The effect of water on stress relaxation of faulted and unfaulted sandstone

A series of stress relaxation experiments have been carried out on faulted and intact Tennessee sandstone to explore the influence of pore water on strength at different strain rates. Temperatures employed were 20, 300 and 400°C, effective confining pressure was 1.5 kb and strain rates as low as 10^–10 sec^–1 were achieved. Most samples were prefaulted at 2.5 kb confining pressure and room temperature. This is thought to have secured a reproducible initial microstructure.The strength of the dry rock was almost totally insensitive to strain rate in the range 10^–4 to 10^–10 sec^–1. In contrast, the strength of the wet rock decreased rapidly with strain rate at rates less than 10^–6 sec^–1. Brittle fracture of the quartz grains which constitute this rock is the most characteristic mode of failure under the test conditions used.The experimental data are discussed in terms of the possible deformation rate controlling processes, and it is suggested that in the wet experiments at intermediate to high strain rates (10^–7 to 10^–4 sec^–1) the observed deformation rate is controlled by the kinetics of water assisted stress corrosion, whilst deformation at low strain rates (ca. 10^–9 sec^–1) is controlled by a pressure solution process.The results have implications for the rheology of fault rocks at depths of perhaps 10 to 15 km in sialic crust.     

The characteristics of rain electricity in Nigeria. I — Magnitudes and variations

The characteristics of rain and point charges based on routine measurements extending over four rainy seasons are presented. An average rain current density of (1.0±0.1)×10^–10 A m^–2 and charge per unit volume of rain water of (0.43±0.02)×10^–4 C m^–3 for the locality are obtained, which are compared with data obtained elsewhere by other workers. The point-discharge current measurements lead to a revised estimate of (0.86±0.08)×10^–9 A m^–2 for the average point discharge current below storm clouds.     

Effect of displacement rate on the real area of contact and temperatures generated during frictional sliding of Tennessee sandstone

Summary The real area of contact has been determined, and measurements of the maximum and average surface temperatures generated during frictional sliding along precut surfaces in Tennessee sand-stone have been made, through the use of thermodyes. Triaxial tests have been made at 50 MPa confining pressure and constant displacement rates of 10^–2 to 10^–6 cm/sec, and displacements up to 0.4 om. At 0.2 cm of stable sliding, the maximum temperature decreases with decreasing nominal displacement rate from between 1150° to 1175°C at 10^–2 cm/sec to between 75° to 115°C at 10^–3 cm/sec. The average temperature of the surface is between 75 and 115°C at 10^–2 cm/sec, but shows no rise from room temperature at 10^–3 cm/sec. At 0.4 cm displacement, and in the stick-slip mode, as the nominal displacement rate decreases from 10^–3 to 10^–6 cm/sec, the maximum temperature decreases from between 1120° to 1150°C to between 1040° to 1065°C. The average surface temperature is 115° to 135°C at displacement rates from 2.6×10^–3 to 10^–4 cm/sec.With a decrease in the displacement rate from 10^–2 to 10^–6 cm/sec, the real area of contact increases from about 5 to 14 percent of the apparent area; the avergge area of asperity contact increases from 2.5 to 7.5×10^–4 cm². Although fracture is the dominate mechanism during stick-up thermal softening and creep may also contribute to the unstable sliding process.     

New terrestrial heat flow values from Hungary

Summary Temperature and conductivity measurements show, that in the Southern part of Transdanubia (the part of Hungary which lies Westwards from Danube) the heat flow is about 2–2.4·10^–6 cal/cm² sec. Eastward from the Danube, in the Hungarian Plain estimates are even higher, and vary between 2.3·10^–6 and 2.8·10^–6 cgs. The gradient of temperature is everywhere quite high, 5.0 resp. 5.8·10^–4 deg. C/cm on the average. Thus, at a depth of 1000 m, the virgin rock temperature is about 60–70 deg. C, at 2000 m about 110–130 deg. C.     

Zero-frequency tidal distrotion due to the orbital elements of tide-forming bodies

Summary It has been proved that orbital elements of perturbing bodies should be taken into account when 10^–10 accuracy is required for zero-frequency tidal distortion in the second zonal Stokes parameter of the geopotential. A solution at the 10^–15 level of magnitude has been presented. The zero-frequency tidal distortion in the fourth zonal Stokes parameter has been derived as 1·3×10^–10 if the secular Love number is unity. It should be reflected in the geopotential models respecting the 10^–10 level of magnitude.     

