Episodic Cenozoic volcanism and tectonism in the Andes of Peru

Radiometric and geologic information indicate a complex history of Cenozoic volcanism and tectonism in the central Andes. K-Ar ages on silicic pyroclastic rocks demonstrate major volcanic activity in central and southern Peru, northern Chile, and adjacent areas during the Early and Middle Miocene, and provide additional evidence for volcanism during the Late Eocene. A provisional outline of tectonic and volcanic events in the Peruvian Andes during the Cenozoic includes: one or more pulses of igneous activity and intense deformation during the Paleocene and Eocene; a period of quiescence, lasting most of Oligocene time; reinception of tectonism and volcanism at the beginning of the Miocene; and a major pulse of deformation in the Middle Miocene accompanied and followed through the Pliocene by intense volcanism and plutonism. Reinception of igneous activity and tectonism at about the Oligocene-Miocene boundary, a feature recognized in other circum-Pacific regions, may reflect an increase in the rate of rotation of the Pacific plate relative to fixed or quasifixed mantle coordinates. Middle Miocene tectonism and latest Tertiary volcanism correlates with and probably is genetically related to the beginning of very rapid spreading at the East Pacific Rise.     

Permian volcanism and its relation to the tectonic development of Sumatra

The Permian volcanic deposits covering an extensive area southeast of Lake Singkarak, Central Sumatra, have been mapped in detail and studied in relation to the structural development of the pre-Tertiary Sumatra orogene. During the Permian time, Central Sumatra was occupied by an elongated sea basin in which thick sequence of bathyal and neritic sediments were deposited. Pelitic sediments dominated, but at the same time volcanic activity started in this area. The volcanic products comprise mainly flow of hornblende and augite andesites with their tuffs. In some parts of Southern Sumatra the volcanic activity lasted till Cretaceous time. The main phase of folding took place about 120 m.y. ago accompanied by an emplacement of granitic rocks. After an uplift in younger Cretaceous time, the area was strongly attacked by erosion. The regional geologic history of Sumatra revealed clearly a divergent behaviour from the classical concept of magmatic evolution in an orogenic belt as demonstrated by the dominantly andesitic character of the geosynclinal volcanism. Other examples of volcanism associated with geosynclinal subsidence, possessing an andesitic rather than a basaltic character, could also be observed in the Tertiary Sunda mountain system of Sumatra. The Indonesian examples and similar occurrences of andesitic volcanism during a geosynclinal subsidence in other parts of the world, show that much work still will have to be done in order to gain a better undestanding regarding the relationship between volcanism and orogenesis.     

The relation of mesozoic-cainozoic volcanism to tectonics in the Afro-Arabian dome

The Mesozoic-Cainozoic volcanism of NE Africa and Arabia is described in terms of four major magmatic provinces. These are dominantly basaltic and vary in general composition from strongly alkaline to tholeiitic, with some overlap between provinces. Most, but not all, of this volcanism has taken place within the confines of the Afro-Arabian dome and its attendant rifts, but the magmatism is not explicable in terms of a ‘mantle plume’ beneath Afar because such a plume explains neither the spatial nor temporal distribution of volcanicity. Instead, I propose a multi-stage model in which basaltic magmas of transitional composition were generated in Paleogene time as a response to regional extension. With subsequent evolution of the Red Sea and Gulf of Aden into proto-oceans, these magmas took on the character of oceanic tholeiites. In the Neogene, alkalic activity related to epeirogenic doming was superimposed on this extension-related volcanism. A third, independent magma-generating mechanism appears to have operated NW of the Afro-Arabian dome, where small volumes of alkaline basalt magma have been erupted intermittently since early in the Cretaceous. This activity may result from periodic tapping of deep asthenosphere by tensile fracturing and upwarping of the northward-drifting African plate.     

