Journal of Petrology

Journal of Petrology Advance Access published online on December 17, 2008
Journal of Petrology, doi:10.1093/petrology/egn060

This Article Abstract FREE Full Text (PDF) Supplementary Data All Versions of this Article: 49/12/2123    most recent egn060v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Maria, A. H. Articles by Luhr, J. F. Search for Related Content GeoRef GeoRef Citation Social Bookmarking          What's this?

© The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

Lamprophyres, Basanites, and Basalts of the Western Mexican Volcanic Belt: Volatile Contents and a Vein–Wallrock Melting Relationship

Anton H. Maria^1,* and James F. Luhr^2,

¹Geology and Physics Department, University of Southern Indiana, 8600 University Boulevard, Evansville, IN 47712, USA
²Department of Mineral Sciences, Smithsonian Institution, PO BOX 37012, NHB-119, Washington, DC 20013-7012, USA

Received November 14, 2007; Revised typescript accepted October 30, 2008

	ABSTRACT

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
We present geochemical data for Quaternary basalts, basanites, and lamprophyres within the Colima and Mascota volcanic fields at the western end of the Mexican Volcanic Belt. On the basis of data for 11 whole-rock samples plus 124 glass inclusions and olivine host crystals, we evaluate a vein–wallrock melting relationship between the lamprophyres (vein-dominated melts) and the mafic calc-alkaline rocks (diluted by partial melting of peridotite wall-rock after exhaustion of phlogopite and other vein minerals). Whole-rock Fe³⁺/Fe²⁺ and glass-inclusion %S⁶⁺ indicate relatively high fO₂ in these magmas, up to several log units above the Ni–NiO buffer. The highest concentrations of water and most other volatile elements (7% H₂O, 1460 ppm CO₂, 2% SO₃^Total, 2400 ppm Cl, and 1% F) were recorded for a glass inclusion from a Colima minette with 48·2 wt % SiO₂, 6·0 wt % K₂O, and 1·2 wt % P₂O₅ (normalized anhydrous). This sample's volatile composition corresponds to a depth of entrapment of 24 km (calculated pressure of 6660 bars). This inclusion (trapped within olivine with Mg-number 91·5) represents the most primitive melt in this study and has a composition that can be attributed to partial melting of phlogopite-pyroxenite veins in the mantle wedge with minor dilution (possibly as little as 25%) by partial melts from the surrounding peridotite wall-rock. However, there are indications that even this inclusion has undergone degassing, suggesting that primary vein melts have even higher H₂O and CO₂ contents. Further dilution of the vein-dominated lamprophyre melts by wallrock melts yields basanites and ultimately calc-alkaline basalts. Mafic calc-alkaline whole-rock and glass-inclusion compositions are consistent with formation through mixing of 5% vein melts with 95% peridotite wallrock melt. Among the calc-alkaline glass inclusions, the Mascota basaltic andesite has the highest concentrations of water and most other volatile elements with 49·6 wt % SiO₂, 1·0% K₂O, 0·3% P₂O₅ (normalized anhydrous), 2·8% H₂O, 296 ppm CO₂ (1425 bars pressure and 5·2 km depth of entrapment), 0·8% SO₃^Total, 870 ppm Cl, and 720 ppm F. Such mafic calc-alkaline melts are envisioned as parental to the volumetrically dominant andesites of western Mexico.

KEY WORDS: igneous petrology; subduction; lamprophyre; mantle volatiles; vein–wallrock melting