A portable instrument for the measurement of point-discharge currents in the atmosphere using junction FET's

A contiuous recording instrument with field effect transistors has been developed to study the point discharge currents in the atmosphere. The chief advantages of the present instrument are that it is portable and has a range of measurement from 10^–9 A to 10^–7 A for positive currents and 10^–9 A to 10^–5 A for negative currents on a linear scale.     

Distribution ofBe andBe in the Pacific Ocean

The vertical distributions of¹⁰Be and⁹Be at three locations in the Pacific (25°N, 170°E; 17°N, 118°W; 3°S, 117°W) are presented. The results show that both isotopes exhibit nutrient-like profiles. From the surface to the bottom, the increase for¹⁰Be is two- to threefold and that for⁹Be is about fivefold. While the inter-station variations in surface water concentrations may reach a factor of two, deep-water values tend to be much more uniform averaging about 2000 atoms/g for¹⁰Be and 30 pM for⁹Be. A similar situation applies to the¹⁰Be/⁹Be ratio; it varies approximately from 1 to 3 × 10⁻⁷ (atom/atom) at shallow depths but tends toward a value close to 1.1 × 10⁻⁷ in the deep ocean. The variation of¹⁰Be/⁹Be can be viewed as resulting from the fact that¹⁰Be in a given parcel of water consists of two components: recycled and primary. The recycled component is that part of¹⁰Be which has reached tracer equilibrium with⁹Be, as opposed to the primary component which, upon entering the sea from the atmosphere, has yet to equilibrate with⁹Be through particle cycling and mixing processes. It is estimated that 70% to nearly 100% of¹⁰Be at the three stations are being recycled, and the recycled beryllium bears an atomic ratio of¹⁰Be/⁹Be close to 1 × 10⁻⁷. The oceanic residence time of Be is of the order of 1000–4000 years, comparable to or slightly longer than the ocean mixing time.     

Drifter observations of the Hebrides slope current and nearby circulation patterns

The mean flow at and around the Hebrides and Shetland Shelf slope is measured with ARGOS tracked drifters. Forty-two drifters drogued at 50 m were deployed in three circles over the Hebrides slope at 56.15°N in two releases, one on 5th December, 1995 and the second on 5–9th May, 1996. The circles span a distance of some 20 km from water depths of 200 m to 1200 m. Drifters are initially advected poleward along-slope by the Hebrides slope current at between 0.05 and 0.70 m s^–1 in a laterally constrained (25–50 km wide) jet-like flow. Drifters released in winter remained in the slope current for over 2000 km whilst summer drifters were lost from the slope current beyond the Wyville-Thomson Ridge, a major topographic feature at 60°N. Dispersion from the slope region into deeper waters occurs at bathymetric irregularities, particularly at the Anton Dohrn Seamount close to which the slope current is found to bifurcate, both in summer and winter, and at the Wyville-Thomson Ridge where drifters move into the Faeroe Shetland Channel. Dispersion onto the continental shelf occurs sporadically along the Hebrides slope. The initial dispersion around the Hebrides slope is remarkably sensitive to initial position, most of the drifters released in shallower water moving onto the shelf, whilst those in 1000 m or more are mostly carried away from the slope into deeper water near the Anton Dohrn Seamount. The dispersion coefficients estimated in directions parallel and normal to the local direction of the 500 m contour, approximately the position of the slope current core, are approximately 8.8 × 10³ m² s^–1 and 0.36 × 10³ m² s^–1, respectively, during winter, and 11.4 × 10³ m² s^–1 and 0.36 x 10³ m² s^–1, respectively, during summer. At the slope there is a minimum in across-slope mean velocity, Reynolds stress, and across-slope eddy correlations. The mean across-slope velocity associated with mass flux is about 4 × 10^–3 m s^–1 shelfward across the shelf break during winter and 2 × 10^–3 m s^–1 during summer. The drifters also sampled local patterns of circulation, and indicate that the source of water for the seasonal Fair Isle and East Shetland currents are the same, and drawn from Atlantic overflows at the Hebrides shelf.     

