Search

The South Aegean volcanic arc consists of five volcanic fields, with products that range from medium- and high-K calc-alkaline basalts to rhyolites. Parental magmas are generated by variable proportions of decompression and flux melting of a mantle source metasomatized by sediment melts and aqueous fluids released from the subducted slab. Fluid/sediment ratios are lowest in Santorini (Greece) where high lithospheric extension results in a predominance of decompression melting, shallower magma storage, and more mafic volcanism than elsewhere in the arc. Contributions from slab sediment melt decrease from west to east. With the lowest convergence rate and surface heat flux of any continental arc worldwide, the South Aegean is an ideal natural laboratory for studying arc magmatism at low magma production rates.

Kolumbo is the largest of twenty submarine volcanic cones, tectonically aligned in the transtentional Anydros basin, one of the most seismically active zones in the South Aegean Volcanic Arc, whose magmatism is related to the subduction of the African Plate beneath the Aegean microplate. Kolumbo explosively erupted in 1650 CE, causing the death of 70 people on Santorini, which is only 7 km SW of Kolumbo. Explorative cruises employing ROVs discovered a high temperature (220°C) hydrothermal field with CO2-rich discharges and accumulation of acidic water at the bottom of the crater (505 m b.s.l.), increasing the related hazard. A possible magma chamber was recognized below the crater at depth 9-6 km by seismic data [Dimitriadis et al. 2009]. Geochemical data [Klaver et al. 2016] suggest that Kolumbo have a different mantle source and storage system from Santorini. It is fundamental to understand the behaviour of this volcano, and how its storage and plumbing system works, to correctly assess risk for nearby islands.We present petrographic, geochemical and isotopic data of samples collected during the cruises and by divers. Most samples represent the juvenile products of the 1650 CE activity, characterizing different magmas interacting before the eruption. They consist of white rhyolitic pumices with grey and black bands, also including basaltic-andesitic enclaves. Plagioclase, biotite, pyroxenes are the main mineral phases; olivine is found in the mafic enclaves. Minerals show quite complex zoning and a large compositional variability. Fresh lithic lavas were sampled; they also have amphibole and can be subdivided in three groups with distinctive petrographic textures that are well reflected in their different chemical compositions. They give information on the early history of the volcano and on how the rhyolitic magma could have been generated.Our data suggest the presence of a complex storage system where the most evolved magma differentiated by assimilation and fractional crystallization, undergoing several inputs of mafic magmas. Early batches of new melts initially mixed with the resident ones, whereas later arrivals only mingled with the rhyolitic magma, thus possibly representing the final trigger of the eruption.

Processes of crystal-mush remobilization by mafic magma recharges are often related to the outpouring of large volumes of silicic melt during caldera-forming eruptions. This occurred for the Campanian Ignimbrite (CI) eruption (Campi Flegrei, Italy), which produced a voluminous trachy-phonolitic ignimbrite in southern-central Italy about 40 ka ago. We focussed on the proximal-CI deposits at San Martino that are composed of a main sequence of early-erupted, crystal-poor units and a late-erupted (post-caldera collapse) crystal-rich Upper Pumice Flow Unit (UPFU). Detailed micro-analytical geochemical data were performed on glasses and crystals of pyroclasts from these deposits and coupled with Sr-Nd isotopic measurements on glasses. Results show that the CI eruption was fed by two distinct melts for the early-erupted units and the late UPFU, respectively. The glasses of the early-erupted units have negative Eu anomalies and show more evolved compositions and higher Nd isotope ratios than those of the UPFU, which have positive Eu/Eu*. The magmas of the early units formed the main volume of eruptible melt of the CI reservoir, and are interpreted as having been extracted from cumulate crystal-mush without a vertical geochemical gradient within the magma reservoir. The data indicate that the generation of the distinctive UPFU melts involved the injection of a new batch of mafic magma into the base of the CI reservoir. The mafic magma allowed heating and reactivation of the CI crystal-mush by melting of low-Or sanidines (+/− low-An plagioclases), leaving high-An plagioclases and high-Mg# clinopyroxenes as residual phases and a crystal-mush melt, made of 20% of the initial mush interstitial melt (with a composition similar to the early-erupted units) and 80% of sanidine melt. When the mush crystallinity was sufficiently reduced, the mafic magma was able to penetrate into the reactivated crystal-mush, mixing with variable proportions of crystal-mush melt and generating cooler hybrid melts, which underwent further crystallization of high-Or sanidine at variable degrees (10–25%). Finally, possibly a short time before the eruption, the UPFU magmas were able to mix and mingle with the crystal-poor eruptible melts still persisting in the CI reservoir at the time of UPFU emission. We suggest that the complex mechanisms described for the magma evolution feeding the CI eruption may occur whenever a crystal-mush is reactivated by new mafic magma inputs

The complex processes occurring in the initial phases of an eruption are often recorded in the products of its opening stage, which are usually characterized by small volume and limited dispersal, and thus generally poorly studied. The 2010 eruption of Eyjafjallajokull (Iceland) represents a unique opportunity for these investigations thanks to the good preservation of tephra deposits within the ice/snow pack. A detailed geochemical investigation on the glassy groundmass of single ash clasts disclosed a population of fragments with unusual high 87 Sr/ 86 Sr (up to 0.70668) for Icelandic magmatism, and anomalous elemental composition with respect to most of the juvenile material of the eruption. This suggests that during its rise, before intruding into the ice cover, magma at a dyke tip selectively assimilated hydrothermal minerals with seawater-related, high-Sr isotopic ratios (zeolites, silica phases, anhydrite) hosted in altered volcanic/epiclastic rocks. According to the observed precursory seismicity, only restricted to few hours before the onset of the eruption, this process could have accompanied subcritical aseismic fracture opening during the days before the eruption, possibly related to stress corrosion-cracking processes, which enhanced the partial dissolution/melting and subsequent selective assimilation of the host rocks.

Monte Amiata is a small volcano composed by trachytic to olivine latitic lava flows and domes emplaced in a very short time between 305 and 231 ka. The main petrographic features are represented by the occurrence of i) abundant rounded magmatic enclaves increasing in dimension and quantity passing from early to late erupted Monte Amiata volcanic rocks, ii) large sanidine megacrysts, mainly confined in the second stage of activity characterised by the emplacement of exogenous domes and massive lava flows, and iii) mafic olivine latitic lava flows, with intermediate compositions between the early silica-rich volcanic rocks and the most mafic rounded magmatic enclaves hosted by the Monte Amiata volcanic rocks. The occurrence of rounded magmatic enclaves testifies fresh magma injection and stirring within a differentiated magma reservoir. This triggered the pouring out of the viscous trachydacitic resident magma. A reverse differentiation pathway is observed with time of magma emplacement, which is accompanied by the decrease of silica contents and increase of MgO and compatible elements passing from early trachydacites to final olivine-latites. The same timely reverse differentiation pathway is observed among magmatic enclaves, with the most mafic terms hosted by final olivine-latitic lava flows. Fine-grained rounded magmatic enclaves, indeed, range in composition from potassic trachybasalt (absarokite) to olivine-latite. The overall geochemical and isotopic features agree with a mixing process between a highly differentiated (i.e., high silica), and partially crystallised, high-K calc-alkaline end- member and a mafic ultrapotassic magma possibly leucite-bearing. Absence of leucite in the Amiata rocks and enclaves is due to high-silica activity of derived magmas caused by the high-silica end-member of the mixing process.

In 2011–2012, Santorini was characterized by seismic-geodetic-geochemical unrest, which was unprecedented since the most-recent eruption occurred in 1950 and led to fear an eruption was imminent. This unrest offered a chance for investigating the processes leading to volcanic reactivation and the compositional characteristics of involved magma. We have thus analyzed the He-Ne-Ar-isotope composition of fluid inclusions in olivines and clinopyroxenes from cumulate mafic enclaves hosted in cogenetic dacitic lavas of the 1570–1573 and 1925–1928 eruptions of Nea Kameni. These unique data on Aegean volcanism were compared with those of gases collected in quiescent periods and during the unrest. The 3He/4He ratios (3.1–4.0 Ra) are significantly lower than the typical arc-volcano values (R/Ra ∼ 7–8), suggesting the occurrence of magma contamination in Santorini plumbing system, which would further modify the 3He/4He ratio of parental magmas generated in the local metasomatized mantle. The 3He/4He values of enclaves (3.1–3.6 Ra) are comparable to those measured in gases during quiescent periods, confirming that enclaves reflect the He-isotope signature of magma residing at shallow depths and feeding passive degassing. A significant increase in soil CO2 flux from Nea Kameni and anomalous compositional variations in the fumaroles were identified during the unrest, accordingly with previous studies. Simultaneously, 3He/4He ratios up to 4.0 Ra were also measured, demonstrating that the unrest was due to the intrusion into the shallow plumbing system of a more-primitive 3He-rich magma, which is even volatile richer and less contaminated than mafic magma erupted as enclaves. This new intrusion did not however trigger an eruption.

