High concentrations of vanadium cause very unusual coloration in hibonite (purple) and grossite (light violet) crystals in an exotic mineral assemblage from Sierra de Comechingones (Argentina). In the hibonite (CaAl₁₂O₁₉) structure vanadium ions, in various valence states (divalent, trivalent, and tetravalent), may be distributed over five crystallographic sites with coordinations corresponding to different polyhedra, namely, three unequal octahedra [M1 (D_3d), M4 (C_3ν), and M5 (C_s)], one M3 tetrahedron (C_3ν), and one unusual fivefold-coordinated trigonal bipyramid M2 (D_3h). Possible locations of vanadium ions in grossite (CaAl₄O₇) are limited to two crystallographically distinct sites (T1 and T2, both C₁) in tetrahedral coordination.

The combination of single-crystal X-ray diffraction and absorption spectroscopy techniques aided by chemical analyses has yielded details on the nature of the vanadium-induced color in both hibonite and grossite crystals. In hibonite, both M4 face-sharing octahedral and M2 trigonal bipyramid sites of the R-block are partially occupied by V³⁺. Strongly polarized bands recorded at relatively low energies in optical absorption spectra indicate that V²⁺ is located at the M4 octahedral site of the hibonite R-block. Chemical analyses coupled with an accurate determination of the electron densities at structural sites in hibonite suggest that the vanadium ions occupy about 10 and 5% of the M4 and M2 sites, respectively. For grossite, polarized optical absorption spectra reveal no indications of V²⁺; all observed absorption bands can be assigned to V³⁺ in tetrahedral coordination. Although not evident by the observed electron densities at the T sites of grossite (due to the low-V content), longer bond distances, and a higher degree of polyhedral distortion suggest that V³⁺ is located at the T2 site.

The crystal structure of turneaureite, ideally Ca₅(AsO₄)₃Cl, was studied using a specimen from the Brattfors mine, Nordmark, Värmland, Sweden, by means of single-crystal X-ray diffraction data. The structure was refinedto R1 = 0.017 on the basis of 716 unique reflectios with F_o > 4σ(F_o) in the P6₃/m space group, with unit-cell parameters a = 9.9218(3), c = 6.8638(2) Å, V = 585.16(4) ų. The chemical composition of the sample, determined by electron-microprobe analysis, is (in wt%; average of 10 spot analyses): SO₃ 0.22, P₂O₅ 0.20, V₂O₅ 0.01, As₂O₅ 51.76, SiO₂ 0.06, CaO 41.39, MnO 1.89, SrO 0.12, BaO 0.52, PbO 0.10, Na2O 0.02, F 0.32, Cl 2.56, H₂O_calc 0.58, O(≡F+Cl) –0.71, total 99.04. On the basis of 13 anions per formula unit, the empirical formula corresponds to (Ca_4.82Mn_0.17Ba_0.02Sr_0.01)∑_5.02 (As_2.94P_0.02S_0.02Si_0.01)_∑2.99O₁₂[Cl_0.47(OH)_0.42F_0.11]_∑1.00.Turneaureite is topologically similar to the other members of the apatite supergroup: columns of face-sharing M1 polyhedra running along c are connected through TO₄ tetrahedra with channels hosting M2 cations and X anions. Owing to its particular chemical composition, the studied turneaureite can be considered as a ternary calcium arsenate apatite; consequently it has several partially filledanion sites within the anion columns. Polarized single-crystal FTIR spectra of the studied sample indicate stronger hydrogen bonding and less diverse short-range atom arrangements around (OH) groups in turneaureite as compared to the related minerals johnbaumite and svabite. An accurate knowledge of the atomic arrangement of this apatite-remediation mineral represents an improvement in our understanding of minerals able to sequester and stabilize heavy metals such as arsenic in polluted areas.

