Intrinsic oxygen fugacities of diogenites and mesosiderite clasts

Oxygen fugacities of diogenite and mesosiderite clast material were measured with the double ZrO₂ cell technique between 800° and 1150°C. The samples were taken from large clasts in the diogenites Johnstown (En₇₃) and Tatahouine (En₇₅), and the mesosiderites Estherville (En₈₁), West Point (Fo₈₈) and Emery (En₆₈). Fugacity values for all except Emery plot near the wüstite-iron buffer curve and are interpreted as indicating similar source regions and environments of crystallization for the two suites. Emery orthopyroxene records a lower fugacity, close to the fayalite-quartz-iron buffer curve, probably as a result of equilibration with the mesosiderite matrix assemblage. The similarity of redox conditions experienced by mesosiderite orthopyroxenite and diogenites is not sufficient to require a single parent body and, if the common achondrites were derived from Vesta, mesosiderites probably came from a different body.     

Petrology, Mineralogy and Geochemisty of Antarctic Mesosiderite GRV 020175: Implications for Its Complex Formation History

<正>GRV 020175 is an Antarctic mesosiderite,containing about 43 vol%silicates and 57 vol% metal.Metal occurs in a variety of textures from irregular large masses,to veins penetrating silicates, and to matrix fine grains.The metallic portion contains kamacite,troilite and minor taenite.Terrestrial weathering is evident as partial replacement of the metal and troilite veins by Fe oxides.Silicate phases exhibit a porphyritic texture with pyroxene,plagioclase,minor silica and rare olivine phenocrysts embedded in a fine-grained groundmass.The matrix is ophitic and consists mainly of pyroxene and plagioclase grains.Some orthopyroxene phenocrysts occur as euhedral crystals with chemical zoning from a magnesian core to a ferroan overgrowth;others are characterized by many fine inclusions of plagioclase composition.Pigeonite has almost inverted to its orthopyroxene host with augite lamellae, enclosed by more magnesian rims.Olivine occurs as subhedral crystals,surrounded by a necklace of tiny chromite grains(about 2-3μm).Plagioclase has a heterogeneous composition without zoning. Pyroxene geothermometry of GRV 020175 gives a peak metamorphic temperature(～1000℃) and a closure temperature(～875℃).Molar Fe/Mn ratios(19-32) of pyroxenes are consistent with mesosiderite pyroxenes(16-35) and most plagioclase compositions(An_(87.5_96.6)) are within the range of mesosiderite plagioclase grains(An_(88-95)).Olivine composition(Fo_(53.8)) is only slightly lower than the range of olivine compositions in mesosiderites(Fo_(55-90)).All petrographic characteristics and chemical compositions of GRV 020175 are consistent with those of mesosiderite and based on its matrix texture and relatively abundant plagioclase,it can be further classified as a type 3A mesosiderite.Mineralogical, penological,and geochemical studies of GRV 020175 imply a complex formation history starting as rapid crystallization from a magma in a lava flow on the surface or as a shallow intrusion.Following primary igneous crystallization,the silicate underwent varying degrees of reheating.It was reheated to 1000℃,followed by rapid cooling to 875℃.Subsequently,metal mixed with silicate,during or after which,reduction of silicates occurred;the reducing agent is likely to have been sulfur.After redox reaction,the sample underwent thermal metamorphism,which produced the corona on the olivine, rims on the inverted pigeonite phenocrysts and overgrowths on the orthopyroxene phenocrysts,and homogenized matrix pyroxenes.Nevertheless,metamorphism was not extensive enough to completely reequilibrate the GRV 020175 materials.     

Cooling rate based on schreibersite growth for the Emery mesosiderite

The Emery mesosiderite contains large Ni-rich grains of the phosphide schreibersite, which have exsolved from kamacite. Computer simulation of diffusion-controlled growth of this schreibersite indicates that exsolution occurred during cooling at the rate of 0.1°C/Myr. This determination supports the cooling rate estimated for mesosiderites by Powell (Geochim. Cosmochim. Acta33, 789–810), using taenite-kamacite data.     

