Search

Petrologic and geochronologic data for metapelitic lower crustal xenoliths from New Mexico (USA) and Chihuahua (Mexico) states provide evidence for both a magmatic and collisional component to the enigmatic Mesoproterozoic Picuris orogeny. These garnet-sillimanite-bearing metapelites are found within the southern Rio Grande rift at Kilbourne Hole and Potrillo Maar in southern New Mexico and northern Chihuahua. Geothermobarometry and rutile with Quaternary U-Pb dates indicate equilibration in the local lower crust, which is actively undergoing ultra-high temperature (UHT) metamorphism (Cipar et al., 2020). The samples contain older detrital zircons dating back to the Paleoproterozoic, marking their deposition at the surface. Coupled zircon U-Pb dates and trace-element ratios (e.g., Gd/Yb) show a clear transition from oscillatory-zoned, low-Gd/Yb detrital magmatic zircon to featureless, high-Gd/Yb metamorphic zircon between 1500 and 1400 Ma, marking the transition from subduction to collision during this period. Metamorphic zircon and monazite grew in two major intervals. The first, between ca. 1450 and 1350 Ma, documents the journey of the sediments to depth within the orogen and provides evidence of extended Mesoproterozoic metamorphism in the region. The second corresponds with UHT metamorphism that commenced at ca. 32 Ma and is associated with the Rio Grande rift. Whereas nearly all garnets are homogeneous in both major and trace elements, a single garnet from one sample has a core defined by abundant quartz and acicular sillimanite inclusions. The core and rim of this garnet is homogeneous in major and most trace elements, but the rim is enriched in the slowest diffusing elements, Zr and Hf, which likely indicates rim growth at higher temperatures. We interpret the garnet core to have grown at the time of emplacement of the sediments into the lower crust. Because this occurred in the sillimanite stability field and because the metamorphic zircon and monazite all have negative Eu anomalies, indicating their equilibration with feldspar (stable at depths of

We integrated new and existing bedrock and detrital zircon dates from the Zealandia Cordillera to explore the tempo of Phanerozoic arc magmatism along the paleo-Pacific margin of southeast Gondwana. We found that episodic magmatism was dominated by two high-magma-addition-rate (MAR) events spaced ∼250 m.y. apart in the Devonian (370–368 Ma) and the Early Cretaceous (129–105 Ma). The intervening interval between high-MAR events was characterized by prolonged, low-MAR activity in a geographically stable location for more than 100 m.y. We found that the two high-MAR events in Zealandia have distinct chemistries (S-type for the Devonian and I-type for the Cretaceous) and are unlikely to have been related by a repeating, cyclical process. Like other well-studied arc systems worldwide, the Zealandia Cordillera high-MAR events were associated with upper-plate deformation; however, the magmatic events were triggered by enhanced asthenospheric mantle melting in two distinct arc-tectonic settings—a retreating slab and an advancing slab, respectively. Our results demonstrate that dynamic changes in the subducting slab were primary controls in triggering mantle flare-up events in the Phanerozoic Zealandia Cordillera.

Accurate knowledge of tidal turbine impacts on the far-field hydrodynamic conditions, which extend from 3 to 20 diameters downstream the turbine, is essential for the estimation of tidal resource, farm layout design and environmental impact. For this purpose, tidal turbine operation is modelled within coastal models, as enhanced bottom friction, or momentum sinks. In Delft3D-FLOW, a state-of-the-art coastal model, turbine operation is usually represented via momentum losses, using the Porous Plate tool. However, the Porous Plate tool presents significant limitations to accurately represent energy extraction and geometry of tidal turbines. Recently, a new tool (Actuator Disc) based on the Momentum Actuator Disc Theory (MADT) was developed in Delft3D-FLOW, overcoming the aforementioned limitations and showing excellent results against laboratory data. The aim of this work is to compare the behaviour of the Actuator Disc and Porous Plate on the far-field hydrodynamics. Overall, significant differences were found, with the Porous Plate significantly underestimating the impact on instantaneous and residual flow velocities and turbulence conditions, when the turbine operates at its rated power. Consequently, MADT appears as the best alternative to investigate the far-field hydrodynamic impacts of tidal turbine operation and previous research based on the Porous Plate tool should be revisited.

The modelling of ocean waves is an integral part of coastal and offshore engineering. Both theoretical and experimental modelling methods are available and are commonly used in support of each other. In particular, due to the difficulty in measuring the velocity throughout the water column, the wave kinematics are often derived, by means of wave theory, from a measurement of the free surface elevation. However, wave theory is often based upon idealistic conditions, such as infinite spatial domains and time lengths. The question therefore arises, how well are the wave kinematics in an experimental wave tank described by wave theory? The present paper compares theoretical solutions against experimental data, for the free surface elevation and the velocity throughout the water column, to assess the ability of Stokes’ wave theory to describe the kinematics of regular waves in a short, physical, wave flume. Experimentally, the free surface elevation is measured with a set of resistive wave probes, while wave kinematic data is acquired with particle image velocimetry (PIV). For this study, ten different regular waves, of varying steepness, are generated in a 35 m long and 0.7 m deep wave flume. The theoretical solutions are computed based on Stokes 2 nd order wave theory. The presented results show error values of the order of 10–20%, indicating validity of the employed wave theory as a function of the reflection coefficient achieved in the physical wave flume. These result highlight the potential inaccuracies incurred in any wave tank if the wave theory is used to derive the kinematics from the free surface elevation without having detailed knowledge of the reflection characteristics at the point of interest in the tank.

Who can contemplate the surface of the ocean and not notice its incessant, seemingly random motion? This motion represents the net effect of long swells bearing memories of distant storms, shorter waves driven up by winds blowing nearby and over the surrounding area, and very small ripples caused by surface tension, with the pattern constantly changing and its constituents continuously traveling in different directions. Closer to shore, the ocean surface seems dominated by waves of steeper slopes with aligned crests mostly traveling towards the shore. As the crests approach the shore they get steeper, slightly disintegrating near their tops but still getting taller until they turn into plunging, roaring breakers, finally giving up their energy as they dissipate (with some reluctance) over sandy beaches or rocky shores. Whether maneuvering (or simply observing) a small craft on its way through the waves, or watching the waves crash loudly onto the beach, one can hardly fail to be impressed by their unrelenting display of energy. Waves as Energy Carriers Waves carry energy from one point to another. One observes an excellent illustration of this phenomenon when the relatively calm surface of a pond or a small lake is disturbed by a pebble or a small stone thrown into the water. A pattern of waves emanates from where the stone fell, traveling radially out and creating ever-expanding circles with alternating crests and troughs that get smaller, the farther the disturbance travels. In this case, waves carry away a portion of the kinetic energy of the stone over the water surface. A roughly similar phenomenon occurs when wind blows over a region of the ocean surface and generates waves that are higher and longer, the faster and longer the wind blows. In deep waters, the dissipation rate of the energy contained in waves is very small, enabling them to last longer and travel farther than the wind that created them. This happens for wind blowing over different regions, leading to an averaging, storing, and concentration of energy, both spatially and temporally. Because the density of sea water is about three orders of magnitude greater than that of air, energy delivered by waves is also more concentrated than wind energy.

