The molybdenum (Mo) isotope composition (defined as S98Mo measured per mil relative to NIST-3134) of many modern arc systems and the upper continental crust is heavier than the mantle and most subducting slab lithologies. This observation has led to a model whereby fluids leaving the slab transfer isotopically heavy Mo preferentially to the mantle wedge, leaving the residual slab isotopically lighter. We explore this model via an Mo isotope study of the metasedimentary and melange lithologies of the Catalina Schist in California. These rocks record subduction zone metamorphism over a wide range of high-pressure/low-to-medium temperature conditions. Mo isotope compositions of the metasedimentary rocks decrease with increasing metamorphic grade, from a S98Momean of -0.04 + 0.13 %o (1a, n = 3) for the lawsonite albite facies to -0.38 + 0.07 %o (1a, n = 5) for the epidote-amphibolite facies. The highest grade [amphibolite facies] samples show a slight uptick in S98Mo values, with a mean of -0.19+ 0.14 %o (1a, n = 4). When coupled with major and trace element variations, the changes in S98Mo appear to reflect metamorphic effects rather than sedimentary source rock heterogeneity. The positive relationship between S98Mo and [Mo] and negative relationship with Ce/Mo argue for progressive loss of Mo with isotope fractionation during increasing P-T conditions; the reversal of this trend, seen in the increase in S98Mo between the epidote amphibolite and amphibolite facies may reflect a subsequent partial re-fertilization by fluids carrying isotopically heavy Mo. S98Mo values in the melange across all grades are highly variable, ranging from -1.75 to -0.19%o, consistent with large scale mobilization of Mo with isotope fractionation. A metasomatized amphibolite block and reaction rind show similar light isotope compositions and signs of Mo depletion and transport. Together these three whole-rock data sets demonstrate pervasive open system behavior and mobility of Mo in the Catalina Schist. LA-ICP-MS Mo measurement of minerals in the differing metasedimentary metamorphic grades indicates that Fe oxides and/or hydroxides, titanite, rutile, and epidote are important reservoirs of Mo but when comparing mineral Mo content and modal abundance to whole-rock Mo concentrations, several samples were found to have "missing Mo", something that has been observed in other Mo inventory studies. We attribute this to sample heterogeneity between the thin section and whole-rock powder scales.

We investigated nitrogen (N) sources and incorporation in the shallow forearc of an active subduction zone by studying samples from two serpentinite seamounts (South Chamorro and Conical; Ocean Drilling Program Legs 195 and 125) along the Mariana convergent margin. We report N concentrations and delta N-15 values for serpentinized peridotites, serpentinite muds, and metabasaltic clasts. All are enriched in N relative to likely ultramafic and mafic protoliths. A modest positive correlation between N and delta N-15 for the serpentinized peridotites could reflect extents of N infiltration by N-bearing fluids. Mixing calculations identify the source of the fluids as dehydrating metasediments and/or altered oceanic crust from the subducting Pacific slab. Such addition is consistent with that of other fluid mobile elements (B, Li, As, and Cs), but their concentrations show little correlation with N. The general lack of correlation in concentrations of N and major and trace elements complicates the identification of mineral hosts and N residency (e.g., loose adsorption on serpentine minerals). The enrichment of N in these samples indicates that serpentinized mantle wedges could be a globally significant N reservoir. If subducted, this reservoir could provide significant amounts of isotopically heavy N to sub-arc depths and beyond.

