Mineral Chemistry and Metasomatic Growth of Aluminous Enclaves in Gedrite--Cordierite-Gneiss from Southwestern New Hampshire, USA

The aluminous enclaves occur in gedrite-cordierite-gneissesof the Middle Ordovician Ammonoosuc Volcanics, and are composedof combinations of the aluminous minerals sillimanite (Sill),kyanite, corundum (Cor), staurolite (St), sapphirine (Sa), andspinel (Sp), which are set in a matrix of cordierite (Crd) orplagioclase (Plag). Generally, where plagioclase is present,both it and the aluminous minerals are separated from gedrite(Ged) and rare hornblende (Hbl) by cordierite. The enclavesarc interpreted to have formed near the peak of Acadian (Devonian)metamorphism at sillimanite-staurolite-muscovite grade by reactionsthat were encountered during the pressure decrease which accompaniedthe rise of gneiss domes in the region. The enclaves are divided into two main types: (1) enclaves ofcordierite surrounding aluminous minerals; and (2) enclavesof cordierite and plagioclase surrounding aluminuous minerals.Sapphirine grains contain between 9?2 and 9?3 Al atoms per formulacalculated to 14 cations. Staurolites from the enclaves areMg-rich and have (Fe²⁺+ Mn)/(Fe²⁺+Mn+Mg) ratios of 0-59–0?64. The textures and mineralogy of the enclaves suggest that theserocks originally consisted of Ged+Sill?Qz?Hbl?Sp?Plag. Theseminerals reacted to form Crd+Aluminous Minerals?Plag. The mineralogyof both main types of enclaves can be explained by two analogoussets of continuous Fe-Mg reactions:The structure of the enclavessuggests that the mineral growth by the above reactions wasdiffusion controlled, which would have resulted from oversteppingthe above reactions (i.e. the P change exceeded the reactionrate). Therefore, chemical potential gradients (relative mobilityof diffusing components) between gedrite and sillimanite controlledthe location of mineral growth. The Fe-Mg ratio of the bulkcomposition and the proportions of non-Fe-Mg minerals (quartzand sillimanite) appear to determine which continuous Fe-Mgreactions were encountered. Examples of mineral sequences in the cordierite enclaves are:Sill (core)/St+Crd/Ged (matrix); Cor+Crd (core)/Ged (matrix),and Sill (core)/St+Crd/Sa+Crd/Ged (matrix). Examples of themineral sequences in the cordierite-plagioclase enclaves are:Sill (core)/St+Plag/Plag+Crd/Hbl+Ged (matrix); Cor+Plag (core)/St+Plag/Sa+Plag/Ged+ Hbl (matrix); and St+Plag (core)/Plag+Crd/Ged+Hbl (matrix). P–µ_FeMg–1 diagrams proved to be an importanttool for understanding and illustrating the development of theenclaves. These diagrams allow one to view simultaneously allthe discontinuous and continuous Fe-Mg reactions along a P–µ_H2O(or T) rock path. With this information it is possible to determinequalitatively which reactions and what sequence of reactionsmight be encountered by bulk compositions with variable Fe-Mgratios and modal proportions of phases.     

Sapphirine+Forsterite and Sapphirine+Humite-Group Minerals in an Ultra-Magnesian Lens from Kuhi-lal, SW Pamirs, Tajikistan: Are these Assemblages Forbidden?

Sapphirine occurs with humite-group minerals and forsteritein Precambrian amphibole-facies rocks at Kuhi-lal, SW PamirMountains, Tajikistan, a locality also for talc+kyanite magnesiohornblendewhiteschist. Most of these sapphirine-bearing rocks are graphiticand sulfidic (pyrite and pyrrhotite) and contain enstatite,clinohumite or chondrodite, spinel, rutile, gedrite, and phlogopite.A phlogopite schist has the assemblage with X_Fe = Fe/(Fe+Mg)increasing as follows: chlorite (0-003)<phlogopite (0.004–0.005)sapphirine (0.004–0.006) enstatite (0-006)forsterite (0-006–0-007)<spinel (0-014). This assemblage includes the incompatiblepair sapphirine+forsterite, but there is no textural evidencefor reaction. In one rock with clinohumite, X_Fe increases asfollows: clinohumite (0-002) <sapphirine (0-003) <enstatite(0-004–0-006) <spinel (0-010). Ion microprobe and wet-chemicalanalyses give 0-57–0-73 wt.% F in phlogopite and 0-27wt.% F in chlorite in the phlogopite schist; 0-04, 1.5–1.9,and 4.4 wt.% F in forsterite, clinohumite, and chondrodite,respectively; and 0-0-09 wt.% BeO and 0-05–0-21 wt.% B₂O₃in sapphirine. Stabilization of sapphirine+clinohumite or sapphirine+chondroditeinstead of sapphirine+phlogopite is possible at high F contentsin K-poor rocks, but minor element contents appear to be toolow to stabilize sapphirine as an additional phase with forsterite+enstatite+spinel.Although sapphirine+forsterite is metastable relative to spinel+enstatitein experiments conducted at a_H2O=1 in the MgO-Al₂O₃-SiO₂-H₂Osystem, it might be stabilized at a_H2O0.5, P4 kbar, T650–700C.Textures in the Kuhi-lal whiteschists suggest a polymetamorphicevolution in which the rocks were originally metamorphosed atT650C, P 7 kbar, conditions under which sapphirine+clinohumiteand sapphirine+chondrodite are inferred to have formed, andsubsequently affected by a later event at lower P, similar T,and lower a_H2O. The latter conditions were favorable for sapphirine+forsteriteto form in a rock originally containing chlorite+forsterite+spinel+enstatite.     

