The Liquid Line of Descent of Anhydrous, Mantle-Derived, Tholeiitic Liquids by Fractional and Equilibrium Crystallization--an Experimental Study at 1{middle dot}0 GPa

Two series of anhydrous experiments have been performed in anend-loaded piston cylinder apparatus on a primitive, mantle-derivedtholeiitic basalt at 1·0 GPa pressure and temperaturesin the range 1060–1330°C. The experimental data provideconstraints on phase equilibria, and solid and liquid compositionsalong the liquid line of descent of primary basaltic magmasdifferentiating in storage reservoirs located at the base ofthe continental crust. The first series are equilibrium crystallizationexperiments on a single basaltic bulk composition; the secondseries are fractionation experiments where near-perfect fractionalcrystallization was approached in a stepwise manner using 30°Ctemperature steps and starting compositions corresponding tothe liquid composition of the previous, higher-temperature glasscomposition. Liquids in the fractional crystallization experimentsevolve with progressive SiO₂ increase from basalts to dacites,whereas the liquids in the equilibrium crystallization experimentsremain basaltic and display only a moderate SiO₂ increase accompaniedby more pronounced Al₂O₃ enrichment. The principal phase equilibriacontrols responsible for these contrasting trends are suppressionof the peritectic olivine + liquid = opx reaction and earlierplagioclase saturation in the fractionation experiments comparedwith the equilibrium experiments. Both crystallization processeslead to the formation of large volumes of ultramafic cumulatesrelated to the suppression of plagioclase crystallization relativeto pyroxenes at high pressures. This is in contrast to low-pressurefractionation of tholeiitic liquids, where early plagioclasesaturation leads to the production of troctolites followed by(olivine-) gabbros at an early stage of differentiation. KEY WORDS: liquid line of descent; tholeiitic magmas; equilibrium crystallization; fractional crystallization     

Petrology and Mineral Chemistry of Lower Crustal Intrusions: the Chilas Complex, Kohistan (NW Pakistan)

Mineral major and trace element data are presented for the mainrock units of the Chilas Complex, a series of lower crustalintrusions emplaced during initial rifting within the MesozoicKohistan (paleo)-island arc (NW Pakistan). Detailed field observationsand petrological analysis, together with geochemical data, indicatethat the two principal units, ultramafic rocks and gabbronoritesequences, originate from a common parental magma, but evolvedalong different mineral fractionation trends. Phase petrologyand mineral trace element data indicate that the fractionationsequence of the ultramafic rocks is dominated by the crystallizationof olivine and clinopyroxene prior to plagioclase, whereas plagioclaseprecedes clinopyroxene in the gabbronorites. Clinopyroxene inthe ultramafic rocks (with Mg-number [Mg/(Fe_tot + Mg] up to0·95) displays increasing Al₂O₃ with decreasing Mg-number.The light rare earth element depleted trace element pattern(Ce_N/Gd_N 0·5–0·3) of primitive clinopyroxenesdisplays no Eu anomaly. In contrast, clinopyroxenes from thegabbronorites contain plagioclase inclusions, and the traceelement pattern shows pronounced negative anomalies for Sr,Pb and Eu. Trace element modeling indicates that in situ crystallizationmay account for major and trace element variations in the gabbronoritesequence, whereas the olivine-dominated ultramafic rocks showcovariations between olivine Mg-number and Ni and Mn contents,pointing to the importance of crystal fractionation during theirformation. A modeled parental liquid for the Chilas Complexis explained in terms of mantle- and slab-derived components,where the latter component accounts for 99% of the highly incompatibleelements and between 30 and 80% of the middle rare earth elements.The geochemical characteristics of this component are similarto those of a low percentage melt or supercritical liquid derivedfrom subducted mafic crust. However, elevated Pb/Ce ratios arebest explained by additional involvement of hydrous fluids.In accordance with the crystallization sequence, the subsolidusmetamorphic reactions indicate pressures of 0·5–0·7GPa. Our data support a model of combined flux and decompressionmelting in the back-arc. KEY WORDS: Kohistan; Island arc; gabbro; trace element modelling; lower crustal intrusion     

Cyclicity in the Main and Upper Zones of the Bushveld Complex, South Africa: Crystallization from a Zoned Magma Sheet