The ultramafic rocks of the Ulten Zone tectonic melange experienced a well-documented multi-stage metamorphic history, from residence in a hot mantle wedge to down-dragging and trapping in a Variscan slab accompanied by amphibolitisation, and finally two-stage exhumation accompanied by chloritisation and serpentinisation. We use these rocks as a natural laboratory to investigate whether volatile element fluxes in continental subduction zones promote long-term volatile element storage in the overlying mantle wedge. Here, we obtained new data on the chemical composition, iron speciation, carbon concentrations and isotopic compositions of ultramafic bulk rocks, and the carbon-oxygen isotope composition of carbonates in samples from >10 ultramafic lenses, which we combine with previously published data for additional insights. The carbonate stable isotope compositions show a distinct provinciality, whereby rocks from the little retrogressed ultramafic lenses in the NE Ulten Zone domain have lower average delta C-13(V-PDB) and delta O-18(V-SMOW) of -16.8 parts per thousand to -5.7 parts per thousand and +8.0 parts per thousand to +17.8 parts per thousand, respectively, than those in the more retrogressed SW domain (-11.2 parts per thousand to 0.0 parts per thousand and +12.9 parts per thousand to +20.7 parts per thousand, respectively), suggesting influx of distinct crustal fluids. Bulk-rock carbon contents range from 130 to 28,000 mu g g(-1), exceeding estimates for the convecting mantle, and are on average higher in rocks from the NE domain (median 880 mu g g(-1)), which can be modelled as Rayleigh-style dolomite addition at similar to 800-700 degrees C. Rocks from the SW domain have lower C contents (median 570 mu g g(-1)) which correlate positively with delta C-13 and can be modelled as Rayleigh-style calcite addition at similar to 500-400 degrees C. The lowest delta C-13 and C contents point to dedolomitisation during low-temperature (400 degrees C) serpentinisation, and furthermore suggest the contribution of a C-13-depleted phase to the bulk-rock compositions. After melt depletion during formation of the pre-Variscan continental lithosphere, the sulfur inventory was replenished during amphibolitisation near peak-metamorphic conditions, via sulphidation during interaction with siliceous fluid. Sulfur was markedly (re-)depleted during processes related to exhumation, reflecting low sulfur fugacity during chloritisation and serpentinisation. The available data suggest that the high bulk-rock Fe3+/Sigma Fe (median 0.18) resulted from reduction of some aqueous sulphate during amphibolitisation, accompanied by redox-neutral carbonation, and from carbonate reduction during chloritisation when sulfur fugacity was low. Ignoring exhumation-related C loss and taking near-peak metamorphic conditions as representative, significant amounts of C in carbonates and water in amphiboles may be stored in continental mantle wedges. These are subsequently stabilised below collisional orogens, which cover vast areas of Earth's continents and possibly constitute particularly volatile element-rich lithospheric mantle reservoirs.

Deep-Earth cycling of nitrogen (N) along the subduction pathway, remains relatively poorly understood, particularly that related to deep subduction of continental crust. The ultrahigh-pressure (UHP) Dora-Maira Massif whiteschists in the Western Alps and their lower -P-T equivalent leucophyllites in the Eastern Alps show a decrease in N concentration and an increase in 815N relative to their presumed country rock protoliths. On average, the whiteschists contain 20.7 +/- 10.5 ppm N (mean +/- 1a) with 815Nair = +2.7 +/- 2.5%o (mean +/- 1a) and the leucophyllites contain 59.8 +/- 12.3 ppm N with 815Nair = +4.3 +/- 2.0%o, whereas the metagranitic country rocks from Dora-Maira contain 41.3 +/- 12.5 ppm N with 815Nair =-1.9 +/- 3.2%o and the country rocks from the Eastern Alps contain 91.1 +/- 42.5 ppm N with 815Nair = +1.7 +/- 2.1%o. The loss of N and isotope shift in the whiteschists and leucophyllites from these two localities was accompanied by loss of LILE (Rb, Ba), Sr, CaO, Na2O, and FeOtotal. The other stable isotope systems applied to study both suites (Mg, Fe, O, Li and Ba) show no obvious correlation with variations in 815N. Infiltration by and interaction with a serpentinite-derived fluid depleted in N and LILE could have led to leaching of N from the country rock protoliths and production of the +4.6%o (Dora-Maira Massif) and +2.6%o (Eastern Alps) isotope shifts adequately explained using a Rayleigh distillation model. The N exchange and leaching process is best explained as involving N2-NH4 or NH3-NH4 at about 550-600 degrees C, the latter the approximate temperatures at which the fluid infiltration is thought to have occurred. These results provide new insights regarding N behavior and isotopic fractionation during fluid-rock interactions occurring in UHP rocks and point to the importance of H2O-rich fluids from serpentinite dehydration as agents of metasomatism along deep subduction interfaces. Furthermore, the data for the metagranites indicate the potential of this lithology to retain significant amounts of N to at least 100 km and beyond sub-arc depths.