The development of a water accounting framework that enables consistent and contextual reporting of minerals operations' water use, represents a key sustainable development objective that many Australian minerals companies have committed to. A future water accounting framework will require: • key terms and definitions for water metrics for use in the Australian minerals industry that are consistent with the wider water sector, • a suite of metrics representing key components of an operation's water balance, • a calculation method for some of the metrics, and • a reporting protocol. This paper presents the methodology that was used to derive these four components. Existing accounting frameworks were reviewed and key terms and definitions that were consistent with current regulatory requirements and voluntary approaches were proposed. A calculation methodology was developed to obtain some of the key indicators and variables that cannot be easily measured directly (eg water loss through evaporation). The adopted terms and definitions typically require a combination of empirical and estimated data. To test the proposed definitions and calculation methodology, values for water metrics were derived for eight sites, selected to test a variety of commodity and biophysical situations. The proposed framework and definitions could deal with the specific situations selected as case studies. The associated calculation methodology produced complete accounts at various levels. Some elements of the accounts could be directly checked against available data and simulations. The framework captured well variations in the reuse efficiencies; reuse efficiency could even vary for similar processes. It is concluded that this preliminary accounting framework provides a consistent language and metrics for quantifying and communicating water management, both at the intersection of the site with the surrounding landscape (through detailed reporting of site inputs and outputs) and within operational activities. Using this consistent framework for quantifying and documenting water use within an operation will provide a platform for compiling and comparing minerals industry water use with other sectors. This may support benchmarking activities within regions or companies and help demonstrate leadership amongst water users in water resource stewardship. The approach presented here is applicable to a wide variety of mining and processing activities, and biophysical environments, and following a period of stakeholder engagement and third party review, is being refined as a first step towards developing nationally consistent water use reporting for the Australian minerals industry. © 2010 Australasian Institute of Mining and Metallurgy.

One class of models for the early history of the Earth requires the present-day inventory of siderophile elements in the mantle to have been established by equilibrium partitioning between core-forming metal and mantle minerals at high pressures and temperatures deep inside the Earth. We have accordingly carried out reconnaissance experiments on the partitioning of nickel between model lower mantle magnesiowustite (Mg9 = 85 and 1.3 wt% NiO) and a model core-forming alloy. Fe 94 Ni 6 ( approximately 7 wt% Ni) at pressures between 1-40 GPa and temperatures ranging from 1200 degrees C to >2000 degrees C. Reversal experiments were also attempted. Our results highlight the difficulty of attaining equilibrium partitioning in this system and imply that partition coefficients derived from unreversed experiments should accordingly be viewed with reservation. Our data nevertheless imply that the concentration of NiO in lower mantle magnesiowustite in equilibrium with core-forming metal with approximately 7 wt% Ni would be extremely low, e.g., about 0.2 wt% NiO. Moreover, equilibrium seems to be fairly insensitive to the effects of either pressure or temperature, and so it is unlikely that magnesiowustite could acquire 1.3 wt% NiO simply by equilibrating with core-forming metal under special high P-T conditions early in Earth history. Alternative hypotheses for the present-day siderophile element inventory of the mantle are accordingly preferred.

Recent seismic evidence suggests that subducted slabs experience resistance to further descent when they encounter the 660 km seismic discontinuity. Several possible causes of this resistance are evaluated. It is concluded that the chemical composition of the lower mantle is similar to that of the upper mantle, and that compositional change is therefore unlikely to be the cause of resistance to slab penetration. The proposal that a large increase of viscosity at the 660 km discontinuity impedes descending slabs is also rejected. However, three other factors are identified, each of which is capable of causing substantial resistance to descending slabs: (1) the negative slope of the transformation of silicate spinel to Mg-perovskite + magnesiowu¨stite; (2) differentiation of oceanic lithosphere into basaltic and depleted peridotitic layers, causing the slab to be buoyant compared with surrounding mantle pyrolite between depths of 660–800 km; (3) the accumulation of former oceanic crust to produce a gravitationally stable layer of garnetite (about 50 km thick) on top of the 660 km discontinuity. The combined effects of these sources of resistance provide a filter for subducted slabs. Those slabs with seismic zones extending below 600 km may possess sufficient negative buoyancy and strength to overcome the barriers and penetrate into the lower mantle. However, the resistance causes strong buckling and plastic thickening of these slabs, which accumulate to form huge blobs or ‘megaliths’ underneath the 660 km discontinuity. In contrast, slabs with seismic zones extending no deeper than 300

An important role for the recycling of terrigenous sediments in the mantle was proposed by Armstrong [1,2]. Aspects of this hypothesis have been tested via an experimental study of phase equilibria in a composition similar to that of average upper continental crust, over the pressure interval 6–24 GPa. This composition is sufficiently close to those of the principal argillaceous and siliceous classes of pelagic sediments to use the results to interpret the behaviour of these latter sediments, in addition to terrigenous sediments, during subduction. The subsolidus phase relationships displayed by lithologies derived from these materials are complex, with major roles being played by K-hollandite, garnet, Na-clinopyroxene and stishovite. Continentally derived lithologies achieve densities similar to, or greater than, surrounding mantle at depths greater than 200 km. The geodynamical implications of these buoyancy relationships are discussed. A reconnaissance study has also been made to assess the melting and element partitioning behaviours in the subducted continental crust lithologies. During partial melting at relatively low pressures (5–10 GPa), orthoclase, wadeite and K-hollandite are eliminated near the solidus, whereas the stability field of Na-clinopyroxene extends to temperatures well above the solidus. Liquids formed by small to moderate degrees of partial melting of subducted, continentally derived lithologies in this pressure interval have high K Na ratios and high SiO2 contents. At higher pressures (16–24 GPa), the stability fields of K-hollandite and stishovite extend towards the liquidus, whereas sodium-bearing phases are eliminated much closer to the solidus. Resultant partial melts consequently possess relatively low K Na ratios and SiO2 contents. Lead possesses a high crystal-liquid partition coefficient in K-hollandite, whereas uranium is excluded. The broad stability field of K-hollandite during partial melting of these lithologies at pressures above 15 GPa therefore has the capacity to cause fractionation of lead from uranium. The petrological and geochemical implications of these results are discussed in the context of the partial melting of terrigenous materials during subduction and the interactions of the partial melts with the surrounding mantle.

An integrated analysis of molluscs, ostracods and pollen is presented from early Holocene lake marl at Llangorse Lake, Powys, South Wales. The mollusc and ostracod records reflect a littoral environment with increasing eutrophic conditions and a progressive shallowing of lake waters, although the record does not begin until around 9600 BP, probably due to the isolated location. The pollen profile provides evidence of both local and regional vegetational changes over the first two millennia of the present interglacial. Following climatic amelioration at the beginning of the Holocene, there was rapid colonization by trees and shrubs, with juniper widespread by 9800 BP, birch woodland by 9500 BP, and pine by 9300 BP. Late in the ninth millennium BP, hazel became established, to be followed during the course of the eighth millennium by elm and oak. There are indications in both the pollen and molluscan/ ostracod sequences of a relatively dry climatic regime during the early Holocene, while the pollen record suggests a short-lived climatic cooling in the middle of the tenth millennium BP.

