FIRE and MUD Contents

Mineral and Glass Compositions in June 15, 1991, Pumices: Evidence for Dynamic Disequilibrium
in the Dacite of Mount Pinatubo

By James F. Luhr1 and William G. Melson1

1 Department of Mineral Sciences, NHB-119, Smithsonian Institution, Washington, DC 20560.


ABSTRACT

Whole-rock compositions, textures, and mineral and glass compositions were investigated in two pumices from the June 15 eruption: white, phenocryst-rich, and gray, phenocryst-poor. The gray pumice is slightly poorer in SiO2 and K2O and richer in V, Cr, Co, Ni, Cu, and Zn compared to the white pumice, although more extensive sample suites described elsewhere in this volume show the two pumice types to be virtually identical in composition. The mineral assemblage is similar in both cases (plagioclase, hornblende, titanomagnetite, ilmenite, cummingtonite, biotite, quartz, apatite, anhydrite, sulfides, rare zircon, and xenocrystic olivine plus chromite, augite, and orthopyroxene), but the textures and mineral abundances are very different. The white pumice has large, euhedral phenocrysts surrounded by vesiculated, compositionally homogeneous matrix glass. In the gray pumice, however, most phenocrysts show broken margins, and the vesiculated matrix glass is compositionally heterogeneous and contains abundant tiny crystals; some of these appear to be microlites, but others are clearly broken fragments. Cummingtonite rims are common on hornblende in the white pumice but are rare in the gray pumice.

Compositions of coexisting Fe-Ti oxides indicate temperatures of 845-893°C and 836-842°C for the white and gray pumices, respectively, on the basis of two different published algorithms. Calculated oxygen fugacities fall about 2.3 log units above the Ni-NiO solid oxygen buffer and thus reflect highly oxidized conditions that are consistent with the presence of primary anhydrite. The Fe-Ti oxide temperatures are not consistent with the presence of cummingtonite in the pumices of Mount Pinatubo; experimental studies have shown that cummingtonite is not stable above about 800°C. This discrepancy may reflect a late-stage heating event that caused the Fe-Ti oxides to reequilibrate but was too rapid for the kinetically inhibited reaction of cummingtonite. Alternatively, this discrepancy may reflect inadequate calibration of the Fe-Ti oxide geothermometers for the highly oxidized conditions of the magmas of Mount Pinatubo.

Plagioclase phenocrysts in the white pumice show complex zoning patterns with abrupt outward rises in calcium followed by more gradual declines. Some of these calcium spikes can be correlated from crystal to crystal and may record major magma-mixing events in the earlier history of the magma. Detailed traverses across the outer 100 micrometers of 15 plagioclase growth rims show considerable variability, but the outermost rims are reasonably homogeneous in composition at An40.2+-1.8. Surprisingly, application of plagioclase-melt models by using this rim composition and the average matrix glass of the white pumice shows that the plagioclase rims are too calcic to have been in equilibrium with the matrix glass at any likely temperature. These anomalously calcic rims may have grown under disequilibrium conditions induced by water loss from the magma system during eruptive ventings preceding the June 15 climax.

The gray pumice appears to record quenched disequilibrium that resulted as a dacitic melt-rich magma violently invaded the main, coarsely crystalline dacitic body and began to blend with its high-silica rhyolite melt. The violence of mixing may have caused shattering of phenocrysts from the main dacite. Small microlites were growing from the partially blended melt when the system was frozen upon eruption. As indicated by the plagioclase-melt modeling discussed above, even the dominant coarse-textured dacite from the June 15 eruption may have been out of chemical equilibrium prior to eruption.

The 1991 dacite of Mount Pinatubo and the 1982 trachyandesite of El Chichón are examples of an important class of hydrous, sulfur-rich, anhydrite-bearing magmas that erupt from subduction-zone volcanoes. These eruptions have the potential to inject large quantities of sulfur gases into the stratosphere, where they can play a significant role in modifying the Earth's climate. Recognition of ancient, high-sulfur eruptions in the geologic record is complicated by the rapid dissolution of anhydrite from vesicular pumices in surface waters. The pumices of Mount Pinatubo and El Chichón are strikingly different in whole-rock composition and mineralogy but share several features that might allow analogous eruptions to be identified in the ancient record: high oxidation states, low preeruption temperatures, high crystal contents, abundant hornblende (plus biotite), sulfate-rich apatite, and most importantly, inclusions of anhydrite that are trapped in other phenocrysts and thereby preserved.

Note to readers: Figures and tables open in separate windows. To return to the text, close the figure or table's window or bring the text window to the front.

INTRODUCTION

The climactic eruption of Mount Pinatubo on June 15, 1991, was remarkable both for the mass of magma ejected and for the mass of sulfur gases released to the atmosphere. The eruption ejected at least 4 km3 of dacitic magma, an amount that makes it one of the largest eruptions of the 20th century (Scott and others, 1991); it ranks behind the 1912 Novarupta-Katmai (Alaska) eruption but is of roughly similar magnitude to the 1932 Cerro Azul-Quizapu (Chile) and 1902 Santa María (Guatemala) eruptions, given the uncertainty of tephra volume estimates (Fierstein and Nathenson, 1992). Approximately 20 million tons of SO2 were lofted into the stratosphere by the June 15 eruption (Bluth and others, 1992), the largest stratospheric sulfur injection since satellite-based measurements began in 1978. The stratospheric SO2 combined with water vapor to form a cloud of submicrometer-sized sulfuric acid aerosol droplets that were predicted to lower surface air temperatures by an average of about 0.5°C through 1993 (Hansen and others, 1992b), a prediction that has been borne out by globally integrated air temperature measurements to date (Hansen and others, 1992a; Kerr, 1993). Although quantitative estimates are complicated by the heavy rainfall caused by Typhoon Yunya during the June 15 eruption, studies of leachable sulfate from fresh ash-fall deposits (Bernard and others, this volume) indicate that a substantial additional mass of sulfur gases was released by the eruption but quickly precipitated as tiny sulfate minerals that were adsorbed onto ash particles and carried to the ground. Thus, the total release of sulfur by the eruption was significantly greater than the 20 million tons of SO2 measured in the stratosphere.

Fresh pumices from the 1991 eruption contain euhedral-subhedral microphenocrysts of the water-soluble sulfate mineral anhydrite (Bernard and others, 1991; see also papers in this volume: Bernard and others, Fournelle and others, Hattori, McKibben and others, Pallister and others, Rutherford and Devine). Although anhydrite has long been reported in volcanic rocks, it has usually been interpreted as xenocrystic (Yoshiki, 1933; Kôzu, 1934; Kwano, 1948; Taneda, 1949; Katsui, 1958; Taylor, 1958; Nicholls, 1971; Yagi and others, 1972; Arculus and others, 1983). Anhydrite was first recognized as an important, stable igneous mineral after the 1982 eruption of the Mexican volcano El Chichón (Luhr and others, 1984). The El Chichón eruption injected about 7 million tons of SO2 into the stratosphere, which is second only to Pinatubo in the 14-year record of satellite-based SO2 measurements (Bluth and others, 1992). Leachate studies of fresh El Chichón ashes indicated that a roughly equivalent mass of sulfur gases was rapidly converted to sulfates, adsorbed onto ash particles, and carried to the Earth's surface (Varekamp and others, 1984). El Chichón and Mount Pinatubo represent a newly recognized class of subduction-zone volcanoes whose eruptions of water- and sulfur-rich, oxidized, anhydrite-bearing magmas are capable of playing an important role in short-term modification of the Earth's climate.