Relationships between mineralization and silicic volcanism in the Central Andes

Studies of late Tertiary silicic volcanic centres in the Western and Eastern Cordilleras of the Central Andes show that three volcanic environments are appropriate sites for mineralization: (1) ring-fracture extrusions post-dating large calderas; (2) similar extrusions within ignimbrite shields; and (3) isolated, small silicic volcanoes. Subvolcanic tin mineralization in the Eastern Cordillera is located in silicic stocks and associated breccias of Miocene age. The Cerro Rico stock, Potosi, Bolivia, contains tin and silver mineralization and has an intrusion age apparently millions of years younger than that of the associated Kari Kari caldera. Similar age relationships between mineralization and caldera formation have been described from the San Juan province, Colorado. The vein deposits of Chocaya, southern Bolivia, were emplaced in the lower part of an ignimbrite shield, a type of volcanic edifice as yet unrecognized in comparable areas of silicic volcanism. The El Salvador porphyry copper deposit, Chile, is related to silicic stocks which may have been intruded along a caldera ring fracture. Cerro Bonete, Chile, provides a modern example of the volcanic superstructure which may have overlain isolated mineralized stocks and breccia pipes such as that of Salvadora at Llallagua, Bolivia.Existing models for the genesis of porphyry copper deposits suggest that they formed in granodioritic stocks located in the infrastructure of andesitic stratovolcanoes. Sites of porphyry-type subvolcanic tin mineralization in the Eastern Cordillera of Bolivia are distinguished by the absence of such andesitic structures. The surface expression of a typical subvolcanic porphyry tin deposit was probably an extrusive dome of quartz latite porphyry, sometimes related to a larger caldera structure. Evidence from the El Salvador porphyry copper deposit in the Eocene magmatic belt in Chile suggests that it too may be more closely related to a silicic volcanic structure than to an andesitic stratovolcano.The dome of La Soufriere, Guadeloupe is proposed as a modern analog for the surface expression of subvolcanic mineralization processes, the phreatic eruptions there suggesting the formation of hydrothermal breccia bodies in depth. Occurrence of mineralized porphyries, millions of years after caldera formation, does not necessarily indicate that intrusions and mineralization are not genetically related to the sub-caldera pluton, but may be a consequence of the long thermal histories (1–10 million years) of the lowermost parts of large plutons. Caldera formation can only inhibit mineralization by dispersal of ore metals when these are of magmatic origin, and ignimbrites should not be taken as being unlikely to be associated with porphyry mineralization. Whether ore metals are of wall rock or magmatic origin, the key to understanding the relationships between silicic volcanism and mineralization lies in the fractionation of trace elements within large zoned magma chambers during their igneous history, and their subsequent hydrothermal migration. Small, highly mineralized intrusions formed late in a caldera cycle (such as the Cerro Rico) may be due to the introduction of fresh supplies of mafic magma into the lower parts of the main pluton.     

Upper Cenozoic volcanism in the Central Andes — Ages and volumes

Fifty-two new K-Ar dates for Upper Cenozoic volcanic rocks from north Chile and southwest Bolivia are presented, together with a compilation of previously available dates from this region. These dates are combined with calculations of volumes of lava and ignimbrite for a segment of the volcanic province (19°30′S to 22°30′S) to identify fluctuations in the level of volcanic activity during the last 24 Ma. Histograms of volumes against time have been plotted for each half-degree quadrant. In the southern half of the study area, there were peaks of activity in the periods 12 to 9 Ma and 6 to 3 Ma. In the northern half, a large proportion of the material was erupted in the period 6 to 0 Ma. This regional variation suggests that localized factors may govern the rate of volcanic output and complicates attempts at correlation with “pulses” of volcanic activity recognized elsewhere in the Pacific region. There is no conclusive evidence for volcanic episodes synchronous over such wide areas. A simple correlation between changes in spreading behaviour and changes in levels of volcanic activity is unlikely, in view of the complexity of the interactions at destructive plate margins. The rate of continental crust accretion from volcanic processes must be much less than that due to intrusive processes to account for the thickening of the Andean crust since the Jurassic.     

Drivers of peatland water table dynamics in the central Andes,Bolivia and Peru

Cushion plant dominated peatlands are key ecosystems in tropical alpine regions of the Andes in South America. The cushion plants have formed peat bodies over thousands of years that fill many valley bottoms, and the forage produced by the plants is critical for native and nonnative domesticated mammals. The sources and flow paths of water supporting these peatlands remain largely unknown. Some studies have suggested that glacier meltwater streams support some peatlands, and that the ongoing loss of glaciers and their meltwaters could lead to the loss or diminishment of peatlands. We analysed the hydrologic regime of 10 peatlands in four mountain regions of Bolivia and Peru using groundwater monitoring. Groundwater levels in peatlands were relatively stable and within 20 cm of the ground surface during the rainy season, and many sites had water tables 40–90 cm below the ground surface in the dry season. Topographic and groundwater elevations in the peatlands demonstrated that the water source of all 10 peatlands was hillslope groundwater flowing from lateral moraines, talus, colluvium, or bedrock aquifers into the peatlands. There was little to no input from streams, whether derived from glacier melt or other sources, and glacier melt could not have recharged the hillslope aquifers supporting peatlands. We measured the stable water isotopes in water samples taken during different seasons, distributed throughout the catchments, and the values are consistent with this interpretation. Our findings indicate that peatlands in the study region are recharged by hillslope groundwater discharge rather than stream water and may not be as vulnerable to glacial decline as other studies have indicated. However, both glaciers and peatlands are susceptible to changing thermal and precipitation regimes that could affect the persistence of peatlands.     

