Barium-rich silicates were recently found in the lower Cambrian black shale sequence of Anhui Province, South China. The Ba silicates can be divided into Ba-poor K-feldspar, Ba-rich K-feldspar and cymrite, on the basis of occurrence and composition. The Ba-poor K-feldspar have low BaO contents (<1.00 wt.%) and occur as sporadic grains in mudstone. In contrast, the Ba-rich K-feldspar have higher BaO contents (1.36–20.51 wt.%) and occur mainly as rims around grains of Ba-poor K-feldspar in mudstone, as void fillings in dolostone, and as grains dispersed in chert. The cymrite is observed only in chert and occurs as dispersed euhedral tabular crystals. Analyses of the Ba-poor K-feldspar and Ba-rich K-feldspar provide compositions that total <100 wt.% and show non-stoichiometry characterized by (Na + K + Ba + Ca)apfu less than 1.00 and (Al + Si)apfu slightly higher than 4.00. Raman and infrared (IR) spectroscopy has revealed the presence of ${\rm NH}_{\rm 4}^{\rm +} $ and H2O in the Ba-rich K-feldspar, which accounts for the unusual composition of these feldspars. We suggest that the formation processes of the Ba-poor K-feldspar, Ba-rich K-feldspar and cymrite were closely associated with multi-stage hydrothermal fluids enriched in Ba and NH4, consistent with an extensional tectonic setting during the early Cambrian. The Ba-poor K-feldspars might have formed through interactions between pre-existing K-feldspars and hydrothermal fluids. The Ba-rich K-feldspars in mudstone and dolostone formed mainly by precipitation from hydrothermal fluids that infiltrated the sediments, and we suggest the Ba-rich K-feldspars and cymrite in the chert are of exhalative hydrothermal origin.

Overgrowths of whiskers of hedenbergite (Ca(Fe2+,Mg)Si2O6) on the hydrous pyroxenoid babingtonite (Ca2Fe2+Fe3+[Si5O14(OH)]) have been observed at Arvigo in Switzerland and Kreimbach/Kaulbach in Germany, and we have studied them with transmission electron microscopy in order to understand their structural relationships and formation. The boundaries between babingtonite and hedenbergite are sharply defined, and the two minerals are in direct contact with no additional phases present. The relationships of babingtonite (Bab) and hedenbergite (Hd) were determined as Bab[100]//Hd[112] in the Arvigo specimen and Bab[$\bar 1$00]//Hd[1$\bar 1$2] in the Kreimbach/Kaulbach specimen. Diffraction derived from Bab(031) and Hd(02$\bar 1$) in the Arvigo samples and Bab(031) and Hd(021) in the Kreimbach/Kaulbach samples were observed in identical positions. The reciprocity between the babingtonite and hedenbergite structures is governed by the direction of the SiO4-tetrahedral chains, and the related configuration of octahedra. Thus, hedenbergite is apparently an epitaxial phase grown on a base of {010} plates of babingtonite. The defined orientation relationship is also consistent with that shown in topotaxial intergrowths of other clinopyroxenes and pyroxenoids. The topotaxial intergrowths may result from diffusion-controlled solid-state reactions, whereas rapid whisker growth is characteristic of supersaturated solutions or a vapour medium. The epitaxial growth of hedenbergite whiskers on babingtonite with an abrupt but coherent change of structure at the interface represents an ideal example where the similar chemical compositions of host and guest contribute strongly to the close structural relationship.

The mineralogy of nine recently discovered dykes (VL1:VL8 and VL10) in the vicinity of Vattikod village, Nalgonda district in Telangana State is described. The mineral assemblage present and their compositions are comparable to those of bona fide lamproites in terms of the presence of phlogopite (Ti-rich, Al-poor phlogopite and tetraferriphlogopite); amphiboles (potassic-arfvedsonite, potassic-richterite, potassic-ferro-richterite, potassic-katophorite, Ti-rich potassic-katophorite, Ti-rich potassic-magnesio-katophorite); Al-poor clinopyroxenes; feldspars (K-feldspar, Ba-K-feldspar and Na-feldspar) and spinels (chromite-magnetite and qandilite-ulvöspinel-franklinite). These dykes have undergone diverse and significant degrees of deuteric alteration as shown by the formation of secondary phases such as: titanite, allanite, hydro-zircon, calcite, chlorite, quartz and cryptocrystalline SiO2. On the basis of their respective mineralogy: the VL4 and VL5 dykes are classified as pseudoleucite-phlogopite lamproite; VL2 and VL3 dykes as pseudoleucite-amphibole-lamproite; and VL6, VL7 and VL8 as pseudoleucite-phlogopite-amphibole-lamproite. VL10 is extensively altered but contains fresh euhedral apatite microphenocrysts together with pseudomorphs after leucite and is classified as a pseudoleucite-apatite-(phlogopite?) lamproite. The mineralogy of the Vattikod lamproite dykes is compared with that of the Ramadugu, Somavarigudem and Yacharam lamproite dykes which also occur in the Ramadugu lamproite field. The lamproites from the Eastern Dharwar Craton are considered as being possible expressions of ancient subduction-related alkaline magmatism along the Eastern Ghats mobile belt.

