The Rocky Beach Metamorphic Melange contains metre-scale phacoids of high-P low-T metamorphic rocks embedded in chlorite-actinolite schist. The phacoids include eclogite, glaucophane schist and omphacitite and provide evidence for four episodes of metamorphism with mineral assemblages: M1 = actinolite-glaucophane-titanite-apaite, M2 = almandine-omphacite-lawsonite ±quartz, M3 = phengite- glaucophane-K-feldspar-quartz, and M4 = chlorite-actinolite-calcite-quartz-titanite-white mica ± albite ± talc. M1-M3 occurred at a Neoproterozoic-Early Palaeozoic convergent plate boundary close to the eastern margin of Gondwana. Peak metamorphic conditions were attained during the static phase M2, with temperatures of ~560°C and pressures in excess of 1.8 GPa, equivalent to a depth of burial of at least 54 km.

The chemical and structural alteration of metamict zircon crystals from a 619 ±17 (2σ) Ma old, posttectonic granite in the southern part of the Eastern Desert, Egypt was studied. The crystals show simple oscillatory growth zones with metamictization–induced fractures, which provided pathways for fluid infiltration. Electron and ion microprobe analyses reveal that metamict, i.e. U and Th–rich, areas are heavily enriched in Ca, Al, Fe, Mn, LREE, and a water species, and have lost Zr and Si as well as radiogenic Pb. These chemical changes are the result of an intensive reaction with a low–temperature (120—200°C) aqueous solution. The chemical reactions probably occurred within the amorphous regions of the metamict network. During the zircon–fluid interactions the metamict structure was partially recovered, as demonstrated by micro-Raman and -infrared measurements. A threshold degree of metamictization, as defined empirically by an α–decay dose, Dc, was necessary for zircons to undergo hydrothermal alteration. It is proposed that Dc marks the first percolation point, where the amorphous domains start to form percolating clusters in the metamict network and where bulk chemical diffusion is believed to increase dramatically. The time of the hydrothermal alteration is determined by a lower intercept age of a U-Pb SHRIMP discordia of 17.9 (2σ) Ma, which is in good agreement with an apatite fission track age of 22.2 (2σ) Ma. The hydrothermal alteration event occurred contemporaneously with the main rifting phase of the Red Sea and widespread low- temperature mineralizations along the Red Sea coast.

The North Qôroq centre comprises a series of concentric nepheline syenite intrusions and forms part of the Igaliko Nepheline Syenite Complex, in the rift-related Gardar Province of South Greenland. The North Qôroq syenites range from mildly undersaturated augite syenite to strongly peralkaline agpaitic nepheline syenite. Extensive in situ fractional crystallization has been postulated for the chemical variation both within units and throughout the centre. Many of the rocks have been affected by metasomatic fluids associated with the emplacement of younger syenite units, and this complicates their interpretation. In this study, the trends and compositions exhibited by pyroxene and amphibole from North Qôroq are examined and related to either primary crystallization or metasomatic activity (e.g. controls of fO2, peralkalinity). Implications thus drawn are used to interpret the chemical processes inherent in the chemical and fluid evolution of alkaline magmas, and, in particular, the transition from miaskitic to agpaitic magmatism. In general, the major phases of the North Qôroq syenites records the increasing evolution of the units by crystal fractionation, towards peralkaline compositions. The composition of olivine, in the least evolved syenites, also points to a relatively high state of fractionation of the parent magma, whilst pyroxene and amphibole record an overall decrease in Mg/Mg+Fe), and a general increase in Fe3+ and alkali content, with increased fractionation. The increasing peralkalinity of the magma also governs the evolution of pyroxene and, to a lesser degree, amphibole towards higher Zr and Ti contents in the more Na-rich compositions. The trends for pyroxene from metasomatized syenite show similar patterns, but lower Fe2+ enrichment, suggesting the source of the metasomatic fluids is similar to the evolving syenites. The presence of amphiboles in metasomatic rocks, and high F contents attest to the F-rich nature of the metasomatic fluids, which is in agreement with results previously reported for metasomatic fluorapatite.