The purpose of this study is to contribute toward documentation of the mineralogy, petrology, and geochemistry of the two major types of pumice ejected during the climactic eruption of Mount Pinatubo on June 15, 1991. Our focus is on (1) geochemical and textural comparisons of the white and gray pumices, (2) estimation of preeruptive temperature and oxygen fugacity from analysis of coexisting Fe-Ti oxide compositions and whole-rock ferric/ferrous ratios, (3) compositional zoning profiles in plagioclase crystals, (4) estimation of preeruptive water contents in the melt from plagioclase-glass compositional relations, and (5) glass compositions from the pumice matrices and inclusions in phenocrysts. We emphasize the similarities and differences between the dacite of Mount Pinatubo and the trachyandesite of El Chichón, which we hope will aid in the recognition of sulfur-rich eruptions of this type in ancient deposits from which the primary anhydrite crystals have long since been leached by surface waters.

SAMPLES STUDIED

Our study is based on two pumices that were collected on about July 18, 1991, from the surface of the June 15 pyroclastic-flow deposit just outside the rear gate of Clark Air Base (the area called Mactan in other reports: 15°10.2'N., 120°29.0'E.). The deposit was about 5 m thick; at the time of collection the interior still had a temperature of several hundred degrees centigrade, but the surface was cool and rainwashed. One sample is white and has relatively large and abundant phenocrysts (USNM# 116534-2) and will subsequently be referred to as the "white" pumice. It is equivalent to the "phenocryst-rich" pumices discussed elsewhere in this volume. Prior to our studies, the white-pumice sample was approximately 12x12x6 cm in size and weighed about 300 g. The other specimen (USNM# 116534-1) is light gray with phenocrysts that are both smaller and less abundant than in the white pumice; it will be referred to as the "gray" pumice and is equivalent to the "phenocryst-poor" types discussed elsewhere in the volume. The sample was 10x8x6 cm in size and weighed about 200 g. As discussed by Pallister and others (this volume), the white pumices are the dominant variety and account for about 85 percent of the June 15 lapilli, with the gray types, which typically have breadcrust surfaces, accounting for the remainder. Inclusions of white pumice are commonly found in gray pumices, and the two types can be mingled together to form banded pumices. Investigations of vertical sections through the tephra-fall sequence demonstrate that the two pumice types were ejected together throughout the eruption, with the gray pumices decreasing slightly in abundance with time from about 20 percent of the lapilli near the base of the sequence to about 13 percent near the top (David and others, this volume).

ANALYTICAL TECHNIQUES

Slabs were cut from the two pumices for preparation of polished sections and for whole-rock analysis. The powders were prepared by grinding small pumice chips in a shatterbox with alumina puck and container until the powder passed completely through a 100-mesh nylon seive. The powders were dried at 110°C and then at 1,000°C, with weight losses reported as H2O- and LOI (loss on ignition), respectively, in table 1. The dehydrated powder was then combined with Li-tetraborate and fused to a glass disk for X-ray fluorescence (XRF) determination of TiO2, Al2O3, Fe2O3total, MnO, Na2O, K2O, and P2O5 abundances using the Smithsonian Institution's Philips PW-1480 spectrometer (table 1). Values for SiO2, FeO, CaO, and MgO were determined by wet chemical techniques on the original powder. Analyses of S, Cl, and nine other trace elements were performed by XRF on separate disks prepared by pressing rock powder dried at 110°C with cellulose. Twenty additional trace element abundances were determined by instrumental neutron activation (INA) analysis of the original powders at Washington University (Lindstrom and Korotev, 1982).

Electron microprobe analyses were conducted on the Smithsonian's 9-spectrometer ARL-SEMQ instrument, with 15 kV accelerating potential, a specimen current of 15 nA on brass, natural and synthetic standards, on-peak background corrections, and Bence-Albee interelement corrections. The beam was focused and stationary for analysis of Fe-Ti oxides and plagioclase, focused and manually moved for analysis of all glass inclusions and matrix glass in the gray pumice, and defocused and manually moved for the matrix glass in the white pumice.

Table 1. Whole-rock analyses of June 15, 1991, pumices.
["White" is phenocryst-rich white pumice, USNM# 116534-2. "Gray" is phenocryst-poor gray pumice, USNM# 116534-1]

WHOLE-ROCK COMPOSITIONS

The June 15 pumices from table 1 both are classified as medium-K dacites by the scheme of Gill (1981). The whole-rock sulfur contents of the white and gray pumices (0.17 and 0.03 wt% SO3, respectively) can be used to estimate the amount of anhydrite in the pumices, assuming that anhydrite (with 59 wt% SO3) contains virtually all of the sulfur in the samples. This calculation indicates 0.29 and 0.05 wt% anhydrite for the white and gray pumices, respectively. Other whole-rock pumice analyses from Bernard and others (this volume), Fournelle and others (this volume), and Pallister and others (this volume) show 0.25-0.48 wt% SO3 (0.42-0.81 wt% anhydrite) for white pumices and 0.13-0.35 wt% SO3 (0.22-0.59 wt% anhydrite) for gray pumices; these values are all higher than those measured in our study. The relatively large ranges for SO3 may reflect primary variations in the otherwise compositionally homogeneous dacites of Mount Pinatubo. Given the rapid dissolution of anhydrite in surface waters and the fact that Typhoon Yunya was crossing Luzon at the time of the climactic eruption, however, this spread of sulfate contents is perhaps more likely a result of different degrees of anhydrite dissolution, with the samples analyzed in our study having undergone the greatest posteruption leaching.

In comparing the white and gray pumice analyses in table 1, the white pumice is seen to be slightly less mafic: richer in Ba, Ta, and U and poorer in MgO, V, Ni, Cu, Sc, Cr, and Co. When evaluated against the larger data set of Pallister and others (this volume), however, no systematic differences between the white and gray pumices are found.

MINERAL ASSEMBLAGES AND TEXTURES

Both pumice types contain the same assemblage of stable minerals: plagioclase, hornblende, titanomagnetite, ilmenite, cummingtonite, biotite, quartz, apatite, anhydrite, sulfides, and zircon. Cummingtonite occurs exclusively as rims on hornblende; it is common in the white pumice but is rare in the gray pumice. The typical rounding of quartz crystals indicates that they may have become unstable shortly before the eruption. The fact that major element compositions of glass inclusions within quartz crystals are virtually indistinguishable from compositions of glass inclusions within other phenocrysts and the matrix glass of the white pumice (see table 4; Westrich and Gerlach, 1992; Gerlach and others, this volume) demonstrates that the quartz crystals grew from the dacite of Mount Pinatubo and are not xenocrystic. Hattori (this volume) gives a detailed treatment of sulfides in the 1991 products; those present in the dacites are Cu-Fe sulfides and pyrrhotite. Minor zircon has been identified as inclusions in both anhydrite and Fe-Ti oxides (Bernard and others, 1991; Matthews and others, 1992). Rare xenocrysts in the dacites include olivine (with chromite inclusions), augite, and orthopyroxene. Pallister and others (this volume) determined that these xenocrysts do not have appropriate compositions to have originated in the andesites and basalts erupted during the days prior to June 15 and suggested that they may represent crystals from the underlying Zambales ophiolite complex, a source also advocated by Fournelle and others (this volume).

