EXAFS spectroscopy has been used to monitor changes in divalent cation site geometries across the P2/c-P1̄ phase transition in the sanmartinite (ZnWO4)-cuproscheelite (CuWO4) solid solution at ambient and liquid nitrogen temperatures. In the ZnWO4 end member, Zn occupies axially-compressed ZnO6 octahedra with two axial Zn-O bonds at approximately 1.95 Å and four square planar Zn-O bonds at approximately 2.11 Å. The substitution of Zn by Cu generates a second Zn environment with four short square planar Zn-O bonds and two longer axial Zn-O bonds. The proportion of the latter site increases progressively as the Cu content increases. Cu EXAFS reveals that the CuO6 octahedra maintain their Jahn-Teller axially-elongate geometry throughout the majority of the solid solution and only occur as axially-compressed octahedra well within the stability field of the Zn-rich phase with monoclinic long-range order.

Low-temperature experiments in the ‘dry’ ternary Cu-Bi-S system, conducted by using sulphidation methods down to 120°C produced a new metastable solid solution series Cu10Bi2S13-Cu5Bi2S8 at 178°C coexisting with CuS. This transformed slowly at 190–200°C to an assemblage of either CuS-(Cu,Bi)8S9 or CuS-Bi2S3 or both, depending on available sulphur. Sulphidation experiments on Cu3BiS3 similarly revealed a solid solution range for the phase (Cu,Bi)8S9 of up to Cu/Bi = 3/2 at 178–190°C, and a lower stability limit of 138°C Isothermal sections of the system were constructed at 200 and 300°C, based on the new information collected but excluding the metastable series.

Metasomatic interaction on a cm scale between calc-silicate pods and the enclosing sillimanite + biotite + tourmaline gneiss at Partridge Breast Lake, northern Manitoba, Canada, led to the development of an inner (by calc-silicate rock), hornblende-rich reaction zone and an outer, biotite-rich zone. The boundary between the reaction zones is interpreted as the original calc-silicate/metapelite interface. Compared with its metapelitic protolith, the biotite zone shows a two- to twenty-fold depletion in the concentrations of incompatible trace elements (notably the light rare earths, U, Th, Nb, Ta, Zr and Hf). In contrast, the relative concentrations of trace elements remained nearly constant during the mineralogical transformation of the calc-silicate rock to the hornblende zone. The depletion of trace elements in the biotite zone is attributed to the dissolution of accessory phases (e.g. monazite). Although stable at the metamorphic conditions (∼600–650°C at ∼ 4.5 kbar) prevalent during metasomatism, Mg-rich tourmaline is absent in the biotite zone, suggesting that either the pH or composition (e.g. the (Al + Si)/(Ca + Mg + Fe) ratio) of the aqueous fluid phase was inappropriate for the preservation of this mineral.

A 65 × 107 µm grain of euhedral tetrataenite (ordered FeNi) attached to a similarly sized grain of troilite occurs within an impact-melt rock clast in the Jelica LL6 chondrite breccia. After impact melting, immiscible metallic Fe-Ni and troilite droplets formed within the silicate melt progenitor of the clast. At ⩾1200°C while the surrounding silicate was still partly molten, euhedral taenite with ∼ 50 wt.% Ni began crystallizing in one of the metal-troilite droplets. Troilite nucleated at one edge of the euhedral taenite grain and began to crystallize at ∼870°C. At 320°C the metal phase underwent an ordering reaction and formed tetrataenite. The unrecrystallized clast-host boundary and the differences in olivine composition and degree of polycrystallinity of troilite between the clast and Jelica host indicate that the clast was incorporated into Jelica during a late-stage brecciation event.

