belt is a tectonic melange zone composed of metamorphic rocks, ultramafic rocks, greenstones and sedimentary rocks. The metamorphic rocks are of typical high-pressure/low-temperature type and ophiolitic rocks which recorded low-pressure ocean-floor metamorphism are distributed in the Horokanai area, northern part of the belt (Ishizuka et al., 1983). Ultramafic rocks form the basal member of the ophiolite in the northern part and are exposed as peridotite complexes of various size in the southern part of the Kamuikotan belt. They are generally in fault contact with surrounding Kamuikotan metamorphic rocks, sediments and mafic volcanic rocks (e.g. Niida and Katoh, 1978). The peridotite complexes mainly consist of harzburgite and dunite which suffered serpentinization to various extent. The Takadomari complex from the northern part, and the Iwanai-dake complex from the southern part of the belt consist of harzburgite, dunite and small amounts of orthopyroxenite. The Nukabira complex (southern part of the belt) consists of lherzolite and harzburgite with dunite and pyroxenites. Podiform chromitite deposits sometimes accompany dunite in the latter two complexes. Primary hydrous minerals are sometimes included in chromian spinel of the former two complexes. The Cr# of chromian spinel and Fo (forsterite) content of olivine in peridotites from the Takadomari complex are high (Cr#=0.64-0.92, Fo=91.9- 94.0), and are higher in dunite than in harzburgite on average. The Cr# and Fo are also high (Cr#=0.43-0.87, Fo=90.8-93.5) in the Iwanai-dake complex and have wider ranges (Cr#=0.18-0.86, Fo=89.0-93.2) in the Nukabira complex. The Cr# of spinel and Fo of olivine in dunite complexes are sometimes similar to or even lower than the values in harzburgite from Iwanai-dake and Nukabira complexes. The mineral chemistry of dunites in the Iwanai-dake and Nukabira complexes depends on their thickness within harzburgite or lherzolite bodies. Thicker dunite layers tend to show higher Cr# of chromian spinel and Fo content of olivine. In addition to this, there are clear systematic variations in lithology across dunite layers and surrounding peridotites. In the case of thick dunite layers, the Cr# of chromian spinel and the Fo content of olivine gradually increase from the surrounding peridotite to the central part of the dunite layer (Cr#>0.6 and Fo>91.5 in dunite). In contrast to this fact, these values in thin dunite layers are similar to the surrounding peridotite, lherzolite or harzburgite (Cr#=0.4- 0.6 and Fo=89-91). These values are generally lower in thinner dunite layers than in the thick ones compared at the center. The harzburgites and dunites in the Takadomari complex are interpreted to be a series of refractory residue after extraction of high-Mg andesite magma; partial melting of the fertile peridotite possibly occurred under hydrous conditions. Harzburgite and lherzolite are mostly simple residue in the Iwanai-dake and Nukabira complexes but their dunites are not simple residue because the Cr# and Fo content are not systematically higher in the dunite than in the harzburgite (e.g. Arai, 1987; 1994). The dunite may be produced by a reaction between wall peridotite and a melt formed at higher pressures (e.g. Quick, 1981; Fisk, 1986; Kelemen, 1990). The differences in thickness and mineral chemistry of dunite layers are due to the difference of the melt/wall peridotite volume ratio. Supplying a large amount of melt caused formation of thick dunite layers and generated high-Mg andesitic magma by the reaction process.

New evidence for ultrahigh‐pressure metamorphism (UHPM) in the Eastern Alps is reported from garnet‐bearing ultramafic rocks from the Pohorje Mountains in Slovenia. The garnet peridotites are closely associated with UHP kyanite eclogites. These rocks belong to the Lower Central Austroalpine basement unit of the Eastern Alps, exposed in the proximity of the Periadriatic fault. Ultramafic rocks have experienced a complex metamorphic history. On the basis of petrochemical data, garnet peridotites could have been derived from depleted mantle rocks that were subsequently metasomatized by melts and/or fluids either in the plagioclase‐peridotite or the spinel‐peridotite field. At least four stages of recrystallization have been identified in the garnet peridotites based on an analysis of reaction textures and mineral compositions. Stage I was most probably a spinel peridotite stage, as inferred from the presence of chromian spinel and aluminous pyroxenes. Stage II is a UHPM stage defined by the assemblage garnet + olivine + low‐Al orthopyroxene + clinopyroxene + Cr‐spinel. Garnet formed as exsolutions from clinopyroxene, coronas around Cr‐spinel, and porphyroblasts. Stage III is a decompression stage, manifested by the formation of kelyphitic rims of high‐Al orthopyroxene, aluminous spinel, diopside and pargasitic hornblende replacing garnet. Stage IV is represented by the formation of tremolitic amphibole, chlorite, serpentine and talc. Geothermobarometric calculations using (i) garnet‐olivine and garnet‐orthopyroxene Fe‐Mg exchange thermometers and (ii) the Al‐in‐orthopyroxene barometer indicate that the peak of metamorphism (stage II) occurred at conditions of around 900 °C and 4 GPa. These results suggest that garnet peridotites in the Pohorje Mountains experienced UHPM during the Cretaceous orogeny. We propose that UHPM resulted from deep subduction of continental crust, which incorporated mantle peridotites from the upper plate, in an intracontinental subduction zone. Sinking of the overlying mantle and lower crustal wedge into the asthenosphere (slab extraction) caused the main stage of unroofing of the UHP rocks during the Upper Cretaceous. Final exhumation was achieved by Miocene extensional core complex formation.

THE GENESIS OF PYROXENITE-RICH PERIDOTITE AT CABO ORTEGAL (NW SPAIN). INFERENCES FROM GEOCHEMICAL, MINERAL AND PB-SR ISOTOPIC DATA