Stromboli is famous for its persistent volcanic activity consisting of periodic discrete explosions alternating with lava effusion and more violent explosions. This paper presents a detailed reconstruction of the geological history of Stromboli and description of the characteristics and distribution of the volcanic units and structural features. Six main growth stages (Eruptive Epochs 1–6), in addition to the c. 200 ka activity of Strombolicchio, are recognized between c. 85 ka and the present day, displaying a magma composition ranging from calc-alkaline to potassic series which usually varies with changing Eruptive Epochs. The Epochs are subdivided into sequences of eruptions and characterized by dominant central-vent summit activity with episodic phases of flank activity along fissures and eccentric vents. The activity was repeatedly interrupted by erosional and destructive phases driven by recurrent vertical caldera-type (cc1–5) and sector (and flank) collapses (sc1–7) and generally associated with significant quiescences. The different serial character of the Stromboli rocks is associated with largely variable trace element contents and isotope ratios. These petrochemical characteristics together with our new stratigraphy indicate that magmas, generated in a heterogeneous mantle wedge, underwent complex differentiation processes during their ascent. Magmas are characterized by polybaric evolution residing in small magma reservoirs that are alternatively tapped by the different collapses. DVD: The 10 000 scale geological map of Stromboli is included on the DVD in the printed book and can also be accessed online at http://www.geolsoc.org.uk/Memoir37-electronic. Also included is a full geochemical dataset for Stromboli.

Based on detailed field, petrographic, chemical, and isotopic data, this paper shows that the youngest magmas of the active Nisyros volcano (South Aegean Arc, Greece) are an example of transition from rhyolitic to less evolved magmas by multiple refilling with mafic melts, triggering complex magma interaction processes. The final magmatic activity of Nisyros was characterized by sub-Plinian caldera-forming eruption (40 ka), emplacing the Upper Pumice (UP) rhyolitic deposits, followed by the extrusion of rhyodacitic post-caldera domes (about 31–10 ka). The latter are rich in magmatic enclaves with textural and compositional (basaltic–andesite to andesite) characteristics that reveal they are quenched portions of mafic magmas included in a cooler more evolved melt. Dome-lavas have different chemical, isotopic, and mineralogical characteristics from the enclaves. The latter have lower 87Sr/86Sr and higher 143Nd/144Nd values than dome-lavas. Silica contents and 87Sr/86Sr values decrease with time among dome-lavas and enclaves. Micro-scale mingling processes caused by enclave crumbling and by widespread mineral exchanges increase from the oldest to the youngest domes, together with enclave content. We demonstrate that the dome-lavas are multi-component magmas formed by progressive mingling/mixing processes between a rhyolitic component (post-UP) and the enclave-forming mafic magmas refilling the felsic reservoir (from 15 wt.% to 40 wt.% of mafic component with time). We recognize that only the more evolved enclave magmas contribute to this process, in which recycling of cumulate plagioclase crystals is also involved. The post-UP end-member derives by fractional crystallization from the magmas leftover after the previous UP eruptions. The enclave magma differentiation develops mainly by fractional crystallization associated with multiple mixing with mafic melts changing their composition with time. A time-related picture of the relationships between dome-lavas and relative enclaves is proposed, suggesting a delay between a mafic magma input and the relative dome outpouring. We also infer that the magma viscosity reduction by re-heating allows dome extrusion without explosive activity.

In situ Sr-isotope data by microdrilling, cou- pled with major and trace element analyses, have been performed on plagioclase and clinopyroxene from seven samples collected during the 2002–2003 eruptive crisis at Stromboli volcano (Aeolian Islands, Italy). On 28 December 2002, the persistent moderate explosive activity was broken by an effusive event lasting about 7 months. A more violent explosion (paroxysm) occurred on 5 April 2003. Two magma types were erupted, namely a volatile- poor and highly porphyritic magma (HP-magma) poured out as scoria or lava and a volatile-rich, phenocryst-poor magma (LP-magma) found as pumice. LP-magma differs from the HP-magma also for its slightly less-evolved chemistry, the groundmass composition and the lower Sr- isotope ratios. Micro-Sr-isotope data show the presence of zoned minerals in strong isotope disequilibrium, as previ- ously found in products erupted in 1984, 1985 and 1996 AD, with 87Sr/86Sr values generally decreasing from cores to rims of minerals. Only some outer rims testify for equilibrium with the host groundmass. The internal mineral zones with high Sr-isotope ratios (0.70665–0.70618) are interpreted as ‘antecrysts’, crystallised during the previous activity and recycled in the present-day system since the opening shoshonitic activity of the Recent Period, which occurred at about 2.5 ka ago. This result has implications for the dynamics of the present-day plumbing system of Stromboli at intermediate pressure (about 2–3 km depth) and allows us to propose a model whereby an HP-magma reservoir is directly interconnected at the bottom with a cumulate crystal much reservoir. Efficient mixing between residing HP- and input LP-magmas can occur in this res- ervoir, due to more similar rheological characteristics of the two magmas than in the conduit, where crystallisation is enhanced by degassing. Antecrysts (and possibly melts) re-enter in the HP-magma reservoir both from the bottom, recycled by ascending LP-magmas crossing the crystal mush, and from the top, recycled by descending degassed and dense HP-magma, residual of the periodic Strombolian explosions at the surface. The isotope variation measured in the groundmasses allows calculating the proportion of the LP-magma entering the shallow HP-magma reservoir at *20%. From this proportion, we estimate that the total volume of LP-magma input during 2002–2003 closely matches the magma volume erupted in the effusive event, suggesting a steady-state system at broadly constant vol- ume. The comparison with estimates of the LP-magma volume ejected by the paroxysm indicates that the LP-magma amount directly reaching the surface during the 5 April paroxysm is minimal with respect to that entering the system.

The Secche di Lazzaro (SDL) phreatomagmatic activity, with the associated Neostromboli sector collapse, represents the most powerful activity of the last 6 ky at Stromboli. As revealed by its present-day activity, Stromboli is one of the most eruptive volcanoes in Italy, and flank instability, along its NW flank, is a common process. Volcano instability is often dramatised by explosive eruptions, thus it is of crucial importance to understand the linking between volcano collapse and the plumbing system itself. The possible role of pre-eruptive magmatic processes as triggers of explosive eruptions can be mainly preserved by minerals and revealed by petrochemical studies. We studied the juvenile components (scoria and pumice) of the pyroclastic deposits from the SDL phreatomagmatic activity with the aim to understand the eruption–collapse link. The SDL pyroclastic sequence has been investigated in three different outcrops: at Secche di Lazzaro (SDL-SW sequence), at Vallone del Monaco (SDL-S) and along the old path from Stromboli village to the active craters (SDL-N). Juvenile components are KS-shoshonites similar to the most evolved lavas erupted at the end of the Neostromboli potassic period. Our study indicates that the SDL system is strongly heterogeneous and made up by at least two different sequences (SDL-N and SDL-SW/SDL-S) with different textural characteristics of the deposits and composition of the juvenile components. Mineral chemistry data show a variable presence of recycled crystals (antecrysts), possibly deriving from the previous Neostromboli activity (up to 13 ky ago). Whole rocks are characterised by small but significant differences between the SDL-N and the southern samples, pointing out the presence of compositionally different potassic magmas and the crystallization and recycling of slightly different mineral assemblages. A clear isotope heterogeneity has been also found and explained by the presence of 10% minimum of less radiogenic antecrysts, whereas crustal contamination would have only affected the higher Sr-radiogenic residual melt. These data suggest that the SDL phreatomagmatic eruptions were fed by compositionally distinct potassic magmas, which rested and evolved in two shallow magma conduits when the magma-supply ceased in the waning phase of the Neostromboli activity. Our results also rule out the occurrence of new arrivals of refreshing magmas in the SDL feeding system, as well as other internal magma-related causes, as triggers for the SDL explosive eruptions. Accordingly, external collapse-related causes seem effectively more suitable for triggering the SDL phreatomagmatic eruptions by decompression of the magmatic systems. We also suggest the SDL-N and SDL-SW/SDL-S sequences are generated by two respective explosive events, separated in time and/or space. They can be respectively related to multiple failures of Sciara del Fuoco which successively depressurised the SDL-N and the SDL-SW/SDL-S magma conduits.