Concentration data are reported for 18 trace elements in chalcopyrite from a suite of 53 samples from 15 different ore deposits obtained by laser-ablation inductively-coupled plasma-mass spectrometry. Chalcopyrite is demonstrated to host a wide range of trace elements including Mn, Co, Zn, Ga, Se, Ag, Cd, In, Sn, Sb, Hg, Tl, Pb and Bi. The concentration of some of these elements can be high (hundreds to thousands of ppm) but most are typically tens to hundreds of ppm. The ability of chalcopyrite to host trace elements generally increases in the absence of other co-crystallizing sulfides. In deposits in which the sulfide assemblage recrystallized during syn-metamorphic deformation, the concentrations of Sn and Ga in chalcopyrite will generally increase in the presence of co-recrystallizing sphalerite and/or galena, suggesting that chalcopyrite is the preferred host at higher temperatures and/or pressures. Trace-element concentrations in chalcopyrite typically show little variation at the sample scale, yet there is potential for significant variation between samples from any individual deposit. The Zn:Cd ratio in chalcopyrite shows some evidence of a systematic variation across the dataset, which depends, at least in part, on temperature of crystallization. Under constant physiochemical conditions the Cd:Zn ratios in co-crystallizing chalcopyrite and sphalerite are typically approximately equal. Any distinct difference in the Cd:Zn ratios in the two minerals, and/or a non-constant Cd:Zn ratio in chalcopyrite, may be an indication of varying physiochemical conditions during crystallization.

Chalcopyrite is generally a poor host for most elements considered harmful or unwanted in the smelting of Cu, suggesting it is rarely a significant contributor to the overall content of such elements in copper concentrates. The exceptions are Se and Hg which may be sufficiently enriched in chalcopyrite to exceed statutory limits and thus incur monetary penalties from a smelter.

Synthesis and solubility studies of brizziite, NaSbO3, have been undertaken to determine the possible role of this rare secondary phase in the immobilization of Sb under supergene conditions and the conditions responsible for its formation in the supergene zone. Solubility studies were undertaken at T = 298.15 K. A value of ΔGfө) (NaSbO3, s, 298.15 K) = –806.66 ± 1.4 kJ mol–1 was derived. Calculations involving tripuhyite, FeSbO4, byströmite, MgSb2O6, ordoñezite, ZnSb2O6 and rosiaite, PbSb2O6, show that brizziite is a thermodynamically stable phase only at negligible activities of Pb2+(aq) at high pH and high salinity. Calculations involving mopungite Na[Sb(OH)6] combined with reported mineral associations suggest that mopungite is the thermodynamically unstable precursor to brizziite and its presence in natural settings must be due to kinetic stability. This explains why brizziite is such a rare secondary phase and therefore why it cannot exert any significant influence on the dispersion of Sb in the supergene environment.

Solid phase reaction synthesis of wollastonite-2M by a natural rock precursor as the source of amorphous silica and CaCO3 is reported. Chemical treatments were carried out on a diatomitic rock from Crotone (Calabria, Italy) in order to measure its reactive silica and CaCO3 contents. Four series of synthesis were performed at 1000°C at ambient pressure by mixing, at different stoichiometry, the diatomitic rock with a natural limestone as a source of additive CaCO3, and sodium carbonate (Na2CO3) as triggering agent.

Wollastonite-2M was characterized by chemo-physical, crystallographical and morphological-microtextural analyses. All these characterizations, together with infrared and nuclear magnetic resonance (29Si) responses provide values comparable to literature data. Estimation of the amorphous phase in the synthesis powders was performed through quantitative phase analysis using the combined Rietveld and reference intensity ratio methods, resulting in a final product of 96.3% wollastonite-2M.