The major element composition of plagioclase, pyroxene, olivine,and magnetite, and whole-rock ⁸⁷Sr/⁸⁶Sr data are presented forthe uppermost 2·1 km of the layered mafic rocks (upperMain Zone and Upper Zone) at Bierkraal in the western BushveldComplex. Initial ⁸⁷Sr/⁸⁶Sr ratios are near-constant (0·7073± 0·0001) for 24 samples and imply crystallizationfrom a homogeneous magma sheet without major magma rechargeor assimilation. The 2125 m thick section investigated in drillcore comprises 26 magnetitite and six nelsonite (magnetite–ilmenite–apatite)layers and changes up-section from gabbronorite (An₇₂ plagioclase;Mg# 74 clinopyroxene) to magnetite–ilmenite–apatite–fayaliteferrodiorite (An₄₃; Mg# 5 clinopyroxene; Fo₁ olivine). The overallfractionation trend is, however, interrupted by reversals characterizedby higher An% of plagioclase, higher Mg# of pyroxene and olivine,and higher V₂O₅ of magnetite. In the upper half of the successionthere is also the intermittent presence of cumulus olivine andapatite. These reversals in normal fractionation trends definethe bases of at least nine major cycles. We have calculateda plausible composition for the magma from which this entiresuccession formed. Forward fractional crystallization modelingof this composition predicts an initial increase in total iron,near-constant SiO₂ and an increasing density of the residualmagma before magnetite crystallizes. After magnetite beginsto crystallize the residual magma shows a near-constant totaliron, an increase in SiO₂ and decrease in density. We explainthe observed cyclicity by bottom crystallization. Initiallymagma stratification developed during crystallization of thebasal gabbronorites. Once magnetite began to crystallize, periodicdensity inversion led to mixing with the overlying magma layer,producing mineralogical breaks between fractionation cycles.The magnetitite and nelsonite layers mainly occur within fractionationcycles, not at their bases. In at least two cases, crystallizationof thick magnetitite layers may have lowered the density ofthe basal layer of melt dramatically, and triggered the proposeddensity inversion, resulting in close, but not perfect, coincidenceof mineralogical breaks and packages of magnetitite layers. KEY WORDS: layered intrusion; mineral chemistry; isotopes; magma; convection; differentiation     

Partial Crystallization of Mid-Ocean Ridge Basalts in the Crust and Mantle

Pressures at which partial crystallization occurs for mid-oceanridge basalts (MORB) have been examined by a new petrologicalmethod that is based on a parameterization of experimental datain the form of projections. Application to a global MORB glassdatabase shows that partial crystallization of olivine + plagioclase+ augite ranges from 1 atm to 1·0 GPa, in good agreementwith previous determinations, and that there are regional variationsthat generally correlate with spreading rate. MORB from fast-spreadingcenters display partial crystallization in the crust at ridgesegment centers and in both mantle and crust at ridge terminations.Fracture zones are likely to be regions where magma chambersare absent and where there is enhanced conductive cooling ofthe lithosphere at depth. MORB from slow-spreading centers displayprominent partial crystallization in the mantle, consistentwith models of enhanced conductive cooling of the lithosphereand the greater abundance of fracture zones through which theypass. In general, magmas that move through cold mantle experiencesome partial crystallization, whereas magmas that pass throughhot mantle may be comparatively unaffected. Estimated pressuresof partial crystallization indicate that the top of the partialmelting region is deeper than about 20–35 km below slow-spreadingcenters and some ridge segment terminations at fast-spreadingcenters. KEY WORDS: MORB; olivine gabbro; partial crystallization; partial melting; ridge segmentation; fracture zones; crust; mantle; lithosphere     

Metamorphic Evolution of Iron-rich Mafic Cumulates from the Otztal-Stubai Crystalline Complex, Eastern Alps, Austria

Olivine-rich rocks containing olivine + orthopyroxene + spinel+ Ca-amphibole ± clinopyroxene ± garnet are presentin the central Ötztal–Stubai crystalline basementassociated with eclogites of tholeiitic affinity. These rockscontain centimetre-sized garnet layers and lenses with garnet+ clinopyroxene ± corundum. Protoliths of the olivine-richrocks are thought to be olivine + orthopyroxene + spinel dominatedcumulates generated from an already differentiated Fe-rich () tholeiitic magma that was emplaced into shallowcontinental crust. Protoliths of the garnet-rich rocks are interpretedas layers enriched in plagioclase and spinel intercalated ina cumulate rock sequence that is devoid of, or poor in, plagioclase.U–Pb sensitive high-resolution ion microprobe dating ofzircons from a garnet layer indicates that emplacement of thecumulates took place no later than 517 ± 7 Myr ago. Aftertheir emplacement, the cumulates were subjected to progressivemetamorphism, reaching eclogite-facies conditions around 800°Cand >2 GPa during a Variscan metamorphic event between 350and 360 Ma. Progressive high-P metamorphism induced breakdownof spinel to form garnet in the olivine-rich rocks and of plagioclase+ spinel to form garnet + clinopyroxene ± corundum inthe garnet layers. Retrogressive metamorphism at T 650–680°Cled to the formation of Ca-amphibole, chlorite and talc in theolivine-rich rocks. In the garnet layers, högbomite formedfrom corundum + spinel along with Al-rich spinel, Ca-amphibole,chlorite, aspidolite–preiswerkite, magnetite, ilmeniteand apatite at the interface between olivine-rich rocks andgarnet layers at P < 0·8 GPa. Progressive desiccationof retrogade fluids through crystallization of hydrous phasesled to a local formation of saline brines in the garnet layers.The presence of these brines resulted in a late-stage formationof Fe- and K-rich Ca-amphibole and Sr-rich apatite, both characterizedby extremely high Cl contents of up to 3·5 and 6·5wt % Cl, respectively. KEY WORDS: cumulates; Variscan metamorphism; SHRIMP dating; högbomite; saline brines     

