Journal of Petrology Advance Access originally published online on January 21, 2005
Journal of Petrology 2005 46(5):999-1012; doi:10.1093/petrology/egi008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Distinguishing Melting of Heterogeneous Mantle Sources from Crustal Contamination: Insights from Sr Isotopes at the Phenocryst Scale, Pisgah Crater, California
DEPARTMENT OF EARTH AND SPACE SCIENCES, UNIVERSITY OF CALIFORNIA LOS ANGELES, LOS ANGELES, CA 90095-1567, USA
RECEIVED APRIL 15, 2004; ACCEPTED DECEMBER 6, 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Compositionally heterogeneous basaltic centers from a variety of tectonic environments, including Pisgah Crater in the Mojave Desert region of California, exhibit secular changes in their chemistry that might be explained by the sequential melting of ultramafic to mafic mantle sources. We have analyzed phenocrysts from alkali basalts and hawaiites erupted at Pisgah Crater to investigate the effects of open-system modifications imposed on basaltic systems. We present 87Sr/86Sr data for individual phenocrysts of amphibole and clinopyroxene and the first published results of single olivine grains, in addition to plagioclase. Each mineral phase exhibits a range in Sr isotope composition that may only partially overlap the isotopic composition of the other mineral phases, suggesting an interplay between two magmatic end-members that continued up to the time of eruption. Limited 87Sr/86Sr variability in minerals from early and intermediate lavas indicates only moderate syn-crystallization open-system modification, whereas minerals in late-erupted lavas have much higher 87Sr/86Sr, consistent with extensive open-system modification. Rimward increases in 87Sr/86Sr of plagioclase confirm that these changes occurred within the stability field of plagioclase and, therefore, at crustal or near-crustal depths. The major element compositions of olivine-hosted melt inclusions indicate that an Al-rich component of andesitic composition (87Sr/86Sr
0·7056), possibly derived from plagioclase-rich cumulates or pelites, was assimilated by magma generated from asthenosphere or young lithosphere with 87Sr/86Sr 
0·7038. The results clearly demonstrate the utility of measuring the 87Sr/86Sr of individual minerals and indicate that Pisgah Crater basalts probably acquired isotopically enriched geochemical signatures from crustal contamination, rather than from mixing of heterogeneous mantle melts. KEY WORDS: assimilation; basalts; melt inclusions; minerals; Sr isotopes
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Basaltic magmatism primarily results from the melting of ultramafic peridotitic rocks residing in the upper mantle. However, there has been increasing recognition that melts of pyroxenite or eclogite may make significant volumetric contributions to the source of potassic basalts (Foley, 1992
, and references within) or subordinate contributions, as in the case of mid-ocean ridge basalt (MORB) (Hirschmann & Stolper, 1996). Reiners (2002) compiled evidence from a variety of tectonic settings, including Pisgah Crater in the Mojave Desert of California, for secular geochemical changes during mafic eruptions that could result from the progressive, deep-level mixing of melts generated from garnet peridotite and garnet pyroxenite mantle sources. At Pisgah, early erupted basalts have lower 87Sr/86Sr, higher 143Nd/144Nd and higher incompatible element abundances whereas late-erupted basalts have higher 87Sr/86Sr, lower 143Nd/144Nd and lower incompatible element abundances (Glazner & Farmer, 1992). Contamination by typical intermediate or silicic crust cannot account for these variations and, thus, they have been attributed to mixing between a depleted upper-mantle magma and a relatively more evolved melt generated from a mafic source, possibly either a mid-crustal gabbro or mantle websterite in pervasively veined lithosphere (Glazner et al., 1991). Reiners (2002) suggested that Pisgah Crater geochemical variations result from the progressive, deep-level mixing of melts generated from separate and geochemically distinct garnet peridotite and garnet pyroxenite mantle sources without the influence of crustal contamination.
To more fully examine the processes and scales involved in the geochemical and isotopic variations of Pisgah Crater basalts, we present Sr isotope results from individual minerals, melt inclusions, groundmass, and whole-rocks representing the eruptive history of Pisgah Crater. Mineral-scale chemical and isotopic analyses provide insights into the progressive evolution of the magmas, and the timing of processes involved as a result of the inherent ability of minerals to record changes in the composition of the magma from which they grow. These changes may be exhibited as chemical or isotopic variability in the minerals themselves or in the melt inclusions they contain. Chemical heterogeneity in melt inclusions has been utilized to identify a diverse range of primitive magma compositions involved in the formation of MORB and ocean-island basalt (OIB) (Sobolev & Shimizu, 1993; Saal et al., 1998; Sobolev et al., 1999; Kamenetsky et al., 2002), and mineral-scale Sr isotope heterogeneity has been utilized to identify and constrain open-system processes such as recharge in intermediate composition rocks (Davidson et al., 1998, 2001) and crustal contamination in basaltic (Wolff & Ramos, 2002; Tollstrup, 2003) and rhyolitic rocks (Wolff et al., 1999; Wolff & Ramos, 2003). Such mineral-scale Sr isotope analyses have, however, generally focused on minerals characterized by high Sr contents, such as plagioclase. In this study, we show how results from more technically demanding analyses of mafic minerals including clinopyroxene, amphibole and melt inclusion-bearing olivine, in addition to plagioclase, provide detailed insights into the effects of open-system modification in basalts and suggest that crustal contamination, rather than the mixing of disparate melts of differing mantle sources, better accounts for secular chemical and isotopic changes in the Pisgah Crater lavas.
|
BACKGROUND |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Pisgah Crater is a Pleistocene cinder cone surrounded by a series of morphologically young flows located
50 km east of Barstow, California (Fig. 1). The cone and flows are Quaternary in age (Luedke & Smith, 1987) and have been inferred to be as young as Holocene (Chesterman, 1982). Wise (1969) separated the Pisgah Crater flows into three units based on flow morphologies. All three units contain olivine and plagioclase with progressively increasing phenocryst sizes from unit 1 (1 mm) to unit 3 (5 mm). Titanaugite is also present, except in the unit 3 flows, which instead contain rare amphibole. Mineral crystallization is postulated to have initiated at 1225–1150°C, with olivine (Fo87–83) followed by plagioclase (An69) at between 1150 and 1100°C, and titanaugite below 1100°C based on crystallization experiments using a unit 1 lava (Ussler & Glazner, 1989). Lava compositions range from early alkali basalt (unit 1) to late hawaiite (unit 3). Less mafic rocks have higher CaO, Al2O3, and SiO2 contents but lower TiO2, Na2O, K2O, and P2O5 contents (e.g. Fig. 2). Incompatible trace element concentrations such as Rb, Sr, Ba, La, and Zr also decrease with increasing differentiation. Whole-rock 87Sr/86Sr and 143Nd/144Nd vary from more depleted values in the more mafic lavas (0·7036 and 0·51294, respectively) to more enriched values (0·7050 and 0·51275) in the more differentiated lavas. 206Pb/204Pb and 208Pb/204Pb decrease slightly in the eruption sequence, whereas 207Pb/204Pb is generally uniform. The oxygen isotope compositions of all whole-rocks fall within the range (
18O = 6·1–6·7
) observed in the earliest erupted unit 1 lavas (Glazner et al., 1991).
