The September 2007 ash eruption at Oldoinyo Lengai terminated a period of >30 y of relatively quiet extrusion of natrocarbonatite lavas. Ash erupted on the 24th September comprised nuclei of ijolitic phases, surrounded by finer peralkaline, silica-undersaturated material. The mantling material, as well as that forming nucleus-free particles, is dominated by nepheline, Ti-andradite, Na-melilite, combeite and a Na-Ca-phosphate-carbonate with minor amounts of K-Fe-sulphide and Mn-magnetite in a sparse matrix of alkali carbonate. The absence of clinopyroxene precludes this material from being termed either nephelinite or melilitite. We propose that the material represented by the mantles is an extremely undersaturated, peralkaline hybrid magma resulting from interaction between natrocarbonatite and nephelinite.

A multidisciplinary approach has been used to study Au occurrences within pyrite and arsenopyrite in four refractory Au ores from Colombia, France (Le Châtelet and Villeranges) and Portugal (Neves Norte). The Au was characterized by optical and scanning electron microscopy and analysed using electron and ion microprobes to determine Au distribution, with particular attention to spectral interferences in electron and ion microprobes, background measurements in electron probes, and quantitative analysis using external standardization in ion probes. The ionic emission rate is proven to be dependent on the Au status; combined Au has a greater ion emission than metallic Au. Invisible Au occurrences are closely linked to As distribution. Gold bonding in arsenopyrite, examined by transmission electron microscopy, is shown to be dispersed within the FeAsS crystal structure. Typical growth patterns and As-Au diffusion zoning in pyrite and arsenopyrite may account for the very irregular distribution of Au in these minerals.

Electron microprobe analyses are presented for astrophyllite-group minerals from hydrothermal veins and pegmatites of the Ilimaussaq complex, South Greenland. The analyses fall mainly into two groups: (1) niobophyllites with the highest Nb/(Nb+Ti) ratios yet recorded (∼0.9), occurring only in the veins, and (2) an essentially continuous sequence from astrophyllite to niobophyllite with Nb/(Nb+Ti) up to 0.6, found in veins and pegmatites. It is highly likely that there is complete solid solution between astrophyllite and niobophyllite. More limited substitution of Mn for Fe has resulted in the formation of kupletskite in some rocks. Altered zones in certain astrophyllites and niobophyllites have compositional features similar to the type ‘hydroastrophyllite’. The astrophyllite-group minerals in the hydrothermal veins crystallized at temperatures below 400°C at 1 kbar and under high pH and low oxygen fugacity, whereas those in the pegmatites were formed from water-rich melts which were hotter (≥450°C), less basic and more oxidized.

Iron played a decisive part when uranyl phosphates (‘yellow U ores’) formed during supergene alteration of the aplitic and pegmatitic rocks of the Hagendorf Pegmatite Province. Three different supergene U mineral successions, referred to in the current study as pathways I to III, were identified. Pathway I began with the release of PO42– by the dissolution of rockbridgeite and ended after shortdistance transport under alkaline reducing or slightly acidic oxidizing conditions in the precipitation of bassetite. Pathway II is an advanced form of pathway I that did not stop with the precipitation of bassetite, but progressed by acidic Cu-bearing meteoric waters under oxidizing conditions up to the torbernite precipitation stage. Minor amounts of Mn or Ca may have led to a deviation from the normal pathway into the stability fields of lehnerite or autunite, respectively, both of which may occur either as solid solution series or in a layered intergrowth with torbernite. Limonite-cored torbernite has been described for the first time and only exists in pathway III. Unlike its counterpart pathways I and II, which appeared at the end of a complex polystage element recycling process of secondary Fe phosphates under fluctuating redox and pH conditions, limonite-cored torbernite resulted from a monostage transformation of primary ‘black Fe-U ore minerals’ under strongly oxidizing conditions and short-distance element transport. These restricted physicochemical conditions caused the immediate stabilization of the Fe-U-P system and by doing so the U-Pb ratios of the black ore progenitor were well preserved in the limoniticc ore. Torbernite was analysed for U, Th and Pb isotopes by laser ablation inductively coupled plasma mass spectrometry techniques. For one domain the data yielded a formation age of 4.55±0.02 Ma, which corresponds to Miocene-Pliocene weathering and geomorphological processes in the study area. A second domain gave a discordia with an upper intercept age of 549±12 Ma, interpreted to represent a thermal event at the Precambrian-Cambrian boundary. Ferrocolumbite is found exclusively in the mineralization of pathway III. Due to the proximity of ferrocolumbite to torbernite, limonite-cored torbernite probably inherited the 549 Ma age from ferrocolumbite during supergene alteration.