Distribution of Ferric Iron in some Upper-Mantle Assemblages

The distribution of ferric iron among the phases of upper-mantlerocks, as a function of pressure (P), temperature (T) and bulkcomposition, has been studied using ⁵⁷Fe Mssbauer spectroscopyto determine the Fe³⁺/Fe ratios of mineral separates from 35peridotite and pyroxenite samples. The whole-rock Fe³⁺ complementof a peridotite is typically shared approximately evenly amongthe major anhydrous phases (spinel and/or garnet, orthopyroxeneand clinopyroxene), with the important exception of olivine,which contains negligible Fe³⁺. Whole-rock Fe³⁺ contents areindependent of the T and P of equilibration of the rock, butshow a well-defined simple inverse correlation with the degreeof depletion in a basaltic component. Fe³⁺ in spinel and inboth pyroxenes from the spinel Iherzolite facies shows a positivecorrelation with temperature, presumably owing to the decreasein the modal abundance of spinel. In garnet peridotites, theFe³⁺ in garnet increases markedly with increasing T and P, whereasthat in clinopyroxene remains approximately constant. The complexnature of the partitioning of Fe³⁺ between mantle phases resultsin complicated patterns of the activities of the Fe³⁺ -bearingcomponents, and thus in calculated equilibrium f_O2, which showlittle correlation with whole-rock Fe³⁺ or degree of depletion.Whether Fe³⁺ is taken into account or ignored in calculatingmineral formulae for geothermobarometry can have major effectson the resulting calculated T and P. For Fe-Mg exchange geothermometers,large errors must occur when applied to samples more oxidizedor reduced than the experimental calibrations, whose f_O2 conditionsare largely unknown. Two-pyroxene thermometry is more immuneto this problem, and probably provides the most reliable P—Testimates. Accordingly, the convergence of P—T valuesderived for a given garnet peridotite assemblage may not necessarilybe indicative of mineral equilibrium. The prospects for thecalculation of accurate Fe³⁺ contents from electron microprobeanalyses by assuming stoichiometry are good for spinel, uncertainfor garnet, and distinctly poor for pyroxenes. KEY WORDS: mantle; oxidation; partitioning; peridotite; thermobarometry ^*Corresponding author. Present address: School of Earth and Ocean Sciences, University of Victoria, P.O. Box 1700, Victoria, B.C., V8W 2Y2, Canada     

Phase Relations on the Actinolite-Pargasite Join

Phase relations along the join Ca₂Mg₄Fe²⁺Si₈O₂₂ (OH)₂ (Actinolite)-NaCa₂Mg_3?2Fe_0?8²⁺AlSi₆Al₂O₂₂(OH)₂ (Pargasite) have been studied at P_H2O = 1 kb andthe oxygen fugacities defined by the iron-wustite(IW) buffer. Actinolite and bornblende are separated by a solvus and thefield of actinolite+hornblende+vapor is present in the regionbetween Ac₈₅Pa₁₅ and Ac₅₅ Pa₄₅ at 680 ?C. Complete miscibilityis achieved at 720 ?C. At temperatures higher than the solvusthere is a continuous solid solution series between the twoend members. The stability field of amphibole solid solutiongradually increases with increasing pargasite content in actinolite.The phase assemblages at temperatures higher than those of asolid solution series between the two end members change withincreasing pargasite content in the bulk composition as follows;Act+Cpx+Qz+V, ActHbl+Cpx+Opx+Qz+V, Hbl+Cpx+Opx+Pl+V and Hbl+Cpx+Pl+Ol+V. In comparison with the Fe-free system, the extent of the miscibilitygap between actinolite and hornblende is reduced by an increasein the Fe²⁺ content. The present study should provide an adequatebasis for the interpretation of actinolite-hornblende pairsin metamorphic rocks.     

Chemistry of Kornerupine and Associated Minerals, a Wet Chemical, Ion Microprobe, and X-Ray Study Emphasizing Li, Be, B and F Contents

Kornerupine and associated minerals in 31 samples of high-graderocks relatively rich in Al and Mg were analysed by wet chemistry,ion microprobe mass analyser, electron microprobe and X-raypowder diffraction. For 11 samples of kornerupine and threesamples of biotite (F only) analysed by both wet chemical andion microprobe methods, the best agreement was obtained forB₂O₃, whereas the ion microprobe Li₂O values were systematicallysomewhat higher than the wet chemical values. The wet chemicalmethods give Li₂O=0–0?19 wt.%; BeO=0–0?032 wt.%;B₂O₃=0–4?01 wt.%; and F=0?07–0?77 wt.% in kornerupine,whereas ion microprobe analyses on other kornerupines give valuesup to 0?35 wt.% Li₂O, O066 wt.% BeO, and 4?72 wt.% B₂O₃. Thesum B+Al+Fe³⁺+Cr is close to 6?9 atoms per 22 (O, OH, F) or21?5 (O) in kornerupine. In general, Li/Fe ratios decrease as follows: kornerupine ?sapphirinebiotite> Crd (Na<0?03 per 18 oxygens)>tourmaline, garnet,orthopyroxene. However, for cordierite with Na>004, Li/Fedecreases as follows: cordierite>kornerupine. Sapphirineand sillimanite are the only associated minerals to incorporatesignificant boron (0?1–0?85 wt.% B₂O₃) and then only whenthe single site for B in kornerupine is approaching capacity.Sillimanite B₂O₃ contents increase regularly with kornerupineF. Fractionation of fluorine increases as follows: kornerupine<biotite<tourmaline,and K^krn-BtD=(F/OH)^Krn/(F/(OH)^Bt (assuming ideal anion composition)increases with biotite Ti. Kornerupine B₂O₃ content is a measureof B₂O₃ activity in associated metamorphic fluid, whereas sillimaniteB₂O₃ content increases with temperature, exceeding 0?4 wt.%whenT=900?C at very low water activities. New data on 11 kornerupines and literature data indicate thatthe unit cell parameters a, c, and V decrease with increasingB content and b, c, and V increase with increasing Fe³⁺ content.In Fe³⁺-poor kornerupines, b increases with Mg and with (Mg+ Fe²⁺) but the effect of Mg on b via the substitution ^VIMg+^IVSi=^VIAl+^IVAloverwhelms the effect of Fe²⁺=Mg substitution.     

Crystallization of Chromite and Chromium Solubility in Basaltic Melts

The equilibrium between chromite and melt has been determinedon four basalts at temperatures of 1200–1400?C over arange of oxygen fugacity (f_o2) and pressures of 1 atm and 10kb. The Cr content of chromite-saturated melts at 1300?C and1 atm ranges from 0?05 wt.% Cr₂O₃ at a log f_o2= –3 to1?4 wt.% at a log f_o2=–12?8. The Cr²⁺/Cr³⁺ of melt increaseswith decreasing f_o2 and is estimated by assuming a constantpartitioning of Cr³⁺ between chromite and melt at constant temperature.The estimated values of Cr²⁺/Cr³⁺ in the melt are at f_o2 valuesof 4–5 orders of magnitude lower than the equivalent Fe²⁺/Fe³⁺values. The Cr/(Cr+Al) of chromite coexisting with melt at constanttemperature changes little with variation of f_o2 below log f_o2=–6.Five experiments at 10 kb indicate that Cr₂O₃ dissolved in themelt is slightly higher and the Cr/(Cr + Al) of coexisting chromiteis slightly lower than experiments at 1 atm pressure. Thus variationin total pressure cannot explain the large variations of Cr/(Cr+ Al) that are common to mid-ocean ridge basalt (MORB) chromite. Experiments on a MORB at 1 atm at f_o2 values close to fayalite-magnetite-quartz(FMQ) buffer showed that the Al₂O₃ content of melt is highlysensitive to the crystallization or melting of plagioclase,and consequently coexisting chromite shows a large change inCr/(Cr + Al). It would appear, therefore, that mixing of a MORBmagma containing plagioclase with a hotter MORB magma undersaturatedin plagioclase may give rise to the large range of Cr/(Cr +Al) observed in some MORB chromite.     