New occurrences of derbylite, Fe_x²⁺Fe³⁺_4–2xTi_⁴⁺3+xSb³⁺O₁₃(OH), and graeserite, Fe_x²⁺Fe³⁺_4–2xTi⁴⁺_3+xAs³⁺O₁₃(OH), have been identified in the Monte Arsiccio mine, Apuan Alps, Tuscany, Italy. Derbylite occurs as prismatic to acicular black crystals in carbonate veins. Iron and Ti are replaced by V (up to 0.29 atoms per formula unit, apfu) and minor Cr (up to 0.04 apfu). Mössbauer spectroscopy confirmed the occurrence of Fe²⁺ (up to 0.73 apfu), along with Fe³⁺. The Sb/(As+Sb) atomic ratio ranges between 0.73 and 0.82. Minor Ba and Pb (up to 0.04 apfu) substitute. Derbylite is monoclinic, space group P2₁/m, with unit-cell parameters a = 7.1690(3), b = 14.3515(7),c = 4.9867(2) Å, β = 104.820(3)° and V = 495.99(4) ų. The crystal structure was refined to R₁ = 0.0352 for 1955 reflections with F_o > 4σ(F_o). Graeserite occurs as prismatic to tabular black crystals, usually twinned, in carbonate veins or as porphyroblasts in schist. Graeserite in the first kind of assemblage is V rich (up to 0.66 apfu), and V poor in the second kind (0.03 apfu). Along with minor Cr (up to 0.06 apfu), this element replaces Fe and Ti. The occurrence of Fe²⁺ (up to 0.68 apfu) is confirmed by Mössbauer spectroscopy. Arsenic is dominant over Sb and detectable amounts of Ba and Pb have been measured (up to 0.27 apfu). Graeserite is monoclinic, space group C2/m, with unit-cell parameters for two samples: a = 5.0225(7), b = 14.3114(18), c = 7.1743(9) Å,β = 104.878(3)°, V = 498.39(11) ų; and a = 5.0275(4), b = 14.2668(11), c = 7.1663(5) Å, β = 105.123(4)° and V = 496.21(7) ų. The crystal structures were refined to R₁ = 0.0399 and 0.0237 for 428 and 1081 reflections with F_o > 4σ(F_o), respectively. Derbylite and graeserite are homeotypic. They share the same tunnel structure, characterised by an octahedral framework and cuboctahedral cavities, hosting (As/Sb)O₃ groups and (Ba/Pb) atoms.

The crystal structure of svabite, ideally Ca₅(AsO₄)₃F, was studied using a specimen from the Jakobsberg mine, Värmland, Sweden, by means of single-crystal X‑ray diffraction data. The structure was refined to R₁ = 0.032 on the basis of 928 unique reflections with F_o > 4s(F_o) in the P6₃/m space group, with unit-cell parameters a = 9.7268(5), c = 6.9820(4) Å, V = 572.07(5) Å3. The chemical composition of the sample, determined by electron-microprobe analysis, is (in wt%, average of 10 spot analyses): SO₃ 0.49, P₂O₅ 0.21, V₂O₅ 0.04, As₂O₅ 51.21, SiO₂ 0.19, CaO 39.31, MnO 0.48, SrO 0.03, PbO 5.19, Na₂O 0.13, F 2.12, Cl 0.08, H₂O_calc 0.33, O (≡ F+Cl) –0.91, total 98.90. On the basis of 13 anions per formula unit, the empirical formula corresponds to (Ca_4.66Pb_0.16Mn_0.04Na_0.03)_Σ4.89(As_2.96S_0.04Si_0.02P_0.02)_Σ3.04O₁₂[F_0.74(OH)_0.24Cl_0.01]. Svabite is topologically similar to the other members of the apatite supergroup: columns of face-sharing M1 polyhedra running along c are connected through TO₄ tetrahedra with channels hosting M2 cations and X anions. The crystal structure of synthetic Ca₅(AsO₄)₃F was previously reported as triclinic. On the contrary, the present refinement of the crystal structure of svabite shows no deviations from the hexagonal symmetry. An accurate knowledge of the atomic arrangement of this apatite-remediation mineral represents an improvement in our understanding of minerals able to sequester and stabilize heavy metals such as arsenic in polluted areas.

