Suite Of Xenoliths Research Articles

Spinel to garnet lherzolite transition in relation to high temperature palaeogeotherms, eastern Australia∗

Eastern Australian xenolith suites and lithospheric transition zones are re‐evaluated using new mineral analyses and thermo‐barometry. Some suites, including that defining the southeastern Australian geotherm, are not fully equilibrated. New pressure‐temperature estimates, based on experimental calibrations that allow for Cr and Ti in pyroxenes, differ from earlier results by up to 0.6 GPa and 250°C. The preferred Brey and Köhler 1990 thermo‐barometer indicates a shallower cooler garnet lherzolite transition under Mesozoic New South Wales (50 km depth at 980° C) than for Tertiary Tasmania (60 km depth at 1090°C). Deviations between palaeogeotherms may reflect: (i) higher temperature gradients for Tasmania and New South Wales (by 100°C/0.1 GPa) related to abnormally hot mantle; (ii) higher temperature gradients linked to more voluminous magmatism, largely Cenozoic in age; and (iii) complex temperature perturbations linked to different levels of magmatic intrusion. These deviations blur reconstructions of lithospheric assemblages, where temperature is determinable and pressure comes from an assumed geotherm. Potential errors in locating spinel lherzolite and crust‐mantle transition assemblages may reach 15 km in depth. The highest Tertiary geothermal gradients in Tasmania and northeastern New South Wales match those from regions of active lithospheric extension. The young southeastern Australian geotherm is decaying from a higher temperature equilibration, based on experimental work, and Mesozoic New South Wales geotherms trend towards the lower gradients of bounding cratons.

A petrologic re-investigation of the Adak volcanic center, central Aleutian arc, Alaska

The Adak volcanic center is located in the central part of the Aleutian arc and consists of three main volcanic vents. Andrew Bay Volcano, the oldest center, has been mostly removed by erosion. The next youngest vent, Mount Adagdak, was built in three major volcanic stages whereas Mount Moffett, the largest volcanic edifice, consists of a main cone and a parasitic cone each with several magmatic phases. Adak is unique compared to other modern Aleutian volcanic centers in that it contains two xenolith suites (Conrad and Kay, 1984; Debari et al., 1987). One suite consisting predominantly of mafic xenoliths occurs on Mount Moffett whereas an assemblage of ultramafic and mafic xenoliths is found on Mount Adagdak. Lavas erupted at Adak span the compositional range from 48.4 to 65.0 wt.% SiO2 and are characterized by significant variations in Al2O3, MgO, Sr, Ni and Cr. On Harker diagrams, this variability produces compositional trends with significant scatter. The Adak suite has total REE contents that vary from 32 to 154 ppm but do not correlate systematically with silica. (LaYb)n ratios range from 2.41 to 21.72 with the majority of lavas between 2.41 and 6.06. On process identification diagrams, the Adak suite plots as steeply sloping trends that contrast with the horizontal patterns of most other Aleutian centers. Measured isotopic ranges are large and nearly equal to those for the entire arc. Although they span similar silica ranges, subtle geochemical and isotopic differences distinguish the different volcanic vents of Adak. On Mount Moffett, a geochemically and isotopically distinct group of andesites (55.5–57.9 SiO2), the mafic andesites, occur on its NE flank. These lavas have elevated MgO, Ni and Cr but are depleted in Al2O3 relative to other Mount Moffett andesites with similar silica. They also have more heterogeneous REE abundances and isotopic ratios than most of the other andesites. Significant compositional differences exist between Adak and the other volcanic centers of the central Aleutian arc. Although these differences are characteristic of all geochemical systems, they are greatest for major and rare-earth elements and isotopic ratios. The lack of coherent relationships on major- and trace-element Harker diagrams, the isotopic variability, as well as the steeply sloping trends on REE process identification diagrams suggest that the Adak volcanic suite was not formed predominantly by closed-system crystal fractionation, but must be the product of a complex open-system process(es). The significant isotopic variability displayed by the suite suggests that contamination by an isotopically distinct contaminant must also have been an important petrologic component in the evolution of the suite. REE data are also suggestive of a role for magma mixing. Such a complex petrologic evolution is consistent with an immature lithospheric plumbing system. Based on REE systematics, the xenolith suites of Adak cannot, as previously proposed, be related to the host lavas or the rest of the Adak suite through crystal fractionation schemes. Rather they are probably accidental fragments derived from various depths along lithospheric conduits. In light of their relation to xenolith-bearing units, the mafic andesites of Adak presumably represent hybrid magmas formed during the interaction of ascending magmas with lithospheric wall rock. They are, therefore, characteristic of immature volcanic centers and unlikely to be related directly to the magmatic processes responsible for the generation of primary arc magmas. Because of the close proximity of the vents and the subtle compositional differences between their lavas, the Adak volcanic center was probably supplied by a single, deep lithospheric plumbing system that fed separate crustal magma chambers. The absence of historic volcanic activity on Adak suggests this plumbing system was abandoned before complete conduit development. This decline in magmatism may reflect a re-adjustment of volcano spacing within this part of the Aleutian arc.

