Experimental and major element constraints on the evolution of lavas from Lihir Island,Papua New Guinea

The major element chemistry of SiO₂-undersaturated arc lavas from Lihir Island, Papua New Guinea, and 1 atmosphere experiments on an alkali basalt from this island show complex polybaric fractionation affected this suite of lavas. Low Ni and MgO are typical of these arc lavas and result from olivine fractionation, probably at high pressure. Fractionation at low pressure (<5 kb) produces two evolutionary trends. Separation of clinopyroxene, plagioclase and minor olivine from the primitive lavas results in increasing normative nepheline contents and major element trends similar to those of the experiments. In contrast, addition of magnetite and amphibole to the fractionating assemblage in the evolved lavas results in decreasing normative nepheline and major element trends which are markedly different from those of the experiments. The composition of experimental glasses and 1 atmosphere liquid lines of descent, derived from anhydrous melting experiments run at the fayalite-magnetite-quartz (FMQ) buffer and at higher oxygen fugacities, are displaced from the lavas on oxide-oxide plots. HighfO₂ produces high Fe³⁺/Fe²⁺ and the early crystallization of abundant magnetite, and high H₂O contents are responsible for crystallization of amphibole. Crystal fractionation of these phases and the high Fe³⁺/Fe²⁺ are responsible for the displacement of the lavas and experimental glasses in mineral projection schemes from the 1 atmosphere olivine-clinopyroxene-plagioclase saturation boundary of Sack et al. (1987).     

Upper mantle source for some hawaiites,mugearites and benmoreites

Chemical analyses of over seventy lavas or dykes containing spinel lherzolite inclusions of high pressure mineralogy, show that most host magmas are of alkali olivine basalt or basanite composition with relatively rare olivine nephelinites, and olivine melilitites. The 100 Mg/Mg+Fe⁺⁺ ratios of host magmas display a strong maximum at about Mg₇₀ consistent with partial melting of source peridotite with olivine of Fo_88–90. In contrast to these primary magmas, there occur some host magmas with 100 Mg/Mg+Fe⁺⁺<60 and with chemical compositions resembling those of classical hawaiite, mugearite, and nepheline benmoreite magmas. It is inferred that these magmas have been produced by crystal fractionation, within the upper mantle, of parental basanites or alkali olivine basalts. The presence of kaersutitic hornblende xenocrysts accompanying the lherzolite inclusions, and the nature of the chemical variation between associated basanites and nepheline benmoreites suggests that crystal fractionation has been dominated by kaersutitic hornblende, together with olivine and, in some cases, probably clinopyroxene. The mantle-derived nepheline benmoreite magmas also show similarities to some plutonic nepheline syenites.     

Petrogenetic relationships between melilitite and lamproite

A detailed petrological study of the S. Venanzo olivine melilitite and Cupaello kalsilitite, located at the NE margin of the Roman Comagmatic Region of Italy has provided new data to evaluate their genetic relationships and related tectonic/magmatic conditions. Early crystallized olivine (Fo₉₂) from S. Venanzo is compatible with crystallization from near primary mantle melts, while late-crystallized olivine from both rocks (Fo_88–79) reflects primarily the high Ca of the host liquids, attained under the volcanic crystallization regime. Magnesiochromite inclusions in the early-crystallized olivine are consistent with near-primary melts close to lamproite in composition. Nepheline and kalsilite from both lavas contain high alkalis (+Ca), relative to Al, and thus correspond more closely with nepheline compositions from carbonatite-related assemblages, than with those from a wide compositional range of Alban Hills lavas. Coexisting melilite has high (Na+K)/Al, reflecting the Peralkalinity Index of the bulk rocks. Diopside and phlogopite from both lavas are characteristic of lamproites and groundmass kimberlites in their high Mg/(Mg+Fe²⁺) ratio (0.86–0.95; 0.80–0.90, respectively) and T-site (Si+Al) deficiencies. Götzenite, Ca₂Na[Zr, Ti]Si₂O₇(O, OH, F)₂, identified in both lavas, is typical of nephelinite-ijolite assemblages. On the other hand, khibinskite, K₄Zr₂Si₄O₁₄, found in the Cupaello lava, may be regarded as a Si-poor variety of wadeite, a mineral characteristic of lamproites. Clinopyroxene and monticellite, coexisting as late-crystallized phases in both lavas, suggest a common P-T liquid path of thermal descent in the system CaMgSi₂O₇–CO₂ in the presence of excess CO₂, but with different intersections with the åkermanite stability field. Substantial differences in SiO₂ saturation combined with high Mg number and liquidus temperatures experimentally determined at atmospheric pressure in both lavas (1276° C and 1260° C, respectively) indicate that a parent-daughter relationship is unlikely under the volcanic P-T regime. In La Roche's Rm-Ri-Rs diagram, the S. Venanzo composition fall close to the Oldoinyo Lengai alkalic pyroxenite trend of Donaldson and Dawson, while the Cupaello compositions follow a lamproitic trend, consistent with that indicated by the quartz-normative glass of the Gaussberg lava. Mantle compositions corresponding to wehrlite-clinopyroxenite and enriched in H₂O, CO₂, F, and LILE, are favoured as potential sources for the lavas. Their origin is probably related to subcrustal fluid transfer promoted by the Tyrrhenian mantle doming.     