Hemimorphic crystals of four M1+-M3+-diarsenates (M1+ = Li, Na; M3+ = Al, Sc, Ga) were prepared by mild hydrothermal methods (T = 493 K) at very low pH values of ~1–1.5. The crystalst ructures of all compounds were determined using single-crystal X-ray diffraction data (Mo-Kα, CCD detector, 293 K) and were found to be isotypic with gittinsite (CaZrSi2O7), and the later described synthetic compound LiFeAs2O7. Crystaldat a for LiAlAs2O7, LiScAs2O7, LiGaAs2O7 and NaScAs2O7, respectively, are: space group C2 (no. 5), a = 6.583(1), 6.790(1), 6.638(1), 6.959(1), b = 8.007(2), 8.532(2), 8.181(2), 8.712(2), c = 4.635(1), 4.845(1), 4.696(1), 4.817(1) Å, β = 104.13(3), 103.75(3), 104.01(3), 102.05(3)°, V = 236.92(9), 272.64(9), 247.43(9), 285.6(1) Å3; Z = 2. The structure models were refined anisotropically to R1(F) = 0.0152, 0.0157, 0.0144, 0.0199, respectively.

The non-centrosymmetric framework structure contains isolated As2O7 groups (in staggered configuration; As–O–As angles show a narrow range between 130.32(16) and 139.8(2)°) which share their O ligands with MO6 (M = Al, Sc, Ga) octahedra and distorted Li/NaO4 tetrahedra. The strongly distorted LiO4 tetrahedra exhibit an average Li–O bond length of 2.071 Å, but four longer Li–O bonds (between 2.781–3.042 Å) also contribute to give a reasonable bond-valence sum for the Li atoms, and reflect an effective [4+4]-coordination. The Na compound is also characterized by four short and four long Na–O bonds; the Na atom appears slightly overbonded. The average Sc–O bond length in the two Sc representatives is 2.083 and 2.097 Å.

Details of the topology and crystal chemistry are discussed in comparison with those of gittinsite and other diarsenates, and correlations are noted. A reevaluation of the geometry of As2O7 groups in 58 presently known and structurally characterized inorganic diarsenates shows a mean As–O–As angle of 124.2° (excluding 180° angles). The As–O bond lengths to the central bridging O atom are elongated to a mean value of 1.755 Å, whereas all other bond lengths are shortened (mean value 1.666 Å); the overall mean As–O bond length is 1.688 Å.

The crystal structure of bornemanite, ideally Na6☐BaTi2Nb(Si2O7)2(PO4)O2(OH)F, a = 5.4587(3), b = 7.1421(5), c = 24.528(2) Å, α = 96.790(1), β = 96.927(1), γ = 90.326(1)°, V = 942.4(2) Å3, space group (P1̄), Z = 2, Dcalc. = 3.342 g cm–3, from the Lovozero alkaline massif, Kola Peninsula, Russia, has been solved and refined to R1 = 6.36% on the basis of 4414 unique reflections (Fo >4sF). Electron microprobe analysis yielded the empirical formula (Na6.07Mn2+0.23Ca0.06☐0.64)Σ7.00 (Ba0.73K0.13Sr0.06☐0.08)Σ1.00(Ti2.05Nb0.80Zr0.02Ta5+0.01Fe3+0.03Al0.02Mn2+0.06Mg0.01)Σ3.00(Si2O7)2(P0.97O4)O2 [F1.27(OH)0.74]Σ2.01. The crystal structure of bornemanite is a combination of a TS (titanium silicate) block and an I (intermediate) block. The TS block consists of HOH sheets (H-heteropolyhedral, O-octahedral). The TS block exhibits linkage and stereochemistry typical for Group III (Ti = 3 a.p.f.u.) of Ti-disilicate minerals: two H sheets connect to the O sheet such that two (Si2O7) groups link to the trans edges of a Ti octahedron of the O sheet. The O sheet cations give Na3Ti (4 a.p.f.u.). The TS block has two different H sheets, H1 and H2, where (Si2O7) groups link to [5]Ti and [6]Nb polyhedra, and there are two peripheral sites which are occupied by Ba and Na, respectively. There are two I blocks: the I1 block is a layer of Ba atoms; the I2 block consists of Na polyhedra and (PO4) tetrahedra.