Oxygen isotope composition of garnet and spinel peridotites in the continental mantle: Evidence from the Vitim xenolith suite, southern Siberia

Spinel and garnet fades peridotites from the Vitim Volcanic Field in Siberia comprise a suite of fertile mantle xenoliths, which are LREE depleted and have strongly depleted strontium and neodymium isotopic compositions. Determination of 18O16O ratios by conventional and laser-assisted fluorination techniques yield a very restricted range of whole-rock δ18O values for both spinel peridotites (+ 5.4 to + 5.8‰) and garnet lherzolites (+ 5.5 to + 5.8‰) equivalent to that observed for lunar rocks and midocean ridge basalts (MORB). Mineral δ18O values for the xenoliths are: olivine = + 5.1 to + 5.8%., orthopyroxene = + 5.7 to + 6.0%., clinopyroxene = + 5.5 to + 6.2%., garnet = + 5.5 to +6.0%., and spinel = + 4.9 to + 5.5%. Similarly, δ18O ranges for silicate mineral pairs vary from only 0.5 to 0.7%. The sixteen peridotite xenoliths studied exhibit equilibrium O-isotope fractionations between minerals of a magnitude expected from theoretical and experimental considerations, in particular a temperature-controlled 18O distribution between olivine and pyroxenes in spinel-bearing peridotites. Because the Vitim garnet and spinel peridotites are rich in basaltic components and show long-term depletion in incompatible elements, they are derived from a MORB-like mantle source which makes up, according to current geochemical models, a predominant part of the Earth's upper mantle. The Vitim xenolith suite may, therefore, reflect its bulk O-isotope character. With increasing temperature and degree of partial melting, such a mantle source could generate basaltic liquids with δ18O values varying from C. + 6.1 to + 5.7%. Thus, the broad δ18O range of c. + 4.5 to + 7.5%. reported in the literature for spinel peridotite xenoliths is probably a very restricted disequilibrium phenomenon that likely reflects recent metasomatic overprinting in highly localized zones of the uppermost continental mantle affected by basaltic magmatism and, therefore, should not be considered as representative of the upper mantle, in general.

Carbonation Reactions in Eclogites at High Pressure: Implications for Carbonic Fluids in the Mantle

Carbonation reactions for olivine-bearing ultramafic rocks at conditions appropriate for the Earth's mantle have been extensively studied over the past three decades (of. review by Luth et al., 1993, and references therein). In these assemblages, the carbonation reactions invariably involve olivine reacting with CO2. For example, carbonation in harzburgites (ol + opx) proceeds via: Enstatite + 2 Magnesite = 2 Forsterite + 2 CO2. In lherzolites (ol + opx + cpx), carbonation would occur initially via: Dolomite + 2 Enstatite = 2 Forsterite + Diopside + 2 CO2, with an exchange reaction between pyroxene and carbonate stabilizing magnesite at the expense of dolomite at higher pressure. Eelogite, a garnet + clinopyroxene rock that lacks olivine, is well-represented in suites of mantle-derived xenoliths, and is inferred to be a significant lithology in the mantle. The above carbonation reactions for olivine-bearing assemblages won't work for these olivine-frce assemblages, and little is known about possible carbonation reactions for these rocks. In a model (Fe-free) eclogite, an assemblage of epx + gar + coes + Mg, Ca carbonate is observed (Luth, unpub, data). A carbonation reaction consistent with this assemblage would be: Dolomite + Coesite = Diopside + CO2 (DCDV). This reaction is of interest aside from as a potential c a r b o n a t i o n react ion for eclogites. With decreasing Co2, DCDV intersects the carbonsilicate-carbonate reaction: Dolomite + Coesite = Diopside + 2 Diamond/Graphite + 02. This reaction may define the coexistence of elemental carbon and carbonate in eclogites and garnet pyroxenites (Luth, 1993). The lower-pressure, quartz-bearing analog of DCDV has been studied in CO2 and H20-CO2 fluids (Eggler et al., 1976; Wyllie et al., 1983; Slaughter et al., 1975; Eggert and Kerrick, 1981; Jacobs and Kerrick, 1981; LtRtge and Metz, 1991, 1993). The DCDV reaction has not been studied experimentally to date. It's location in pressuretemperature space may be calculated with the internally-consistent thermodynamic databases of Holland and Powell (1990) and Berman (1988, 1991). The two calculated curves are discrepant, markedly so at > 3 GPa, possibly resulting from the required extrapolation of the thermodynamic data to these conditions. Dalton and Wood (1993) found that these databases have difficulties reproducing reaction boundaries for carbonatebearing reactions for peridotites, which were attributed to uncertainties in the thermochemical data for carbonates. The experimental determination of the DCDV reaction serves to constrain the stability of carbonate in eclogitic assemblages, and in addition will provide an informative test of the currently-available databases for a carbonation reaction not involving magnesite.