Phase Relationships of Hydrous Alkalic Magmas at High Pressures: Production of Nepheline Hawaiitic to Mugearitic Liquids by Amphibole-Dominated Fractional Crystallization Within the Lithospheric Mantle

Experimental melting studies were conducted on a nepheline mugearitecomposition to pressures of 31 kbar in the presence of 0–30%added water. A temperature maximum in the near-liquidus stabilityof amphibole (with olivine) was found for a water content of3·5 wt % at a pressure of 14 kbar. This is interpretedto have petrogenetic significance for the derivation of nephelinemugearite magmas from nepheline hawaiite by amphibole-dominatedfractional crystallization at depth within the lithosphericmantle. Synthetic liquids at progressively lower temperaturesrange to nepheline benmoreite compositions very similar to thoseof natural xenolith-bearing high-pressure lavas elsewhere, andsupport the hypothesis that continued fractional crystallizationcould lead to high-pressure phonolite liquids. Independent experimentaldata for a basanite composition modeled on a lava from the sameigneous province (the Newer Basalts of Victoria) permit theinference that primary asthenospheric basanite magmas undergopolybaric fractional crystallization during ascent, and mayevolve to liquids ranging from nepheline hawaiite to phonoliteupon encountering cooler lithospheric mantle at depths of 42–50km. Such a model is consistent with the presence in some evolvedalkalic lavas of both lithospheric peridotite xenoliths indicativeof similar depths and of megacryst suites that probably representdisrupted pegmatitic segregations precipitated from precursoralkalic magmas in conduit systems within lithospheric mantle. KEY WORDS: experiment; high pressure; alkalic magmas; amphibole; nepheline mugearite; basanite; lithosphere     

Commingling and mixing of S-type peraluminous, ultrapotassic and basaltic magmas in the Cayconi volcanic field, Cordillera de Carabaya, SE Peru

The Cayconi district of the Cordillera de Carabaya, SE Peru, exposes a remnant of an upper Oligocene–Lower Miocene (22.2–24.4 Ma) volcanic field, comprising a diverse assemblage of S-type silicic and calc-alkaline basaltic to andesitic flows, members of the Picotani Group of the Central Andean Inner Arc. Basaltic flows containing olivine, plagioclase, clinopyroxene, ilmenite and glass, and glassy rhyolitic agglutinates with phenocrystic quartz, cordierite, plagioclase, sanidine, ilmenite and apatite, respectively exhibit mineralogical and geochemical features characteristic of medium-K mafic and Lachlan S-type silicic lavas. Cordierite-bearing dacitic agglomerates and lavas, however, are characterized by dispersed, melanocratic micro-enclaves and phenocrysts set in a fine-grained quartzo-feldspathic matrix. They contain a bimodal mica population, comprising phlogopite and biotite, as well as complexly zoned, sieve-textured plagioclase grains, sector-zoned cordierite, sanidine, quartz, irregular patches of replaced olivine, clinopyroxene and orthopyroxene and accessory phases including zircon, monazite, ilmenite and chromite. The coexistence of minerals not in mutual equilibrium and the growth/dissolution textures exhibited by plagioclase are features indicative of magmatic commingling and mixing. Trachytic-textured andesite flows interlayered with olivine+plagioclase–glomerophyric, calc-alkaline basalts have a phenocrystic assemblage of resorbed orthopyroxene and plagioclase and exhibit melanocratic groundmass patches of microphenocrystic phlogopite, Ca-rich sanidine, ilmenite and aluminous spinel. The mineralogical and mineral chemical relationships in both the dacites and the trachytic-textured andesites imply subvolcanic mixing between distinct ultrapotassic mafic melts, not represented by exposed rock types, and both the S-type silicic and calc-alkaline mafic magmas. Such mixing relationships are commonly observed in the Oligo-Miocene rocks of the Cordillera de Carabaya, suggesting that the S-type rocks in this area and, by extension, elsewhere derive their unusually high K₂O, Ba, Sr, Cr and Ni concentrations from commingling and mixing with diverse, mantle-derived potassic mafic magmas.     