Despite their similarities in mineralogy, the white and gray pumices differ greatly in texture. The white pumices have large, relatively unshattered phenocrysts of plagioclase, hornblende, and quartz surrounded by clear vesiculated glass that is largely free of microlites. Some plagioclase and quartz grains exceed 3 mm in diameter. In the gray pumices, most crystals show at least one broken margin, and crystal sizes are generally less than 1 mm. The vesiculated glassy matrix is charged with small crystals. Back-scattered electron photomicrographs of the vesiculated matrix glasses from the white and gray pumices are shown in figure 1 at two different magnifications; similar photos and observations are found in Pallister and others (this volume). Figure 1A shows a euhedral plagioclase phenocryst surrounded by vesiculated, crystal-free glass. The largest glass patch from this scene is magnified in figure 1C. Figure 1B shows a broken plagioclase phenocryst with truncated compositional zonation surrounded by vesiculated glass containing many small plagioclase and hornblende crystals. One of the largest glassy areas is magnified in figure 1D. The small crystals in the gray-pumice matrix glass appear to include both euhedral microlites, such as the one labeled hornblende with slight swallow-tail terminations in figure 1D, and anhedral fragments broken from larger crystals. Although hornblende is common among these tiny crystals, we have not identified cummingtonite either as euhedral microlites or as crystal fragments. It seems to be restricted to overgrowths on the rims of hornblende phenocrysts.

Figure 1. Back-scattered electron images of white (A and C) and gray pumices (B and D). Abbreviations: g, glass; p, plagioclase; h, hornblende. The white "p" and black "p" on the plagioclase of B identify darker (calcium-rich) and lighter (sodium-rich) compositional zones, respectively, that are separated by a planar interface and truncated by the broken crystal margin. C is the largest glass patch from A at higher magnification; D is one of the largest glassy areas of B at higher magnification.

Mineral abundances were determined for the two pumices by both point counting and least-squares modeling, with results given in table 2. A single polished section of each pumice was point counted under combined transmitted-reflected light. Following the criteria of Wilcox (1954), phenocrysts (>0.3 mm), microphenocrysts (<0.3 mm, >0.03 mm), and groundmass (<0.03 mm) were distinguished and counted separately. The point-counting results (volume percent) were then converted to weight percent for comparison with weight percent modes calculated by least-squares modeling (see table 2 for methods).

Table 2. Modal abundances by point counting and least-squares calculation.
[Mineral abbreviations: Plag, plagioclase; Hbd, hornblende; Oxid, Fe-Ti oxides (titanomagnetite plus ilmenite); Anhy, anhydrite; Apat, apatite; Qtz, quartz. Textural abbreviations: ph, phenocrysts (>0.3 mm); mp, microphenocrysts (>0.03 mm, <0.3 mm: after Wilcox, 1954); tr, trace. Both samples also contain trace amounts of cummingtonite rims on hornblende, biotite, sulfides, zircon inclusions, and xenocrysts of olivine with chromite inclusions, augite, and orthopyroxene]

The grain-size differences of the two pumice types are clearly reflected in the point-count data. The white pumice contains 44.2 vol% phenocrysts and 47.2% groundmass glass plus crystals, whereas the gray pumice has only 7.1% phenocrysts and 71.4% groundmass material. Although the point-counted and modeled modes are quite similar for the white pumice, the results are quite discordant for the gray pumice. This difference stems from the fact that the groundmass material for the white pumice is largely homogeneous glass, whereas for the gray pumice it consists of heterogeneous glass with abundant crystal fragments and microlites. An attempt was made to count vesicles, although distinguishing thin glass septa from vesicles, even in reflected light, is very difficult. A more reliable estimate of the vesicularity can be obtained from the average densities calculated for vesicle-free pumices from modes and phase densities in table 2 (~2.65 g/cm3), coupled with actual pumice densities measured by Pallister and others (this volume): white = 0.819 g/cm3 and gray = 0.977 g/cm3. These data indicate vesicularities of 31 vol% and 37 vol% for the white and gray pumices, respectively, compared with the point-counted values of 46 vol% and 53 vol% in table 2.

Fe-Ti OXIDE COMPOSITIONS AND ESTIMATES OF TEMPERATURE AND OXYGEN FUGACITY

Discrete crystals of titanomagnetite and ilmenite are present in both pumices. These Fe-Ti oxides are homogeneous and unzoned within each sample and show no significant differences between the two pumices. Electron microprobe analyses of the oxides in both samples are given as mean and 1lc sigma values in table 3, along with recalculations of Fe2O3, FeO, and mineral formulas after Stormer (1983). These mean analyses were used to calculate temperatures and oxygen fugacities based on two different algorithms, Andersen and others (1993) and Ghiorso and Sack (1991), with the results listed in table 3; these methods yield temperatures of 845°C and 893°C, respectively, for the white pumice, and 836°C and 842°C, respectively, for the gray pumice. Similar temperatures are obtained from the same geothermometers when using Fe-Ti oxide data in Pallister and others (this volume) and Rutherford and Devine (this volume). As noted in both of those studies, however, these temperatures are inconsistent with the experimentally based upper temperature limit of about 790°C for cummingtonite stability at pressures of 2-3 kbar (Geschwind and Rutherford, 1992).

Table 3. Electron microprobe analyses of titanomagnetite and ilmenite. [Mineral abbreviations: tmt, titanomagnetite; ilm, ilmenite; ulv, ulvöspinel. n indicates the number of individual spot analyses included in the mean and 1lc sigma values. T and -log fO2 values were calculated from the mean analyses using programs provided by Andersen and others (1993: QUILF Version 4.1, selected reactions: FeMgIlSp, FeMnIlSp, FeTi, and MH) and Ghiorso and Sack (1991). DNNO values are deviations from the Ni-NiO buffer of Huebner and Sato (1970), calculated for P=2,000 bar. Their expression for the NNO buffer is: logfO2=-24930/T+9.36+0.046x(P-1)/T, with T in Kelvin and P in bars]

This discrepancy might reflect a late-stage heating event that caused the Fe-Ti oxide compositions to reequilibrate but was unable to melt the cummingtonite or convert it to orthopyroxene + quartz as a result of sluggishness of those reactions. This heating might have been related to the intrusion of basalt into the Pinatubo system, which has been invoked as a trigger to the 1991 eruptions (Pallister and others (1992; this volume). Experimental data in Fonarev and Korolkov (1980) show that metastable cummingtonite compositions persisted up to 21 days at temperatures of 760-780°C. A late-stage heating event might also explain the corroded outlines of quartz crystals in the dacites of Mount Pinatubo; the experimental data of Rutherford and Devine (this volume) show the upper thermal stability limit of quartz in water-saturated systems to be 790°C at PH2O=3 kbar and 820°C at PH2O=2 kbar. Late-stage heating might also provide an explanation for the apparent disequilibrium between plagioclase and matrix glass, discussed in a later section.

An alternative explanation for the discrepancy between Fe-Ti oxide temperatures and the thermal stability limit of cummingtonite has been put forward by Geschwind and Rutherford (1992) and Rutherford and Devine (this volume). They noted that the highly oxidized conditions pertinent to the magmas of Mount Pinatubo lie outside the experimental calibration of the Fe-Ti oxide geothermometers and showed that for experimental charges equilibrated at these high-fO2 conditions, the geothermometer of Andersen and others (1993) yields temperatures that are about 30°C too high. They have suggested, therefore, that Fe-Ti oxide temperatures calculated from this geothermometer for highly oxidized systems should be adjusted downward by 30°C. Although the analyses in table 3 would still yield temperatures for the geothermometer of Andersen and others (1993) that are up to 20°C too high for cummingtonite stability, this difference is not significant when considering likely errors for Fe-Ti oxide geothermometry. The formulation of Ghiorso and Sack (1991), however, yields an unacceptably high corrected temperature of 963°C for the white pumice.