Metasomatism in Two Natural Peridotites: Effects of Low-aSiO2 Melts

Metasomatism of upper mantle peridotites is a crucial process in the genesis and evolution of magmas in the mantle, yet the knowledge about actual metasomatic agents and processes during metasomatism is still imperfect. No clarity exists about chemical signatures enabling one to distinguish between possible agents (fluids, silicate melts, carbonatite melts) interacting with the peridotites. We are studying two occurrences of peridotites showing metasomatic reaction relationships, in which the responsible agents are preserved or can be shown to be alkaline silicate melts. These are the Inagli Dunite, Siberia, and the Gees xenolith suite, West Eifel.

Geochemical, Sr-Nd Isotopic, and P-T Data on a Central Mexican Xenolith Suite: Evidence for Upper Mantle and Lower Crust Compositions

Northeast of San Luis Potosi (Central Mexico, 22~ 101~ Quaternary basanites and tuff breccias have sampled a suite of ultramafic upper mantle to lower crustal xenoliths, associated with seven phreatomagmatic maars. Xenolith distribution range from spinel lherzolites (55%), spinel pyroxenites (10%), hornblende pyroxenites (18%) to metaigneous and metapelitic granulites (17%). The SLP volcanic field is located in the back-arc region of two major volcanic belts along the southwestern border of the North American craton (Fig.l): The Oligocene calc-alkaline volcanics from the Sierra Madre Occidental in the east and the late Miocene to recent basaltic to rhyolitic volcanics f rom the Transmexican Volcanic Belt. Analysis of the geochemical and Sr-Nd isotopic data together with p-T-conditions of the xenolith profile gives hints for the thermal evolution and composition of the upper mantle segment and lower crust in Central Mexico. The latter is of

The nature of the lower continental crust of Europe: petrological and geochemical evidence from xenoliths

Studies of lower crustal xenoliths in alkaline volcanics of Europe have revealed that a wide range of rock types exists in the lower crust. These range from eclogites, mafic garnet-bearing and garnet-free granulites, through felsic granulites to high-grade metasediments. Xenolith suites, and by inference the lower crust, are very variable in chemistry and mineralogy. Estimates of lower crustal composition vary according to the proportions of different rock types included in a xenolith suite, from basic (< 50 wt. % SiO 2) to intermediate (≈ 55 wt. % SiO 2). Ages within the lower crust are very poorly constrained, but zircons from crustal xenoliths are often younger than the oldest rocks exposed at the surface. This suggests that much of the mafic lower crust has been added to the base of the pre-existing crust by processes of intrusion and underplating. Such additions to the crust have probably occurred during periods of subduction, rifting and orogenic collapse. Evidence from radiogenic and stable isotopes suggests that extensive contamination of mafic magmas by the pre-existing crust has occurred. Seismic velocity constraints indicate that in some parts of Europe, e.g. Germany, mafic granulites are confined to the base of the crust, but in other regions (northwest Scotland, Hungary) they may occupy much of the lower crust. In the northern part of the Baltic shield (Kola) the thick crust contains variably retrogressed and metasomatized eclogites at depths between 30 and 48 km.

Mafic and ultramafic mantle and deep crustal xenoliths from Banks Peninsula, South Island, New Zealand

Abstract Mafic and ultramafic xenoliths of spinel lherzolite, wehrlite, websterite, pyroxenite, and metagabbro (mafic granulite) occur in a basanite plug forming Le Bons Bay Peak on the eastern flanks of Akaroa volcano, Banks Peninsula. Spinel lherzolites are assigned to the Cr‐diopside xenolith suite and are interpreted to represent fragments of depleted upper mantle derived from a depth of c. 60 km. Pyroxenites belong to the wehrlite series Al‐augite xenolith suite and may represent basaltic cumulates derived from differentiation of mafic magma intruded at or near the crust/ mantle boundary. Mafic granulite of gabbroic composition represents a tectonised and partly recrystallised lower crustal basaltic cumulate. Mineral and bulk rock chemistry of the xenoliths suggests that chemical and mineralogical variation in Banks Peninsula magmas could reflect the development of a subcrustal basaltic “underplate”. This might explain the trend from extensively modified to relatively unmodified mantle‐derived magmas...