Phase-X, a potential sink for K in the mantle, is a synthetic hydrous K-rich silicate formed by the breakdown of K-amphibole at high pressure. It has the general formula A2–xM2Si2O7Hx where A = K, Na, Ca, □ (vacancy), and M = Mg, Al, or Cr. No other isomorphic substitutions, either for the A or the M site, were reported for such a synthetic compound. Here we report the crystal structure and chemical composition of a crystal of phase-X containinglarg e amounts of trivalent Fe. This crystal was synthesized in the model system garnet lherzolite–K2CO3 at lower pressure (P = 7 GPa) and higher temperature (T = 1450–1650°C) with respect to the stability range reported in the literature (i.e. P = 9–17 GPa and T = 1150–1400°C). Quantitative analysis led to the following formula: . The lattice parameters (hexagonal setting) are: a = 5.005(1), c = 13.148(2) Å, V = 285.23(9) Å3, and Z = 2. The structure was refined in space group P63cm to R = 5.06% using199 independent reflections and consists of MO6 octahedra layers stacked alongthe c axis and linked together by Si2O7 groups. The Si2O7 groups form pillars in the layer that contains A atoms in the cavities between the pillars. The Raman spectrum in the OH-stretching region indicates the existence of two different OH environments. However, the position of H could not be determined. The substitution of Fe3+ for Mg shortens octahedral bond distances. In addition, the entry of Fe3+ in M induces geometrical changes to the adjacent A site. The crystal-chemical characteristics are compared with published data on synthetic Fe-free phase-X.

Electron microprobe analysis (EMPA) and microscopic investigations were performed on monazite grains from the black sand deposits of northern Sinai beach. Electron microprobe backscattered electron images, X-ray mapping and wavelength dispersive spectroscopy line scans showed some grains with sector zoning and others with thorite inclusions. Based on the EMPA data, the studied monazite is grouped into monazite-(Ce) and Th-rich monazite. Monazite-(Ce) is enriched in REE and P, whereas Th-rich monazite is enriched in Th, U, Ca, Y, Si and Fe. The compositional variations of monazite are governed by the substitution of REE by Th, U, Ca and Y. The monazite grains show enriched chondrite-normalized REE patterns represented mainly by LREE and only Gd from the HREE. These patterns demonstrate negative Nd and Euanomalies. The floods associated with the pluvial periods which prevailed in Egypt during the Pleistocene, were able to erode the source rocks and liberate heavy minerals including monazite. The mineral grains were transported through several wadis and tributaries to the main channel of the River Nile. At the confluence sites, these heavy minerals were mixed with Ethiopian and central African heavy mineral assemblages. The grains continued to move together downriver until being deposited in their current locations. The analytical results suggest that pegmatites and granites of the Eastern Desert are the most likely source of the monazite. However, due to the large area of the Nile watershed, other undiscovered sources are possible.

The systematics of cation ordering in binary spinel solid solutions have been investigated using an interatomic potential model combined with Monte Carlo simulations. The formalism to describe a system containing three cation species ordering over two non-equivalent sub-lattices is developed and the method applied to the MgAl2O4-FeAl2O4 binary solid solution. Our results compare favourably with experimental measurements of site-occupancy data, although the experiments display a slightly larger degree of non-ideality than the simulations. A possible kinetic origin of the non-ideal behaviour was examined by performing simulations in which only exchange of Mg and Fe2+ between tetrahedral and octahedral sites was permitted below the Al-blocking temperature of 1160 K. This approach improves the agreement with the experimental site occupancies, and suggests that the blocking temperature for moving Mg and Fe2+ between tetrahedral and octahedral sites is significantly lower than for moving Al