Pumpellyite-Actinolite Facies Schists of the Taveyanne Formation near Lo?che, Valais, Switzerland

Electron microprobe analyses are presented for new-formed mineralsfrom a small exposure of semi-schistose Taveyanne Formationof the pumpellyite-actinolite facies near Lo?che, Valais. Comparisonsare drawn with minerals of other low-grade metamorphic areas,especially in southern New Zealand. Sphene shows considerablesubstitution of Ca(Al,Fe)SiO₄(OH) for CaTiSiO₅. Epidotes aresharply divided into early pistacitic (Ps = 0.28–0.37)and later clinozoisitic varieties (Ps = 0.11–0.19). Pumpellyitesrange from pumpellyite-(Fe) to pumpellyite-(Al) and are generallyless Fe-rich than those of zeolite and prehnite-pumpellyitefacies. Pumpellyite inclusions in albitized plagioclase areparticularly low in Mg. Actinolites are low in A1₂O₃, TiO₂,and Na₂O, essentially identical compositions being nucleatedon detrital augite, hornblende, and in the matrix. Phengitesare also extremely low in Na₂O and TiO₂. Chlorites are ripidolites.Albitized clastic plagioclase has the composition An0.7–1.6and albite in clinozoisite-calcite-albite-phengite-chloriteveins An2.1–2.3. Calcites carry minor Mn > Fe ? Mg.New-formed iron oxides are absent, whereas pyrrhotite and minorpyrite occur in one rock, buffering fs₂ and indicating low fo₂. Ratios Mg: Fe^* (Fe^* = total Fe) in coexisting chlorites andA1, Na-poor actinolites vary sympathetically both in the Lo?cheand southern New Zealand rocks here considered, giving K_D =(Mg/Fe^*) _actlnolIte/(Mg/Fe^*)_chlorle = 1.72. Mg/Fe^* ratios inpumpellyites tend to vary sympathetically with those of coexistingchlorites and actinolites but are more variable. Substitutionof (Fe, Mg)Si for A1₂ in phengitic micas and chlorites variessympathetically in the same suites between mafic volcanic andmore pelitic extremes. Various minor elements also behave ina consistent fashion, indicating an encouraging tendency towardsequilibrium. Variable (though small) A1₂O₃ contents of actinolite,Fe: Al ratios in epidotes and pumpellyites, and Mg: Fe^* ratiosin phengites, even within a single grain, are evidence of short-rangedisequilibrium; metamorphic equilibration is evidently easierbetween some crystal structures and structural sites than betweenothers. In phase rule analysis of assemblages in such rocks it is commonlynecessary to treat Fe²O₃, FeO, and MgO as separate componentsand it may also be necessary to regard CO² as an inert componentand/or to interpret observed assemblages as of low variance.The presence of the Ca-Al silicates and sphene indicates verylow X_co2 in the metamorphic fluids in all rocks examined exceptan albite-chlorite-calcite-quartz-anatase assemblage. But higherAn in albites than in isofacial and in greenschist facies rocksof southern New Zealand can be ascribed to significantly higherXco₂ at Lo?che, especially in the veins, than in New Zealand. Pumpellyite and epidotes of the pumpellyite-actinolite faciestend to be lower in Fe and richer in Al than those of lowergrade facies. Important reactions include those of the formpumpellyite-(Fe³⁺)+chlorite+quartz+H₂=pumpellyite-(Al)+actinolite,and pumpellyite+chlorite+quartz- ‘epidote’+actinolite+water.Careful selection of pumpellyite and chlorite compositions isrequired for experimental and chemographic analysis of pumpellyitestability. In the absence of critical data, temperatures ofabout 250–350? and pressures of several kilobars are provisionallysuggested for the Lo?che metamorphism.     

Transition from the Zeolite to Prehnite-Pumpellyite Facies in the Karmutsen Metabasites, Vancouver Island, British Columbia

The upper Triassic Karmutsen metabasites from northeast VancouverIsland, B.C., are thermally metamorphosed by the intrusion ofthe Coast Range Batholith. The amygdaloidal metabasites developedin the outer portion of the contact aureole show a progressivemetamorphism from zeolite to prehnite-pumpellyite facies. Thesize of an equilibrium domain is extremely small for these metabasites,and the individual amygdule assemblages are assumed to be inequilibrium. Two major calcite-free assemblages (+chlorite+quartz)are characteristic: (i) laumontite+pumpellyite+epidote in thezeolite facies and (ii) prehnite+pumpellyite+epidote in theprehnite-pumpellyite facies. The assemblages and compositionsof Ca-Al silicates are chemographically and theoretically interpretedon the basis of the predicted P-T grid for the model basalticsystem, CaO-MgO-A1₂O₃-Fe₂O₃-SiO₂-H₂O. The results indicate:(1) local equilibrium has been approached in mineral assemblagesand compositions; (2) the X_Fe³⁺ values in the coexisting Ca-Alsilicates decrease from epidote, through pumpellyite to prehnite;(3) with increasing metamorphic grade, the Fe³⁺ contents ofepidotes in reaction assemblages decrease in the zeolite facies,then increase in the prehnite-pumpellyite facies rocks. Suchvariations in the assemblages and mineral compositions are controlledby a sequence of continuous and discontinuous reactions, andallow delineation of T-X_Fe³⁺ relations at constant pressure.The transition from the zeolite to prehnite-pumpellyite faciesof the Karmutsen metabasites is defined by a discontinuous reaction:0·18 laumontite+pumpellyite+0·15 quartz = 1·31prehnite+ 0·78 epidote+0·2 chlorite+ 1·72H₂O, where the X_Fe³⁺ values of prehnite, pumpellyite and epidoteare 0·03, 0·10 and 0·18, respectively.These values together with available thermodynamic data andour preliminary experimental data are used to calculate theP-T condition for the discontinuous reaction as P = 1·1±0·5 kb and T = 190±30°C. The effectsof pressure on the upper stability of the zeolite facies assemblagesare discussed utilizing T-X_Fe³⁺ diagrams. The stability of thelaumontite-bearing assemblages for the zeolite facies metamorphismof basaltic rocks may be defined by either continuous or discontinuousreactions depending on the imposed metamorphic field gradient.Hence, the zeolite and prehnite-pumpellyite facies transitionboundary is multivariant.     