Reaction Between Ultramafic Rock and Fractionating Basaltic Magma I. Phase Relations, the Origin of Calc-alkaline Magma Series, and the Formation of Discordant Dunite

This paper presents results of modelling reaction between peridotiteand fractionating tholeiitic basalt in simple and complex silicatesystems. Reactions are outlined in appropriate binary and ternarysilicate systems. In these simple systems, the result of reactionsbetween ‘basalt’ and ‘peridotite’ maybe treated as a combination of Fe-Mg exchange and mass transferreactions at constant Fe/Mg. Fe-Mg exchange in ternary and higher-ordersystems is nearly isenthalpic, and involves a slight decreasein magma mass at constant temperature. Mass transfer reactions,typically involving dissolution of orthopyroxene and consequentcrystallization of olivine, are also nearly isenthalpic in ternaryand higher-order silicate systems, and produce a slight increasein the magma mass at constant temperature. The combined reactionsare essentially isenthalpic and produce a slight increase inmagma mass under conditions of constant temperature or constantenthalpy. Initial liquids saturated in plagioclase+olivine will becomesaturated only in olivine as a result of near-constant-temperaturereaction with peridotite, and crystal products of such reactionswill be dunite. Liquids saturated in clinopyroxene+olivine willremain on the cpx-ol cotectic during reaction with peridotite,but will crystallize much more olivine than clinopyroxene asa result of reaction, i.e., crystal products will be clinopyroxene-bearingdunite and wehrlite rather than olivine clinopyroxenite, whichwould be produced by cotectic crystallization. The Mg/Fe ratioof crystal products is ‘buffered’ by reaction withmagnesian peridotite, and dunites so produced will have high,nearly constant Mg/Fe. Production of voluminous magnesian dunitein this manner does not require crystal fractionation of a highlymagnesian olivine tholeiite or picrite liquid. Combined reaction with ultramafic wall rock and crystal fractionationdue to falling temperature produces a calc-alkaline liquid lineof descent from tholeiitic parental liquids under conditionsof temperature, pressure, and initial liquid composition whichwould produce tholeiitic derivative liquids in a closed system.Specifically, closed-system differentiates show iron enrichmentat near-constant silica concentration with decreasing temperature,whereas the same initial liquid reacting with peridotite producessilica-enriched derivatives at virtually constant Mg/Fe. Reaction between fractionating basalt and mafic to ultramaficrock is likely to be important in subduction-related magmaticarcs, where tholeiitic primary liquids pass slowly upward throughhigh-temperature wall rock in the lower crust and upper mantle.Although other explanations can account for chemical variationin individual calc-alkaline series, none can account as wellfor the characteristics shared by all calc-alkaline series.This process, if it is volumetrically important on Earth, hasimportant implications for (Phanerozoic) crustal evolution:sub-arc mantle should be enriched in iron, and depleted in silicaand alumina, relative to sub-oceanic mantle, acting as a sourcefor sialic crust It is probable that inter-occanic magmaticarcs have basement similar to alpine peridotite, in which sub-oceanicmantle has been modified by interaction with slowly ascendingbasaltic liquids at nearly constant temperature. Discordantdunite bodies in alpine pendotite may record extraction of sialiccrust from the Earth's upper mantle. ^*Present address: Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543     

Fe-rich tholeiitic liquids and their cumulate products in the Pleasant Bay layered intrusion, coastal Maine

Fe-rich tholeiitic liquids are preserved as chilled pillows and as the chilled base of a 27 meter thick macrorhythmic layer in the Pleasant Bay mafic-silicic layered intrusion. The compositions of olivine (Fo₁) and plagioclase (An₁₃₋₈) in these extremely fine grained rocks suggest that they represent nearly end stage liquids that formed by fractionation of tholeiitic basalt. Their major element compositions (∼17.5 wt% FeO_T and 54 wt%SiO₂) closely resemble highly evolved glasses in the Loch Ba ring dike and some recent estimates of end-stage liquids related to the Skaergaard layered intrusion, and are consistent with recent experimental studies of tholeiite fractionation. Their trace element compositions are consistent with extensive earlier fractionation of plagioclase, olivine, clinopyroxene, ilmenite, magnetite and apatite. The mineral assemblage of the chilled rocks (olivine, clinopyroxene, quartz, ilmenite and magnetite), apatite saturation temperatures, and very low Fe³⁺/Fe²⁺indicate conditions of crystallization at temperatures of about 950 °C and f _{O 2} about two log units below FMQ. Cumulates that lie about 3 meters above the chilled base of the macrorhythmic layer contain cumulus plagioclase, olivine, clinopyroxene, ilmenite, apatite and zircon. This mineral assemblage and the Fe-Mg ratio in clinopyroxene cores suggest that this cumulate was in equilibrium with a liquid having a composition identical to that of the chilled margin which lies directly beneath it. The high FeO_T and low SiO₂ concentrations of this cumulate (23.3 and 45.8 wt%, respectively) are comparable to those in late stage cumulates of the Skaergaard and Kiglapait intrusions. This association of a chilled liquid and cumulate in the Pleasant Bay intrusion suggests that late stage liquids in tholeiitic layered intrusions may have been more SiO₂-rich than field-based models suggest and lends support to recent experimental studies of tholeiite fractionation at low f _O2 which indicate that saturation of an Fe-Ti oxide phase should cause FeO_T to decrease in the remaining liquid. Received: 17 January 1997 / Accepted: 10 June 1997     