Mesosiderites—I. Compositions of their metallic portions and possible relationship to other metal-rich meteorite groups

The metal from 17 mesosiderites has been analyzed for Ni, Ga, Ge and Ir by the techniques of atomic-absorption spectrometry and neutron activation. Most mesosiderite metal samples fall in a narrow compositional range: Ni, 7·0–9·0 per cent; Ga, 13–16 ppm; Ge, 47–58 ppm; and Ir, 2·4–4·4 ppm. Most of those falling outside these ranges belong to Powell's (1971) least-metamorphosed type. Mesosiderite metal falls in the same general composition range as IIIAB irons, IIIE irons, pallasites and H-group chondrite metal. There are distinct differences in detail, however, and firm evidence for a close genetic relationship between any of these groups and the mesosiderites is lacking. Metallic portions of Weekeroo-type irons tend to have slightly higher Ni, Ga, Ge and Ir contents than found in mesosiderite metal, and the two groups tend to form a single trend on all plots. The Weekeroo-type silicates closely resemble mesosiderites in terms of orthopyroxene composition and oxygen-isotope ratio. We interpret these similarities to indicate that the silicate and metallic portions of these two groups are closely related; if the mesosiderite silicates and metal were initially formed in separate parent bodies, these were of similar composition and formed at about the same distance from the Sun.     

Olivines and olivine coronas in mesosiderites

Olivines and their surrounding coronas in mesosiderites have been studied texturally and compositionally by optical and microprobe methods. Most olivine is compositionally homogeneous but some is irregularly zoned. It ranges from Fo_58–92 and shows no consistent pattern of distribution within and between mesosiderites. Olivine occurs as large single crystals or as partially recrystallized mineral clasts, except for two lithic clasts. One is in Emery, the other in Vaca Muerta, and they are both shock-modified olivine orthopyroxenites. $\text{FeO}\text{MnO}$ ratios in olivine exhibit a variety of differing trends and range from 22–46, most commonly 35–40. These values are lower than those in olivine from diogenites sensu stricto (45–50) and have therefore experienced a different history. Some of the olivine clasts could have coexisted with some of the large orthopyroxene clasts as equilibrium assemblages, but some could not. Much of the olivine may be derived from mesosiderite olivine orthopyroxenites, which differ from diogenites sensu stricto. More magnesian olivine may be a residue from one or more source rocks, with varying degrees of melting. These events probably occurred in a highly evolved and differentiated parent body.Fine-grained coronas surround olivine, except for those in impact-melt group mesosiderites (Simondium, Hainholz, Pinnaroo) and those without tridymite in their matrices (Bondoc, Veramin). Coronas consist largely of orthopyroxene, plagioclase, clinopyroxene, chromite, merrillite and ilmenite and are similar to the matrix, but lack metal and tridymite. Coronas contain abundant orthopyroxene but are unusually rich in chromite (up to 7%) and merrillite (up to 20%). The outer parts of the corona grade into the matrix, but have little or no metal and tridymite. Texturally the innermost part of the corona can be divided into three stages of development: I Radiating acicular; II Intermediate; III Granular. Stage I is the result of the greatest disequilibrium between olivine and matrix orthopyroxene and Stage III has the least disequilibrium. Coronas are the result of the reaction olivine + tridymite = orthopyroxene, probably because FeO (and MgO) diffuse from olivine to tridymite in the matrix. Absence of metal and concentration of chromite in the corona are probably the result of an FeO potential gradient away from the olivine. Merrillite concentrations are a result of P₂O₅ migration into the corona but are controlled by the availability of calcic pyroxene, or possibly plagioclase. Although the coronas are texturally similar to terrestrial and lunar counterparts, they are unique and represent different kinds of reactions marked by a large degree of intra-corona diffusion under dry conditions. Opaque oxide-silicate-metal buffer assemblages yield apparent equilibration conditions of about 840°C and fO₂ near 10^?20. Poikiloblastic pyroxene textures in some coronas suggest a closing of reaction systems between 900 and 1000°C and such systems may record a higher temperature stage of development.     

Mesosiderites and howardites: igneous formation and possible genetic relationships

We report neutron activation data for major, minor and trace elements determined in whole rock howardites and silicates from mesosiderites. Compositions of howardites and mesosiderites are similar, and intermediate between those of eucrites and diogenites. Relative to howardites mean mesosiderite abundances are slightly nearer the diogenites. Literature data indicate that mesosiderites have a higher normative silica component than howardites. It appears that this partly results from a higher content of a highly evolved igneous component, and partly from in situ reduction of FeO to Fe followed by magnetic separation of metal prior to analysis. Removal of a portion of the FeO in this manner yields a higher normative SiO₂ component for the nonmagnetic fraction. Petrographic observations demonstrate the formation of SiO₂ which may have resulted from a combination of various factors including accretion of a reducing agent together with the Fe-Ni metal, extensive reaction during the long cooling period, and catalysis by the finely divided metal.In the mesosiderites Mincy, Lowicz and Veramin the light rare earth elements (REE) are enriched. The resulting REE pattern is qualitatively similar to that in terrestrial basalts thought to have been formed by small degrees of partial melting. Of several partial melting models tested, the best match to the REE patterns is provided by one involving ~2–4% partial melting of a source containing low REE abundances. It appears that the light REE enrichment is not associated with the hypothetical silica enriched igneous phase.Since numerous properties separate mesosiderite silicates from howardites, it is clear that they are not composed of precisely the same material. Whether or not they originated on the same parent body is unresolved. If parent body regoliths were mixed vertically and horizontally on a planet-wide basis, then separate bodies would be required.     