The compositions of coexisting and individual cooperite (ideally PtS) and braggite (ideally (Pt,Pd)S) grains from the Merensky Reef of the Bushveld Complex, as well as cooperite, braggite and vysotskite (ideally PdS) grains from the UG-2 of the Bushveld Complex were investigated. There is a clearly defined miscibility gap between cooperite and braggite, but no evident gap between braggite and vysotskite. Partition coefficients between cooperite and braggite are determined on coexisting phases. The KDbraggite/cooperite in atomic ratios are estimated to be 0.54 for Pt, 15.81 for Pd and 5.93 for Ni. For Rh and Co the KDbraggite/cooperite are estimated to be > 1.40 and > 1.46 respectively. No systematic behaviour is detected for Fe and Cu. Coupled substitutions of Pd + Ni for Pt in cooperite and braggite/vysotskite are indicated. Within the cooperite of the Merensky Reef, the Pd:Ni ratio is approximately 9:11. The substitution trend in braggite, which extends to vysotskite in the UG-2, is dependent on the base-metal sulphide (BMS) association. If pentlandite is the dominant Ni-bearing BMS, the Pd:Ni ratio is about 7:3 in the Merensky Reef and in the UG-2. Millerite as the dominant Ni-bearing BMS in the UG-2 changes this ratio to 3:1. It is concluded that the Ni-content in braggite/vysotskite from BMS assemblages does not depend on the NiS activity, but rather on temperature of formation.

A thermodynamic prediction of the Gibbs free energy of formation (ΔGfo) of nukundamite (empirical composition Cu5.5FeS6.5) was made in order to specify whether the nukundamite + chalcopyrite or the bornite + pyrite assemblage is stable in the Cu-Fe-S system. The results of calculations using previously reported data of ΔGfo values of some Cu-Fe-sulphide minerals in equilibrium with nukundamite indicate that the total free energy of the nukundamite + chalcopyrite assemblage is appreciably higher than that of the bornite + pyrite assemblage in the temperature range 250–400°C. This means that nukundamite + chalcopyrite is a metastable assemblage under common ore-forming conditions.

The occurrence of nukundamite is not uncommon in the Fijian kuroko deposits in contrast to the Japanese kuroko deposits. A thermochemical treatment for this phenomenon leads to the interpretation that the black ore containing nukundamite in the Fijian deposit was formed under relatively highsulphidation and low-pH conditions. This suggestion is in good agreement with the present experimental result that the bornite + pyrite assemblage was produced in the temperature range 350–250°C by using near-neutral hydrothermal solutions.

Cobalt-rich spinel is found as a ∼200 µm inclusion, together with a glassy phase, in a gem-quality blue sapphire from Bo Ploi, Thailand. This is the first reported natural occurrence of such a spinel. Its composition is directly analogous with that of cochromite, a previously reported rare cobalt-rich chromite. The compositional ranges for the cobalt-rich spinel, obtained using electron microprobe and proton microprobe (methods described below), are Al2O3 48.18–61.27 %, CoO 19.7–22.84 %, Cr2O3 0–12.28 %, FeO 8.64–9.67 %, MgO 6.04–6.89, TiO2 0.49–0.73 %, Ni 2251–2532 p.p.m., Zn 335–371 p.p.m., Mn < 177–849 p.p.m., Ga 113–153 p.p.m., Nb 24–1252 p.p.m., Zr <4–167 p.p.m., Sn 22–428 p.p.m., As <4–56 p.p.m., Sr <4–59 p.p.m., Ag 13–64 p.p.m. Transition elements decrease in abundance from core to rim of the spinel while the other elements increase. Crystal chemical considerations suggest that a vacancy-creating substitution mechanism may be operative in the cobalt-rich spinel despite the small scale, i.e. 3Co2+ = 2Al3+ + [4]□. The glassy phase coexisting with the spinel is likely to be the product of heating by the host basaltic magma. The mode of occurrence of the cobalt-rich spinel prevents further physical investigation. This unusual spinel is considered to be the result of a complex magma mixing process in the lower crust.