Petrogenetic relationships between peralkaline rhyolite dykes and mafic rocks in the post-Variscan gabbroic complex from Bocca di Tenda (northern Corsica, France)

The Lower Permian complex from Bocca di Tenda (Corsica island, France) consists of a gabbroic sequence crosscut by chilled dykes ranging in composition from basalt to trachyandesite and peralkaline rhyolite. The gabbroic sequence is mostly composed of olivine gabbronorites, quartz gabbronorites/diorites locally displaying high ilmenite amounts, and hornblende-rich tonalites. The quartz gabbronorites/diorites and the hornblende-rich tonalites have similar initial ε_Nd values (+0.9 to ?1.1) and record a fractional crystallization process driven by separation of plagioclase, pyroxene, and ilmenite. The olivine gabbronorites have slightly higher initial ε_Nd than the quartz gabbronorites/diorites and the hornblende-rich tonalites, thereby documenting that the early evolution of the melts that gave rise to the gabbroic sequence was controlled by concomitant fractional crystallization and crustal assimilation. The trachyandesite dykes are rare and rich in dark mica. The selected trachyandesite has initial ε_Nd of +0.4, which is slightly lower than the ε_Nd of the basalt dykes. The basalt and the trachyandesite dykes are most likely genetically related through a process of fractional crystallization controlled by segregation of plagioclase, clinopyroxene and minor ilmenite, and assimilation of crustal material. The peralkaline rhyolites have initial ε_Nd values ranging from +0.3 to ?0.3. Whole-rock chemical variations and trace element compositions of Na-amphibole (arfvedsonite) indicate that the peralkaline rhyolite dykes record a process of fractional crystallization mainly controlled by separation of alkali feldspar and minor ilmenite and arfvedsonite. A plausible petrogenetic hypothesis for the genesis of the peralkaline rhyolite melts implies a protracted process of fractional crystallization from the trachyandesitic melts. This fractionation process would be initially ruled by separation of plagioclase, dark mica, and minor ilmenite. An alternative hypothesis for the origin of the peralkaline rhyolite melts implies partial melting of nearly coeval amphibole-rich mafic intrusives, which formed by crustally contaminated mantle-derived melts. The genesis of the peralkaline rhyolites is in any case correlated with mantle-derived melts that experienced extensive crustal contamination.     

Igneous garnet and amphibole fractionation in the roots of island arcs: experimental constraints on andesitic liquids

To evaluate the role of garnet and amphibole fractionation at conditions relevant for the crystallization of magmas in the roots of island arcs, a series of experiments were performed on a synthetic andesite at conditions ranging from 0.8 to 1.2 GPa, 800–1,000°C and variable H₂O contents. At water undersaturated conditions and fO₂ established around QFM, garnet has a wide stability field. At 1.2 GPa garnet + amphibole are the high-temperature liquidus phases followed by plagioclase at lower temperature. Clinopyroxene reaches its maximal stability at H₂O-contents ≤9 wt% at 950°C and is replaced by amphibole at lower temperature. The slopes of the plagioclase-in boundaries are moderately negative in space. At 0.8 GPa, garnet is stable at magmatic H₂O contents exceeding 8 wt% and is replaced by spinel at decreasing dissolved H₂O. The liquids formed by crystallization evolve through continuous silica increase from andesite to dacite and rhyolite for the 1.2 GPa series, but show substantial enrichment in FeO/MgO for the 0.8 GPa series related to the contrasting roles of garnet and amphibole in fractionating Fe–Mg in derivative liquids. Our experiments indicate that the stability of igneous garnet increases with increasing dissolved H₂O in silicate liquids and is thus likely to affect trace element compositions of H₂O-rich derivative arc volcanic rocks by fractionation. Garnet-controlled trace element ratios cannot be used as a proxy for ‘slab melting’, or dehydration melting in the deep arc. Garnet fractionation, either in the deep crust via formation of garnet gabbros, or in the upper mantle via formation of garnet pyroxenites remains an important alternative, despite the rare occurrence of magmatic garnet in volcanic rocks.     