Melcherite (IMA2015-018), ideally Ba2Na2Mg[Nb6O19]·6H2O, occurs as a vug mineral in the carbonatite of the Jacupiranga mine, Cajati county, São Paulo state, Brazil, associated with dolomite, calcite, magnetite, pyrrhotite, tochilinite, ‘pyrochlore’ and fluorapatite. This is also the type locality for zirkelite, quintinite, menezesite and pauloabibite. The mineral forms irregular, tabular crystals up to 200 µm in maximum dimension. Melcherite is transparent and displays a vitreous lustre; it is beige with a white streak. It is non-fluorescent. The mineral displays perfect cleavage on {001}. Chemical composition varies from Ba2Na2Mg[Nb6O19].6H2O to (BaK)(NaCa)Mg[Nb6O19].6H2O. Empirical formulae for the first and the second compositions are: (Ba1.75K0.19)Σ1.94(Na1.80Ca0.19)Σ1.99(Mg0.96Mn0.02Al0.02)Σ1.00Nb6.02O19.00·6H2O and (Ba0.99K1.00)Σ1.99(Na1.02Ca0.96)Σ1.98(Mg0.95Mn0.05)Σ1.00Nb6.02O19.00·6H2O, respectively. Data for a single crystal with the second composition are: trigonal, R$\bar 3$, a = 9.0117(6) Å, c = 23.3986(16) Å, V = 1645.64(19) Å3 and Z = 3. Calculated density for this formula is 3.733 g/cm3, and the calculated mean refractive index is 1.924. Melcherite is a hexaniobate that has structural layers parallel to the xy plane that stack along the c axis with simultaneous 1/3 [110] displacement so as to produce an R lattice. The melcherite structure is built by layers of [(Ba,K)(O,H2O)9] polyhedra and the [Nb6O19]8− super-octahedron (Lindqvist anion) interconnected by [(Na,Ca)O6] polyhedra. Cations of Mg2+ are bonded to six water molecules each and are not associated with Lindqvist oxygen ions. The mineral is named in honour of Geraldo Conrado Melcher (1924–2011), a pioneer in Jacupiranga carbonatite studies.

A new mineral species, fluorlamprophyllite (IMA2013-102), ideally Na3(SrNa)Ti3(Si2O7)2O2F2, has been found in the Poços de Caldas alkaline massif, Morro do Serrote, Minas Gerais, Brazil. Alternatively, the idealized chemical formula could be written as (SrNa)[(Na3Ti)F2][Ti2(Si2O7)2O2], setting the large interlayer cations before the cations of the layer. Fluorlamprophyllite is the F-analogue of lamprophyllite. It is associated with aegirine, analcime, natrolite, nepheline and microcline. Fluorlamprophyllite crystals are brownish-orange and bladed. The mineral is transparent with a pale yellow streak and an adamantine lustre. It is brittle and has a Mohs hardness of ~3; cleavage is perfect on {100} and no parting was observed. The calculated density is 3.484 g/cm3. Optically, fluorlamprophyllite is biaxial (+), with α = 1.735(7), β = 1.749(7) and γ = 1.775(9) and 2Vmeas = 72(3)°. An electron microprobe analysis produced an average composition (wt.%) (9 points) of Na2O 10.63(30), K2O 0.47(3), SiO2 30.51(13), SrO 18.30(24), MgO 0.81(17), Al2O3 0.23(2), CaO 1.11(7), MnO 5.03(38), TiO2 27.41(87), Fe2O3 2.45(37), F 2.86(23), plus H2O 1.00 (added to bring the total close to 100%), –O = F –1.20, with the total = 98.61%. The elements Nb and Ba were sought, but contents were below microprobe detection limits. The resultant chemical formula was calculated on the basis of 18 (O + F) atoms per formula unit. The addition of 1.00 wt.% H2O brought [F+(OH)] = 2 pfu, yielding (Na2.63Sr1.35Mn0.54Ca0.15Mg0.15K0.08)Σ4.90(Ti2.63Fe0.24Al0.04)Σ2.91Si3.89O16[F1.15(OH)0.85]Σ2.00. The mineral is monoclinic, with space group C2/m and unit-cell parameters a = 19.255(2), b = 7.0715(7), c = 5.3807(6) Å, β = 96.794(2)° and V = 727.5(1) Å3. The structure is a layered silicate inasmuch as the O atoms are arranged in well-defined, though not necessarily close-packed layers.