Phase transformations in a representative oceanic crust composition (MORB) have been studied at pressures up to 28 GPa, at 1200°C and 1500°C using a multi-anvil apparatus. The assemblage of majorite + CaSiO 3 -rich perovskite + stishovite which is displayed by the MORB composition at pressures above 20 GPa was found to transform further into an assemblage containing a new aluminous phase at pressures above 25 GPa. This Al-phase has a complex composition and is also relatively sodium rich. Its X-ray diffraction pattern is similar to that observed for MgAl 2 O 4 (calcium ferrite type), and is tentatively indexed on the basis of an orthorhombic cell, yielding a zero-pressure density of 3.87(2) g/cm 3 . Majorite garnet was found to remain as the most abundant phase at the highest pressure studied (28 GPa) and does not appear likely to transform to perovskite until pressures exceed 29–30 GPa. These results imply that the basaltic crust component of subducted oceanic lithosphere would be 0.1–0.2 g/cm 3 less dense than surrounding mantle pyrolite between the depths of 660–800 km, thereby confirming an earlier hypothesis of Anderson. Moreover, the underlying layer of harzburgite would be ∼ 0.05 g/cm 3 less dense than pyrolite throughout most of this depth interval. These buoyancy relationships contribute to a barrier which impedes and may sometimes prevent the entry of slabs into the lower mantle.

Rare earth fractionation in kimberlites implies that they were produced by partial melting in the presence of residual garnet, in accordance with the widely held belief that kimberlites were formed by small degrees of partial melting of a garnet lherzolite lithology in the upper mantle. However, recent discoveries in some kimberlites of diamond xenocrysts containing syngenetic inclusions of majorite, and of xenoliths which originally contained majoritic garnet are suggestive of a deeper, transition zone origin for kimberlites. Experiments on a synthetic Group I kimberlite were carried out using an MA-8 apparatus to evaluate this possibility. At 16 GPa and 1650°C, majorite garnet (13% Al2O3) and β-M2SiO4 crystallize together on the liquidus, showing that this kimberlite magma could have been produced by a small degree of partial melting of a majorite +β-M2SiO4 (or γ-M2SiO4) assemblage in the transition zone (400–650 km). However, the first appearance of garnet well below the liquidus at 10 GPa implies that this typical kimberlite composition could not have been produced by a small degree of partial melting of garnet peridotite at depths shallower than 300 km, and casts doubt on conventional models of kimberlite petrogenesis. Isotopic, trace element and geochemical similarities imply a genetic relationship between kimberlites and ocean island basalts (OIBs). However, kimberlites were derived from a source possessing a higher Mg-number, and lower Na2O, Al2O3 and CaO contents than the OIB source. It is proposed that the ultimate source regions both of kimberlites and OIBs lie in the transition zone, in a boundary layer comprised of mixed domains of subducted former harzburgite and aesthenospheric pyrolite. The boundary layer was refertilized by partial melts derived from garnetite (former subducted oceanic crust) trapped on the 650 km discontinuity.

Lunar volcanic glasses (LVG) contain high abundances of many volatile elements as compared to mare basalts. Whereas the latter were derived from devolatilized source regions in the outer few hundred kilometers of the Moon, LVG are believed to come from a globally extensive source region at greater depths, which is relatively rich in volatile components. Abundances of volatile elements in LVG and their source regions are reviewed and in many cases found to resemble those in the Earth's mantle. Likewise, siderophile element abundances (including the highly siderophile elements Ir, Os, Au and Re) are similar in the Earth's mantle and in the source regions of LVG. The close geochemical similarity between the LVG source region, (i.e., the lunar lower mantle) and the terrestrial mantle constitutes strong evidence that protolunar material was derived primarily from the Earth's mantle. Since highly siderophile elements are believed to have been added to the Earth as an oxidised veneer at the end of its accretion, and then convectively mixed into the mantle, the removal of protolunar material from the Earth must have occurred after this late addition which may represent carbonaceous chondritic or cometary material from the outer solar system. It is widely believed that the Moon was formed by the impact of a giant planetesimal on the protoearth. The planetesimal is presumed to have formed in the inner solar system and to have differentiated internally into a mantle and core. Numerical modelling of the giant impact shows that most of the material now in the Moon would have been derived from the mantle of the impactor. However, this is contradicted by the geochemical evidence provided by LVG which implies that the Moon is composed mainly of material derived from the terrestrial mantle. A possible answer to this dilemma is provided by the observation that the LVG source region reflects the geochemistry of the “late, oxidised veneer” which accreted on the Earth. This suggests that protolunar material may have been removed from the Earth's mantle by impacts from one or more high-velocity (∼ 30 km/s) planetesimals composed mainly of ices, and derived from the outer solar system. The volatiles from the impacting planetesimals were lost after these collisions. The impacts probably ejected material from the Earth's mantle as jets of molten matter which was rapidly quenched, thereby retaining a large proportion of its terrestrial volatile inventory.

The solubilities of several oxide components of the mantle in molten iron have been measured at 16 GPa and 1700–2000°C. Their relative solubilities are as follows: FeO > Cr2O3 > MnO, V2O3, TiO2 > SiO2 > MgO > Al2O3. Oxide contents reach significant levels, varying from ∼ 10 wt% FeO to ∼ 1 wt% SiO2 at the respective metal-oxide eutectics, but are very small for MgO and Al2O3. The solubilities of these oxides are expected to increase substantially at temperatures above 2000°C and at pressures above 16 GPa and are relevant to the core-formation process in the Earth. Individual oxide species dissolve quasi-congruently under conditions which maintain oxygen fugacities near the iron-wustite buffer. However, the dissolution of mantle mineral phases is highly incongruent. If a mixture of metallic iron and mantle silicates were subjected to increasing pressures and temperatures (above 16 GPa and 2000°C) the most soluble species, FeO, would first be extracted into the metallic liquid. When the FeO activity had been lowered sufficiently, significant amounts of SiO2 would then enter this melt. At extremely high temperatures and pressures, appreciable amounts of MgO could even dissolve. These results provide a background for interpreting recent diamond-anvil experiments in which molten iron was observed to react with mantle silicates at temperatures of 2700–3700°C and pressures of 20–70 GPa. They also elucidate the nature of chemical reactions between the core and mantle which may occur in the “D” layer of the lower mantle. The differing solubilities of mantle oxides in molten iron place constraints on models of accretion of the Earth and of accompanying core-formation processes. The observation that TiO2 is undepleted in the Earth's mantle (relative to CI chondrites) makes it improbable that the core contains a substantial amount of dissolved Si or SiO2. It therefore seems likely that oxygen (as dissolved FeO) is the principal light element component of the core. This conclusion is reconciled more readily with a modified version of homogeneous accretion of the Earth than with some current models of heterogeneous accretion. Accretion of the Earth from a population of giant planetesimals would have led to complete melting of the mantle and core at very high temperatures. Under these conditions, nearly all FeO and other transition metal oxides would have been extracted from the mantle into the core, together with a significant amount of SiO2. It is difficult to reconcile the geochemical consequences of this accretion scenario with the present composition of the mantle.