Hydrothermal vent waters at 13°N on the East Pacific Rise: isotopic composition and gas concentration

Gas concentrations and isotopic compositions of water have been measured in hydrothermal waters from 13°N on the East Pacific Rise. In the most Mg-depleted samples ( 5 × 10⁻³ moles/kg) the gas concentrations are: 3–4.5 × 10⁻⁵ cm³ STP/kg helium, 0.62–1.24 cm³ STP/kg CH₄, 10.80–16.71 × 10⁻³ moles/kg CO₂. The samples contain large quantities (95–126 cm³/kg) of H₂ and some carbon monoxide (0.26–0.36 cm³/kg) which result from reaction with the titanium sampling bottles. δ¹³C in methane and CO₂ (−16.6 to −19.5 and −4.1 to −5.5 respectively) indicate temperatures between 475 and 550°C, whereas δ¹³C_CO is compatible with formation by reduction of CO₂ on Ti at 350°C close to the sampling temperature.³He/⁴He are very homogeneous at (7.5 ± 0.1)R_A(³He/⁴He = 1.0 × 10⁻⁵) and very similar to already published data as well as CH₄/³He ratios between 1.4 and 2.1 × 10⁶.¹⁸O and D in water show enrichments from 0.39 to 0.69‰ and from 0.62 to 1.49‰ respectively. These values correspond to W/R ratios of 0.4–7. The distinct¹⁸O enrichments indicate that the isotopic composition of the oceans is not completely buffered by the hydrothermal circulations. The³He-enthalpy relationship is discussed in terms of both hydrothermal heat flux and³He mantle flux.     

Magnetic torque and convection in the earth's core

Summary The problem of expressing analytically the magnetic torque, acting on the electrically conducting part of the Earth's mantle, is treated as a function of the system of convection on the surface of the core. The changes of velocities in the system of convection are estimated for decadic changes of the Earth's rotation and for the perturbation of the Earth's rotation in 1897. As regards the decadic changes of the Earth's rotation a change of velocity in the system of convection at the surface of the core of the order of 10^–4 m/s corresponds, and as regards the perturbation of the Earth's rotation in 1897 (10^–3 s/year) a change of velocity of 10^–3 m/s reduced to the whole surface of the core corresponds, and 10^–2 m/s corresponds for the region of the focus of the world geomagnetic anomaly (dimension of this region is 10⁶ m).     

Mobility spectrometry of the atmospheric ions in relation to atmospheric pollution

Summary The dynamic spectra of the atmospheric ions with mobilities between 3.2×10^–4 m²/V sec and 1.3×10^–8 m²/V sec were obtained in the clean and the polluted atmosphere. Based on the results, the size spectra of the submicron aerosols were deduced which show that the distribution function in Junge's expression increases in proportion to the radius of particle. Very rapid progress of the atmospheric pollution by the submicron aerosols was inferred at a rural country. The concentration of submicron aerosol particles at the site was found to have been increasing at the rate of being doubled every three years since 1964.

---

Zusammenfassung Die dynamischen Spektren atmosphärischer Ionen, deren Beweglichkeiten zwischen 3.2·10^–4 m²/(V sec) und 1.3·10^–8 m²/(V sec) lagen, wurden in reiner und verschmutzter Luft gemessen. Auf der Grundlage dieser Messergebnisse wurden die Grössenspektren der Aerosolteilchen mit Radien unterhalb 10^–6 m abgeleitet. Sie zeigen dass die verteilungsfunktion in der Formel von Junge im Verhältnis zum Teilchenradius anwächst. Sehr starkes Anwachsen der Verschmutzung durch so kleine Teilchen wurde in einer ländlichen Umgebung festgestellt. Es wurde gefunden, dass seit 1964 die Zahl der Verschmutzungsteilchen dieser Grösse an der Messtelle alle drei Jahre sich verdoppelt hat.

     

On the validity of Boltzmann's distribution law for the charges of aerosol particles in electrical equilibrium