The algorithms of Andersen and others (1993) and Ghiorso and Sack (1991) yielded estimates for oxygen fugacity that are listed in table 3 both as -log fO2 values and as log-unit deviations from the Ni-NiO (NNO) solid oxygen buffer. The latter method of comparison is convenient in that it removes temperature-dependent variations in the calculated oxygen fugacities. The two algorithms are very consistent in indicating oxygen fugacities about 2.3 log units above NNO. These values are consistent with the presence of anhydrite in the 1991 pumices, given the experimentally determined lower fO2 limit for anhydrite stability of 1-1.5 log units above NNO (Carroll and Rutherford, 1987).

Sack and others (1980) and Kilinc and others (1983) presented an alternative method for estimating magmatic oxygen fugacity based on the Fe3+/Fe2+ value and major element composition of glass quenched at known temperature. We have not determined the Fe3+/Fe2+ values for matrix glasses in this study but, rather, use the whole-rock compositions from table 1 to make these calculations. The algorithm of Kilinc and others (1983) yields oxygen fugacities 2.5 and 2.7 log units above NNO for the white and gray pumices, respectively, which are indistinguishable from those estimates that are based on Fe-Ti oxide compositions.

On the basis of phase equilibrium experiments, Rutherford and Devine (this volume) have modeled the magma of Mount Pinatubo as a vapor-saturated system at a pressure of about 2 kbar. At that pressure and an oxygen fugacity 2.3 log units above NNO, thermodynamic data of Helgeson and others (1978) can be used to calculate the fugacity ratio fSO2/fH2S for the gas phase, which increases from about 5 to 32 over the temperature range from 800°C to 900°C. This calculation shows that SO2 was the dominant sulfur gas species in the magma, consistent with the release of some 20 million tons of stratospheric SO2 as detected by the satellite-borne Total Ozone Mapping Spectrometer (TOMS) shortly after the eruptions (Bluth and others, 1992). The trachyandesite erupted by El Chichón volcano in 1982 was slightly less oxidized (NNO + 1 log unit), and although it was anhydrite saturated and also produced a large cloud of SO2 in the stratosphere, calculations similar to those described above indicate that H2S was the dominant sulfur species, with the fugacity ratio fSO2/fH2S of about 0.06 (Luhr, 1990). Comparison of the Pinatubo and El Chichón systems demonstrates the sensitivity of sulfur speciation to oxygen fugacity and the fact that SO2 is not necessarily the dominant sulfur species in the gas phase of all anhydrite-saturated magmas. Regardless of the original sulfur speciation, however, the ultimate fate of gaseous sulfur carried to the stratosphere is to become oxidized and hydrated to form aerosol droplets of sulfuric acid (McKeen and others, 1984).

MATRIX GLASS COMPOSITIONS

As shown in figures 1A and C, the matrix of the white pumice contains relatively broad septa of homogeneous, crystal-free glass that could be analyzed easily by microprobe. Analysis 1 in table 4 gives the mean and 1lc sigma values for 11 moving, defocused spot analyses of the matrix glass. As shown on figure 2, a plot of SiO2 versus K2O, this mean composition is very similar to white-pumice matrix glass analyses reported elsewhere (Gerlach and others, this volume; Matthews and others, 1992; Pallister and others, this volume; Rutherford and Devine, this volume), and this similarity demonstrates the homogeneous nature of the matrix glass.

Table 4. Electron microprobe analyses of glass and plagioclase (in weight percent).

Figure 2. Matrix glass (solid symbols) and glass inclusion (open symbols) analyses from this study and literature sources (Gerlach and others, this volume; Matthews and others, 1992; Pallister and others, this volume; Rutherford and Devine, this volume) shown as SiO2 versus K2O, with all data normalized anhydrous using Fe3+=0.44xFetotal, which is appropriate for log fO2=NNO+2.3 (Kilinc and others, 1983). Stippled balloons enclose all analyses for matrix glasses from the white and gray pumices. Estimates of glass content on the right-hand y-axis were calculated assuming that both pumices contain 39 wt% plagioclase (with 0.28 wt% K2O) and 13.5 wt% hornblende (with 0.22 wt% K2O) and that all remaining K2O in the whole-rock analyses (1.38 wt% K2O=1.52 wt% (table 1) minus 0.14 wt% from plagioclase and hornblende) resides in matrix glass. Thus, matrix glass content (wt%)=100x1.38/(K2O in glass). The horizontal arrows mark the lowest and highest estimated glass contents for the white (44-48 wt%) and gray (43-57 wt%) pumices.

The white-pumice matrix glass analysis listed in table 4 has 1.35 wt% corundum in the CIPW norm; normative corundum is characteristic of melt compositions that have undergone considerable hornblende fractionation (Cawthorn and Brown, 1976; Zen, 1986). Some workers have interpreted normative corundum in glass analyses as a reflection of sodium loss during analysis and have corrected the analyses by adding sodium until the normative corundum disappears (Merzbacher and Eggler, 1984). Although this technique may be appropriate for hornblende-free systems, it is clearly inappropriate for hornblende-rich rocks such as the dacites of Mount Pinatubo.

Glass inclusions are a conspicuous feature of many phenocrysts in the white pumice. Analyses 2, 3, and 4 in table 4 are moving, focused spot analyses on large glass inclusions in titanomagnetite, quartz, and plagioclase, respectively. As shown in figure 2, the glass inclusion in quartz and several other glass-inclusion analyses from the literature are virtually identical to the matrix glass, whereas other glass-inclusion analyses from this study and the literature fall to both higher and lower SiO2 contents (fig. 2: open circles). In general, though, glass inclusions trapped in phenocrysts of the white pumice show compositions very similar to the enclosing matrix glass.

In contrast to the matrix glasses in the white pumice, gray-pumice matrix glasses range widely in composition (analyses 5-7: table 4). Analysis 5 is virtually identical to the white-pumice matrix glasses, whereas analyses 6 and 7 and three other gray-pumice matrix glass analyses from Pallister and others (this volume) range downward over nearly 10 wt% in SiO2 (fig. 2: filled squares). Two analyses of glass inclusions in plagioclase from the gray pumice (analyses 8 and 9: table 4) fall in the high-silica part of this range (fig. 2: open squares). A similar relation was noted in some of the pumices erupted from Nevado del Ruiz in 1985 (Melson and others, 1990) and interpreted to indicate remelting and partial assimilation of a highly crystalline, cooler carapace by a hotter andesitic magma, which itself was probably heated by and partially mixed with invading basaltic magma.

The K2O contents of the whole-rock pumices (1.5 wt%: table 1) and matrix glasses (table 4) can be used to estimate the abundance of glass because most of the K2O is partitioned into glass during crystallization. Among the major crystalline phases, only plagioclase and hornblende contain significant K2O: 0.28 and 0.22 wt%, respectively. The least-squares model for the white pumice shown in table 2 indicates 39 wt% plagioclase and 13.5 wt% hornblende. For simplicity we have assumed the same amounts of plagioclase and hornblende in both the white and gray pumices; the calculated glass abundances are very insensitive to variations in these modal estimates. The right-hand y-axis to figure 2 shows the glass contents estimated for various matrix-glass K2O values, with horizontal arrows marking the limits for the white (44-48 wt% glass) and gray pumices (43-57 wt% glass). The gray pumices appear to represent mixtures of different magma batches with quite different crystallinities.