	INTRODUCTION

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
Quaternary volcanism in the western portion of the Mexican Volcanic Belt (MVB) displays a high degree of compositional diversity relative to other parts of the MVB and most subduction-related volcanic arcs worldwide (Luhr, 1997; Luhr et al., 2006). Although calc-alkaline basalts to rhyolites are volumetrically dominant (andesite is the most commonly analyzed rock of the western MVB, Luhr et al., 2006), there are also hypersthene-normative intraplate-type basalts, and a wide variety of hypersthene- to nepheline-normative alkaline rocks, including lamprophyres (Luhr & Carmichael, 1981; Wallace & Carmichael, 1989, 1992; Lange & Carmichael, 1990, 1991; Carmichael et al., 1996; Luhr, 1997; Luhr et al., 2006; Vigouroux et al., 2008). This diversity of magma compositions has been attributed to the superposition of active continental rifting atop an environment of plate convergence associated with subduction of the Rivera Plate beneath the SW corner of the North American Plate (Luhr et al., 1985; Allan et al., 1991). The continental rifting is manifest in a structural triple junction involving the Colima Rift Zone, the Chapala Rift Zone, and the Tepic–Zacoalco Rift Zone, which intersect SSW of Guadalajara (Fig. 1) and define the NE corner of the Jalisco Block. Although not well constrained, rifting apparently began in the Pliocene, which also is the age of the earliest alkaline magmatism in this area (Luhr et al., 1985).

The alkaline magmas of Colima and Mascota form scoria cones and lava flows in close temporal and spatial association with the calc-alkaline rocks of the large central volcanoes of this area; they are distinguished from basalts by an absence of plagioclase phenocrysts and high concentrations of Ba, Sr, K, and P (Luhr & Carmichael, 1981; Carmichael et al., 1996). These rocks form transitional series marked by an increase in abundance of hydrous minerals from basanite to lamprophyre (i.e. the presence of phlogopite distinguishes the minettes from the basanites). High Mg-numbers [atomic 100 MgO/(MgO + FeO^total)], ranging from approximately 70 to 90, and relatively high concentrations of incompatible elements (Luhr & Carmichael, 1981; Carmichael et al., 1996) as well as somewhat radiogenic Sr, Nd, and Pb isotope ratios (Luhr, 1997) can be reconciled by derivation from a mantle source enriched metasomatically by fluids from a subducting slab (Luhr et al., 1989; Wallace & Carmichael, 1989, 1992; Allan et al., 1991; Carmichael et al., 1996; Luhr, 1997).

The coexistence of potassic and calc-alkaline magmas at subduction zones has been documented worldwide (e.g. Nicholls & Whitford, 1983; Ellam et al., 1988; Bloomer et al., 1989; Koloskov et al., 1999), yet the relationship between these magmas is still poorly understood. For the potassic lamprophyres and calc-alkaline basalts of the western MVB, similar element abundance patterns and Sr, Nd, and Pb isotopic ratios (Luhr, 1997) have prompted the proposal of a genetic connection reflecting a vein–wallrock melting relationship (Luhr et al., 1989; Carmichael et al., 1996; Luhr, 1997), following the model of Foley (1992). According to this view, the lamprophyres represent partial melts of phlogopite–pyroxenite veins in the sub-arc mantle wedge, and are able to ascend to the surface with little differentiation because of the active continental rifting. When the same vein-melt component is diluted by partial melting of the surrounding mantle peridotite, first basanites, and ultimately calc-alkaline basalts can be produced. The basalts differentiate to form the volumetrically dominant andesites and the rest of the calc-alkaline suite.

In this study we focus on glass inclusions trapped in olivine phenocrysts from the Colima Volcanic Complex and the Mascota Volcanic Field. These rocks and their glass inclusions provide a rare window into the volatile-rich environment of the mantle wedge, yielding insights into the origins of subduction zone volcanism. We present major-, trace-, and volatile-element compositions for 124 glass inclusions: 24 from basanites, 69 from minettes, six from hornblende lamprophyres, and 25 from calc-alkaline basalts and basaltic andesites. Volatile contents of glass inclusions were measured by electron microprobe (EMP; S,%S⁶⁺, Cl), Fourier-transform IR spectroscopy (FTIR; H and C species), and secondary-ion mass spectrometry (SIMS; H, C, F, S, Cl). Using these data, we further explore the nature of the mantle source, the origin of the primary melts, ascent and degassing of magmas, and the relationships between lamprophyres and calc-alkaline basalts and basaltic andesites of the western MVB.

