Orbicules in diorite from the Grenville Front zone of eastern Labrador consist of biotite- and/or hornblende-studded, dioritic cores enclosed by fine-grained shell structures alternately enriched and depleted in biotite. The orbicules occur in a mesocratic, quartz-bearing matrix. Epidote of inferred magmatic origin occurs in all parts of the rock. Plagioclase in the matrix is relatively sodic, and biotite more ferroan than in the orbicules, suggesting that the matrix material has the most evolved composition, and crystallized last.

The diorite is unusually aluminous (orbicules: 24.9-27.4 wt.% Al2O3; matrix: 22.4-23.6% Al2O3) and calcic (orbicules: 7.0-8.4 wt.% CaO; matrix: 6.0-6.9% CaO); it shows a positive Eu anomaly, and has elevated Sr concentrations (1800-2500 ppm Sr), demonstrating that, compositionally, it resembles a plagioclase cumulate. Mass-balance calculations suggest that the orbicule cores had a crystal/melt ratio of ≤5. This accounts for the extreme fractionation of the rock (e.g., in orbicules, Zr <5 ppm). Compared with fractional crystallization patterns, variation diagrams show counter-trends (e.g. the siliceous matrix contains elevated TiO2) or scatter for several components, suggesting that the crystal/melt ratio governed some of the geochemical characteristics of the diorite.

The presence of coarse mafic clots containing primary epidote, biotite and/or hornblende testify to an elevated water content in the orbieule cores. The shell magma apparently formed as a result of the interaction of supercooled orbicule core fluids with the matrix magma, and tended to serve as a reservoir for alkalis and Fe. Alkalis and Ca diffused in opposite directions, possibly as a result of a temperature gradient at the orbicule/matrix interface. This, however, requires decoupling of the thermodiffusional behaviour of alkalis and femic components in hydrated intermediate magma, which contrasts with documented Soret diffusion in mafic systems.

The solidification of the shell magma prior, to the orbicule cores and matrix is attributed to dewatering, consistent with the fine grain size of the shell structures. Except where remobilized core material has disrupted the shells, the cores crystallized in isolation from the matrix, which fractionated toward a more evolved composition.

The Kef Lakhal amphibolitcs and associated amphibolitc and olivine-rich enclaves are dcscribcd and their major and trace element chemistry indicates that both amphibolites were evolved medium to high alumina tholeiitic basalts with talc-alkaline affinities probably formed within plate settings. The olivine-rich enclaves are disrupted periodotites of the type lherzolite-harzburgite and probably represent mantle residua after melting.

Compositional data—that is data where concentrations are expressed as proportions of a whole, such as percentages or parts per million—have a number of peculiar mathematical properties which make standard statistical tests unworkable. In particular correlation analysis can produce geologically meaningless results. Aitchison (1986) proposed a log-ratio transformation of compositional data which allows inter-element relationships to be investigated. This method was applied to two sets of geochemical data—basalts from Kilauea Iki lava lake and grantic gneisses from the Limpopo Belt—and geologically 'sensible' results were obtained. Geochemists are encouraged to adopt the Aitchison method of data analysis in preference to the traditional but invalid approach which uses compositional data.

A horizon containing graphite in the Early Silurian slates of the Eastern Central System, Central Spain, has been studied. The carbonaceous slates contain quartz, muscovite, graphitized organic matter, plagioclase, pyrite and apatite; and the slates above the graphite horizon contain quartz, phengite, plagioclase and chlorite. The origin of the graphite is related to a period of very slow accumulation rate, anoxic seafloors deep-water upwellings and plankton blooms. The chlorite and graphite are mutually exclusive and by' using the chlori te AlIV content and the graphite Co parameter, the conditions of metamorphism were calculated to be 300 ± 20°C and 1.5-2 kbar.

Geochemical investigation of samples from 20 granitic intrusions in six tectonic zones of the Hellenides shows that both I-type and S-type granites occur in the region. The I-type granites from four of the zones, namely the Rhodope Massif (RM), the Serbomacedonian Massif (SMM), the Perirhodope Zone (PRZ) and the Attico-cycladic Zone (ACZ), show some systematic differences in their geochemistry. In particular, the Rb, Y, Nb, K and Ti contents increase in the sequence PRZ, SMM, RM and ACZ. The PRZ granites are of Jurassic age, those of the SMM and RM are Eocene to Oligocene and the ACZ ones are Miocene. The differences between zones are attributed to a combination of differences in partial melting and high-pressure fractionation processes. Geochemical differences within zones are explained by variable degrees of amphibole and apatite fractionation and accumulation.