A new compositional variety of zirconolite characterized by high Mg, Al, Y2O3 and REE, and low Fe is described from a sapphirine granulite xenolith entrained in an intrusive norite body which was emplaced into the late Archaean (2520–2480 Ma) Vestfold Hills high-grade terrain during the early Proterozoic. The zirconolite, and similarly Mg-Al rich perrierite-(Ce), formed as a result of sanidinite facies partial melting of the particularly magnesian and aluminous sapphirine granulite xenolith during its incorporation into the c. 1170°C basic magma at c. 2240 Ma. The high REE compared to Al cations require that a previously unrecognized coupled substitution:

occurs in this zirconolite. Full chemical analyses are presented for zirconolite and perrierite from this unique occurrence.

Detailed examination of ‘staringite’ by X-ray precession photography and high-resolution transmission electron microscopy shows it to consist of a sub-microscopic intergrowth of cassiterite and tapiolite. ‘Staringite’ is discredited as a valid mineral species.

Nifontovite and olshanskyite, two rare hydrous calcium borate minerals, have been found in crystalline limestone near gehlenite-spurrite skarns at Fuka, Okayama Prefecture. Nifontovite occurs as aggregates of tabular crystals up to 5 cm long and 1.5 cm wide, and rarely as euhedral crystals up to 1 mm long. Olshanskyite occurs as anhedral masses, or as micro-twinned platy crystals up to 1 cm long. Wet chemical analyses give the empirical formulae Ca3.052B5.991O6.038(OH)12·1.96H2O and Ca2.888B3.997(OH)18 on the basis of O = 20 for nifontovite and OH=18 for olshanskyite, respectively. The formulae are consistent with those from type localities.

The X-ray powder data for these minerals were determined with accuracy. The unit cell parameters of nifontovite agree closely with those published previously. X-ray studies show that olshanskyite is triclinic with the possible space group P1̄ or P1 and a = 9.991(5), b = 14.740(11), c = 7.975(3) Å, α = 94.53(4), β = 69.08(3), γ = 112.44(5)° and Z = 3. The density 2.19 g cm−3 (meas.) obtained for olshanskyite agrees with the estimated ideal value 2.31 g cm−3 (calc.). Nifontovite was formed by hydrothermal alteration of an anhydrous borate, and olshanskyite was formed by hydrothermal alteration of nifontovite and the anhydrous borate.

In specimens of Mn skarn from the type locality of Långban, hyalotekite, (Ba,Pb,K)4Ca2−(Si,B,Be)12O28F, occurs in a matrix consisting mostly of aegirine (⩽22 mol.% CaMnSi2O6), andradite (⩽27 mol.% Mn3Fe2Si3O12), hematite, pectolite, quartz, calcite, baryte, barylite, and hedyphane. Melanotekite, plumbian taramellite, ferrian K-feldspar (to 6.5 wt.% Fe2O3), rhodonite, a talc-like mineral, apophyllite, and several Pb-As-Sb-O minerals are found in trace amounts. In a single specimen of reedmergnerite-microcline pegmatite from Dara-i-Pioz, hyalotekite occurs in close association with microcline. Other accessory minerals are albite, aegirine, pyrochlore, eudialyte, and polylithionite. The optical constants for hyalotekite from Långban and Dara-i-Pioz are, respectively, α = 1.656, 1.646, β = 1.659–1.660, 1.649, γ = 1.670–1.671, 1.659 (all ± 0.002), 2Vγ (mcas) = 57.2–60.5 ± 0.5°, 57.0 ± 1.1°(λ = 589 nm). Cell parameters of the Dara-i-Pioz hyalotekite for a body-centred triclinic cell are: a = 11.284(2), b = 10.930(1), c = 10.272(8) Å, α = 90.35(2)°, β = 90.11(3)°, γ = 89.98(1)°. Electron and ion microprobe data show that Långban hyalotekite is heterogeneous even within a given sample; the most important substitutions are Pb = Ba, K and coupled B = Si and B = Be. Minor constituents include Mn in the Långban hyalotekite and Na in the Dara-i-Pioz hyalotekite, which also differs in its significantly higher Ba/Pb ratio. Conditions suggested for hyalotekite formation at Långban and Dara-i-Pioz are P ⩽ 4 kbar, T ⩾ 500°C, silica saturation, peralkalinity, and relatively high oxygen fugacities and low sulphur fugacities. These temperatures must have been sufficiently high to allow for miscibility of Pb with Ba (and K) despite the lone pair of electrons of Pb2+.