Although rare earth elements (REE) are now considered as emerging contaminants, the mechanisms controlling the mobility of REE in geochemical systems remain elusive. The complexity and multi-element characteristics of REE including potential synergistic and antagonistic interactions with environmental surfaces make the prediction of REE fate in nature a challenging task. In this study, a comprehensive set of batch and column transport experiments were conducted to examine the interactions of REE group, as well as some co-occurring or chemically analogous elements (Sc, Y, Th and U), with 100–300 μm quartz sand particles. Results from batch experiments showed that middle REE (MREE) and heavy REE (HREE) are preferentially adsorbed at low and high REE loadings, which showed the occurrence of two different types of binding sites. A surface complexation model has been developed, which successfully predicted sorption of REE, Sc, Y, Th and U. Experimental data and reactive transport modeling evidenced the importance of the strong sites, and highlighted the competitive binding of MREE with other REE and Y for the quartz surface sites. These results may have strong implications for the development of new prediction tools for accurately assessing the reactive transport of REE in natural systems.

Mélanges are mixtures of subducted materials and serpentinized mantle rocks that form along the slab-mantle interface in subduction zones. It has been suggested that mélange rocks may be able to ascend from the slab-top into the overlying mantle, as solid or partially molten buoyant diapirs, and transfer their compositional signatures to the source regions of arc magmas. However, their ability to buoyantly rise is in part tied to their phase equilibria during melting and residual densities after melt extraction, all of which are poorly constrained. Here, we report a series of piston-cylinder experiments performed at 1.5–2.5 GPa and 500–1050 °C on three natural mélange rocks that span a range of mélange compositions. Using phase equilibria, solidus temperatures, and densities for all experiments, we show that melting of mélanges is unlikely to occur along the slab-top at pressures ≤ 2.5 GPa, so that diapirism into the hotter mantle wedge would be required for melting to initiate. For the two metaluminous mélange compositions, diapir formation is favored up to pressures of at least 2.5 GPa. For the peraluminous mélange composition investigated, diapir buoyancy is possible at 1.5 GPa but limited at 2.5 GPa due to the formation of high-density garnet, primarily at the expense of chlorite. We also evaluate whether thermodynamic modeling (Perple_X) can accurately reproduce the phase equilibria, solidus temperatures, and density evolution of mélange compositions. Our analysis shows good agreement between models and experiments in mélange compositions with low initial water contents and low-pressure (≤ 1.5 GPa) conditions. However, discrepancies between the thermodynamic models and experiments become larger at higher pressures and high-water contents, highlighting the need for an improved thermodynamic database that can model novel bulk compositions beyond the canonical subducting lithologies. This study provides experimental constraints on mélange buoyancy that can inform numerical models of mélange diapirism and influence the interpretations of both geophysical signals and geochemical characteristics of magmas in subduction zones.