|
|
, unit 1 flows; gray diamonds, unit 2 flows; 
, unit 3 flows. K2O represents typical decreasing concentrations of incompatible elements in successive Pisgah Crater eruptions compared with an index of differentiation (MgO). The linear trend in 1/Sr–87Sr/86Sr is interpreted to result from simple mixing between two end-members. Arrow (FC) represents a general fractionation trend in these diagrams. |
SAMPLE PREPARATION AND ANALYSIS |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Minerals, groundmass, and whole-rocks from one of each of the three Pisgah Crater flow units were analyzed for Sr isotopes. Rock samples comprising
0·5 m2 of the upper surface (6–15 cm thick) of the flows were collected, crushed, and sieved. After cleaning in 1% HNO3, olivine, plagioclase, titanaugite, and amphibole grains (0·25–0·45 mm) were hand picked to minimize adhering glass and/or groundmass. Individual grains or pairs of grains were weighed and, where applicable, spiked for Sr. A single plagioclase crystal (unit 3) was also sampled by microdrilling [four sites from core to rim; following the procedures of Davidson et al. (1998)]. Samples of groundmass were also obtained by microdrilling fine-grained portions of thin-section billets. In addition, rock chips were inspected and used to make whole-rock powders. Splits of these powders were then leached in cold 6N HCl for 10 min to assess the isotopic composition of readily leachable components.
Minerals, groundmass, and whole-rock samples were fully dissolved over at least a 5 day period in Teflon digestion vessels using high-purity HF, HNO3, and HCl. Sr was separated using low-blank, microsample chromatography (Wolff et al., 1999). Sr isotope ratios were measured on a VG Sector 54 mass spectrometer using five Faraday collectors in dynamic mode at UCLA. Sr elemental abundances were determined for approximately half of the samples of each mineral phase, with the remainder analyzed for isotope composition only. In this way, we ensured that spike-contributed 87Sr and 86Sr did not inadvertently bias spike-corrected 87Sr/86Sr ratios obtained for minerals. 87Sr and 86Sr contributions to the minerals and groundmass from the spike were, in any case, minor as a result of the high purity of the 84Sr spike used and our optimization of the spike-to-sample ratio in the analyzed mixtures. In addition, the effect of the Sr blank was stripped from the 87Sr/86Sr results using the average of measured total process blank 87Sr/86Sr ratios and Sr concentrations (Tables 1 and 2). For Sr-rich minerals such as plagioclase and clinopyroxene, blank effects were minor (0·00002), but for Sr-poor minerals such as olivine the effects were significant. For olivines containing 1 ng Sr, 87Sr/86Sr corrected results were lower by 0·0002; however, for the majority of olivines with >2 ng Sr, 87Sr/86Sr ratios decreased by <0·0001. Such variations are small compared with the overall range of 87Sr/86Sr signatures displayed by the data suite.
|
|
In addition to Sr isotopes, major element compositions of selected plagioclase and olivine, and of spinel and melt inclusions within olivine, from unit 3 lavas were determined using a Cameca electron microprobe at Washington State University (Tables 3 and 5). To allow for the possibility of post-entrapment modification of melt inclusions resulting from olivine crystallization, model melt inclusion compositions were also calculated by adding between 0 and 10% of equilibrium olivine using the calculator of Danyushevsky et al. (2000)
with olivine incrementally added until a KD [(glass Fe/Mg)/(olivine Fe/Mg)] of 0·3 was obtained (Roeder & Emslie, 1970) with (Fe2+/Fe3+) calculated as described in the appendix of Danyushevsky et al. (2000). Table 4 displays recalculated melt inclusion compositions and the amount of olivine added to individual melt inclusions.
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
87Sr/86Sr of minerals and groundmass
Minerals and groundmass from the three Pisgah Crater flow units have variable Sr concentrations and 87Sr/86Sr disequilibrium within and between flow units (Tables 1 and 2; Fig. 3). As expected from partitioning relationships, Sr concentrations generally decrease in the sequence plagioclase > groundmass > cpx/amph > olivine. Minerals exhibit minor compositional changes over the eruptive history such as decreasing anorthite contents in plagioclase (from An72–65 in unit 1 to An67–63 in unit 3 flows) and decreasing magnesium contents in olivine (from Fo85–73 in unit 1 to Fo81–75 in unit 3 flows).
|
The Sr isotope compositions of leached whole-rocks (Table 1) confirm previous evidence (Glazner et al., 1991
) for increasing 87Sr/86Sr in the Pisgah Crater eruption sequence but are more limited in their range (from 0·70401 to 0·70472 for unit 1 to unit 3, respectively). Differences between leached and unleached 87Sr/86Sr ratios in whole-rock powders are apparent and a leachate from sample P4 (unit 3) has 87Sr/86Sr of 0·70835, although significant decreases in Sr concentrations from leaching are not observed.
Not surprisingly, the Sr isotope compositions of minerals and groundmass from the three Pisgah Crater flow units increase sympathetically with increasing whole-rock Sr isotopic compositions from unit 1 to unit 3. Within flows, individual mineral phases and groundmass are characterized by distinctly different ranges in Sr concentrations and 87Sr/86Sr (Tables 1 and 2; Fig. 3). 87Sr/86Sr signatures of specific mineral phases and groundmass in unit 1 and 2 flows commonly overlap. The majority of plagioclase and clinopyroxenes, as well as a few olivines, in units 1 and 2 have similar 87Sr/86Sr to that of unit 1 whole-rocks (0·7040; Figs 3 and 5), whereas groundmass 87Sr/86Sr is generally higher than that of the majority of minerals. The 87Sr/86Sr signatures of minerals and groundmass in unit 3 flows, however, are much higher and vary extensively, with only limited overlap between minerals and groundmass (Fig. 3). The 87Sr/86Sr within individual mineral phases and groundmass isolated from single flow units is variable and characterized by, in order of increasing ranges of 87Sr/86Sr, plagioclase < cpx/amph < groundmass < olivine. In general, 87Sr/86Sr signatures in plagioclase and clinopyroxene range to the lowest values in any individual flow (e.g. <0·7038 in unit 1 flows), whereas olivine exhibits the highest 87Sr/86Sr (e.g. >0·7056 in unit 3 flows). Overall, the majority of mineral phases from a given flow are not in isotopic equilibrium with each other or with the groundmass.