Parisite-(La) (IMA2016-031), ideally CaLa2(CO3)3F2, occurs in a hydrothermal vein crosscutting a metarhyolite of the Rio dos Remédios Group, at the Mula mine, Tapera village, Novo Horizonte county, Bahia, Brazil, associated with hematite, rutile, almeidaite, fluocerite-(Ce), brockite, monazite-(La), rhabdophane-(La) and bastnäsite-(La). Parisite-(La) occurs as residual nuclei (up to 5 mm) in steep doubly-terminated pseudo-hexagonal pyramidal crystals (up to 8.2 cm). Parisite-(La) is transparent, yellow-green to white, with a white streak and displays a vitreous (when yellow-green) to dull (when white) lustre. Cleavage is distinct on pseudo-{001}; fracture is laminated, conchoidal, or uneven. The Mohs hardness is 4 to 5, and it is brittle. Calculated density is 4.273 g cm−3. Parisite-(La) is pseudo-uniaxial (+), ω = 1.670(2) and ε = 1.782(5) (589 nm). The empirical formula normalized on the basis of 11 (O + F) atoms per formula unit (apfu) is Ca0.98(La0.83Nd0.51Ce0.37Pr0.16Sm0.04Y0.03)Σ1.94C3.03O8.91F2.09. The IR spectrum confirms the absence of OH groups. Single-crystal X-ray studies gave the following results: monoclinic (pseudo-trigonal), space group: C2, Cm, or C2/m, a = 12.356(1) Å, b = 7.1368(7) Å, c = 28.299(3) Å, β = 98.342(4)°, V = 2469.1(4) Å3 and Z = 12. Parisite-(La) is the La-dominant analogue of parisite-(Ce).

During systematic characterization of amphiboles that still lack a complete mineral description, fluoro-tremolite was identified in a specimen from the Limecrest-Southdown quarry, Sparta, New Jersey, USA, which was provided by the Franklin Mineral Museum. The ideal formula of fluoro-tremolite is A□ BCa2CMg5TSi8O22WF2 and the empirical formula derived for the holotype specimen, based on the results of electron-microprobe analysis and single-crystal structure refinement, is A(Na0.28K0.02)Σ0.30B(Ca1.99Na0.01)Σ2.00C(Mg4.70${\rm Fe}_{{\rm 0}{\rm. 28}}^{{\rm 2 +}} $Zn0.01${\rm Ti}_{{\rm 0}{\rm. 01}}^{{\rm 4 +}} $)Σ5.00T(Si7.68Al0.32)Σ8.00O22W(F1.16OH0.84)Σ2.00. The unit-cell dimensions in space group C2/m are a = 9.846(2), b = 18.050(3), c = 5.2769(14) Å, β = 104.80(2)° and V = 906.7 (3) Å3 and Z = 2; the a:b:c ratio is 0.545:1:0.292. Fluoro-tremolite is biaxial (+), with α = 1.5987(5), β = 1.6102(5), γ = 1.6257(5), 2V(meas.) = 85(1)o and 2V(calc.) = 82o. The strongest ten reflections in the powder X-ray pattern [d values (in Å), I, (hkl)] are: 2.706, 100, (151); 3.126, 67, (310); 2.531, 59, ($\bar 2$02); 3.381, 57, (131); 2.940, 43, ($\bar 1$51, 221); 3.276, 37, (240); 2.337, 36, ($\bar 3$51); 2.592, 35, (061); 2.731, 34, ($\bar 3$31); 2.163, 34, (261). Both the mineral and the mineral name have been approved by the Commission on New Minerals, Nomenclature and Classification of the International Mineralogical Association (IMA2016–018); the holotype has been deposited at the Franklin Mineral Museum (32 Evans Street, Franklin, 07416 New Jersey, US), under the catalogue number 7710.

Comparison with new data on tremolite and synthetic fluoro-tremolite provides a more sound crystal-chemical model of the end-member compositions and their solid-solution.