Summary The validity ofBoltzmann's energy distribution law as applied to the charge distribution of monodisperse aerosols, i.e. aerosols which contain particles of one size only, was studied with improved equipment now available.Since monodisperse aerosols with particle size of about 10^–6 cm cannot yet be produced and because aerosols in general contain particles of different sizes, polydisperse aerosols stored in a large gasometer were used for the investigation. The composition and the average radius of these aerosols heterogeneous in particle size were determined by the Exhaustion Method employing a diffusion battery without end-pieces or connecting tubing.The experimental curve found for the ratio of the uncharged (N ₀) to the charged (N) nuclei versus radius (r) of the particles deviates for all investigated radii between 0.5 and 4.0·10^–6 cm from the theoretical curve of a monodisperse aerosol computed according to the Boltzmann law. For radii smaller than about 1.4·10^–6 cmN ₀/N is smaller than that given by Boltzmann's law, for radii greater than 1.4·10^–6 cm larger, or, in other words, forr<1.4·10^–6 cm the number of charged nuclei found in the aerosols investigated is greater than that predicted byBoltzmann's law, and forr>1.4·10^–6 cm smaller.The deviations from the theoretical curve forr>1.4·10^–6 cm can be fully explained by the polydispersity of the aerosols used; forr<1.4·10^–6 cm the deviations are too big to be attributed to polydispersity. From this it must be concluded thatBoltzmann's distribution law is not valid for the charge distribution of homogenous aerosols containing nuclei with radii smaller than about 1.4·10^–6 cm.The equivalent radius as deduced from a substitute monodisperse aerosol in charge equilibrium for which theBoltzmann law is assumed valid, deviates in the range of 1.0·10^–6<r<3.0·10^–6 cm (or 90%>100N ₀/Z>55%) on the average by up to 35% from the actual mean radius of the investigated polydisperse aerosols computed from the radii of their singly sized components.

---

Zusammenfassung Die Gültigkeit desBoltzmannschen Gesetzes der Energieverteilung für die Ladungsverteilung von monodispersen Aerosolen, das heisst Aerosolen, welche Teilchen von nur einer Grösse enthalten, wurde mit den jetzt zur Verfügung stehenden verbesserten Mitteln studiert.Da monodisperse Aerosole mit Teilchengrössen von ungefähr 10^–6 cm noch nicht erzeugt werden können und Aerosole im allgemeinen Teilchen von verschiedenen Grössen enthalten, wurden für die jetzige Untersuchung polydisperse Aerosole, welche in einem grossen Gasometer gespeichert waren, benützt. Die Zusammensetzung und der mittlere Radius dieser Aerosole mit heterogener Teilchengrösse wurden nach der Exhaustions-Methode unter Benützung einer Diffusionsbatterie ohne Endstücke oder Verbindungsröhren bestimmt.Die experimentell gefundene Kurve, welche das Verhältnis der ungeladenen (N ₀) und geladenen (N) Kerne als Funktion des Radius (r) darstellt, weicht für alle untersuchten Radien zwischen 0.5 und 4.0·10^–6 cm von der theoretischen Kurve eines monodispersen Aerosols, wie sie nachBoltzmann's Gesetz berechnet wird, ab. Für Radien kleiner als ungefähr 1.4·10^–6 cm ist das VerhältnisN ₀/N kleiner als es durchBoltzmann's Gesetz gegeben wird, für Radien grössen als 1.4·10^–6 cm grösser oder, mit anderen Worten, fürr<1.4·10^–6 cm ist die Zahl der geladenen Teilchen, die in den untersuchten Aerosolen gefunden wurde, grösser als die vonBoltzmann's Gesetz vorausgesagte, und fürr>1.4·10^–6 cm kleiner.Die Abweichungen von der theoretischen Kurve fürr>1.4·10^–6cm können vollkommen durch die Polydispersität der benützten Aerosole erklärt werden; fürr<1.4·10^–6 cm sind die Abweichungen zu gross, als dass sie der Polydispersität zugeschrieben werden könnten. Daraus muss geschlossen werden, dassBoltzmann's Verteilungsgesetz für die Ladungsverteilung eines homogenen Aerosols, welches Kerne mit Radien kleiner als ungefähr 1.4·10^–6 cm enthält, nicht gültig ist.Der äquivalente Radius, wie er von einem im Ladungsgleichgewicht befindlichen, monodispersen Ersatz-Aerosol, für welchesBoltzmann's Gesetz als gültig angenommen wird, abgeleitet werden kann, weicht im Bereich von 1.0·10^–6<r<3.0·10^–6 cm (oder 90%>N ₀/Z>55%) durchschnittlich bis zu 35% vom tatsächlichen mittleren Radius des untersuchten polydispersen Aerosols ab.

---

---

The research reported in this article has been supported in part by the Geophysics Research Directorate of the Air Force Cambridge Research Laboratories, through the European Office of the Air Force Research Division, United States Air Force under Contract AF 61(052)-26, by the United States Department of Army, through its European Research Office, Contract DA-91-591-EUC-1282 & 1657 and by the Instrumentation Engineering Physics & Analysis Laboratory of the General Electric Co., Schenectady, New York under Retainer Agreement.     

