X-ray absorption spectroscopy (XAS) has been used to study the local structures of Fe, Co, Zn, and Mg in synthetic and natural staurolites. Results for the near-edge features (XANES) and refined EXAFS were used to deduce the crystallographic site(s) occupied for each element. The least squares refined, mean first shell metal-O bond lengths, coordination numbers, and Debye-Waller factors are in the ranges: Fe 1.97–1.99 Å, 3.5–4.1, 0.013–0.023 Å2; Co 1.97–1.98 Å, 3.5–3.9, 0.009–0.014 Å2; Zn 1.95–1.96 Å, 3.7–3.8, 0.008–0.009 Å2; and Mn 1.99 Å, 3.8, 0.012 Å2. No significant differences were found which depended on element concentration or whether the sampes are synthetic or natural. The refined bond length for Mg in end-member Mg-staurolite is 2.00 Å but EXAFS data did not allow refinement to give reliable coordination number and Debye-Waller factor estimates. The EXAFS data, together with staurolite and model compound XANES features, suggest that > about 90% of Fe, Co, Zn and Mn are concentrated in the tetrahedral T2 site, while in Mg-staurolite > about 75% of the total Mg is in T2 with the remainder in octahedral sites. In natural staurolite from Pizzo Forno (PF2), a greater proportion of Mg appears to be in octahedral coordination.

The Lower Mafic Succession of the Great Dyke is a 700 m thick sequence of gabbroic rocks which shows remarkably regular mineral compositional trends and trace element contents in whole rocks. Such chemical trends are strongly indicative of undisturbed fractionation having taken place within the magma chamber and contrast with the major development of cyclic units which characterize the underlying Ultramafic Sequence of the Great Dyke. The style of fractionation is quite different to that in the equivalent Main Zone of the Bushveld Complex with the latter possibly reflecting a ‘leaky’ input system, whereas in the Great Dyke the magma chamber was sealed. Major compositional reversals at the interface between the websterite layer (the topmost unit of the Ultramafic Sequence) and the base of the Lower Mafic Succession indicate a change in crystallization conditions at this level. Modal percentages of plagioclase and Al2O3 content of pyroxenes show the same trends indicating a strong control by temperature and magma composition.

Modelling of the fractionation processes and the influence of trapped liquid was carried out for Mg#, Cr2O3, and NiO in pyroxenes and for Zr in whole rock. The lowermost gabbroic rocks are adcumulates with effectively zero trapped liquid which contrasts with 10–15% trapped liquid in the underlying websterite There is a gradual rise in the amount of trapped liquid upwards in the Lower Mafic Succession. These results have implications for the mechanisms by which porosity is reduced in mafic cumulates. An injection of a small amount (10%) of new magma at the interface of the Ultramafic-Mafic Sequences of the Great Dyke was of a composition slightly different to that which gave rise to the cyclic units of the Ultramafic Sequence.

Intruded into the Palaeogene lava field and underlying Moine (Neoproterozoic) crystalline basement rocks around Loch Scridain, Isle of Mull, Scotland, is a suite of high-level, inclined, xenolithic sheets, ranging in composition from basalt, through andesite and dacite, to rhyolite. These sheets, associated with the Mull central volcano, were emplaced post 55 Ma. As well as numerous crustal xenoliths, the more basic members of the complex contain a diverse suite of ultrabasic and basic xenoliths. Xenolith types include feldspathic peridotite with cumulus olivine, pyroxenite, gabbro with cumulus plagioclase and cumulus clinopyroxene, and pure anorthosite. Mineralogical data, coupled with whole-rock major- and trace-element data from a small number of the xenoliths suggest that the xenoliths represent early-formed cumulates cognate with their host basalts. Sr and Nd isotope data from the xenoliths confirms the cognate origin, and also shows that the basic magmas suffered crustal contamination at an early stage.