Airborne dust generated by primary blasting was collected in Taffs Well Quarry, just north of Cardiff, Wales. Collections of airborne particulate matter were also made in the nearby village of Morganstown at the same time as the blasting collections. The explosions were recorded on a motor-driven camera and a digital video camera. These images show that the dust clouds generated by the explosions consist of three distinct components; a reddish-grey dust cloud, followed by a light grey dust cloud, and finally a pale grey cloud that stayed near the blast face. It is believed that the reddish-grey cloud was composed mostly of mineral grains, as evidenced by the chiefly red colour of the dolomitic limestone rock in the quarry. The whiter clouds contained more explosive combustion particles (diesel soot). The samples were studied by analytical scanning electron microscopy, very high-resolution field emission scanning electron microscopy, and image analysis. The two different components (minerals and diesel soot) can be readily seen under high-resolution electron microscopy. Any consideration of the possible adverse health effects or nuisance value of this dust needs to consider both of these components. A size distribution of the quarry particles shows that soot particles dominate the assemblage under 2 mm, whereas the mineral grains are more abundant over 2 mm. This contrasts with the Morganstown particle sizes, where the two components show similar size distributions. The determination of the mineralogy of the quarry dust and Morganstown particles has shown highly complex and heterogeneous mixtures, though some distribution patterns are emerging. It is concluded that much of the dust in Morganstown probably originated from sources other than the quarry, such as other local industries, roads and construction sites.

Based on petrography, mineral chemistry, and petrology, the physico-chemical evolution of an agpaitic dyke was found to be very similar to that of the neighbouring Ilímaussaq complex. Various mineral assemblages were used to reconstruct the crystallization conditions of the dyke rock for different stages during cooling. The early magmatic phenocryst assemblage is alkali feldspar + nepheline + augite + olivine + magnetite and indicates liquidus temperatures of ∼850ºC, silica activities of ∼0.5, and oxygen fugacities of FMQ –1.5 to –3. The groundmass assemblage albite + microcline + nepheline + sodalite + arfvedsonite + aegirine + aenigmatite + astrophyllite indicates lower temperatures of between 600 and 450ºC, at silica activities of 0.25, and oxygen fugacities around the FMQ buffer. Amphibole composition strongly responds to fluorite saturation and proves crystallization occurred in a system closed to oxygen. Late-stage hydrothermal conditions are indicated by the conversion of nepheline and sodalite to analcime and the growth of aegirine on arfvedsonite. These late-stage reactions are constrained to temperatures of <300°C, water activities of between 0.5 and unity, and oxygen fugacities above MH. The dyke has to be regarded as a small equivalent of the larger Ilímaussaq complex, in which identical differentiation processes proceeded at a scale very different in terms of magma volume and cooling times.

Work was undertaken in 1997 as part of a UK government programme to monitor air quality in the Cornish china clay area. These studies showed that concentrations of airborne particulate matter with an aerodynamic diameter <10 μm (PM10) did not exceed UK National Air Quality Strategy (NAQS, 1995) guidelines but were, however, greater than those normally associated with rural areas. Subsequently, in the summer of 2000, airborne dust sampling was conducted at three sites around the IMERYS (formerly ECCI) Kernick mica dam, a potential source of resuspended airborne particulates from china clay waste minerals. Analytical scanning electron microscopy examination of the collected material suggested that up to 61% of the airborne PM10 monitored next to the mica dam could be associated with china clay extraction compared with 46% 2 km to the south. Results showed the presence of substantial percentages of respirable kaolinite, muscovite and quartz; the major components of china clay waste. The health effects of the particular mineral dust emissions associated with china clay mining are reviewed.

Blackhead Quarry exploits a small Miocene (~10 Ma) basanitic volcanic centre in the Dunedin Volcanic Group, New Zealand. Vesicles near the quarry top contain Ca– and Na-rich zeolites, abundant calcite and rare pyrite resulting from localized low-T hydrothermal alteration (<100°C). Mineral assemblages were characterized by EDS and laser Raman spectroscopy with the latter the most useful in determining the identity of fibrous zeolite species. Secondary mineral assemblages crystallized during final stages of lava cooling from aqueous solutions enriched in Ca, Na, K, Ba and (CO3)2–leached from surrounding calcareous rocks and basanite. Phillipsite-K (generation I) crystallized first (in some places directly on fresh basanite) followed by the Ca-rich zeolites, chabazite-Ca, calcian phillipsite–K (generation II), gismondine and thomsonite. Later Na-rich fluids crystallized gonnardite and natrolite, and finally calcite from late Ca-rich fluids. Zeolite composition is not reflected by morphology. For example, both phacolitic and pseudocubic chabazite are chabazite-Ca, and although all phillipsite crystals have a similar habit, their composition varies widely. Various lithologies comprising the Blackhead volcanic centre have unique secondary mineral paragenetic sequences controlled largely by the rock structure and solution chemistry.

