Occurrence and mineral association

The new mineral occurs at No. 3 orebody (also known as Matsumaezawa Pit) of the Tanohata Mine, Shimohei District, Iwate Prefecture, Japan (39°55''N, 141°54''E). This manganese orebody is located in the upper-Jurassic accretionary complex chert and underwent contact metamorphism by the intrusion of granodiorite (Nambu et al., Reference Nambu, Tanida and Kitamura1969a). The orebody has a metasomatic origin and zoned structure. The dominant zones are composed mainly of braunite, rhodonite, quartz, natronambulite and serandite, while a thin amphibole-rich zone is wedged between them (Nishio-Hamane et al., Reference Nishio-Hamane, Minakawa and Okada2014). A more detailed geological description of the Tanohata Mine was reported by Nambu et al. (Reference Nambu, Tanida and Kumagai1969b). Tanohata Mine is the type locality for mangano-ferri-eckermannite (first described as ‘kôzulite’ by Nambu et al., Reference Nambu, Tanida and Kitamura1969a: see Discussion section), natronambulite (Matsubara et al., Reference Matsubara, Kato and Tiba1985), potassic-ferri-leakeite (first described as ‘potassicleakeite’ by Matsubara et al., Reference Matsubara, Miyawaki, Kurosawa and Suzuki2002), watatsumiite (Matsubara et al., Reference Matsubara, Miyawaki, Kurosawa and Suzuki2003), tanohataite (Nagase et al., Reference Nagase, Hori, Kitamine, Nagashima, Abduriyim and Kuribayashi2012) and iwateite (Nishio-Hamane et al., Reference Nishio-Hamane, Minakawa and Okada2014).

The specimen containing mangani-eckermannite was obtained by the senior author in the early 2000s as ‘kôzulite’ and was stored in his systematic mineral collection until the recent investigation showed it to be a new mineral.

Mangani-eckermannite is associated closely with braunite, vittinkiite, albite and quartz.

General appearance and physical properties

Mangani-eckermannite occurs as prismatic crystals up to 0.3×0.2 mm and their aggregates up to 1 mm intergrown with braunite, vittinkiite and quartz (Figs 1 and 2). The colour of the new mineral is cherry-red to very dark red; in thicker grains it is reddish-brown. Mangani-eckermannite is translucent with a pinkish white streak and vitreous lustre. It is brittle, fracture is stepped along crystal elongation and uneven across a crystal. Cleavage is perfect on {110}. No fluorescence is observed under long- or short-wave ultraviolet light. Mohs hardness is 6 based on scratch tests. Density measured by flotation in Clerici solution is 3.16(2) g/cm³; the density value calculated using the empirical formula and unit-cell volume obtained from single-crystal X-ray diffraction (XRD) data is 3.186 g/cm³.

Figure 1. A 0.7 cm thick vein consisting of dark red mangani-eckermannite intergrown with black braunite, pink vittinkiite and colourless albite in a rock mainly consisting of colourless to pale salmon-reddish quartz. Holotype sample #5774/1, size: 2 × 1.5 × 1.5 cm. Photo: M. Milshina.

Figure 2. Anhedral grains of mangani-eckermannite (Meck) closely intergrown with braunite (Bnt), albite (Ab) and vittinkiite (Vtk). Polished section. Back-scattered electron image.

The new mineral is brown in transmitted light. Thicker grains are reddish brown. It is optically biaxial (–), with α = 1.645(3), β = 1.668(2) and γ = 1.675(3) (589 nm). The estimated 2V angle based on the conoscopical isogyre curve in the sections perpendicular to optical axes is 60(10)°, the calculated 2V value is 57°. Dispersion of optical axes is not observed but could be masked by the saturated colour of the mineral. Mangani-eckermannite is weakly pleochroic in brown colours, with Z ≈ Y > X. Twinning was observed with the twinning plane parallel to crystal elongation. The elongation is positive. The crystals which show twinning have a parallel extinction.