Accessory tourmaline in metasediments from the Sinai crystalline basement exhibits textural and chemical signatures that relate to the evolution of regional metamorphism and deformation during the Pan-African orogeny and testifies to different P-T path segments. Tourmaline inclusions in various porphyroblasts were formed during the prograde phase of metamorphism; acicular to prismatic crystals in the matrix, oriented sub-parallel to, and enveloped by, the main foliation crystallized syntectonically under prograde and peak metamorphic conditions; tourmaline cross-cutting the main foliation may have formed just after the peak or during the retrograde phase of metamorphism. Some of the cores in tourmaline crystals, showing different colours, are interpreted as former detrital grains. The abundance of tourmaline decreases with increasing peak metamorphic conditions. The tourmaline investigated belongs to the schorl-dravitess group, generally with XMg of 0.42–0.73 and XCa = Ca/(Ca+Na+K+□) of 0.02–0.24, typical of tourmalines in metapelites and metapsammites; whereas detrital cores have been derived from various sources, including former tourmaline-quartz and pre-existing high-metamorphic rocks. Tourmaline of the Sinai metasediments was formed during metamorphism of the sedimentary precursors, essentially in a closed system, where clay minerals and organic matter, together with detrital tourmaline, served as the source of boron. Although a metamorphic facies should be defined by characteristic mineral assemblages present in metamorphic rocks, tourmaline chemistry is a good monitor of P-T conditions in the metapelites and semi-metapelites investigated, showing an increase in XMg with increasing metamorphic grade, where XturMg = 0.60 distinguishes between greenschist and lower-amphibolite facies, while XturMg = 0.65 could distinguish lower- from middle- to upper-amphibolite facies. The results of tourmaline-biotite geothermometry compare well with our former temperature estimates using conventional geothermometry and phase-diagram modelling.

Kutnohorite with moderate and bright orange-red cathodoluminescence (CL) was studied by CL microscopy and spectroscopy. This mineral was found in fossiliferous concretions composed mainly of rhodochrosite from the Mn-carbonate mineralization at Úrkút, Hungary. The CL microscopy reveals that kutnohorite occurs as impregnations, layers and veinlets. X-ray diffraction, infrared spectroscopy and electron microprobe studies indicate that the luminescent kutnohorite has excess Ca (72.9–80.0 mol.% CaCO3, 16.3–20.5 mol.% MnCO3, 3.3–5.6 mol.% MgCO3 and 0.0–0.5 mol.% FeCO3). Transmission electron microscopy shows that the luminescent carbonate has a dolomite-type structure, with modulated and mosaic microstructures. The CL spectra of this Ca-rich kutnohorite have a single emission band at 630 nm that is characteristic of Mn2+ substitution in the structure. Our results provide evidence for moderate-to-bright cathodoluminescence of Mn-rich natural carbonates even at 8–10 wt.% Mn and up to 2400 ppm Fe. The self-quenching of Mn appears incomplete in the case of Ca-rich kutnohorite from Úrkút.

The occurrence is reported of spectacular oriented needles of rutile within quartz veins and pods in granulite-facies rocks that have undergone partial melting under ultrahigh-temperature (UHT) conditions in the Karur region in southern India. Laser Raman spectroscopy and electron microprobe analysis confirm that (1) the mineral inclusions within quartz are pure rutiles, and (2) a secondcategory of quartz contains pure hematite inclusions. The rutile never coexists with the hematite inside the same quartz specimens. Fluidinclusions in the quartz are characterized as a CO2 + H2O mixture. Application of two geothermometric calibrations utilizing Ti in quartz yields minimum estimates of ~1190ºC, suggesting the formation of the rutile-quartz under UHT metamorphic condition. It is proposed that such Ti-rich quartz segregations couldbe another indicator for extreme crustal metamorphism at very high temperatures.

Micas of the muscovite–lepidolite series are main constituents of the evolved pegmatites from the Okatjimukuju-Kaliombo portion of the Karibib belt, Namibia. The compositional variations shown by the micas from the intermediate zones are mainly controlled by the Li3Al-1□-2 and SiLi2Al-2□-1 substitution schemes, whereas for the micas from the core margins and the replacement bodies, only the first of these two exchange vectors seems to operate. The chemical composition of the micas not only depends on the degree of pegmatite evolution, but also on the position in the internal zonation of the pegmatite. Micas from the core margins and the replacement units are generally richer in F, Li, Rb, Cs and Zn than those from the intermediate zones. In general, the contents of these elements increase with decreasing K/Rb ratio. However, some data departing from this general trend are also observed, which could be related to subsolidus processes. Some pegmatite bodies show a complete internal evolution, developed from the margins to the core zone, which is reflected in the chemical composition of the micas. The regional distribution of pegmatites does not define a zonation, because an overlapping of pegmatites with different degrees of evolution occurs. This could be due to the high level of evolution attained by most of the rare-element pegmatites, and to their topography with respect to a dome structure of the basement.