We examined exposures of meta-ophiolitic breccias from the French/Italian Alps that experienced peak P-T conditions (1.0-2.4GPa; 300-550 °C) similar to those encountered in some modern subduction zones, with the goal of assessing degrees of carbon (C) loss via decarbonation reactions.The targets were contacts of clasts with carbonate cement at which calc-silicate phases growing as reaction rims could reflect decarbonation driven by interaction with H 2 O-rich fluids.A field, petrographic and isotopic approach for two of six localities studied thus far provides hints of such decarbonation.Carbonate C-O isotope data (87 microdrilled samples) for Lago Nero (LN) and Ubaye Valley (UV) suggest interaction with low-δ 18 O (likely H 2 O-rich) fluids as evidenced by δ 18 O VSMOW shifts toward (~ +14‰) from values for ocean-floor protoliths (~ +30‰).These shifts are consistent with interactions with fluids from mafic and ultramafic sources [1].A subset of samples from UV falls toward higher δ 18 O and much lower δ 13 C, showing a trend consistent with interaction with fluids from nearby calc-schists [2].The array of C-O isotope values for LN resembles a Rayleigh fractionation trend, perhaps an aggregate result of precipitation from H 2 O-rich fluids containing C released during decarbonation over larger volumes in the slab/interface section.Both localities show modest mineralogical evidence of decarbonation as reaction rims at the carbonate-silicate contacts, with pumpellyite, clinozoisite/epidote, and Ca/Na-amphiboles at metabasaltic clast rims at UV and tremolite/actinolite crystallized at rims of ultramafic clasts at UV and LN.These reaction rims are abundant but fine-grained (thicknesses of 20-200 μm), thus providing evidence of decarbonation insufficient to shift cement δ 13 C to lower values except at very local scales.Additional thermodynamic modeling (PERPLE_X) will reveal whether the calc-silicate phases observed are plausibly the results of decarbonation reactions produced at/near peak P-T by interaction of H 2 O-rich/low-XCO 2 fluids.Thus far, our data suggest that meta-ophiolitic breccias in the Western Alps interacted extensively with H 2 O-rich fluids resulting in modest decarbonation and impressive retention of carbonate in such breccias to depths of at least ~50km in the forearc.

Mars exploration is focused on seeking evidence of habitable environments and microbial life. Terrestrial glassy basalts may be the closest Mars‐surface weathering analog and observations increasingly indicate their potential to preserve biogeochemical records. The textures, major and trace element geochemistry, and N concentrations and isotopic compositions of subaerial, subglacial and continental lacustrine hyaloclastites from Antarctica, Iceland, and Oregon, respectively, were studied using micro‐imaging and chemical methods, including gas‐source mass spectrometry. Alteration by meteoric‐sourced waters occurred in circum‐neutral, increasingly alkaline low‐temperature conditions of ∼60°C–100°C (Iceland) and ∼60°C–170°C (Antarctica). Incompatible large ion lithophile element (LILE) enrichments compared to mid‐ocean ridge basalt (MORB) are consistent with more advanced alteration in Antarctic breccias consisting of heulandite‐clinoptilolite, calcite, erionite, quartz, and fluorapophyllite. Granular and tubular alteration textures and radial apatite represent possible microbial traces. Most samples contain more N than fresh MORB or ocean island basalt reflecting enrichment beyond concentrations attributable to igneous processes. Antarctic samples contain 52–1,143 ppm N and have δ15Νair values of −20.8‰ to −7.1‰. Iceland‐Oregon basalts contain 1.6–172 ppm N with δ15Ν of −6.7‰ to +7.3‰. Correlations between alteration extents, N concentrations, and concentrations of K2O, other LILEs, and Li and B, reflect the siting of secondary N likely as NH4+ replacing K+ and potentially as N2 in phyllosilicates and zeolites. Although much of the N enrichment and isotope fractionation presented here is not definitively biogenic, given several unknown factors, we suggest that a combination of textures, major and trace element alteration and N and other isotope geochemical compositions could constitute a compelling biosignature in samples from Mars' surface/near‐surface.