Partitioning of fluorine and chlorine between apatite and aqueous fluids at high pressure and temperature: implications for the F and Cl content of high P-T fluids

Experiments in a piston-cylinder apparatus equilibrated natural fluorapatite with aqueous HCl, NaCl, NaOH, Na 2CO 3 and CO 2 H 2O mixtures at 1–2.0 GPa and 950–1050°C. Post-experiment apatite compositions were determined by electron microprobe and fluid compositions were calculated from mass-balance. This enabled calibration of the fluorine and chlorine contents of apatite coexisting with a variety of fluids at high P and T, and provides better constraints on the halogen composition of aqueous fluids which may exist in the deep crust and upper mantle. Results obtained at 2.0 GPa and 1050°C indicate that the ratio X ClAp/X OHAp in apatite ( X FAp , X ClAp and X OHAp are the mole fractions of fluor-, chlor- and hydroxyapatite) is proportional to the chlorine content of the fluid up to ∼ 1 (H 2O CO 2), ∼ 5 (H 2O HCl) and ∼ 15 (H 2O NaCl) wt% Cl. In contrast, the ratio X FAp/X OHAp was found to vary in proportion to the fluorine content of the fluid only at concentrations of ≤ 0.15 (H 2O HCl) and ≤ 0.4 (H 2O NaCl) wt% F. Arising from changes in the activities of aqueous HCl and HF, the results of this study demonstrate that fluid composition, in addition to the absolute abundances of F and Cl in the fluid, will determine the amounts of these elements incorporated into coexisting apatite. Specifically, relatively low Cl and F concentrations in acidic (H 2O HCl) fluids are required to obtain high ratios of X ClAp/X OHAp and X FAp/X OHAp in apatite, whereas much higher abundances of Cl and F in basic fluids (i.e., Na-bearing) are required to achieve the same result. In addition, resulting from dilution of the aqueous phase with CO 2 and the subsequent reduction in the activity of H 2O, apatites with high ratios of X ClAp/X OHAp and X FAp/X OHAp may be produced by CO 2 H 2O fluids with relatively low levels of Cl or F. Application of these results to apatites from a suite of ultramafic xenoliths from southeastern Australia indicates chlorine abundances of 500–3500 ppm in the fluid phase and points to chlorine as a potentially important component of some mantle fluids.

Formation and composition of the lower continental crust: Evidence from Scottish xenolith suites

Crustal xenoliths occur in a variety of Carboniferous and Tertiary alkali basaltic rocks across Scotland and provide indications of the nature of the deep crust beneath the major lithotectonic zones from the Archean Lewisian of the far NW to the Paleozoic terranes in the SE. Sr, Nd, and Pb isotopic and representative chemical compositions are presented for 35 mafic xenoliths from the different age and tectonic zones of Scotland and for 21 felsic, gneissose xenoliths, from the Midland Valley and Southern Uplands in the south. The coexistence of mafic granulite and spinel lherzolite xenoliths at several of these localities, together with seismic velocity data, provide evidence that the mafic xenoliths, and possibly some of the felsic xenoliths, may be derived from 20–30 km depth. The absence of garnet from the mafic assemblages suggests that in most areas the crust was probably not much more than 30 km thick in the Permo‐Carboniferous. The extremely variable Al2O3 (up to 27 wt %) and MgO contents at any one location, together with positive Eu and Sr anomalies in chondrite‐ and mantle‐normalized diagrams, provide evidence that many of the mafic xenoliths represent cumulates from basaltic magmas. The Loch Roag mafic xenoliths, erupted through the depleted high‐grade metamorphic Lewisian Complex, are especially rich in K (average 1.3 % K2O) and Rb (average 54 ppm) with K/Rb extending as low as 100. The K/U ratios for the mafic xenoliths are high, ranging to over 100,000 at Loch Roag and in the SW Midland Valley. The relative concentrations of U, K, Rb, Pb, Nd, Sm, Eu and Sr in most of the mafic Xenoliths, while displaying regional characteristics, display intra‐site variations that are consistent with crystal‐liquid fractionation. There is little evidence of U or Rb depletion via fluid or melt extraction during metamorphism. The concentrations of U in the mafic xenoliths display as much intrasite uniformity as the rare earth elements and correlate with Zr concentrations. U/Zr ratios are similar to those found in ocean island basalts. The mafic xenoliths display a marked decrease in ϵNd and 206Pb/ 204Pb with increasing age of the upper crust from SE to NW, as also observed in the ∼400 Ma Caledonian granites. Sm‐Nd crustal residence ages for the Loch Roag suite are particularly variable ranging from 1.6 to 3.6 Ga. However, these isotopic compositions and model ages appear to be, in part, the product of crustal assimilation by the igneous precursors to the mafic xenoliths. Model ages, mineral ages and scattered whole rock isochrons indicate that many of the precursors to these Xenoliths were in fact formed and metamorphosed during the late Proterozoic and Paleozoic and provide evidence of widespread deep crustal magma reservoirs, underplating, crustal assimilation and arc accretion at these times. Nd and Sr isotopic data for coexisting pyroxenite xenoliths at Loch Roag indicate that they may represent contaminated ultramafic cumulates that were cogenetic with the precursors to the mafic xenoliths. The Pb isotopic compositions of the Midland Valley mafic and felsic xenoliths are similar and relatively uniform despite large variations in Nd and Sr isotopic compositions. This is explicable if the crust in this region was relatively uniform and similar in composition to the mantle and/or if crustal Pb dominated in arc environments. The average major element composition of the Scottish mafic xenolith suite is similar to that reported from other regions such as the Sierra Nevada. This is taken as evidence that the processes of crystal fractionation and formation of mafic and ultramafic cumulates during underplating are relatively consistent between different regions. Average U/Pb ratios of mafic and felsic xenoliths are effectively identical (0.068 and 0.064, respectively). The average Sm/Nd ratio (0.20) for the entire suite is lower and average K/U ratio (27,000) higher than estimates of average lower crust based on Queensland mafic xenoliths. These compositions as well as those of basement rocks, sedimentary supracrustal rocks and granites provide evidence of systematic chemical differences between the Scottish and Australian lithosphere. The compositional features of the Paleozoic Scottish crust such as low Sm/Nd and high K/U may have been sustained in part by recycling of Archean and early Proterozoic lithosphere through sedimentary processes, subduction and arc volcanism. The average Sm‐Nd crustal residence age and Pb isotopic composition of the mafic xenoliths at Loch Roag is very similar to that of the host Lewisian. Conservative mass balance models imply an average ratio of Lewisian‐derived Nd to parent magma Nd in the xenoliths of about 2. Such large degrees of contamination indicate that although underplating may be significant process in reconstituting Archean lithosphere, it does not necessarily result in substantial changes to the average age of the crust.