The Minastira granite, a c. 25 Ma subvolcanic plug of fine-grained granitic rock in the Cordillera Oriental of SE Peru, has preserved textures indicative of a history involving mixing of at least two magmas, a volumetrically dominant felsic component and a less voluminous mafic one. The felsic component is represented by variably fractured, altered and embayed crystals of quartz, feldspar, biotite with minor coarsegrained melt- and fluid-inclusion rich apatite, and possible cordierite (now a pseudomorphous Fe-Mg phase), whereas the mafic component is represented by calcic plagioclase. The process of magma mixing is reflected by: (1) ubiquitous sieved-textured plagioclase with complex textural relationships; (2) a large range in plagioclase compositions with reversals and spike patterns in profiles; (3) embayed and internally fractured (thermal shock?) quartz; (4) the rare occurrence of pyroxene coronas on quartz; and (5) textures within biotite suggestive of its incipient breakdown. The lack of mafic enclaves indicates that physico-chemical conditions of the mixing were conducive to homogenization (i.e. chemical diffusion) and a superficially homogeneous rock is now observed. The association of glomeroclasts of crystals originating from both the mafic and felsic end members and a quenched quartz-feldspar matrix indicate that the mixing occurred in an underlying magma chamber.

The chemistry and textures of the minerals in the layered syenites of Nunarssuit and West Kûngnât record a complex post-crystallisation history. The initial cumulus phases were alkali feldspar, pyroxene and olivine, apatite and opaque oxides. Grain size distributions of the pyroxenes and olivines have been modified since crystallisation, possibly during transport of these grains in a density current, prior to deposition in layers. Olivines and pyroxenes re-equilibrated with intercumulus melt. The lower ratio of mafic phases to intercumulus fluid in leucocratic layers has resulted in pyroxene and olivine in these layers undergoing more re-equilibration that in melanocratic layers, consequently pyroxenes and olivines in leucocratic layers are more ferroan. Pyroxenes reacted with either intercumulus or deuteric fluid to form amphibole. In West Kûngnât pyroxene and amphibole re-equilibrated via a later fluid. Olivines reacted with deuteric fluid to produce biotites, probably at temperatures < 550°C. After crystallisation alkali feldspars exsolved, the exsolution textures coarsened in the presence of deuteric fluids with a magmatic origin, possibly at temperatures as low as 450°C. Zoning patterns of apatites and zircons were modified during interactions with either, or both intercumulus or deuteric fluids. Leucocratic layers underwent more modification than melanocratic ones, probably due to the higher concentration of intercumulus melt in the former leading to a higher concentration of deuteric fluids as water exsolved from the magma during cooling. The syenites underwent little compaction during solidification.

In the Halls Creek Orogen of north-western Australia, the distance of melt migration through migmatitic metasedimentary rocks and adjacent metabasites is partly constrained by relationships of leucosomes and small mafic magma veins to rock boundaries and structural elements. Stromatic leucosomes in metasediments are cut by a network of small extensional fractures and shear zones, oriented steeply during melt migration. These shear zones allowed cm- to 10 m-scale migration of felsic magma derived by in situ anatexis. In the adjacent metabasite layers, a similar shear array allowed injection of H2O-undersaturated mafic to ultramafic magma, locally dehydrating and chemically modifying these rocks. However, these mafic to ultramafic veinlets are too mafic to be explained by in situ anatexis, necessitating an external magma source. Also, the lack of felsic veinlets cutting metabasites, and mafic veinlets cutting metasediments, requires that vertical inter-connectivity of these fracture systems was restricted. We propose along-layer migration of mafic to ultramafic magma through the metabasite, assisted by horizontal connection of the shear zones. This migration occurred independantly of metre-scale felsic magma migration in the adjacent metasediments, even though these two deformation-assisted magma migration systems may have been operating at the same time.