The microdrilled plagioclase crystal from unit 3 is characterized by 87Sr/86Sr variations from core to rim (0·7043 to 0·7046), demonstrating that 87Sr/86Sr increased during mineral growth and that isotopic variations occur on an intracrystal as well as intercrystal scale. In addition, Sr concentrations decrease from 860 ppm in the core to 720 ppm near the rim.
Besides reflecting the broadest range and highest 87Sr/86Sr, olivine grains have Sr concentrations (5–9 ppm) that greatly exceed estimates (0·05 ppm) predicted from applying experimentally determined partition coefficients (Beattie, 1994) to Pisgah Crater groundmass Sr abundances. Such elevated concentrations could be related to inclusions of melt in the olivine.
Mineral and melt inclusion compositions
Mineral and melt inclusion compositions (unit 3) are variable and magnesium contents of olivine (Fo81–75) are similar to those in crystallization experiments conducted by Ussler & Glazner (1989; Fo87–83) on a unit 1 lava. Major element compositions of olivine, and of their spinel and melt inclusions, are shown in Tables 3–5; the olivine grains were obtained from the same separate used for unit 3 olivine Sr analyses. The melt inclusions are fine grained and were not experimentally rehomogenized (Nielson et al., 1998). Rather, average compositions were obtained using multiple electron microprobe spots from across the exposed area of individual inclusions. No daughter crystals of minerals significantly greater in size than those of the groundmass are apparent within or near the margins of melt inclusions, suggesting that post-entrapment crystallization of the melt inclusion was limited. In addition, no major element variations are detected in the host olivine approaching melt inclusion walls. Melt inclusions are found in olivine with relatively uniform Fo contents both within individual and between olivine crystals (Fig. 4a) and have recalculated compositions (see procedure in ‘Sample preparation and analysis’) ranging from 47 to 57% SiO2 and from 1 to 5·5% MgO. Such MgO contents approach those of the host basalt (5·9%) and extend to significantly lower abundances. Al2O3 contents are generally elevated (15–25%), with inclusions with the highest SiO2 also having high Al2O3 contents. In addition, K2O and Na2O contents vary from 0·7 to 2·7% (Fig. 4b) and from 3·5 to 6·3%, respectively, with Na2O/K2O ratios ranging from 2 to 9 in the melt inclusion suite (Table 4).
|
Plagioclase from early erupted unit 1 flows contains 72–65% anorthite, whereas that in late-erupted unit 3 flows has generally lower anorthite contents of 67–63%. These compositions are consistent with more extensive differentiation in late-erupted Pisgah Crater basalts and are similar to plagioclase compositions predicted by the Ussler & Glazner (1989)
experiments.
Geothermometry
Olivine host and spinel inclusion pairs from unit 3 flows reflect equilibration temperatures (Ghiorso, 1991) similar to those predicted by Ussler & Glazner (1989). Nine calculated temperatures cluster within a 30°C range from 1060 to 1090°C; two others are lower (1040°C). Olivine–spinel pairs from the core and rim portions from a single olivine yield higher temperatures than the other grains (1145°C and 1119°C) and the olivine contains a melt inclusion with intermediate SiO2 (55%) and high alkali contents (e.g. 2·7% K2O). In addition, olivines characterized by temperatures between 1040°C and 1090°C have melt inclusions with variable SiO2 and K2O (Tables 4 and 5; Fig. 4c).
|
EFFECTS OF OPEN-SYSTEM BEHAVIOR ON ISOTOPE HETEROGENEITY |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Detailed analyses of individual rock components are extremely valuable in identifying and isolating the effects of magmatic processes (e.g. Simonetti et al., 1996
; Davidson et al., 1998, 2001; Wolff et al., 1999; Wolff & Ramos, 2003). This is in contrast to the view whole-rocks provide, wherein the effects of all magmatic processes involved in the genesis of those rocks are summed. Sr isotope heterogeneity in Pisgah Crater minerals, melt inclusions, and groundmass clearly demonstrates that open-system processes have occurred. In addition, melt inclusion compositions suggest the presence of magmatic components with diverse chemistries. Major element compositions of olivine–spinel pairs offer additional temperature constraints that, when coupled with diffusion considerations, help constrain crystal residence times. In this section, we discuss intra- and intergrain isotopic and chemical heterogeneities in more detail, and in the following section we use these results to evaluate two competing hypotheses (Glazner et al., 1991; Reiners, 2002) that might account for the geochemical variations exhibited in Pisgah Crater basalts.
Intra- and intergrain isotopic heterogeneity
Linearly correlated whole-rock 87Sr/86Sr and inverse Sr abundances in the Pisgah Crater lavas are suggestive of simple two-component mixing (Fig. 2; Glazner et al., 1991) although, as discussed below, this correlation is not observed at the scale of microdrilled groundmass samples (Fig. 5). In general, 87Sr/86Sr varies within each type of mineral, between coexisting minerals, between minerals and host groundmass, and within groundmass (Fig. 3). The similarity between the lowest 87Sr/86Sr of the earliest erupted unit 1 minerals and groundmass suggests crystallization of labradorite, titanaugite, and olivine from a relatively uncontaminated magma with 87Sr/86Sr of 0·7037–0·7040. In addition, 87Sr/86Sr ranges that partially overlap in all three mineral suites are consistent with nucleation and growth of the majority of these different minerals prior to extensive open-system modification. Homogeneous 87Sr/86Sr in labradorite and only minor 87Sr/86Sr heterogeneity in titanaugite probably attest to the incipient effects of open-system modification in unit 1 lavas. More extensive modification of the groundmass results in variable 87Sr/86Sr ratios that are higher than the entire range of unit 1 minerals except for a few olivines that have even higher 87Sr/86Sr.