An Experimentally Constrained Petrogenetic Grid in the Silica-Saturated Portion of the System KFMASH at High Temperatures and Pressures

Experiments in the quartz-saturated part of the system KFMASHunder fO₂ conditions of the haematite–magnetite bufferand using bulk compositions with X_Mg of 0·81, 0·72,0·53 define the stability limits of several mineral assemblageswithin the P–T field 9–12 kbar, 850–1100°C.The stability limits of the mineral assemblages orthopyroxene+ spinel + cordierite ± sapphirine, orthopyroxene + garnet+ sapphirine, sapphirine + cordierite + orthopyroxene and garnet+ orthopyroxene + spinel have been delineated on the basis ofP–T and T–X pseudosections. Sapphirine did not appearin the bulk composition of X_Mg = 0·53. A partial petrogeneticgrid applicable to high Mg–Al granulites metamorphosedat high fO₂, developed in our earlier work, was extended tohigher pressures. The experimental results were successfullyapplied to several high-grade terranes to estimate P–Tconditions and retrograde P–T trajectories. KEY WORDS: KFMASH equilibria; experimental petrogenetic grid at high fO₂     

Petro-tectonic Imprints in the Sapphirine Granulites from Anantagiri, Eastern Ghats Mobile Belt, India

Sapphirine granulite occurring as lenses in charnockite at Anantagiri,Eastern Ghat, India, displays an array of minerals which developedunder different P-T-X conditions. Reaction textures in conjunctionwith mineral chemical data attest to several Fe-Mg continuousreactions, such as

1. spinel+rutile+quartz+MgFe_–1=sapphirine+ilmenite
2. cordierite=sapphirine+quartz+MgFe_–1
3. sapphirine+quartz=orthopyroxene+sillimanite+MgFe_–1
4. orthopyroxene+sapphirine+quartz=garnet+MgFe_–1
5. orthopyroxene+sillimanite=garnet+quartz+MgFe_–1
6. orthopyroxene+sillimanite+quartz+MgFe_–1=cordierite.

Calculated positions of the reaction curves in P-T space, togetherwith discrete P-T points obtained through geothermobarometryin sapphirine granulite and the closely associated charnockiteand mafic granulite, define an anticlockwise P-T trajectory.This comprises a high-T/P prograde metamorphic path which culminatedin a pressure regime of 8?3 kb above 950?C, a nearly isobariccooling (IBC) path (from 950?C, 8?3 kb, to 675?C, 7?5kb) anda terminal decompressive path (from 7?5 to 4?5 kb). Spinel,quartz, high-Mg cordierite, and sapphirine were stabilized duringthe prograde high-T/P metamorphism, followed by the developmentof orthopyroxene, sillimanite, and garnet during the IBC. Retrogradelow-Mg cordierite appeared as a consequence of decompressionin the sapphirine granulite. Deformational structures, reportedfrom the Eastern Ghat granulites, and the available geochronologicaldata indicate that prograde metamorphism could have occurredat 30001?00 and 2500?100 Ma during a compressive orogeny thatwas associated with high heat influx through mafic magmatism. IBC ensued from P_max and was thus a direct consequence of progrademetamorphism. However, in the absence of sufficient study onthe spatial variation in P-T paths and the strain historiesin relation to time, the linkage between IBC and isothermaldecompression (ITD) has remained obscure. A prolonged IBC followedby ITD could be the consequence of one extensional mechanismwhich had an insufficient acceleration at the early stage, orITD separately could be caused by an unrelated extensional tectonism.The complex cooled nearly isobarically from 2500 Ma. It sufferedrapid decompression accompanied by anorthosite and alkalinemagmatism at 1400–1000 Ma.     

Reaction Textures in a Suite of Spinel Granulites from the Eastern Ghats Belt, India: Evidence for Polymetamorphism, a Partial Petrogenetic Grid in the System KFMASH and the Roles of ZnO and Fe2O3

Spinel granulites, with or without sapphirine, occur as lensesin garnetiferous quartzofeldspathic gneisses (leptynites) nearGokavaram in the Eastern Ghats Belt, India. Spinel granulitesare mineralogically heterogeneous and six mineral associationsoccur in closely spaced domains. These are (I) spinel–quartz–cordierite,(II) spinel–quartz–cordierite–garnet–orthopyroxene–sillimanite,(III) spinel–cordierite–orthopyroxene–sillimanite,(IV) spinel–quartz–sapphirine–sillimanite–garnet,(V) spinel–quartz-sapphirine–garnet and (IV) rhombohedral(Fe–Ti) oxide–cordierite–orthopyroxene–sillimanite.Common to all the associations are a porphyroblastic garnet(containing an internal schistosify defined by biotite, sillimaniteand quartz), perthite and plagioclase. Spinel contains variableamounts of exsolved magnetite and is distinctly Zn rich in thesapphirine-absent associations. X_Mg in the coexisting phasesdecreases in the order cordierite–biotite–sapphirine–orthopyroxene–spinel–garnet–(Fe–Ti)oxides. Textural criteria and compositional characteristicsof the phases document several retrograde mineral reactionswhich occurred subsequent to prograde dehydration melting reactionsinvolving biotite, sillimanite, quartz, plagioclase and spinel.The following retrograde mineral reactions are deduced: (1)spinel + quartz cordierite, (2) spinel + quartz garnet + sillimanite,(3) garnet + quartz cordierite + orthopyroxene, (4) garnet+ quartz + sillimanite cordierite, (5) spinel + cordierite orthopyroxene + sillimanite, (6) spinel + sillimanite + quartz sapphirine, (7) spinel + sapphirine + quartz garnet + sillimanite,and (8) spinel + quartz sapphirine + garnet. A partial petrogeneticgrid for the system FeO–MgO–Al₂O₃–SiO₂–K₂O–H₂Oat high fo₂, has been constructed and the effects of ZnO andFe₂O₃ on this grid have been explored Combining available experimentaland natural occurrence data, the high fo₂ invariant points inthe partial grid have been located in P–T space. Geothermobarometricdata and consideration of the deduced mineral reactions in thepetrogenetic grid show that the spinel granulites evolved throughan anticlockwise P–T trajectory reaching peak metamorphicconditions >9 kbar and 950C, followed by near-isobaric cooling(dT/dP = 150C/kbar). This was superimposed by an event of near-isothermaldecompression (dT/dP = 15C/kbar). The studied spinel granulites,therefore, preserve relic prograde mineral associations andreaction textures despite being metamorphosed at very high temperatures,and bear evidence of polymetamorphism. KEY WORDS: spinel granulite; Eastern Ghats; India; polymetamorphism; geothermometry; geobarometry Corresponding author     

Structure and Petrology of the Alpine-type Peridotite at Burro Mountain, California, U.S.A.