Rb-Sr and Sm-Nd isotopic variations in dissected crustal xenoliths

Four crustal xenoliths from Scotland have been dissected in order to evaluate the effects of magma-xenolith interaction on Rb-Sr and Sm-Nd isotopic systematics, and hence the use of the xenoliths for accurate determination of source compositions. Each xenolith was sampled from the edge to the center. The Nd isotopic compositions are virtually uniform across each xenolith; however, there are significant variations in Rb, Sr, and REE concentration, as well as Rb Sr and Sm Nd ratios and Sr isotopic composition. Most of this variation appears to be inherited from the protolith. However, one sample shows systematic variations in Rb and Sr concentration and Sr isotopic composition with distance from the xenolith edges due to interaction with the host magma. This sample displays petrographic evidence for grain boundary melt infiltration. These data are consistent with theoretical considerations of transport times and diffusivities that place limits on the amount of modification that is possible via diffusion alone. The calculated effects of transport of even large amounts of crustal xenoliths on the compositions of xenoliths and host magmas will be small, provided transport times are short. Overall, the isotopic variations within the xenolith suite mirror the source rocks from which they were extracted and are inferred to reflect variations in the deep crust of Scotland. However, small samples, such as most crustal xenoliths, may be of little use in defining meaningful crustal residence ages because of modal variations in metamorphic minerals and isotopic re-equilibration. The one xenolith studied has extremely variable Sm Nd ratios due to the presence of garnet and yields a broad range of geologically meaningless Nd model ages, but yields an Sm-Nd isochron age of 396 ± 49 Ma. The Loch Roag xenolith has a large range of Rb Sr ratio, and the Sr isotopic data define an isochron age of 416±28 Ma. Both ages are taken to represent the timing of metamorphism and/or cooling from 600°C at the end of the Caledonian Orogeny in Scotland.

Petrology of ultramafic and mafic xenoliths in picrite of Kahoolawe Island, Hawaii

A picrite lava (22 wt% MgO; 35 vol.% ol) along the western shore of the∼1.3–1.4 Ma Kahoolawe tholeiitic shield, Hawaii, contains small xenoliths of harzburgite, lherzolite, norite, and wehrlite. The various rock types have textures where either orthopyroxene, clinopyroxene, or plagioclase is in a poikilitic relationship with olivine. The Mg#s of the olivine, orthopyroxene, and clinopyroxene in this xenolith suite range between 86 and 82; spinel Mg#s range from 60 to 49, and plagioclase is An75–80. A 87Sr/86Sr ratio for one ol-norite xenolith is 0.70444. In comparison, the host picrite has olivine phenocrysts with an average Mg# of 86.2 (range 87.5–84.5), and a whole-rock 87Sr/86Sr ratio of 0.70426. Textural and isotopic information together with mineral compositions indicate that the xenoliths are related to Kahoolawe tholeiitic magmatism, but are not crystallization products of the magma represented by their host picrite. Rather, the xenoliths are crystalline products of earlier primitive liquids (FeO/MgO ranging ∼1 to 1.3) at 5–9 kbar in the cumulate environment of a magma reservoir or conduit system. The presence of ultramafic xenoliths in picrite but not in typical Kahoolawe tholeiitic lava (6–9 wt% MgO) is consistent with replenishment of reservoirs by dense Mg-rich magma emplaced beneath resident, less dense tholeiitic magma. Mg-rich magmas have proximity to reservoir cumulate zones and are therefore more likely than fractionated residual liquids to entrain fragments of cumulate rock.