A clinoenstatite-bearing cumulate Olivine Pyroxenite from Howqua,Victoria

The Howqua Olivine Pyroxenite of eastern Victoria, Australia, intrudes a metamorphosed sequence of Cambrian high-Mg lavas. It crystallized an unusual mineral assemblage: Cr-rich magnesiochromite, olivine (Fo₉₄), and protoenstatite (now inverted to polysynthetically-twinned clinoenstatite). Residual liquid crystallized strongly-zoned interstitial pyroxenes followed by pargasite. Pargasite, often showing quench habit, crystallized in interstitial glass which is now altered to serpentine.The extremely refractory nature of the cumulus phases indicates a very high temperature of crystallization for liquidus olivine and chromite from a high-MgO, low-Al₂O₃ parent liquid similar in some respects to Archaean peridotitic komatiites. The suggested origin by hydrous melting of depleted mantle peridotite, plus other compositional and mineralogical similarities (especially the olivine-liquid reaction producing protoenstatite) indicate that the parent magma of the Howqua Olivine Pyroxenite had many features in common with the high-SiO₂, high-MgO clinoenstatite-bearing boninitic lavas of the Western Pacific. It is interpreted as a more extreme melt with affinities to boninite and it demonstrates that ultramafic magmas existed in the Cambrian.     

The evolution of young silicic lavas at Medicine Lake Volcano,California: Implications for the origin of compositional gaps in calc-alkaline series lavas

At Medicine Lake Volcano, California, the compositional gap between andesite (57–62 wt.% SiO₂) and rhyolite (73–74 wt.% SiO₂) has been generated by fractional crystallization. Assimilation of silicic crust has also occurred along with fractionation. Two varieties of inclusions found in Holocene rhyolite flows, hornblende gabbros and aphyric andesites, provide information on the crystallization path followed by lavas parental to the rhyolite. The hornblende gabbros are magmatic cumulate residues and their mineral assemblages are preserved evidence of the phases that crystallized from an andesitic precursor lava to generate the rhyolite lavas. The andesitic inclusions represent samples of a parental andesite and record the early part of the differentiation history. Olivine, plagioclase and augite crystallization begins the differentiation history, followed by the disappearance of olivine and augite through reaction with the liquid to form orthopyroxene and amphibole. Further crystallization of the assemblage plagioclase, amphibole, orthopyroxene, magnetite, and apatite from a high-SiO₂ andesite leads to rhyolite. This final crystallization process occurs on a cotectic that is nearly horizontal in temperature-composition space. Since a large amount of crystallization occurs over a limited temperature interval, a compositional gap develops between rhyolite and high SiO₂ andesite.Liquidus surfaces with shallow slopes in temperature-composition space are characteristic of several late-stage crystallization assemblages in the andesite to rhyolite compositional range. Experimentally produced plagioclase+ amphibole+orthopyroxene+magnetite and plagioclase+ augite+low-Ca pyroxene+magnetite cotectics have liquidus slopes that are nearly flat. At other calc-alkaline volcanic centers crystallization processes involving large compositional changes over small temperature intervals may also be important in the development of bimodal volcanism (i.e. the existence of a composition gap). At Mt. Mazama and Mt. St. Helens, USA and Aso Caldera and Shikotsu, Japan the amphibole-bearing assemblage was important. At Krakatau, Indonesia and Katmai, USA, an augite+orthopyroxene-bearing assemblage was important. In addition to its role in the production of a compositional gap between intermediate and rhyolitic lavas, the crystallization process increases the H₂O content of the residual liquid. This rapid increase in residual liquid volatile content which results from the precipitation of a large proportion of crystalline solids may be an important factor among several that lead to explosive silicic eruptions.     