The alpine-type peridotite at Burro Mountain is a partiallyserpentinized harzburgite-dunite body approximately 2 km indiameter. It lies in a chaotic mélange derived from theFranciscan Formation (Upper Jurassic to Upper Cretaceous) ofthe southern Coast Ranges of California. The peridotite is boundedon the east by a vertical fault in the Nacimiento fault zonethat brings sedimentary rocks of Taliaferro's (1943b) AsuncionGroup (Upper Cretaceous) into contact with the peridotite. Theperidotite appears to be one of a number of tectonic lenses,having a wide range in size, that make up the mélange.These lenses include metagraywacke, metachert, greenstone, amphibolite,and blueschist, as well as ultramafic rocks, and represent awide range of pressure-temperature environments. The outer shell of the peridotite is a sheared serpentinitezone 10–15 m thick. The peridotite was tectonically emplacedat its present level as a cold solid mass and had little effecton the mineral assemblages of the Franciscan Formation. Localdevelopment of lawsonite and aragonite in shear zones may berelated to the peridotite emplacement. Foliated harzburgite forms approximately 60 per cent of theperidotite. It is a lithologically uniform rock that has anolivine: orthopyroxene ratio of approximately 75:25. Accessoryclinopyroxene and chromian spinel generally make up less than5 per cent of the harzburgite. Dunite, composed of olivine,accessory chromian spinel (< 5 per cent), and trace amountsof pyroxene, makes up approximately 40 per cent of the peridotiteand occurs as dikes, sills, and irregular bodies in the harzburgite. Olivine and pyroxene show small but significant compositionalvariations and chromian spinel shows a large range in the cationratio Cr/(Cr+Al+ Fe³⁺). The compositional variations in theseminerals are related to original differences in bulk chemicalcomposition. The following compositional ranges were determinedfor minerals in the harzburgite: olivine, Fo_91.1–Fo_91.4;orthopyroxene, En_89.8–En_91.1; clinopyroxene, Ca_47.0Mg_50.0Fe_3.0–Ca_48.7Mg_48.2Fe_3.1;chromian spinel, Cr/(Cr+Al+Fe³⁺) 0.37–0.55. The pyroxeneshave a range in A1₂O₃ content of 1.3–3.0 wt per cent.Olivine from dunite ranges from Fo₉₁ to Fo_{92 7} and the chromianspinel has a range in the Cr/(Cr+Al+Fe³⁺) ratio of 0.30–0.75.Although all the dunites are lithologically similar, three distincttypes are recognized on the basis of composition of coexistingolivine and chromian spinel. Structural relations between thethree types of dunite suggest three periods of emplacement (possiblyoverlapping) of dunite into harzburgite. The evidence indicatesthat the dunite, and probably also the harzburgite crystallizedfrom an ultramafic magma, probably in the upper mantle. After the magmatic episode and crystallization, the peridotitewas subjected to a deep-seated plastic deformation and recrystallization.The first phase of the deformation produced a pervasive, planarstructural element (S₁) that crosscuts many harzburgite-dunitecontacts. It is probable that some of the dunite sills wereemplaced during this deformation. The foliation, S₁, is definedby layers of different orthopyroxene content in harzburgite,and by discontinuous layers of chromian spinel in dunite. Flowor slip along S₁ produced slip folds in harzburgite—dunitecontacts with axial planes parallel to S₁. At a later stage,isoclinal folds developed in S₁, and the present olivine microfabricwas probably formed by recrystallization in the stress fieldthat produced the isoclinal folding. In the olivine microfabric,X tends to be perpendicular to the axial planes (S₂) of theisoclinal folds and Y and Z tend to form double maxima in S₂approximately 90° apart. Mg–Fe²⁺ distribution betweencoexisting mineral pairs yields a calculated temperature offormation of approximately 1200 °C. Although this temperatureis only a nominal value, it indicates that the mineral pairsequilibrated at a significantly high temperature. In view ofthe deformation and recrystallization, the calculated temperaturepossibly represents subsolidus re-equilibration of the mineralsduring this event. The deformation and recrystallization probablyoccurred shortly after crystallization while the peridotitewas still at a high temperature. A later deep-seated deformation produced small scattered kinkfolds in S₁ that tend to disrupt the major olivine microfabric.The kink folding was accompanied or followed by the developmentof kink bands in olivine that reflect intragranular glidingon the system T = [Okl], t = [100]. The kink bands probablyformed at a minimum temperature of 1000 °C. Following the deep-seated deformation, which probably took placein the mantle, the peridotite mass was tectonically detachedand moved upward to its present level in the crust. Cleavages,joints, and faults provided channels for water to pervade theperidotite and allow alteration of the primary minerals.     