	GEOLOGICAl SETTING

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
Colima Volcanic Complex
The Colima Volcanic Complex includes three north–south-aligned calc-alkaline volcanic centers that have migrated southward by 25 km during the Quaternary (Cortés-Cortés et al., 2005). Volcán Cántaro, in the north, is an eroded andesite–dacite volcanic complex that rises to 2925 m from a base at about 1600 m. Allan (1986) reported four K–Ar ages of 1· 66–0·95 Ma for Cántaro lavas and domes. Cántaro is the only volcano in the chain to have erupted biotite-bearing products, an important correlation tool. Following the formation of Volcán Cántaro, the focus of magmatism shifted southward by about 16 km, and the massive andesitic cone of Nevado de Colima began to grow (Fig. 1). Robin et al. (1984, 1987, 1990) discussed the geological and petrological development of Nevado, and presented seven K–Ar ages (ranging from 0·53 to 0·14 Ma) to support their interpretation of a three-stage evolution punctuated by two episodes of caldera formation. Nevado de Colima's youngest cone is now only slightly eroded and forms the summit, the highest peak (4320 m) in western Mexico. Toward the late stages of activity at Nevado, the focus of magmatism shifted southward again by 5 km, with growth of the ancestral Volcán de Colima (or Paleofuego). Volcán de Colima collapsed to form a 5 km diameter caldera, open toward the south, during one or more events. The collapse history is not fully resolved (Luhr & Prestegaard, 1988; Komorowski et al., 1997), but the most recent collapse occurred 3000 years ago, based on ¹⁴C analysis of charcoal, and was followed by growth of the modern cone of Volcán de Colima (Luhr & Carmichael, 1980, 1990; Luhr, 2002; Zobin et al., 2002).

Eleven mafic scoria cones on the NE and NW flanks of Nevado de Colima (Luhr & Carmichael, 1981; Hooper, 1995; Carmichael et al., 2006) include two that are calc-alkaline and nine that are alkaline (Luhr & Carmichael, 1981). Carmichael et al. (2006) published ⁴⁰Ar/³⁹Ar ages for all of these scoria cones. Of the nine alkaline scoria cones, which include ne-normative basanites, leucite basanites, and minettes (phlogopite-bearing lamprophyres), eight of them, including all of the minettes, lie on the western side of the Cántaro–Colima chain. The basanitic Cerro Apaxtepec, centered 14 km NE of Nevado, is the exception. Cerro Apaxtepec has the most youthful cone and lava-field morphologies, and the youngest ⁴⁰Ar/³⁹Ar age at 62 ka. Of the eight alkaline cones along the west side of the chain, the oldest is a basanite cone, dated at 473 ka, which lies 32 km WNW of Nevado. The remaining seven, with ages from 137 to 240 ka, lie 15 km west of Cántaro.

The two calc-alkaline cones include Cerro Tezontle, a basaltic scoria cone that lies 22 km ENE of Nevado and is dated at 545 ka, and a trachybasalt cone that lies 21 km north of Cántaro and has an age of 1·22 Ma. With the exception of this trachybasalt cone, which is not considered further in this study, the ages of the flanking scoria cones indicate that they erupted during growth of Nevado de Colima. A small number of alkaline scoria deposits sampled from the slopes of Volcán Colima (Luhr & Carmichael, 1982; Calanchi et al., 1995) have magma-mingling characteristics also consistent with coeval eruption of basanite–minette-type magmas and the volumetrically dominant calc-alkaline magmas.

Mascota Volcanic Field
In the northern part of the Jalisco Block (Fig. 1), several Plio-Pleistocene volcanic fields are associated with grabens that cut Cretaceous pyroclastic deposits. Within the Mascota, Los Volcanes, and San Sebastián volcanic fields, calc-alkaline basalts to andesites erupted contemporaneously with mafic alkaline magmas including lamprophyres. Mascota is the youngest of the three fields, and is mostly confined to two NNW–SSE-trending grabens (Carmichael et al., 1996, fig. 3). Carmichael et al. (1996) published K–Ar ages ranging from 0·49 ± 0·08 Ma to 0·07 ± 0·08 Ma for Mascota minettes based on four phlogopite separates and one whole-rock sample. On the basis of morphological evidence, the youngest eruption took place at Volcán Malpais, north of the town of Mascota, and produced a fresh-looking basaltic andesite lava flow that lacks soil development, supports only sparse vegetation, and may be just a few thousand years old. Similar but older rocks are found to the east in the Los Volcanes volcanic field (1·6–3·4 Ma: Wallace & Carmichael, 1989, 1992), and to the north in the San Sebastián field (0·26–0·48 Ma: Lange & Carmichael, 1990, 1991).