The single-crystal X-ray structure of a sulphate-bearing buttgenbachite, Cu36Cl7.8(NO3)1.3(SO4)0.35(OH)62.2.5.2H2O, from Likasi, Democratic Republic of Congo, has been determined at 100 and 288 K. The basic framework of the structure is the same as has been previously reported for buttgenbachite, except for the identification of a hydrogen-bonded chloride ion (occupancy 0.6) at the origin instead of a Cu ion with partial occupancy. The nature of nitrate positional disorder along channels in the c direction and how this relates to the presence of other species such as chloride ions and water molecules and, most importantly, sulphate ions has been elucidated. One nitrate ion, with an occupancy of 0.18, lies at 2/3,l/3,l/4 and shares the site with a chloride ion (occupancy 0.30) and also a sulphate ion (occupancy 0.09); a second nitrate, with an occupancy of 0.24, lies at 2/3,1/3,0.084 and shares the site with a water molecule (occupancy 0.06). As a result, a formula of Cu36Cl7.8(NO3)1.3(SO4)0.35(OH)62.2.5.2H2O is obtained. Re-refinement of deposited data for a supposed connellite crystal from the Toughnut mine,Tombstone, Arizona gives a related, but different, pattern of anion substitution. No sulphate could be detected in the structure and it is evident that this structure refers to buttgenbachite. A nitrate nitrogen atom and a chloride ion are disordered at 2/3,l/3,l/4, the overall site being fully occupied. A chloride ion with ∼0.5 occupancy is sited at the origin and the formula Cu36Cl7.9(NO3)1.1(OH)63.4H2O is indicated. Re-refinement of a deposited data set for another buttgenbachite crystal from the Likasi mine reveals a partially occupied nitrate centred at 2/3,l/3, z and a partially occupied chloride at 2/3,l/3,l/4. Either 0.5Cl–, OH–, H2O or H3O+ is located at the origin. If the latter is the case, the stoichiometry for this buttgenbachite is Cu36Cl6.5(NO3)1.5(OH)64.5.5H2O. The present study has highlighted the fact that a range of compositions for buttgenbachite exists, depending on the pH and relative activities of chloride, nitrate and sulphate ions in solutions from which the mineral crystallizes.

We have imaged the spatially-preserved microtexture of biogenic apatite, retained together with its collagen template, in non-demineralized human bone using high-resolution transmission electron microscopy. Using ion-beam thinning, a specimen preparation method generally employed for inorganic minerals rather than for biological materials, we have imaged a composite nanostructure of bone not previously reported, and we propose a model for this nano-architecture that involves a boxconstruction of apatite plates and apatite sheets. This observation provides a new understanding of bone strength at the nanometre scale and suggests how post mortem enhancement of this texture by recrystallization probably accounts for the durability of ancient bone. Modern sheep bone (a close analogue for recently dead human bone) imaged in the same way also shows evidence of this composite architecture.

Hydroxylian pseudorutile in an adamellite from Cameroon has been identified using EPMA, FT-IR and single-crystal XRD analysis. TiO2 and Fe2O3 contents range between 60.6 and 67.9 wt.% and 22.2 and 36.5 wt.%, respectively, and totals vary from 90.1 to 98.1 wt.%. The SiO2 (0.03 —1.40 wt.%), Al2O3 (0.02—0.68 wt.%) and CaO (0.03—0.44 wt.%) contents increase systematically and MnO (0.04—0.69 wt.%) decreases slightly with decreasing Fe3+/Ti values. The MgO content is <0.06 wt.%. H2O contents (1.3—7.5 wt.%) are greater in parts of crystals with lower Fe3+/Ti, suggesting solid solution between the end-members Fe3+2Ti3O9 and Fe3+Ti3O6(OH)3. The lattice parameters are a = 2.85 Å and c = 4.57 Å with hexagonal symmetry and point group 6/mmm.