COMPOSITIONAL ZONING IN PLAGIOCLASE

Plagioclase compositions were studied in considerable detail in the white pumice. Care was taken to identify crystals with prominent growth bands outward to a euhedral face in contact with vesiculated glass. A few plagioclase analyses were also made for the gray pumice, but no large phenocrysts are present, and the broken margins of most crystals made the identification of growth rims difficult.

For about a dozen euhedral crystals in the white pumice, automated traverses were made from rim to rim at a spacing of 2 lc mum. Three examples of relatively symmetrical compositional traverses are shown in figure 3. Traverse 105 was made across a 215-m euhedral crystal. The core of An40-42 is surrounded by a mantle of An45-47, from which the composition drops sharply outward to An37 before rising to An40 at the very rim. Traverse 109 was made across a 450-lc mum euhedral crystal. It shows most of the same general features found in traverse 105 (as indicated by dashed correlation lines) but also shows a core area that rises up to An48, apparently reflecting plagioclase crystallization prior to nucleation of the crystal shown in traverse 105. A still earlier stage of plagioclase growth is evident in traverse 110, taken across a 1,100-m euhedral phenocryst. Its core area ranges from An31 to An39. An abrupt increase to An45 is followed outward by a general decline to An35 before another abrupt increase to An50-52. This latter zone is tentatively correlated with the core of the crystal in traverse 109. Abrupt outward increases in An content such as those in traverses 109 and 110 were observed in many other plagioclase crystals. They may reflect mixing of calcium-rich mafic magma into a more evolved magma body, such as the scenario advocated by Pallister and others (1992; this volume) for triggering of the 1991 eruption. If so, these recurrent calcium spikes provide evidence for repeated mixing events in the history of the 1991 dacite.

Figure 3. Rim-to-rim compositional traverses across three euhedral plagioclase crystals in the white pumice, with a point spacing of 2 lc mum. Traverse A (105) is across a 215-µm euhedral crystal; Traverse B (109) is across a 450-lc mum crystal; Traverse C (110) is across a 1,100-lc mum crystal. Dashed lines show tentative correlations of calcium-rich zones. Mol% An=100xCa/(Ca+Na).

Although plausible correlations could be made from zone to zone among the three crystals shown in figure 3, other crystals could not be correlated with them easily. Our study of this problem suffers from reliance on conventional polished sections. A proper investigation of correlated zoning changes in plagioclase would involve traverses made across crystals that were separated from the rock, oriented crystallographically, and polished down to a plane through the crystal center. Our results offer encouragement to a future study of this type, which might reveal important details of the early magmatic history of the dacite of Mount Pinatubo. Also, the detailed stratigraphy revealed by combined electron microprobe and Normarski- and laser-interference imaging as used by Pearce and others (1987) would be informative.

Coexisting plagioclase and glass compositions can be used to estimate the preeruptive water contents of the melt (Housh and Luhr, 1991) and, accordingly, one of the main purposes of analyzing plagioclase in the 1991 pumices was to determine the composition of plagioclase that was in equilibrium with the matrix glass just prior to eruption. In order to address this question, 15 traverses of 100-lc mum length and 2-lc mum spacing were made across euhedral plagioclase rims in the white pumice. These traverses are shown in figure 4. Traverses A and B, C and D, and N and O are for different rims of the same crystal. These pairs show broadly similar patterns of compositional variation, but not all details of the traverses can be correlated. The problem is even greater when comparing traverses from one crystal to another. Although most of the traverses in figure 4 were intentionally placed above and below others of broadly similar form, when taken as a group, it is clear that these crystal rims show quite a diversity of compositional zoning. It appears that zoning in plagioclase may be affected mostly by the local environment, with only major compositional changes (such as the large Ca spikes of fig. 3) correlatable from crystal to crystal. We note, however, that crystals from quite different preeruption crystallization sites could have been mixed during eruption and possibly before eruption by convection and other movements in the magma.

Figure 4. Compositional traverses across the outermost 100 lc mum of 15 different plagioclase rims in the white pumice, with a point spacing of 2 lc mum. Lines connecting traverses A, and B, C and D, and N and O indicate that these pairs are for different faces of the same crystal. Small labels give the traverse number and the An content of the outermost rim. Mol.% An=100xCa/(Ca+Na).

Despite the dissimilar zoning profiles in plagioclase crystal rims, the outermost rim compositions show relatively little variation. The composition of the outermost plagioclase rim for each traverse is labeled on figure 4. These range from An37.2 to An43.0. The mean An content is 40.2 with 1lc sigma of 1.8. The mean analysis and 1lc sigma values are given in table 4.

PLAGIOCLASE-MELT EQUILIBRIA AND ESTIMATED WATER CONTENTS IN THE MELT

Housh and Luhr (1991) gave two different algorithms for calculating water contents of melts that are based on solution of albite and anorthite exchange reactions between coexisting plagioclase and melt. The calculated melt water contents are strongly dependent upon the assumed temperature of equilibration, which must be known independently. Figure 5 shows the results of these calculations for the white-pumice matrix-glass analysis from table 4, with plagioclase composition plotted against calculated melt water content. The solid lines show predicted values for the albite equation, and the dashed lines show solutions for the anorthite exchange reaction. Different temperatures of solution are indicated, and they bracket the temperature range discussed above, which is from 780°C to 900°C. The large dots show values of mole fraction anorthite in plagioclase (XAn) and melt H2O content for which albite and anorthite solutions converge. Housh and Luhr (1991) estimated the errors on melt water content to be +-0.54 wt% and +-0.33 wt% for the albite and anorthite equations, respectively, so an exact match of the two curves is an unrealistic expectation. The lightly stippled field bounded by the other dotted line shows the range of permissible An-H2O solutions, given these error limits. The vertical line and bounding stippled field indicate the mean plagioclase rim composition for the white pumice (table 4) and 1lc sigma variations.

Figure 5. Plagioclase composition (Mol% An=100xCa/(Ca+Na)) versus melt water contents calculated from the albite (solid lines) and anorthite (dashed lines) exchange reactions between plagioclase and melt from Housh and Luhr (1991) at the indicated temperatures. The matrix glass composition for the white pumice from table 4 was used in all calculations. The large black dots and connecting dotted line indicate the locus of T-XAnpl-H2O points for which the two equations give the same melt water contents. The dotted line bounding the lightly stippled field is the locus of points of maximum An content agreement for the two equations when the errors given by Housh and Luhr (1991) for the albite (+-0.54 wt% H2O) and anorthite (+-0.33 wt% H2O) reactions are considered. Thus, the lightly stippled field represents acceptable agreement between the two equations. The vertical line and bounding heavily stippled field show the mean plagioclase rim composition and 1lc sigma variations from table 4. Note the minimal overlap of the two stippled fields, which indicates apparent lack of equilibrium between the matrix glass and plagioclase rims.