Magma evolution of Quaternary minor volcanic centres in southern Peru,Central Andes

Minor centres in the Central Volcanic Zone (CVZ) of the Andes occur in different places and are essential indicators of magmatic processes leading to formation of composite volcano. The Andahua–Orcopampa and Huambo monogenetic fields are located in a unique tectonic setting, in and along the margins of a deep valley. This valley, oblique to the NW–SE-trend of the CVZ, is located between two composite volcanoes (Nevado Coropuna to the east and Nevado Sabancaya to the west). Structural analysis of these volcanic fields, based on SPOT satellite images, indicates four main groups of faults. These faults may have controlled magma ascent and the distribution of most centres in this deep valley shaped by en-echelon faulting. Morphometric criteria and ¹⁴C age dating attest to four main periods of activity: Late Pleistocene, Early to Middle Holocene, Late Holocene and Historic. The two most interesting features of the cones are the wide compositional range of their lavas (52.1 to 68.1 wt.% SiO₂) and the unusual occurrence of mafic lavas (olivine-rich basaltic andesites and basaltic andesites). Occurrence of such minor volcanic centres and mafic magmas in the CVZ may provide clues about the magma source in southern Peru. Such information is otherwise difficult to obtain because lavas produced by composite volcanoes are affected by shallow processes that strongly mask source signatures. Major, trace, and rare earth elements, as well as Sr-, Nd-, Pb- and O-isotope data obtained on high-K calc-alkaline lavas of the Andahua–Orcopampa and Huambo volcanic province characterise their source and their evolution. These lavas display a range comparable to those of the CVZ composite volcanoes for radiogenic and stable isotopes (⁸⁷Sr/⁸⁶Sr: 0.70591–0.70694, ¹⁴³Nd/¹⁴⁴Nd: 0.512317–0.512509, ²⁰⁶Pb/²⁰⁴Pb: 18.30–18.63, ²⁰⁷Pb/²⁰⁴Pb: 15.57–15.60, ²⁰⁸Pb/²⁰⁴Pb: 38.49–38.64, and δ ¹⁸O: 7.1–10.0‰ SMOW), attesting to involvement of a crustal component. Sediment is absent from the Peru–Chile trench, and hence cannot be the source of such enrichment. Partial melts of the lowermost part of the thick Andean continental crust with a granulitic garnet-bearing residue added to mantle-derived arc magmas in a high-pressure MASH [melting, assimilation, storage and homogenisation] zone may play a major role in magma genesis. This may also explain the chemical characteristics of the Andahua–Orcopampa and Huambo magmas. Fractional crystallisation processes are the main governors of magma evolution for the Andahua–Orcopampa and Huambo volcanic province. An open-system evolution is, however, required to explain some O-isotopes and some major and trace elements values. Modelling of AFC processes suggests the Charcani gneisses and the local Andahua–Orcopampa and Huambo basement may be plausible contaminants.     

Lead isotopic evidence for evolutionary changes in magma-crust interaction,Central Andes,southern Peru