	SAMPLES STUDIED

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
We analyzed glass inclusions from 11 samples of scoria, six from the Colima field and five from the Mascota field (Table 1). At Colima the six samples represent five cones: Cerro Tezontle, Cerro Apaxtepec, Cerro Telcampana, and Volcáns San Isidro and La Erita (Luhr & Carmichael, 1981; Luhr, 1997). The five samples studied at Mascota represent four cones: La Esperanza, Volcán Molcajete Galope, Volcán Novillero, and Volcán Molcajete.

View this table: [in this window] [in a new window]   Table 1: Classification and location of whole-rock samples studied

 

View larger version (39K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 1. Location map for Colima and Masocota volcanic fields, marked by triangles, within the Western Mexican Volcanic Belt (after Carmichael et al., 1996). Other Quaternary volcanic centers are also labeled. The large dashed rectangle represents the perimeter of the Michoacan–Guanajuato Volcanic Field. The bold dashed lines represent rift zones. Offshore, the bold black lines represent segments of the East Pacific Rise mid-ocean ridge.

 
Tephra samples were collected from quarry walls and road cuts. At each location we collected small scoria clasts (<3 cm and preferably of the order of 1 cm in diameter) as well as finer materials that passed through a centimeter-size sieve. Our goal was to collect rapidly quenched samples with melt inclusions that were frozen quickly to glass. For the scoria samples (all those from Colima plus MAS02-6A, -6B), one fraction of 250 g was cleaned, dried, and powdered for whole-rock analysis. A second fraction of 50 g was cleaned, dried, gently crushed with an agate mortar and pestle, and then sieved to retain the 0·35–0·70 mm fraction. Samples containing ash-sized material (MAS02-1, -2, and -3) were sieved without crushing to retain the same size range.

Olivine crystals were separated and processed to expose glass inclusions following procedures described by Luhr (2001) and discussed below. Basic descriptions of selected glass inclusions studied are given in Table 2 (see Electronic Appendix 1, http://www.petrology.oxfordjournals.org, for descriptions of all inclusions studied), and representative thin-section photographs are shown in Fig. 2. Glass inclusions are not abundant in olivine crystals from any of these samples. The majority of the inclusions are ellipsoidal in shape, but many other forms are also present, including some best described as irregular. We measured or estimated size in three dimensions for each glass inclusion. The average largest dimension is 200 µm. Some of the largest inclusions (e.g. 600 µm) might be described as aggregates, as they appear to represent several inclusions that merged. Bubble size was also measured, and by far the most common situation is a single spherical bubble, representing 8 ± 8% of the inclusion volume on average. It is notable that all (n = 8) of the glass inclusions from the Mascota minette cone Novillero are characterized by multiple bubbles (Fig. 2b) and orange–brown glass with very low H, C, and S contents. Eleven other glass inclusions, one from Mascota basaltic andesite MAS02-3, and the others from Colima basanites and minettes, have multiple or large vapor bubbles (>8 vol. %). Of these 19 glass inclusions with high numbers of bubbles, all but one have SiO₂ > 52% and characteristics of degassed melt entrapped at relatively shallow crustal levels.

View this table: [in this window] [in a new window]   Table 2: Description of melt inclusions in the studied samples

 

View larger version (78K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 2. Photomicrographs of glass inclusions. (a) Glass inclusion within an olivine crystal from La Erita minette sample VF99-08J, as viewed under plane light in immersion oil. (b) Olivine crystal and glass inclusion with multiple bubbles from Novillero minette sample MAS02-1B.