In-situ cosmogenic 3He exposure ages of pyroxene phenocrysts from basalts from the Upper Neostromboli formation in southwest Stromboli date its eruption at 7.0 ± 0.3 ka (1σ, n = 3, Ginostra site) and 6.8 ± 0.2 ka (1σ, n = 10, Timpone del Fuoco site) respectively. Correlation of our new data to previous K/Ar and palaeomagnetic ages from the northwestern Neostromboli phase suggests that it erupted within a confined period between roughly 6 and 14 ka. The low uncertainty on the 3Hecos ages as well as on individual exposure ages (4.4–8.7%) demonstrates that 3Hecos exposure dating is a viable tool for dating Holocene basalt lavas. The ages compare favourably to uncertainties obtained for radiocarbon dating of similar rocks.

Detailed stratigraphic and geochemical studies were carried out on remnants of the partially collapsed Pizzo-o-Sopra-la-Fossa (Pizzo) scoria cone, which represents the best viewpoint for observing the present-day activity on top of the Stromboli volcano, Italy. The main outcrop is above the active craters and is strongly altered by acid fumarolic gases; for this reason, the Pizzo activity has previously been poorly studied (for age and composition).We focus mainly on the well-preserved Le Croci fallout succession. We show that the Pizzo activity occurred in the Roman age (∼2.4–1.8 ka) and was fed by two compositionally and genetically distinct magma groups, each occupying a separate reservoir. Shoshonitic basalts (Pizzo-SHO) were stored at lower depths than high-K basalts to high-K basaltic andesites (Pizzo-HKCA). Although Pizzo- SHO and Pizzo-HKCA magmas were affected by multiple mafic magma inputs (with different dynamics as reflected in variable mineral zoning), they did not interact with each other. They were, however, erupted simultaneously suggesting they were emitted by two unconnected conduits and vents located in the summit area. Stromboli historic activity was thus quite different from present-day activity. In the Roman age, the volcano was characterised by paired activity taking place on the summit with similar explosive characteristics, from two coeval vents each fed by different and non-parental magmas. While the Pizzo-SHO magmas still persist in the present-day shallow reservoir, the Pizzo-HKCA magmas could be quiescent and re-activated in the future, possibly from an eccentric vent/fissure.

The subaerial part of the Stromboli stratovolcano was built up in the last 100 kyr through six periods of activity; the erupted magmas record the largest compositional variation of all the Aeolian arc volcanoes (calc-alkaline, shoshonitic, and potassic alkaline magma series).The trace element characteristics of the less evolved magmas of each period of activity are coherently correlated with their radiogenic isotope (Sr, Nd, Pb) composition, and are typical of volcanic arc rocks. In terms of U-series isotopes, samples from the different magma series have both U and Th excesses, and this distinctive feature provides additional constraints on source enrichment processes within the mantle wedge and on the mechanism of partial melting. Overall the complete set of data demonstrates that the genesis of the different magma series at Stromboli can be accommodated in a mantle source that experienced two distinct enrichment processes by different parts of the subducting oceanic crust of the Ionian slab.The first was caused by supercritical liquids originating from the basaltic and sedimentary parts of the subducting slab at 45 GPa and 9008C.The second was induced by aqueous fluids, again originating from the basaltic and sedimentary parts of the slab, released from a shallower part of the subducted Ionian slab (55 GPa and 8008C). U^Th disequilibria constrain the timing of the first metasomatic event (Stage I: supercritical liquids) at 4435 ka, whereas the second event (Stage II: aqueous fluids) occurred at 100 ka.The high-angle dip of the Ionian slab ( 708) caused the superimposition of the metasomatizing agents of the two enrichment processes in the same volume of the mantle wedge, explaining the occurrence of such different magma series in a single volcanic edifice. The U^Th disequilibria provide evidence for dynamic melting of the metasomatized mantle wedge combined with an ageing effect resulting from the restoration of secular equilibrium after the perturbation caused by the U-rich aqueous fluids of Stage II.The trace element and radiogenic isotope (U,Th, Sr, Nd, Pb) signature of the mantle source of the magmas at Stromboli is thus dependent upon the amount of supercritical liquids and aqueous fluids released by the two components of the subducted slab, whereas the distinctive U and Th excesses of the magmas result from a combination of mantle ageing and time-dependent dynamic melting. The geochemical and radiogenic isotope signature of the mantle source beneath Stromboli places important constraints on the isotopic polarity from Southern Latium to the Aeolian arc attributed to the effect of a HIMU mantle component following either lateral inflow of foreland mantle material or upwelling of a mantle plume in the centre of the Tyrrhenian basin. Our geochemical model demonstrates that the high Pb/Pb of the putative ‘HIMU’ mantle component could be equally formed during metasomatism of the pre-existing mantle wedge by either the supercritical liquid (Stage I) or aqueous fluid (Stage II) released by the subducted altered basalt of the Ionian plate.

Stratigraphic changes in bulk-rock and mineral chemistry within a volcanic sequence can be used to understand the evolution of the magma storage system beneath the volcano; in particular, changing eruptive styles and major changes in magma chemistry can be linked to open- or closed-system behaviour. Changes in phenocryst mineral chemistry, often expressed as core–rim zonation patterns or the development of reaction rims, combined with thermodynamic analysis of appropriate mineral equilibria, can provide powerful insights into the timing of inputs of new, more primitive magma. The Vancori period (26–13.8 ky) of the eruptive activity of Stromboli has a relatively simple stratigraphy, comprising three sub-periods, Lower, Middle and Upper Vancori, each one beginning and ending with an explosive volcanic event, marked by breccias. Electron microprobe analyses of phenocryst and groundmass minerals (olivine, clinopyroxene, plagioclase, Fe–Ti oxides) in representative samples from the Lower, Middle and Upper Vancori phases of activity allow tracking of the evolution of the magma chamber system with time. Thermodynamic analysis of olivine–clinopyroxene mineral equilibria is used to constrain changes in temperature within the magmatic system. The main stratigraphic discontinuities within the Vancori sequence are clearly reflected by changes in bulk-rock and mineral chemistry, and evidence of textural and thermodynamic disequilibrium (calculated temperatures) in the eruptive products. On the basis of these data the Lower and Middle Vancori periods of the eruptive history of Stromboli appear to be characterised by an open magma chamber system repeatedly re-filled with basaltic magma, which was continuously being erupted. At the end of the Middle Vancori period the system became closed, corresponding to a period of dormancy of the volcano, and fractional crystallisation dominated. Following a major flank collapse of the volcanic edifice, the Upper Vancori eruptive products show the most extreme range in bulk-rock chemistry as the chamber was progressively flushed with new basaltic magma. D 2005 Elsevier B.V. All rights reserved.

Volcanic activity of Stromboli in the last 100 years was characterised by regular Strombolian eruptions with some occurrences of major explosions and paroxysms at the summit crater and lava flows down into the Sciara del Fuoco. Two types of juvenile fragments, shoshonitic to high-K basalts in composition, are peculiarly outpoured during major explosions: black scoriae, similar to those erupted by the normal Strombolian activity, representing a highly porphyritic (45–60 vol%), volatile-poor magma (HP magma) and a small volume of light pumice, representing a low-porphyritic (

Subduction-related and continental intraplate type magmatism coexisted in the San Pedro–Ceboruco (SPC) graben in western Mexico in a rather unusual close association. The magmatic systems belong to the Trans-Mexican Volcanic Belt (TMVB), the active volcanic arc linked to subduction of the Rivera and Cocos plates beneath the North American plate. Three different magmatic series are recognized in the San Pedro–Ceboruco graben: calc-alkaline, Na-alkaline, and transitional. The transitional series can be further subdivided into: Low-Ti; High-Ti and Amado Nervo groups. The different magmatic series are characterized by variable LILE/HFSE and LREE/HFSE ratios. The Na-alkaline series has the most radiogenic Nd compositions and the lowest 87 Sr/ 86 Sr (0.70320–0.70344). The calc-alkaline series shows the lowest eNd and the highest 87 Sr/ 86 Sr (0.70395– 0.70402) overlapping the values for the transitional High-Ti group (0.70390–0.70404). Sr isotopic compositions for transitional Low-Ti (0.70367–0.70394) and transitional Amado Nervo (0.70351–0.70389) groups are intermediate between those of the Na-alkaline and calc-alkaline rocks. All the studied rocks show similar 207 Pb/ 204 Pb (15.58–15.61) and 208 Pb/ 204 Pb (38.41– 38.65), but 206 Pb/ 204 Pb discriminates well the Na-alkaline series (18.90–19.03) from all the others (18.68–18.75). Compositional and isotopic data suggest that the different series derive from distinct parental magmas, which were generated by partial melting of a heterogeneous mantle source characterized by two different components. A depleted mantle (DM) component with low 206 Pb/ 204 Pb and an enriched (EM) component characterized by high 206 Pb/ 204 Pb. These two components