Lithium is of great interest as a tracer of metamorphic reactions and related fluid-mineral interactions because of its potential to isotopically fractionate during inter- and intracrystalline diffusional processes. Study of its transfer through subduction zones, based on study of arc volcanic and metamorphic rocks, can yield insight regarding ocean-to-mantle chemical cycling. We investigated major- and trace-element concentrations and delta Li-7 in garnet in ultrahigh-pressure (UHP) Lago di Cignana metasedimentary rocks, relating these observations to reconstructed prograde devolatilization history. In all garnet crystals we studied, heavy rare earth elements (HREEs), Y, and Li showed strong zoning, with elevated concentrations in cores (15-50 ppm Li) and marked high-concentration anomalies (up to 117 ppm Li, 5500 ppm Y; little or no major-element shift) as growth annuli, in which some crystals showed subtle elevation in delta Li-7 greater than analytical error of similar to 3 parts per thousand (2 sigma). Rutile inclusions appeared abruptly at annuli and outward toward rims, accompanied by inclusions of a highly zoned, Ca- and rare earth element-rich phase and decreased Nb concentrations in garnet. These relationships are interpreted to reflect prograde garnet-forming reaction(s), in part involving titanite breakdown to stabilize rutile, which resulted in delivery of more abundant Y and HREEs at surfaces of growing garnet crystals to produce annuli. Co-enrichments in Li and Y + REEs are attributed to mutual incorporation via charge-coupled substitutions; thus, increased Li uptake was a passive consequence of elevated concentrations of Y + REEs. The small-scale fluctuations in delta(7) Li (overall range of similar to 9 parts per thousand) observed in some crystals may correlate with abrupt shifts in majorand trace-element concentrations, suggesting that changes in reactant phases exerted some control on the evolution of delta(7) Li. For one garnet crystal, latestage growth following partial resorption produced deviation in major- and trace-element compositions, including Li concentration, accompanied by a 10 parts per thousand-15 parts per thousand negative shift in delta(7) Li, perhaps reflecting a change in the mechanism of incorporation or source of Li. These results highlight the value of measuring the major- and trace-element and isotope compositions of garnets in high-pressure and UHP metamorphic rocks in which matrix mineral assemblages are extensively overprinted by recrystallization during exhumation histories. Lithium concentrations and isotope compositions of the garnets can add valuable information regarding prograde (and retrograde) reaction history, kinetics of porphyroblast growth, intracrystalline diffusion, and fluid-rock interactions. This work, integrated with previous study of devolatilization in the Schistes Lustres/Cignana metasedimentary suite, indicates retention of a large fraction of the initially subducted sedimentary Li budget to depths approaching those beneath volcanic fronts, despite the redistribution of this Li among mineral phases during complex mineral reaction histories.

Ultramafic and carbonate-rich rocks juxtaposed in a sedimentary melange exposed in Champorche Valley, Italy, show evidence of metasomatic alteration at their contacts and the reactions that occurred at these contacts afford an assessment of the sources and compositions of fluids associated with the alteration and the degree to and scale at which this reaction mobilized carbon. At these contacts, carbonate-rich rocks display calcite replacement by diopside and tremolite along foliations and stylolite planes as a result of carbonation and/or carbonate dissolution. Elsewhere, the associated ultramafic bodies record serpentine replacement by carbonates (carbonation) and formation of carbonate and of carbonate + diopside + chlorite metamorphic veins. These two sets of observations point to coeval, coupled decarbonation and carbonation resulting in C mobility along the deep subduction interface (at about 60 km depth) but conceivably without net loss and large-scale transport.The d 18 O V-SMOW values of all the samples analyzed in this study are lower than expected for likely oceanic protoliths (for marine limestone, carbonate with values of +28 to +30‰; for seafloor ophicarbonate somewhat lower), suggesting pervasive interaction of these rocks with externally-derived fluids, as has been observed throughout the region (for calc-schists such as the Schistes Lustres, lowering values to +20 ± 2‰). The calc-schist unit at Champorcher has d 18 O falling into this regionally-developed range; however, the other rocks at this locality (carbonate mélange/broken formation, metasomatic rinds, and veins) tend to have lower values (to as low as +13.5‰). These lower values appear to require interaction with a fluid with d 18 O lower than that affecting the calc-schist on a regional scale and more consistent with derivation in a mafic/ultramafic (ophiolitic) source. Some metamorphic veins showing carbonate d 18 O values 1 to 2‰ lower than their hosts, and the occurrence of metamorphic veins and metasomatic horizons with anomalously high 87 Sr/ 86 Sr point to some circulation of isotopically distinct external fluids enriched in radiogenic Sr within the Champorcher suite, perhaps involving a source in devolatilizing terrigenous rocks.Juxtaposition of rocks at scales observed in these melange units could enhance the mobilization of C via decarbonation reactions if this deformation is accompanied by the infiltration of the interface by H 2 O-rich fluids capable of driving them. Information regarding the metasomatism within this hybrid carbonate-ultramafic unit bears on the question of C mobilization along subduction interfaces and whether the magnitude of any loss or gain at the scales investigated could significantly influence whole-margin C cycling.