 

	ANALYTICAL TECHNIQUES

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
A scoria clast from each sample was selected for thin sectioning and modal analysis. Over 1000 points were counted for each sample (Table 3). Designations of phenocryst and microphenocryst correspond to crystals larger than 0·3 mm and crystals between 0·3 and 0·03 mm, respectively. Matrix includes glass and microlites (<0·03 mm).

View this table: [in this window] [in a new window]   Table 3: Point-counted modes for samples investigated (values are vol. %)

 
Whole-rock scoria samples were analyzed for major and trace elements by X-ray fluorescence (XRF) spectrometry using the Smithsonian's Philips PW 1480 instrument (Table 4). Concentrations of FeO were determined by titration with K-dichromate following a modified version of the method of Peck (1964). Fe₂O₃ was then calculated from the XRF value for total iron. The same powdered whole-rock scoria samples were also analyzed for trace elements by inductively coupled plasma mass spectrometry (ICP-MS) at Washington State University (Table 4). Detailed discussion of analytical techniques has been given by Luhr & Haldar (2006), and estimates of analytical precision are listed in Table 4.

View this table: [in this window] [in a new window]   Table 4: Whole-rock compositions of scoria samples measured by XRF and ICP-MS

 
In addition to analysis of whole-rock samples, we measured major- and trace-element compositions as well as volatile contents of glass inclusions trapped in olivine hosts. Olivine crystals were separated from the 0·35–0·7 mm sieved fraction using heavy liquids (Na-polytungstate), and then leached in fluoroboric acid to remove any adhering matrix glass. Glass inclusions were detected by observing the crystals in immersion liquid (index of refraction = 1·678) with a microscope. Selected crystals were then oriented and mounted in epoxy to position inclusions for grinding and polishing in preparation for analysis.

The major and minor element compositions of 124 glass inclusions and their host olivine crystals (data for selected inclusions are presented in Table 5; see Electronic Appendix 2, http://www.petrology.oxfordjournals.org, for complete dataset), as well as of the glassy or microcrystalline groundmass of the 11 scoria samples (Table 6) were measured using the Smithsonian's JEOL 8900 EMP. Standards included a combination of natural glasses and minerals (Jarosewich et al., 1980). Following procedures described by Luhr (2001), the conditions selected for analysis of glass inclusions and groundmass glass were 10 nA beam current, 15 kV accelerating potential, and 10 µm beam diameter. The search for crystal-free glass for groundmass analysis was conducted in back-scatter electron mode. Volatile elements S and Cl were analyzed using a significantly higher beam current of 80 nA. Analysis of S in glass inclusions included a peak-seek routine, calibrated with anhydrite and troilite, and followed techniques established by Carroll & Rutherford (1988), Wallace & Carmichael (1994), and Matthews et al. (1999), to estimate the S⁶⁺/S^2– ratio (expressed as %S⁶⁺: Table 5 and Electronic Appendix 2). Sulfur concentration was measured at the mean peak position determined for each inclusion, using a 60 s count time and the Smithsonian scapolite standard (Jarosewich et al., 1980). For analysis of olivine host crystals, instrumental conditions were set at 20 nA beam current, 15 kV accelerating potential, and beam diameter of 1 µm. Detailed discussion of these analytical techniques has been presented by Luhr (2001), and Luhr & Haldar (2006).

View this table: [in this window] [in a new window]   Table 5: Electron microprobe data for olivine-hosted glass inclusions, adjacent olivine, and corrections for post-entrapment crystallization

 

View this table: [in this window] [in a new window]   Table 6: Electron-microprobe data for glassy–microcrystalline groundmass material

 
Following EMP analysis, we attempted to doubly polish each glass inclusion (ideally producing a rectangular window >25 µm on a side through only glass) for analysis of hydrogen (hydroxyl: OH^–; molecular water: H₂O_molecular; total water: H₂O^total) and C (carbonate: CO₃^2–) species by FTIR spectroscopy using the Smithsonian's Bio-Rad Excalibur FTIR bench with attached MA-500 microscope. A KBr beam splitter and liquid-nitrogen-cooled MCT detector were used for all analyses. Procedures following Nakamoto (1978), Stolper (1982), and Dixon et al. (1995), described by Luhr (2001), were used for this work. We obtained data for H₂O^total in 89 glass inclusions and for CO₃^2– in 17 of those (Table 7 and Electronic Appendix 3, http://petrology.oxfordjournals.org). For 10 of the inclusions an unfortunate combination of high water content and low wafer thickness resulted in saturation of the mid-IR total water peak at 3535 cm^–1 and masking of the near-IR peaks by fringes on the spectra.