A carbonatite sample from Phalaborwa, South Africa, consists of apatite, magnetite and a calcitedolomite ‘perthite’ which is interpreted as being due to exsolution of dolomite from a high-Mg calcite precursor. Carbon and oxygen isotope data indicate that the carbonates are equilibrated. In situ ionmicroprobe analyses for Fe, Mn, Na, Si, Y, the REEs, Pb, Th and U give the following average concentrations (in ppm) in the sequence apatite, calcite, dolomite: Fe 98, 1680, 8190; Mn 61, 510, 615; Na 1171, 627, 125; Si 368; 1.6, 0.2; Sr 4447, 5418, 2393; Ba 37, 2189, 75; La 1245, 300, 67; Y 121, 50, 5.8; Pb 16, 5.4, 1.4; Th 20, 0.02, 0; U 2.4, 0, 0.01. The concentrations are reasonably uniform in both apatite and dolomite, but in calcite are more variable. Na, Si, Y, the REEs, Pb, Th and U partition into apatite relative to both carbonates (and, hence, the precursor carbonate); KD ap/cc for REE decreases from ∽4 for La to ∽2 for Tm. There is almost equal partitioning of Sr between apatite and calcite. During separation of dolomite from calcite, Sr and Ba partition strongly into calcite and all the other analysed elements, except Fe and Mn, also preferentially enter calcite. The REEs prefer calcite relative to dolomite, and the KD dol/cc is reasonably constant, only varying from 0.23 to 0.17. Sr, Ba and Pb in the carbonates, and their partitioning between the calcite and dolomite, differ from other carbonatite carbonates reported in the literature.

Rhönite occurs in lower crustal pyroxenite xenoliths erupted in phonolite from the Mount Sidley composite volcano, Marie Byrd Land, Antarctica, as a localized breakdown product, with plagioclase, clinopyroxene, ± olivine ± Ti-magnetite + melt, of kaersutite, and as microphenocrysts (with olivine, plagioclase, clinopyroxene) in pockets of basanitic melt. Rhö nite after kaersutite has a more NaSi-rich/ CaAl-poor composition, lower Ti, and formed at higher oxidation (∼NNO) conditions than rhönite occurring as microphenocrysts in basanite. Comparison with experimentally determined rhönite stability in understaturated alkali basalt and as a reaction product after Ti-amphibole indicates that the Mount Sidley rhönite (and associated minerals) formed between 1090 and 1190°C at <0.5 kbar, presumably during temporary residence of the xenoliths in a shallow magma chamber below the volcanic edifice.

The PM10 (airborne particulate matter with aerodynamic diameter <10 mm) in Beijing has a distinct seasonality, with industrial, domestic and natural sources providing a heterogeneous cocktail of airborne particulate matter (PM). Collections were made during late winter, summer and high wind dust storms to determine composition and probable sources of this PM. The concentration of the PM during winter (174 μg m–3) was approximately four times higher than summer (37 μg m–3) with dust storms raising the concentration further (200 μg m–3). During the winter the PM was dominated by combustion products (66% filter area). During the summer combustion products and loess contributed ~35% to the filter area each, but during elevated wind speeds (>10 mph) loess completely dominated the collections (96% filter area). The majority of the PM10 collected was in the respirable (PM2.5) size range (winter 99.7%, summer 96.6%, dust storms 82.3%). The loess in Beijing comprises quartz, feldspar, calcite, chlorite and mica and is in the coarse silt to sand (20–60 mm) size range. The collections are therefore likely to be made up of finer silt and clay, primarily derived from of the erosion of cultivated land. Using a plasmid assay, the Beijing particulate matter was found to have little or no surface free radical activity.

A LREE-, Si-, S-enriched fluorapatite with ∼7.7 wt.% LREE2O3, 5.3 wt.% SiO2 and 3.5 wt.% SO3 from the Mushugai-Khuduk REE deposit, southern Mongolia, has been experimentally metasomatized in pure H2O at 900°C and 1000 MPa. Despite the high LREE content, inclusions or rim grains of monazite did not form during fluid-induced metasomatism. Under high-contrast BSE imaging, the reacted fluorapatite presents a mottled appearance with light, moderate and dark areas and hosts large fluid inclusions containing anhydrite as the solid phase. Dark areas show moderate depletion in all the LREE. Measurable HREE such as Gd and Y remain unchanged. In the areas of intermediate brightness, Ce and, to a lesser extent, La are moderately enriched relative to the original fluorapatite. In the light areas, Ce and La display strong enrichment, whereas the other LREE, Gd and Y remain little changed. In the dark areas, depletion in LREE is matched by depletion in Si and Na, satisfying the coupled substitution reactions Si4+; + (LREE)3+ = P5+ + Ca2+ and Na+ + (LREE)3+ = 2Ca2+. In the light areas, enrichment in LREE and P is accompanied by depletion in Si, Na and S, implying operation of two different coupled substitution reactions, namely: 2P5+ = Si4+ + S6+ and 2P5+ + LREE3+ = 2S6+ + Na+. In all areas, strong enrichment in F is accompanied by strong depletion in Cl and OH. The formation of large, fluid and anhydrite-filled inclusions associated with the LREE-enriched light areas is evidence of a negative volume change, combined with additional local dissolution of the fluorapatite. This occurred during the metasomatism of the original Cl- and OH-bearing fluorapatite to a nearly pure end-member fluorapatite. Local dissolution of fluorapatite in the space now occupied by the voids, combined with element remobilization from the body of the fluorapatite as a whole, provided the Ca and S needed for the formation of anhydrite as well as the P and LREE required for the enrichment of the bright areas.