After decades of repose, Puyehue-Cordon Caulle Volcano (Chile) erupted in June 2011 following a month of continuously increasing seismic activity. The eruption dispersed a large volume of rhyolitic tephra over a wide area and was characterized by complex dynamics. During the initial climactic phase of the eruption (24–30 h on 4–5 June), 11–14-km-high plumes dispersed most of the erupted tephra eastward towards Argentina, reaching as far as the Atlantic Ocean. This first eruptive phase was followed by activity of lower intensity, leading to the development of a complex stratigraphic sequence, mainly due to rapid shifts in wind direction and eruptive style. The resulting tephra deposits consist of 13 main layers grouped into four units. Each layer was characterized based on its dispersal direction, sedimentological features, and on the main characteristics of the juvenile fraction (texture, density, petrography, chemistry). The lowest part of the eruptive sequence (Unit I), corresponding to the tephra emitted between 4 and 5 June, is composed of alternating lapilli layers with a total estimated volume of ca. 0.75 km3; these layers record the highest intensity phase, during which a bent-over plume dispersed tephra towards the southeast-east, with negligible up-wind sedimentation. Products emitted during 5–6 June (Unit II) signaled an abrupt shift in wind direction towards the north, leading to the deposition of a coarse ash deposit in the northern sector (ca. 0.21 km3 in volume), followed by a resumption of easterly directed winds. A third phase (Unit III) began on 7 June and resulted in tephra deposits in the eastern sector and ballistic bombs around the vent area. A final phase (Unit IV) started after 15 June and was characterized by the emission of fine-grained white tephra from ash-charged plumes during low-level activity and the extrusion of a viscous lava flow. Timing and duration of the first eruptive phases were constrained based on comparison of the dispersal of the main tephra layers with satellite images, showing that most of the tephra was emitted during the first 72 h of the event. The analyzed juvenile material tightly clusters within the rhyolitic field, with negligible chemical variations through the eruptive sequence. Textural observations reveal that changes in eruption intensity (and consequently in magma ascent velocity within the conduit) and complex interactions between gas-rich and gas-depleted magma portions during ascent resulted in vesicular clasts with variable degrees of shear localization, and possibly in the large heterogeneity of the juvenile material.

In the Kirka-Afyon and Isparta areas, potassic and ultrapotassic alkaline magmatism followed Upper Eocene - Middle Miocene calc-alkaline volcanism, which was related to the northward subduction of the African Plate beneath Eurasia. The alkaline volcanism lasted from 14.8 Ma (in the north) to 4 Ma (in the south) and was associated with a presently still active extensional tectonic regime. It is mainly located along a N-S fault, which marks the limit between the Aegean Are and the Cyprus Are (Isparta Angle). In the Kirka-Afyon area the potassic to ultrapotassic rocks range from silica saturated (trachybasalts to trachytes) to highly undersaturated (phonolitic leucitites). In the Isparta area ultrapotassic rocks occur together with rhyolitic lavas. In both areas, the ultrapotassic magmatism is transitional between lamproitic and Roman type, with a more lamproitic character for the Isparta rocks. The potassic suites of both the Kirka-Afyon and Isparta areas are characterised by rocks with Sr isotope ratios increasing from mafic to silicic rocks. The rocks from Kirka-Afyon area have an orogenic affinity, whereas those from Isparta area exhibit orogenic (the potassic suite) and within-plate (the ultrapotassic suite) affinities. Extreme enrichments in LIL elements characterise the Afyon phonolitic leucitites, whereas enrichments in Sr and REE are observed in Isparta potassic rocks. A residual, probably lithospheric mantle metasomatised by fluids and/or melts of different origin (subduction-related for the Kirka-Afyon rocks; deep astenospheric origin for the Isparta rocks), is considered to be the source of the ultrapotassic magmas. A more "fertile" and probably deeper mantle, enriched by subduction-related components, is believed to be the source of the potassic rocks. Lower degrees of partial melting, in the mantle, at higher pressure, and probably at higher X-CO2, may account for the genesis of the phonolitic leucitite parental magmas. Partial melting of the mantle took place simultaneously in both sources, probably due to post-collision extensional tectonics, which was also responsible for the uprise of deep astenospheric mantle.

Glass inclusions hosted by clinopyroxene and matrix glasses of pumices from the Plinian fall deposit associated with the Campanian Ignimbrite eruption were analysed in order to investigate processes that occurred at various stages of the magmatic evolution and to characterise the magma-withdrawal dynamics during the eruption. Petrographic, SEM–EDS, and electron microprobe analyses were performed on pumice samples, glass inclusions, and host minerals. Primary non-evolved glass inclusions were found both in homogeneous salite and in zoned diopside-salite clinopyroxenes. All the studied glasses are peralkaline alkali-trachytic in composition. The variation trends of glass inclusions and matrix glasses overlap for most elements except for alkalies and Cl. The least evolved compositions, however, are shown by glass inclusions found in salitic clinopyroxene at interfaces with diopside from the early erupted Plinian pumices (Lower Fall), whereas the most evolved compositions were found in the matrix glasses of the late-erupted Plinian pumices (Upper Fall). Excluding the most mafic glass inclusions, the observed compositional range is closely comparable to the range for ignimbrite juvenile clast compositions, indicating that the two eruptive phases drained the same magmatic reservoir. The compositional evolution of the Campanian Ignimbrite magmatic system seems to have resulted from different processes occurring at different stages: pre-eruptive magma mixing, due to an input of mafic magma into the reservoir, and crystal fractionation starting from different hybrid parental magmas to produce the most evolved compositions. These processes could have generated a vertically and laterally zoned magma chamber with multiple layers of variably hybrid magma extending from the input points of mafic magma. This type of zoning seems to better account for all the complex chemical and mineralogical characteristics of the Campanian Ignimbrite magmatic system, instead of a simple vertically zoned magmatic reservoir. A syn-eruptive mingling between the least evolved hybrid magmas and the most evolved melts characterised the first phase of Plinian eruption. This uncommon emptying of the magma chamber could be explained by the particular geometry of the reservoir.

Stromboli is the type locality of continuous and moderately explosive volcanic activity. Monitoring temporal variations in the composition of the material erupted allows constraints to be made on the magma chamber dynamics and volcanic hazard. Here we present an Sr isotope survey of scoriae and lavas erupted from Stromboli volcano during this century. The material erupted is transitional between shoshonite and High-K basalt, with relatively constant major and trace element composition. This implies no substantial physical separation of minerals during crystallization (ca. 50 vol% minerals in both lavas and scoriae) and hence a relatively homogeneous reservoir. 87Sr/86Sr values are constant from 1900 to ca. 1980, then, beginning prior to the major lava flow eruption of December 1985, there is a smooth decrease. The Sr isotope decrease records the arrival of a new feeding magma, and allows estimation of the magma residence time and the volume of the reservoir beneath Stromboli volcano. Our results, along with a critical assessment of magma flux estimates, are best reconciled with steady state conditions, and establish the existence of a relatively small, ∼0.3 to 0.04 km3, reservoir with a magma residence time (τ) of ca. 19 years. Monitoring the temporal variation of isotopic ratios in active volcanoes appears a successful tool in forecasting some major volcanic eruptions.

Periodic lava sampling was carried out at the active vents during the entire duration ofthe 28 December 2002 to 22 July 2003 effusive eruption. Major and trace element bulk rock analyses were performed at different laboratories, thereby acquiring four independent sets of analysis. Nd and Sr isotope ratios were obtained on whole rocks and groundmasses, together with micro-Sr isotope analyses on plagioclase and clinopyroxene by microdrilling technique. Crystal size distribution, mineral, and glassy matrix chemistry were analyzed on selected samples. The products show a fairly homogeneous composition, close to that of the crystal-rich scoria that erupted in the previous years. Slight variations of trace elements and isotope ratios between products that erupted before and after the 5 April paroxysm are likely accounted for by limited mixing between the fresh, volatile-rich magma that erupted during the paroxysm and the volatile-poor magma feeding the lava flow. Micro-Sr isotope data show large isotopic disequilibria pointing to the persistence of highly Sr-radiogenic xenocrysts or crystal cores in the shallow magmatic system, probably recycled from the previous activity. Data rule out important changes in the dynamics of the plumbing system shortly before the eruption. A discrete input of deep magma into the lower part of the shallow system some months before the eruption may be at the origin of the increase of the magmatostatic pressure in the conduits, leading to the effusive eruption. An alternative hypothesis considers a nearly steady-state feeding system undergoing gradual, long-term pressure increase in its upper part, eventually leading to periodic lava effusions.

