Magmatic rocks from the Dolomites, Carnic and Julian Alps, Italy, have been sampled to investigate the origin and geodynamic setting of Triassic magmatism in the Southern Alps. Basaltic, gabbroic and lamprophyric samples have been characterized for their petrography, mineral chemistry, whole-rock major and trace elements, and Sr, Nd and Pb isotopic compositions. Geothermobarometric estimates suggest that the basaltic magmas crystallized mostly at depths of 14–20 km. Isotopic data show variable degrees of crustal contamination decreasing westwards, probably reflecting a progressively more restitic nature of the crust, which has been variably affected by melting during the Permian period. Geochemical and isotopic data suggest that the mantle source was metasomatized by slab-derived fluids. In agreement with previous studies and based on geological evidence, we argue that this metasomatism was not contemporaneous with the Ladinian–Carnian magmatism but was related to previous subduction episodes. The lamprophyres, which likely originated some 20 Ma later by lower degrees of melting and at higher pressures with respect to the basaltic suite, suggest that the mantle source regions of Triassic magmatism in the Dolomites was both laterally and vertically heterogeneous. We conclude that the orogenic signatures of the magmas do not imply any coeval subduction in the surrounding of Adria. We rather suggest that this magmatism is related to the Triassic rifting episodes that affected the western Mediterranean region and that were ultimately connected to the rifting events that caused the break-up of Pangea during the Late Triassic – Early Jurassic period.

Curie temperatures indicating non-titaniferous magnetite are common in Icelandic basalts of all ages, especially Tertiary ones. Yet, microprobe analyses of such samples have shown high titanium in the magnetite. To resolve this paradox, and the mechanism at work, the magnetic mineral fraction of eight basalt samples with Js-T curves characteristic for pure magnetite was subjected to a multi-disciplinary analysis including Mössbauer spectroscopy and X-ray diffraction. In most of the samples titanium in the magnetite, as analysed with the microprobe, ranged between 16 and 28 wt.%, indicating submicroscopic solvus exsolution in the titanomagnetite, beyond the power of resolution for the microprobe. More unexpectedly in view of the reversible Js-T curves, Mössbauer spectroscopy showed appreciable proportion of maghemite in the magnetic fraction. A three-stage mechanism is proposed for the formation of the mineral assemblages observed: (1) limited high-temperature oxyexsolution; (2) solvus exsolution during low-temperature hydrothermal alteration; and (3) maghemitization of the magnetite. Finally, the maghemite may transform to hematite with time. It is concluded that maghemite is much more common in Icelandic rocks than hitherto believed.

Ultramafic xenoliths enclosed in Plio-Quaternary alkali basalts from Tallante near Cartagne (southern Spain) are composed mainly of spinel lherzolites which are probably upper mantle residues. In many xenoliths, the spinel lherzolite is cut by pyroxenite or gabbroic anorthosite veinlets generally 0.2–3 cm thick. The clinopyroxenite veinlets were formed by high-pressure crystal-liquid segregation from alkali basalt magmas formed earlier than the host basalts, whereas mantle metasomatism played a role in the genesis of gabbroic anorthosites. Close to the contact with the veinlets, the spinel lherzolites are enriched in Ca, Fe, and some incompatible elements including light REE due to the migration of a fluid from the veinlets into the surrounding lherzolites. The host alkali basalts were derived from a heterogeneous, incompatible element-enriched upper-mantle source probably similar in composition and nature to the composite xenoliths, but were formed in a garnet stability field.