The moho as a magnetic boundary revisited

We have re‐examined the question of whether it is generally true that in continental regions a relatively magnetic crust overlies a relatively non‐magnetic mantle. Whereas the original conclusion was based on limited data, the present study is based on about 400 globally‐distributed xenolith samples. While it is now clear that the “Moho” represents a maximum velocity gradient in a complex transition from crust to mantle, and while crustal magnetization is highly variable, the original concept of a “magnetic crust” overlying a non‐magnetic mantle is strongly supported by the new data. A large upper mantle xenolith suite of ultramafic composition is characterized by non‐magnetic chrome spinels and magnesian ilmenites. Such rocks are magnetic only where demonstrably altered. The most significantly magnetic rocks within the crust are mafic types prograded to granulite grade or mafic melts of basaltic composition which crystallized deep in the crust at high pressure and temperature. Prograded rocks have characteristic magnetite Curie points, and are strongly magnetic, especially the more mafic compositions, while the mafic rocks formed by crystallization at depth have Curie points between 400°–570°C and are strongly magnetic. Induced magnetizations are commonly several A/m, and can readily account for long‐wavelength magnetic anomalies measured by satellite and aircraft; thus, there is no need to search for exotic sources for “missing” crustal magnetization. No metal or other peculiar magnetic species have been observed by us in any of our (unaltered) samples. Mineral assemblages (including magnetic components) are consistent with oxidizing conditions close to the FMQ buffer.

Contrasting styles of Pre‐Cenozoic and Mid‐Tertiary crustal evolution in northern Mexico: Evidence from deep crustal xenoliths from La olivina

The principal deep crustal rock types found at the La Olivina xenolith locality in southeastern Chihuahua, Mexico, are mafic granulites, paragneisses, and intermediate‐ to silicic‐composition orthogneisses. These granulite facies xenoliths are interpreted in terms of two age groups, pre‐Cenozoic and mid‐Tertiary, based on previous ion probe dating of zircons from the xenoliths and on isotopic comparisons of the xenoliths to rocks of known age. The mafic granulites have Pb, Nd, and Sr isotopic compositions identical to those of Oligocene volcanic rocks from the La Olivina region. Compositionally, they are olivine‐normative gabbroic cumulates, and they precipitated from two or more mid‐Tertiary basalt to dacite or rhyolite assimilation/fractional crystallization series. Mineral assemblages in the xenoliths record pressures of ≤7.2 kbar or depths of &lt;25 km. If these are the maximum pressures the rocks experienced and if the crust was &gt;35 km thick in Oligocene time as inferred from regional tectonic considerations, then the mafic granulites cannot be samples of basaltic magmas underplated near the crust‐mantle boundary. The cumulate protoliths for the mafic granulites probably formed in magma chambers well above the Moho. The mafic granulites are plausibly representative of kilometers of new crust that formed in mid‐Tertiary time beneath the extensive ignimbrite fields of Mexico. Most orthogneiss xenoliths are pre‐Cenozoic, and they are rocks associated with the late Paleozoic Ouachita Orogeny and older events (e.g., Proterozoic basement and Paleozoic arc rocks). The Ouachita Orogeny was a collision event involving North America and a continental plate to the south, and the results of this study indicate that La Olivina is located above the southern plate. The paragneiss xenoliths overlap in isotopic composition with Carboniferous flysch units exposed in the Marathon uplift of west Texas. These sedimentary rocks and the sedimentary protoliths of the paragneiss xenoliths were derived from the southern plate and deposited before the orogeny in the ocean basin that separated North America and the southern plate. The paragneisses were not metamorphosed to granulite facies until mid‐Tertiary time. Pre‐Cenozoic and mid‐Tertiary crustal evolution followed very different paths in northern Mexico. For example, Nd isotopic evidence for crustal recycling is much more evident in rocks associated with the Paleozoic convergence than in rocks produced during mid‐Tertiary magmatism. Furthermore, mafic rocks are very rare in the pre‐Cenozoic xenolith suite, but they dominate the mid‐Tertiary one. These contrasting characteristics of the pre‐Cenozoic and mid‐Tertiary xenolith suites are interpreted to reflect differences in tectonic environment and crustal thickness. Preceding the collision event, the southern plate had an Andean‐type margin, and the abundant evidence for crustal recycling during this time implies that the crust was thick. The rarity of pre‐Cenozoic mafic xenoliths suggests that Proterozoic and Paleozoic lower crust may have delaminated in response to crustal overthickening associated with the Ouachita collision event.

Proterozoic depletion of the lithosphere recorded in mantle xenoliths from Inner Mongolia

MANTLE xenoliths entrained in alkali basalt provide direct samples from a cross-section of the lithosphere. In most cases, the chemistry and isotope systematics of these objects have suggested a complex history of chemical change by processes such as metasomatism and partial melting (see, for example, ref. 1). Here, however, we present evidence from a suite of xenoliths from alkali basalts in the Chinese province of Inner Mongolia, suggesting that the lithosphere in this region has remained stable and closed to chemical modification since experiencing partial melting in the Proterozoic. Clinopyroxene separates from these xenoliths yield an approximate isochron in the 143Nd/144Nd versus 147Sm/144Nd diagram, suggesting an episode of depletion ∼1.6 Gyr ago. Although other explanations for this correlation, such as mixing, cannot be ruled out definitively, further evidence for the long-term undisturbed nature of the lithosphere in this region comes from neodymium model ages of the clinopyroxenes. In particular, one highly depleted sample with very high 147Sm/144Nd provides a model age of 1.6 Gyr, consistent with the isochron result.