The effect of composition on the ferric-ferrous ratio in basaltic liquids at atmospheric pressure

The effects on the ferric-ferrous ratio of varying individual components in a dry basaltic liquid have been determined at atmospheric pressure and constant oxygen fugacity (fO₂). Experiments were conducted by suspending 100 mg samples from pt loops at 1200°C (fO₂ = 10^?8atm) and 1360°C (fO₂ = 10^?6atm) in an atmosphere controlled by mixtures of CO₂ and H₂. A microanalytical wetchemical technique and the electron microprobe were used to determine the composition of the resulting basaltic glasses. In order of decreasing significance, the addition of oxides of K, Na, Si, Al, or Ca produces an increase in the ferric-ferrous ratio of the melt at 1200°C. The change in the ferric-ferrous ratio produced by component addition is less at 1360°C than at 1200°C.     

Neutron capture records of mesosiderites and an iron meteorite

The Sm and Gd isotopic compositions of silicates from six mesosiderites (Dalgaranga, Estherville, Morristown, Northwest Africa (NWA) 1242, NWA 2932, and Vaca Muerta) and one iron meteorite (Udei Station) were determined to elucidate the cosmic-ray exposure records. All seven samples showed significant ¹⁵⁰Sm/¹⁴⁹Sm and ¹⁵⁸Gd/¹⁵⁷Gd isotopic shifts from neutron capture reactions corresponding to neutron fluences of (1.3-21.8) × 10¹⁵ n cm⁻². In particular, Vaca Muerta showed a significantly higher neutron fluences than the other six samples. The parameter for the degree of neutron thermalization (ε_Sm/ε_Gd) also showed a significant difference between Vaca Muerta (0.76) and the other samples (0.93-1.20). These results suggest a two-stage irradiation of the Vaca Muerta silicates in the parent body (>50 Ma) before formation of the mesosiderite and during its transit to Earth (138 Ma). This is consistent with the ⁸¹Kr-Kr cosmic-ray exposure age data of a Vaca Muerta pebble from a previous noble gas isotopic study.     

On the thermal and redox history of the Yamato diogenite Y-74013

The Yamato diogenite, Y-74013, shows a high degree of textural equilibrium with the apparent crystallization sequence: troilite and metal → orthopyroxene → plagioclase. The position of the large chromite crystals in this sequence is unclear. Except chromite, all other minerals have composition similar to common orthopyroxene achondrites. The chromite is more magnesian than in common diogenites, strongly zoned and, on the whole, intermediate in composition between chromites of diogenites and pallasites. Texture, mineral composition data and an equilibrium thermodynamic analysis of the mineral association strongly indicate that the chromite crystallized earlier than the silicates at a much higher temperature (possibly above 1100°C) and rapidly grew in a medium which was progressively enriched in Mg, Al and Ti. But the chromite failed to reach chemical equilibrium, even at its outermost rim, with the orthopyroxene. The calculated equilibrium log fO₂ of the Yamato diogenite, ?20·21 to ?11·08 for temperatures between 880°C and 1500°C is well within the normal oxygen fugacity range of pyroxene achondrites.     

Conditions of Formation of Iron–Carbon Melt Inclusions in Garnet and Orthopyroxene under <Emphasis Type="Italic">P-T</Emphasis> Conditions of Lithospheric Mantle