Corundum xenocrysts from alkaline basalt fields differ in characteristics and hence lithospheric origins. Trace element, fluid/melt inclusion and oxygen isotope studies on two eastern Australian corundum deposits are compared to consider their origins. Sapphires from Weldborough, NE Tasmania, are magmatic (high-Ga, av. 200 ppm) and dominated by Fe (av. 3300 ppm) and variable Ti (av. 400 ppm) as chromophores. They contain Cl, Fe, Ga, Ti and CO2-rich fluid inclusions and give δ18O values (5.1–6.2‰) of mantle range. Geochronology on companion zircons suggests several sources (from 290 Ma to 47 Ma) were disrupted by basaltic melts (47 ± 0.6 Ma). Gem corundums from Barrington, New South Wales, also include magmatic sapphires (Ga av. 170 ppm; δ18O (4.6–5.8‰), but with more Fe (av. 9000 ppm) and less Ti (av. 300 ppm) as chromophores. Zircon dating suggests that gem formation preceded and was overlapped by Cenozoic basaltic melt generation (59–4 Ma). In contrast, a metamorphic sapphire-ruby suite (low-Ga, av. 30 ppm) here incorporates greater Cr into the chromophores (up to 2250 ppm). Fluid inclusions are CO2-poor, but melt inclusions suggest some alkaline melt interaction. The δ18O values (5.1–6.2‰) overlap magmatic sapphire values. Interactions at contact zones (T = 780–940°C) between earlier Permian ultramafic bodies and later alkaline fluid activity may explain the formation of rubies.

Some picritic lavas from near the base of the volcanic succession in the Lebombo region of the Karoo Province contain phenocrysts of olivine and skeletal microphenocrysts of ilmenite, the latter indicating rapid crystallization from the magma. These ilmenite grains contain up to 6% MgO. The near-liquidus crystallization and the high MgO content of the ilmenite are highly anomalous for ilmenite in tholeiitic rocks. These features reflect the atypical nature (high MgO and TiO2) of the parental magma and/or highly reduced conditions of crystallization.

The spinel peridotite inclusions in basalts from Sardinia are upper-mantle residues affected by metasomatism which led to an enrichment particularly of U and light REE. The metasomatism took place prior to the recrystallization which produced the primary mineral assemblage of the inclusions. The compositional variations imply that the xenoliths are residual after at least two melting events.

Partially melted basalts enclosing amygdales which have been completely melted formed at Scawt Hill adjacent to a Tertiary dolerite plug. Melting of the basalts commenced in a clay-rich mesostasis to produce a feldspathic liquid which then crystallized to an assemblage of dendritic olivine, skeletal hypersthene, opaque oxide and Mg-hercynite in a microcrystalline plagioclase matrix. An original mineral assemblage of zeolite, calcite, and saponite-nontronite in the amygdales melted and quenched to a brown glass now containing complexly zoned pyroxenes with plagioclase and opaque oxide. Melting commenced between 700–800°C, reaching a maximum temperature of 1168°C, and was followed by rapid cooling. The assimilation of remelted basalt may alter the course of crystallization of contaminated magmas.

Amphiboles in dykes from the dominantly silica-undersaturated Monteregian series range from pargasitic megacrysts and xenocrysts to kaersutitic, pargasitic and hastingsitic phenocrysts, groundmass prisms and reaction rims. Amphiboles in dykes and plutons from the silica-oversaturated White Mountain Magma Series range from kaersutite, through hornblende, hastingsite and edenite, to sodic-calcic and sodic varieties. This contrast between the amphiboles from the two series is probably a reflection of differing melt silica activity and is a useful petrologic discriminant. In most cases, pargasitic amphibole megacrysts from Monteregian monchiquites reflect the Mg numbers of their host rocks and are considered cognate. The megacrysts are lower in Ti and higher in Mg and Al than their phenocryst mantles and rims. This is probably a result of higher pressures of formation.