Kalsilite in the lavas of Mt. Nyiragongo (Belgian Congo)

A considerable part of the nephelinite lavas of the volcanoMt. Nyiragongo in the eastern Belgian Congo contains kalsiliteas one of the main constituents. The mineral never occurs asthe only feldspathoid of the rock but is accompanied by nepheline,abundant melilite, and, sometimes, by small to moderate amountsof leucite. Other important constituents of these kalsilite-bearingrocks are clinopyroxene, olivine, perovskite, titanomagnetite,sodalite, &c. The feldspars are lacking. Kalsilite occurs both as complex nepheline-kalsilite phenocrystsin which these phases are strictly co-axial and in the fine-grainedgroundmass as grains separate from those of nepheline. The complex nepheline-kalsilite phenocrysts exhibit a continuousseries of progressing exsolution schematically presented inFig. 5. The series begins with a perthite-like nepheline-kalsilitecore surrounded by a drop-like development of nepheline in themargin of the crystal and ends up with a homogeneous kalsilitecore surrounded by a nepheline margin. The complex phenocrysts occur mostly as aggregates causing atypically glomeroporphyritic texture. Evidence is presentedindicating that, in the very first stages of crystallization,some of the Nyiragongo lavas are able to precipitate small amountsof phenocrysts of approximate composition K₃NaAl₄Si₄O₁₆. Throughcrystal-rise under turbulent currents in the molten lava massthese phenocrysts have been accumulated into aggregates andthus have been preserved until extrusion. Granted sufficientlyslow cooling under static conditions, the phenocrysts wouldhave reacted with the molten lava. The roles of the crystal-riseand of the turbulent currents in lava are illustrated by theoccurrence of the ‘giant’ leucite aggregates foundin the inner walls of the crater and by observations on thelava lake of the mountain. The occurrence of kalsilite in the groundmass is explained bythe existence of a two-phase area in the sub-solidus range inthe nepheline-kalsilite system. The Nepheline Aggregate lavas represent the last extrusionsemitted by the Nyiragongo main crater. The nepheline phenocrystscharacteristic of these lavas range considerably higher in potassiumcontent than the nephelines found in other Nyiragongo flows.The crystals are slightly zoned with a large potassium-richcore coated by a narrow margin with gradually decreasing potassiumcontent. The zoning may be detected only by using special methods.The history of crystallization of the nepheline phenocrystsis considered analogous to that of the complex nepheline-kalsilitephenocrysts with the only difference that the nepheline phenocrystsof the Nepheline Aggregate lavas are less rich in potassiumand, consequently, have not been subjected to exsolution.     

Basalt evolution at low pressure: implications from an experimental study in the system CaO-FeO-MgO-Al2O3-SiO2

New experiments have been performed in the system CaO+MgO+Al₂O₃+SiO₂ (CMAS)+FeO at atmospheric pressure. Most of the experiments were conducted on Fe-rich compositions, in the low-temperature field of the assemblage liq(liquid)+an(anorthite) +aug(augite)+ol(olivine), and mostly along five isotherms. Others were located on, or nearby the assemblage boundaries. These experiments, together with the previously reported high temperature experiments (Shi and Libourel 1991; Libourel et al. 1989), permit contouring the complete liq+an+aug+ol divariant field, and tracing out some of its boundaries. The boundary of the assemplage liq+an+aug+ol consists of six segments, with the appearance of one of the following phases, orthopyroxene, pigeonite, tridymite, bustamite, kirschsteinite, and spinel, as an additional phase. Within the stability field of the assemblage liq+an+aug+ol, the compositions of all the coexisting phases have been described as functions of temperature and silica content in the melt by applying a multiple linear regression method. This allows a quantitative characterization of the divariant assemblage liq+an+aug+ol in the system CMAS+FeO, from 1273°C to 1055°C, with olivine compositions ranging from Mg^*[Mg/(Mg+Fe)]=1 to 0.08. Knowing the composition-temperature relationships, the basic T-X configuration of the assemblage liq+an +aug+ol has been analysed, and mass-balance calculations have been performed to examine the FeO effect on different crystallization processes. Addition of FeO to the system CMAS transforms the thermal divide in the assemblage liq+an+di(diopside)+fo(forsterite) into a thermal ridge. With decreasing temperature, the spine of the thermal ridge moves towards Si-poor compositions at Mg-rich end but towards Si-rich compositions at the Fe-rich end. This indicates that late-stage tholeiitic liquids can follow a trend of silica enrichment without the crystallization of an oxide phase. Crystallization paths of the assemblage liq+an+aug+ol are determined by the detailed T-X relations of the thermal ridge with the melt evolving away from the spine. The boundary reactions with decreasing temperature have also been characterized numerically.     