Olivine-plagioclase corona textures occur in ophitic to sub-ophitic olivine gabbros at Black Hill, South Australia. Contrasting with many corona and symplectite textures previously described, these do not involve spinel or garnet as reaction products and did not form under high-pressure conditions. Rather, the coronas formed at no more than 1 kbar pressure and are composed of a shell of orthopyroxene around the olivine often succeeded by a shell of amphibole or occasionally biotite. Beyond this, a vermicular symplectite of anorthite containing orthopyroxene and rarer amphibole vermicules extends out to host plagioclase of labradorite composition. Textural relations are used to infer a subsolidus igneous origin for all but the orthopyroxene shell which may have formed in the presence of some magma. Compositional zonation is absent from all the constituent phases except the amphibole shell which is strongly zoned in Mg# and may have a late origin. An average maximum corona width of 150- 200 μm indicates a limiting distance for subsolidus chemical diffusion. The corona products involve the reactants olivine and plagioclase in the proportions 1:3 and symplectite formation may have been promoted by a Na potential gradient. The system must also have been open to minor components including H2O and TiO2, with H2O possibly being derived from a hydrothermal system. Such systems may have been set up in the country rocks on intrusion of the magma and subsequently collapsed inwards into the pluton during sub-solidus cooling.

Coronas containing Ca-amphibole with aluminous minerals have been characterised optically and by scanning electron microscopy, analytical transmission electron microscopy and electron-probe microanalysis. The layers nearest to plagioclase are amphibole + epidote + kyanite, followed by amphibole + epidote + staurolite + spinel. These assemblages are consistent with waterundersaturated conditions, possibly at lower metamorphic grade than the commoner assemblage amphibole + spinel. Observed mineral proportions and compositions were used in a seven-layer model of steady-state, diffusion-controlled growth with local equilibrium. This model is not fully realistic, because the observed amphibole is strongly zoned from tschermakitic to actinolitic away from plagioclase, suggesting disequilibrium. However, the four-mineral layer has been successfully modelled assuming local equilibrium, with diffusion coefficients Lii larger for i = FeO and MgO than for SiO2, AlO3/2, CaO and FeO3/2. Retarded grain-boundary diffusion of the latter components is explicable by crystal-chemical effects. The number of minerals per layer is constrained by a modified form of the metasomatic phase rule of Korzhinskii, with the role of 'inert' components played by relatively immobile ones (having relatively small fluxes and relatively small diffusion coefficients).

Minute CaMg-bearing rhodochrosite inclusions were discovered in spessartine crystals of small coticule globules enclosed within redschists that form the country rock of the famous coticule bands of Belgium. This points to a former metamorphic reaction relationship between the two minerals, such as rhodochrosite + pyrophyllite = spessartine + quartz + CO2 + H2O. The coexistence of spessartine with Mg-chlorite in the coticule globule could also indicate more complex reactions such as rhodochrosite + sudoite + quartz = spessartine + Mg-chlorite + CO2 + H2O. Our finding casts some doubt on the suggestion by Kramm (1976) that spessartine formed from Mn-montmorillonite in previously glassy volcanic tufts.

Crystallographic studies of micro-inclusions in diamond coat, carried out by analytical electron microscopic techniques, show that structures possessing rhombohedral symmetry form a significant fraction of the population of well-crystaUised particles. They are, however, less frequent than apatite or biotite, which were identified earlier by the same investigative methods. Thin-foil specimens of diamond coat were prepared by mechanical fine-polishing, and subsequently ion-beam milling, sawn and polished plates of coated diamonds oriented parallel to (100) or (110). The majority of crystalline inclusions analysed were ≤0.2 p,m in diameter. Data obtained on 15 individual inclusions, including composition analysis by energy-dispersive X-ray spectroscopy, are reported and discussed. The a-axis of the hexagonal unit cell, and the c/a ratio, were determined for all specimens, and are believed to be accurate to ≈1% in most cases. In 12 out of 15 specimens, the cations with Z ≥ 11 identified comprised substantially only Mg, Fe and Ca. Values of a in these specimens ranged from 4.81 to 4.92 Å; and for three of them the space group R, corresponding to the dolomite structure, was positively identified. These crystals are classed as ankerites. Representative Mg:Fe:Ca ratios in the ankerites are 28 : 18 : 54, with cell dimensions a = 4.87 Å, c/a = 3.37. The ankerites contained small amounts of Ba, typically 1-3%, and smaller amounts of Sr, typically 0.2-2%. Three out of the 15 specimens contained >10% Ba. In two of these, Ba and Ca concentrations were roughly equal, and in one of the two, which had cell parameters a = 5.11 Å, c/a = 3.50, the space group symmetry Rc, corresponding to the calcite structure, was verified.