Clinopyroxene phenocrysts and groundmass crystals are relict phases in altered basalt and basaltic andesite lavas, and arenites of the Cudgegong Volcanics and Toolamanang Volcanics, Cudgegong-Mudgee district, New South Wales. Petrography, field relationships and clinopyroxene compositions indicate that basaltic blocks in the latter unit are reworked from the Cudgegong Volcanics. Clinopyroxene phenocrysts show a restricted compositional range and minor Feenrichment from core to rim, features considered indicative of a calc-alkaline parent magma. It is proposed that the Cudgegong Volcanics crystallized under hydrous conditions, at least in the later stages, with rising fO2 resulting in a Fe-Ti oxide crystallizing as a primary phase. The clinopyroxenes are considered to have crystallized at moderate (5–6 kbar) and falling pressures and at minimum temperatures in the range 900 to 1000°C. Coupled substitutions affecting the “other” components in the clinopyroxene structural formula indicate that the ivAl-viFe3+, ivAl-viAl and ivAl-viTi4+ couples are important. The Sofala Volcanics, south of the study area, and the Cudgegong Volcanics are similar in age, petrography and stratigraphic position, and contain relict clinopyroxenes which are chemically similar. This suggests that the units are laterally equivalent and adds further evidence to the proposal that an oceanic island arc system was active in central western New South Wales during the Late Ordovician.

The Madagascar Cretaceous igneous province exposed in the Mahajanga basin is represented by basalt and basaltic andesite lavas. New 40Ar–39Ar plateau ages (92.3 ± 2.0 Ma and 91.5 ± 1.3 Ma) indicate that the magmatism in the Mahajanga basin started about 92 Ma ago. Four geochemically distinct magma types (Groups A–D) are present. Group A and C rocks have low to moderate TiO2 (1.2–2.6 wt%), Nb (3–9 μg g−1) and Zr (82–200 μg g−1), and show large variations in ɛNdi (+0.1 to −10.8), 206Pb/204Pb (15.28 to 16.33) and γOs (+11.4 to +7378). The large isotopic variations, particularly in Os, Nd and Pb isotopic compositions, are likely due to crustal contamination. The low Pb isotope ratios observed in the Group A and C rocks suggest involvement of continental crust with low μ (238U/204Pb). Group B and D rocks have moderate to high TiO2 (2.2–4.9 wt%), Nb (8–24 μg g−1) and Zr (120–327 μg g−1). Age-corrected isotopes of Group B and D lavas show a small range in ɛNdi (+1.0 to +4.0) and a wide range in γOs (+128 to +1182). Values of 207Pb/204Pb are within the range for Groups A and C, but the Group D 206Pb/204Pb (16.52–17.08) and 208Pb/204Pb (37.51–38.01) values are higher, indicating a different crustal contaminant. Pb isotopic values of the Group B rocks seem to reflect the isotopic features of their mantle source. The magma groups of Mahajanga display a wide range of trace element and isotopic compositions that cannot be explained only by open-system crystallization processes but, rather, by distinct mantle sources.

Size distribution data obtained from detailed field study of bubbles and amethyst-geodes in the basaltic lava flows of the Serra Geral Formation (Ametista do Sul, Rio Grande do Sul, Brazil) are used in cooling and vesiculation models to infer the origin and the formation processes of the geodic cavities. Coupled field observations and modelling results emphasize that (1) the formation of geodes in the studied lava flow can be explained, qualitatively and quantitatively, by the exsolution of magmatic gas from the supersaturated melt with no need for external surface water supply; (2) the vertically elongated habits of the geodes result from higher cooling rate of the magma in contact with the accumulating bubbles; and (3) the abnormal metre-sized geodes with their branching habits result not only from the diffusive/decompressive bubble growth but also from the coalescence of inwards-progressing tubular cavities.