Lead isotopic measurements were made on Andean igneous rocks of Jurassic to Recent age in Moquegua and Tacna Departments, southernmost Peru, to clarify the petrogenesis of the rocks and, in particular, to investigate the effect of crustal thickness on rock composition. This location in the Cordillera Occidental is ideal for such a study because the ca. 2 Ga Precambrian basement rocks (Arequipa massif) have a distinct Pb isotopic signature which is an excellent tracer of crustal interaction, and because geomorphological research has shown that the continental crust was here thickened drastically in the later Tertiary.Seven samples of quartz diorites and granodiorites from the Ilo and Toquepala intrusive complexes, and seven samples of Toquepala Group subaerial volcanics were analyzed for Pb isotopic compositions. The plutonic rocks range in age from Jurassic to Eocene; the volcanic rocks are all Late Cretaceous to Eocene. With one exception, the Pb isotopic ratios are in the ranges ²⁰⁶Pb/²⁰⁴Pb= 18.52–18.75, ²⁰⁷Pb/²⁰⁴Pb= 15.58–15.65, and ²⁰⁸Pb/²⁰⁴Pb= 38.53–38.74. The data reflect very little or no interaction with old continental material of the Arequipa massif type.Lead from four Miocene Huaylillas Formation ash-flow tuffs, two Pliocene Capillune Formation andesites and five Quaternary Barroso Group andesites has lower ²⁰⁶Pb/²⁰⁴Pb than that in the pre-Miocene rocks, but relatively high ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb (²⁰⁶Pb/²⁰⁴Pb= 18.16–18.30, ²⁰⁷Pb/²⁰⁴Pb= 15.55–15.63, ²⁰⁸Pb/²⁰⁴Pb= 38.45–38.90). Tilton and Barreiro [9] have shown that contamination by Arequipa massif granulites can explain the isotopic composition of the Barosso Group lavas, and the new data demonstrate that this effect is evident, to varying degrees, in all the analysed Neogene volcanic rocks. The initial incorporation of such basement material into the magma coincided with the Early Miocene uplift of this segment of the Cordillera Occidental [32], and thus with the creation of a thick crustal root. The data strongly imply a relationship between crustal thickness and degree of crustal contamination of magmas in this area, but a rigorous relationship is not yet established.     

Petrology of seamounts northwest of Samoa and their relation to Samoan volcanism

Petrological and geochemical data on dredged samples from five submarine volcanos northwest of Samoa indicate that three of these volcanos belong to the Samoan volcanic province (Field, Lalla Rookh, and Combe banks), and two belong to separate magmatic zones (Wallis Islands and Alexa Bank). The Samoan volcanic province increases in age westward and both shield-building tholeiitic and alkalic lavas (Combe Bank) and strongly undersaturated (post-erosional?) melilitites or nephelinites and ankaramites (Field and Lalla Rookh banks) are present. The age progression and petrochemical character of these rocks is consistent with a fixed hotspot beneath eastern Samoa. Slightly askew from this trend is Alexa Bank where dredged lavas are ocean-island tholeiites; however, its radiometric age and compositional characteristics apparently preclude its association with Samoa by a fixed-hotspot model. Dredged volcanic rocks from near the Wallis Islands are geochemically, petrologically, and temporally different from Samoan volcanism, but are similar in these respects to Quaternary volcanism in Rotuma and Fiji and may be related to plate reorganization accompanying opening of the North Fiji Basin.     

Trace element distribution in the cainozoic lavas of Nevado Coropuna and Andagua Valley,Central Andes of Southern Peru

Major and minor elements have been determined on 26 samples of andesitie to rhyolitic lavas from Nevado Coropuna and Andagua valley in Southern Peru. Nevado Coropuna dating back since late Miocene is the highest stratovolcano of Peru. It is located at 150 km NW of Arequipa and at 110 km E of the Pacific coast. Andagua valley is situated at about 30 km E of Coropuna. The magmatic activity there, as shown by the presence of several cones, is more recent than that of Corpouna and is related to the tectonic graben characterizing this valley. The geological position of the valley is very important because it is near the transverse line separating the zone of rather flat subduction of the Nazca plate from another one dipping more steeply to the SE. The lavas from Andagua show higher Ti, P, Sr and alkali contents than those from Coropuna, and several display some alkaline tendency with Na affinity. No shoshonitic rocks have been found in the area. According to their geochemistry, Corpuna and Andagua andesites do not seem to have been originated by a single process. In particular, the distribution of Ni, Cr, Ti, Zr, Y, P, Nb, and Sr would exclude either adirect origin from pyrolitic materials, or aprogressive crustal contamination as the most important factors for their origin. Calculations of mineral/melt equilibria for Coropana andesites suggests crystallization processes at depth less than 35 km and H₂O-understurated conditons at the time of the phenocrysts precipitation, indicating a possible high undersaturation at depth of the source zone. As lar as the rhyolites are concerned, their geochemical characteristics do not preclude a crustal origin. A statistical study of the chemical zonation of the Plio-Quaternary lavas of southern Peru has shown an increase of Ti and P contents eastward of the Chile-Peru trench.     