Spectroscopical studies

Infrared spectroscopy

In order to obtain the infrared (IR) absorption spectrum, a powdered mangani-eckermannite sample was mixed with anhydrous KBr, pelletised, and analysed using an ALPHA FTIR spectrometer (Bruker Optics) in the range of 360–3800 cm^–1, at a resolution of 4 cm^–1. A total of 64 scans were collected. The IR spectrum of an analogous pellet of pure KBr was used as a reference. The assignment of absorption bands in the IR spectrum of mangani-eckermannite (Fig. 3) is as follows: weak bands in the range of 3600–3800 cm^–1 correspond to O–H stretching vibrations; and bands of Si–O stretching vibrations are observed in the range of 900–1140 cm^–1. The ranges of 580–790 and 370–460 cm^–1 correspond to overlapping bands of O–Si–O and M⋅⋅⋅O–H bending modes and overlapping bands of M⋅⋅⋅O stretching and Si–O–Si bending modes, respectively where M = Mg, Mn or Fe and ‘⋅⋅⋅’ denotes a noncovalent (ionic) bond.

Figure 3. Powder infrared absorption spectrum of mangani-eckermannite.

The assignment of the bands of O–H stretching vibrations was carried out taking into account occupancies of sites coordinating OH groups (Mg_0.655Mn_0.345 for M1 and Mg_0.812Mn_0.188 for M3: see below) and data taken from literature sources.

For potassium-fluor-richterite and K-rich richterites, the band of stretching vibrations of OH groups belonging to the cluster MgMgMg–OH–(K) is observed in the range from 3694 to 3730 cm^–1 (Della Ventura et al., Reference Della Ventura, Parodi, Maras and Mottana1992, Reference Della Ventura, Robert and Hawthorne1998; Gottschalk and Andrut, Reference Gottschalk and Andrut1998). In the IR spectrum of synthetic end-member potassium richterite this band is observed at 3734 cm^–1 (Hawthorne, Reference Hawthorne1995).

As a result of annealing of Mn–bearing arfvedsonite in air, the OH stretching band of the (MgMgMg)–OH–A–OH configuration shifts downward from 3730 to near 3700 cm^–1 with formation of the (MgMgMg)–OH–A–O configuration due to dehydrogenation of OH at the O3 site (Ishida and Hawthorne, Reference Ishida and Hawthorne2001). According to Ishida and Hawthorne (Reference Ishida and Hawthorne2001), wavenumbers of O–H stretching bands of richteritic amphiboles (M = Fe or Mn) are 3644–3640 cm^–1 for the (MgR²⁺R²⁺)–OH–□ configuration and 3643 cm^–1 for the (MgMgR³⁺)–OH–□ configuration. According to Oberti et al. (Reference Oberti, Cámara, Della Ventura, Iezzi and Benimoff2006), the assignments of OH-stretching bands for parvo-mangano-edenite is as follows:

3692 cm^–1 for AlSi–(MgMgMn)–OH–^ANa,

3693 cm^–1 for SiSi–(MgMgMn)–OH–^ANa,

3672 cm^–1 for SiSi–(MgMgMg)–OH–^A□,

3677 cm^–1 for SiSi–(MnMnMn)–OH–^ANa,

3659 cm^–1 for SiSi–(MgMgMn)–OH–^A□

3641 cm^–1 for SiSi–(MgMnMn)–OH–^A□

A similar band assignment was made by Hawthorne and Della Ventura (Reference Hawthorne, Della Ventura, Hawthorne, Oberti, Della Ventura and Mottana2007). Generally, substitution of M²⁺ for M³⁺ in the local environment of the OH group results in the 30–40 cm^–1 shift of the OH-stretching band towards lower frequencies.

On the basis of these data and trends and taking into account relative band intensities, OH-stretching bands in the IR spectrum of mangani-eckermannite have been assigned to the following configurations:

3714 cm^–1 – (MgMgMg)–OH–^ANa

3703 cm^–1 – (MgMgMn²⁺)–OH–^ANa

3644 cm^–1 – (MgMn²⁺Mn²⁺)–OH–^ANa and/or (MgMgMn³⁺)–OH–^ANa

3612 cm^–1 – (MgMn²⁺Mn³⁺)–OH–^ANa and/or (MgMn²⁺Fe³⁺)–OH–^ANa

Raman spectroscopy

The Raman spectrum of mangani-eckermannite (Fig. 4) was obtained by means of a Horiba Labram HR Evolution spectrometer. This dispersive, edge-filter-based system is equipped with an Olympus BX 41 optical microscope, a diffraction grating with 600 grooves per millimetre, and a Peltier-cooled, Si-based charge-coupled device (CCD) detector. The Raman signal was excited by 532 nm laser with the beam power of 10 mW at the sample surface. Raman signal was collected in the range of 100–4000 cm^–1 with a 50× objective and the system being operated in the confocal mode, beam diameter was ~2.6 μm and the axial resolution ~5 μm. Time acquisition was 60 s per spectral frame and 4 accumulations were applied. Wavenumber calibration was done using the Rayleigh line and low-pressure Ne-lamp emissions. The wavenumber accuracy was ~0.5 cm^–1, and the spectral resolution was ~2 cm^–1. Band fitting was done after appropriate background correction, assuming combined Lorentzian–Gaussian band shapes using the Voight function (PeakFit; Jandel Scientific Software).