|
Sr isotope modification that largely postdates the bulk of crystallization may also explain the broad range of 87Sr/86Sr exhibited in unit 1 olivine. Sr is not readily incorporated into olivine. As a result, Sr contents and 87Sr/86Sr of olivine almost certainly reflect Sr in melt included within it. If locally extensive 87Sr/86Sr heterogeneity existed within the magma, as suggested by the groundmass, olivine could potentially trap melt inclusions with highly variable 87Sr/86Sr. Once isolated, Sr in these melt inclusions could not mix or re-equilibrate with Sr in the surrounding magma. As such, olivine-hosted melt inclusions, possibly formed along wallrock–magma interfaces during the initial stages of mixing when undercooling and rapid crystallization could occur (Danyushevsky & Leslie, 2002
), might be expected to retain some of the isotopic heterogeneity that existed prior to subsequent magma evolution or continued mixing. In comparison with olivine, groundmass exhibits less Sr isotope variability but a similar mean 87Sr/86Sr. High 87Sr/86Sr in olivine suggests that if plagioclase and clinopyroxene co-precipitated with olivine, the high rimward 87Sr/86Sr in plagioclase and clinopyroxene was buffered by the lower 87Sr/86Sr of their uncontaminated crystal cores, and/or olivine crystallized in magma not sampled by either plagioclase or clinopyroxene. Because Sr is not readily incorporated into olivine, buffering towards pre-contaminated signatures may be minimal, allowing olivine to record a wider spectrum of 87Sr/86Sr than the groundmass. In addition, olivine, plagioclase, clinopyroxene, and amphibole are not thought to be xenocrystic because distinct populations of minerals with different major element compositions are not seen. In contrast, Fo contents of olivine and An contents of labradorite in both early and later flows are relatively uniform and apparent DSr values (labradorite 2; titanaugite 0·11; calculated using mineral and groundmass Sr concentrations) are similar to those of other basaltic systems (Rollinson, 1993), suggesting that all minerals originated in basalts similar to the current hosts. 87Sr/86Sr signatures in flow unit 2 labradorite and titanaugite are also variable and generally higher than those of the same phases in unit 1, with olivine again characterized by the most variable 87Sr/86Sr (Fig. 3). The variable and elevated groundmass 87Sr/86Sr compared with labradorite and titanaugite is consistent with 87Sr/86Sr modification of the magma after the onset of crystallization of these minerals.
Compared with those of earlier flows, unit 3 minerals and groundmass have consistently higher 87Sr/86Sr than the same phases in units 1 or 2. Labradorite 87Sr/86Sr is generally uniform and Sr contents are lower than in labradorite in flow units 1 and 2. Rare amphibole has 87Sr/86Sr and Sr contents intermediate between those of labradorite and olivine, and its presence suggests increasing H2O contents in late-erupted lavas. Groundmass 87Sr/86Sr is variable and higher than for all minerals except olivine, and olivine retains more variable 87Sr/86Sr than groundmass. These variations are consistent with further Sr isotope modification of the magma after the bulk of crystallization of labradorite and amphibole, as neither of these phases retains 87Sr/86Sr as high as that seen in either the groundmass or olivine.
In contrast to minerals, groundmass and leached rock powders produce results that differ from those for the whole-rocks. Microdrilled groundmass samples do not lie along the trend defined by Pisgah Crater whole-rocks (87Sr/86Sr vs 1/Sr; Fig. 5). Such decoupling could reflect the likelihood that variable mixtures of crystals and melt are sampled at the scale of microdrilling, resulting in the potential sampling of more than two possible mixing end-members. In addition, leached whole-rock powders have lower 87Sr/86Sr ratios (Tables 1 and 2) than unleached powders and identical Sr concentrations (within error), with a single leachate sample characterized by 87Sr/86Sr substantially higher (0·7083) than values for all the Pisgah Crater basalts. Only minor differences in concentrations between leached and unleached powders, however, suggest little leachable Sr and post-eruption alteration, with differences in 87Sr/86Sr probably resulting from 87Sr/86Sr heterogeneity in whole-rock powders. The absence of post-eruption alteration is supported by the youthful (<30 ka) and pristine nature of these rocks, and the observation that mineral signatures, which are inherently more resistant to post-eruption alteration, define even broader 87Sr/86Sr variations than both groundmass and whole-rocks.
Effects of diffusional relaxation and post-entrapment crystallization on crystals and melt inclusions
The presence of 87Sr/86Sr disequilibrium in Pisgah Crater minerals creates an isotopic gradient and an opportunity for diffusion to modify 87Sr/86Sr. Diffusion would act to minimize Sr isotope contrasts between minerals and groundmass. Therefore, variations could have been even greater than currently observed if diffusion was significant. The effects of diffusion would be to progressively decrease the range of 87Sr/86Sr in minerals in the sequence olivine (Morioka, 1983), labradorite (Giletti & Casserly, 1994), and titanaugite (Sneeringer et al., 1984), and to drive the resulting 87Sr/86Sr toward groundmass values. For melt inclusions in olivine, the low diffusion rate associated with Sr moving from the inclusion into the olivine host is expected to further decrease Sr diffusion between the melt inclusion and the outside melt. Olivine, however, is characterized by the broadest range of 87Sr/86Sr followed by titanaugite and then labradorite, which is inconsistent with diffusion control. Thus, mineral signatures indicate that diffusion plays an only minor role in generating the 87Sr/86Sr signature of Pisgah Crater minerals.
Based on the temperatures calculated from olivine–spinel geothermometry in the unit 3 flow (1040–1090°C, Table 5), a simple diffusion model can be used to evaluate the maximum residence time for labradorite in a unit 3 flow (Fig. 6). This model assumes (1) an initial step-function distribution of 87Sr/86Sr between the mixed or otherwise contaminated melt and the edge of the crystal, and (2) that the edge of the crystal was initially characterized by the same 87Sr/86Sr as the core (0·70430) of the crystal prior to diffusion. We use a simple diffusion model (Crank, 1975) to determine the 87Sr/86Sr at 0·5 mm distance within the labradorite crystal. Diffusional exchange occurring over 950 years could have generated the measured rim (R4; 0·70454) labradorite 87Sr/86Sr signature (Fig. 6) from diffusion of Sr from the melt. This offers a maximum limit of the residence age of the plagioclase in the unit 3 flow, as the plagioclase could have grown from a melt of intermediate Sr isotope composition.
|
Just as with Sr isotopes, major elements of olivine-hosted melt inclusions in unit 3 flows record a wide range of compositions (Table 3). These inclusions have MgO contents that range from
0·5 to 3% as well as variable Na2O/K2O ratios. Mg has been shown to readily diffuse through olivine at magmatic temperatures allowing re-equilibration of melt inclusions with the host magma (Gaetani & Watson, 2002), with such re-equilibration expected to increase MgO in melt inclusions in this case. Thus, if partial re-equilibration occurred, melt inclusions would have been even more evolved with respect to MgO at the time of entrapment than currently observed (Fig. 4b). In contrast, post-entrapment decreases in MgO contents in melt inclusions would occur if olivine crystallized from within the inclusion itself, possibly along the walls of the inclusion. No such olivine has been detected along the inclusion walls and no Fo gradients are apparent in the enclosing olivine when electron microprobe analyses approach the inclusion–olivine interface. Thus, re-equilibration within the olivine may have occurred, but post-entrapment olivine crystallization is not readily evident. None the less, we have recalculated melt inclusion compositions by incrementally adding olivine back into the melt inclusions.