We measured the N concentrations and isotopic compositions of 44 samples of terrestrial potassic and micro- and meso-porous minerals and a small number of whole-rocks to determine the extent to which N is incorporated and stored during weathering and low-temperature hydrothermal alteration in Mars surface/near-surface environments. The selection of these minerals and other materials was partly guided by the study of altered volcanic glass from Antarctica and Iceland, in which the incorporation of N as NH4+ in phyllosilicates is indicated by correlated concentrations of N and the LILEs (i.e., K, Ba, Rb, Cs), with scatter likely related to the presence of exchanged, occluded/trapped, or encapsulated organic/inorganic N occurring within structural cavities (e.g., in zeolites). The phyllosilicates, zeolites, and sulfates analyzed in this study contain between 0 and 99,120 ppm N and have delta N-15(air) values of -34 parts per thousand to +65 parts per thousand. Most of these minerals, and the few siliceous hydrothermal deposits that were analyzed, have delta N-15 consistent with the incorporation of biologically processed N during low-temperature hydrothermal or weathering processes. Secondary ion mass spectrometry on altered hyaloclastites demonstrates the residency of N in smectites and zeolites, and silica. We suggest that geological materials known on Earth to incorporate and store N and known to be abundant at, or near, the surface of Mars should be considered targets for upcoming Mars sample return with the intent to identify any signs of ancient or modern life.

The dynamics of Earth's subduction zones can be examined through a combination of geophysical, geological, and geochemical observations of modern margins, theoretical study, and research on metamorphic rocks representing ancient subduction. No two modern subduction zones are alike—each has its own character and behavior. Recent work on subduction zone dynamics has increasingly examined individual margins rather than making broad generalizations regarding categories of subduction zone properties and processes. Many factors can distinguish one margin from another, to name a few, convergence rate, age and thermal structure of the incoming plate, subduction obliquity, makeup of the upper plate, thicknesses and types of subducting sediment, whether sediments are being accreted or the forearc is being eroded, and the subduction of features such as oceanic plateaus, abandoned or active spreading ridges, and fracture zones. In fact, the hugely varied combinations of these factors lead to the generation and evolution of subduction margins with incredibly diverse behavior and associated hazards to society in the form of earthquakes, tsunamis, and volcanic eruptions. In some cases, variations in behavior of an individual margin, along its strike, can clarify the most important forcing factors in its evolution. The greatest advances in subduction zone science will be provided by multidisciplinary groups assembled to consider the full complexity of margins both modern and ancient.

Melange rocks are physical mixtures of metasediments, eclogites, and serpentinized ultramafic rocks formed from deformation-assisted mixing and fluid-rock interactions along the slab-mantle interface. An increasing number of field, geochemical and experimental studies have argued that these rocks could play a major role in mass transfer from the slab to the overlying mantle and the formation of arc magmas. First, we will present experimental evidence that melting of peridotite hybridized by melange rocks produces melts that carry both the major and trace element signatures observed in a variety of natural arc magmas [1]. Based on a first study we conducted using melange rocks from Syros (Greece), we propose that differences in the nature and relative contributions of melanges hybridizing the peridotite mantle wedge produce a range of primary arc magmas, from tholeiitic to calc-alkaline. Therefore, assimilation of melanges into the wedge may play a key role in transferring subduction signatures from the slab to erupted arc magmas. Second, we will explore the fate and petrophysical properties (solidus, density) of melange rocks as they descend along the slab-top during subduction. New high-pressure, high-temperature experiments will investigate density and mineralogical variations of melange rocks from Syros (Greece) and Catalina Schist (USA) at 1.5 and 2.5 GPa, and provide new constraints on the ability of melange rocks to physically migrate by buoyancy from the slab-top to the overlying mantle.