We agree with Korotev that the abundances of siderophiles in Apollo 16 breccias are anomalously large but do not accept his hypothesis that they were supplied by the impact of a hitherto unknown type of meteorite possessing a grossly non-chondritic composition. Co, Ni, and other siderophiles in Apollo 16 breccias display systematic relationships which imply that a substantial proportion of these siderophiles were present as oxidised species in the lunar crust prior to the impacts of chondritic-type meteorites with the lunar surface. These pre-existing ( = indigenous) lunar siderophiles were reduced to the metallic state, to varying degrees, during impact bombardment and became alloyed with (chondritic) meteoritic metal. In this manner, the metallic particles of complex and unusual compositions now present at the A16 site were formed. The indigenous siderophiles present in the A16 crust prior to meteoritic bombardment may themselves have been derived from two distinct sources, one endogenous and the other exogenous. The first corresponds to basic and ultrabasic magmas formed by partial melting of the Moon's deep interior which were erupted and then and intimately mixed into the lunar crust. The latter represents basic and ultrabasic material ejected from the Earth's primitive crust by early meteorite bombardment, which then fell upon the lunar surface. The siderophiles present in each of these components display “terrestrial-like” abundance patterns because they are both ultimately of terrestrial origin. The anomalously high abundances of siderophiles at the A16 site may reflect an important contribution from ejected terrestrial material. Overall, the siderophile abundance patterns present in lunar highland breccias represent mixtures of terrestrial and chondritic components.

Intraplate magmas worldwide display a set of geochemical and isotopic characteristics which seem to have been produced by a common petrogenetic process connected with the subduction of oceanic lithosphere. Their source regions are to be found in quite diverse locations in the sub-continental and sub-oceanic lithospheres, and also much deeper in the mantle. However, minor-element partition relationships demonstrate that the petrogenetic processes which generated these source regions did not take place under perovskititic-facies conditions in the lower mantle and were therefore almost certainly restricted to the upper mantle. It is proposed that the source regions of intraplate magmas are formed initially in two principal locations within the upper mantle. One of these is situated near the interface between subducted lithosphere and overlying mantle, and is produced as a consequence of slab-mantle interactions mainly at depths between ∼150 and ∼300 km. Partial melting of the eclogitic crust of the slab in this region is caused by water released by bodies of subducted former serpentinite. The melts migrate into adjacent entrained refractory peridotite, hybridizing with it and thereby rendering it fertile. This is essentially an extension of the same processes which are believed to be responsible for the formation of calc-alkaline source regions, and may account for certain similarities in the geochemical and isotopic characteristics of the two associations. The relative depletions of Ti, Nb and Ta displayed by the calc-alkaline association are attributed to the retention of these elements by residual rutile during partial melting of the upper layers of the subducted eclogitic crust at depths of 80–100 km. At greater depths between 150- and 300 km, when the deeper layers of the subducted oceanic crust experience partial melting, rutile is no longer a residual phase. Therefore, Ti, Nb and Ta become highly incompatible and are effectively transferred from the slab into future source regions of intraplate magmas. Subduction may occur at shallow angles and extend for considerable horizontal distances beneath continents, producing source regions by slab-mantle interaction in or near the base of the subcontinental lithosphere. These may initiate the generation of intraplate magmas within continents, not only during subduction, but also for a considerable period following its cessation. A second type of intraplate magma source region is formed when subducted former oceanic crust is buoyantly trapped to form a gravitationally stable layer of garnetite situated on top of the 650-km seismic discontinuity. This layer is believed to be ∼70–100 km thick and to be of global extent. Partial melting of trapped oceanic crust occurs near the top of this layer, and the resultant liquids react and hybridise with an overlying boundary layer of pyrolite, causing the development of characteristic chemical and isotopic signatures in this lithology. Large diapirs eventually ascend from the fertilised boundary layer, undergoing partial melting at shallower depths and producing extensive “hot-spot” volcanism, both within both oceanic and continental regions. Small ascending diapirs from the same source region may be trapped in or near the base of the continental and oceanic lithosphere, causing local chemical heterogeneity in these regions.

The currently popular “giant impact hypothesis” attempts to explain both the origin of the Moon and the high angular momentum density of the earth-moon system in terms of a single event in which the growing Earth was struck by a martian-sized planetesimal. This hypothesis is implausible in its present form because: (1) it possesses a low intrinsic dynamical probability; (2) the collision would have melted and differentiated the Earth causing geochemical signatures quite different to those actually observed; and (3) the collision mechanics cause the Moon to be derived mainly from the mantle of the impactor, whereas geochemical evidence strongly indicates derivation of protolunar material mainly from the terrestrial mantle. These problems can be resolved by requiring that a giant impact occurring after the Earth had accreted to ∼ 70% of its present size was responsible for the high angular momentum density of the Earth-Moon system, but did not itself produce the Moon. Collisions by much smaller (e.g. 0.001–0.01 ME), high-velocity planetesimals at a very late stage of accretion were probably responsible for ejecting protolunar material from the Earth's mantle without causing gross melting and differentiation of the Earth.

Ultrasonic data for the velocities of SiO2-stishovite have been determined as a function of pressure to 10 kbar at room temperature for polycrystalline specimens hot-pressed at pressures P = 120kbar and temperatures T = 900°C. These cylindrical specimens are 2 mm in diameter and 0.9–1.4 mm long and have a grain size less than 10 μm. Compressional and shear wave velocities were measured both parallel and perpendicular to the axis of pressing and were found to be isotropic at 10 kbar with νp = 11.0 ± 0.2km/sec andνs = 6.9 ± 0.3km/sec; this shear velocit is substantially higher than that of Mizutani et al. (1972) perhaps due to the presence of crack orientations in their specimen which affected νs but not νp. The Murnaghan P-V trajectories calculated from the ultrasonic data [bulk modulus Ks = 2.5 ± 0.3Mbar and assuming (∂Ks/∂P)T = 6 ± 2] are consistent with recent hydrostatic compression data and with the shock wave compression data above 600 kbar. The combined evidence from the data of the ultrasonic and hydrostatic compression techniques suggests that the most probable value of the bulk modulus of stishovite at zero pressure is close to the upper limit of the uncertainty of our ultrasonically determined value, K0 = 2.7−2.8Mbar. Elasticity data for rutile-type oxides are not compatible with normal Ks-V0 systematics perhaps due to the neglect of non-central forces in the lattice model. These new stishovite data would make it impossible to satisfy the elasticity-density data of the lower mantle using an oxide mixture with either olivine or pyroxene stoichiometry.