The new mineral bianchiniite, Ba₂(Ti⁴⁺V³⁺)(As₂O₅)₂OF, has been discovered in the Monte Arsiccio mine, Apuan Alps, Tuscany, Italy. It occurs as brown {001} tabular crystals, up to 1 mm across, with a vitreous lustre. It is brittle, with a perfect {001} cleavage. Streak is brownish. In reflected light, bianchiniite is grey, with orange–yellow internal reflections. It is weakly bireflectant, with a very weak anisotropy in shades of grey. Minimum and maximum reflectance data for COM wavelengths [R_min/R_max (%), (λ, nm)] are: 5.0/5.8 (470),5.7/6.5 (546), 5.7/7.0 (589) and 5.2/6.3 (650). Electron microprobe analyses gave (wt.% – average of 10 spot analyses): TiO₂ 10.34, V₂O₃3.77, Fe₂O₃ 3.76,As₂O₃ 44.36, Sb₂O₃ 0.22, SrO 0.45, BaO 34.79, PbO 0.28, F 1.77, sum 99.74, –O=F–0.75, total 98.99. On the basis of 12 anions per formula unit, the empirical formula of bianchiniite is (Ba_2.00Sr_0.04Pb_0.02)_Σ2.06(Ti⁴⁺_1.14V³⁺_0.44Fe³⁺_0.42)_Σ2.00[(As_3.96Sb_0.02)_Σ3.98O₁₀](O1.18F0.82)Σ2.00. Bianchiniite is tetragonal, space group I4/mcm, with unit-cell parameters a = 8.7266(4), c = 15.6777(7) Å, V = 1193.91(12) ų and Z = 8. Its crystal structure was refined from single-crystal X-ray diffraction data to R1 = 0.0134 on the basis of 555 unique reflections with Fo > 4σ(Fo)and 34 refined parameters. The crystal structure shows columns of corner-sharing [Ti/(V,Fe)]-centred octahedra running along c, connected along a and b through (As₂O₅) dimers. A {001} layer of Ba-centred [10+2]-coordinated polyhedra is intercalated between (As₂O₅) dimers. Bianchiniite has structural relations with fresnoite- and melilite-group minerals. The name honours the two mineral collectors Andrea Bianchini (b. 1959) and Mario Bianchini (b. 1962) for their contribution to the knowledge of the mineralogy of pyrite ± baryte ± iron-oxide ore deposits from the Apuan Alps.

A gem-quality purplish-red tourmaline sample of alleged liddicoatitic composition from the Anjanabonoina pegmatite, Madagascar, hasbeen fully characterised using a multi-analytical approach to define its crystal-chemical identity. Single-crystal X-ray diffraction, chem-ical and spectroscopic analysis resulted in the formula: ^X(Na_0.41^□_0.35Ca_0.24)_Σ1.00^Y(Al_1.81Li_1.00

Fe³⁺_0.04Mn³⁺_0.02Mn²⁺_0.12Ti_0.004)_Σ3.00

^ZAl₆[^T(Si_5.60B_0.40)Σ_6.00O₁₈]

(BO₃)₃(OH)₃W[(OH)_0.50F_0.13O_0.37]_Σ1.00, which corresponds to the tourmaline species elbaite having the typical space group R3m and relatively small unit-cell dimensions, a= 15.7935(4) Å, c= 7.0860(2) Å and V= 7.0860(2) ų.Optical absorption spectroscopy showed that the purplish-red colour is caused by minor amounts of Mn³⁺(Mn₂O₃= 0.20 wt.%).Thermal treatment in air up to 750°C strongly intensified the colour of the sample due to the oxidation of all Mn²⁺ to Mn³⁺ (Mn₂O₃ up to 1.21 wt.%). Based on infrared and Raman data, a crystal-chemical model regarding the electrostatic interaction betweenthe X cation and W anion, and involving the Y cations as well, is proposed to explain the absence or rarity of the mineral species ‘liddicoatite’.

Four specimens of the roméite-group minerals oxyplumboroméite and fluorcalcioroméite from the Långban Mn-Fe deposit in Central Sweden were structurally and chemically characterized by single-crystal X-ray diffraction, electron microprobe analysis and infrared spectroscopy. The data obtained and those on additional roméite samples from literature show that the main structural variations within the roméite group are related to variations in the content of Pb²⁺, which is incorporated into the roméite structure via the substitution Pb²⁺ → A²⁺ where A²⁺ = Ca, Mn and Sr. Additionally, the cation occupancy at the six-fold coordinated B site, which is associated with the heterovalent substitution ^BFe³⁺ + ^Y□ →^BSb⁵⁺ + ^YO^2–, can strongly affect structural parameters.

Chemical formulae of the roméite minerals group are discussed. According to crystal-chemical information, the species associated with the name ‘kenoplumboroméite’, hydroxycalcioroméite and fluorcalcioroméite most closely approximate end-member compositions Pb₂(SbFe³⁺)O₆□, Ca₂(Sb⁵⁺Ti)O₆(OH) and (CaNa)Sb₂O₆F, respectively. However, in accord with pyrochlore nomenclature rules, their names correspond to multiple end-members and are best described by the general formulae: (Pb,#)₂(Sb,#)₂O₆□, (Ca,#)₂(Sb,#)₂O₆(OH) and (Ca,#)Sb₂(O,#)₆F, where ‘#’ indicates an unspecified charge-balancing chemical substituent, including vacancies.