Recalculated melt inclusion compositions (Table 4) differ from measured compositions (Table 3) in having generally higher MgO (up to 5·5%) and FeO contents, whereas SiO2, Al2O3, Na2O, and K2O contents generally decrease (typically by 1%). In contrast, Na2O/K2O ratios would not be readily modified by post-entrapment crystallization or diffusion. Thus, wide variations in Na2O/K2O confirm the presence of a diverse suite of melts, irrespective of post-entrapment olivine crystallization. A notable trait of these melt inclusions is their high Al2O3 contents (15–25%), which are relatively uncommon in basalts. Melt inclusions in plagioclase from basalts in Scotland (Dempster et al., 1999) have similarly high Al2O3 contents (up to 30%) and have been inferred to be associated with melting of pelites present in the magma as xenoliths. Alternatively, such high-Al melt inclusions (up to 22%) have been seen in rare mantle xenoliths and been attributed to the presence of small-volume metasomatic mantle melts involved in ocean-island volcanism (Schiano & Clocchiatti, 1994). Na2O/K2O ratios, however, seldom exceed four in these xenoliths. For comparison, the majority of Hawaiian picrites (Norman et al., 2002) have Al2O3 contents 16%. In magmas similar to those of Pisgah Crater (7–10% MgO), melt inclusions from Kerguelen (Borisova et al., 2002) have Al2O3 contents of 12–15% and Na2O/K2O of 1–2. Magmas associated with these inclusions are thought to originate from plume and lithospheric mantle sources and do not attain Al2O3 contents near those of melt inclusions at Pisgah Crater.
To review, Sr isotopes in Pisgah Crater minerals and groundmass indicate little modification of early erupted units but extensive late-stage modification. This is manifest in elevated, highly variable 87Sr/86Sr in olivine and groundmass occurring in all flow units. The absence of a linear correlation in 87Sr/86Sr and 1/Sr in groundmass and minerals may result from multiple mixing components, locally dependent 87Sr/86Sr heterogeneity in the magma conduit or chamber, or because minerals reflect variable time-dependent or spatially dependent 87Sr/86Sr and Sr concentrations. 87Sr/86Sr variations are not consistent with those expected for diffusion, and Sr diffusion modeling suggests plagioclase residence times of 1 kyr, implying that open-system modification occurred rapidly. At Pisgah Crater, more extensive open-system behavior is associated with more radiogenic Sr signatures, more evolved magma and mineral compositions, and lower abundances of incompatible elements.
|
ORIGIN OF ISOTOPE AND COMPOSITIONAL HETEROGENEITY IN PISGAH CRATER LAVAS |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Two hypotheses have been presented to account for chemical and isotope variations of Pisgah Crater lavas. The first attributes them to mixing of basaltic melts generated from chemically and isotopically distinct upwelling mantle garnet peridotite and garnet pyroxenite. The second suggests that these same chemical and isotopic traits result from mantle melts assimilating mafic crustal components. We evaluate these competing hypotheses below.
Mixing of peridotitic and pyroxenitic melts
Reiners (2002) documented the worldwide existence of basaltic eruption sequences that have trace element and isotopic characteristics similar to that observed at Pisgah Crater, and presented magma mixing models that appeal to the melting of heterogeneous upwelling mantle sources composed of garnet peridotite and garnet pyroxenite for their origin. These differing mantle components can be mixed and melted, or melts from these components can be independently generated and mixed prior to eruption. In an attempt to account for linearly correlated trace element variations (e.g. La vs Nd, Nb vs Zr) in these basalt sequences, Reiners concluded that mixing of melts generated from these two mantle components occurs at mantle depths. Such a process could be responsible for many of the geochemical and isotopic variations observed at Pisgah Crater, as long as the proposed garnet peridotite and garnet pyroxenite sources have different isotopic compositions and melt to differing degrees (i.e. pyroxenite >10%; peridotite <2%). Thus, the peridotite-derived magma is more mafic, has higher incompatible element abundances, and is characterized by unradiogenic isotope signatures. In contrast, the pyroxenite-derived magma is more evolved, has lower incompatible element abundances, and has radiogenic isotope signatures. Mixing of such magmas would account for decreasing incompatible element abundances, variable isotopic compositions, and the apparent absence of extensive major element variations in the basalts that compose the Pisgah Crater eruption sequence.
Mineral-scale results from this investigation are generally consistent with this mixing model but require major caveats. For Pisgah Crater basalts, melting and mixing of peridotitic and pyroxenitic melts are consistent with model curves with final depths of melting between 51 and 93 km [mixing lines P–E, P–D and P–C in fig. 7 of Reiners (2002)]. These mixing lines indicate 2% melting of the peridotite and 20–40% melting of the pyroxenite. The initial ‘relatively’ unmixed peridotite melt would, according to Pisgah results, have 87Sr/86Sr 0·7037–0·7039 (unit 1). Thorium excess in a unit 1 basalt (25%; Ramos, 2000) is consistent with its origin as a small-degree partial melt of garnet peridotite. Evolved phenocryst (e.g. olivine Fo83) and whole-rock compositions (e.g. MgO 8%), however, suggest that this magma underwent some differentiation prior to eruption (i.e. unit 1 magmas are not primary mantle melts). Overlapping 87Sr/86Sr in olivine, clinopyroxene, and labradorite in unit 1 lavas suggests that crystallization occurred in this peridotite-derived magma at crustal or near-crustal pressures when plagioclase became a stable liquidus phase, implying pressures that are significantly lower than those associated with the proposed depths of melting (51 km). 87Sr/86Sr contrasts between minerals and groundmass clearly require a second component introduced into the magmatic system after the bulk of crystallization had occurred. If the second component is a pyroxenite-derived melt, it mixed with the peridotite-derived melt just prior to eruption, as labradorite and clinopyroxene do not record the shift to higher 87Sr/86Sr seen in the groundmass. Thus, the addition of the second magma seems to act as a catalyst for eruption and limited mixing, rather than ensuring thorough mixing as proposed by Reiners (2002). Even whole-rocks that lie further along the linear trace element arrays, and that presumably reflect more extensive mixing, are characterized by appreciable 87Sr/86Sr heterogeneity within both minerals and groundmass.