Polarized far- and mid-infrared reflectance spectra were measured at room temperature for the new mineral species jarandolite, CaB3O4(OH)3, for the two principal directions. Thirty three vibration modes for and 32 for were observed and analysed numerically. Symmetry analysis predicts 41 and 40 vibration modes that include lattice and O–H modes. Mode assignment was made based on the structure of jarandolite. The values of the mode frequencies (ωTO), the damping factors (γ) and the oscillator strength (S) of each oscillator were obtained by fitting to a Lorentz model.

In order to better understand the formation and evolution processes of ultrahigh pressure (UHP) felsic rocks, we determined the ages of various domains of zircon and monazite crystals from the diamondiferous quartzofeldspathic rocks of the Saxonian Erzgebirge. According to cathodoluminescence imagery and Th/U ratios, three zircon zones were distinguished. Each was dated using several spot analyses from a sensitive high-resolution ion microprobe analysing Pb, U and Th isotopes. The results were: (1) core zone – 21 analyses: Th/U ≤ 40.023 and 337.0±2.7 Ma (2σ, combined 206Pb/238U-207Pb/235U age); (2) diamond-bearing intermediate zone – 23 analyses: Th/U ≥ 50.037 and 336.8±2.8 Ma; and (3) rim zone – 12 analyses: Th/U = 0.015–0.038 (plus one analysis of 0.164) and 330.2±5.8 Ma. The U-Pb obtained ages are virtually concordant. Furthermore, two oscillatory zoned zircon cores (Th/U ≥ 50.8) yielded (~concordant) ages of ~400 Ma. Six SHRIMP analyses of monazites gave an age of 332.4±2.1 Ma. In addition, Pb, Th and U contents in monazite were analysed with an electron microprobe (EMP). A mean age of 324.7±8.0 (2σ) Ma was acquired from 113 analyses.

By combining the defined ages with previously published P-T conditions, minimum velocities for burial and exhumation were estimated. In addition, we present a likely geodynamic scenario involving age data from the literature as well as this study: beginning 340 million years ago, gneisses at the base of a thickened continentalcrust (~1.8 GPa, 650ºC) were transported to depths of at least 130 km, possibly as deep as 250 km. Here they were heated (>1050ºC) and partially melted and as a result began to rise rapidly. The burial and subsequent ascent back to a depth of 50 km, where zircon rims and monazite formed, took only a few million years and perhaps significantly less.

The elemental compositions of zoned alkali feldspar megacrysts from the Karkonosze pluton have been analysed and Pb isotope ratios determined using LA-ICP-MS, EMPA and TIMS. The results are used to interpret the magmatic environments within which they crystallized. Growth zones in the megacrysts show fluctuating trace element patterns that reflect a systematic relationship between incompatible LREE and compatible Ba. Chemical gradients between zones in the cores and rims of the megacrysts are not accompanied by significant variation in initial Pb isotope composition. The nucleation and crystallization of the megacrysts is interpreted as having occurred in an environment of magmatic hybridization caused by mixing of mantle and crustal components in which effective homogenization of the Pb isotope composition preceded the onset of megacryst growth. The concentrations of LREE in alkali feldspar zones were used to reconstruct hypothetical melt compositions. Some of the zones appear to have crystallized in an homogenous magmatic environment having clear geochemical affinities with end-member magmas in the Karkonosze pluton, whereas others crystallized in heterogeneous domains of magma hybridization. With the exception of Nd, zones crystallized in more homogeneous magma show LREE fractionation under near-equilibrium conditions. Trace element abundances of megacrysts grown in dynamic, homogeneous magmatic environments of the Karkonosze pluton occasionally deviate from the predicted patterns and show LREE impoverishment.

The crystal structures of nine, and the chemical compositions of ten, natural samples of leucophanite, ideally NaCaBeSi2O6F, were investigated. The analysed samples display a large compositional variation with trace-element abundances >50,000 ppm, primarily due to rare earth elements (REE). Fromthese data, we propose a substitution scheme for the incorporation of REE for Ca, with additional Na substituting for Ca and the generation of vacancies to ensure charge balance. Compositional zonation was observed in some samples; this zonation correlates with variations in cathodoluminescence. The crystal structure of the nine analysed samples could all be refined in space group P212121. We found no evidence for a reduction of symmetry with increased trace-element concentration. Various twin combinations were observed and these seem related to crystallization conditions rather than structural or chemical factors.