Mineral species should be identified by an end-member formula and by using the dominant-valency rule as recommended by the IMA–CNMNC. However, the dominant-end-member approach has also been used in the literature. These two approaches generally converge, but for some intermediate compositions, significant differences between the dominant-valency rule and the dominant end-member approach can be observed. As demonstrated for garnet-supergroup minerals, for example, the end-member approach is ambiguous, as end-member proportions strongly depend on the calculation sequence. For this reason, the IMA–CNMNC strongly recommends the use of the dominant-valency rule for mineral nomenclature, because it alone may lead to unambiguous mineral identification. Although the simple application of the dominant-valency rule is successful for the identification of many mineral compositions, sometimes it leads to unbalanced end-member formulae, due to the occurrence of a coupled heterovalent substitution at two sites along with a heterovalent substitution at a single site. In these cases, it may be useful to use the site-total-charge approach to identify the dominant root-charge arrangement on which to apply the dominant-constituent rule. The dominant-valency rule and the site-totalcharge approach may be considered two procedures complementary to each other for mineral identification. Their critical point is to find the most appropriate root-charge and atomic arrangements consistent with the overriding condition dictated by the end-member formula. These procedures were approved by the IMA−CNMNC in May 2019.

Tourmalines from the late-Variscan Arbus pluton (SWSardinia) and its metamorphic aureole were structurally and chemically characterized by single-crystal X-ray diffraction, electron and nuclear microprobe analysis, Mössbauer, infrared and optical absorption spectroscopy, to elucidate their origin and relationships with the magmatic evolution during the pluton cooling stages. The Arbus pluton represents a peculiar shallow magmatic system, characterized by sekaninaite (Fe-cordierite)-bearing peraluminous granitoids, linked via AFC processes to gabbroic mantle-derived magmas. The Fe²⁺-Al-dominant tourmalines occur in: a) pegmatitic layers and pods, as prismatic crystals; b) greisenized rocks and spotted granophyric dikes, as clots or nests of fine-grained crystals in small miaroles locally forming orbicules; c) pegmatitic veins and pods close to the contacts within the metamorphic aureole. Structural formulae indicate that tourmaline in pegmatitic layers is schorl, whereas in greisenized rocks it ranges fromschorl to fluor-schorl. Tourmalines in thermometamorphosed contact aureole are schorl, foitite and Mg-rich oxy-schorl. The main substitution is Na+Fe²⁺↔▢+Al, which relates schorl to foitite. The homovalent substitution (OH)↔F at the O1 crystallographic site relates schorl to fluor-schorl, while the heterovalent substitution Fe²⁺+(OH, F)↔Al+O relates schorl/fluor-schorl to oxy-schorl. Tourmaline crystallization in the Arbus pluton was promoted by volatile (B, F and H₂O) enrichment, low oxygen fugacity and Fe²⁺ activity. The mineralogical evolutive trend is driven by decreasing temperature, as follows: sekaninaite+quartz →schorl+quartz→fluor-schorl+quartz → foitite+quartz. The schorl→foitite evolution represents a distinct trend towards (Al+!) increase and unit-cell volumedecrease. These trends are typical of granitic magmas and consistent with Li-poor granitic melts, as supported by the absence of elbaite and other Li-minerals in the Arbus pluton. Tourmaline-bearing rocks reflect the petrogenetic signi!cance of contribution from a metapelitic crustal component during the evolution of magmas in the middle-upper crust.

Eight spinel single-crystal samples belonging to the spinel sensu stricto-magnesiocoulsonite series (MgAl₂O₄-MgV₂O₄) were synthesized and crystal-chemically characterized by X‑ray diffraction, electron microprobe and optical absorption spectroscopy. Site populations show that the tetrahedrally coordinated site (T) is populated by Mg and minor Al for the spinel sensu stricto compositions, and only by Mg for the magnesiocoulsonite compositions, while the octahedrally coordinated site (M) is populated by Al, V³⁺, minor Mg, and very minor amounts of V⁴⁺. The latter occurs in appreciable amounts in the Al-free magnesium vanadate spinel, T(Mg)M(Mg_0.26V³⁺_1.48V⁴⁺_0.26)O₄, showing the presence of the inverse spinel VMg₂O₄. The studied samples are characterized by substitution of Al³⁺ for V³⁺ and (Mg²⁺+V⁴⁺) for 2V³⁺ described in the system MgAl₂O₄-MgV₂O₄-VMg₂O₄.

