Search

Ultramafic xenoliths are rarely found at convergent plate margins. A notable exception is in the Betic Cordillera of southern Spain, where the eruption of xenolith-bearing alkaline basalts during the Pliocene post-dated the Cenozoic phase of plate convergence and subduction-related magmatism. Mantle xenoliths of the monogenetic volcano of Tallante display extreme compositional heterogeneities, plausibly related to multiple tectono-magmatic episodes that affected the area. This study focuses on two peculiar composite mantle xenolith samples from Tallante, where mantle peridotite is crosscut by felsic veins of different size and mineralogy, including quartz, orthopyroxene, and plagioclase. The veins are separated from the peridotite matrix by an orthopyroxene-rich reaction zone, indicating that the causative agents were alkali-rich hydrous silica-oversaturated melts, which were likely related to recycling of subducted continental crust components. The present study reports new and detailed major and trace elements and Sr-Nd-Pb analyses of the minerals in the composite Tallante xenoliths that confirm the continental crust derivation of the metasomatic melts, and clarifies the mode in which subduction-related components are transferred to the mantle wedge in orogenic areas. The particular REE patterns of the studied minerals, as well as the variation of the isotopic ratios between the different zones of the composite xenoliths, reveal a complex metasomatic process. The distribution of the different elements, and their isotope ratios, in the studied xenoliths are controlled by the mineral phases stabilised by the interaction between the percolating melts and the peridotitic country rock. The persistence of marked isotopic heterogeneities and the lack of re-equilibration suggest that metasomatism of the sub-continental lithospheric mantle occurred shortly before the xenolith exhumation. In this scenario, the studied xenoliths and the metasomatic processes that affected them may be representative of the mantle sources of mafic potassic to ultrapotassic magmas occurring in post-collisional tectonic settings.

The presence of crystal mushes in the feeding system of active volcanoes is generally revealed by antecrysts, representing the crystalline portion of old magmas recycled in the juvenile material, but very little is known about the fate of interstitial liquid hosted within the crystal-rich mush (i.e. antemelt). U-series disequilibria measured in magmas erupted in the past 18 years at Stromboli volcano provide the first geochemical evidence of the involvement of antemelt and help constraining the timescales of the processes occurring in the plumbing system of the volcano. Despite almost constant major and trace element composition, significant variations in isotope ratios are observed. (230Th/232Th) decreases with time, whilst (238U/232Th), (226Ra/230Th) and 87Sr/86Sr are different in the two types of magma erupted. Magma with low phenocryst content (lp) is erupted as pumices during paroxysm and is thought to belong to a deep reservoir. Highly porphyritic magma (hp) is erupted during the normal “Strombolian” activity as scoria and during the effusive events as lavas, and it is considered to derive from the former one within a shallow reservoir through degassing-driven crystallisation, mixing and incorporation of antecrysts. The distinct (238U/232Th) of lp and hp magma requires the involvement of a component with high 87Sr/86Sr and (238U/232Th) deriving from older magmas erupted earlier in the volcano history (up to 2.5 ka). The incompatibility of U and Th in major mineral phases limits the possible effect of antecrysts, hence requiring the involvement of a U- and Th-rich antemelt. The decrease of 226Ra-excess from lp to hp magmas provides further and independent evidence for the involvement of a few thousands years old antemelt. The variation with time of (230Th/232Th) within lp and hp magmas is exploited to constrain the residence time of magmas in the deep and shallow reservoir of the volcano to 55 yrs (inferred reservoir volume 0.5 km 3 ) and 2–10 yrs (inferred reservoir volume 0.02 – 0.09 km 3 ), respectively. Our results show the occurrence of magmatic processes operating at different timescale within the feeding system of a so-called steady state volcano, such as Stromboli. We show that, while most of the magma is erupted within few years, a portion of it, made of both crystals (antecrysts) and residual liquid (antemelt), can be stored in the plumbing system for thousands of years to be eventually rejuvenated and mixed back into the shallow reservoir. The presence of antemelts may also affect the eruptive mechanism by promoting heat transfer in locked crystal mushes, hence favouring their remobilisation.

In the Italian peninsula leucite-free and -bearing ultrapotassic rocks occur intimately associated but well separated in time. Silica-saturated ultrapotassic and associated shoshonitic magmas erupted during the Pliocene to Lower Pleistocene. Silica under-saturated leucite-bearing ultrapotassic rocks, mainly leucitites, were emplaced sometime later, during the Middle–Upper Pleistocene. The transition from leucite-free to -bearing rocks is diachronous and in some cases is complicated by the occurrence of crustal-derived magmas coeval with early leucite-free magma. The Monte Cimino–Vico area is a key locality for studying the transition from leucite-free to -bearing ultrapotassic rocks. There a volcanic complex is overlain by some 500 ka younger leucite-bearing rocks of Vico stratovolcano. Most of the potassic and ultrapotassic rocks of the Monte Cimino volcanic complex are characterised by high Mg-# (66.4–77.8), and compatible element abundances in spite of their high silica contents (52.9–58.0 wt.%). Ultrapotassic leucite-free rocks transitional to lamproites are olivine-bearing orthopyroxene- and plagioclase-free latites and K-rich basaltic trachyandesite and they are confined in the final mafic activity of the Monte Cimino volcanic complex. Two-pyroxene parageneses are observed in the early Monte Cimino volcanic complex giving equilibration temperature as high as 1050 °C and suggesting a sub-crustal derivation of their parental magmas. A two-pyroxene high-K calc-alkaline magma is then inferred for the genesis of the early Monte Cimino activity (e.g. dome complex and silicic lava flows and ignimbrites), which was modelled through crystal-fractionation from a high-K basaltic andesitic magma plus mixing with crustal anatectic rhyolitic magma. Final mafic lavas fill the isotopic and chemical gap between shoshonite-like and lamproite-like rocks, and each lava flow could be considered as a mantle-derived term from a heterogeneous sub-continental lithospheric mantle. Final mafic lavas are high-silica and -MgO ultrapotassic rocks due to crystallisation of primitive magmas generated by partial melting of a vein metasomatised network within a lithospheric upper mantle source, at low pressure and high P H2O . The shift to the younger Vico volcano, in which leucite-bearing magmas prevail, is thought to be due to the arrival in the mantle source of newly formed metasomatic agent from the undergoing slab characterised by the increasing of a sedimentary carbonate recycled component. This new component appears in the Middle Pleistocene and it is responsible for the composition of the magmatism of the Roman region. The recycling of carbonate-bearing pelitic sediments within the mantle wedge produces a carbonate-rich metasomatic agent. The latter is responsible for stabilisation of phlogopite-bearing wehrlitic veins within the mantle wedge. Partial melting of this newly formed vein network under high X CO2 generated silica under-saturated, and then leucite-bearing, ultrapotassic primary melts. A further shift is observed among the Vico post-caldera magmas, in which the appearance of a deep asthenospheric component is argued to be channelled late through slab-tears.