Geochemical structure of the Yoko-Dovyren layered dunite-troctolite-gabbro-norite massif,northern Baikal area

The application of the principle and algorithm of the cluster analysis of rock compositions in magmatic complexes, which were described elsewhere, made it possible to reveal the spaceless and spatial geochemical structure of the Yoko-Dovyren layered mafic-ultramafic massif. The diversity of rocks composing this intrusion was demonstrated to comprise eleven discrete geochemical types (clusters): dunites, harzburgites, melanotroctolites, troctolites, two types of olivine gabbro, two types of olivine gabbronorites, quartz gabbronorites, and granophyres. These geochemical types of rocks and the corresponding fractionation parameters (the iron atomic fraction f of mafic minerals and the anorthite concentration An of plagioclase) define a succession corresponding to the tendencies in the crystallization of a magma of respective composition. This geochemical succession is in complete agreement with the succession in which rocks were formed in the intrusion (from dunite in its bottom part to quartz gabbronorites and granophyres near its roof) and is complicated by cyclical repetitions. The main tendency revealed in the cyclic layering is as follows: cyclical intercalations consist of rocks corresponding to the neighboring members of the rock succession (plagiodunites and melanotroctolites, melanotroctolites and troctolites, troctolites and olivine gabbro, olivine gabbro and olivine gabbronorites). These tendencies are closely similar to those identified in the Kivakka intrusion, a fact suggesting that these tendencies can be common for all layered complexes of mafic and ultramafic rocks. Original Russian Text ? A.A. Yaroshevskii, S.V. Bolikhovskaya, E.V. Koptev-Dvornikov, 2006, published in Geokhimiya, 2006, No. 10, pp. 1027–1039.     

Dunite Formation Processes in Highly Depleted Peridotite: Case Study of the Iwanaidake Peridotite, Hokkaido, Japan

Dunite formation processes in highly depleted peridotites arediscussed based upon a detailed study of the Iwanaidake peridotite,Hokkaido, Japan, which consists mainly of harzburgite with asmall amount of dunite. In the harzburgites, the Mg# [= 100x Mg/(Mg + Fe²⁺)] of olivine ranges from 91·5 to 92·5,and the Cr# [= 100 x Cr/(Cr + Al)] of spinel from 30 to 70;in the dunites, the Mg# of olivine ranges from 92·5 to94 and the Cr# of spinel from 60 to 85, respectively. The NiOwt % of olivine in harzburgites ranges from 0·38 to 0·44,and in dunites from 0·35 to 0·37. The Mg# andCr# are higher and NiO wt % is lower in the dunites than inthe harzburgites surrounding the dunites. The Mg# and Cr# exhibitnormal depletion trends expected from simple partial melting,whereas the NiO wt % shows an abnormal trend. On the basis ofmass balance calculations, dunites are considered to be derivedfrom the harzburgites by a process involving incongruent meltingof orthopyroxene (orthopyroxene olivine + Si-rich melt). Hydrousconditions were necessary to lower the solidus, and thus meltingof harzburgite was probably triggered by the introduction ofhydrous silicate melt. The dunite in this massif may have formedin the mantle wedge above a subduction zone. KEY WORDS: depleted peridotite; hydrous melt; incongruent melting; residual dunite; Iwanaidake peridotite     

The Composition of Near-solidus Partial Melts of Fertile Peridotite at 1 and 1{middle dot}5 GPa: Implications for the Petrogenesis of MORB

We have determined the near-solidus melt compositions for peridotiteMM-3, a suitable composition for the production of mid-oceanridge basalt (MORB) by decompression partial melting, at 1 and1·5 GPa. At 1 GPa the MM-3 composition has a subsolidusplagioclase-bearing spinel lherzolite assemblage, and a solidusat 1270°C. At only 5°C above the solidus, 4% meltis present as a result of almost complete melting of plagioclase.This melting behaviour in plagioclase lherzolite is predictedfrom simple systems and previous experimental work. The persistenceof plagioclase to > 0·8 GPa is strongly dependenton bulk-rock CaO/Na₂O and normative plagioclase content in theperidotite. At 1·5 GPa the MM-3 composition has a subsolidusspinel lherzolite assemblage, and a solidus at 1350°C.We have determined a near-solidus melt composition at 2% meltingwithin 10°C of the solidus. Near-solidus melts at both 1and 1·5 GPa are nepheline normative, and have low normativediopside contents; also they have the highest TiO₂, Al₂O₃ andNa₂O, and the lowest FeO and Cr₂O₃ contents compared with higherdegree partial melts. Comparison of these near-solidus meltswith primitive MORB glasses, which lie in the olivine-only fieldof crystallization at low pressure, indicate that petrogeneticmodels involving aggregation of near-fractional melts formedduring melting at pressures of 1·5 GPa or less are unlikelyto be correct. In this study we use an experimental approachthat utilizes sintered oxide mix starting materials and peridotitereaction experiments. We also examine some recent studies usingan alternative approach of melt migration into, and entrapmentwithin ‘melt traps’ (olivine, diamond, vitreouscarbon) and discuss optimal procedures for this method. KEY WORDS: experimental petrology; mantle melting; near-solidus; fertile peridotite; MORB     