RELEVANT TARGETS FOR MARS SAMPLE-RETURN Matthew P. Nikitczuk1 , Gray E. Bebout 1,3, Charles A. Geiger2, Tsutomu Ota3, Tak Kunihiro3, Ryoji Tanaka3, John F. Mustard4, Roberta L. Flemming5, Eizo Nakamura3. 1Lehigh University (Dept. Earth and Environmental Sciences, Bethlehem, PA 18015, USA; mpn217@lehigh.edu), 2Universität Salzburg (Fachbereich Chemie und Physik der Materialien, Salzburg, Austria), 3Okayama University (Institute for Planetary Materials, Misasa, Tottori 682-0193, Japan), 4Brown University (Department of Earth, Environmental and Planetary Sciences, Providence, Rhode Island, USA), 5Universtiy of Western Ontario (Dept. Earth Sciences, London, ON, N6A 3K7, Canada)

The Ivrea Zone, exposed in the Southern Alps, affords an examination of nitrogen mobilization and storage in deep continental crust, in rocks spanning the amphibolite to granulite facies transition and that experienced multiple devolatilization reactions and partial melting. Such information is important for considerations of whole-Earth N cycling, as continental crust is thought to contain up to similar to 15% of global N, much of it organic in origin and thus tied to ancient-Earth biogeochemical processes. In these rocks, N is redistributed among silicate phases during devolatilization and both H2O-saturated and dehydration melting, showing particular affinity for potassic phases such as biotite, muscovite and alkali feldspar. In metasedimentary rocks, N concentrations decrease strongly with increasing grade, from the Kinzigite Zone (amphibolite facies) with 29 to 181 mu g/g to the Stronalite Zone (granulite facies) containing only 10 to 15 mu g/g. The decrease in N is intermediate in magnitude to that of Cs, showing somewhat greater loss, and that of Rb that shows smaller degrees of loss. The loss of N with increasing grade is due to breakdown of biotite, which is the main mineral host of N. It also reflects a strong preference of N for the fluid during dehydration and a moderate preference of N for the melt during partial melting. Accompanying this loss of N is a shift by up to 9 parts per thousand in delta N-15 towards lower values. The direction of this shift contrasts with what would be expected for loss of N into fluids speciated as either N-2 or NH3 and likely reflects complex isotope fractionation effects during equilibration between melt, fluid, and rock, perhaps during evolving redox conditions. The lowest delta N-15 values for metabasaltic rocks in this suite are near that of the mantle (-5 parts per thousand) and the array of values across the traverse could represent initial protolith heterogeneities and superimposed effects of magmatic degassing of the protolith. The leucosomes broadly overlap in N concentrations and delta N-15 values with their host rocks and lack any systematic relationships with metamorphic grade or type of host lithology. As has been noted in previous work, estimation of the size and isotopic composition of the N reservoir in continental crust is complicated its great lithologic heterogeneity and by the paucity of data for metamorphic rocks representing greater depths. Superimposed on the lithologic heterogeneity associated with varying protoliths is the ability of highly devolatilized and partially melted deep crustal rocks to store N over long time periods. We used our new data of deep crustal rocks from the Ivrea Zone, combined with published N data and based on lithological proportions in the crust, to provide a first estimate for the N content in the middle continental crust of 56 +/- 4 mu g/g. For the lower continental crust, we estimate a N content of 26 +/- 4 mu g/g N but note that the data base of granulite-facies samples is still small. Calculation of the total continental crust N content yields 74 +/- 4 mu g/g, which is in between previous estimates of 56, 60 and 88 mu g/g N. Finally, we estimate the N isotopic composition of the bulk continental crust as delta N-15 = 4.3 +/- 0.5.