Of great importance in the problem of redox evolution of mantle rocks is the reconstruction of scenarios of alteration of Fe⁰- or Fe₃C-bearing rocks by oxidizing mantle metasomatic agents and the evaluation of stability of these phases under the influence of fluids and melts of different compositions. Original results of high-temperature high-pressure experiments (P = 6.3 GPa, T = 1300–1500°С) in the carbide–oxide–carbonate systems (Fe₃C–SiO₂–(Mg,Ca)CO₃ and Fe₃C–SiO₂–Al₂O₃–(Mg,Ca)CO₃) are reported. Conditions of formation of mantle silicates with metallic or metal–carbon melt inclusions are determined and their stability in the presence of CO₂-fluid representing the potential mantle oxidizing metasomatic agent are estimated. It is established that garnet or orthopyroxene and CO₂-fluid are formed in the carbide–oxide–carbonate system through decarbonation, with subsequent redox interaction between CO₂ and iron carbide. This results in the formation of assemblage of Fe-rich silicates and graphite. Garnet and orthopyroxene contain inclusions of a Fe–C melt, as well as graphite, fayalite, and ferrosilite. It is experimentally demonstrated that the presence of CO₂-fluid in interstices does not affect on the preservation of metallic inclusions, as well as graphite inclusions in silicates. Selective capture of Fe–C melt inclusions by mantle silicates is one of the potential scenarios for the conservation of metallic iron in mantle domains altered by mantle oxidizing metasomatic agents.     

Formation of the Fe,Mg-Silicates,Fe<Superscript>0</Superscript>, and Graphite (Diamond) Assemblage as a Result of Cohenite Oxidation under Lithospheric Mantle Conditions

Experimental studies in the Fe₃C–SiO₂–MgO system (P = 6.3 GPa, T = 1100–1500°C, t = 20–40 h) have been carried out. It has been established that carbide-oxide interaction resulted in the formation of Fe-orthopyroxene, graphite, wustite, and cohenite (1100 and 1200°C), as well as a Fe–C–O melt (1300–1500°C). The main processes occurring in the system at 1100 and 1200°C are the oxidation of cohenite, the extraction of carbon from carbide, and the crystallization of metastable graphite, as well as the formation of ferrosilicates. At T ≥ 1300°C, graphite crystallization and diamond growth occur as a result of the redox interaction of a predominantly metallic melt (Fe–C–O) with oxides and silicates. The carbide–oxide interaction studied can be considered as the basis for modeling a number of carbon-producing processes in the lithospheric mantle at fO₂ values near the iron–wustite buffer.     

Experimental determination of the partitioning of gallium between solid iron metal and synthetic basaltic melt: Electron and ion microprobe study

The distribution of Ga between solid Fe metal and synthetic basaltic melt has been investigated experimentally at two temperatures over a limited range of oxygen fugacities. Reversal experiments were conducted, indicating a close approach to equilibrium. Analysis of run products was performed using an electron and an ion microprobe. At 1 bar total pressure, the solid metal/silicate melt partition coefficient. D(Ga), is given by: 1190°C: logD(Ga) = ?0.92 logfO₂ ? 11.91330°C: logD(Ga) = ?0.77 logfO₂ ? 8.8. For the common assumption of a valence state for Ga of 3 in silicates and assuming ideal solution of Ga in both phases, a slope of ?0.75 is predicted. The slope obtained at 1300°C (?0.77) is indistinguishable from this value, while the slope at 1190°C (?0.92) is somewhat higher. Henry's law is obeyed over the concentration range 0.007 wt.% to 0.15 wt.% Ga in metal, a factor of 20 in concentration. These partition coefficients may be utilized to evaluate metal/silicate fractionation processes in the Earth, Moon and Eucrite Parent Body. The lunar mantle appears to be depleted in Ga by a factor of 20–40 relative to CI abundances. This depletion is consistent with extraction of Ga into a geophysically plausible lunar core if Ga was initially present at a subchondritic concentration. A similar explanation probably accounts for the depletion of Ga in the Eucrite Parent Body. The upper mantle of the Earth appears to be depleted in Ga by a factor of 4– 7 relative to CI abundances. This depletion is far smaller than would be expected as a result of core formation, but is consistent with two quite different hypotheses: (i) a small amount of solid metal and sulfur-bearing metallic liquid was retained in the mantle after core formation; or (ii) addition of a late-stage chondritic component involving 5% to 10% of the upper mantle following core formation.     