Figure 4. Raman spectrum of mangani-eckermannite excited by a 532 nm laser: in the 100–1250 cm^–1 region and in the 2500–4000 cm^–1 region (inset in the upper right corner). The measured spectrum is shown by dots. The curve matched to the dots is a result of spectral fit as a sum of individual Voigt peaks shown below the curve.

The assignment of Raman bands of mangani-eckermannite is analogous to the assignment of its IR absorption bands:

Bands at 3727 and 3705 and 3675 cm^–1 correspond to O–H stretching vibrations in the prevailing (MgMgMg)–OH–^ANa and (MgMgMn²⁺)–OH–^ANa local environments, respectively. The broad band at 3675 cm^–1 observed as a shoulder in the overall spectrum, may be due to the superposition of bands corresponding to the (MgMn²⁺Mn²⁺)–OH–^ANa and (MgMgMn³⁺)–OH–^ANa configurations.

A very weak peak at 3783 cm^–1 cannot be associated to an O–H stretching fundamental mode, but theoretically it may correspond to a combination (lattice + O–H stretching) mode. Intensities of bands corresponding to other local environments are below the detection limits.

Bands of Si–O stretching vibrations are observed in the range of 800–1100 cm^–1.

The ranges of 500–700 and 370–470 cm^–1 correspond to overlapping bands of O–Si–O and M⋅⋅⋅O–H bending modes and overlapping bands of M⋅⋅⋅O stretching and Si–O–Si bending modes, respectively.

Bands below 370 cm^–1 are due to lattice modes involving Na⋅⋅⋅O stretching vibrations and bending vibrations of Na- and M-centred polyhedra.

Chemical composition

Seven electron-microprobe analyses were carried out with a Cameca SX-100 electron microprobe (WDS mode with an accelerating voltage of 15 kV, a beam current on the specimen of 10 nA and a beam diameter of 4 μm). Peak counting times (CT) were 20 s for all elements; CT for each background was one-half of the peak time. The raw intensities were converted into concentrations using X-PHI (Merlet, Reference Merlet1994) matrix-correction software. The H₂O content was calculated from hydrogen content measured using secondary-ion mass spectrometry (SIMS) at Cameca IMS 4F. Lithium, beryllium and boron were sought after but their contents measured by SIMS were found to be below the detection limits of this method. The presence of OH^– groups and the absence of B–O, C–O and N–O bonds in the mineral is also confirmed by IR and Raman spectroscopic data.

Analytical data and standards used are given in Table 1.

Table 1. Chemical composition of mangani-eckermannite.

*Apportioned from 19.24 wt.% MnO measured by electron microprobe analysis

**For total Mn calculated as MnO.

***Calculated from 0.172 wt.% H (mean of 5 spots; range 0.170–0.174; S.D. 0.002) determined by SIMS.

S.D. = Standard deviation

The empirical formula, calculated on the basis of 24 anions (O+OH) per formula unit, is Na_2.26K_0.24Ca_0.24Mg_2.54Mn²⁺_1.69Mn³⁺_0.71Fe³⁺_0.26Ti_0.04Al_0.03Si_7.97O_22.48(OH)_1.52. All Fe was considered as Fe³⁺ due to the presence of Mn³⁺ and redox potential arguments, whereas the Mn²⁺:Mn³⁺ ratio was calculated from charge-balance requirements. At the next step, we filled the T and C sites in order to yield 13 cations: Si and Al were placed in T sites while Mg, Fe, Ti, Mn³⁺ and partly Mn²⁺ where placed in C, in accordance with the common scheme for amphiboles. After these positions were filled, some Mn²⁺ still remained and was placed in the B position – in very good agreement with the structure data that indicated more electrons than Na+Ca could give there. All the above results in the following chemical formula re-written according to the requirements of the current amphibole nomenclature (Hawthorne et al., Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012): ^A(Na_0.74K_0.24□_0.02)_Σ1.00 ^B(Na_1.52Ca_0.24Mn²⁺_0.24)_Σ2.00 ^C(Mg_2.54Mn²⁺_1.45Mn³⁺_0.71Fe³⁺_0.26Ti_0.04)_Σ5.00 ^T(Si_7.97Al_0.03)_Σ8.00O₂₂ ^W[(OH)_1.52O_0.48]_Σ2.00.