Even considering the errors indicated by the lightly stippled field, the plagioclase rim compositions are much too calcic to have been in equilibrium with the mean glass composition at any likely temperature and melt water content. The model of Housh and Luhr (1991) was calibrated with experiments that bracket the plagioclase and melt compositions found in the pumices of Mount Pinatubo, and thus is expected to be applicable. It might be argued that the zone of plagioclase that actually equilibrated with the matrix glass composition under preeruptive conditions is not the outermost layer but lies inward from the rim. In this scenario, the outermost layers would have grown after or during eruptive degassing. In that case, however, one would expect strong normal zoning of the rims if there was an approach to equilibrium crystallization, which is not seen in most crystals (figure 4). The lack of consistency in the zoning patterns of plagioclase rims makes it very difficult to evaluate this hypothesis. Clearly, zones of appropriate composition (An31-34) can be found in individual plagioclase crystals, but these cannot be correlated from one crystal to another.

If we accept a temperature of 780°C for the magma prior to eruption, consistent with the stability of cummingtonite, and use the mean glass and plagioclase analyses from table 4, the albite equation indicates about 6.2 wt% H2O in the melt, whereas the anorthite equation yields just 4.2 wt% H2O. As discussed below, the albite value is actually quite close to melt water content estimates based on ion-microprobe analysis of trapped glass inclusions. The dilemma posed by these results is the wide discrepancy between the albite and anorthite solutions, a discrepancy that is much greater than estimated errors for the technique. These problems in application of the Housh and Luhr (1991) model to the white pumices of Mount Pinatubo appear to indicate that the plagioclase rims and matrix glass were not in chemical equilibrium at the time of eruptive quenching, a surprising conclusion, given the compositional homogeneity of both phases.

We can propose one viable mechanism to account for these seemingly anomalous calcic rim compositions. During cooling experiments, Lofgren (1980) produced reversely zoned plagioclase margins with disequilibrium compositions by suddenly lowering temperature to a new static value, which induced crystallization. He noted that high degrees of undercooling can also be produced by sudden loss of water from the system, a relevant process for the Pinatubo case. Growth zoning under such conditions is controlled by interface-kinetic and diffusion-rate factors and can lead to compositions far from those expected under equilibrium crystallization. Smith and Brown (1988) reviewed the problem of plagioclase zone compositions at large degrees of undercooling and found that initial plagioclase compositions are consistently more calcic than expected from equilibrium relations. The calcic rims of the plagioclases in the white pumice of Mount Pinatubo may reflect such a process, induced by rapid water loss during the eruptive events preceding June 15, 1991.

Two papers in this volume reported estimates for preeruptive water contents of the melt based on direct ion microprobe measurement of H2O in trapped glass inclusions: Gerlach and others, and Rutherford and Devine. Many host crystals develop cracks during eruptive cooling, and when these intersect the melt/glass inclusions, leakage of H2O and other volatiles can result. In addition, minerals with strong cleavages, such as hornblende and plagioclase, can be particularly prone to leakage. Thus, the most reliable data are expected for strong minerals with poor cleavage, such as quartz or olivine. All three of the above studies reported ion microprobe H2O values for glass inclusions in quartz, and these range from 4.4 to 6.3 wt%; it is expected that the highest values are the most reliable. As seen on figure 5, melt water contents approaching 6 wt% would only be consistent with both the albite and anorthite solution models at 780°C if the plagioclase rim composition was in the range of An30-33. As shown in figures 3 and 4 and discussed above, however, plagioclase of this composition is not common in the outer 100 lc mum of plagioclase crystals in the white pumice.

DISCUSSION

ORIGIN OF THE GRAY PUMICES: MAGMA MIXING AND DISEQUILIBRIUM

The abundance of broken crystals, the crystal-rich nature of the matrix glass, and the strong compositional heterogeneity in that glass are important clues for understanding the origin of the gray pumices of June 15. Pallister and others (1992) interpreted the gray pumices to represent a thermal boundary layer between normal white-pumice-type dacite and intruding basalt. Inclusions of the latter were found in mixed andesitic scoriae erupted on June 12, 1991, just prior to the climactic eruption. In their view, the gray pumices are reheated white-pumice-type dacites in which many crystals dissolved and drove the melt to a more mafic composition. This early model has been revised in Pallister and others (this volume); it now recognizes the importance of crystal fragmentation in the gray pumices in producing the shattered phenocrysts and crystal-rich matrix glass. They advocated fragmentation of the gray-pumice crystals upon ascent of the dacite in the conduit system during the early phase of the June 15 eruption, with fragmentation caused either by choked flow in the conduit or by violent vesiculation of the magma.

We envision an important role for magma mixing in the origin of the gray pumice, but we believe that the mixing was between the white-pumice dacite and a more melt-rich magma, at either higher temperature and (or) higher water content, with a dacitic melt composition close to that of analysis 7 (table 4). This melt-rich magma invaded and mixed with the dominant, white-pumice variety just prior to eruption. The lack of equilibration in the gray pumices, indicated most dramatically by the wide range of matrix glass compositions, suggests that this mixing event happened just hours or at most a few days before the cataclysmic eruption. It may ultimately be possible to correlate this proposed mixing event with characteristics of the preeruptive seismicity (Harlow and others, this volume; White and others, this volume). The minute quench crystals present in the gray-pumice matrix (fig. 1D) may reflect crystallization brought on by the intimate and possibly highly turbulent mixing of the two magmas. The mixing appears to have been sufficiently violent that most crystals in the gray pumices, except for the smallest quench crystals, were fragmented, even down to the scale of several micrometers (fig. 1D). The very similar compositions of Fe-Ti oxides (table 3) and other minerals between the white and gray pumices may indicate that most of the minerals in the latter grew in white-pumice-type dacite prior to the mixing event and that only the silica-poor matrix glass compositions from the gray pumices provide direct evidence for the nature of the invading magma. Failure of the Housh and Luhr (1991) plagioclase-melt model when applied to the white pumice from June 15 indicates that even the main phenocryst-rich dacite from Pinatubo was not in chemical equilibrium upon eruptive quenching. The anomalously calcic plagioclase rims may have grown under the influence of kinetic factors attendant on the initiation of supercooling during H2O degassing.