New data on the Late Cenozoic basaltic volcanism in Syria,applied to its origin

New data on geology and 21 K–Ar dates of the Late Oligocene–Quaternary basalts in Syria, combined with analysis of the new and previous data are used to reconstruct the volcanic history and relations between it and tectonic events. Volcanism began at the end of Oligocene (26–24 Ma) and was concentrated in the Late Oligocene–Early Miocene along a N-trending band, which stretches from the Jebel Arab (Harrat Ash Shaam) up to Kurd Dagh and southern Turkey. Activity waned in the Middle Miocene (17–12 Ma), but was resumed in the same band in the Tortonian and increased in the Messinian and Early Pliocene (6.3–4 Ma), when volcanism spread to the Shin Plateau and its coastal extension. After a brief hiatus ～ 4–3.5 Ma, volcanism became still more intensive and spread from the N-trending band to the east into the northern margin of the Mesopotamian Foredeep and to the west into the Dead Sea Transform zone. Additional eruptions continued into the Holocene.Volcanism lasted > 25 million years in the Jebel Arab Highland and > 15 million years in the Aleppo Plateau. The long duration of volcanism in the same parts of the moving Arabian plate and absence of records of one-way migration of the activity mean that the magmatic sources moved together with the plate, i.e., they were situated within the lithosphere mantle. Coincidence of the tectonic and volcanic stages of the Arabian plate development proves that volcanic activity depended on the geodynamic situation, caused by the plate motion. Situated within the lithosphere, magmatic sources within this transverse band were possibly caused by thermal and deforming influences of the asthenospheric lateral flow, moved laterally from the Ethiopia–Afar deep superplume.     

Hydrocarbon generation in the interiors of Kamchatka and its relationship to volcanism and hydrothermal activity

This paper considers oil-and-gas provinces and the epochs of hydrocarbon generation in Kamchatka. We distinguish the following epochs of hydrocarbon generation: the Early Mesozoic, the Early Cretaceous, the Late Cretaceous, the Early Paleogene, the Late Paleogene, the Miocene, and the Pliocene-Quaternary. We emphasize the fact that all gas generation epochs were related to deep processes, including volcanism and hydrothermal activity.     

Geologic Setting, Field Survey and Modeling of the Chimbote, Northern Peru, Tsunami of 21 February 1996

—Whereas the coast of Peru south of 10°S is historically accustomed to tsunamigenic earthquakes, the subduction zone north of 10°S has been relatively quiet. On 21 February 1996 at 21:51 GMT (07:51 local time) a large, tsunamigenic earthquake (Harvard estimate M _w = 7.5) struck at 9.6°S, 79.6°W, approximately 130 km off the northern coast of Peru, north of the intersection of the Mendaña fracture zone with the Peru–Chile trench. The likely mechanism inferred from seismic data is a low-angle thrust consistent with subduction of the Nazca Plate beneath the South American plate, with relatively slow rupture characteristics. Approximately one hour after the main shock, a damaging tsunami reached the Peruvian coast, resulting in twelve deaths. We report survey measurements, from 7.7°S to 11°S, on maximum runup (2–5m, between 8 and 10°S), maximum inundation distances, which exceeded 500 m, and tsunami sediment deposition patterns. Observations and numerical simulations show that the hydrodynamic characteristics of this event resemble those of the 1992 Nicaragua tsunami. Differences in climate, vegetation and population make these two tsunamis seem more different than they were. This 1996 Chimbote event was the first large (M _w >7) subduction-zone (interplate) earthquake between about 8 and 10°S, in Peru, since the 17th century, and bears resemblance to the 1960 (M _w 7.6) event at 6.8°S. Together these two events are apparently the only large subduction-zone earthquakes in northern Peru since 1619 (est. latitude 8°S, est. M _w 7.8); these two tsunamis also each produced more fatalities than any other tsunami in Peru since the 18th century. We concur with Pelayo and Wiens (1990, 1992) that this subduction zone, in northern Peru, resembles others where the subduction zone is only weakly coupled, and convergence is largely aseismic. Subduction-zone earthquakes, when they occur, are slow, commonly shallow, and originate far from shore (near the tip of the wedge). Thus they are weakly felt, and the ensuing tsunamis are unanticipated by local populations. Although perhaps a borderline case, the Chimbote tsunami clearly is another wake-up example of a "tsunami earthquake."     

The dynamic relation between activities in the Northern and Southern solar hemispheres

The north–south (N/S) asymmetry of solar activity is the most pronounced phenomenon during 11-year cycle minimums. The goal of this work is to try to interpret the asymmetry as a result of the generalized synchronization of two dynamic systems. It is assumed that these systems are localized in two solar hemispheres. The evolution of these systems is considered in the topological embeddings of a sunspot area time series obtained with the use of the Takens algorithm. We determine the coupling measure and estimate it on the time series of daily sunspot areas. The measurement made it possible to interpret the asymmetry as an exchangeable dynamic equation, in which the roles of the driver–slave components change in time for two hemispheres.     