Chromian tourmaline, in association with other Cr-bearing minerals (amphibole, mica, epidote, chlorite, titanite, rutile and chromian spinel), occurs in fine calc-schist levels within metacarbonate rocks from the Nevado-Filabride Complex, SE Spain. Electron microprobe analyses of tourmaline and coexisting minerals document both chemical differences dependent on the host-rock type and an irregular distribution of Cr at grain scale. Tourmaline is Na-rich dravite, with average Mg/(Mg+Fe) ratios of 0.83 and 0.63 a.p.f.u. and Cr contents of 0.32 and 0.18 a.p.f.u., in dolomitic and ankeritic marbles, respectively. Tourmaline contains small but significant concentrations of Zn (av. 0.01 a.p.f.u.) and in ankeritic marble it also contains Ni (av. 0.04 a.p.f.u.). Zn-rich chromian spinel appears as small relict inclusions in silicates, with average Cr, Fe, Al and Zn contents of 1.201, 1.241, 0.411 and 0.107 a.p.f.u., respectively. Amphibole, epidote, mica and chlorite show average Cr contents of 0.088, 0.138, 0.115 and 0.267 a.p.f.u., respectively, in dolomitic marbles, and 0.103, 0.078, 0.065 and 0.185 a.p.f.u., respectively, in ankeritic marbles. Cr-silicates formed through metamorphic reactions involving detrital Cr-rich spinel, in addition to clay minerals and carbonates. The B necessary to form tourmaline was probably derived from the leaching of underlying evaporitic rocks.

Knowledge of the solid-phase speciation of As in Bengali sediments associated with hazardous As-rich groundwaters is crucial to understanding the processes controlling As release. The local coordination environment of As in such a sediment has been probed using K-edge As EXAFS. This revealed that As exists predominantly in its oxidized form, As(V), probably adsorbed as bidentate arsenate tetrahedra on metal (Fe and/or Al) oxide/hydroxide surfaces, although incorporation of As into a metal oxide structure cannot be ruled out. Arsenic was found to occur in several different coordination environments and this, together with the low concentration (<10 μg g–1) of As in the sediment prevented the unambiguous assignment of the second coordination sphere. The EXAFS analysis of the sediment after incubation under anaerobic conditions in the presence of added electron donor for metal reduction indicated changes in the relative concentrations of different solid-phase As species, providing circumstantial evidence for differential susceptibility to microbial action.

The ∽16 Ma Rangwa Caldera Complex, part of the large Kisingiri nephelinite-carbonatite volcano, Homa Bay District, western Kenya (0º34’S; 34º09’E) contains carbonatitic lapilli and ash tuffs, agglomerate and tuffisite, and a number of intrusive calcite carbonatites. A detailed petrographic and electron microprobe study has been performed on 20 fresh samples from the collection at The Natural History Museum, London.

Most of the juvenile lapilli and ash particles are either predominantly composed of devitrified silicate glass (now biotite/phlogopite but probably also originally potassic silicate) or calcite carbonatite, which suggests that two molten liquids were erupted simultaneously. Some 10 mm-diameter lapilli contain quench-textured calcite crystals set in devitrified glass. They are interpreted as having crystallized from a molten silicate-carbonate melt at, or very near, the surface.

The extrusive carbonate is mostly composed of calcite, consistent with intrusive calcite compositions at Rangwa. Other key minerals are magnetite, two types of mica (magnesian-biotite phenocrysts and phlogopite xenocrysts) and fluorapatite.