Helium and carbon isotopes in fluorites: implications for mantle carbon contribution in an ancient subduction zone

The concentrations of helium and carbon in fluorite associated with Cretaceous to Neogene (90–13 Ma) granitic magmatism in the Japanese arc have been measured. Concentrations of Li, U, Th and Gd were measured to correct for secondary generated ³He. The CO₂/³He of fluorites are almost uniform (1.5×10¹⁰–4×10¹⁰) and in fair agreement with the range of present island arc volcanic gases. The calculated mantle C contribution in the Mesozoic subduction zone appear to have been identical to the present one (7–19%) indicating that the C flux from the mantle in supra-subduction zone environments has remained fairly constant during the past 70 million years.     

Measurement of lower-stratospheric aerosol by a balloon-borne photometer

Height distribution of the stratospheric aerosol extinction coefficient was measured in the altitude range 10 to 20 km by a balloon-borne multi-color sunphotometer in May 1978. It is demonstrated that detailed structures of the distribution of stratospheric aerosol can be remotely measured by the solar occultation method as well as by lidar andin situ particle counter observations. In the aerosol layer appearing at 18 km altitude the extinction coefficient at 800–1000 nm wavelength reached to 3×10^–7 m^–1, which was reasonable compared with lidar observations. Wavelength dependence of the aerosol optical depth was crudely estimated to be proportional to ^–1.5.     

Temporal and spatial variation on codaQ −1 associated with the North Palm Springs earthquake of July 8, 1986

A temporal and spatial change of codaQ ^–1 associated with the occurrence of the North Palm Springs earthquake of July 8, 1986 was studied by using 242 small local earthquakes in the vicinity of the mainshock. We found that the codaQ ^–1 of earthquakes which occurred before the mainshock was significantly higher than that of the aftershocks in the mainshock area while the codaQ ^–1 for the surrounding area remained almost constant throughout 1986. CodaQ ^–1 was determined separately for the lapse time windows of 10 to 20 sec. and 15 to 40 sec. for the period from 1981 to 1987. The result for the time window 10 to 20 sec. showed a peak in codaQ ^–1 before the time of mainshock at all frequencies. The peak appeared earlier at lower frequencies. There was no significant change in codaQ ^–1 for the time window 15 to 40 sec., probably because the change was restricted to a small area.     

A fracture mechanics study of subcritical tensile cracking of quartz in wet environments

Load relaxation and cross-head displacement rate-change experiments have been used to establish log₁₀ stress intensity factor (K) versus log₁₀ crack velocity (v) diagrams for double torsion specimens, of synthetic quartz cracked on thea plane in liquid water and moist air.For crack propagation normal toz and normal tor at 20°C,K _Ic (the critical stress intensity factor) was found to be 0.852±0.045 MN·m^–3/2 and 1.002±0.048 MN·m^–3/2, respectively.Subcritical crack growth at velocities from 10^–3 m·s^–1 to 10^–9 m·s^–1 at temperatures from 20°C to 80°C is believed to be facilitated by chemical reaction between the siloxane bonds of the quartz and the water or water vapour of the environment (stress corrosion). The slopes, of isotherms in theK-v diagrams are dependent upon crystallographic orientation. The isotherms have a slope of 12±0.6 for cracking normal tor and 19.9±1.7 for cracking normal toz. The activation enthalpy for crack propagation in the former orientation in liquid water at temperatures from 20°C to 80°C is 52.5±3.8 kJ·mole^–1.A discussion is presented of the characteristics of theK-v diagrams for quartz.     

Climatological and hydrographic influences on nearshore food webs off the southeastern United States: bacterioplankton dynamics