At the Ashanti concession, Ghana, gold-bearing quartz veins and disseminated sulphide lodes occur in narrow (1-3 m) shear zones with altered argillites and metatholeiite host rocks. The mineralisation is concealed by up to 10 m of kaolinite-mica forest ochrosol soils, beneath which is a saprolitic zone of leached rock extending down 60-70 m to the hypogene ore zone. In the unweathered hypogene orebody, gold occurs as free grains in quartz, as sub-microscopic inclusions in the disseminated arsenopyrite, as gold tellurides and as aurostibite. The gold is released from the hypogene orebody by physical dissaggregation and chemical dissolution, the latter involving hydroxyl, thiosulphate, cyanide, and fulvate complexing. Dissolution and reprecipitation of the gold appears to have taken place largely in situ with little evidence of supergene enrichment. Consequently, the gold mineralogy of the soils is complex with residual and secondary gold grains exhibiting widely different textural and compositional characteristics. Residually enriched grains display pitted, rounded surfaces and have silver-depleted rims, while supergene gold grains are compositionally homogenous and have unpitted surfaces. The supergene grains display platelet, dendritic, irregular and octahedral habit. A fine grained spongy form of gold has also been observed from weathered telluride-bearing quartz veins. Much of the secondary gold is intergrown with iron oxides and hydroxides. The gold mineralogy of the Ashanti soils appears to be controlled by physico-chemical processes active during the lateritic pedogenesis producing residual and supergene enrichment of gold.

The Mobale gold mine (Kivu, Zaire) has been worked since 1923, but the presence of aurostibite in the ore was recognized only very recently. The discovery of the mineral has also led to the recognition of its heretofore unsuspected reactivity during the amalgamation ore dressing operations. At the surface of aurostibite a layer of up to 30 μm thickness forms which has an overall brown colour and a low reflectivity, and is composed of several variegated, concentric zones. These zones contain gold, mercury, antimony and oxygen, with traces of iron and arsenic. Oxygen is present with a concentration range of 2-4 wt.% (20-30 at.%). The zones form at the expense of aurostibite, according to micrographical criteria. In complex grains formed of gold-arsenopyrite-galena-aurostibite, only aurostibite is affected by the formation of the external brown zones. Later pure gold veneers are deposited on the outermost surface of the brown layers during the processing of the ore. The newly discovered features of processed aurostibite could explain some of the difficulties met with during the dressing of antirnonian gold ores.

Electron microprobe analyses of chromite ores from Baramiya, Seifein, Siwigat, Ashayer, Um Salatit and Ras Shait, Egypt, reveal two compositional groups. The unaltered chromites from Baramiya and Seifein have high Al and low Cr contents [Cr/(Cr + Al) = 0.56] whereas the remainder have low Al and high Cr contents [Cr/(Cr + Al) = 0.73]. Such bimodality characterises Alpine-type peridotites. The second group probably crystallised at higher T and lower P than the first group and has a composition characteristic of a type III alpine-type peridotite, i.e. of arc or possibly ocean plateau origin but not of mid-ocean ridge origin. The crystallisation setting of the first group is less certain but not inconsistent with the same environment as the second group so that considering the likely higher pressure of crystallisation of the first group overall favours plutonic crystallisation in the roots of an arc for both groups.

Rim, patchy and fracture alteration of the chromite occurred possibly partly of late magmatic (deuteric) origin but mainly connected with fluid movements, serpentinisation and tectonism. The final composition of the resultant ferritchromite is variable and depends largely on the original chromite composition; the composition of the ferritchromite developed in highly cataclased chromites deviates markedly from that of the original chromite presumably due to unmixing and migration.

Cowlesite specimens from nine new occurrences in Northern Ireland and from eight localities previously reported in the literature, were crystal-chemically characterised. The chemical data for a total of 25 samples indicate a composition close to the stoichiometric formula for cowlesite and a very limited compositional range. X-ray powder diffraction spectra were performed on 8 samples. The crystallographic unit cell permit a proper indexing of the X-ray powder diffraction patterns as orthorhombic, unit cell constants close to a 23.3, b 30.6, c 25.0 Å and doubled with respect to those originally reported in literature. The powder pattern and the TG curve arc reported for an unidentified mineral species associated with cowlesite in one of the samples from Northern Ireland.

Th- and Zr-bearing perovskite and Zr-rich Ti garnet are described from the Polino calcite carbonatite, Italy. The garnets contain 6.05-7.50 wt.% ZrO2 in cores and up to 15.80 wt.% in rims. Silicon is low in these garnets and substantial Al and Fe 3+ has been assigned to the tetrahedral sites. A strong correlation of Si with Zr from published data suggests the possibility that some Zr may also be tetrahedrally coordinated. However, the chemistry alone does not constrain the site distribution of Fe, Ti, Zr and Mg unambiguously. The perovskites are very unusual in containing 2.81-3.27 wt.% ZrO2 and 1.48-1.71 wt.% ThO2. For comparison, new analyses are presented of the zirconium garnet kimzeyite and coexisting perovskite from the type locality at Magnet Cove.