View this table: [in this window] [in a new window]   Table 7: FTIR data for hydrogen and carbon species in glass inclusions

 
The whole-rock powder of Mascota minette MAS02-1A, devolatilized in a 1 atm furnace at 1330°C with oxygen fugacity equivalent to the synthetic Ni–NiO (NNO) buffer (Huebner & Sato, 1970), was used as a background reference sample for the carbonate peaks. FTIR background corrections were performed using a least-squares peak-fitting routine, written by S. Newman (written communication to J. Luhr), and the Solver tool in Excel to subtract the spectrum of devolatilized glass. Absorption coefficients for the 1430, 1515, 1630, 4500, and 5200 cm^–1 peaks were calculated using the major-element compositions of the glass inclusions as measured by EMP and the equations reported by Dixon & Pan (1995) and Dixon et al. (1995), augmented by the approach described by Luhr (2001).

As a test of the FTIR methodology, we also analyzed four calc-alkaline andesite–rhyolite glasses for which H₂O^total has been determined by H-manometry (C. Mandeville, personal communication), with results listed in Table 7 and Electronic Appendix 3. Considering the typical 1 uncertainties associated with calculation of species abundances from the Beer–Lambert Law (absorbance measurement 5%, glass density 0·5%, wafer thickness 6%, and molar absorptivity 6%) the 1 uncertainties for single FTIR measurements are estimated as 10%.

Following FTIR analysis, the same doubly polished wafers were prepared for analysis by SIMS. The olivine + glass wafers were plucked out of the mounts, cleaned in acetone, and remounted with tiny drops of superglue on a polished 1 inch round aluminum disk. Concentrations of five volatile elements (H, C, F, S, Cl) were measured for 22 glass inclusions (Table 8) using the Cameca IMS 6f ion microprobe at the Department of Terrestrial Magnetism, Carnegie Institution of Washington, following the procedures described by Hauri et al. (2002). Analytical uncertainties for measurement of volatiles are estimated as 10%.

View this table: [in this window] [in a new window]   Table 8: Comparison of volatile elements in glass inclusions as determined by FTIR, SIMS, and EMP

 
The last analytical step was determination of trace-element concentrations in glass inclusions (n = 14: Table 9) by laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) at the University of Maryland, following the procedures discussed by Michael et al. (2002). In preparation, each wafer was removed from its mount, cleaned in acetone, flipped, and remounted. The highest quality data were produced by passing a 55 µm spot back and forth over the larger inclusions. For smaller glass inclusions, stationary spots with diameters of 55, 30, and 15 µm were used, with progressively greater analytical errors. Most of the analyses were calibrated using ⁴³Ca as an internal standard; two samples were calibrated with ⁴⁹Ti. NIST 612 was used as a calibration standard, with analyses of VG-2, VG-99, VG-568, and BCR-26 serving as secondary standards.

View this table: [in this window] [in a new window]   Table 9: LA-ICP-MS data for trace elements in glass inclusions

 
The attrition rate was particularly high for glass inclusions from the Mascota samples. Only four of the Mascota inclusions survived to be analyzed by SIMS (three from basaltic andesite MAS02-3B and one from hornblende lamprophyre MAS02-2B), and only two of these, one from each sample, were analyzed by LA-ICP-MS.