Geology and petrology of Floreana Island, Galapagos Archipelago, Ecuador

Floreana is the sixth largest of the Galapagos Islands and lies at the extreme southern margin of the Galapagos Platform. In sharp contrast to the other major islands, Floreana lacks a well-developed volcanic center, but instead is dominated by pyroclastic vents and derivative ash and cinder deposits. The majority of exposed lavas are xenolith-bearing alkali-olivine basalts that constitute a coherent suite referred to as the Main Series. The Flank Series comprises a group of isolated vents on the southern flank of the island which erupted xenolith-poor basaltic lavas that are generally poorer in incompatible trace elements and richer in plagioclase. Paleomagnetic measurements, supported by K-Ar age determinations, record a relatively limited timespan for subaerial volcanism. Shield development reached maximum dimensions by 1.0 Ma, with the final phases of pyroclastic activity at an end by 0.3 Ma. Main Series rocks are phenocryst-poor and typically contain clusters of xenocrysts in an extremely fine-grained, granular matrix. Petrogenetic modelling indicates they represent products of partial melting of a LREE-enriched spinel-lherzolite source, little modified by subsequent fractional crystallization. Porphyritic Flank Series lavas experienced rather different ascent histories and may have resided in shallow-level magma chambers prior to eruption. Isotopic, geochemical and petrographic data demonstrate that Flank Series lavas are not cogenetic in any simple way with Main Series flows. Floreana lavas contain a diverse suite of ultramafic xenoliths. Al-augite group megacrysts are likely to be cognate and comagmatic with host alkali-olivine basalts. Transition metal contents of megacrysts and host lavas, however, preclude clinopyroxene fractionation as a mechanism of magmatic differentiation. It is proposed that megacrysts were products of passive crystallization and were never effectively separated from host magmas. Cr-diopside group nodules may have crystallized at very high temperatures from advanced partial melts and are less likely to bear any direct genetic connection to host basalts.

Thermal history of the upper mantle beneath a young back-arc extensional zone: ultramafic xenoliths from San Luis Potos�, Central Mexico

At the San Luis Potosi (SLP) volcanic field (Central Mexico), Quaternary basanites and tuff breccias have sampled a suite of ultramafic xenoliths, predominately spinel lherzolites, spinel-olivine websterites, spinel pyroxenites, and hornblende-rich pyroxenites. Spinel lherzolites from the La Ventura maars have protogranular to equigranular textures, those from the Santo Domingo maars are strongly sheared. Both spinel-lherzolite types show similar whole-rock major and trace-element abundances. They are fertile to slightly depleted with mineralogical and geochemical heterogeneities induced by partial melting processes. Pyroxenites with either magmatic or metamorphic textures are high-pressure cumulates. Hornblende-rich pyroxenites are genetically linked to the host basanites. Most of the protogranular spinel lherzolites contain veinlets of glass along grain boundaries. These glasses are chemically homogeneous and have trachybasaltic to trachyandesitic compositions. Mg- and Fe2+-partitioning between olivine and glass suggests chemical equilibrium between the melts represented by the glasses and the spinel-lherzolite mineral assemblage at about 1,000°C and 10 to 15 kbar. The melts are interpreted to be of upper mantle origin. They may have been formed by in-situ partial melting in the presence of volatiles or represent percolating melts chemically buffered by the spinel-lherzolite mineral assemblage at uppermost mantle conditions. Mineral chemistry in all rock types of the whole xenolith suite reveals distinct disequilibrium features reflecting partial re-equilibration stages towards lower temperatures estimated to be from 1,050°C to 850°C at 9 to 15 kbar. The presence of similar zoning and exsolution features mainly documented in pyroxenes along with similar maximum and minimum temperatures requires all sampled xenoliths to have undergone the same temperature regime within the upper mantle. The sheared spinel lherzolites from the Sto. Domingo field are interpreted as formerly protogranular material which was sheared during uplift and cooling. The estimated mantle temperatures are higher than those predicted by low heat-flow measurements at the SLP fild, indicating that surface heat flow has not equilibrated to elevated temperatures at depth. This strongly supports a young perturbation event beneath the SLP area and connects the onset of uplift and cooling of the SLP-mantle segment with the back-arc extensional regime of the Quaternary volcanic cycle of the Transmexican Volcanic Belt.