Solubility of the buffer assemblage pyrite + pyrrhotite + magnetite in NaCl solutions from 200 to 350°C

In the design of hydrothermal solubility studies it is important that the system be completely defined chemically. If the solubilities of minerals containing m metallic elements are to be determined in hydrothermal NaCl solutions, the phase rule requires that a total of m + 6 independent intensive parameters be controlled or measured in order to determine completely the system.In this study the solubility of the univariant assemblage pyrite + pyrrhotite + magnetite has been determined in vapor saturated hydrothermal solutions from 200 to 350°C for NaCl concentrations ranging from 0.0 to 5.0 molal. At any temperature, oxygen and sulfur fugacities were buffered by the chosen assemblage. System pH was determined from excess CO₂ partial pressures and computed ionic equilibria. Equilibrium constants were calculated by regression analysis of solubility data. The results show that more than 10 ppm of each mineral can dissolve in typical hydrothermal solutions under geologically realistic conditions. Solubilities were best represented by the species Fe²⁺ and FeCl⁺ at 200 and 250°C; Fe²⁺, FeCl⁺ and FeCl₂⁰ at 300°C; and Fe²⁺ and FeCl₂⁰ at 350°C. Ore deposition would occur by lowering temperature, diluting chloride concentration, or by raising pH through wall rock alteration reactions.     

Experimental investigation of the partitioning of phosphorus between metal and silicate phases: implications for the Earth,Moon and Eucrite Parent Body

The solid metal/silicate melt partition coefficient for P, D(P), has been determined experimentally at 1190°C and 1300°C. The dependence of the partition coefficient on oxygen fugacity has been investigated, and is consistent with a valence state of 5 for P in the silicate melt. The experimental partition coefficients are given by: $\text{log D(P) = ?1.21 log ?O}{}_2\text{? 15.95 at 1190°C}$$\text{log D(P) = ?1.53 log ?O}{}_2\text{? 17.73 at 1300°C}$The experimentally determined partition coefficients may be used to interpret the low $\text{P}\text{La}$ ratios of the Earth, Moon and eucrites relative to C1 chondrites. The low $\text{P}\text{La}$ ratios in the eucrites may be explained by partitioning of P into 5% to 25% of a sulfur-bearing metallic liquid assuming equilibration and separation of the liquid metal from the silicates at low degrees of partial melting of the silicates. The low $\text{W}\text{La}$ ratios in the eucrites compared to C1 chondrites require the separation of an additional 2% to 10% solid metal.The lowering of both $\text{P}\text{La}$ and $\text{W}\text{La}$ ratios in the Moon may be explained by partitioning of P and W into metal during formation of a small core by separation of liquid metal from silicate at low degrees of partial melting of the silicates. The $\text{W}\text{La}$ ratios in the Earth and Moon are virtually indistinguishable, while $\text{P}\text{La}$ ratios differ by a factor of two. The concentrations of FeO also appear to be different. These observations are difficult to reconcile with the hypothesis of a terrestrial origin of the Moon following formation of the Earth's core, but are consistent with independent formation of the Earth and Moon.     

Phlogopite: High temperature solution calorimetry,thermodynamic properties,Al-Si and stacking disorder,and phase equilibria

Methods have been developed for solution calorimetry of hydrous phases in molten lead borate near 700°C. These involve thermochemical cycles using dissolution and decomposition reactions of hydrous silicates and hydroxides. Preliminary results suggest that H₂O derived from the decomposition of hydroxides dissolves in molten 2PbO-B₂O₃ with an exothermic enthalpy of solution of −5.7 ±0.7 kcal mol⁻¹. Hydroxyphologopite persists metastably at 714°C and its heat of solution in 2PbO·B₂O₃ has been measured. From these new data, the standard enthalpy of formation of phlogopite from the elements at 25°C is −1485.5 ±1.5 kcal mol⁻¹. The standard free energy of formation is -1394.6 ±1.5 kcal mol⁻¹, assuming complete tetrahedral Al-Si disorder.Two structural features complicate the thermodynamics of synthetic and natural micas. The first is a varying degree of tetrahedral Al-Si disorder. Raman spectroscopic study of phlogopite synthesized above 600°C suggests a disordered Al-Si distribution. Calculations of the P-T locus of the geologically important equilibrium: Phl + 3Qtz = 3En + Sa + H₂O, using our thermochemical data, agree within experimental error with the results of calculations based on the best available phase equilibrium data only if a tetrahedrally disordered phlogopite is assumed. Such calculations are very sensitive to uncertainties in ΔH° and ΔG°, and reversed phase equilibrium experiments remain essential to obtaining reliable estimates of thermodynamic properties. In contrast to these Al-Si disordered phlogopites, some biotites of low temperature parageneses (<600°C) may have substantial Al-Si order. A variable Al-Si distribution has a substantial effect on the configurational entropy and therefore on the free energy of the mica in question. Because of these and other questions, applications of biotite equilibria to determining volatile fugacities in igneous and metamorphic petrogenesis are subject to large uncertainties.The second structural complication is stacking disorder, which is present in phlogopite synthesized at 650°C but not in the 850°C sample. The enthalpy difference between these two samples, determined by solution calorimetry, is smaller than the experimental uncertainty of ±1.0 kcal mol⁻¹. Thus there appears to be little driving force for ordering, and micas with disordered stacking sequences may persist in many geologic environments. The effect of stacking disorder on thermodynamic properties is probably very small.     