The geochemistry of the Dunedin Volcano,East Otago,New Zealand: Rare earth elements

A variety of alkaline lavas from the Dunedin Volcano have been analyzed for the rare earth elements (REE) La-Yb. The compositions analyzed were: basalt-hawaiite-mugearite-benmoreite; basanite, nepheline hawaiite, nepheline trachyandesite and nepheline benmoreite; trachyte; phonolite. The series from basalt to mugearite shows continuous enrichment in the REE, consistent with a crystal fractionation model involving removal of olivine and clinopyroxene. From mugearite to benmoreite there is a depletion in the REE which is explained by the appearance of apatite as a liquidus phase. The chondrite normalized REE patterns for the phonolites are characterized by strong enrichment and fractionation coupled with a sharp depletion in Eu. Removal of plagioclase from benmoreite magma is suggested for the derivation of the phonolites. The series basanite-nepheline hawaiite, and basanite-nepheline hawaiite-nepheline benmoreite appear to be high pH₂O analogues of the series basalt-ben-moreite, with enrichment of the REE being achieved by removal of clinopyroxene, kaersutite and olivine. Compared with other lavas the trachyte has low REE abundances and is characterized by a striking positive Eu anomaly.     

Supercritical pE-pH diagrams,with applications to the stability of biotite

A pE-pH diagram for supercritical aqueous fluids at 500° C and 2 kb total pressure is calculated from thermodynamic data assuming ideal mixing of gas species. The experimental data on the stability of annite-phlogopite solid-solutions (Wones and Eugster 1965) and on the stability of aluminous biotite (Rutherford 1968) have been used to calculate the stability of these biotite solid solutions in sanidine, muscovite and Al₂SiO₅ bearing assemblages. The resulting pE-pH diagrams show that Al increases the stability field of annite more per atom than does Mg. However, the addition of Al to biotite does not increase the stability of annite in very acid and alkaline solutions. Environments sufficiently acid to render aluminous biotite metastable are probably not found in nature, but the mildly alkaline environments are attainable. At constant f _{O 2}, f _{H 2} and the Al content of biotite is at a minimum in muscovite + sanidine bearing assemblages, and increases relatively rapidly with increases in pH in the sanidine field, and increases more slowly with decreases in pH through the muscovite and Al₂SiO₅ fields.These diagrams show that the composition of biotite solid-solutions containing more aluminum than ideal annite-phlogopite and coexisting with sanidine and magnetite cannot be used to infer intensive parameters (T f _{O 2}, ) prevailing during the crystallization of the biotite, and that in no case can any biotite composition in muscovite or Al₂SiO₅ bearing assemblages be used to extract this information.     

Kalsilite in the lavas of Mt. Nyiragongo (Belgian Congo)

A considerable part of the nephelinite lavas of the volcanoMt. Nyiragongo in the eastern Belgian Congo contains kalsiliteas one of the main constituents. The mineral never occurs asthe only feldspathoid of the rock but is accompanied by nepheline,abundant melilite, and, sometimes, by small to moderate amountsof leucite. Other important constituents of these kalsilite-bearingrocks are clinopyroxene, olivine, perovskite, titanomagnetite,sodalite, &c. The feldspars are lacking. Kalsilite occurs both as complex nepheline-kalsilite phenocrystsin which these phases are strictly co-axial and in the fine-grainedgroundmass as grains separate from those of nephe-line. The complex nepheline-kalsilite phenocrysts exhibit a continuousseries of progressing exsolution schematically presented inFig. 5. The series begins with a perthite-like nepheline-kalsilitecore surrounded by a drop-like development of nepheline in themargin of the crystal and ends up with a homogeneous kalsilitecore surrounded by a nepheline margin. The complex phenocrysts occur mostly as aggregates causing atypically glomeroporphyritic texture. Evidence is presentedindicating that, in the very first stages of crystallization,some of the Nyiragongo lavas are able to precipitate small amountsof phenocrysts of approximate composition K₃NaAl₄Si₄O₁₆. Throughcrystal-rise under turbulent currents in the molten lava massthese phenocrysts have been accumulated into aggregates andthus have been preserved until extrusion. Granted sufficientlyslow cooling under static conditions, the phenocrysts wouldhave reacted with the molten lava. The roles of the crystal-riseand of the turbulent currents in lava are illustrated by theoccurrence of the ‘giant’ leucite aggregates foundin the inner walls of the crater and by observations on thelava lake of the mountain. The occurrence of kalsilite in the groundmass is explained bythe existence of a two-phase area in the sub-solidus range inthe nepheline-kalsilite system. The Nepheline Aggregate lavas represent the last extrusionsemitted by the Nyiragongo main crater. The nepheline phenocrystscharacteristic of these lavas range considerably higher in potassiumcontent than the nephelines found in other Nyiragongo flows.The crystals are slightly zoned with a large potassium-richcore coated by a narrow margin with gradually decreasing potassiumcontent. The zoning may be detected only by using special methods.The history of crystallization of the nepheline phenocrystsis considered analogous to that of the complex nepheline-kalsilitephenocrysts with the only difference that the nephe-line phenocrystsof the Nepheline Aggregate lavas are less rich in potassiumand, consequently, have not been subjected to exsolution.     