The ideal formula is NaNa₂(Mg₄Mn³⁺)Si₈O₂₂(OH)₂,which requires Na₂O 11.18, MgO 19.38, Mn₂O₃ 9.48, SiO₂ 57.79, H₂O 2.17, total 100 wt.%.

The Gladstone–Dale compatibility index calculated based on the empirical formula and unit-cell parameters from the single-crystal XRD data is 1 – (K _p/K _c) = 0.007 using D _meas and 0.015 using D _calc. Both values are rated as superior and indirectly but convincingly confirm the correctness of our chemical data, including the calculation of the Mn²⁺/Mn³⁺ ratio, the presence of iron in trivalent state and the determination of H₂O content. For details on the application of the compatibility concept in mineralogy see (Mandarino, Reference Mandarino1981) and references therein.

X-ray diffraction data and crystal structure

The powder XRD data (Table 2) were obtained using RKU-86 mm camera with CrKα radiation, V-filter and using Ge as an internal standard. The unit-cell parameters refined from the powder data using UNITCELL software by Holland and Redfern (Reference Holland and Redfern1997) are as follows: mangani-eckermannite is monoclinic, space group C2/m, a = 9.920(5), b = 18.112(4), c = 5.2920(4) Å, β = 103.97(4)°, V = 923(1)ų and Z = 2.

Table 2. Powder X-ray diffraction data (d in Å) of mangani-eckermannite.

To collect the single-crystal XRD data a crystal of mangani-eckermannite, 0.04 × 0.11 × 0.22 mm in size, extracted from the polished section analysed using electron microprobe and SIMS, was mounted on a glass fibre and examined with a Supernova Rigaku-Oxford Diffraction diffractometer equipped with a micro-source MoKα radiation (λ = 0.71073 Å; 50 kV and 0.8 mA) and a Pilatus 200K Dectris detector. The sample-to-detector distance was set to 68 mm. The data were collected by 1855 frames over 43 runs; the exposure time was 30 s per frame. The data were processed by CrysAlisPro 1.171.41.123a software (Rigaku Oxford Diffraction). The mineral is monoclinic, C2/m, a = 9.9533(4), b = 18.1440(7), c = 5.2970(2) Å, β = 103.948(4)°, V = 928.39(6) ų and Z = 2.

The crystal structure of mangani-eckermannite was solved from single-crystal X-ray data and refined with the use of the SHELX software package (Sheldrick, Reference Sheldrick2015) to R ₁ = 0.0264 for 1001 independent reflections with I > 2σ(I). Crystal data, data collection information and structure refinement details are given in Table 3, coordinates and equivalent isotropic displacement parameters of atoms in Table 4, selected interatomic distances in Table 5. Bond-valence sums are given in Table 6. Assigned site occupancies based on the results of the crystal structure refinement are given in Table 7. The crystallographic information file has been deposited with the Principal Editors of Mineralogical Magazine and is available as Supplementary material (see below).

Table 3. Single-crystal X-ray diffraction data collection information and structure refinement parameters for mangani-eckermannite.

*w = 1/[σ²(F _o²) + (0.0371P)² + 0.4615P]; P = ([max of (0 or F _o²)] + 2F _c²)/3

Table 4. Atom coordinates, equivalent isotropic displacement parameters (in Ų) and site occupancy factors (s.o.f.) for mangani-eckermannite.

* U _iso

** The position of H atom was found from difference Fourier synthesis and O3–H distance was restrained softly to be 0.90(1) and U _isoH = 1.2U _eqO3. The site occupancy factor for the H atom was fixed according to the chemical formula.

Table 5. Selected interatomic distances (in Å) in the crystal structure of mangani-eckermannite.

Table 6. Bond-valence analysis (values given in valence units) for mangani-eckermannite. Bond valence parameters from Gagné and Hawthorne (Reference Gagné and Hawthorne2015).*

* Note: Bond-valence calculations were performed taking into account the assigned site occupancies of Table 7.