	RESULTS

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
Classification and major elements
Classification of the Colima and Mascota alkaline rocks is problematic. Although modern rock classification generally follows the IUGS guidelines established by Le Bas et al. (1986), the TAS diagram (Fig. 3a) does not reflect the mineralogical criteria used in naming lamprophyres and related rocks. The uniqueness of these rocks can only be appreciated in their petrography, particularly for those characterized by the presence of phlogopite or amphibole in the absence of early formed feldspars. In Table 1 we list rock names for each sample. Two samples from this study are typical, subduction-related, mafic calc-calkaline rocks. For these samples we adopt the terminology of the TAS diagram (Fig. 3a). Thus, SAY-22F is a basalt and MAS02-3 is a basaltic andesite. The alkaline rocks are named based on the original classification scheme used by Luhr & Carmichael (1981) and Carmichael et al. (1996). Hornblende lamprophyre is characterized by hornblende, olivine, and clinopyroxene phenocrysts and an absence of early formed plagioclase. Minettes are similar, with phenocrysts of olivine, and augite, but contain phlogopite as the hydrous phase. Basanites have nepheline in the CIPW norm, and compositions transitional to minette, but lack phlogopite phenocrysts. We continue to describe Colima's Apaxtepec deposits as basanitic, maintaining consistency with earlier literature, despite the fact that our scoria samples (VF9901-A and -B) have hypersthene in the norm instead of nepheline. It is possible that the slightly lower concentrations of Na₂O in these samples is due to leaching. For the same reason, we describe Colima's La Erita as a minette, although phlogopite is restricted to the groundmass in our sample.

View larger version (30K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 3. Total alkalis vs silica wt %. (a) The 11 whole-rock samples from this study (Table 4) and their respective groundmass compositions (Table 6) connected by tie lines, with classification boundaries after Le Bas et al. (1986). (b) The glass inclusions from this study (Table 5 and Electronic Appendix 2).

 
The glass inclusions from Colima and Mascota exhibit a range of compositions, containing 44·0–58·6 wt % SiO₂, 1·6–8·8% MgO, 0·7–8·0% K₂O, and 0·15–1·5% P₂O₅ (Table 5 and Electronic Appendix 2). On a TAS plot (Fig. 3b), compositional fields for inclusions within the alkaline rocks overlap (Fig. 3b), yet remaining distinct from the field for inclusions from calc-alkaline samples. For Fig. 3b and all other figures depicting major-element glass inclusion data, compositions were corrected to account for possible post-entrapment crystallization of the host olivine. Major-element compositions of the inclusions were adjusted to values in equilibrium with the adjacent host-olivine compositions (Luhr, 2001) using the olivine–melt Fe–Mg-exchange relations of Ulmer (1989). The largest source of uncertainty in this calculation is in estimating the %Fe³⁺ in the melt. We have assumed that 20% of total Fe is present as Fe³⁺ for glass inclusions from calc-alkaline basalt SAY-22F and basaltic andesite MAS02-3, and 40% Fe³⁺ for glass inclusions from basanites and lamprophyres. Although these assumptions are not entirely consistent with estimates of %S⁶⁺ (e.g. MAS02-3), these corrections (as high as 8% added olivine, but average of 3%; Table 5 and Electronic Appendix 2) have little effect on glass compositions beyond SiO₂ and MgO. Increasing the assumed %Fe³⁺ in the melt from 20% to 40% reduces melt Fe²⁺/Mg by an amount equivalent to a reduction of calculated post-entrapment olivine crystallization by 5 wt %.

The compositions of glass inclusions may also have been affected by loss of Fe to the olivine host by diffusion. Correction of Fe loss by diffusion (Danyushevsky et al., 2000) can be achieved by adjusting the Fe compositions of inclusions so that they are consistent with trends in MgO vs FeO^T for whole-rock and groundmass samples from a given location. Unfortunately, in this study we do not have sufficient whole-rock and groundmass data to correct for possible diffusive loss of Fe.