Shonkin Sag laccolith,Montana

A variety of rock types is developed in the Shonkin Sag laccolith, with extreme compositions represented by the porphyritic pseudoleucite-bearing chilled margin, shonkinitic in composition, and by the final differentiate, chemically a nepheline syenite. During differentiation the pyroxenes changed in composition from Ca-rich varieties through aegirine-augite to acmite; there is no evidence of an immiscibility gap between Ca-rich and Na-rich pyroxenes. Olivine compositions changed from approximately Fa₂₀ to Fa₄₀, but in marked contrast the coexisting biotites exhibit a more extensive compositional range, from annite₂₄ to annite₁₀₀; the crystal margins of annites in the most evolved rocks are manganophyllite-rich. Titanomagnetites are TiO₂-poor varieties. Arfvedsonite and melanite occur in the most evolved syenites.From mineralogical and thermodynamic data initial and final temperatures of crystallization of the various rock types have been calculated. At an estimated total vapor pressure of 310 bars, the temperature of intrusion was 985° C. Final crystallization of the laccolith took place below 700° C, and crystallization intervals for most rock types are of the order of 170° C. Over the total magmatic temperature range the activity of silica in the melt decreased from 0.13 to 0.09. Oxygen fugacity falls with temperature approximately parallel to the synthetic fayalite-magnetite-quartz oxygen buffer until olivine disappears. The crystallization of Na-rich pyroxenes does not demand an increase in the fugacity of oxygen, but rather requires that the oxygen fugacity fall less rapidly with temperature than would be the case if olivine and magnetite were present.     

The effects of FeO on the system CMAS at low pressure and implications for basalt crystallization processes

The common occurrence of olivine, clinopyroxene and plagioclase in natural basalts may suggest that crystallization of these phases exerts important controls on the evolution of natural basalts at low pressure. In order to understand the effects of such an assemblage on the evolution of basalt, an experimental study of the divariant assemblage 1+an+cpx+ol in the system CMAS + FeO was undertaken at 1 bar and under very reducing conditions. Experiments have been performed along three isothernal sections. Detailed electron microprobe data have been collected on all the coexisting phases. Combining these data with available data in the system CMAS, the compositions of all the coexisting phases have been described as functions of temperature and CaO content in the melt by applying a multiple linear regression method. This allows a quantitative characterization of the divariant assemblage 1+an+epx+ol in the system CMAS + FeO, from 1275°C to 1160°C, with liquid compositions ranging from mg#=1 to mg#=0.28. Knowing the composition-temperature relationships, the basic T-X configuration of the assemblage 1+an+cpx+ol has been analysed, and mass balance calculations have been performed to examine the FeO effect on different crystallization processes. Addition of FeO to the system CMAS transforms the thermal divide in the assemblage 1+an+cpx+ol to a thermal ridge, and shifts the spine of the thermal ridge towards SiO₂-poor compositions with decreasing temperature. This indicates that tholeiitic basalts can be relatively silica-poor, and evolve towards slightly more silica-poor compositions. With increasing FeO content, pigeonite replaces ortheonstatite as a crystallizing phase along the silica-rich boundary of the assemblage 1+an+cpx+ol. In some iron, and silica rich composition, olivine possibly terminates its crystallization or even starts resorption prior to pigeonite precipitation; this suppresses silica enrichment in the melt. Crystallization paths of the assemblage 1+an+cpx+ol are determined by the detailed T-X relations of the thermal ridge with the melt evolving away from the spine. The liquid evolution trend depends strongly on the initial compositions. A tholeiitic liquid can follow a moderate silica enrichment path at the same very reducing condition of there is a high modal proportion of olivine. ‘A Fenner mechanism therefore is at work at the same place as a Bowen mechanism’.     