Table 7. Site occupancies proposed for mangani-eckermannite based on the results of the crystal structure refinement.

SSF_exp and SSF_calc are the experimental and calculated site-scattering factors normalised to unity, respectively.

The crystal structure of mangani-eckermannite, similar to other amphiboles, is based on double chains of corner-sharing tetrahedra [T ₄O₁₁]^∞ running along the c axis. Adjacent double chains are linked via the strips built by M1–3-centred octahedra. M4-centred octahedra are located at the periphery of M1–3 octahedral strips. The A cations occupy the centre of the large cavity and are moved from the centre of the cavity with point symmetry 2/m to the A2 and Am sites, which is typical for amphiboles (Hawthorne et al., Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012).

The refinement of occupancies for the M1–3 sites was performed for Mg vs. Mn and results in the following values of the refined number of electrons (e _ref): 16.49, 22.43 and 14.44 for M1, M2 and M3, respectively. Refinement of the M4 site was performed for Na vs. Mn and gave e _ref = 13.72. Site occupancy factors of the Am and A2 sites were refined using the Na scattering curve and gave e _ref values of 5.80 and 0.52, respectively. Final site occupancies are given in Table 7. They are based on the results of the crystal structure refinement, analysis of interatomic distances, bond valence calculations and chemical composition.

Relationship to other species and the status of mangano-ferri-eckermannite

Mangani-eckermannite is another new member of the amphibole supergroup (Hawthorne et al., Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012). At lower hierarchical levels, it belongs to the sodium amphiboles subgroup within the group of ^W(OH,F,Cl)-dominant amphiboles. It is the Mn³⁺-dominant analogue of eckermannite ^ANa^BNa₂ ^C(Mg₄Al) ^TSi₈O₂₂ ^W(OH)₂ (Oberti et al., Reference Oberti, Boiocchi, Hawthorne, Ball and Harlow2015b) and magnesio-arfvedsonite ^ANa^BNa₂ ^C(Mg₄Fe³⁺)^TSi₈O₂₂ ^W(OH)₂ (Oberti et al., Reference Oberti, Boiocchi, Hawthorne, Ball and Harlow2015a). For the comparison of the three species see Table 8.

Table 8. Comparative data for mangani-eckermannite, eckermannite and magnesio-arfvedsonite.

*For the type specimen described by Oberti et al. (Reference Oberti, Boiocchi, Hawthorne, Ball and Harlow2015a)

Though manganese is a rather common component in amphiboles, in most cases it is present just as a small admixture. ‘Mangani-amphiboles’ i.e. those having Mn³⁺ as a dominant component amongst trivalent cations in C positions, are rare. Mangani-eckermannite is the seventh member of the amphibole supergroup with Mn³⁺ as a species-defining cation, after potassic-mangani-leakeite KNa₂(Mg₂Mn³⁺₂Li)Si₈O₂₂(OH)₂ (Armbruster et al., Reference Armbruster, Oberhänsli, Bermanec and Dixon1993 as “kornite”; renamed by Hawthorne et al., Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012), mangano-mangani-ungarettiite NaNa₂(Mn²⁺₂Mn³⁺₃)Si₈O₂₂O₂ (Hawthorne et al., Reference Hawthorne, Oberti, Cannillo, Sardone, Zanetti, Grice and Ashley1995 as “ungarettiite”; renamed by Hawthorne et al., Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012), mangani-obertiite NaNa₂(Mg₃Mn³⁺Ti⁴⁺)Si₈O₂₂O₂ (Hawthorne et al., Reference Hawthorne, Cooper, Grice and Ottolini2000 as “obertiite”; renamed by Williams et al., Reference Williams, Hatert, Pasero and Mills2014), mangani-dellaventuraite NaNa₂(MgMn³⁺₂Ti⁴⁺Li)Si₈O₂₂O₂ (Tait et al., Reference Tait, Hawthorne, Grice, Ottolini and Nayak2005 as “dellaventuraite”; renamed by Hawthorne et al., Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012), oxo-mangani-leakeite NaNa₂(Mn³⁺₄Li)Si₈O₂₂O₂ (Oberti et al., Reference Oberti, Boiocchi, Hawthorne, Ball and Ashley2016) and mangani-pargasite NaCa₂(Mg₄Mn³⁺)(Si₆Al₂)O₂₂(OH)₂ (Hålenius et al., Reference Hålenius, Bosi and Jonsson2020). All these amphiboles occur mainly in metamorphosed manganese deposits.