The present data in conjunction with data from the literature provide a basis for quantitative analyses of two solid-solution series MgAl₂O₄-MgV₂³⁺O₄ and MgV₂³⁺O₄-V⁴⁺Mg₂O₄. Unit-cell parameter increases with increasing V³⁺ along the series MgAl₂O₄-MgV₂O₄ (8.085–8.432 Å), but only slightly increases with increasing V³⁺ along the series VMg₂O₄-MgV₂O₄ (8.386–8.432 Å). Although a solid solution could be expected between the MgAl₂O₄ and VMg₂O₄ end-members, no evidence was found. Amounts of V⁴⁺ are nearly insignificant in all synthetic Al-bearing vanadate spinels, but are appreciable in Al-free vanadate spinel.

An interesting observation of the present study is that despite the observed complete solid-solution along the MgAl₂O₄-MgV₂O₄ and MgV₂O₄-VMg₂O₄ series, the spinel structure seems to be unable to stabilize V⁴⁺ in any intermediate members on the MgAl₂O₄-Mg₂VO₄ join even at high oxygen fugacities. This behavior indicates that the accommodation of specific V-valences can be strongly influenced by crystal-structural constraints, and any evaluation of oxygen fugacities during mineral formation based exclusively on V cation valence distributions in spinel should be treated with caution. The present study underlines that the V valency distribution in spinels is not exclusively reflecting oxygen fugacities, but also depends on activities and solubilities of all chemical components in the crystallization environment.

Oxy-foitite, □(Fe²⁺Al₂)Al₆(Si₆O₁₈)(BO₃)₃(OH)₃O, is a new mineral of the tourmaline supergroup. It occurs in high-grade migmatitic gneisses of pelitic composition at the Cooma metamorphic Complex (New South Wales, Australia), in association with muscovite, K-feldspar and quartz. Crystals are black with a vitreous luster, sub-conchoidal fracture and gray streak. Oxy-foitite has a Mohs hardness of ∼7, and has a calculated density of 3.143 g/cm³. In plane-polarized light, oxy-foitite is pleochroic (O= dark brown and E = pale brown), uniaxial negative. Oxy-foitite belongs to the trigonal crystal system, space group R3m, a = 15.9387(3) Å, c = 7.1507(1)Å and V = 1573.20(6)ų,Z = 3. The crystal structure of oxy-foitite was refined to R1 = 1.48% using 3247 unique reflections from single-crystal X-ray diffraction using MoKα radiation. Crystal-chemical analysis resulted in the empirical structural formula: ^X(□_0.53Na_0.45Ca_0.01K_0.01)_Σ1.00^Y(Al_1.53Fe²⁺_1.16Mg_0.22Mn²⁺_0.05Zn_0.01Ti⁴⁺_0.03)_Σ3.00^Z(Al_5.47Fe³⁺_0.14Mg_0.39)_Σ6.00[(Si_5.89Al_0.11)_Σ6.00O₁₈](BO₃)₃^V(OH)₃^W[O_0.57F_0.04(OH)_0.39]_Σ1.00. Oxy-foitite belongs to the X-site vacant group of the tourmaline-supergroup minerals, and shows chemical relationships with foitite through the substitution ^YAl³⁺+^WO^2-→^YFe²⁺+^W(OH)^1–.