Petrology of the Blue Mountain Complex, Marlborough, New Zealand

Blue Mountain is a central-type alkali ultrabasic-gabbro ringcomplex (1?1?5 km) introducing Upper Jurassic sediments, Marlborough,New Zealand. The ultrabasic-gabbroic rocks contain lenses ofkaersutite pegmatite and sodic syenite pegmatite and are intrudedby ring dykes of titanaugite-ilmenite gabbro and lamprophyre.The margin of the intrusion is defined by a ring dyke of alkaligabbro. The plutonic rocks are cut by a swarm of hornblende-biotite-richlamprophyre dykes. Thermal metamorphism has converted the sedimentsto a hornfels ranging in grade from the albite-epidote hornfelsfacies to the upper limit of the hornblende hornfels facies. The rocks are nepheline normative and consist of olivine (Fo_82-74),endiopside (Ca₄₅Mg₄₈Fe₇-Ca₃₆Mg₅₅Fe₉), titanaugite (Ca₄₀Mg₅₀Fe₁₀-Ca₄₄Mg₃₉Fe₁₇),plagioclase (An_73-18), and ilmenitetitaniferous magnetite, withvarious amounts of titaniferous hornblende and titanbiotite.There is a complete gradation between end-iopside and titanaugitewith the coupled substitution R_y^+z+Si(Ti⁺⁴+Fe⁺³)+Al⁺³ and asympathetic increase in CaAl₂SiO₆ (0?2-10?2 percent) and CaTiAl₂O₆(2?1-8?1 per cent) with fractionation. Endiopside shows a small,progressive Mg enrichment along a trend subparallel to the CaMgSi₂O₆-Mg₂Si₂O₆boundary, and titanaugite is enriched in Ca and Fe⁺²+Fe⁺³ withdifferentiation. Oscillatory zoning between endiopside and titanaugiteis common. Exsolved ilmenite needles occur in the most Fe-richtitanaugites. The amphiboles show the trend: titaniferous hornblende(1?0–5?7 per cent TiO₂)kaersutite (6?4 per cent TiO₂)Fe-richhastingsite (18?0–19?1 per cent FeO as total Fe). Biotiteis high in TiO₂ (6?6–7?8 per cent). Ilmenite and titaniferousmagnetite (3?5–10?6 per cent TiO₂) are typically homogeneousgrains; their composition can be expressed in terms of R⁺²RO₃:R⁺²O:R₂⁺³O₄. The intrusion of igneous rocks was probably controlled by subterraneanring fracturing. Subsidence of the country rock within the ringfracture provided space for periodic injections of magma froma lower reservoir up the initial ring fracture to form the BlueMountain rocks at a higher level. Downward movement of the floorof the intrusion during crystallization caused inward slumpingof the cumulates which affected the textural, mineralogical,and chemical evolution of the rocks in different parts of theintrusion. The order of mineral fractionation is reflected by the chemicalvariation in the in situ ultrabasic-gabbroic rocks and the successiveintrusions of titanaugite-ilmenite gabbro and lamprophyre ringdykes, marginal alkali gabbro and lamprophyre dyke swarm. Aninitial decrease, then increase in SiO₂; a steady decrease inMgO, CaO, Ni, and Cr: an initial increase, then decrease inFeO+Fe₂O₃, TiO₂, MnO, and V; almost linear increase in Al₂O₃and late stage increase in alkalis and P₂O₃, implies fractionationof olivine and endiopside, followed by titanaugite and Fe-Tioxides, followed by plagioclase, hornblende, biotite, and apatite.Reversals in the composition of cumulus olivine and endiopsideand Solidification Index, indicate that the ultrabasic-gabbroicsequence is composed of four main injections of magma. The ultrabasic rocks crystallized under conditions of high PH₂Oand fairly high, constant P_O2; P_H2 and P_O2 increased duringthe formation of the gabbroic rocks until fracturing of thechamber roof occurred. The abundance of euhedral amphibole inthe latter injection phases suggests that amphibole accumulatedfrom a hydrous SiO₂ undersaturated magma when an increase inP_O2, stabilized its crystallization. Plutonic complexes similar to Blue Mountain are found withinand beneath the volcanic piles of many oceanic islands, e.g.Canaries, Reunion, and Tahiti, and those intruding thick sedimentarysequences, as at Blue Mountain, e.g. the pipe-like intrusionsof the Monteregian Hills, Quebec.     

Melting of a Hydrous Mantle: I. Phase Relations of Natural Peridotite at High Pressures and Temperatures with Controlled Activities of Water, Carbon Dioxide, and Hydrogen

Four natural peridotite nodules ranging from chemically depletedto Fe-rich, alkaline and calcic (SiO₂ = 43.7–45.7 wt.per cent, A1₂O₃ = 1.6O–8.21 wt. per cent, CaO = 0.70–8.12wt. per cent, alk = 0.10–0.90 wt. per cent and Mg/(Mg+Fe²⁺)= 0.94–0.85) have been investigated in the hypersolidusregion from 800? to 1250?C with variable activities of H₂O,CO₂, and H₂. The vapor-saturated peridotite solidi are 50–200?Cbelow those previously published. The temperature of the beginningof melting of peridotite decreases markedly with decreasingMg/(Mg+SFe) of the starting material at constant CaO/Al₂O₃.Conversely, lowering CaO/Al₂O₃ reduces the temperature at constantMg/(Mg+Fe) of the starting material. Temperature differencesbetween the solidi up to 200?C are observed. All solidi displaya temperature minimum reflecting the appearance of garnet. Thisminimum shifts to lower pressure with decreasing Mg/(Mg + Fe)of the starting material. The temperature of the beginning ofmelting decreases isobarically as approximately a linear functionof the mol fraction of H₂O in the vapor (X_H2O^v). The data alsoshow that some CO₂ may dissolve in silicate melts formed bypartial melting of peridotite. Amphibole (pargasitic hornblende) is a hypersolidus mineralin all compositions, although its P/T stability field dependson bulk rock chemistry. The upper pressure stability of amphiboleis marked by the appearance of garnet. The vapor-saturated (H₂O) liquidus curve for one peridotiteis between 1250? and 1300?C between 10 and 30 kb. Olivine, spinel,and orthopyroxene are either liquidus phases or co-exist immediatelybelow the temperature of the peridotite liquidus. The data suggest considerable mineralogical heterogeneity inthe oceanic upper mantle because the oceanic geotherm passesthrough the P/T band covering the appearance of garnet in variousperidotites. The variable depth to the low-velocity zone is explained byvariable aHjo conditions in the upper mantle and possibly alsoby variations in the composition of the peridotite itself. Itis suggested that komatiite in Precambrian terrane could formby direct melting of hydrous peridotite. Such melting requiresabout 1250?C compared with 1600?C which is required for drymelting. The genesis of kimberlite can be related to partial meltingof peridotite under conditions of X_H2O^v = 0.5–0.25 (X_CO2^v= 0.5–0.75). Such activities of H₂O result in meltingat depths ranging between 125 and 175 km in the mantle. Thisrange is within the minimum depth generally accepted for theformation of kimberlite.     

Sillimanite and Ilmenite from High-grade Metamorphic Rocks of Antarctica and Other Areas

Sillimanite from a variety of high-grade metamorphic rocks containsfrom 0.13 to 1.82 weight per cent Fe₂O₃ and less than 0.1 weightper cent TiO₂. The iron is trivalent and substitutes for Alonly. Ilmenite associated with the sillimanite contains no morethan 0.4 weight per cent Al₂O₃, SiO₂, CaO, and MnO; and MgOdoes not exceed 1.6 weight per cent. It ranges in compositionfrom Ilm₉₉Hem₁ to Ilm₈₅Hem₁₅. A least squares fit of precision unit cell data on 10 analyzedsillimanites gives the following cell dimensions for iron-freesillimanite: a = 7.4830 Á, b = 7.6708 Á, c = 5.7694Á and V = 331.15 Á³. The projected increase incell volume with substitution of 10 mole per cent Fe₂SiO₃ is1.66 per cent. A regular increase in the Fe₂O₃ content of sillimanite withincreasing Fe₂O₃ content of associated ilmenite in 15 of 21samples analyzed suggests that sillimanite and ilmenite crystallizedin equilibrium in the 15 samples. The compositions of the tensillimanite-ilmenite pairs analyzed by the author fit the followingempirical curve (sol;(X_Fe2O3)Il = 1.110 x 10^–3. This regularincrease in Fe₂O₃ contents fits a model of Fe³⁺ substitutionfor Al on two independent sites in sillimanite and a coupledsubstitution of for Fe²⁺ Ti on two sites in ilmenite. Sillimaniteand ilmenite are behaving as ideal solutions over the compositionalrange 0 < X_Fe2SIO3 < 0.013 in sillimanite and 0 < X_Fe2O3< 0.15 in ilmenite. Equations have been derived for expressing the variation inFe₂O₃ content of sillimanite associated with quartz and ilmeniteor hematite as a function of pressure, temperature, and Fe₂O₃content of the oxide minerals. For example, the Fe₂O₃ contentof a sillimanite with 1.5 mole per cent Fe₂SiO₃ coexisting withTi-free hematite is calculated to decrease 11 per cent witha 5 kb increase in pressure. The rate of increase with temperatureof the Fe₂O₃ content of sillimanite is greater in hematite-bearingassemblages than in ilmenite-bearing assemblages.     