In 1962 Dawson described a calcite carbonatite (specimen BD83) from the Tanzanian volcano Oldoinyo Lengai as a sövite, thus implying that at an earlier stage in its evolution this volcano had crystallized magmatic calciocarbonatites as well as the highly alkalic natrocarbonatite lava that has been erupted in more recent times. This proposition is difficult to reconcile with the currently fashionable hypothesis whereby the natrocarbonatite lava separated immiscibly from a type of nephelinite magma, most recently thought to be a wollastonite nephelinite. In 1993 Dawson sought to discredit the magmatic origin of this sövite specimen by arguing that it was derived from natrocarbonatite lava through extreme alteration (calcitization) in which process the original nyerereite was replaced by calcite in near-perfect pseudomorphs. We suggest that the arguments advanced in support of this concept are unconvincing and that the specimen is exactly what it was originally described as, namely a magmatic sövite in which the calcite crystallized from a magma rather than having replaced nyerereite. We do not seek to discredit the liquid immiscibility hypothesis but do believe that whatever process is responsible for the Oldoinyo Lengai natrocarbonatites and silicate rocks must also allow for the crystallization of calciocarbonatite.

Strongly peraluminous, cordierite-bearing anatectic dacites from the offshore Tertiary Erlend volcanic centre, north of the Shetland Isles, are shown to contain graphite which is interpreted as being essentially a restite phase inherited from carbonaceous pelitic source rocks. The form and characteristics of the graphite are documented and graphite geothermometry applied to establish that the graphite records a minimum peak temperature of ∼800°C, confirming that temperatures at which anatexis occurs were attained. The different morphological forms of graphite observed suggest the possibility that minor amounts of fluid-deposited graphite may also be present. The chemistry of the Erlend dacites is compared with that of other known examples of graphite-bearing peraluminous silicic igneous rocks and briefly with experimentally generated peraluminous liquid compositions. The Erlend source rocks were probably subjected to a higher degree of partial melting than has occurred in the petrogenesis of many other anatectic peraluminous silicic rocks.

It is now generally agreed that the term ‘agpaitic’ should be restricted to peralkaline nepheline syenites (and phonolites) containing minerals such as eudialyte and rinkite, that is complex silicates of Zr, Ti, the rare earth elements (REE), and F and other volatiles. There are, however, cases of transition into more common types of nepheline syenites containing zircon, titanite, ilmenite, etc.

The agpaitic rocks are characterized by extremely high contents of rare elements such as Li, Be, Nb, Ta, REE, Zr, Th, etc. and of volatiles, first of all F and Cl. This gives rise to a wealth of mineral species, more than 500 in the Lovozero and Khibina complexes of the Kola peninsula, about 250 in Mont Saint-Hilaire, Quebec, Canada, and about 200 in the type locality, the Ilímaussaq complex, South Greenland.

These rocks have very long melting intervals and solidus temperatures as low as 500 to 400°C. They are accompanied by a gas phase rich in methane and other hydrocarbons and most probably also by sodium-rich fluids as indicated by the presence of minerals such as ussingite (NaAlSi3O8·NaOH) and villiaumite (NaF) and of pegmatites and hydrothermal veins rich in sodium and rare and volatile elements.

Agpaitic nepheline syenites are considered to have been formed by consolidation of melts oversaturated in alkalis, especially sodium, under conditions preventing the volatiles from escaping. These melts have been derived by extreme fractionation processes in alkali basaltic or nephelinitic magmas. The main stage of crystallization of the melts is characterized by minerals such as nepheline (sometimes also sodalite), alkali feldspars, arfvedsonite, aegirine and eudialyte, but the most highly developed, hyperagpaitic lujavrites of the Ilímaussaq complex have been formed from melts with extreme concentrations of sodium and volatiles resulting in the formation of naujakasite instead of nepheline, ussingite instead of sodalite and alkali feldspars, and steenstrupine instead of eudialyte. During the late stages of crystallization, sodium-rich fluids are the cause of late- and postmagmatic alteration and of the formation of hydrothermal mineralizations. The late stages are characterized by water-soluble sodium-rich minerals of which more than 80 have been found in the Khibina and Lovozero complexes.