Liquid Immiscibility and the Evolution of Basaltic Magma

This experimental study examines relationships between alternativeevolution paths of basaltic liquids (the so-called Bowen andFenner trends), and silicate liquid immiscibility. Syntheticanalogues of natural immiscible systems exhibited in volcanicglasses and melt inclusions were used as starting mixtures.Conventional quench experiments in 1 atm gas mixing furnacesproved unable to reproduce unmixing of ferrobasaltic melts,yielding instead either turbid, opalescent glasses, or crystallizationof tridymite and pyroxenes. In contrast, experiments involvingin situ high-temperature centrifugation at 1000g (g = 9·8m/s²) did yield macroscopic unmixing and phase separation. Centrifugationfor 3–4 h was insufficient to complete phase segregation,and resulted in sub-micron immiscible emulsions in quenchedglasses. For a model liquid composition of the Middle Zone ofthe Skaergaard intrusion at super-liquidus temperatures of 1110–1120°C,centrifugation produced a thin, silicic layer (64·5 wt%SiO₂ and 7·4 wt% FeO) at the top of the main Fe-richglass (46 wt% SiO₂ and 21 wt% FeO). The divergent compositionsat the top and bottom were shown in a series of static runsto crystallize very similar crystal assemblages of plagioclase,pyroxene, olivine, and Fe–Ti oxides. We infer from theseresults that unmixing of complex aluminosilicate liquids maybe seriously kinetically hampered (presumably by a nucleationbarrier), and thus conventional static experiments may not correctlyreproduce it. In the light of our centrifuge experiments, immiscibilityin the Skaergaard intrusion could have started already at thetransition from the Lower to the Middle Zone. Thus, magma unmixingmight be an important factor in the development of the Fe-enrichmenttrend documented in the cumulates of the Skaergaard LayeredSeries. KEY WORDS: liquid immiscibility; Skaergaard; layered intrusions; experimental petrology     

Stratigraphy and Petrology of the Mulcahy Lake Layered Gabbro: An Archean Intrusion in the Wabigoon Subprovince, Ontario

The Mulcahy Lake gabbro is an Archean layered intrusion of tholeiiticbulk composition located in the Wabigoon subprovince. The intrusionis 6 km thick at the thickest and is exposed over an area of63 km². It intrudes basaltic to siliceous volcanics of the CrowLake-Savant Lake greenstone belt and is intruded by the Atikwabatholith. Zircon U-Pb data indicate crystallization at 27322+1·0/–0·9m·y. Principal phases are plagioclase, orthopyroxene, augite (andpigeonite in iron-rich rocks), olivine, hornblende and magnetite.Olivine is confined to several horizons. Apatite and then zirconare prominent accessory phases at advanced stages of fractionation.Plagioclase, pyroxenes and olivine are cumulate phases. Hornblendeis invariably an intercumulus phase. Magnetite is ubiquitousthroughout the intrusion, generally as a cumulate phase, andforms centimeter thick layers in fractionated rocks. Fractionationfollowed a tholeiitic trend with iron enrichment in the liquid. The intrusion is divided into lower, mixed, middle, upper andmarginal zones. The lower and middle zones are 2·0 and2·5 km thick respectively. The upper zone is approximately1 km thick, and the marginal zone is measured in hundreds ofmeters. A 200 m thick mixed zone is interposed between the lowerand middle zones. The base of the lower zone consists of ultramaficunits containing olivine of Fo₈₂. The top of the zone has olivineof Fo₂₈. Fractionation of the lower zone, from the floor up,was interrupted by the introduction of pristine liquid whichmixed with more dense and cooler residual liquid in the chamberto form the mixed zone. Further introduction of several minorpulses of liquid constructed the lower part of the middle zone.The upper part of the middle zone was constructed from a majorpulse of liquid plus several minor pulses each of which is representedby reversals in cryptic layering. The upper zone consists ofultramafic to iron-rich gabbro cumulates formed by cooling throughthe roof plus horizons formed by influx of pristine liquid.Marginal zone rocks represent cooling through the walls of theintrusion. Rhythmic layering is well developed in lower and middle zonecumulates. Petrofabric data show that orthopyroxene has a lineationin the plane of layering and parallel to structures suggestiveof flow. Plagioclase laths also have a preferred orientationin many cumulates and in unlayered gabbros as well. Flow, possiblylaminar, of liquid-crystal material is suggested and may belinked to the ultimate development of layering. Pressure during the course of crystallization probably was greaterthan 2 and less than 5 kb. Temperatures estimated from pyroxenesvaried from approximately 1200 to 1000 °C.f_o², is not wellconstrained but was sufficient to allow the formation of thinlocal magnetite cumulates late in the crystallization. The primarymelt was hydrous as indicated by the presence of hornblende.It is very unlikely that the melt was saturated with water duringcrystallization of the cumulate phases.     