Consistently higher 87Sr/86Sr in the groundmass than plagioclase in all Pisgah Crater flow units and internal variation within plagioclase (P4 microdrilled sample) suggest that mixing occurred at crustal or near-crustal pressures. Alternatively, mixing of peridotite- and pyroxenite-derived magmas could have involved an additional third component, although linear trace element mixing arrays would not be expected as there would then be variable three-component mixing. Non-linear trends in microdrilled groundmass samples (Fig. 5), however, do not rule out this possibility. Clearly, extensive 87Sr/86Sr heterogeneity between groundmass and minerals, and within the groundmass itself, records mixing that might not be expected for magmas that are generated and mixed in the mantle.
Recalculated major element compositions of olivine-hosted melt inclusions from unit 3 flows suggest that the more radiogenic mixing component, a presumed 20–40% partial melt of pyroxenite, is andesitic in composition with SiO2 57%, MgO <2·0%, and high Na2O/K2O. Thus, both melt inclusions and evolved mineral compositions indicate that this pyroxenite-derived magma was differentiated. In addition, the three melt inclusions with the highest K2O contents are found in olivines that bracket the span of compositions in unit 3 lavas, requiring a large range of olivine compositions for the incompatible element enriched peridotite-derived magma (Fig. 4a). Such high K2O contents in melt inclusions also exceed those of presumably unmixed, early erupted magmas.
The combined presence of U enrichment (40%; Ramos, 2000) and amphibole in unit 3 lavas suggests that the melting processes generating the pyroxenite-derived magma involved H2O. U enrichment in basalts is typically seen in arc settings where U-rich fluids, originating from the subducted slab, flux the overlying mantle wedge and initiate melting (Turner et al., 2003). In contrast, U enrichment is seldom seen in basalts generated in areas of presumed mantle upwelling such as at ocean islands and mid-ocean ridges (Bourdon & Sims, 2003; Lundstrom, 2003). As a result, U enrichment in unit 3 lavas is somewhat surprising given that Pisgah Crater does not overlie an active subduction zone. It also requires U-rich fluids to have been introduced into the pyroxenitic mantle source during or immediately prior to melting.
Overall, the Reiners (2002) peridotite and pyroxenite magma mixing models could describe the processes generating the chemical and isotopic variations of Pisgah Crater whole-rocks and minerals if (1) pyroxenite-derived magma addition and incomplete mixing occurred shortly before the time of eruption, (2) magmas mixed and crystallized at crustal or near-crustal depths, and (3) pyroxenite sources were fluxed by U-rich fluids during or immediately prior to mantle melting.
Gabbroic crustal contamination
In contrast to mixing of mantle-derived magmas, Glazner et al. (1991) suggested that assimilation of Proterozoic mafic crust may be responsible for the chemical and isotopic variations of Pisgah Crater basalts. Those workers suggested that uncontaminated Pisgah Crater magmas may have stalled in mid- to lower-crustal magma chambers where they experienced variable degrees of differentiation and crustal assimilation prior to eruption. Glazner et al. (1991) also allowed for the possibility that mantle-derived melts could have interacted with websteritic or pyroxenitic mantle to generate the chemical and isotopic variations of Pisgah Crater basalts. The absence of minerals and magmas with primitive compositions at Pisgah Crater shows that the magmas are not undifferentiated melts directly derived from ultramafic mantle sources. 87Sr/86Sr variations in minerals and groundmass are consistent with the interaction of mantle-derived magma with crustal materials at near-crustal or crustal depths. The contaminating agent must have lower incompatible element abundances than the uncontaminated magma to explain decreasing abundances of these elements with increasing 87Sr/86Sr. Melt from old, gabbroic cumulates or granulitic pelites could be relatively depleted in incompatible elements such as Sr yet still exert significant leverage on 87Sr/86Sr, especially with contemporaneous fractionation of plagioclase. Similar processes have been proposed for the generation of geochemical and isotopic signatures in Si-rich arc rocks (e.g. Costa & Singer, 2002; Dungan & Davidson, 2004). However, in contrast to Pisgah Crater, incompatible element concentrations in most of these rocks are constant or increase as a result of assimilation.
In early erupted Pisgah Crater lavas, minor crustal addition could generate the limited 87Sr/86Sr heterogeneity observed in the groundmass, although such contamination would have had to occur immediately prior to eruption, as plagioclase and clinopyroxene phenocrysts are not affected. In later eruptions, however, isotopic variations suggest greater crustal influence. This might be expected if continued heating of the enclosing crust allowed for more extensive melting, further assimilation in late-erupted lavas, and greater depletion of incompatible elements in later crustal additions.
Continued heating may also explain the late apparent increase in H2O as suggested by the appearance of amphibole and 238U excesses (vs 230Th) in unit 3 lavas (Ramos, 2000). Heating may have been responsible for dehydrating nearby crustal materials, which, in turn, allowed for the preferential transport of U into relatively incompatible element depleted lavas, resulting in relative 238U enrichment.
Wide variations in both major element compositions of melt inclusions and 87Sr/86Sr are more easily explained in this scenario (Table 4; Fig. 4). Melt inclusions in olivine may be sampled throughout the magmatic system, including near the magma–crust interface. This would allow for a highly variable range of melts with correspondingly highly variable 87Sr/86Sr to be captured. In addition, the compositions of individual melt inclusions, especially MgO contents, may also vary sporadically and be more evolved than the host magma, depending on the specific chemical characteristics of the assimilant. Such a scenario would be consistent with 87Sr/86Sr variations occurring within the plagioclase stability field at or near crustal depths. Progressive crustal melting and assimilation accompanying cooling and crystallization would be responsible for the rise in mineral 87Sr/86Sr throughout the eruption suite. Radiogenic isotope signatures in melt inclusion-bearing olivine suggest an assimilant with 87Sr/86Sr 0·7056. These variations are accompanied by relatively uniform major element compositions, which are consistent with the assimilation or mixing of a basaltic andesite or andesitic composition melt generated from either (1) hydrated cumulates derived from earlier basaltic magmatism or (2) pelitic crust potentially associated with the Orocopia–Pelona–Rand schist (Miller et al., 1996), with low abundances of incompatible elements and high 87Sr/86Sr (>0·7056).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Late Pleistocene alkali basalts and hawaiites erupted at a single volcanic site, Pisgah Crater, are characterized by progressively greater open-system chemical and isotopic variations. The lava flow sequence exhibits an isotopic transition from OIB-like signatures to those expected for the lithosphere. Sr isotope variations in individual minerals and groundmass provide insight into open-system behavior at the microscale. Early flows are generally enriched in incompatible elements, and labradorite and titanaugite phenocrysts record only minor amounts of syn-crystallization contamination. The isotopic characteristics of these crystals (i.e. 87Sr/86Sr
0·7037–0·7040) establish a baseline starting composition from which later lavas evolved and reflect basalt derivation from a somewhat enriched mantle source, unless open-system processes affected these basalts at conditions other than those recorded by the phenocrysts. Late-erupted lavas exhibit 87Sr/86Sr isotopic variations between minerals and groundmass and within the groundmass of individual flows that are consistent with more extensive syn-crystallization contamination; olivine-hosted melt inclusions range from basaltic to andesitic compositions.