Nisyros island is a calc-alkaline volcano, built up during the last 100 ka. The first cycle of its subaerial history includes the cone-building activity with three phases, each characterized by a similar sequence: (1) effusive and explosive activity fed by basaltic andesitic and andesitic magmas; and (2) effusive andextrusive activity fed by dacitic and rhyolitic magmas. The second eruptive cycle includes the caldera-forming explosive activity with two phases, each consisting of the sequence: (1) rhyolitic phreatomagmatic eruptions triggering a central caldera collapse; and (2) extrusion of dacitic-rhyolitic domes and lava flows. The rocks of this cycle are characteized by the presence of mafic enclaves with different petrographic and chemical features which testify to mixing-mingling processes between variously evolved magmas. Jumps in the degree of evolution are present in the stratigraphic series, accompanied by changes in the porphyritic index. This index ranges from 60% to about 5% and correlates with several teochemical parameters, including a negative correlation with Sr isotope ratios (0.703384–0.705120). The latter increase from basaltic andesites to intermediate rocks, but then slightly decrease in the most evolved volcanic rocks. The petrographic, geochemical and isotopic characteristics can be largely explained by processes occurring in a convecting, crystallizing and assimilating magma chamber, where crystal sorting, retention, resorption and accumulation take place. A group of crystal-rich basaltic andesites with high Sr and compatible element contents and low incompatible elements and Sr isotope ratios probably resulted from the accumulation of plagioclase and pyroxene in an andesitic liquid. Re-entrainment of plagioclase crystals in the crystallizing magma may have been responsible for the lower 87Sr/86Sr in the most evolved rocks. The gaps in the degree of evolution with time are interpreted as due to liquid segregation from a crystal mush once critical crystallinity was reached. At that stage convection halted, and a less dense, less porphyritic, more evolved magma separated from a denser crystal-rich magma portion. The differences in incompatible element enrichment of pre-and post-caldera dacites and the chemical variation in the post-caldera dome sequence are the result of hybridization of post-caldera dome magmas with more mafic magmas, as represented by the enclave compositions. The occurrence of the quenched, more mafic magmas in the two post-caldera units suggests that renewed intrusion of mafic magma took place after each collapse event.

New geochemical and isotopic data are reported for calc-alkaline (CA) volcanics of the Aeolian arc. Three main groups are recognized: the Alicudi and Filicudi volcanics in the western part of the arc; the Panarea, Salina and Lipari (henceforth termed PSL) volcanics in the central part of the arc and the Stromboli suite which makes up the eastern part of the arc. Each group is characterized by distinctive isotopic ratios and incompatible element contents and ratios. 87Sr/86Sr values (0.70352–0.70538) increase from west to northeast, and are well correlated with 143Nd/144Nd (ɛNd from +4.8 to -1.5). Pb isotope ratios are fairly high (6/4=19.15–19.54; 7/4=15.61–15.71; 8/4=38.97–39.36), with a general increase of 7/4 and 8/4 values from Alicudi to PSL islands and Stromboli. LILE contents and some incompatible element ratios (e.g. Ba/La, La/Nb, Zr/Nb, Rb/Sr) increase from the western to the central part of the arc, whereas HFSE and REE abundances decrease. Opposite variations are often observed in the volcanics toward the north-east from PSL islands. To account for these features and the decoupling observed between isotopic compositions and incompatible element abundances and ratios, it is suggested that a mantle source with affinities to the MORB source is “metasomatized” by slab-derived, crustal components. The proportion of crustal material entrained in the mantle source increases from Alicudi to Stromboli, according to the Sr and Nd isotope variations. It is also proposed that slab derived hydrous fluids play an important role, but which is variable in different sectors of the arc. This is attributed to the metasomatizing agent having variable fluid/melt ratios, reflecting different types of mass transfer from the subducted contaminant (probably pelagic sediments) to the mantle wedge. Thus, it is suggested that the slab derived end-member has a high hydrous fluid/melt ratio in the PSL mantle source and a correspondingly lower ratio in the Alicudi and Stromboli sources.

After the 6 month-long effusive event of 2002-2003, a new lava effusion occurred at Stromboli between 27 February and 2 April 2007. Despite the different durations, approximately the same volume of magma was emitted in both eruptions, in the order of 107 m3. A paroxysmal eruption occurred at the summit craters in both the 2002–2003 and 2007 episodes, during which a significant amount of low porphyritic (LP), volatile-rich magma was erupted. In both cases, the paroxysm did not interrupt the lava emission. Here, we present compositional data, including texture, mineralogy, chemistry and Sr and Nd isotope ratios of bulk-rock, groundmass and separated minerals of lavas erupted in 2007, together with chemistry and Sr and Nd isotope composition of the pumices emitted during the 15 March paroxysm. As a whole, the lavas have the same texture and chemistry that characterize the highly porphyritic (HP) products usually erupted at Stromboli during normal Strombolian activity and effusive events. Compared to the previous HP products, the 2007 lavas show minor but systematic mineralogical and isotopic variations which are consistent with a modest increase of the magma supply rate of the volcano. Compositional variations during the entire duration of the event are very modest. Glass chemistry changes in lavas erupted in the second half of March can be explained by the minor mixing between the volatile-rich LP magma rising through the shallow magmatic system during the 15 March paroxysm and the degassed residing HP magma. A first conclusion of this study is that there is no compositional evidence supporting major changes in the magma dynamics of the volcano accompanying the effusive activity, as also suggested for the 2002–2003 event. The activity of Stromboli is controlled by a steady state feeding system in which refilling, mixing, degassing and crystallization at shallow level continuously operate, with modest oscillations in the magma supply rate. Switching between normal Strombolian and effusive activity is related to periods of relatively more vigorous refilling of the shallow system, leading to progressive pressure increase in the upper conduits associated with only minor compositional variations in the erupted products.

[1] Mineralogical, geochemical, and melt inclusion analyses have been performed on scoria samples from a small scoria fall deposit interbedded between the Cape Riva (21 ka) pyroclastic deposits and the Minoan (3.6 ka) Plinian eruptions of Santorini volcano and located near the resort of Imerovigli (25°25′14″/36°26′01″, WGS84), close to the town of Phira. The scoria samples are basaltic andesites with normally zoned olivine, reversely zoned pyroxenes, and variably zoned plagioclase. Olivine contains two types of melt inclusions. Type I inclusions, hosted in cumulus Fo85–90 olivine, are primitive low-K basalts with MgO 6–7 wt % and contain volatile element abundances of about 500 ppm F, 1000 ppm Cl, and 1300 ppm S. Low abundance of incompatible elements and rare earth element patterns indicate an arc tholeiitic composition. The δ18O in host olivine is low (5.3‰) and lower than the value for the groundmass (6.5‰), whereas 87Sr/86Sr of the type I inclusions is rather high (0.70579) and higher than the whole rock and groundmass values (0.70465 and 0.70468, respectively). Type II inclusions, hosted in Fo80–83 olivine rims, are calc-alkaline andesites and have low volatile contents. Interstitial glasses (groundmass) are similar in composition to the type II inclusions. We propose that the tholeiitic, high Sr radiogenic, type I primitive magma was generated by high partial melting degrees of a mid-ocean ridge basalt (MORB)-like mantle wedge metasomatized by subducted sediment melts and minor aqueous fluids. The input amount of sediment melts and the degree of mantle melting decreased with time, providing the parental magmas of the calc-alkaline, less Sr radiogenic, type II andesitic melts. The scoria fall eruption was fed by hybrid magma with multiple generations of olivine, clinopyroxene, and plagioclase growth and occurred a short time after the mixing process. This magma was probably last stored in a small side reservoir, separated from the Cape Riva and maybe from the Minoan reservoirs and was emplaced from a subaerial vent close to Imerovigli by a Strombolian-type ephemeral eruption.