Ancient ivory, from the Chengdu Jinsha and Guanghan Sanxingdui sites in China, has been buried for several thousand years. In order to determine the degradation mechanisms and to provide a scientific basis for protecting them, these ancient ivory samples have been compared with modern ivory using infrared spectroscopy in the frequency range 400–4000 cm–1. By combining chemical analysis data we compare the crystallinity and crystal chemistry of the apatite component, as well as the structural characteristics of the ivory. These investigations showed that the ancient ivory consists almost entirely of hydroxyl-carbonate apatite as the predominant phase. Compared with the modern ivory, the PO43– and CO32– bands are stronger, the PO4RF values are obviously greater, and an extra OH– band at 3569 cm–1 is observed in the ancient ivory. These results indicate that there is a greater degree of apatite crystallinity in the ancient ivory and also imply that there has been incorporation and recrystallization of CO32– in the apatite during burial. Positive correlations have been found between the apatite crystallinity, CO32– and OH– ion contents, and burial time. The organic matter in ancient ivory has been lost or decomposed as the organic bands (e.g. at 1238 cm–1 and 1337 cm–1) have disappeared. This may be the main reason that ancient ivory is easily dewatered and readily friable after being unearthed.

Volcanic activity at Mt Vulture lasted throughout the Middle Pleistocene and produced SiO2- undersaturated volcanics. Deposits from the Monte Vulture stratovolcano have been classified into four subsynthems and clustered into the Barile Synthem. In the present investigation, trioctahedral micas from the uppermost units (the Ventaruolo Subsynthem) of the Barile Synthem are considered. The samples are labelled VUT187. The phlogopitic micas were separated from the host rock (an olivinefoidite) and underwent chemical (electron microprobe analysis - EMPA and C-H-N), structural (singlecrystal X-ray diffraction) and spectroscopic (Mössbauer) investigations.

The EMPA yielded: MgO (17.62–21.89 wt.%), FeOtot (5.98–9.78 wt.%), TiO2 (1.81–3.92 wt.%) and Al2O3 (14.47–17.98 wt.%), with H2O contents = 2.86 (±0.42) wt.% determined by C-H-N analyses. Mössbauer investigation provided [VI]Fe2+ = 12.6%, [VI]Fe3+ = 87.4%. The chemical and structural data are consistent with the occurrence of Ti-oxy, [VI]M2+ + 2(OH)– ⇋ [VI]Ti4+ + 2O2– + H2, and M3+-oxy substitutions, [VI]M2+ + (OH)– ⇋ [VI]M3+ + O2– + 1/2H2, with M3+ = Fe3+, Al3+. In particular, Fe3+-oxy substitution has affected the Fe2+/Fe3+ ratioin the studied sample. This is probably due to the fact that interaction with underground water or a hydrothermal system may have altered the oxygen fugacity and raised the Fe3+ content of VUT187 phlogopite with respect to magmatic values.

Two types of carbonatic cumulate xenoliths occur in alkali basalts of the northern part of the Carpatho-Pannonian region, Central Europe. One is dominated by Ca-Fe-Mg carbonates with randomly distributed bisulphide globules (Fe1+xS2, x = 0–0.1), Mg-Al spinel, augite, rhönite, Ni-Co-rich chalcopyrite, and a Fe(Ni,Fe)2S4 phase. The second, carbonatic pyroxenite xenolith type, is composed of diopside, subordinate fluorapatite, interstitial Fe-Mg carbonates, and accessory K-pargasite, F-Al-rich ferroan phlogopite, Mg-Al spinel, albite and K-feldspar. All accessory minerals occur in ultrapotassic dacite-trachydacite glass in primary silicate melt inclusions in diopside, together with calcio-carbonatite and CO2-N2-CO inclusions. Textural evidence is provided for multiphase fluid-melt immiscibility in both xenolith types. The carbonatic pyroxenite type is inferred to have accumulated from differentiated, volatile-rich, ultrapotassic magma derived by a very low-degree partial melting of strongly metasomatized mantle. Mineral indicators point to a genetic link between the carbonatite xenolith with olivine-fractionated, silica-undersaturated alkalic basalt ponded at the mantle-crust boundary.