The relationship between glaciation and volcanism: Numerical simulation and the Holocene volcanism in Kamchatka

This study is concerned with numerical simulation of the strain due to glaciation and glacial melting, when a magma zone (a layer containing inclusions of magma and magma cumulates) is present at the crust–mantle boundary. According to analytical solutions of this problem that involves viscous relaxation of an uncompensated depression at the place of the molten glacier, the depth to the zone of increased shear stresses beneath the depression is proportional to its width, while the relaxation duration is proportional to viscosity of the lithosphere and is a few thousand years. These fundamental estimates are corroborated by our numerical simulation. According to it, the magma zone at the Moho boundary shields the zone of increased shear stresses, limiting it from below. The maximum values (12–25 MPa) with glacial thickness 500–1000 m are reached at the top of this layer of low viscosity. The directions of maximum compression (s₁) as calculated for the time after the melting indicate that the magma that rises along dikes is displaced from the center of the magma lens toward its periphery. It is found that glacial unloading makes the dipping faults in the crust above the low-viscosity layer attractors for the rising magma. Glacial unloading accelerates, by factors of a few times, the magma generation in the mantle that occurs following the mechanism of adiabatic decompression, as well as facilitating the accumulation of mantle fluids in the zone of increased shear stresses at the boundary of the low viscosity layer. The magma traverses this deep fluid collector and increases the intensity and explosivity of eruptions at the beginning of an interglacial period. Our numerical simulation results are in general agreement with published data on Early Holocene volcanic eruptions that occurred after the second phase of the Late Pleistocene glaciation in Kamchatka.     

Permian to recent volcanism in northern sumatra,indonesia: a preliminary study of its distribution,chemistry, and peculiarities

Sumatra has been a ‘volcanic arc’, above an NE-dipping subduction zone, since at least the Late Permian. The principal volcanic episodes in Sumatra N of the Equator have been in the Late Permian, Late Mesozoic, Palaeogene, Miocene and Quaternary.Late Permian volcanic rocks, of limited extent, are altered porphyritic basic lavas interstratified with limestones and phyllites.Late Mesozoic volcanic rocks, widely distributed along and W of the major transcurrent.Sumatra Fault System (SFS), which axially bisects Sumatra, include ophiolite-related spilites, andesites and basalts. PossiblePalaeogene volcanic rocks include an altered basalt pile with associated dyke-swarm in the extreme NW, intruded by an Early Miocene (19 my) dioritic stock; and variable pyroxene rich basic lavas and agglomerates ranging from alkali basaltic to absarokitic in the extreme SW.Miocene volcanic rocks, widely distributed (especially W of the SFS), and cropping out extensively along the W coast, include calc-alkaline to high-K calc-alkaline basalts, andesites and dacites.Quaternary volcanoes (3 active, 14 dormant or extinct) are irregularly distributed both along and across the arc; thus they lie fore-arc of the SFS near the Equator but well back-arc farther north. The largest concentration of centres, around Lake Toba, includes the >2000 km³ Pleistocene rhyolitic Toba Tuffs. Quaternary volcanics are mainly calc-alkaline andesites, dacites and rhyolites with few basalts; they seem less variable, but on the whole more acid, than the Tertiary. The Quaternary volcanism is anomalous in relation to both southern Sumatra and adjacent Java/Bali: in southern Sumatra, volcanoes are regularly spaced along and successively less active away from the SFS, but neither rule holds in northern Sumatra. Depths to the subduction zone below major calc-alkaline volcanoes in Java/Bali are 160–210 km, but little over 100 km in northern Sumatra, which also lacks the regular K₂O-depth correlations seen in Java. These anomalies may arise because Sumatra — being underlain by continental crust — is more akin to destructive continental margins than typical island-arcs such as E Java or Bali, and because the Sumatran subduction zone has a peculiar structure due to the oblique approach of the subducting plate. A further anomaly — an E-W belt of small centres along the back-arc coast — may relate to an incipient S-dipping subduction zone N of Sumatra and not the main NE-dipping zone to its W. Correlation of the Tertiary volcanism with the present tectonic regime is hazardous, but the extensive W coastal volcanism (which includes rather alkaline lavas) is particularly anomalous in relation to the shallow depth (<100 km) of the present subduction zone. The various outcrops may owe their present locations to extensive fault movements (especially along the SFS), to the peculiar structure of the fore-arc (suggested by equally anomalous Sn- and W-bearing granitic batholiths also along the W coast), or they may not be subduction-related at all.