The Nabwal Hills, northeast of Lake Turkana, contain a record of magmatism associated with the initiation and early development of the East African Rift System in northernmost Kenya. The predominantly volcanic Asille Group, 1400 m thick, directly overlies metamorphic basement and comprises a sequence of basaltic lava flows with significant intervals of rhyolitic pyroclastic units, and minor intercalations of fluviatile sediments. The basement gneisses yield K–Ar cooling ages on biotite of 510 and 522 Ma, typically Pan-African. The 40Ar–39Ar ages on alkali feldspar crystals from the rhyolitic units are concordant and show that the Asille Group spans an interval from at least 34.3 to 15.8 Ma, continuing to at least as young as 13 Ma based on previous measurements. Vertebrate fossil sites, containing primate remains, at Irile and Nabwal are shown to be 17 ± 2 Ma old, Early Miocene, based upon K–Ar age measurements on immediately overlying basalts. Variably reliable whole rock K–Ar ages, determined on basalt samples from low in the sequence, indicate that volcanism commenced as early as 34.8 Ma ago. The overall geochronological results show that magmatism in the Nabwal Hills began about 35 Ma ago in Late Eocene times, interpreted as the time of initiation of crustal extension that led to the development of this segment of the East African Rift System. The Asille Group is tilted about 6° to the SSW. This tilting occurred later than 13 Ma ago, and prior to the eruption of the flat-lying Gombe Group basalts. These basalts may have begun erupting about 6 Ma ago in Late Miocene times, although much of this volcanism occurred between about 3.9 and 4.2 Ma ago in Pliocene times. It is suggested that the main rifting, which continues today, commenced in Late Miocene times, less than 13 Ma ago, and is partly reflected in the tilting of the Asille Group.

A little before 4 Ma ago, deposition of Pliocene and Pleistocene strata described as the Omo Group began in the Turkana and Omo basins of northern Kenya and southern Ethiopia. Soon after, basaltic magma erupted as thin lava flows, and intruded as dykes into the oldest Pliocene strata of the basin. These flows and intrusions are similar petrographically and geochemically, and mark a basaltic magmatic event spanning latitudes from 2° 45′ N to 6° 45′ N at a longitude of about 36° E. By 3.94 Ma, this basaltic magmatic activity had ceased. Previous researchers used these lavas as an important seismic marker in their study of the southern part of the Turkana Basin. Subsequent volcanic eruptions formed North, Central and South islands in Lake Turkana, and the Korath Range in southern Ethiopia. Thus there was a hiatus in basaltic magmatic activity of nearly 4 Ma in the area presently occupied by Lake Turkana and the lower Omo Valley, although volcanism continued on the eastern margin of the basin. Here we review the field occurrences of these basalts, their distinctive petrography, composition, age and significance to Pliocene deposition in the basin.

Major and trace element data on Tertiary continental basalt flows from the Shalatein area, Red Sea coastal plain of Egypt, have been presented and used to obtain more information about their source region and the processes involved in their generation. The rocks are mainly alkali olivine basalt with MgO and Mg no. in the range of 9.8–5 wt % and 65–46, respectively. They display wide variations in incompatible element concentrations, particularly LREE, Zr, Nb, K, Y, Ba and Sr. There is no evidence of significant crustal contamination or a lithospheric mantle signature in these rocks. Normalized trace element patterns and diagnostic elemental ratios are very similar to those of modern ocean-island basalts (OIB) a feature which suggests that the mantle source region was the asthenosphere. Comparison with the different types of OIB indicates that the basalts may be derived from a high U/Pb (HIMU) source with slightly elevated K and Ba contents. The basalts show general trends of increasing incompatible elements (K2O, Nb, Y, Sr and Yb), and decreasing contents of compatible elements (Cr, Ni, Sc and Ca) with decreasing Mg no. Furthermore, TiO2, P2O5, LREE and Th define maxima at about Mg no.=56, suggesting late fractionation of Fe–Ti oxides and apatite. Although these variations are consistent with fractional crystallization processes, the wide variations in LREE contents and the incompatible trace element ratios Ce/Y (1.2–3.8), Zr/Nb (2.3–7.1) and Nb/Y (0.6–4) in the least fractionated samples (Mg no. > 56) suggest that fractional crystallization involving the observed phenocryst assemblage (olivine and clinopyroxene) cannot fully explain such compositional variations. Modelling of the mafic rocks (Mg no. > 56) using REE suggests varying degrees of partial melting of an enriched mantle source region in the garnet stability field. Partial melting is attributed to plume-related mantle upwelling beneath the Red Sea rift system.