Synchrotron radiation study on the high-pressure and high-temperature phase relations of KAlSi3O8

In situ X-ray diffraction study on KAlSi₃O₈ has been performed using the cubic type high pressure apparatus, MAX90, combined with synchrotron radiation. We determined the phase relations of sanidine, the wadeite-type K₂Si₄O₉+kyanite (Al₂SiO₅)+coesite (SiO₂) assemblage, and hollandite-type KAlSi₃O₈, including melting temperatures of potassic phases, up to 11 GPa. Our data on subsolidus phase boundaries are close to the recent data of Yagi and Akaogi (1991). Melting relations of sanidine are consistent with the low pressure data of Lindsley (1966). The breakdown of sanidine into three phases reduces melting temperature, and wadeite-type K₂Si₄O₉ melts first around 1500° C in three phase coexisting region. Melting point of hollandite-type KAlSi₃O₈ is between 1700° C and 1800° C at 11 GPa. If these potassic phases host potassium in the earth's mantle, the true mantle solidus temperature will be much lower than the reported dry solidus temperature of peridotite.     

Oxygen isotope composition of phenocrysts from Tristan da Cunha and Gough Island lavas: variation with fractional crystallization and evidence for assimilation

We have measured the δ¹⁸O values of the major phenocrysts (olivine, clinopyroxene and plagioclase) present in lavas from Tristan da Cunha and Gough Island. These islands, which result from the same mantle plume, have enriched radiogenic isotope ratios and are, therefore, prime candidates for an oxygen isotope signature that is distinct from that of MORB. Consistent differences between the δ¹⁸O values of olivine, pyroxene and feldspar in the Gough lavas show that the phenocrysts in the mafic Gough Island lavas are in oxygen isotope equilibrium. The olivines in lavas with SiO₂ <50 wt% have a mean δ¹⁸O value of 5.19‰, consistent with crystallization from a magma having the same oxygen isotope composition as MORB. Phenocrysts in all the Gough lavas show a systematic increase in δ¹⁸O value as silica content increases, which is consistent with closed-system fractional crystallization. The lack of enrichment in δ¹⁸O of the Gough magmas suggests that the mantle source contained <2% recycled sediment. In contrast, the Tristan lavas with SiO₂ >48 wt% contain phenocrysts which have δ¹⁸O values that are systematically ∼0.3‰ lower than their counterparts from Gough. We suggest that the parental mafic Tristan magmas were contaminated by material from the volcanic edifice that acquired low δ¹⁸O values by interaction with water at high temperatures. The highly porphyritic SiO₂-poor lavas show a negative correlation between olivine δ¹⁸O value and whole-rock silica content rather than the expected positive correlation. The minimum δ¹⁸O value occurs at an SiO₂ content of about 45 wt%. Below 45 wt% SiO₂, magmas evolved via a combination of assimilation, fractionational crystallization and crystal accumulation; above 45 wt% SiO₂, magmas appeared to have evolved via closed-system fractional crystallization. Received: 23 November 1998 / Accepted: 27 September 1999     

Experimental petrology of alkalic lavas: constraints on cotectics of multiple saturation in natural basic liquids

Experimental study of natural alkalic lava compositions at low pressures (pO₂QFM) reveals that crystallization of primitive lavas often occurs in the sequence olivine, plagioclase, clinopyroxene, nepheline without obvious reaction relation. Pseudoternary liquidus projections of multiply saturated liquids coexisting with plagioclase (±olivine±clinopyroxene±nepheline) have been prepared to facilitate graphical analysis of the evolution of lava compositions during hypabyssal cooling. Use of (TiAl₂)(MgSi₂)_–1 and Fe³⁺ (Al)_–1 exchange components is a key aspect of the projection procedure which is succesful in reducing a wide range of compositions to a systematic graphical representation. These projections, and the experiments on which they are based, show that low pressure fractionation plays a significant role in the petrogenesis of many alkalic lava suites from both continental and oceanic settings. However, the role of polybaric fractionation is more evident in the major element chemistry of these lava suites than in many tholeiitic suites of comparable extent. For example, the lavas of Karisimbi, East Africa, show a range of compositions reflecting a polybaric petrogenesis from primitive picrites at 1360° C/18 kb and leading to advanced low pressure differentiates. Evolved leucite-bearing potassic members of this and other suites may be treated in a nepheline-diopside-kspar (+olivine+leucite) projection. Compositional curvature on the plagioclase+clinopyroxene+olivine+leucite cotectic offers a mechanism to explain resorption of plagioclase in alkalic groundmass assemblages and the incompatibility of albite and leucite. This projection is useful for evaluating the extent of assimilation of the alkalic portions of crustal granulites. Assimilation appears to have played some role in the advanced differentiates from Karisimbi.     