87Sr/86Sr variations in minerals are probably generated during growth in a magma characterized by progressively increasing 87Sr/86Sr. Diffusion is presumed to be a minor modifying effect on mineral 87Sr/86Sr signatures and constrains maximum plagioclase residence times to 1 kyr, suggesting that open-system processes occurred relatively quickly. Considered together with the progressive increases in whole-rock 87Sr/86Sr and decreases in whole-rock incompatible element abundances, these variations are consistent with assimilation of an incompatible element depleted, high 87Sr/86Sr mafic cumulate, possibly amphibolitic in grade, or a granulitic pelite. Thus, the isotopic and chemical heterogeneities in Pisgah Crater lavas more probably arise from crustal contamination rather than from mixing of heterogeneous mantle-derived magmas.
|
ACKNOWLEDGEMENTS |
|---|
We would like to express our gratitude to Scotty Cornelius for assistance with electron microprobe analyses, and to Frank Tepley for assistance with microdrilling. Reviews by R. Conrey, L. Farmer, D. Geist, J. Gill, D. Graham, J. Shervais, E. Widom, and J. Wolff improved earlier versions of the manuscript. This research was supported by NSF grant EAR-9418323 awarded to M.R.R. and the UCLA TIMS Laboratory.
|
FOOTNOTES |
|---|
Present address: Department of Geology, Northern Arizona University, Flagstaff, AZ 86001, USA. 
* Corresponding author. Present address: Department of Geological Sciences, Central Washington University, Ellensburg, WA 98926, USA. Telephone: 626 688-4457. Fax: 509 963-2821. E-mail: ramos{at}geology.cwu.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONBACKGROUNDSAMPLE PREPARATION AND ANALYSISRESULTSEFFECTS OF OPEN-SYSTEM BEHAVIOR...ORIGIN OF ISOTOPE AND...CONCLUSIONSREFERENCES |
|---|
Beattie, P. (1994). Systematics and energetics of trace element partitioning between olivine and silicate melts: implications for the nature of mineral/melt partitioning. Chemical Geology 177, 57–71.
Borisova, A. Y., Nikogosian, I. K., Scoates, J. S., Weis, D., Damasceno, D., Shimizu, N. & Touret, J. L. (2002). Melt, fluid, and crystal inclusions in olivine phenocrysts from Kerguelen plume-derived picritic basalts: evidence for interaction with the Kerguelen Plateau lithosphere. Chemical Geology 183, 195–220.[CrossRef][Web of Science]
Bourdon, B. & Sims, K. W. (2003). U-series constraints on intraplate basaltic magmatism. In: Bourdon, B., Henderson, G. M., Lundstrom, G. C. & Turner, S. (eds) U-series Geochemistry. Mineralogical Society of America, Reviews in Mineralogy and Geochemistry 52, 215–254.
Chesterman, C. W. (1982). Potentially active volcanic zones in California. California Division of Mines and Geology Special Publication 63, 9–16.
Costa, F. & Singer, B. (2002). Evolution of Holocene dacite and compositionally zoned magma, Volcan San Pedro, Southern Volcanic Zone, Chile. Journal of Petrology 43, 1571–1593.
Crank, J. (1975). The Mathematics of Diffusion, 2nd edn. Oxford: Clarendon Press.
Danyushevsky, L. V. & Leslie, R. A. (2002). The common occurrence of melt inclusions with exotic major-element compositions in high-Fo olivine phenocrysts: a result of localised wall-rock assimilation? Transactions, America Geophysical Union 47, 1492.
Danyushevsky, L. V., Della-Pasqua, F. N. & Sokolov, S. (2000). Re-equilibration of melt inclusions trapped by magnesian phenocrysts from subduction-related magmas: petrological implications. Contributions to Mineralogy and Petrology 138, 68–83.[CrossRef][Web of Science]
Davidson, J. P., Tepley, F. J. & Knesel, K. M. (1998). Crystal isotope stratigraphy; a method for constraining magma differentiation pathways. EOS Transactions, American Geophysical Union 79, 185–193.
Davidson, J. P., Tepley, F., Palacz, Z. & Meffan-Main, S. (2001). Magma recharge, contamination and residence times revealed by in-situ laser ablation isotopic analysis of feldspar in volcanic rocks. Earth and Planetary Science Letters 184, 427–442.[CrossRef][Web of Science]
Dempster, T. J., Preston, R. J. & Bell, B. R. (1999). The origin of massif-type anorthosite; evidence from interactions between crustal xenoliths and basaltic magma. Journal of the Geological Society, London 156(1), 41–46.
Dungan, M. A. & Davidson, J. P. (2004). Partial assimilative recycling of the mafic roots of arc volcanoes: an example from the Chilean Andes. Geology 32, 733–736.
Foley, S. (1992). Vein-plus-wall-rock melting mechanisms in the lithosphere and the origin of potassic alkaline magmas. Lithos 28, 435–453.[CrossRef][Web of Science]
Gaetani, G. A. & Watson, E. B. (2002). Modeling the major-element evolution of olivine-hosted melt inclusions. Chemical Geology 183, 25–41.[CrossRef][Web of Science]
Ghiorso, M. (1991). Fe–Ti oxide geothermometry; thermodynamic formulation and the estimation of intensive variables in silicic magmas. Macintosh program for determining temperatures using olivine–spinel geothermometry. Contributions to Mineralogy and Petrology 108, 485–510.[CrossRef][Web of Science]
Giletti, B. J. & Casserly, J. (1994). Strontium diffusion kinetics in plagioclase feldspars. Geochimica et Cosmochimica Acta 58, 3785–3793.[CrossRef][Web of Science]
Glazner, A. F. & Farmer, G. L. (1992). Production of isotope variability in continental basalts by cryptic crustal contamination. Science 255, 72–74.