The pyroclastic rocks contain many calcite carbonatite clasts, and fragments of calcite, aegirine and diopside, fluorapatite, magnetite, plus some phlogopite, titanite, K-feldspar, fenite and glimmerite; ijolite lithics are rare. Thus, there is no evidence for a cognate nephelinitic (ijolitic) or melilitic magma nor evidence for a direct relationship with the nephelinites of the Kisingiri volcano.

Two hypotheses are discussed. A rising silicate and K-rich carbonatite liquid may have evolved towards a carbonate-rich K-silicate liquid after crystallization of calcite, phlogopite, apatite and magnetite. Preservation of the the potassic component may be rare, with a more usual scenario being that potassic component separates as fenitizing fluids. The alternative is that the silicate component is remobilized fenite, formed from country rock that was mobilized by supercritical K-rich, fenitizing fluids associated with the carbonatite. Both scenarios require generation of a K-rich carbonatite magma, probably from a carbonated phlogopite-rich metasomatized mantle.

The Chernobyl accident in 1986 resulted in the widespread identification of the post-accident presence of radioactive (or ‘hot’) particles across large areas of Eastern and Central Europe. Such particles arise from direct deposition and also from condensation and interactions on particle surfaces during and following the deposition of soluble fallout. Identification of the presence and nature of hot particles is necessary in order to determine the long-term ecological impact of radioactive fallout. This paper describes several techniques for the identification and characterization of hot particles in soil samples from Belarus. In addition to new results from the use of gamma spectrometry, we include two novel instrumentation approaches that have been developed and applied to Chernobyl fallout-contaminated soils. The first, ‘differential’ autoradiography, utilizes a photographic film sandwich to characterize the nature of the ionizing radiation emitted from samples. In this paper we show that differential autoradiography can not only identify hot particle presence in soil, but can also determine the dominant radionuclide in that particle. The second approach, sector field ICP-MS (ICP-SFMS), can provide rapid, high-precision determination of the actinides, including the transuranic actinides, that characteristically occur in hot particles originating from weapons fallout or fuel matrices. Here, ICP-SFMS is shown to yield sufficiently low detection limits for plutonium isotopes (with the exception of 238Pu) to enable us to confirm negligible presence of plutonium in areas outside the Chernobyl exclusion zone, but with high levels of fission-product contamination.

The interaction of the {104} cleavage surface of calcite with Hg(CH3COO)2 aqueous solutions with concentration of 5 mM Hg(II) (pH ≈3.5), was investigated using microscopic and spectroscopic techniques. In situ atomic force microscopy experiments showed that surface microtopography changes significantly as a result of the interaction, and that the initial rhombic etch pits induced by H2O dissolution are rapidly transformed to deeper etch pits exhibiting an unusual triangular shape. The growth of these etch pits is strongly anisotropic, moving faster along the [21] direction than along the [010] direction (with step-retreat velocities of ∼12 nm s –1 and ∼4 nm s–1, respectively). The modified etch pits are due to Hg(II) sorption in the surface, rather than due to the effect of the acetate anion. The sorption (adsorption and probably absorption also) of Hg(II), in the first minutes of the interaction, is shown by X-ray photoelectron spectroscopy. After ∼2 h, the triangular etch pits are interconnected to form larger hexagonal etch pits, while Hg(II)-bearing phases (confirmed later by SEM-EDS) grow onto the surface through a heterogeneous nucleation process. The crystal growth of orthorhombic (montroydite-type) hydrated Hg(II) oxide (HgO·nH2O) on the surface of calcite was confirmed by XRD patterns and FT-IR spectra from samples exposed for longer times to Hg(CH3COO)2 solution.

Eleven nepheline-kalsilite crystalline solutions with various proportions of K:Na have been studied from room temperature to 1050/11500C by X-ray powder diffraction. Nepheline expansion is relatively high and little affected by composition, whereas kalsilite expansion is lower but affected to a significant degree by K:Na ratio. The generally higher rate of expansion in nepheline is apparently related to the collapse of the tetrahedral framework around the smaller of its two alkali sites. Occupancy of these sites by the relatively small Na ion fürther extends the potential for thermal vibration before the structure is stretched to the critical degree required for phase transformation. Once the structure changes to that of kalsilite, with its single alkali site, an increase in content of the larger K ion limits the degree to which kalsilite can expand. Crucial to the overall expansion behaviour of these minerals are the specific tetrahedral configurations of nepheline vs. kalsilite, the number and geometry of their alkali sites, the occupancies of those sites, and the flexibility inherent in each structure that allows for adjustment with increasing temperature.