Thermal expansion of anhydrous Mg-cordierite between 25 and 950° C

Thermal expansion measurements on synthetic orthorhombic Mg-cordierite (“low cordierite”), Mg₂Al₄Si₅O₁₈, were carried out with a high-temperature X-ray Guinier camera. The measurements confirm previous studies reporting low thermal expansion and suggesting a closer relationship with ring rather than framework silicates. No indication for structural modifications correlated with symmetry changes was observed. However, two discontinuities in the thermal expansion function at 275±25° and 600±50° C were detected and are assumed to represent higher-order phase transitions. The first discontinuity is related to a minimum in specific volume, implying unusual expansion behavior (negativ thermal coefficient) at low temperatures. An estimate of the dP/dT-slope of the established discontiniuties yields a positive sign for the one at 275° C and a negative one for that at 600° C.     

Fluid-bearing and fluid-absent invariant points in the system CaO-MgO-Al2O3-SiO2-CO2-H2O for a greenschist facies assemblage

An outline for a metamorphic grid involving a greenschist facies assemblage is presented in Watts (1973). This grid is derived from a theoretical determination of all possible P-T dependent (solid-solid) reactions and invariant points as well as those in which a fluid, containing CO₂, H₂O and other unspecified components, takes part. The thermodynamic data in Watts (1973) were accidentally calculated at 425° K and not at 425° C. Since 425° K is an unreasonably low temperature for greenschist facies equilibria, these data have been recalculated at 700° K (427° C) and a new topology of fluid-absent (solid-solid) reactions and invariant points in P-T space results, since the slopes of such reactions change under conditions at the higher temperature. The slopes of the fluid-bearing reactions are independent of P, T and remain unchanged. However, for each different P-T grid of solid-solid reactions, a new set of chemographic arrangements is valid for the fluidbearing invariant points. A set of μ_H2O-μ_CO2 diagrams consistent with, and dependent on the new P-T grid is presented.     

Geochemistry of actinolitic hornblendes from tonalitic rocks,Northern Portugal

Chemical analyses are given for actinolitic hornblendes of tonalitic rocks from the Hercynian belt of Northern Portugal. The distribution of elements between amphibole and co-existing biotite is studied. The composition of the amphiboles is analysed in the light of experimental data on amphiboles and the physical conditions of crystallization inferred from the study of the biotite and rock series. The data on the biotites lead to the definition of a temperature of 800°C for the crystallization of actinolitic hornblendes with Mg/(Mg + Fe) ratios of 0·72-0·61 at pressures of about 3 Kb and fO₂ defined by FMQ.     

Structure of mineral glasses—III. NaAlSi3O8 supercooled liquid at 805°C and the effects of thermal history

The distribution of interatomic distances in amorphous NaAlSi₃O₈ has been determined at 805°C by X-ray radial distribution analysis to investigate structural differences between the glass (T < 763°C) and the supercooled liquid (763°C < T < 1118°C). Except for slight differences attributable to thermal expansion, no significant changes were observed. The sample crystallized during the course of the experiment, but at least one crystal-free data set was obtained. The transition from the inferred six-membered ring structure of the supercooled liquid to the four-membered ring structure of the crystal was clearly visible in radial distribution function (RDF's) determined before and after crystallization.RDF's were also determined at 25°C for two NaAlSi₃O₈ glasses with different histories. The first was derived from a melt that had been cooled slowly from 1600 to 32°C above the melting point (T_f = 1118°C) to detect possible repolymerization to a more ‘crystal-like’ structure as the melt approached T_f. The second glass was prepared by holding a single crystal of Amelia albite at 50°C above T_f to see if the crystalline four-membered ring structure was preserved in melts at temperatures just above the liquidus. No significant differences were observed between these two RDF's and one obtained from a glass quenched from 1800°C. These results suggest that in addition to the destruction of formation of a periodic structure, melting and crystallization in NaAlSi₃O₈ also involves a repolymerization of tetrahedra. This would explain the observed kinetic barrier to melting and crystallization in the anhydrous system and the catalytic effect of small amounts of water or alkali oxide.