Glazner, A. F., Farmer, G. L., Hughes, W. T., Wooden, J. L. & Pickthorn, W. (1991). Contamination of basaltic magma by mafic crust at Amboy and Pisgah Craters, Mojave Desert, California. Journal of Geophysical Research 96(B8), 13673–13691.[CrossRef]
Hirschmann, M. M. & Stolper, E. M. (1996). A possible role for garnet pyroxenite in the origin of the ‘garnet signature’ in MORB. Contributions to Mineralogy and Petrology 124, 185–208.[CrossRef][Web of Science]
Kamenetsky, V. S., Sobolov, A. V., Eggins, S. M., Crawford, A. J. & Arculus, R. J. (2002). Olivine-enriched melt inclusions in chromites from low-Ca boninites, Cape Vogel, Papua New Guinea: evidence for ultramafic primary magma, refractory mantle source and enriched components. Chemical Geology 183, 287–303.[CrossRef][Web of Science]
Luedke, R. G. & Smith, R. L. (1987). Map showing distribution, composition, and age of Late Cenozoic volcanic centers in California, Nevada, and Arizona. US Geological Survey Miscellaceous Investigations Series, Map 1-1091-B.
Lundstrom, C. C. (2003). U-series disequilibria in mid-ocean ridge basalts: observations and models of basalt genesis. In: Bourdon, B., Henderson, G. M., Lundstrom, G. C. & Turner, S. (eds) U-series Geochemistry. Mineralogical Society of America, Reviews in Mineralogy and Geochemistry 52, 175–214.
Miller, J. S., Glazner, A. F. & Crowe, D. E. (1996). Muscovite–garnet granite in the Mojave Desert: relation to the crustal structure of the Cretaceous arc. Geology 24, 335–338.
Morioka, M. (1983). Cation diffusion in olivine—III. Mn2SiO4 system. Geochimica et Cosmochimica Acta 47, 2275–2279.[CrossRef][Web of Science]
Nielson, R., Michaul, P. & Sours-Page, R. (1998). Chemical and physical indicators of compromised melt inclusions. Geochimica et Cosmochimica Acta 62, 831–839.[CrossRef][Web of Science]
Norman, M. D., Garcia, M. O., Kamenetsky, V. S. & Nielson, R. L. (2002). Olivine-hosted melt inclusions in Hawaiian picrites: equilibration, melting, and plume source characteristics. Chemical Geology 183, 143–168.[CrossRef][Web of Science]
Ramos, F. C. (2000). Mantle sources generating recent volcanism in the Western US. Ph.D. thesis, University of California, Los Angeles.
Reiners, P. (2002). Temporal–compositional trends in intraplate basalt eruptions; implications for mantle heterogeneity and melting processes. Geochemistry, Geophysics, Geosystems 3(2), paper number 10.1029/2001GC000250.
Roeder, P. L. & Emslie, R. F. (1970). Olivine–liquid equilibrium. Contributions to Mineralogy and Petrology 29, 275–280.[CrossRef][Web of Science]
Rollinson, H. (1993). Using Geochemical Data: Evaluation, Presentation, and Interpretation. New York: Longman.
Saal, A. E., Hart, S. R., Shimizu, N. & Layne, G. D. (1998). Pb isotopic variability in melt inclusions from oceanic island basalts, Polynesia. Science 282, 1481–1484.
Schiano, P. & Clocchiatti, R. (1994). Worldwide occurrence of silica-rich melts in sub-continental and sub-oceanic mantle minerals. Nature 368, 621–624.[CrossRef]
Simonetti, A., Shore, M. & Bell, K. (1996). Diopside phenocrysts from nephelinite lavas, Napak Volcano, eastern Uganda: evidence for magma mixing. Canadian Mineralogist 34, 411–421.[Web of Science]
Sneeringer, M., Hart, S. R. & Shimizu, N. (1984). Strontium and samarium diffusion in diopside. Geochimica et Cosmochimica Acta 48, 1589–1608.[CrossRef][Web of Science]
Sobolev, A. V. & Shimizu, N. (1993). Ultra-depleted primary melt included in an olivine from the Mid-Atlantic Ridge. Nature 363, 151–154.[CrossRef]
Sobolev, A. V., Hofmann, A. W., Nikogosian, I. K., Shimizu, N. & Chaussidon, M. (1999). New evidence for oceanic crustal recycling in the Hawaiian plume: trace elements and isotopes of melt inclusions in olivines from Mauna Loa lavas. European Union of Geosciences, Conference Abstracts 4(1), 344.
Tollstrup, D. L. (2003). Short timescales for crustal residence, transport and contamination of flood basalt magma: crystal isotope stratigraphy of the Columbia River Basalt Group. M.S. thesis, Washington State University, Pullman, WA.
Turner, S., Bourdon, B. & Gill, J. (2003). Insights into magma genesis at convergent margins from U-series isotopes. In: Bourdon, B., Henderson, G. M., Lundstrom, G. C. & Turner, S. (eds) U-series Geochemistry. Mineralogical Society of America, Reviews in Mineralogy and Geochemistry 52, 255–310.
Ussler, W. & Glazner, A. F. (1989). Phase equilibria along a basalt–rhyolite mixing line: implications for the origin of calc-alkaline intermediate magmas. Contributions to Mineralogy and Petrology 101, 232–244.[CrossRef][Web of Science]
Wise, W. S. (1969). Origin of basaltic magmas in the Mojave Desert area, California. Contributions to Mineralogy and Petrology 23, 53–64.[CrossRef]
Wolff, J. A. & Ramos, F. C. (2002). Assimilation yesterday, today, and tomorrow. Transactions, American Geophysical Union 83(47), F1406.
Wolff, J. A. & Ramos, F. C. (2003). Pb isotope variations among Bandelier Tuff feldspars: no evidence for a long-lived silicic magma chamber. Geology 31, 533–536.
Wolff, J. A., Ramos, F. C. & Davidson, J. P. (1999). Sr isotope disequilibrium during differentiation of the Bandelier Tuff: constraints on the crystallization of a large rhyolitic magma chamber. Geology 27, 495–498.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. S. Beard Crystal-Melt Separation and the Development of Isotopic Heterogeneities in Hybrid Magmas J. Petrology, May 1, 2008; 49(5): 1027 - 1041. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. D. Putirka Introduction to Minerals, Inclusions and Volcanic Processes Reviews in Mineralogy and Geochemistry, January 1, 2008; 69(1): 1 - 8. [Full Text] [PDF]
|
|
|
|
|
|
F. C. Ramos and F. J. Tepley III Inter- and Intracrystalline Isotopic Disequilibria: Techniques and Applications Reviews in Mineralogy and Geochemistry, January 1, 2008; 69(1): 403 - 443. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||