A detailed investigation on the final-stage Latera lavas has shown the presence of three groups of magma having different petrochemical affinities: (1) typical high-potassium series (HKS) magmas, represented only by evolved rocks (group A); (2) HKS-like volcanics with slightly lower K2O/Na2O ratios and lower K2O and incompatible trace-element contents than the typical HKS of the Roman Magmatic Province (group B); (3) potassic series (KS) with higher K2O/Na2O ratios and higher SiO2, MgO, Cr, Ni, Rb, Th and LREE contents and lower Ba and Sr contents than the typical Roman KS (group C). Compared to the HKS volcanics, the latter group of volcanics have higher Th contents and Rb/Ba and Ce/Sr ratios, and lower LREE, Sr, Rb and Ba contents, and lower LILE/HFSE and LREE/HFSE ratios. It has been also shown that the group-C lavas display a general composition similar to that of Radicofani volcanics, suggesting a transitional character between the typical Roman-type KS rocks and the Tuscan lamproites. Crystal fractionation was mainly responsible for the petrological variations observed within the two HKS groups, which are believed to derive from two distinct parental magmas, probably a leucite-tephrite liquid (group A) and a leucite-basanite liquid (group B). Limited and variable crustal assimilation processes were associated with crystal fractionation in the evolution of the group-B magmas. On the other hand, the group-C magmas underwent virtually no fractionation during their ascent to surface, thus their composition can be considered almost primary. Different degrees of partial melting, at high-P and high-XCO2 conditions, of a metasomatized phlogopite-bearing peridotitic mantle source are suggested to explain the genesis of the group-A and -B primary magmas. On the other hand, low-P and low-XCO2 conditions during the partial melting of a phlogopite-bearing mantle source, slightly depleted in clinopyroxene, are suggested for the genesis of the group-C magmas. Moreover, the different incompatible-element contents and ratios between the HKS (group-A and -B) and KS (group-C) rocks studied should reflect primary characteristics of the mantle sources.

The 5 April 2003 explosive eruption at Stromboli emplaced typical basaltic scoria, pumice, and lithic blocks. This paper reports a detailed set of mineralogical, geochemical, and isotopic data on the juvenile ejecta and fresh subvolcanic blocks, including micro-Sr isotope analyses and major and dissolved volatile element contents in olivine-hosted melt inclusions. The juvenile ejecta have compositions similar to those of their analogs from previous paroxysms; the 2003 pumice, however, does not contain stable high-MgO olivine, usually typical of large-scale paroxysms and has lower compatible element contents. Texture, composition, and Sr isotope disequilibrium of crystals in pumice indicate that most of them are inherited from the shallow crystal-rich magma and/or crystal mush. The most primitive magma is recorded as rare melt inclusion in olivine Fo 85―86 . It has a typical S/Cl (1.1) and a total volatile content of 3.1 wt % from which the total fluid pressure was evaluated >240 MPa. Hence, moderate pressure conditions can be envisaged for the mechanism triggering the April 2003 paroxysm. The subvolcanic blocks are shoshonitic basalts with 45―50 vol % of phenocrysts (plagioclase + clinopyroxene + olivine). The late-stage crystallization of the crystal-rich magma lead to the formation ofNa-sanidine with plagioclase An 60―25 + olivine Fo 68―49 + Ti-magnetite ± apatite ± phlogopite ± ilmenite assemblage. Mineralogy, chemistry, and Sr―Nd isotopic signatures of the subvolcanic blocks indicate they represent the slowly cooled equivalents of batches of crystal-rich basaltic magma stored in the uppermost subvolcanic feeding system during the last few years. Cooling might be facilitated by short breaks in the summit crater activity.

Stromboli volcano is famous in the scientific literature for its persistent state of activity, which began about 1500 years ago and consists of continuous degassing and mild intermittent explosions (normal Strombolian activity). Rare lava emissions and sporadic more violent explosive episodes (paroxysms) also occur. Since its formation, the present-day activity has been dominated by the emission of two basaltic magmas, differing chiefly in their crystal and volatile contents, whose characteristics have remained constant until now. The normal Strombolian activity and lava effusions are fed by a crystal-rich, degassed magma, stored within the uppermost part of the plumbing system, whereas highly vesicular, crystal-poor light-colored pumices are produced during paroxysms testifying to the ascent of volatile-rich magma batches from deeper portions of the magmatic system. Mineralogical, geochemical, and isotopic data, together with data on the volatile contents of magmas, are presented here with the aim of discussing (1) the relationships between the different magma batches erupted at Stromboli, (2) the mechanisms of their crystallization and transfer, (3) the plumbing system and triggering mechanisms of Strombolian eruptions.

A multidisciplinary study has been conducted on 109 intrusive sheets of Stromboli. The relation between sheet injection conditions with the magmatic and structural variations of the volcano, and the parameters governing the preferential intrusion geometry inside the volcano in a time perspective have been determined. During its subaerial activity (the last 100 ka) Stromboli experienced periods of extrusive growth and sheet intrusions, alternated to caldera and lateral collapses of the NW Sciara del Fuoco flank. The sheet intrusions, similar to the extrusive rocks, span a large compositional range from calc-alkaline to potassic, through high-K calc-alkaline and shoshonitic series. The most evolved compositions of the effusive rocks, however, have not been found as intrusions, whereas some mafic and basic magmas of sheets were never erupted. The repeated time change of location of the intrusions from Strombolicchio to the NE–SW trend passing trough the inferred Paleostromboli summit cone and the southern flank could represent the expression of the shift of the magma chamber towards the southeast, and the subsequent onset of a clear NE–SW-elongated magma chamber located along the main Stromboli weakness zone. The Stromboli evolution in the last 100 ka has been characterized by the persistent N 10–55° E intrusion direction, which reflects the influence of the edifice's state of stress on the magma feeding paths. After 13 ka ago, sheets mainly intruded along to the sector collapse scarps in a circum-amphitheatre geometry, which has been demonstrated to be influenced by the new cone morphostructure and not by different magma rheological characters.

This study focuses on two issues that are still a matter of debate in subduction zones, particularly in western Mexico: (1) the close association within the same volcanic complex of typical amphibole-free andesites to rhyolites and amphibole-bearing andesites to rhyolites, characteristic of the hydrated front of the Mexican arc; and (2) the occurrence of bimodal magmatism without evidence for interaction between mafic and intermediate to silicic magmas, which are in addition characterized by different petrogenetic affinities. Our case study is the San Pedro–Cerro Grande volcanic complex, a Quaternary silicic to intermediate dome complex located in western Mexico. Volcanic activity has been divided into two periods. In the middle Pleistocene, andesitic to dacitic magmas were emplaced along WNW-trending faults in the southern portion of the complex. The Las Cuevas pyroclastic sequence (older than ca. 500 ka) was emplaced during this episode, most likely from a local source. This first period of activity ended before ca. 280 ka with the emplacement of the Cuastecomate Plinian deposit, which is related to the formation of the San Pedro caldera, an ∼4-km-wide subcircular depression that is today partially buried by younger volcanic products. During the second period of activity (ca. 280–30 ka), rhyolitic and dacitic domes were mostly emplaced along the caldera rim and inside the caldera. In addition, hawaiites and mugearites built the Amado Nervo shield volcano on the caldera rim. Intermediate- to high-silica lava and pyroclastic rocks are subalkaline, whereas the Amado Nervo mafic lavas are transitional toward the alkaline series (Na-alkaline). No genetic relationships have been found between subalkaline and transitional Na-alkaline rocks, which are thought to represent different batches of magma from different mantle sources. Petrographic, geochemical, and isotopic variations observed in the transitional Na-alkaline Amado Nervo lavas point to a parental magma from a mantle melt that underwent limited olivine separation during its ascent to the surface. Among subalkaline rocks, two groups showing contrasting petrographical and geochemical features are recognized based on the presence of amphibole. Amphibole-bearing intermediate to silicic rocks are characterized by lower Ce and other incompatible trace element contents and lower 87Sr/86Sr (0.70382–0.70401) compared to amphibole-free rocks (0.70411–0.70424). On the basis of petrological characteristics, the two groups of magmas are interpreted to have evolved in two different magmatic reservoirs under different pressures and water contents in the mid-upper crust. Both groups of magmas were differentiated by open-system processes. We propose that assimilation and equilibrium crystallization (AEC) processes account for the amphibole-bearing rocks. Hotter and less evolved magmas interacted to a higher degree with the crust than the more evolved and colder magmas. This produced the observed higher 87Sr/86Sr in the less differentiated rocks of the amphibole-bearing group. On the other hand, amphibole-free rocks have chemical and isotopic characteristics that can be modeled by assimilation and fractional crystallization (AFC) processes. All data suggest that the two groups of subalkaline rocks have been generated by a common parental hydrous magma, but evolved in two different reservoirs. Amphibole-bearing magmas underwent amphibole fractionation in a mid-upper crustal reservoir and show assimilation of two types of basement: one akin to Oaxaquia and another akin to the Guerrero terrane. Amphibole-free magma only shows assimilation of an Oaxaquia-type basement.