The early thermal state of the Earth provides important constraints on hypotheses relating to its origin and its connection with the Moon. The currently popular giant impact hypothesis of lunar origin requires the Earth’s mantle to have been completely melted during the impact. Differentiation of a molten mantle would have produced strong chemical and mineralogical stratification, causing the mantle to become gravitationally stable and resistant to convective rehomogenization. The resulting composition and mineralogy of the upper mantle and primitive crust would have been dramatically different from those which have existed during the past 3.8 b. y. It is concluded that the Earth’s mantle was not extensively melted at the conclusion of accretion of the planet and therefore the hypothesis that the Moon was formed by the impact of a martian-sized planetesimal on the proto-Earth is probably incorrect. Nevertheless, a wide range of geochemical evidence demonstrates the existence of a close genetic relationship between the Moon and the Earth’s mantle. The key evidence relates to the processes of core formation in planetary bodies and resultant abundance patterns of siderophile elements which remain in their silicate mantles. Because of the complexity of the core formation process within a given body and the multiplicity of chemical and physical processes involved, the mantle siderophile signature is expected to be a unique characteristic. Thus, the siderophile signatures of Mars and of the eucrite parent body are quite distinct from that of the Earth’s mantle. Lunar siderophile geochemistry is reviewed in detail. It is demonstrated that a large group of siderophile elements display similar abundances in the terrestrial and lunar mantles. The similarity implies that a major proportion of the material now in the Moon was derived from the Earth’s mantle after core formation. This implication, however, does not require that the bulk compositions of the lunar and terrestrial mantles should be essentially identical, as is often assumed. Factors which may contribute to significant compositional differences between the two bodies within the context of a close genetic relationship are reviewed. The most promising mechanism for removing terrestrial material from the Earth’s mantle arises from the impacts of a number of large (0.001 to 0.01 ME) but not giant (≥ 0.1 ME) planetesimals after core formation and at the terminal stage of the Earth’s accretion. These impacts evaporated several times their own masses of mantle material and shock-melted considerably more. However, they did not lead to complete or extensive (e.g. > 50%) melting of the entire mantle. Impact-generated clouds of shock-melted spray and vapours were accelerated to high velocities in the presence of a primitive terrestrial atmosphere that co-rotated with the Earth. This provided an effective means of transferring angular momentum from the Earth to the ejected material which condensed to form a ring of Earth-orbiting planetesimals and moonlets. The Moon was formed by coagulation from material derived from the outer regions of this ring. Accretion of the Earth in the presence of the gases of the solar nebula and the co-rotating primitive terrestrial atmosphere may also have provided a mechanism for generating the rapid prograde spin of the proto-Earth.

Summary Recent shock wave data of Jeanloz & Ahrens have established the presence of a phase transformation in Fe0.94 O, occurring at approximately 70 GPa, which has profound implications for the phase behaviour of the geophysically important system FeO—MgO at high pressure. We calculate the phase diagram based on the available shock wave data for FeO and MgO, and suggest that an increase of pressure in the system FeO—MgO will result in a gradual exsolution of an almost pure FeO high-pressure phase (hpp), leaving an iron-depleted (Fe, Mg)O rocksalt (Bl) phase. Using current geophysical data, exsolution of FeO (hpp) from the rocksalt phase in the present-day lower mantle is calculated to occur only in the region immediately overlying the core. Before core segregation took place, however, the increased iron content of the rocksalt phase could place the exsolution region at lesser depths, perhaps 2000 km. Extrapolation of the solubility data of oxygen in liquid iron indicates that the liquid miscibility gap in the system Fe—FeO at atmospheric pressure should close at approximately 4250K. We construct schematic phase diagrams for the system Fe—FeO at high pressure by considering the effect of pressure upon end-member melting points, upon the extent of liquid immiscibility and upon eutectic temperatures and compositions. We suggest that at high pressure (P˜ 90 GPa) the phase diagram of Fe—FeO becomes qualitatively similar to the system Fe—FeS, displaying complete liquid miscibility and a eutectic temperature well below the melting point of pure iron. On the basis of our inferences regarding behaviour in the system Fe—FeO—MgO at high pressure, we have constructed a model for core segregation. Assuming the Earth to have accreted from the primordial solar nebula as a relatively homogeneous mixture of metallic iron and silicate phases, core segregation involving oxygen as the principal light alloying element in the core would commence at a depth where pressure is sufficiently high to cause exsolution of FeO (hpp) from the rocksalt phase, and temperature is sufficiently high to allow formation of an Fe—FeO (hpp) melt without extensively melting the silicate phases. A gravitational instability arises, leading to vertical differentiation of the Earth as molten blobs of the metal phase sink downwards to form the core and the residual depleted silicate material coalesces to form large bodies which rise diapirically upwards to form the mantle.

An hypothesis for mare basalt petrogenesis is presented. In the proposed model, both high- and low-Ti mare basalts are produced by partial melting and hybridization at depth accompanied by equilibrium between the hybrid liquids and the local olivine-pyroxenite residuum. The energy source for mare basalt formation is ultimately provided by radiogenic heating in the primordial lunar interior. The proposed dynamic model explains similarities in the Mg/(Mg+Fe) ratios and Cr2O3 contents of high-Ti and low-Ti basalts and avoids thermal problems inherent in previous models.

The most important petrological problem relating to the development of island arc systems is the origin of the basalt-andesite-dacite-rhyolite volcanic suite. Characteristics of the suite are reviewed. They include distribution of active volcanoes relative to Benioff zones, chemical composition and fractionation trends of magmas, the evolutionary development of island arcs in time, often comprising an earlier tholeiitic stage followed by a later calcalkaline stage, and the trend of magmas to become richer in potassium and other incompatible elements, the greater the height of the volcano above the Benioff zone. Recent researches in experimental petrology suggest three fractionation controls which might be responsible for various characteristics of the suite; amphibole-controlled fractionation, eclogite-controlled fractionation and direct partial melting of mantle pyrolite under conditions of high water vapour pressure. The operation of these processes is reviewed. Amphibole-controlled fractionation is limited to depths smaller than 100 km and causes a calcalkaline trend among residual liquids but with little change in K/Na and rare earth element abundances. Eclogite-controlled fractionation likewise causes a calcalkaline trend but is accompanied by increase of K/Na ratios and fractionation of the rare earth elements. This process probably occurs in the 100–150 km depth interval. Partial melting of the mantle under high P H 2 O is shown to cause the formation of basaltic magmas in the depth interval, 70–100 km. Upon rising, these magmas fractionate towards basaltic andesite and andesite compositions via the crystallization mainly of olivine. Resulting liquids display a tholeiitic differentiation trend. A model for the operation of these processes in the island arc environment is proposed. Amphibolite in the subducted oceanic crust becomes dehydrated at depths of 70–100 km under subsolidus conditions. The water produced causes partial melting to occur in the pyrolite wedge above the Benioff zone. Magmas thus produced differentiate under high P H 2 O to produce the early tholeiite stage of arc development. As the oceanic crust is sub-ducted to depths of 100–150 km, a high P H 2 O is maintained by the dehydration of serpentinite and derivative high pressure hydrated magnesium silicates. Partial melting of the quartz eclogite oceanic crust produces rhyodacite magmas. These react with overlying mantle pyrolite to form pyroxenite. Diapirs of pyroxenite rise upwards from the Benioff zone, partially melting to form magmas which fractionate by eclogite crystallization (80–150 km) and amphibole crystallization (30–100 km) thereby producing the calcalkaline phase of arc development. The residual, refractory eclogite and peridotite in the lithosphere plates which sink below about 150 km have become irreversibly differentiated and never again are able to participate in the formation of basaltic magmas at mid-oceanic ridges. The complementary differentiate is, ultimately, the continental crust, which grows through time by the accretion of island arcs and by the addition of the andesitic volcanic suite, derived as discussed above. It is estimated that about 30–60% of the volume of the mantle has passed through this process of irreversible differentiation.

Experiments on the solubility of FeO in molten iron have been carried out at temperatures between 2100 and 2550°C. The results show that liquid FeO is extensively soluble in molten iron at 2500°C and indicate that they probably become completely miscible above 2800°C. Liquid iron in equilibrium with crystalline magnesiowustite (Mg 0.8 Fe 0.2 )O which is believed to be an important mineral in the lower mantle, would dissolve about 14 mol.% of FeO at 2500°C and 40 mol.% of FeO at 2800°C. The geochemical implications of these results are discussed. It is concluded that the outer core probably contains a large amount of dissolved FeO and that oxygen is probably the principal light element in the outer core.