Transformation of Archaean Lithospheric Mantle by Refertilization: Evidence from Exposed Peridotites in the Western Gneiss Region, Norway

Orogenic peridotites occur enclosed in Proterozoic gneissesat several localities in the Western Gneiss Region (WGR) ofwestern Norway; garnet peridotites typically occur as discretezones within larger bodies of garnet-free, chromite-bearingdunite and are commonly closely associated with pyroxenitesand eclogites. The dunites of the large Almklovdalen peridotitebody have extremely depleted compositions (Mg-number 92–93·6);the garnet peridotites have lower Mg-number (90·6–91·7)and higher whole-rock Ca and Al contents. Post-depletion metasomatismof both rock types is indicated by variable enrichment in thelight rare earth elements, Th, Ba and Sr. The dunites can bemodelled as residues after very high degrees (>60%) of meltextraction at high pressure (5–7 GPa), inconsistent withthe preservation of lower degrees of melting in the garnet peridotites.The garnet peridotites are, therefore, interpreted as zonesof melt percolation, which resulted in refertilization of thedunites by a silicate melt rich in Fe, Ca, Al and Na, but notTi. Previous Re–Os dating gives Archaean model ages forthe dunites, but mixed Archaean and Proterozoic ages for thegarnet peridotites, suggesting that refertilization occurredin Proterozoic time. At least some Proterozoic lithosphere mayrepresent reworked and transformed Archaean lithospheric mantle. KEY WORDS: Archaean mantle; Proterozoic mantle; Western Gneiss Region, Norway; mantle metasomatism; garnet peridotite     

Phenocryst-bulk rock composition relations of abyssal tholeiites and their petrogenetic significance

In order to infer equilibrium phase relations of abyssal tholeiites, olivine, plagioclase, augite, and pigeonite tholeiites from the ocean floor are plotted in terms of the CIPW norm proportions in the tetrahedron olivine-plagioclase-diopside-quartz. The phase relations of abyssal tholeiites have a general similarity in form to those of the experimentally studied relevant systems. Experimental studies on natural basalts allow the pressure of crystallization for abyssal tholeiitic magmas to be evaluated approximately. It appears that the pressure at which the phenocryst-stage crystallization of abyssal tholeiites takes place is as high as 2 or 3 kbar, provided that abyssal tholeiitic magmas are ‘dry’.Abyssal tholeiites could be derived from liquids that are in equilibrium with Ca-poor pyroxene in the pressure range of about 5–8 kbar. Major element chemistry of abyssal tholeiites is incompatible with the view that these tholeiitic basalts are derived from picritic magma by olivine fractionation.     

Melt variability in percolated peridotite: an experimental study applied to reactive migration of tholeiitic basalt in the upper mantle

Melt-rock reaction in the upper mantle is recorded in a variety of ultramafic rocks and is an important process in modifying melt composition on its way from the source region towards the surface. This experimental study evaluates the compositional variability of tholeiitic basalts upon reaction with depleted peridotite at uppermost-mantle conditions. Infiltration-reaction processes are simulated by employing a three-layered set-up: primitive basaltic powder (‘melt layer’) is overlain by a ‘peridotite layer’ and a layer of vitreous carbon spheres (‘melt trap’). Melt from the melt layer is forced to move through the peridotite layer into the melt trap. Experiments were conducted at 0.65 and 0.8 GPa in the temperature range 1,170–1,290°C. In this P-T range, representing conditions encountered in the transition zone (thermal boundary layer) between the asthenosphere and the lithosphere underneath oceanic spreading centres, the melt is subjected to fractionation, and the peridotite is partially melting (T _s ~ 1,260°C). The effect of reaction between melt and peridotite on the melt composition was investigated across each experimental charge. Quenched melts in the peridotite layers display larger compositional variations than melt layer glasses. A difference between glasses in the melt and peridotite layer becomes more important at decreasing temperature through a combination of enrichment in incompatible elements in the melt layer and less efficient diffusive equilibration in the melt phase. At 1,290°C, preferential dissolution of pyroxenes enriches the melt in silica and dilutes it in incompatible elements. Moreover, liquids become increasingly enriched in Cr₂O₃ at higher temperatures due to the dissolution of spinel. Silica contents of liquids decrease at 1,260°C, whereas incompatible elements start to concentrate in the melt due to increasing levels of crystallization. At the lowest temperatures investigated, increasing alkali contents cause silica to increase as a consequence of reactive fractionation. Pervasive percolation of tholeiitic basalt through an upper-mantle thermal boundary layer can thus impose a high-Si ‘low-pressure’ signature on MORB. This could explain opx + plag enrichment in shallow plagioclase peridotites and prolonged formation of olivine gabbros.     