Isotopic relationships of volatile and lithophile trace elements in continental ultramafic xenoliths

The concentrations and isotopic compositions of Sr, Nd, Pb, He and C have been determined for suites of xenoliths from Bullenmerri (Australia), Ichinomegata (Japan), Geronimo (Arizona), and East Africa. The wehrlites and pyroxenites from Bullenmerri have Sr, Nd and Pb isotopic compositions that are generally similar to those found for alkali basalts in the region. The spinel lherzolites, in contrast, have higher 87Sr/86Sr and 206Pb/204Pb and lower 143Nd/144Nd ratios. Whereas the isotopic compositions of He are generally within the range of mid-ocean-ridge basalts (MORB) and do not covary with those of other trace elements, there is an apparent correlation between the 13C/12C and 143Nd/144Nd ratios for each of the two petrologic groups. These relationships, if substantiated for other xenolith suites, greatly limit the possible mechanisms for generating lithophile and volatile isotopic variations in the continental lithosphere. The helium isotopic compositions for all of the xenoliths fall within the range for MORB. This includes those from Ichinomegata, suggesting that the lower 3He/4He ratios found for He sampled at the surface at subduction zones result from mixing mantle He with near-surface crustal He rather than with subducted radiogenic He. Measured C isotopic compositions (relative to Peedee belemnite) for the Ichinomegata xenoliths $$\left( {\mathop \delta \nolimits^{^{13} } C_{PD{\text{B}}} = - 3.2 to - 17.5} \right)$$ include values that are both lighter and heavier than those in MORB, and are compatible with contributions from subducted carbon. The Nd and Sr isotopic compositions of the Ichinomegata xenoliths exhibit a correlation over a substantially greater range of values than typically observed for other light-rare-earth-element (LREE)-depleted xenoliths, and include more radiogenic Sr and less radiogenic Nd compositions. The carbon isotopic compositions found for the East African and Geronimo xenoliths extend to values that are lighter than those typically found for MORB.

Relationships between chamber margin accumulates and pore liquids: evidence from arrested in situ processes in ejecta, Latera caldera, Italy

Biotite- and clinopyroxene-rich mafic nodules occur together with syenite ejecta in ca. 235 to 155 Ka tuffs surrounding the Latera caldera. Clinopyroxenites and leucite monzosyenites crystallized along the lower margins of a crustal magma system, and record complex crystallization histories of potassic magmas that were parental to a range of lava and tuff compositions. The mafic nodules have the mineral assemblage clinopyroxene >biotite>anorthite>orthoclase>leucite>hauyne >titanite>apatite±amphibole±olivine±phlogopite, and comprise mesocumulate and orthocumulate layers that commonly alternate on the scale of several centimeters. Despite the apparently “ultramafic” nature of the early cumulate assemblage, the common occurrence of intercumulate orthoclase, leucite, hauyne, salitic clinopyroxene and vesicular glass indicates that interstitial liquids underwent late-stage differentiation at relatively shallow depths but under highly variable conditions of volatile saturation. Many of the mafic nodules exhibit pronounced variations in the type and abundance of mineral reaction textures. These range from unreacted assemblages to nodules in which cumulate megacrysts are surrounded by well-developed symplectite halos. The textural variability indicates strongly localized disequilibrium between cumulate frameworks and vapor-saturated pore fluids, and is attributed to convective fractionation of liquids through permeable crystal mush. Lithologic discontinuities exhibited by the xenolith suite and host magmas are best explained by variable communication between a compositionally stratified magma reservoir and its partly crystallized lower chamber margins. Periodic replenishment of this system by less evolved magmas subsequently promoted mixing and hybridization that are characteristic features of many Latera lavas and tephra.

Boron geochemistry of the lower crust: Evidence from granulite terranes and deep crustal xenoliths

Boron contents are uniformly low in more than 100 granulites from exposed terranes in India, Norway, and Scotland and from xenolith suites in the western USA. Averages for the terranes (2.5 ppm) and for xenoliths (1.2 ppm) suggest maximum B contents of about 2 ± 1 ppm for the lower crust. Abundance distributions from exposed terranes are skewed to higher values as B has been added to some samples via retrograde fluids during decompression. The samples studied include mafic to felsic lithologies of both igneous and sedimentary origin. There is no correlation of B content with bulk composition or with protolith type in any of the suites studied. Boron is apparently depleted in all granulite protoliths during prograde metamorphism and dehydration. Similar depletions of B and other fluid-mobile elements (e.g., U, Cs) with respect to rare-earth elements (REE) Zr, Ba, Rb and K 2O are seemingly inconsistent with origin of granulites primarily via extraction of silicate melts. These distinctive geochemical features are attributed to selective element transport in fluids released by devolatilization reactions. Anatexis is not necessarily precluded, but dehydration of the subsolidus protolith normally would precede melting and lead to depletion of the fluid-mobile elements. Compositions of any melts and restites eventually formed would reflect the effects of this antecedent process. The systematic depletion of B (and Cs) in granulites contrasts with the highly variable contents of most other incompatible elements determined. For this reason, the estimated lower crust abundance for B is relatively well constrained, whereas abundance estimates for many other elements are model dependent, have large uncertainties and are unlikely to be globally representative.