Substantial indigenous abundances of siderophile elements have been found to be present in the lunar highlands. The abundances of 13 siderophile elements in the parental magma of the highlands crust were estimated by using a simple model whereby the Apollo 16 highlands were regarded as being a mixture of three components (i.e. cumulus plagioclase + intercumulus magma that was parentel to the highlands crust + meteoritic contamination by ordinary chondrites). The parental magma of the highlands was found to possess abundances of siderophile elements that were generally similar to the abundances of the unequivocally indigenous siderophile elements in primitive, low-Ti mare basalts. This striking similarity implies that these estimated abundances in the parental highlands magma are truly indigenous, and also supports the basic validity of our simple model.

Experimental study of the phase boundary for the disproportionation of the inverse spinel Mg2SnO4 into its isochemical mixed oxides indicates a slope dP/dT = 40 ± 10bars/°K. This positive slope is consistent with the large entropies of inverse (relative to normal) spinels predicted from high-temperature entropy-molar volume systematics. Thermodynamic data do not support the existence of a distinctly negative slope for the proposed disproportionation of Mg2SiO4 normal spinel. Evidence from X-ray and phase equilibria studies suggests the possibility that Si4+, Mg2+, and Fe2+ share the octahedral sites in silicate spinels under mantle conditions. The consequences of this partial inverse character are a positive slope for the postulated spinel-mixed oxide phase boundary near 650 km depth, removal of a widely accepted constraint on mantle-wide convection, and anomalous elasticity-density behaviour within the transition zone.

The role of phase transformations in a mantle of pyrolite composition is reviewed in the light of recent experimental data. The pyroxene component of pyrolite transforms to the garnet structure at 300–350 km whilst olivine transforms to beta-Mg 2 SiO 4 near 400 km. Between about 500 and 550 km, beta-Mg 2 SiO 4 probably transforms to a partially inverse spinel structure whilst the CaSiO 3 component of the complex garnet solid solution exsolves and transforms to the perovskite structure. The major seismic discontinuity near 650–700 km is probably caused by disproportionation of Mg 2 SiO 4 spinel into periclase plus stishovite. At a slightly greater depth, the remaining magnesian garnet transforms to the corundum or ilmenite structure. Finally, at a depth probably in the vicinity of 800–1000 km, the (Mg,Fe)SiO 3 component of the ilmenite phase transforms to a perovskite structure whilst stishovite and some of the periclase recombine to form perovskite also. The mineral assemblage so formed is about 4% denser than mixed oxides (MgO + FeO + A1 2 O 3 + CaO + stishovite) isochemical with pyrolite. The above sequence of phase transformations in pyrolite provides a satisfactory general explanation of the elastic properties and density distribution in the mantle. In particular, there is no evidence requiring an increase of FeO/(FeO + MgO) ratio with depth. The depths at which major phase transformations occur in subducted lithosphere differ from those in ‘normal’ mantle. These differences are caused by two factors: (1) Temperatures within sinking plates are much lower than in surrounding mantle to depths of 700 km or more. (2) Irreversible chemical differentiation of pyrolite occurs at oceanic ridges. Lithosphere plates so formed consist of a layer of basaltic rocks underlain successively by layers of harzburgite, lherzolite, and pyrolite slightly depleted in highly incompatible elements (e.g. La, Ba, Rb, U). The phase-transformation behaviour of the first three of these layers differs from that of pyrolite. The effects of these and other factors connected with phase transformations on the dynamics of plate subsidence are discussed. It appears quite likely that plates penetrate the 650–700 km discontinuity, largely because the slope of the spinel disproportionation is probably positive, not negative as generally supposed. The former basaltic oceanic crust probably sinks deeply into the lower mantle, whilst the former harzburgite component of the plate may collect above the perovskite transition boundary. Phase transformations may thus serve as a kind of filter, leading to increased and irreversible mantle heterogeneity with time. The possible roles of phase transformations in causing deep-focus earthquakes and introducing water into the mantle in subduction zones are also briefly discussed.

Abundance patterns of siderophile and volatile elements imply that the Moon was derived from the Earth's mantle after the core had segregated. The relative abundances of siderophile and volatile elements in the Moon and in the Earth's mantle are obtained from a comparison of their abundances in terrestrial ocean-floor basalts and lunar low-Ti mare basalts. The abundances of a group of siderophile elements Ni, Co, W, Ir, Os, P, S and Se are found to be very similar in ocean-floor tholeiites and low-Ti mare basalts and this similarity is believed to extend to their respective source regions in the Earth's mantle and lunar interior. The abundances of the above siderophile elements in the Earth's mantle have been determined by the interaction of several complex processesunique to the Earth, which relate to core formation and non-equilibrium distribution of elements between metallic and silicate phases. Since these factors could not possibly have operated separately within the Moon, the similarity in siderophile element abundances therefore implies thatthe Moon was derived from the Earth's mantle after the Earth's core was formed. Abundance patterns of volatile elements in the Moon differ dramatically from those in ordinary chondrites and from those to be expected from condensation of a nebula of solar composition. These differences imply that the Moon was not formed from components which themselves had condensed directly from the solar nebula. For these reasons, current versions of the capture and binary planet hypotheses of lunar origin, which maintain that the Moon formed independently by accretion in the solar nebula, can be rejected. The drastic depletion of many volatile elements in the Moon compared to the Earth implies that the separation of material from the Earth's mantle to form the Moon occurred at very high temperatures. Depletions of this magnitude are best explained by the hypothesis that material from the Earth's mantle was totally evaporated, and then selectively recondensed, whilst the more volatile components were lost. The Moon then formed in Earth orbit by accretion from this volatile-depleted, mantle-derived condensate.

ScAlO3 has been synthesized at 1000°C and 120 kbar as an orthorhombic perovskite with unit cell dimensions a = 4.933 ± 0.003 A, b = 5.226 ± 0.003 A, and c = 7.193 ± 0.005 A. Ions as small as 0.73 A in radius can thus occupy the A site of this ABO3 structure type. Refinement of the structure from X ray powder diffraction data shows that the oxygen atom coordination is fourfold and fivefold, the average Al-O distance is 1.93 A, and Sc3+ has six near-neighbor oxygen atoms at an average distance of 2.12 A and two more at 2.64 A to give an average eight-coordinate distance of 2.25 A. It is inferred that both Mg2+ and Fe2+, with radii of 0.72 and 0.77 A, respectively, as well as Mn2+ (0.82 A) and Ca2+, could occupy the A sites of silicate perovskites to give mantle assemblages denser than those available to any other known structural type. These densities are deduced from the systematics of perovskite unit cell volumes, and by comparison with Hugoniot data for shocked high-pressure assemblages they are consistent with enstatite, bronzitite, fayalite, and dunite compositions forming perovskite or perovskite assemblages at very high pressures.