Thérèsemagnanite was originally described from the Cap Garonne mine, Var, France. Its ideal formula was reported as (Co,Zn,Ni)6(SO4)(OH,Cl)10·8H2O; without crystal structure data, only the powder X-ray diffraction pattern was given. Revision of the holotype material revealed that thérèsemagnanite is identical to ‘cobaltogordaite’ (IMA2014-043), recently described from the Blue Lizard mine, Utah, USA. Thérèsemagnanite is thus redefined in accordance with the new data obtained for the neotype specimen from Blue Lizard (formerly the holotype specimen of ‘cobaltogordaite’) and ‘cobaltogordaite’ has been discredited by the International Mineralogical Association Commission on New Mineral Nomenclature and Classification (IMA CNMNC). Thérèsemagnanite has the ideal, end-member formula NaCo4(SO4)(OH)6Cl·6H2O. The empirical formulae of the holotype (Cap Garonne) and the neotype (Blue Lizard), both based on microprobe analyses and calculated on the basis of 17 O + Cl atoms per formula unit (with fixed 6 OH groups and 6 H2O molecules; H content is calculated by stoichiometry) are (Na0.64K0.09)Σ0.73(Co2.35Zn1.22Ni0.50)Σ4.07S1.02O3.98(OH)6Cl1.02·6H2O and Na1.01(Co1.90Zn1.37Ni0.48Cu0.15Mn0.05)Σ3.95S1.03O4.09(OH)6Cl0.91·6H2O, respectively. Thérèsemagnanite is trigonal, P$\overline 3 $, a = 8.349(3), c = 13.031(2) Å, V = 786.6(4) Å3 and Z = 2 (neotype). The strongest powder X-ray diffraction lines are [dobs in Å (hkl) (Irel)]: 13.10 (001)(100), 6.53 (002)(8), 4.173 (110)(4), 3.517 (112)(5), 2.975 (104, 10$\overline 4 $)(4), 2.676 (211)(5) and 2.520 (12$\bar 2$)(5) (neotype). Thérèsemagnanite is a cobalt analogue of gordaite, NaZn4(SO4)(OH)6Cl·6H2O. These minerals represent the gordaite group, accepted by the IMA CNMNC.

Evidence of crystallization by particle attachment in synthetic materials is described in numerous contributions. However, efforts to establish the contribution of the particle attachment mechanism to inorganic crystallization in natural environments have barely begun. Here, we show, for the first time, evidence that confirms oriented particle attachment as a crystal growth mechanism that is relevant in sedimentary environments. In these natural settings, oriented particle attachment operates during the formation of highly anisotropically structured clay minerals, which constitute one of the most extensively distributed groups of minerals in the Earth's crust. High-resolution transmission electron microscopy images show that the clay minerals aggregation process occurs in different manners. Smectites aggregate by semi-oriented attachment, while kaolinite, sepiolite and palygorskite aggregate by oriented attachment.

The crystal structure of ceruleite, CuAl4[AsO4]2(OH)8(H2O)4, has been solved to an R1 of 0.0307, using the world's largest crystals from the Cap Garonne mine, France. Ceruleite crystallizes in space group P21/n, with the unit cell a = 7.2000(14), b = 11.345(2), c = 9.856(2) Å, β = 105.57(3)°, V = 775.6(3) Å3 and Z = 1. Ceruleite has a unique structure that consists of Al(O,OH)6 octahedra that are sharing edges to form rhombus-shaped tetramers. AsO4 tetrahedra share two corners with one such rhombus and the other two corners with each of two other rhombi, linking them into a very open mesoporous framework. Cu(OH)2(H2O)2 squares lie in the channels and link Al4 rhombi along || b. H2O molecules are also located in the channels.

Clino-suenoite, ideally □${\rm Mn}_{2}^{2 +} $Mg5Si8O22(OH)2 is a new amphibole of the magnesium-iron-manganese subgroup of the amphibole supergroup. The type specimen was found at the Lower Scerscen Glacier, Valmalenco, Sondrio, Italy, where it occurs in Mn-rich quartzite erratics containing braunite, rhodonite, spessartine, carbonates and various accessory minerals. The empirical formula derived from electron microprobe analysis and single-crystal structure refinement is: ANa0.04B(${\rm Mn}_{1.58}^{2 +} $Ca0.26Na0.16)Σ2.00C(Mg4.21${\rm Mn}_{0. 61}^{2 +} {\rm Fe}_{0.04}^{2 +} $Zn0.01Ni0.01${\rm Fe}_{0.08}^{3 +} $Al0.04)Σ5.00TSi8.00O22W[(OH1.94F0.06)]Σ=2.00. Clino-suenoite is biaxial (+), with α = 1.632(2), β = 1.644(2), γ = 1.664(2) and 2Vmeas. = 78(2)° and 2Vcalc. = 76.3°. The unit-cell parameters in the C2/m space group are a = 9.6128(11), b = 18.073(2), c = 5.3073(6) Å, β = 102.825(2)° and V = 899.1(2) Å3 with Z = 2. The strongest ten reflections in the powder X-ray diffraction pattern [d (in Å), I, (hkl)] are: 2.728, 100, (151); 2.513, 77, ($\bar 2$02); 3.079, 62, (310); 8.321, 60, (110); 3.421, 54, (131); 2.603, 42, (061); 2.175, 42, (261); 3.253, 41, (240); 2.969, 40, (221); 9.036, 40, (020).