IDENTIFICATION OF ANCIENT HIGH-SULFUR ERUPTIONS

The 1991 dacite of Mount Pinatubo and the 1982 trachyandesite of El Chichón are two examples of anhydrite-bearing magmas that released large masses of sulfur gases to the Earth's atmosphere. Given the poor preservation potential of primary igneous anhydrite, it is important to find other parameters that can be used to identify similar eruptions in ancient deposits. The magmas of Mount Pinatubo and El Chichón differ in many other elemental and mineralogical aspects. The magma of Mount Pinatubo is a medium-K dacite that is also relatively poor in other incompatible trace elements (table 1). In sharp contrast, the magma of El Chichón is a high-K trachyandesite that is strongly enriched in other incompatible elements (Luhr and others, 1984). Among the minerals present in June 15 dacites of Mount Pinatubo, quartz, cummingtonite rims on hornblende, ilmenite, zircon, and olivine, chromite, and orthopyroxene xenocrysts are not found in the trachyandesite of El Chichón, and the latter has stable sphene and augite that are not present in the former. Those differences aside, the Pinatubo and El Chichón rocks show other important similarities. Both rock types are very crystal rich (58 wt% for El Chichón, 55 wt% for white pumices of Mount Pinatubo: table 2), have relatively low preeruption temperatures of ~800°C, are dominated by phenocrysts of complexly zoned plagioclase and hornblende, contain minor biotite, and have apatites that are unusually rich in sulfate: 0.34 wt% SO3 for apatites from the trachyandesite of El Chichón (Luhr and others, 1984) and up to 0.78 wt% SO3 for apatites from the dacite of Mount Pinatubo (Imai and others, this volume; Pallister and others, 1993). Hornblende is by far the most abundant mafic phenocryst in both cases. Both rocks also have high values of Fe3+/Fetotal: 0.44 for the white pumice of Mount Pinatubo (table 1) and 0.41 for the trachyandesite of El Chichón. The abundance of hornblende and the presence of biotite, along with the high crystal contents and low temperatures, demonstrate the water-rich nature of both magmas. The high Fe3+/Fetotal values are also consistent with the abundance of hydrous mafic minerals as argued by Carmichael (1967). Thus, the magmas of Mount Pinatubo and El Chichón were alike only in being strongly oxidized and very water- and sulfur-rich. Ancient analogs can be identified by their high crystal contents, high ratios of hornblende to anhydrous mafic minerals, low calculated preeruption temperatures, and sulfate-rich apatites. Their expected high Fe3+/Fetotal values will be quickly overprinted by posteruption oxidation. In pumices from both El Chichón (Luhr and others, 1984) and Pinatubo (Fournelle and others, this volume), anhydrite occurs as occasional inclusions within phenocrystic minerals. These inclusions will survive in the geologic record and provide convincing evidence for ancient high-sulfur eruptions.

ACKNOWLEDGMENTS

We thank Chris Newhall for providing specimens from the June 15 eruption. Gene Jarosewich and Joe Nelen performed the wet chemical and XRF analyses and provided essential assistance in electron microprobe analyses. John Pallister, Malcolm Rutherford, and Chris Newhall reviewed the manuscript and offered many useful comments and suggestions to help resolve earlier conflicts with other studies in this volume.

REFERENCES CITED

Andersen, D.J., Lindsley, D.H., and Davidson, P.M., 1993, QUILF: A PASCAL program to assess equilibria among the Fe-Mg-Mn-Ti oxides, pyroxenes, olivine, and quartz: Computers and Geosciences, v. 19, p. 1333-1350.

Arculus, R.J., Johnson, R.W., Chappell, B.W., McKee, C.O., and Sakai, H., 1983, Ophiolite-contaminated andesites, trachybasalts, and cognate inclusions of Mount Lamington, Papua New Guinea: Anhydrite-amphibole-bearing lavas and the 1951 cumulodome: Journal of Volcanology and Geothermal Research, v. 18, p. 215-247.

Bernard, A., Demaiffe, D., Mattielli, N., and Punongbayan, R.S., 1991, Anhydrite-bearing pumices from Mount Pinatubo: Further evidence for the existence of sulfur-rich silicic magmas: Nature, v. 354, p. 139-140.

Bernard, A., Knittel, U., Weber, B., Weis, D., Albrecht, A., Hattori, K., Klein, J., and Oles, D., this volume, Petrology and geochemistry of the 1991 eruption products of Mount Pinatubo.

Bluth, G.J.S., Doiron, S.D., Schnetzler, C.C., Krueger, A.J., and Walter, L.S., 1992, Global tracking of the SO2 clouds from the June, 1991 Mount Pinatubo eruptions: Geophysical Research Letters, v. 19, no. 2, p. 151-154.

Carmichael, I.S.E., 1967, The iron-titanium oxides of salic volcanic rocks and their associated ferromagnesian silicates: Contributions to Mineralogy and Petrology, v. 14, p. 36-64.

Carroll, M.R., and Rutherford, M.J., 1987, The stability of igneous anhydrite: Experimental results and implications for sulfur behavior in the 1982 El Chichón trachyandesite and other evolved magmas: Journal of Petrology, v. 28, p. 781-801.

Cawthorn, B.G., and Brown, P.A., 1976, A model for the formation and crystallization of corundum-normative calc-alkaline magmas through amphibole fractionation: Journal of Geology, v. 84, p. 467-476.

David, C.P.C., Dulce, R.G., Nolasco-Javier, D.D., Zamoras, L.R., Jumawan, F.T., and Newhall, C.G., this volume, Changing proportions of two pumice types from the June 15, 1991, eruption of Mount Pinatubo.

Fierstein, J., and Nathenson, M., 1992, Another look at the calculation of fallout tephra volumes: Bulletin of Volcanology, v. 54, p. 156-167.

Fonarev, V.I., and Korolkov, G.J., 1980, The assemblage orthopyroxene + quartz. The low-temperature stability limit: Contributions to Mineralogy and Petrology, v. 73, p. 413-420.

Fournelle, J, Carmody, R., and Daag, A.S., this volume, Anhydrite-bearing pumices from the June 15, 1991, eruption of Mount Pinatubo: Geochemistry, mineralogy, and petrology.

Gerlach, T.M., Westrich, H.R., and Symonds, R.B., this volume, Preeruption vapor in magma of the climactic Mount Pinatubo eruption: Source of the giant stratospheric sulfur dioxide cloud.

Geschwind, C.-H., and Rutherford, M.J., 1992, Cummingtonite and the evolution of the Mount St. Helens (Washington) magma system: An experimental study: Geology, v. 20, p. 1011-1014.

Ghiorso, M.S., and Sack, R.O., 1991, Fe-Ti oxide geothermometry: Thermodynamic formulation and estimation of intensive variables in silicic magmas: Contributions to Mineralogy and Petrology, v. 108, p. 485-510.

Gill, J.B., 1981, Orogenic andesites and plate tectonics: New York, Springer-Verlag, 391 p.

Hansen, J., Lacis, A., Lerner, J., Rind, D., Sato, M., Wilson, H., and McCormick, P., 1992a, Climate impact of Mount Pinatubo eruptionm [abs.]: Eos, Transactions, American Geophysical Union, v. 73, no. 43, p. 632.

Hansen, J., Lacis, A., Ruedy, R., and Sato, M., 1992b, Potential climate impact of Mount Pinatubo eruption: Geophysical Research Letters, v. 19, no. 2, p. 215-218.

Harlow, D.H., Power, J.A., Laguerta, E.P., Ambubuyog, G., White, R.A., and Hoblitt, R.P., this volume, Precursory seismicity and forecasting of the June 15, 1991, eruption of Mount Pinatubo.

Hattori, K., this volume, Occurrence and origin of sulfide and sulfate in the 1991 Mount Pinatubo eruption products.

Helgeson, H.C., Delany, J.M., Nesbitt, H.W., and Bird, D.K., 1978, Summary and critique of the thermodynamic properties of rock-forming minerals: American Journal of Science, v. 278A, p. 1-229.

Housh, T.B., and Luhr, J.F., 1991, Plagioclase-melt equilibria in hydrous systems: American Mineralogist, v. 76, p. 477-492.

Huebner, J.S., and Sato, M., 1970, The oxygen fugacity-temperature relationships of manganese oxide and nickel oxide buffers: American Mineralogist, v. 55, p. 934-952.

Imai, A., Listanco, E.L., and Fujii, T., this volume, Highly oxidized and sulfur-rich dacitic magma of Mount Pinatubo, Philippines: Implication for metallogenesis of porphyry copper mineralization in the western Luzon arc.

Katsui, Y., 1958, Groundmass anhydrite in the olivine basalt from the Rishiri volcano, Hokkaido: Journal of the Japanese Association of Mineralogy, Petrology, and Economic Geology, v. 42, p. 188-191.