Petrology of Franciscan Metabasites Along the Jadeite-Glaucophane Type Facies Series, Cazadero, California

The Cazadero blueschist allochthon lies within the Central MelangeBelt of the Franciscan assemblage in the northern Coast Rangeof California. Mineral compositions and assemblages of morethan 200 blueschists from Ward Creek were investigated. Theresults delineate lawsonite-, pumpellyite-, and epidote-zones.The lawsonite and pumpellyite zones are equivalent to the TypeII metabasites of Coleman & Lee (1963) and are characterizedby well-preserved igneous textures, relict augite, and pillowstructures, whereas epidote zone rocks are equivalent to theType III strongly deformed and schistose metabasites. Chlorite,phengite, aragonite, sphene, and minor quartz and albite areubiquitous. The lawsonite zone metabasites contain lawsonite ( < 3 wt.per cent Fe₂O₃), riebeckite-crossite, chlorite, and Ca-Na-pyroxene;some rocks have two distinct clinopyroxenes separated by a compositionalgap. The clinopyroxene of the lowest grade metabasites containsvery low X_jd. In pumpellyite zone metabasites, the most commonassemblages contain Pm + Cpx + Gl + Chl and some samples withhigher Al₂O₃ and/or Fe₂O₃ have Pm + Lw + Cpx + Chl, Actinolitejoins the above assemblage in the upper pumpellyite zone wherethe actinolite-glaucophane compositional gap is well defined.The epidote zone metabasites are characterized by the assemblagesEp + Cpx + two amphiboles + Chl, Lw + Pm + Act + Chl, and Ep+ Pm + two amphiboles + Chl depending on the Fe₂O₃ content ofthe rock. In the upper epidote zone, winchite appears, Fe-freelawsonite is stable, pumpellyite disappears and omphacite containsvery low Ac component. Therefore, the common assemblages areEp + winchite + Lw, and Lw + Omp + winchite. With further increasein metamorphic grade, epidote becomes Al-rich and lawsoniteis no longer stable. Hence Ep + winchite + omphacite ? garnetis characteristic. Mineral assemblages and paragenetic sequences delineate threediscontinuous reactions: (1) pumpellyite-in; (2) actinolite-in;and (3) epidote-in reactions. Using the temperatures estimatedby Taylor & Coleman (1968) and phase equilibria for Ca-Na-pyroxenes,the P–T positions of these reactions and the metamorphicgradient are located. All three metabasite zones occur withinthe aragonite stability field and are bounded by the maximumpressure curve of Ab = Jd + Qz and the maximum stabilities ofpumpellyite and lawsonite. The lawsonite zone appears to bestable at T below 200?C with a pressure range of 4–6?5kb; the pumpellyite zone between 200 and 290?C and the epidotezone above 290?C with pressure variation between 6?5 and 9 kb.The metamorphic field gradient appears to have a convex naturetowards higher pressure. A speculative model of underplatingseamounts is used to explain such feature.     

The stability of sapphirine-quartz and hypersthene-sillimanite-quartz assemblages: an experimental investigation in the system FeO−MgO−Al2O3−SiO2 under H2O and CO2 conditions

Phase relations and mineral chemistry involving the phases garnet (Gt), spinel (Sp), hypersthene (Hy), sapphirine (Sa), cordierite (Cd), sillimanite (Sil) and quartz (Qz) have been experimentally determined in the system FMAS (FeO−MgO−Al₂O₂−SiO₂) under low fO₂ and for various H₂O/CO₂ conditions. Several compositions were studied with 100 (Mg/Mg+Fe) ratio ranging from 64 to 87 with excess quartz and sillimanite. Our data do not show any differences in Gt−Cd stability and composition as a function of H₂O, CO₂ and H₂O−CO₂ (±CH₄) content, in good agreement with a previous experimental study at lower temperature (Aranovich and Podlesskii 1983). At 1,000° C and 11 kbar, under CO₂-saturated conditions, cordierite grew from a crystalline mix unseeded with cordierite. Thus, under water-absent conditions, cordierite will have a high-P stability field in the presence of CO₂. If water has a pressure stabilizing effect on cordierite, then our results would indicate that the effects of H₂O and CO₂ are of the same magnitude at high temperature. Our data support the theoretical P-T grid proposed by Hensen (1986) for high-T metapelites and are largely consistent with the high-temperature experimental data of Hensen and Green (1973). The univariant boundary Gt+Cd=Hy+Sil+Qz, which marks the disappearance of Hy−Sil−Qz assemblages, has a negative dP/dT slope above 1,000° C and a positive one below this temperature. Extrapolation of our data to iron-free systems shows that the high-P breakdown limit of Mg-cordierite has a negative slope in the range 1,025–1,300° C and probably positive below 1,000° C. This indicates a maximum of stability for Mg-cordierite at around 1,000° C and 13 kbar. Because of the curvature of the univariant reactions En+Sil=Py+Qz, Mg−Cd=En+Sil+Qz and Gt+Cd=Hy+Sil+Qz, the iron-free invariant point involving the phases Py, En, Cd, Sil and Qz probably does not exist. Sapphirine—Qz-bearing assemblages are stable only at temperatures above 1,050° C. At 1,075° C, the joint Gt−Sa is stable up to 11 kbar. At higher pressure, garnet, sapphirine and quartz react according to the reaction Gt+Sa+Qz=Hy+Sil. Reequilibrated sapphirines are more aluminous than the theoretical endmember Mg₂Al₄SiO₁₀ due to AlAl=MgSi substitutions [100(Al₂O₃/Al₂O₃+FeO+MgO) in experimental sapphirines ranges from 50.5 to 52.2]. Sapphirine in the assemblage Sa−Cd−Sil−Qz shows a decrease in Al content with decreasing temperature and pressure, such that the alumina isopleths for sapphirine have a slight negative dP/dT slope. A similar decrease in Al content of sapphirine with temperature is also observed in Sa−Sil−Qz assemblages.     