Mineral chemistry and petrogenesis of an ultrapotassic-ultramafic volcanic rock

Compositions of the major phenocryst minerals (olivine, phlogopite) and groundmass minerals (olivine, phlogopite, kalsilite), and a glass phase have been determined from a biotite mafurite occurring as an ejected block in the highly K-rich ultramafic rocks of south west Uganda. Comparison of the phenocryst mineral compositions with those determined from recent high pressure experiments on biotite mafurite composition suggests this rock may have formed by partial melting of a K-enriched mantle source containing both H₂O and CO₂ at approximately 1,250 ° C and 30 kb. The absence of crystalline leucite but its presence as a major component of the glass phase and textural relations in the groundmass indicate that the final consolidation of the biotite mafurite took place at pressures greater than atmospheric. The presence of phlogopite, olivine, kalsilite, and glass mainly of leucite composition may suggest that consolidation took place under the conditions where these phases were in equilibrium. Based on the experimentally determined conditions for the reaction of phlogopite break down to olivine+kalsilite +liquid+vapor, a crude estimation of the consolidation conditions for ejected blocks of biotite mafurite are 1,150 °–1,180 ° C at a of 1–2 kb.     

Stability of the assemblage albite plus forsterite at high temperatures and pressures with petrologic implications

The maximum limits of the assemblage albiteforsterite have been determined experimentally at high pressures and temperatures. At subsolidus temperatures, albite plus forsterite is replaced at high pressures by jadeitic clinopyroxene and enstatitic orthopyroxene. The boundary for this reaction lies within experimental uncertainity of that for jadeite=albite+nepheline. Melting of albite+forsterite at high pressures produces enstatite+liquid, which is different from the low-pressure eutectic behavior. Melting rates are very slow and several hundred hours are required to establish equilibrium near the solidus. The subsolidus boundary for albite plus forsterite lies near that for sanidine plus forsterite, but with a shallower slope which more closely matches that of anorthite plus forsterite. Both albite plus forsterite and anorthite plus forsterite are replaced at high pressures by an assemblage containing clinopyroxene plus orthopyroxene, unlike sanidine plus forsterite, which is replaced by a feldspathoid plus orthopyroxene. The presence of sodium enlarges the depth region over which plagioclase lherzolite can stably exist; it may also stabilize alkali feldspar plus olivine in crustal rocks.     

A regular solution model for met-aluminous silicate liquids: Applications to geothermometry,immiscibility, and the source regions of basic magmas

Adopting a set of multioxide components and using published compositional data on olivineand plagioclase-liquid equilibria we have developed a 17 component regular solution model for met-aluminous silicate liquids. The partial molar excess free energies predicted from this model can be used together with phenocryst compositions as an effective geothermometer, with an approximate error of 20 °C (30 °C for olivine, 12 °C for plagioclase). The regular solution formulation is also successful in predicting liquid immiscibility at (1) high mole fractions of silica commonly observed in phase diagrams, and at (2) lower temperatures in lunar basalts and intermediate lavas. The model yields activities of silica which are consistent with those obtained from solid-liquid silica buffers in rocks which contain olivine and enstatite or quartz. From predicted activities of KAlSi₃O₈ in liquids coexisting with plagioclase a value is obtained for the limiting Henry's law activity coefficient of KAlSi₃O₈ in the solid. This coefficient agrees well with that inferred from plagioclase-sanidine equilibrium phenocryst assemblages in rhyolites. The activities of silica obtained from this model are used to place constraints on the pressure-temperature regions where various types of basic magmas are generated. In conjunction with plagioclase geothermometry an application is given where the pressure, temperature, and water content of an olivine andesite is predicted from the activity of silica.