On 28 December 2002, the persistent Strombolian activity at Stromboli was interrupted by the sudden onset of lava emission onto the Sciara del Fuoco slope, a horseshoe-shaped depression on the NW flank of the volcano. The effusive episode went on until 22 July 2003 and produced a cumulative volume of lavas of 11 × 106 m3 ranking the event as the largest occurred in the past 30 years. The eruptive vents were mainly located in the NE sector of the Sciara del Fuoco depression, at an elevation of 550–600 m a.s.l. On 30 December, the eastern portion of the Sciara del Fuoco collapsed producing a tsunamigenic landslide. On 5 April 2003, a paroxysmal eruption occurred at the summit craters during which crystal-poor pumiceous products were emitted. The paroxysm did not interrupt the lava emission. The Strombolian activity at the summit craters gradually resumed starting from March 2003 and fully recovered by the end of July. Periodic sampling of the lava at the active vents was carried out during the entire effusive event. All the analysed samples are shoshonitic basalts (SiO2 48.5–50.4 wt.%; K2O 2.1–2.4 wt.%) in the range of composition observed in the products erupted during the past 20 years. They bear about 50 vol.% zoned crystals of plagioclase An90–60, diopside–augite and olivine Fo70–73 in a compositionally homogeneous shoshonitic groundmass. Bulk rock major and trace element contents measured with different methodologies at different laboratories show only minor variations. Sr and Nd isotope ratios are close to those of the crystal-rich scoriae erupted in the previous years. Despite of small compositional variations related to the emptying of the zoned topmost portion of the conduit, the data show that the switch from Strombolian explosive to effusive activity is not associated with changes of the textures and composition of the erupted products. Slight but somehow systematic variation of trace elements and isotope ratios between products erupted before and after the 5 April eruption are likely accounted by limited mixing between the fresh, volatile-rich, crystal-poor, magma erupted as pumice during the paroxysm, and the volatile-poor, crystal-rich magma feeding the lava flow. The uniform composition of the erupted lava indicates the presence of a large volume of well-mixed, crystal-rich, homogeneous magma residing in the shallow plumbing system of the volcano. Two possible trigger mechanism of the effusive event are here proposed: (i) a discrete input of fresh magma into the lower part of the shallow magmatic system occurred some months before the eruption and was followed by crystallisation, degassing and mixing with the crystal-rich shallow magma which re-homogenised the system. These processes eventually led to the rise of the magma at a higher level and failure of the conduit walls. (ii) The onset of the effusion may represent the consequence of a gradual rise of the magma level in the conduits occurred in the past two decades. The process of progressive refilling, initiated after the 1985 effusive eruption, ultimately culminated on 28 December 2002 with the failure of the conduit wall and the opening of vents 100–150 m below the summit craters.

The exceptionally persistent activity of Stromboli volcano has lasted for at least 1400 years. The normal strombolian activity is periodically interrupted by more energetic explosions (1–2 per year) and by sporadic effusive episodes (every 10–20 years). Normal activity and effusive episodes are characterized by crystal-rich high-K to shoshonitic basalts issuing from a volatile-poor shallow system. Crystal-poor pumice are emitted only during more violent explosions, and are thought to derive from deep pulses of volatile-rich magma. Shallow level degassing induces massive crystallization of deep pulses of feeding magma that, continuously mixing with the resident one, produces the crystal-rich shoshonite of the persistent activity. We examined the crystallization history of the crystal-rich, shallow reservoir using plagioclase Crystal Size Distribution (CSD) analysis of scoriae and lavas emitted in the past twenty years. CSDs show a linear dependence from crystal size in the size interval 0.06–1.2 mm; number density of larger crystals is biased by right hand truncation effects. CSDs slopes and intercepts are quite constant during the whole considered time span revealing a system that is close to the equilibrium also from a kinetic point of view. The linear crystal size distribution are reached by the system through episodes of growth and resorption, respectively occurring in the degassed and undegassed magma during the continuous mixing in the feeding system. Plagioclase net growth rate (2 × 10 − 11 cm/s) results from a balance of growth (10 − 10 cm/s) and resorption episodes which induce spectacular zoning and resorption textures in crystals larger than 200 μm. CSDs of mafic phases cannot be accurately acquired on each single sample due to poor counting statistics; the evaluation of pyroxene and olivine CSD on the whole data set, however, confirms the conclusions acquired from plagioclase CSDs.

The volcanic rocks of the South Aegean arc (SAAVA) form a chain from the Gulf of Saronikos (Susaki, Egina, Poros, Methana) at West, to an area close to the Anatolian coast at East (Kos, Nisyros and minor islands), through the central part (Milos and Santorini island groups). The volcanic activity began in the Lower Pliocene at Egina (4.7 Ma) and lasted until present days, with the still active Methana, Milos, Santorini and Nisyros volcanoes. The beginning of volcanism is younger in the central sector of the arc. Volcanic center location was controlled by large tectonic lineaments, most of them still active, trending E-W to NW-SE for the western part and mainly NE-SW for the central and eastern parts of the arc. Volcanic fields developed along ellipse shaped areas with the longest axis oriented perpendicular to the subduction front. In the western volcanic fields (Susaki, Egina-Poros-Methana and Milos), volcanic centers are mostly monogenetic and no composite volcanic structures are present. In the eastern sector of the arc, Santorini and Nisyros are important composite volcanoes with caldera structures. Volcanic rocks belong to the calc-alkaline and high-K calc-alkaline (mainly Pliocenic in age) series. Basalts are mainly present in Santorini island group. Magmas underwent to complex differentiation processes, dominated by crystal fractionation, often associated to crustal contamination and mixing-mingling. Large compositionally zoned magma chambers often fed highly explosive eruptions, especially in the central and eastern sectors of the arc. In the western and eastern parts, potassium content of erupted magmas decreased with time, probably due to the increase of partial melting degree of the mantle source. Trends of evolution tend to pass from calc-alkaline to tholeiitic from West to the Santorini volcanic field, back again to calc-alkaline toward the Nisyros volcanic field. Incompatible trace element contents are lower in Santorini mafic magmas. From West to East, Sr and Pb isotope ratios decrease, whereas Nd isotope ratios increase. Partial melting of a MORB-like astenospheric mantle, metasomatised by prevailing subducted sediments, is thought to produce the entire spectrum of parental magmas of SAAVA. Slab-derived fluids are generally reduced. Low Ba/La values suggest the occurrence of even lower fluid contents during magma genesis at Santorini. Total amount of subducted sediments involved in the magma genesis decreases from West to East. Alternatively, the West-East 87 Sr/ 86 Sr and 144 Nd/ 143 Nd variation of the Mediterranean sediments leads to suggest a similar variation in the subducted sediment composition. Most of the geochemical characteristics of the Santorini magmas can be explained by higher partial melting degrees of the mantle source, probably triggered by the greater lithosphere extension, inducing adiabatic upwelling of the mantle. The higher lithosphere extension also caused Santorini magmas to stop at shallower levels, thus preventing amphibole crystallisation and allowing a higher amount of mafic magmas to reach the surface. From the beginning of SAAVA magmatism, occurred in the external parts of the arc, the partial melting degrees of the magma mantle source seem to have increased with decreasing time and going towards the central sector of the arc.

Over the last several hundred years, Stromboli has been characterized by steady-state Strombolian activity. The volcanic products are dominated by degassed and highly porphyritic (HP-magma) black scoria bombs, lapilli and lava flows of basaltic shoshonitic composition. Periodically (about one to three events per year), more energetic explosive eruptions also eject light coloured volatile-rich pumices with low phenocryst content (LP-magma) that have more mafic compositions than the HP-magma. An in situ major and trace element and Sr isotope microanalysis study is presented on four samples chosen to characterize the different modes of activity at Stromboli: a lava flow (1985-1986 effusive event), a scoria bomb from the 'normal' present-day activity of Stromboli (April 1984), and a scoria and coeval pumice sample from a recent more explosive eruption (September 1996). Plagioclase (An