Most countries intend to dispose of their high-level radioactive wastes by converting them into a solidified wasteform which is to be buried within the earth. SYNROC is a titanate ceramic wasteform which has been designed for this purpose on the basis of geochemical principles. It comprises essentially rutile TiO2, ‘hollandite’ Ba(Al,Ti)Ti6O16, zirconolite CaZrTi2O7, and perovskite CaTiO3. The latter three phases have the capacity to accept the great majority of radioactive elements occurring in high-level wastes into their crystal lattice sites. These minerals (or their close relatives) also occur in nature, where they have demonstrated their capacity to survive for many millions of years in a wide range of geological environments. The properties of SYNROC and the crystal chemistry of its constituent minerals are reviewed in some detail and current formulations of SYNROC are summarized. A notable property of SYNROC it its extremely high resistance to leaching by groundwaters, particularly above 100°C. In addition, it can be shown that the capacity of SYNROC minerals to immobilize high-level waste elements is not markedly impaired by high levels of radiation damage. Current investigations are focused on developing a satisfactory production technology for SYNROC and progress towards this objective is described. The high leach resistance of SYNROC at elevated temperatures increases the range of geological environments in which the waste may be finally interred; in particular, SYNROC is well adapted for disposal in deep drill-holes, both in continental and marine environments. The fact that SYNROC is comprised of minerals which have demonstrated long-term geological stability is significant in establishing public confidence in the ability of the nuclear industry to immobilize high-level wastes for the very long periods required.

An experimental and theoretical investigation of the effect of pressure on the solubility of FeO in molten iron has been carried out. Analyses of shock-wave compression data on iron oxides combined with measurements of the Fe O bond length in “metallic” oxides suggest that the partial molar volume of FeO(V*) dissolved in molten iron is substantially smaller than that of molten wustite. Hence the effect of high pressure should be to increase the solubility of FeO in molten iron at a given temperature. This inference is confirmed by an experimental investigation of the effect of pressure on the position of the Fe FeO eutectic. Thermodynamic calculations based on these experiments yield an estimate forV* which is in reasonable agreement with the theoretical estimates. The experimental value ofV* is used to calculate the effect of high pressure upon the Fe FeO phase diagram. Solubility of FeO in molten iron increases sharply with pressure, the liquid immiscibility region contracts and disappears around 20 GPa and it is predicted that the Fe FeO phase diagram should resemble a simple eutectic system above about 20 GPa. Analogous calculations predict that the solubility of FeO in molten iron in equilibrium with magnesiowustite (Mg0.8Fe0.2)O at 2500°C increase from 14 mol.%(P = 0) to above 25 mol.% at 20 GPa. If the core formed by segregation of metallic iron originally dispersed throughout the earth, it seems inevitable that it would dissolved large amounts of FeO, thereby accounting for the observation that the density of the outer core is substantially smaller than that of pure iron under correspondingP, T conditions.

Abundance patterns of siderophile elements in the Earth's mantle have been determined by complex processes connected with internal segregation of a metallic core amounting to 32% of the Earth's mass. Although the lunar core may be only 2% of the lunar mass and formed in very different conditions to the Earth's core, the lunar siderophile pattern is closely related to that of the Earth's mantle, implying that the material in the Moon derived mainly from the Earth's mantle after core formation. New models for the accretion of the Earth which require impacts by large planetesimals in the presence of a corotating primitive terrestrial atmosphere provide a mechanism for transferring material from the Earth's mantle into geocentric orbit as a ring of proto-lunarplanetesimals, from which the Moon later accreted.

Chondritic interplanetary dust particles and also Halley's Comet dust particles from the outer parts of the solar system are enriched in volatile elements and silicon relative to Type 1 carbonaceous chondrites (asteroid belt). Rietmeijer has suggested that these chemical differences imply that there has been a radial chemical differentiation between the two source regions in the inner and outer solar system. The existence of this radial chemical zonation raises the question of whether C1 chondrites can be regarded as having primordial compositions. It is well known that the Earth's upper mantle has an Mg/Si atomic ratio (1.27) which is greater than the C1 ratio (1.05). Petrochemical studies of Venusian basalts and meteorites representing the eucrite parent body (EPB) indicate that the mantles of both Venus and the EPB are similar to Earth in that they are rich in olivine relative to pyroxene. Moreover Venus and the EPB possess Mg/Si ratios that are also similar to Earth and higher than the C1 value. Spectral evidence shows the S-type asteroids which predominate in the inner asteroid belt possess high olivine/pyroxene and Mg/Si ratios. These characteristics seem to predominate throughout the inner solar system. During formation of the terrestrial planets, elements more volatile than silicon were depleted and may have been transported outwards to recondense in the lower-temperature environment of the outer asteroid belt. Some Si may also have been lost in this manner, although not enough to alter planetary Mg/Si ratios. However, recondensation of this Si on the relatively small mass of dust particles in the asteroid belt would have caused a substantial enrichment of Si relative to Mg, which was inherited by C1 chondrites that subsequently formed in this region. It thus seems likely that it is the Mg/Si ratio of the inner planets ( ∼ 1.27), which is more representative of the solar nebula value rather than the C1 ratio. An analogous process of radial chemical fractionation may also have occurred in the outer solar nebula, with volatile elements and silicon lost from the growing giant planets being recondensed onto cosmic interplanetary dust particles and cometary bodies further out from the Sun.

Partition coefficients for some key elements between MgSiO3 perovskite and liquid have been obtained via direct measurements of compositions of coexisting crystals and liquid at 24.5 GPa and 2300°C in a modified chondritic bulk composition. The results confirm our earlier work [1] in showing that the MgSiO3 perovskite-liquid distribution coefficients for Sc, Zr and Hf are substantially greater than unity whilst those for Ca and Sm are substantially smaller than unity. These results, in conjunction with the observed composition of the upper mantle contradict hypotheses which maintain that the Earth's mantle has experienced gross differentiation controlled by the fractionation of MgSiO3 perovskite.

A polycrystalline specimen of anorthite has been hot-pressed at P = 15kbar and T = 1000°C in a piston-cylinder apparatus. Compressional (ν p )and shear(ν s ) velocities are determined as a function of pressure to 7.5 kbar at room temperature by an ultrasonic pulse transmission technique. The specimen is less than 0.5% porous and is elastically isotropic within 1%. The velocities at 7.5 kbar are ν p = 7.29km/secandν s = 3.85km/sec . These data are consistent with those for most terrestrial and lunar plagioclase rocks but not for certain anisotropic rocks and single crystals. The measured velocities demonstrate, moreover, that it is impossible to distinguish between rocks of gabbro, anorthositic gabbro, or anorthosite compositions for the 20–55 km layer of the lunar crust on the basis of seismic data alone. The mean composition of the crust could well be that of a gabbro (17% Al 2 O 3 ) rather than of an anorthositic gabbro (∼25%Al 2 O 3 ) as assumed in some current models.