The new mineral bytízite was found in the dump of shaft No. 16, one of the mines in the Příbram uranium and base-metal district, central Bohemia, Czech Republic. Bytízite is associated with chaméanite, příbramite, giraudite, berzelianite, umangite, eskebornite, hakite, tetrahedrite, bukovite, crookesite and uraninite in a calcite-dominant gangue. The new mineral occurs as anhedral grains up to 40 µm, growing together in aggregates up to 300 µm across. Bytízite is steel-grey in colour and has a metallic lustre. Mohs hardness is ca. 2–3; the calculated density is 6.324 g cm–3. In reflected light bytízite is grey with a yellowish hue, yellowish and brownish. Bireflectance and pleochroism are weak. Anisotropy is strong with grey to brownish rotation tints. Internal reflections were not observed. The empirical formula, based on electron-microprobe analyses, is (Cu3.00Fe0.01Ag0.01)3.02(Sb0.97As0.06)1.03Se2.94. The ideal formula is Cu3SbSe3, which requires Cu 34.71, Sb 22.16 and Se 43.13, total 100.00 wt.%. Bytízite is orthorhombic, Pnma, a = 7.9594(12), b = 10.5830(14), c = 6.8240(11) Å, with V = 574.82(15) Å3 and Z = 4. The strongest reflections of the calculated powder X-ray diffraction pattern [d, Å (I)(hkl)] are: 3.73(37)(210), 3.27(62)(211), 2.867(40)(022), 2.698(100)(122) and 2.646(37)(040). According to the single-crystal X-ray diffraction data (Robs = 0.0437), bytízite is isostructural with synthetic Cu3SbSe3. The structure of bytízite contains two Cu, one Sb, and two Se sites (the latter is occupied both by Se and S atoms). In the structure of both synthetic Cu3SbSe3 and bytízite, there are groups of three cis-edge-sharing tetrahedra [Cu3Se8], which are interlinked to a 3D framework by SbSe3 groups. Bytízite is named after its type locality, the Bytíz deposit, near the village Bytíz.

A P21/c synthetic (Ca0.2Co0.8)CoSi2O6 pyroxene was synthesized by slow cooling from melt at high pressure. Single crystals suitable for X-ray diffraction were obtained and refined. The results were compared to those of C2/c pyroxenes along the series CaCoSi2O6–Co2Si2O6. Strong similarities in the crystal chemical mechanism of the transition with the synthetic CaFeSi2O6–Fe2Si2O6 and CaMgSi2O6–Mg2Si2O6 pyroxenes, both at an average and local level are apparent.

The results, examined together with two new refinements of pigeonite in the ureilites ALHA77257 and RKPA80239 and with a set of natural and synthetic C2/c and P21/c pyroxenes, show that the average cation radius in the M2 site is the driving force for the phase transition from C2/c to P21/c. The longest M2–O3 distances and the O3–O3–O3 angles follow the same trend, dictated only by the ionic radius in M2, in either synthetic or natural pyroxenes, regardless of the ionic radius of the M1 cations. The transition also affects the difference between bridging and non-bridging oxygen atoms and the extent of tetrahedral deformation, whereas the M1–O, M2–O1 and M2–O2 distances are unaffected by the transition and are determined only by the ionic radius of the bonding cation. The structural changes between the ionic radius and the high temperature C2/c and P21/c transitions are similar, and different to the high-pressure transition.

Analysis of natural and synthetic pyroxenes shows that the transition with composition occurs in strain free pyroxenes for a critical radius of 0.85 Å. Increasing strain stabilizes the P21/c structure to a higher temperature and larger cation radius.

Finally, our results show that the monoclinic P21/c Ca-poor clinopyroxene, i.e the mineral pigeonite, crystallizes only at conditions where the structure is HT-C2/c, and changes to the P21/c symmetry during cooling.