Several quartzite bodies outcrop along the Ferreira-Ficalho Thrust Fault (South Portugal), a major accident of the Iberian Variscan Orogen. The sediment is a very pure quartz sandstone, with trace amounts of ultra-resistant heavy minerals and chromite. Chemical characterization (microprobe analyses and Mössbauer spectroscopy) showed the chromite to be unique: besides being Zn-rich, complexly zoned and a cation deficient spinel, all the iron was found to be fully oxidized to Fe3+. Structure refinement of single-crystal X-ray diffraction intensities unambiguously identifies the mineral as a chromite and the Mössbauer data are consistent with tetrahedrally coordinated Fe3+ in the spinel structure. Current geodynamical models see the Ferreira-Ficalho Thrust Fault as a first-order suture resulting from a complex collision of two distinct continental blocks with partial obduction of the intervening oceanic crust. The chromite grains could be envisaged as remnants of an early erosion of this obducted oceanic crust, but its unique chemical character does not allow any definite conclusion. Yet, the complete quartzite heavy mineral contents and its petrographic features are not consistent with their deposition within a continental collision situation.

The weathering of the Mt. Prinzera serpentinites (Parma province, Northern Apennines, Italy) produced dominant smectite and minor Fe-hydroxides. The mobility of the elements during weathering has been estimated using the ratio Ki = MiS/MiR, where Mi indicates the mass of a generic component i before (R) and after (S) the weathering, and using TiO2 as a practically immobile component. For prevalently to tendentially mobile elements, the degree of mobility, which in our case increases as Ki decreases, is in the order Mn ≈ Cr = Fe < V ≈ Zn ≈ Co < Ni < Si < Mg < Ca. The elements Ti, Ga, Al and Zr are prevalently immobile. The mass chemically removed by the circulating waters during weathering may reach very high values (on average 52% of the original mass of serpentinite) and the main contribution to the mass loss is due to Si and Mg. The composition of the perennial and ephemeral springs in the area is in agreement with the degree of the element mobility at least for Cr, Mg and Ca.

This paper is based on microprobe analysis (Ta, Nb, Fe, Mn, Ti, W, Sn) and Fe3+ EPR spectra of cassiterite samples from 14 tin deposits from the Iberian Hercynian Massif. The interpretation of microprobe data, atomic ratios (Fe+Mn)/(Nb+Ta), and Fe3+ EPR spectra reveal that the coupled substitution: 3Sn4+ = (Fe,Mn)2+ + 2(Nb,Ta)5+ is dominant in cassiterites from pegmatites, and high-temperature quartz veins and in the dark zones of cassiterites.

The mechanisms:

2Sn4+ = Fe3+ + (Nb,Ta)5+ and

Sn4+ + O2− = Fe3+ + OH−

are dominant in those from medium-low temperature quartz veins and in the light zones of cassiterites.

The formation of sulphide and cassiterite-bearing chlorite-rich greisens in the Navalcubilla granite has been modelled theoretically. Numerical simulation on the reaction of a hydrothermal fluid with a granitic rock predicts assemblages very similar to those found in nature, with progressive formation of muscovite, quartz, chlorite, microcline and plagioclase zones. The hydrothermal alteration of the rock produces a neutralization of the inflowing acid fluid, a drop in the fS2 and, to a lesser degree, an increment in fO2. During hydrothermal alteration, fS2 and fO2 change abruptly between metasomatic zones, but chlorite seems to control their major changes. Scheelite and cassiterite are concentrated in the internal zones, while sulphides are related to the more external zones. Fluid-rock reactions seem to be very effective for precipitating cassiterite and scheelite, even from very Sn and W-poor fluids. Appreciable amounts of sulphides are only expected in systems with high concentrations of base metals. Boiling and simple cooling of the fluids acidifies and oxidizes them but chemical changes are not strong enough to induce significant precipitation of ore minerals, at least when the temperature changes are small. Continued circulation of fluids along fractures with previously precipitated quartz + wolframite produces replacement of wolframite by scheelite and sulphides.