Iron-bearing oxy-dravite was thermally treated in air and hydrogen atmosphere at 800 °C to study potential changes in Fe, Mg and Al ordering over the octahedrally coordinated Y and Z sites and to explore possible applications to intersite geothermometry based on tourmaline. Overall, the experimental data (structural refinement, Mössbauer, infrared and optical absorption spectroscopy) show that heating Fe-bearing tourmalines results in disordering of Fe over Y and Z balanced by ordering of Mg at Y, whereas Al does not change appreciably. The Fe disorder depends on temperature, but less on redox conditions. The degree of Fe³⁺–Fe²⁺ reduction is limited despite strongly reducing conditions, indicating that the f_O2 conditions do not exclusively control the Fe oxidation state at the present experimental conditions. Untreated and treated samples have similar short- and long-range crystal structures, which are explained by stable Al-extended clusters around the O1 and O3 sites. In contrast to the stable Al clusters that preclude any temperature-dependent Mg–Al order– disorder, there occurs Mg diffusion linked to temperaturedependent exchange with Fe. Ferric iron mainly resides around O²⁻ at O1 rather than (OH)⁻, but its intersite disorder induced by thermal treatment indicates that Fe redistribution is the driving force for Mg–Fe exchange and that its diffusion rates are significant at these temperatures. With increasing temperature, Fe progressively disorders over Y and Z, whereas Mg orders at Y according to the order–disorder reaction: ^YFe + ^ZMg → ^ZFe + ^YMg. The presented findings are important for interpretation of the post-crystallization history of both tourmaline and tourmaline host rocks and imply that successful tourmaline geothermometers may be developed by thermal calibration of the Mg– Fe order–disorder reaction, whereas any thermometers based on Mg–Al disorder will be insensitive and involve large uncertainties.

Natural Mg-rich lucchesiite was thermally treated in air and hydrogen atmosphere up to 800 °C to study potential changes in Fe-, Mg- and Al ordering over the octahedrally coordinated Y-  and Z -sites, and to explore possible applications to intracrystalline geothermometry based on tourmaline. Overall, the experimental data (structural refinement, Mössbauer, infrared and optical absorption spectroscopy) show that thermal treatment of lucchesiite results in an increase of Fe_tot contents at Z balanced by an increase of Mg and Al at Y . This process is accompanied by a significant deprotonation of the O3 anion site. The Fe order–disorder reaction depends more on temperature, than on redox conditions. During heat treatment in H₂ ,reduction of Fe³⁺ to Fe²⁺ was not observed despite strongly reducing conditions, indicating that the f_O2  conditions do not exclusively control the Fe oxidation state at the present experimental conditions. On the basis of this and previous studies, the intersite order–disorder process induced by thermal treatment indicates that Fe redistribution is an important factor for Fe–Mg–Al-exchange and is significant at temperatures around 800 °C. As a result, Fe–Mg–Al intersite order–disorder is sensitive to temperature variations, whereas geothermometers based solely on Mg–Al order–disorder appear insensitive and involve large uncertainties. The presented findings are important for interpretation of the post-crystallization history of both tourmaline and tourmaline host rocks, and indicate that successful tourmaline geothermometers may be developed by thermal calibration of the Fe-Mg–Al order–disorder reaction.

Measurement of oxygen isotope ratios in common silicate minerals such as olivine, pyroxene, feldspar, garnet, and quartz is increasingly performed by Secondary Ion Mass Spectrometry (SIMS). However, certain mineral groups exhibit solid solution series, and the large compositional spectrum of these mineral phases will result in matrix effects during SIMS analysis. These matrix effects must be corrected through repeated analysis of compositionally similar standards to ensure accurate results. In order to widen the current applicability of SIMS to solid solution mineral groups in common igneous rocks, we performed SIMS homogeneity tests on new augite (NRM-AG-1) and enstatite (NRM-EN-2) reference materials sourced from Stromboli, Italy and Webster, North Carolina, respectively. Aliquots of the standard minerals were analysed by laser fluorination (LF) to establish their δ¹⁸O values. Repeated SIMS measurements were then performed on randomly oriented fragments of the same pyroxene crystals, which yielded a range in δ¹⁸O less than ± 0.42 and ± 0.58‰ (2σ) for NRM-AG-1 and NRM-EN-2, respectively. Homogeneity tests verified that NRM-AG-1 and NRM-EN-2 do not show any crystallographic orientation bias and that they are sufficiently homogeneous on the 20 μm scale to be used as routine mineral standards for SIMS δ¹⁸O analysis. We subsequently tested our new standard materials on recently erupted pyroxene crystals from Merapi volcano, Indonesia. The δ¹⁸O values for Merapi pyroxene obtained by SIMS (n = 204) agree within error with the LF-derived δ¹⁸O values for Merapi pyroxene but differ from bulk mineral and whole-rock data obtained by conventional fluorination. The bulk samples are offset to higher δ¹⁸O values as a result of incorporation of mineral and glass inclusions that in part reflects crustal contamination processes. The Merapi pyroxene SIMS data, in turn, display a frequency peak at 5.8‰, which allows us to estimate the δ¹⁸O value of the primary mafic magma at Merapi to ~ 6.1‰ when assuming closed system differentiation.