Kerr, R.A., 1993, Pinatubo global cooling on target: Science, v. 259, p. 594.

Kilinc, A., Carmichael, I.S.E., Rivers, M.L., and Sack, R.O., 1983, The ferric-ferrous ratio of natural silicate liquids equilibrated in air: Contributions to Mineralogy and Petrology, v. 83, p. 136-140.

Kôzu, S., 1934, The great activity of Komagataké in 1929: Mineralogische und Petrographische Mitteilung, v. 45, p. 133-174.

Kwano, Y., 1948, A new occurrence of phenocrystic anhydrite crystals in the glassy rocks from Himeshima, Oita Prefecture, Japan: Geological Survey of Japan Report, v. 126, p. 1-7.

Lindstrom, D.J., and Korotev, R.L., 1982, TEABAGS: Computer programs for instrumental neutron activation analysis: Journal of Radioanalytical Chemistry, v. 70, p. 439-458.

Lofgren, G., 1980, Experimental studies on dynamic crystallization of silicate melts, in Hargraves, R.B., ed., Physics of magmatic processes: Princeton, N.J., Princeton University Press, p. 487-551.

Luhr, J.F., 1990, Experimental phase relations of water- and sulfur-saturated arc magmas and the 1982 eruptions of El Chichón Volcano: Journal of Petrology, v. 31, no. 5, p. 1071-1114.

Luhr, J.F., Carmichael, I.S.E., Varekamp, J.C., 1984, The 1982 eruptions of El Chichón Volcano, Chiapas, México: Mineralogy and petrology of the anhydrite-bearing pumices: Journal of Volcanology and Geothermal Research, v. 23, p. 69-108.

Matthews, S.J., Jones, A.P., and Bristow, C.S., 1992, A simple magma-mixing model for sulfur behaviour in calc-alkaline volcanic rocks: Mineralogical evidence from Mount Pinatubo 1991 eruption: Journal of the Geological Society of London, v. 149, p. 863-866.

McKeen, S.A., Liu, S.C., and Kiang, C.S., 1984, On the chemistry of stratospheric SO2 from volcanic eruptions: Journal of Geophysical Research, v. 89, no. D3, p. 4873-4881.

McKibben, M.A., Eldridge, C.S., and Reyes, A.G., this volume, Sulfur isotopic systematics of the June 1991 Mount Pinatubo eruptions: A SHRIMP ion microprobe study.

Melson, W.G., Allan, J.F., Jerez, D.R., Nelen, J., Calvache, M.L., Williams, S.N., Fournelle, J., and Perfit, M., 1990, Water contents, temperatures, and diversity of the magmas of the catastrophic eruption of Nevado del Ruiz, Colombia, November 13, 1985: Journal of Volcanology and Geothermal Research, v. 41, p. 97-126.

Merzbacher, C., and Eggler, D.H., 1984, A magmatic geohygrometer: Application to Mount St. Helens and other dacitic magmas: Geology, v. 12, p. 587-590.

Nicholls, I.A., 1971, Calcareous inclusions in lavas and agglomerates of Santorini volcano: Contributions to Mineralogy and Petrology, v. 30, p. 261-276.

Pallister, J.S., Hoblitt, R.P., and Reyes, A.G., 1992, A basalt trigger for the 1991 eruptions of Pinatubo volcano?: Nature, v. 356, p. 426-428.

Pallister, J.S., Hoblitt, R.P., Meeker, G.P., Knight, R.J., and Siems, D.F., this volume, Magma mixing at Mount Pinatubo: Petrographic and chemical evidence from the 1991 deposits.

Pallister, J.S., Meeker, G.P., Newhall, C.G., Hoblitt, R.P., and Martinez, M., 1993, 30,000 years of the "same old stuff" at Pinatubo [abs.]: Eos, Transactions, American Geophysical Union, v. 74, no. 43, p. 667-668.

Pearce, T.H., Russel, S.K., and Watson, I., 1987, Laser-interference and Normarski-interference imaging of zoning profiles in plagioclase phenocrysts from the May 18, 1980 eruption of Mount St. Helens, Washington: American Mineralogist, v. 72, p. 1131-1143.

Pier, J.G., Luhr, J.F., Podosek, F.A., and Aranda-Gómez, J.J., 1992, The La Breña-El Jagüey Maar Complex, Durango, Mexico: II. Petrology and geochemistry: Bulletin of Volcanology, v. 54, p. 405-428.

Rutherford, M.J., and Devine, J.D., this volume, Preeruption pressure-temperature conditions and volatiles in the 1991 dacitic magma of Mount Pinatubo.

Sack, R.O., Carmichael, I.S.E., Rivers, M., and Ghiorso, M.S., 1980, Ferric-ferrous equilibria in natural silicate liquids at 1 bar: Contributions to Mineralogy and Petrology, v. 75, p. 369-376.

Scott, W.E., Hoblitt, R.P., Daligdig, J.A., Besana, G., and Tubianosa, B.S., 1991, 15 June 1991 pyroclastic deposits at Mount Pinatubo, Philippines [abs.]: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 61-62.

Smith, J.V., and Brown, W.L., 1988, Feldspar minerals, vol. 1: New York, Springer-Verlag, p. 436-445.

Stormer, J.C., Jr., 1983, The effects of recalculation on estimates of temperature and oxygen fugacity from analyses of multicomponent iron-titanium oxides: American Mineralogist, v. 68, p. 586-594.

Taneda, S., 1949, On the anhydrite from Himeshima lava: Journal of the Japanese Association of Mineralogy, Petrology, and Economic Geology, v. 33, p. 69-73.

Taylor, G.A., 1958, The 1951 eruption of Mount Lamington, Papua: Bureau of Mineral Resources Geology and Geophysics Bulletin, v. 38, p. 1-117.

Varekamp, J.C., Luhr, J.F., and Prestegaard, K.L., 1984, The 1982 eruptions of El Chichón volcano (Chiapas, México): Character of the eruptions, ash-fall deposits, and gas phase: Journal of Volcanology and Geothermal Research, v. 23, p. 39-68.

Westrich, H.R., and Gerlach, T.M., 1992, Magmatic gas source for the stratospheric SO2 cloud from the June 15, 1991, eruption of Mount Pinatubo: Geology, v. 20, p. 867-870.

White, R.A., this volume, Precursory deep long-period earthquakes at Mount Pinatubo, Philippines: Spatio-temporal link to a basalt trigger.

Wilcox, R.E., 1954, Petrology of Parícutin volcano, Mexico: U.S. Geological Survey Bulletin 965-C, p. 281-349.

Yagi, K., Takeshita, H., and Oba, Y., 1972, Petrological study of the 1970 eruption of Akita-Komagatake volcano, Japan: Journal of the Faculty of Science, Hokkaido University, ser. IV, v. 15, p. 109-138.

Yoshiki, B., 1933, Activity of Akita-Komagatake volcano: Journal of the Japanese Association of Mineralogy, Petrology, and Economic Geology, v. 9, p. 153-160.

Zen, E-an, 1986, Aluminum enrichment in silicate melts by fractional crystallization: Some mineralogic and petrographic constraints: Journal of Petrology, v. 27, p. 1095-1117.

FIRE and MUD Contents

PHIVOLCS | University of Washington Press | U.S.Geological Survey

This page is <https://pubs.usgs.gov/pinatubo/luhr/>
Contact: Chris Newhall
Last updated 06.11.99