Distributions of major (Si, Al, Ca, Mg, Fe) and minor elements (Cs, Rb, K, Na, Sr, Ba, Pb, Cr, Sc, Y, Yb, Ho, Sm, Nd, La, Ti, Zr, Hf, U, Th and Nb) between majorite garnet, MgSiO3-perovskite, CaSiO3-perovskite and coexisting liquids have been determined experimentally in ultrabasic and basic compositions at pressures of 15–25 GPa and temperatures of 1400–2200°C using an MA-8 apparatus. The results demonstrate the capacity of silicate perovskites to accept a wide range of normally “incompatible” elements possessing diverse ionic radii and charges into their crystal structures. All of the minor elements investigated are preferentially partitioned in Mg- or Ca-perovskites as compared to majorite garnet under subsolidus conditions. In subsolidus assemblages containing garnet (gnt) and Mg-perovskite (mpv), Sc, Ti, Zr, Hf and Nb are preferentially partitioned into perovskite with Dmpv/gnt values of 3–14 whereas there is little fractionation of rare earth elements (REE), all of which have Dmpv/gnt values of 1–2. In subsolidus assemblages containing Mg-perovskite and Ca-perovskite, Sc, Nb, Zr, and Hf are preferentially partitioned into Mg-perovskite whereas all other minor elements are strongly partitioned into Ca-perovskite. Majorite garnet which appears on the liquidus in ultrabasic compositions at ∼ 16 GPa and ∼ 2100°C is enriched in Al2O3 compared to the coexisting ultrabasic liquid by factors of 2–3 and is depleted in CaO and TiO2 by similar factors. The partition coefficients Dgnt/liq which are applicable to liquidus majorite garnet in ultrabasic compositions are Sc (1.7), Yb (1.4), Y (1.3), Hf (0.8), Zr (0.6), Sm (0.2) and less than 0.1 for K, Sr, La, Th, U, Ba, Rb and Cs. Partition coefficients (Dmpv/liq) between Mg-perovskite and ultrabasic liquids were obtained by combining garnet/liquid and subsolidus garnet/Mg-perovskite partition data obtained at similar temperatures. It was found that Sc, Hf, Zr, Ti and HREE are enriched in liquidus Mg-perovskite with Dmpv/liq values ranging between 2 and 14. Partition coefficients between Ca-perovskite (cpv), garnet and liquid were determined in a basaltic composition at ∼ 20 GPa and ∼ 2000°C.Dcpv/liq values for U, Th and Pb are remarkably high (10–20) and are also high for REE and Sr (2–6) whereas K, Rb, Cs and Ba behave as incompatible elements with Dcpv/liq values less than 0.5. The results provide strong constraints upon hypotheses which maintain that the mantle experienced extensive melting during formation of the earth followed by fractional crystallization—differentiation processes involving majorite garnet and/or MgSiO3-perovskite. The measured partition coefficients show that fractionation of garnet would be accompanied by sharp decreases of Al/Ca and Sc/Sm ratios in residual liquids. Likewise, fractionation of Mg-perovskite would cause marked variations of Lu/Hf and especially of Sc/Sm and Hf/Sm ratios in residual liquids. The observation that the present upper mantle possesses near-chondritic relative abundances of Ca, Al, Sc, Yb, Sm, Zr and Hf categorically excludes models which propose that the bulk mantle once possessed chondritic relative abundances of Mg, Si, Al, Ca and other lithophile elements, but has differentiated to form a perovskitic lower mantle (Mg/Si< 1) and a peridotitic upper mantle (Mg/Si< 1). The chemical and isotopic compositions of ancient (∼ 4.2 Ga) Western Australian zircons imply that even if the mantle were extensively melted and differentiated around 4.5 Ga, it must somehow have become effectively rehomogenised by solid state convection by 4.2 Ga. Since this scenario appears implausible on dynamic grounds, it is concluded that the mantle probably did not experience extensive melting during the formation of the Earth.

Mantle pyrolite differentiates at mid-oceanic ridges to form a layered lithosphere consisting of a basaltic crust, immediately underlain by harzburgite and further underlain by pyrolite which has experienced depletion only of highly incompatible elements (e.g., Rb, light REE). The body forces driving subduction are concentrated mainly in the upper cool, brittle layer of lithosphere, comprised of basalt and harzburgite. The lower layer of relatively ductile pyrolite is stripped off during subduction and resorbed into the upper mantle. This material, which is depleted in highly incompatible elements, provides a future source region of mid-ocean ridge basalt magmas on a timescale of 10⁹ years. The Nd, Sr, and Pb isotopic characteristics of MORBs are explained on the basis of this model. The slab, which sinks to ~600 km, is comprised mainly of former basalt and harzburgite. These differentiated layers undergo a significantly different series of phase transformations to those experienced by mantle pyrolite. Th...

Phase equilibria have been investigated in the system CaGeO 3 MgGeO 2 Al 2 O 3 and on germanium analogues of basalt and a pyrolite-related system at pressures between 2 and 11 GPa and at temperatures between 850 and 1200°C. Solid solubility relationships between garnet, ilmenite and perovskite-type phases were determined and compared with those in the corresponding silicate systems. There are some interesting similarities where the germanate equilibria are able to provide insight into the phase chemistry of corresponding silicate systems. However, there are also some significant differences that limit the ability of germanate models to predict the probable behaviour of corresponding silicate systems at higher pressures.

The partitioning of La, Sm, Yb, Sc and major elements between majorite garnet and ultrabasic liquid has been studied experimentally at 16 GPa and 2000°C. Liquidus majorite possesses a Si/Mg ratio of 1.05 as compared to the Cl chondritic ratio of 0.95, so that extraction of more than 50% of majorite from a chondritic mantle composition would be required to obtain the pyrolite Si/Mg ratio of 0.78 in the upper mantle. Liquidus majorite is enriched in Al, Sc and Yb compared to the melt by almost a factor of two and is similarly depleted in Ca, Sm and Ti. Thus extensive majorite fractionation would necessarily imprint non-chondritic Sm/Yb, Sm/Sc, Al/Ca and Al/Ti ratios on the coexisting liquid. These partition relationships demonstrate that the present upper mantle, which possesses these elements approximately in chondritic relative abundances, cannot have experienced prior separation of more than about 10% of a majorite garnet phase.

Differentiation at mid-ocean ridges generates a layered lithosphere consisting of a basaltic crust, immediately underlain by harzburgite and further underlain by pyrolite which has experienced depletion only of highly incompatible elements. The body forces driving subduction are concentrated mainly in the upper half of the lithosphere which is relatively cool and brittle. During subduction, the lower layer of relatively ductile, slightly depleted pyrolite is stripped off and resorbed into the upper mantle, thereby providing a future source region for MORB magmas. The slab which sinks to ~ 600 km is comprised mainly of differentiated former basalt and harzburgite which undergo a different series of phase transformations to those experienced by mantle pyrolite. In consequence, the former basaltic crust remains denser than surrounding mantle whereas former harzburgite becomes relatively buoyant below the 650 km seismic discontinuity. The resulting non-uniformity in stress distribution causes the slab to buckle at this depth and accrete to form a large, relatively cool ovoid “megalith” of mixed former harzburgite and basaltic crust. Heating of the megalith occurs over 1–2 b.y., leading to partial melting of the former basaltic crust. The resultant liquids contaminate regions of former harzburgite, rendering them fertile in the sense of future capacity to produce basaltic magmas. After thermal equilibration, the newly fertile, former harzburgite becomes buoyant, leading to the separation of diapirs which rise into the upper mantle. Such diapirs rising beneath sub-oceanic lithosphere experience small degrees of partial melting to produce ocean island basalts, mainly of the alkaline suite. Diapirs of fertile former harzburgite rising beneath continents become incorporated into the sub-continental lithosphere. This is a cumulative process and is ultimately responsible for the development of the chemical, physical and isotopic characteristics of the sub-continental lithosphere, including its capacity to produce alkaline magmas when subjected to small degrees of partial melting.