The development of the morphology of baryte crystals grown in a silica gel medium is related to the changes in supersaturation. Nucleation occurs when the supersaturation in the system exceeds a threshold value which depends on the supersaturation rate, i.e. the rate at which the supersaturation is changing at the nucleation site. During growth the supersaturation decreases, although a high level of supersaturation is maintained throughout. These experimental results compare well with the morphological development predicted from theoretical models based on the periodic bond chain theory. The ontogeny of natural harytes corresponds, in considerable detail, to the experimental observations made here and strongly suggests that similarly high supersaturations have prevailed in nature. This correspondence indicates that the technique of crystal growth in a porous silica gel medium provides a good experimental model of studying natural growth processes.

Coaquandite, a new antimony oxy-sulphate hydrate, occurs as spheroidal knobs of silky fibres or, rarely, as tiny transparent colourless lamellar crystals on stibnite at the Pereta mine, Tuscany, Italy; it is associated with klebelsbergite, peretaite, valentinite, sulfur, gypsum, stibiconite, and senarmontite. Coquandite is triclinic P1, with a 11.434(7), b 29.77(4), c 11.314(4) Å, α 91.07(7)°, β 119.24(3)° γ 92.82(1)° . It has a cell volume of 3352(5)Å3 with Z = 12 and a calculated density of 5.78 g cm−3. The crystals, elongated along [001] and flattened on {010}, display polysynthetic twinning with (010) as the twin plane. Optically, they are biaxial (+) with z ≈ c, 2V ≫ 60° n = 2.08(5). The strongest lines of the X-ray powder pattern are [d in Å, (I), (hkl)] 14.84(50)(020), 9.27(41)(111, 110), 6.81(67)130, 3.304(93)(090), 3.092(100)(330).

Coquandite has also been found at the Cetine mine, Tuscany, Italy, and at the Lucky Knock mine, Tonasket, Okanogan County, Washington, USA. 22 microprobe chemical analyses (elemental microanalysis for H) gave Sb2O3 88.91, SO3 8.35, CaO 0.04, Na2O 0.03, H2O 1.43, total 98.76 wt.%, corresponding to the empirical formula (Sb + S = 7) Sb5.98Ca0.01Na0.01O7.96(SO4)1.02.0·78H2O, and to the idealised formula Sb6O8(SO4).H2O. The I.R. spectrum, which confirms the presence of water in the formula, is given.

A partial structural arrangement is also given: the Sb polyhedra lie in nine layers perpendicular to [010] and form 'hexagonally' shaped groups surrounded by SO4 tetrahedra.

Cannonite, ideally Bi2O(OH)2SO4, is a new mineral from the Tunnel Extension mine, Ohio Mining District, Marysvale, Utah, USA. It occurs mostly as intergrown crystalline aggregates (<1 mm) of subhedral to euhedral equant to prismatic crystals (<200 μm) in cavities in quartz gangue. Other associated minerals are cuprobismutite, bismuthinite, and covelline. Cannonite is colourless and transparent with an adamantine lustre and white streak. It is brittle with an uneven to conchoidal fracture. In reflected light it is low reflecting, weakly to moderately bireflectant and weakly anisotropic. Internal reflections (colourless to white) are abundant. Measured reflectance values in air and oil are tabulated. Colour values relative to the CIE illuminant C for R1 and R2 in air respectively are: Y% 10.4, 11.7; Lambdad 475,475; Pe% 2.6, 3.0. Calculated refractive indices at 589 nm: R1 1.91 and R2 1.99. VHN100 229 (range 183-280); calculated Mohs hardness is 4.

X-ray studies show that cannonite is monoclinic with space group P21/c and a 7.700(3), b 13.839(6), c 5.686(2) Å, β 109.11(3) ° It has a cell volume of 572.5(4) Å3 with Z = 4. Dcalc. is 6.515 g/cm3. The strongest six lines of the X-ray powder pattern are [d in Å (I) (hkl)] 3.206 (100) (221); 1.984 (90) (340, 152); 2.924 (70) (131); 3.644 (60) (111); 3.466 (60) (040); 2.782 (50) (112). Averaged probe analyses gave the empirical formula Bi1.99O(OH1.04)2S0.99O4 on the basis of 7 oxygen atoms. The name is for Benjamin Bartlett Cannon of Seattle, Washington, United States of America.