The geochemistry of repetitive cyclical volcanism from basalt through rhyolite in the uchi-confederation greenstone belt,canada

Archean volcanic rocks in the Confederation Lake area, northwestern Ontario, Canada, are in three mafic to felsic cycles collectively 8,500 to 11,240 m thick. Each cycle begins with pillowed basalt and andesite flows and is capped with andesitic to rhyolitic pyroclastic rocks and minor flows. Seventy five samples from this succession were analyzed for major and trace elements including the rare earth elements. In two cycles, tholeiitic basalts are overlain by calcalkaline andesite to rhyolite. In the third, cycle, the tholeiitic basalts are overlain by tholeiitic rhyolites. Fe enrichment in basalts is accompanied by depletion of Ca, Al, Cr, Ni, and Sr, and enrichment in Ti, P, the rare earth elements, Nb, Zr, and Y. This is interpreted as open system fractionation of olivine, plagioclase, and clinopyroxene. Si enrichment in dacites and rhyolites is attributed to fractional crystallization of plagioclase, K-feldspar, and biotite. Tholeiitic basalt liquids are believed to be mantle-derived. Intercalated andesites with fractionated rare earth patterns appear to be products of mixing of tholeiitic basalt and rhyolite liquids and, andesites with flat rare earth patterns are probably produced by melting of previously depleted mantle. Felsic magmas are partial melts of tholeiitic basalt or products of liquid immiscibility in a tholeiitic system perhaps involving extreme fractionation in a high level magma chamber, and assimilation of sialic crust. It is concluded that Archean cyclical volcanism in this area involves the interplay of several magmatic liquids in processes of fractional crystallization, magma mixing, liquid immiscibility, and the probable existence of compositionally zoned magma chambers in the late stages of each cycle. The compositionally zoned chambers existed over the time period represented by the upper felsic portion of each cycle.     

A mafic-ultramafic cumulate sequence derived from boninite-type melts (Niagara Icefalls,northern Victoria Land,Antarctica)

The layered sequence from Niagara Icefalls (northern Victoria Land, Antarctica) is related to the Cambrian-Early Ordovician Ross Orogeny. The sequence consists of dunites, harzburgites, orthopyroxenites, melagabbronorites and gabbronorites of cumulus origin. The Mg# of olivine, spinel, orthopyroxene and clinopyroxene from these rocks yields positive correlations, thus indicating formation from melts that mainly evolved through fractional crystallisation. The following fractionation sequence was identified: olivine (up to 94 mol% forsterite) + Cr-rich spinel → olivine + orthopyroxene ± spinel → orthopyroxene → orthopyroxene + anorthite-rich plagioclase ± clinopyroxene. Clinopyroxenes retain the peculiar trace element signature of boninite melts, such as extremely low concentrations of HREE and HFSE, and LILE enrichment over REE and HFSE. U–Pb isotope data on zircons separated from a gabbronorite have allowed us to constrain the age of emplacement of the Niagara Icefalls sequence at ∼514 Ma. The occurrence of inherited zircons dated at ∼538 Ma indicates that the boninitic melts experienced, at least locally, crustal contamination. The Niagara Icefalls sequence can be related to a regional scale magmatic event that affected the eastern margin of the Gondwana supercontinent in the Middle Cambrian. We propose that the formation of the sequence was associated with the development of an embryonic back-arc basin in an active continental margin. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

High and low pressure phase equilibria of a mildly alkalic lava from the 1965 Surtsey eruption: Experimental results

Melting experiments have been performed on a primitive, mildly alkalic glassy lava (10 wt.% MgO) from the 1965 eruption of the Surtsey volcano located at the tip of the south-eastern propagating rift zone of Iceland. At atmospheric pressure, approximately on the FMQ oxygen buffer, olivine (Fo₈₁) crystallizes from 1240°C, followed by plagioclase (An₇₀) from 1180°C and augite from 1140°C. The experimental glasses coexisting with olivine, plagioclase and augite are ferrobasaltic enriched in FeO (13.6–14.2 wt.%) and TiO₂ (4.0–4.4 wt.%). In high pressure, piston-cylinder, graphite-controlled runs, olivine occurs as the liquidus phase until 14 kbar, above which augite is the liquidus phase. Low-Ca pyroxene is not a liquidus phase at any pressure. The high pressure liquids are, relative to the one atmosphere liquids, significantly enriched in Al₂O₃ and Na₂O and depleted in CaO as a result of changes in the crystallizing assemblages. Furthermore, liquidus augite is dominantly subcalcic and shows significant enrichment in Al and depletion in Ti. Subliquidus plagioclase is enriched in sodium relative to low pressure phase compositions. Evaluated in normative projections, contrasting liquid lines of descent are revealed as a function of pressure. At one atmosphere, the multisaturated liquids are located close to the thermal divide defined by the plane olivine-plagioclase-augite, but appear, with advanced degrees of crystallization, to be moving away from the thermal divide toward normative quartz. The augites crystallizing in the one atmosphere experiments are calcic and slightly nepheline normative. In the 10 and 12.5 kbar experiments, the augites become subcalcic and dominantly hypersthene normative. Because of this shift in augite compositions, transitional basaltic liquids may at high pressure evolve from the tholeiitic side of the olivine-plagioclase-diopside normative divide onto the alkalic side. With increasing pressure above 15 kbar, the liquidus augite compositions move back toward the olivine-plagioclase-diopside normative divide.