As mentioned above, the specimen of the new mineral described in this paper was initially received with the label ‘kôzulite’. The latter was described from Tanohata Mine by Nambu et al. (Reference Nambu, Tanida and Kitamura1969a) without structure description. Kitamura and Morimoto (Reference Kitamura and Morimoto1972) presented the structure refinement of a kôzulite crystal with the composition (Na_2.54K_0.27Ca_0.19)(Mn_3.69Mg_0.63Fe³⁺_0.33Al_0.31)Si_8.00O_21.789[(OH)_2.18F_0.04], however, they did not report detailed structure information, such as atomic coordinates and displacement parameters, as well as the valence of Mn. Later, Barkley et al. (Reference Barkley, Yang and Downs2010) refined the structure of kôzulite giving the ideal formula as NaNa₂[Mn²⁺₄(Fe³⁺,Al)]Si₈O₂₂(OH)₂. After the adoption of the current amphiboles nomenclature (Hawthorne et al., Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012), kôzulite was redefined to mangano-ferri-eckermannite (see, e.g., Oberti et al., Reference Oberti, Boiocchi, Hawthorne, Ball and Harlow2015b). The official IMA-CNMNC List of Mineral Names (Pasero, Reference Pasero2023, http://cnmnc.main.jp/) includes mangano-ferri-eckermannite as a valid mineral species referring to Barkley et al. (Reference Barkley, Yang and Downs2010) and Hawthorne et al. (Reference Hawthorne, Oberti, Harlow, Maresch, Martin, Schumacher and Welch2012). However, the empirical chemical formula was reported by Barkley et al. (Reference Barkley, Yang and Downs2010) as (K_0.20Na_0.80)_Σ1.00(Na_1.60Ca_0.18Mn²⁺_0.22)_Σ2.00(Mn²⁺_2.14Mn³⁺_0.25Mg_2.20Fe³⁺_0.27Al_0.14)_Σ5.00(Si_7.92Al_0.06Ti_0.02)_Σ8.00O₂₂[(OH)_1.86F_0.14]_Σ2.00, i.e. it has Mg > Mn²⁺ and Fe³⁺ > Mn³⁺ > Al as C-cations and, thus, formally corresponds to magnesio-arfvedsonite but not mangano-ferri-eckermannite. This questions the existence of mangano-ferri-eckermannite as a valid mineral species (although, it can exist as a so-called ‘named amphibole’), at least, as applied to the sample studied by Barkley et al. (Reference Barkley, Yang and Downs2010). It is obvious from the data that this sample was not a part of the original material studied by Nambu et al. (Reference Nambu, Tanida and Kitamura1969a) and Kitamura and Morimoto (Reference Kitamura and Morimoto1972) whose ‘kôzulite’ is obviously a ‘mangano-eckermannitic’ amphibole with only 0.63 Mg vs. 3.69 Mn apfu. However, the absence of structural data and the Mn²⁺:Mn³⁺ ratio does not allow assigning of their mineral to a particular mineral species.

It should be also noted that the data reported by Barkley et al. (Reference Barkley, Yang and Downs2010) show very close Mg and Mn²⁺ values (2.20 vs. 2.14 apfu) as well as Fe³⁺ and Mn³⁺ values (0.27 vs. 0.25 apfu). This means that their sample might contain local compositions corresponding to different amphibole species including the mangani-eckermannite described here. More recently, Nishio-Hamane et al. (Reference Nishio-Hamane, Momma, Ohnishi, Norimasa, Minakawa, Okada and Imai2022) studied another sample of ‘kôzulite’ from Tanohata Mine and found it to be mangano-mangani-ungarettiite and ‘mangano-mangani-eckermannite’, a potentially new mineral. Thus ‘kôzulites’ should be subjected to chemical and structural studies to be correctly assigned to a particular mineral species. Nishio-Hamane et al. (Reference Nishio-Hamane, Momma, Ohnishi, Norimasa, Minakawa, Okada and Imai2022) also stated that all samples of ‘kôzulite’ investigated earlier had dark red and reddish black colours, whereas mangano-ferri-eckermannite might exist but should be sought after in orange-coloured specimens.