Vico volcano has erupted potassic and ultrapotassic magmas, ranging from silica-saturated to silica-undersaturated types, in three distinct volcanic periods over the past 05Myr. During Period I magma compositions changed from latite to trachyte and rhyolite, with minor phono-tephrite; during Periods II and III the erupted magmas were primarly phono-tephrite to tephri- phonolite and phonolite; however, magmatic episodes involving leucite-free eruptives with latitic, trachytic and olivine latitic compositions also occurred. In Period II, leucite-bearing mag- mas (87Sr/86 Srinitial ˆ 071037---071115) were derived from a primitive tephrite parental magma. Modelling of phonolites with different modal plagioclase and Sr contents indicates that low-Sr phonolitic lavas differentiated from tephri-phonolite by fractional crystallization of 7% olivine‡27% clinopyroxene‡ 54% plagioclase ‡ 10% Fe---Ti oxides ‡ 4% apatite at low pressure, whereas high-Sr phonolitic lavas were generated by fractional crystallization at higher pressure. More differen- tiated phonolites were generated from the parental magma of the high-Sr phonolitic tephra by fractional crystallization of 10---29% clinopyroxene ‡ 12---15% plagioclase ‡ 44---67% sanidine ‡ 2---4% phlogopite ‡ 1---3% apatite ‡ 7---10% Fe---Ti oxides. In contrast, leucite-bearing rocks of Period III (87Sr/86Srinitial ˆ 070812---070948) were derived from a potassic trachybasalt by assimilation---fractional crystallization with 20---40% of solid removed and r ˆ 04---05 (where r is assimilation rate/crystallization rate) at different pressures. Silica-saturated magmas of Period II (87Sr/86Srinitial ˆ 071044---071052) appear to have been generated from an olivine latite similar to some of the youngest erupted products. A primitive tephrite, a potassic trachybasalt and an olivine latite are inferred to be the parental magmas at Vico. These magmas were generated by partial melting of a veined lithospheric mantle sources with different vein---peridotite/wall-rock proportions, amount of residual apatite and distinct isolation times for the veins.

A fluid-inclusion study has been performed on quartzite nodules of stromboli volcano hosted by calc-alkaline lavas of both the Strombolicchio (200 ka) and Paleostromboli II (60 ka) periods. The nodules are mainly composed of quartz crystals with subordinate plagioclase and K-feldspar. Small interstitial minerals such as plagioclase, K-feldspar, clinopyroxene, biotite, and quartz are also found, together with glass. Muscovite, epidote and zircon occur as accessory minerals. Different quartzite nodules occur: (1) equigranular polygonal granoblastic quartzite nodules forming a polygonal texture with clear triple points; (2) inequigranular polygonal granoblastic quartzite nodules; and (3) break-up nodules with strongly resorbed quartz. These quartzites are restites from partial melting, involving felsic crustal rocks at the magma/wall rock contact. Restitic quartz re-crystallises at variable and generally high temperatures, leading to the formation of quartzites with different textures. Quartz grains contain five types of fluid inclusions distinguished on the basis of both fluid type and textural/phase relationships at room temperature. Type I are two-phase (liquid+vapour) CO2-rich fluid inclusions. They are primary and subordinately pseudosecondary in origin and have undergone re-equilibration processes. Type II mono-phase/two-phase (vapour/liquid+vapour) CO2-rich fluid inclusions are the most common and, based on their spatial distribution and shape, they can be divided into two subclasses: type IIa and type IIb. Type II inclusions are secondary or pseudosecondary and they are assumed to have formed after decrepitation of type I inclusions and cracking of the host quartz. Type III inclusions are mono-phase (vapour); they possibly contain CO2 at very low density and surround the inner rims of quartz grains. Type IV two-phase silicate-melt inclusions contain glass±CO2-rich fluid. Some of them are cogenetic with type II inclusions. Finally, type V two-phase (liquid+vapour) aqueous inclusions are both vapour-rich and liquid-rich aqueous inclusions. Microthermometric experiments were performed on both type I and II inclusions. Type I inclusions homogenise to liquid between 20 and 30.5 °C. Type IIa inclusions homogenise to vapour in the 24 to 30 °C range, with a maximum peak of frequency at 29 °C. Type IIb inclusions also homogenise to vapour between 14 and 25 °C. There appears to be no difference in homogenisation temperature distribution between the Strombolicchio and Paleostromboli II samples. The trapping pressures of the fluid inclusions have been obtained by combining the microthermometric data of the Strombolicchio and Paleostromboli II samples with the pressure–temperature–volume (i.e. density) characteristics for a pure CO2 system. The data on the early inclusions (type I) suggest an important magma rest at a pressure of about 290 MPa (i.e. about 11-km depth). Type IIa CO2 inclusions suggest that a second magma rest occurred at a pressure of about 100 MPa (i.e. about 3.5-km depth), whereas type IIb inclusions were trapped later at a shallower depth during the final magma upwelling. No pressure/depth differences seem to occur between the Strombolicchio and Paleostromboli II periods, indicating the same polybaric rests for the calc-alkaline magmas of Stromboli, despite their significantly different ages. This persistence in magma stagnation conditions from 200 to 60 ka suggests a similar plumbing system for the present-day Strombolian activity.

Mantle plumes or decompression melting of a heterogeneous mantle have been proposed to explain oceanic-island basalt (OIB) type lavas found in the western Trans-Mexican volcanic belt. We show that mantle plumes cannot account for several geologic observations, whereas decompression melting is unrealistic given the low extension rate in the volcanic arc. On the basis of new geologic and geochemical studies, we propose an alternative model that reconciles geologic data, plate tectonic history, and petrology. Since 8.5 Ma, the volcanic front has migrated trenchward ∼80 km, and OIB began to be emplaced in the rear half of the arc after a period of stalled subduction of the Rivera plate between 7.2 and 4.8 Ma. As a whole, OIB accounts for ∼5% of the total volume of volcanism. Their trace element signatures require an enriched-mantle source, akin to the deep asthenosphere, but also indicate a minor involvement of a subduction component. We propose that during the period of very low convergence at the end of the Miocene, the leading edge of the Rivera slab started to sink into the mantle and continued rolling back until it reached the present dip of 45°. The sinking of the slab may have induced small amounts of asthenosphere to flow laterally into the opening mantle wedge, from the Gulf of California rift on the west and through the subducted Rivera-Cocos plate boundary on the east. This mechanism would produce veins of relatively enriched mantle, which would be able to produce melts with an OIB composition once the flux of fluids released from the subducted slab resumed.

The present paper reports the results of a detailed stratigraphical, petrological and geochemical investigation on the island of Stromboli, Aeolian arc, Southern Tyrrhenian sea. Major and trace element data determined on a large quantity of samples from well-established stratigraphic positions indicate that the magmatological evolution of the island through time was more complex than previously known. The activity of the exposed part of Stromboli, which occurred over a time span of about 100 000 years, started with the emission of high-K calc-alkaline (HKCA) volcanics, which were covered by calc-alkaline (CA), shoshonitic (SHO), high-K calc-alkaline (HKCA) and potassic (KS) products. The most recent activity consists of HKCA lavas and the present-day SHO-basaltic volcanics emitted by mildly explosive “strombolian” activity. Most of the products are lavas, with minor amounts of pyroclastic rocks emplaced mainly during the early stages of activity. The transition from the SHO to the KS cycle was associated with the collapse of the upper part of the volcanic apparatus; the transition from KS to the present-day SHO activity has been found to have occurred at the time of the sliding of the western portion of the volcano that generated the “Sciara del Fuoco” depression. The rock series cropping out at Stromboli show variable enrichment in potassium, incompatible trace elements and radiogenic Sr which increase from CA through HKCA, and SHO up to KS rocks. Major, trace element and Sr-isotopic data agree in indicating that the HKCA and SHO series evolved by crystal/liquid fractionation starting from different parental liquids, whereas crustal assimilation appears to have been the leading process during the evolution of KS volcanics. Mixing processes also played a role although they can be well documented only when they occurred between magmas with different isotopic and geochemical characteristics. Geochemical modelling based on trace element and isotopic data indicates that the mafic magmas of the different volcanic series may be generated by melting of an upper mantle heterogeneously enriched in incompatible elements and radiogenic Sr by addition, via subduction, of different amounts of crustal material. Geochemical data, however, are also in agreement with the alternative hypothesis that the most mafic magmas of the different series have been generated by combined processes of fractional crystallization, assimilation and mixing of a CA magma in a deep-sited magma chamber; the mafic magmas formed by these complex processes were successively emplaced in a shallow reservoir where they evolved by simple fractional crystallization (HKCA and SHO series) and by assimilation of crustal material (KS). The occurrence of changes in the geochemical signatures of the magmas at the time of the structural modification of the volcano is believed to favour the hypothesis that the variable composition observed in the volcanic rocks of Stromboli is the result of processes occurring within the volcanic system.

In this paper we report crystal/groundmass partition coefficients of a large set of major and trace elements for the main phenocryst minerals occurring in the potassium-rich volcanic rocks from the Roman comagmatic province, Central-Southern Italy. Based on these data, the patterns of major and trace element variation during the low-pressure evolution of alkaline potassic magmas are briefly discussed.