In contrast with igneous and metamorphic rocks, classical petrochemical calculation methods cannot be used for tropical weathering components (saprolite, bauxitic, ferruginous, siliceous and calcareous laterite) in converting whole-rock chemical analyses into normative mineralogical weight percentages. Weathering profiles are characterized by a mixture of primary and secondary minerals, which are not considered in the classical methods of mineralogical norm calculation. A new petrochemical calculation algorithm is proposed for the conversion of whole-rock chemical analyses into weathering norm (WN) for several components of the tropical weathering profiles. The normative minerals are represented by three primary minerals, six secondary minerals, four primary/secondary mineral pairs, and five minerals which can have both primary and secondary origin. This algorithm has been used in MINNOR, a WINDOWS application written in Visual Basic, which calculates the mineralogical norm. In order to test the program, several types of different chemical weathering profiles from South America and Africa have been selected. Special attention is paid to the weathering profile from Omai, Guyana, South America.

The Qassiarsuk (formerly spelled Qagssiarssuk) complex is located in a roughly E–W trending graben structure between Qassiarsuk village and Tasiusaq settlement in the northern part of the Precambrian Gardar rift, South Greenland. The complex comprises a sequence of alkaline silicate tuffs and extrusive carbonatites interlayered with sandstones, and their subvolcanic equivalents, which represent possible feeders for the extrusive rocks. The Rb-Sr, Sm-Nd and Pb isotopic characteristics of 65 samples of extrusive carbonatite- and silicate tuffs and carbonatite diatremes have been determined by mass spectrometry. The Qassiarsuk complex can be dated to c. 1.2 Ga by Rb-Sr and Pb-Pb isochrons on whole-rocks and mineral separates, agreeing with previous isotopic ages for the volcanic rocks of the Eriksfjord formation in the Eriksfjord area of the Gardar rift, but not with previous, indirect age estimates of >1.31 Ga for assumed Eriksfjord equivalents in the Motzfeldt area further east. Recalculated isotopic compositions at 1.2 Ga indicate that the Qassiarsuk carbonatite- and alkaline-silicate magmas were comagmatic and derived from a depleted mantle source (εNd>4, εSr<−13, time-integrated, single- stage 238U/204Pb ≤ 7.4). The mantle-derived magmas were contaminated with crustal material, equivalent to the local, pre-Gardar granites and gneisses and sediments derived from these. The crustal component has a depleted mantle Nd model age of 2.1-2.6 Ga; at 1.2 Ga it was characterized by εSr = +76, εNd = −8.4, time-integrated, single- stage 238U/204Pb = 8.2−8.3. Strong decoupling of the Pb from the Sr and Nd isotopic systems suggests that the contamination happened only after carbonatitic and alkaline-silicate magmas had evolved from a common parent, by processes such as liquid immisicibility and/or fractional crystallization. Post-magmatic hydrothermal alteration (oxidation, hydration of mafic silicates, carbonatization of melilite) may have contributed further to the contamination of the carbonatite and alkaline silicate rocks of the Qassiarsuk complex.

The dacite of El Joyazo contains abundant metapelitic xenoliths. These can be divided into two main types: garnet-biotite-sillimanite and spinel-cordierite xenoliths. In the xenoliths the widespread occurrence of rhyolitic glass as interstitial films, foliation-parallel layers and primary melt inclusions in all mineral phases indicates that these assemblages developed in the presence of a melt phase, i.e. during anatexis. The composition of the interstitial glass is comparable to that of the melt inclusions, suggesting that melt was locally produced. Phase equilibria indicate that anatexis occurred at P-T conditions of 5–7 kbar and 850±50°C.

Several microstructural lines of evidence show that melt extraction was assisted by deformation during foliation development, and that on the scale of the xenoliths (up to 50 cm) melt escaped mainly by flow along foliation planes. The development of a syn-anatectic foliation also suggests that metapelitic rocks were involved in high-grade metamorphism and partial melting prior to fragmentation and dispersion in the host dacite.

Mass balance calculations, based on the chemical composition of interstitial glass and melt inclusions in minerals, the bulk xenoliths and representative samples of potential pelitic sources support a model wherein the xenoliths represent restites after the extraction of 30 to 55 wt.% melt from graphitic metapelite protoliths similar to the rocks constituting the surrounding Alpujarride metamorphic complex.