Reactions Leading to the Disappearance of Pumpellyite in Low-grade Metamorphic Rocks of the Sanbagawa Metamorphic Belt in Central Shikoku, Japan

The major mineral assemblages of the metabasites of the Omoiji-Nagasawaarea in central Shikoku are hematite+epidote+chlorite+actinolite,riebeckitic actinolite+epidote+chlorite, epidote+chlorite+actinolite,and pumpellyite+epidote+chlorite+actinolite. The constituentminerals are often heterogeneous and assemblages in the fieldof a thin section sometimes do not obey the phase rule, butif grains apparently in non-equilibrium with others are excludedand domains of chemical equilibrium are appropriately chosenthe assemblages approximately obey the phase rule. The stability of hematite, pumpellyite, and epidote associatedwith chlorite and actinolite can be dealt with in terms of aternary system with appropriate excess phases. By fixing theFe²⁺/(Fe²⁺ +Mg) ratio of chlorite, it is dealt with in termsof stability relations in the system Ca₂Al₃Si₃O₁₂(OH)–Ca₂AlFe₂Si₃O₁₂(OH)with excess chlorite, actinolite, quartz, and controlled P_H2O.The maximum and minimum Fe³⁺ contents of epidote in this modelsystem are determined by hematite+epidote+chlorite+actinoliteand pumpellyite+epidote+chlorite+actinolite assemblages. Themaximum Fe³⁺ of the three phase assemblage epidote+chlorite+actinoliteis insensitive to temperature, but the minimum Fe³⁺ contentof epidote is sensitive to temperature and can be used to definethe metamorphic grade by a continuous quantity related to temperature.The phase relations expected for the model system are in goodagreement with the parageneses of the Sanbagawa terrain in centralShikoku and offer an explanation to the rule of Miyashiro &Seki (1958a) that the compositional range of epidote enlargeswith increasing temperature. The model also makes it possibleto estimate semi-quantitatively the temperature range in whichthe assemblage pumpellyite+epidote+chlorite+actinolite is stable.The possible maximum range is about 120 ?C, but the assemblageis stable in metabasite only for about 90 ?C. The higher temperaturelimit of the pumpellyite-actinolite facies defined by the disappearanceof pumpellyite in metabasite corresponds to the temperatureat which epidote with Fe³⁺/(Fe³⁺ +Al) = 0.10 0.15 coexistswith pumpellyite, actinolite, and chlorite. The compositions of epidotes in the metabasites of the Omoiji-Nagasawaarea cluster around Fe³⁺/(Fe³⁺ +Al) = 0.33. The grade of thisarea is close to the lower temperature stability limit of thepumpellyite+epidote+chlorite+actinolite assemblage.     

The Range of Spinel Compositions in Terrestrial Mafic and Ultramafic Rocks

Compositional fields for spinels from a wide variety of mafic–ultramaficigneous rock types and tectonic environments have been determinedfrom a global database of over 26 000 analyses. These fieldsare defined using contoured data density plots based on thespinel prism, and plots of T iO₂ vs ferric iron, for mantlexenoliths, ophiolitic rocks, continental layered intrusions,alkalic and lamprophyric rocks, tholeiitic basalts, Alaskanultramafic complexes and komatiites. Several trends appear regularlyin the various environments: a trend of widely variable Cr/(Cr+ Al) at low Fe²⁺/(Mg + Fe²⁺) (the Cr–Al trend); increasingFe³⁺, Fe²⁺/(Mg + Fe²⁺) and T iO₂ at constant Cr/(Cr + Al) (Fe–Ti trend); a trend found primarily in kimberlites, similar toFe–T i but at constant Fe²⁺/(Mg + Fe²⁺); and an unusualtrend of increasing Al found only in layered intrusions. TheCr–Al and Fe–T i trends are both found to varyingdegrees in tholeiitic basalts. The Cr–Al trend is prevalentin rocks that have equilibrated over a range of pressures, whereasthe Fe–T i trend is dominantly due to low-pressure fractionation.The most Cr-rich chromites found in nature occur in boninites,diamond-bearing kimberlites, some komatiites and ophioliticchromitites. Exceptionally reduced chromites are found in somekomatiites and in ophiolitic chromitites. Detrital chromitesfrom the Witwatersrand conglomerates are of komatiitic provenance. KEY WORDS: basalt; chromite; kimberlite; ophiolite; spinel     

Liquidus Phase Relations in the System CaO-MgO-Al2O3-SiO2 at 2{middle dot}0 GPa: Applications to Basalt Fractionation, Eclogites, and Igneous Sapphirine

To model magmatic crystallization processes for mafic to intermediatecompositions at high pressure, liquidus phase relations in theforsterite–anorthite–diopside–silica (FADS)tetrahedron within the CaO–MgO–Al₂O₃–SiO₂system have been determined at 2·0 GPa. Compositionsof five liquidus invariant points have been determined and theapproximate compositions of five others have been inferred.These involve primary phase volumes for forsterite (fo), enstatite(en), diopside (di), high quartz (qz), spinel (sp), sapphirine(sa), garnet (gt), anorthite (an), and corundum (cor). The determined(with wt % coefficients) and inferred reactions (without coefficients)that define each isobaric invariant point are as follows: 23 en + 68 di + 9 sp = 84 liq + 16 fo 37 di + 63 sa = 47 liq + 40 sp + 13 en 100 gt = 21 liq + 27 sa + 55 en + 18 di 1 di + 59 en + 41 an = 43 liq + 57 gt 18 di + 21 qz + 15 en + 47 an = 100 liq di + an + gt = liq + sa an + gt = liq + sa + en sa + an + di = liq + sp sa + an = liq + cor + sp di + cor = liq + an + sp. These phase relations provide a diverse range of constraintson igneous processes at pressures near 2 GPa. They show thatfractional crystallization of a model basalt gives a residualliquid strongly enriched in SiO₂, strongly depleted in MgO,and mildly enriched in Al₂O₃. Such a trend is consistent withthe calc-alkaline fractionation trend observed at subductionzones, but is in disagreement with suggestions that fractionationof tholeiitic basalt in this pressure range yields an alkalicbasalt. Both trends may occur for natural basalts dependingon the Na₂O content of the parental magma. Also, the data showthat the minimum pressure for the formation of cumulate eclogitesand garnet pyroxenites is about 1·8–1·9GPa. The lower limit of pressure at which sapphirine can crystallizefrom a liquid in the FADS tetrahedron is estimated to be 1·1–1·5GPa and the upper limit is >3 GPa. Sapphirine crystallizesfrom magmas intermediate in composition between basalt and andesite.Probable igneous sapphirine in mafic associations is rare, butit occurs as part of a pyroxenite xenolith from Delegate, Australia,that we suggest is a cumulate assemblage and in a sapphirinenorite at Wilson Lake, Labrador, Canada. KEY WORDS: basalt; eclogite; sapphirine; fractional crystallization