Apatite in most igneous intrusions has a high Cl/F ratio. However, chlor-apatite has been reported in the lower portions of the Bushveld and Stillwater Complexes. This has been used as evidence supporting the early separation of a Cl-rich discrete hydrous fluid in these intrusions. Mineralogical evidence is presented here to demonstrate that the Bushveld Complex, at least, formed from a nearly anhydrous magma, and did not release a hydrous fluid before apatite began to crystallize. It is suggested that apatite in the earliest cumulates equilibrated with trapped interstitial liquid, which converted it from the typical F-rich composition of cumulus apatite to a Cl-rich composition. This is an analogous process to that in which cumulus mafic minerals may become more Fe-rich on cooling and reaction with interstitial liquid.

A study of apatite crystals from the Asio mine, Japan, showed sectoral texture related to the growth of the crystal, and with optically biaxial properties within the sectors. Wet chemical analysis gave a composition Ca5(PO4)3(F0.64,OH0.38,Cl0.01)1.03 for the specimen.

Additional diffraction spots were not observed in precession and oscillation X-ray photographs and electron diffraction photographs. Since the internal textures correlate with the surface growth features, it is suggested that the internal textures and the unusual optical properties were produced during nonequilibrium crystal growth. The fluorine/hydroxyl sites in hexagonal apatite are symmetrically equivalent in the solid crystal but, at a growth surface, this equivalence may be lost, resulting in a reduction of crystal symmetry. Heating of the apatite to about 850°C results in the almost complete disappearance of the optical anomalies due to disordering, which may be related to the loss of hydroxyl from the crystal.

In the arid, Late Precambrian terrain of southern Israel, a complex suite of minerals and amorphous species were deposited in host gneiss from fluids under near-neutral conditions within 1 m of the surface. The morphology of secondary gold appears to relate to its host mineral (skeletal-dendritic with quartz; multi-faceted crystals with arsenates; spherical droplets with iron oxide). The gold is very fine-grained, and was most likely complexed as a thiosulphate.

Three amorphous phases are present (iron oxide, chrysocolla, Cu-Mn-(Fe-As) silicate). At least in part, gold and baryte appear to have crystallized out of a metal-Fe-oxide gel. Other minerals, including apatite, anglesite, and conichalcite, may have grown from appropriate crystallites present in the gel.

The conichalcite occurs mainly as bladed to acicular radial spherulites. In the presence of lead, a solid solution phase between duftite and conichalcite (‘Pb-conichalcite’) was formed.

A suite of orthopyroxenes from spinel Iherzolite xenoliths associated with basanites occurring in the Victorian (Australia) post-Pliocene ‘Newer Volcanics’ province was investigated by means of a crystal chemical methodology which provides accurate site occupancy and site configuration parameters.

The M1 configuration is essentially constrained by AlVI rather than Fe2+. In addition, Fe3+, Cr3+ and Ti4+ are confined to M1 (Molin, 1989) and AlIV to TB. M2 is controlled by FeM22+ ⇌ MgM2, constrained by (Fe2+ + Ca)M2 > 0.14 atoms per formula unit (p.f.u.). Cation substitution in TB and M2 constrains the sum of the volumes of the respective polyhedra VTB+VM2 to remain essentially constant. Therefore, M2 favours the retention of the large Fe2+ up to melting-point, causing non-ideality of this iron-depleted orthopyroxene. As a consequence, the investigated orthopyroxene can be considered an ultimate Fe2+ carrier during partial mantle melting.