Porphyritic natrocarbonatite lavas erupted from the Oldoinyo Lengai volcano (Tanzania) on 17 October 1995 and 15–19 December 1995 differ from previously studied lavas in that they preserve textures indicative of groundmass carbonate-carbonate immiscibility. The immiscible fractions are considered to involve: a Na-K-Ca-CO2-Cl-rich, F-bearing fluid crystallizing gregoryite, sodian sylvite, potassium neighborite as well as a complex Ba-rich carbonate; and a Na-rich, Cl-poor carbonate liquid approximating to a nyerereite-gregoryite cotectic composition. Compositional data are given for potassium neighborite, this mineral being the first recognized occurrence of a fluorine-based perovskite group mineral in a magmatic environment. New compositional data are also given for a previously recognized potassium iron sulphide which indicate that this phase is probably a solid solution between the ternary sulphides, KFe3S4, K2Fe3S4, and KFe2S3. Textural and paragenetic data are interpreted to suggest that these recent lavas are more evolved than previously investigated Oldoinyo Lengai lavas and that natrocarbonatite is a highly evolved rather than a primitive magma.

Platinum-group minerals (PGM) from placer deposits in Colombia, California, Oregon and Alaska were investigated with the electron microprobe, proton microprobe (μ-PIXE) and ion probe to analyse their major and trace element contents and 187Os/186Os isotopic ratios. Most of the grains in the samples investigated proved to be essentially homogeneous alloys of Pt-Fe and Os-Ir-Ru although a few of them contained inclusions of other PGM such as cooperite and laurite. Detailed analyses were undertaken on the Os-Ir-Ru alloy phases.

The 187Os/186Os isotope ratios fell into a range from 1.005 to 1.156 and are consistent with data published on PGM from other placer deposits from these regions. The ratios, together with the trace element data (and in particular the low rhenium content) determined by ion probe and μ-PIXE, indicate that crustal osmium was not incorporated in the grains and that no significant evolution of the 187Os/186Os ratios occurred during their history. These data, along with mineralogical and textural evidence, are consistent with a mantle origin for the grains through ultramafic intrusions, although the data do not entirely rule out alternative interpretations.

The Grønnedal-Íka ring complex (1299 ± 17 Ma) in the Gardar province, South Greenland is composed of a range of layered nepheline syenites which were intruded at a late stage by xenolithic syenite and a plug of carbonatite. The complex was subsequently intruded by a variety of basic dykes, including olivine dolerites, kersantites, vogesites, spessartites, camptonites and an alnöite, and then extensively faulted. The nepheline syenite magmas, produced by fractional crystallisation of basic magmas, show a range in δ13C (−3.86 to −7.57‰) and δ18O (8.27 to 15.12‰), distinctly different to the carbonatites which form a tight group with average δ13C = −4.31 + 0.22 ‰, (1 s.d.) and average δ18O = 7.18 ± 0.41‰ (1 s.d.). Initial 87Sr/86Sr isotope ratios (typically 0.703) suggest the syenites and carbonatites have not assimilated crustal rocks, and therefore the C and O isotope variation within each group is a result of isotopic evolution during fractional crystallisation. A suite of lamprophyre dykes (δ13C −3.86 to −7.86‰ and δ18O 9.12 to 10.81‰) form a coherent group whose stable isotope compositions overlap part of the syenite field, and again are distinctly different from the carbonatites. A single alnöite has δ13C = −3.32‰ and δ18O = 12.34‰ C and O isotope ratios are consistent with origins of syenitic and lamprophyric magmas from a similar source. Despite geochemical evidence which suggests a genetic link between nepheline syenites and carbonatites, C and O isotopic evidence shows that they are not related directly by liquid immiscibility. Comparisons are made between similar rock types from Grønnedal-Íka and from the Gardar Igaliko Dyke Swarm. The possible role of F in controlling δ13C and δ18O during crystallisation of calcite from carbonatite magmas is discussed.