Bacterioplankton productivity, numbers, and cell specific activity were studied in nearshore waters of the southeastern U.S. continental shelf during seasons of maximum freshwater discharge. In April 1984, coastal waters were stratified from normal spring discharge and typical northeastward wind stress. In April 1985, shelf waters were vertically homogeneous due to below normal runoff and southwestward wind stress. In 1984, nearshore bacterial productivity ranged from 7.0 to 14.7 × 10⁶ cells l⁻¹ h⁻¹ and midshelf rates were 40–50% less. In 1985, nearshore productivity ranged from 0.9 to 2.4 × 10⁶ cells 1⁻¹ h⁻¹, and productivity was extremely patchy over the entire shelf. The cell-specific activity (thymidine incorporation per cell) suggests that although productivity was high in 1984, only a fraction of the bacterioplankton was actively growing or incorporating thymidine (0.9–2.9 × 10⁻²¹ mol cell⁻¹ h⁻¹). In 1985, a higher percentage of cells appeared to be active and incorporating thymidine (5–13 × 10⁻²¹mol cell⁻¹h⁻¹) even though productivity was low. Hydrographic conditions along the southeastern coastline may have had a significant impact on the overall community structure and carbon flow through the microbial food web. When coastal waters were stratified in 1984, bacterial biomass was a significant percentage (35–320%) of the phytoplankton biomass. During vertically homogeneous conditions of 1985, bacterial production and biomass were a small percentage (2–13%) of the phytoplankton production and biomass across the shelf. The interannual variation in the microbial food web was attributed to the interannual variability of the southeastern U.S. hydrology due to changes in freshwater discharge and wind direction and intensity. The ecological implications of these results extend to the potential impact of seasonal microbial food webs on nearshore allochothonous and autochothonous organics before removal from the southeastern U.S. coastline.     

The large volume multi-anvil press as a highP-T deformation apparatus

The rheological properties of mantle materials are being investigated up to pressures of 16 GPa and temperatures of 1600°C for times up to 24 h, using a new sample assembly for the 6–8 multi-anvil apparatus. Al₂O₃ pistons, together with a liquid confining medium, are used to generate deviatoric stress in the specimen. Strain rates are estimated by monitoring the relative displacement of the guide blocks of the multi-anvil apparatus, scaled to the total axial strain of the sample. The applied stress on the sample is estimated using grain size piezometry. Strain rates and flow stresses of approximately 10^–4 to 10^–6 s^–1 and 50 to 250 MPa respectively, are presently attainable.Preliminary results on San Carlos olivine single crystals, partially dynamically recrystallized to a grain size of 10 to 300 m, indicate that the effective viscosity of polycrystalline olivine is consistent with values obtained from olivine single crystal creep laws. Assuming a dislocation creep mechanism (n3.5) with (010)[001] as the dominant slip system, the data are best fit using a creep activation volume of 5 to 10×10^–6 m³ mol^–1.     

The 1999 and 2000 eruptions of Mount Cameroon: eruption behaviour and petrochemistry of lava

Mount Cameroon (4,095 m high and with a volume of ~1,200 km³) is one of the most active volcanoes in Africa, having erupted seven times in the last 100 years. This stratovolcano of basanite and hawaiite lavas has an elliptical shape, with over a hundred cones around its flanks and summit region aligned parallel to its NE--SW-trending long axis. The 1999 (28 March–22 April) eruption was restricted to two sites: ~2,650 m (site 1) and ~1,500 m (site 2). Similarly, in the eruption in 2000 (28 May–19 June), activity occurred at two sites: ~4,095 m (site 1) and ~3,300 m (site 2). During both eruptions, the higher vents were more explosive, with strombolian activity, while the lower vents were more effusive. Accordingly, most of the lava (~8×10⁷ m³ in 1999 and ~6×10⁶ m³ in 2000) was emitted from the lower sites. The 1999–2000 lavas are predominantly basanites with low Ni (5–79 ppm), Cr (40–161 ppm) and mg numbers (34–40). Olivine (Fo_77–85, phenocrysts and Fo_68–72, microlites), clinopyroxene (Wo₄₇En₄₁Fs₁₀ to Wo₅₁En₃₄Fs₁₅), plagioclase (An_49–67) and titanomagnetite are the principal phenocryst and groundmass phases. The lavas contain xenocrysts of olivine and clinopyroxene, which are interpreted as fragments of intrusive rocks disrupted by magma ascent. Major and trace element characteristics point to early fractionation of olivine. The clinopyroxenes (Al₂O₃ 1.36–7.83 wt%) have high Al^iv/Al^vi ratios (1.3–1.8) and are rich in TiO₂, characteristics typical of low pressure clinopyroxenes. Geochemical differences between the 1999–2000 lavas and those from previous eruptions, such as higher Nb/Zr of the former, suggest that different eruptions discharged magmas that evolved differently in space and time. Geophysical and petrological data indicate that these fractionated magmas originated just below the geophysical Moho (at 20–22 km) in the lithospheric mantle. During ascent, the magmas disrupted intrusions and earlier magma pockets. The main ascent path is below the summit, where newly arrived magma degasses. Degassed magma simultaneously intrudes the flank rift zones where most lava is extruded.An erratum to this article can be found at