Glass inclusions from the Colima and Mascota volcanic rocks exhibit a range in SiO₂ of 7%, and a comparable range in host olivine Mg-number (Table 5 and Electronic Appendix 2; Fig. 4). A plot of glass Mg-number vs SiO₂ yields very nearly the same pattern, indicating coherence with respect to Mg-number between olivine and inclusions from a given sample. The highest olivine Mg-numbers (88·0–92·1) are for Colima minettes and the basanite from Telcampana. The basanites from the Colima cone Cerro Apaxtepec and the calc-alkaline basalts from the Colima cone Tezontle are more evolved and contain olivine with Mg-number 81·6–88·5. In contrast, the Mg-number of olivine from the Mascota samples is much less variable, ranging from 84·4 to 88·9. Most glass inclusions have SiO₂ of 52%, but concentrations reach 58–59% SiO₂, with maxima at olivine Mg-number 90–92 for the Colima cones San Isidro (minette) and Telcampana (basanite), and at Mg-number 86–89 for samples from Mascota cones Molcajete (minette) and La Esperanza (basaltic andesite), and the youngest Colima cone Apaxtepec (basanite). There is no correlation between SiO₂ and Mg-number (of glass or olivine); low-silica inclusions (<52%) and high-silica inclusions (>52%) are found in basaltic andesites, basanites, and lamprophyres, alike. However, relatively low concentrations of H₂O, CO₂, and SO₃^Total suggest that the glasses with >52% SiO₂ represent sampling of shallow-level degassed melts. Given our focus on melting and volatile conditions within the mantle, our attention is centered on the 76 glass inclusions with <52% SiO₂. Unfortunately, this criterion diminishes the relevance of the Mascota minette glass inclusions, as all but one have >52% SiO₂.

View larger version (18K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 4. Olivine Mg-number vs wt % SiO2 of glass inclusions (olivine-corrected and normalized to 100%). Corrections to glass compositions for post-entrapment crystallization were made assuming 20% Fe3+ in the melt for calc-alkaline basalt SAY22-F and calc-alkaline basaltic andesite MAS02-3, and 40% Fe3+ in the melt for all the lamprophyres.

 
On Harker diagrams (Fig. 5), the samples can be split into low- and high-Ti groups (Wallace & Carmichael, 1989), separated by a value of 1·35 wt % TiO₂ (Fig. 5a). The low-Ti group includes samples from the calc-alkaline Colima basalt SAY-22F and Mascota's basaltic andesite MAS02-03, as well as the Mascota hornblende lamprophyre and Colima's basanite cone Telcampana. The high-Ti group includes all minettes and the Colima basanite cone Cerro Apaxtepec. The Al₂O₃ plot (Fig. 5b) shows this same spread, but inverted. MgO concentrations (Fig. 5c) are relatively constant for glass inclusions with >52% SiO₂. In other words, similarly primitive glasses (5% MgO) are found in calc-alkaline basalt, basaltic andesite, hornblende lamprophyre, and minette. The alkaline and calc-alkaline sample sets are distinct with respect to K₂O (Fig. 5d), with inclusions from alkaline rocks plotting above the two calc-alkaline samples. The P₂O₅ plot (Fig. 5e) looks very much like the K₂O plot (Fig. 5d) for glass inclusions with <52% SiO₂. At higher SiO₂ values, the inclusions trend to lower P₂O₅, possibly as a result of apatite precipitation from the evolving groundmass melts.

View larger version (23K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 5. Harker variation diagrams for glass inclusions using olivine-corrected compositions normalized to 100% (for the major-element sum SiO2, TiO2, Al2O3, FeOT, MnO, MgO, CaO, Na2O, K2O, P2O5, SO3Total, and Cl): SiO2 vs (a) TiO2, (b) Al2O3, (c) MgO, (d) K2O, and (e) P2O5. Whole-rock and groundmass compositions are depicted by open fields.

 
Whole-rock and groundmass compositional fields are also plotted for the alkaline rocks (Fig. 5), revealing a notable difference between samples from Colima and Mascota. For most oxides, compositions of Mascota inclusions trend toward the groundmass field, consistent with melts captured during ascent-related degassing and crystallization. This relationship does not hold for K₂O (Fig. 5d), however; Mascota's minettes trend toward higher K₂O with increasing SiO₂ to a point 4 wt % higher in K₂O than the groundmass. In contrast, whereas the Mascota inclusions are bounded by the whole-rock and groundmass compositions, with respect to SiO₂, SiO₂ values for inclusions from Colima range to lower than the whole-rock values and higher than the groundmass values. The glass inclusions from Colima's San Isidro minette pass through the groundmass composition and beyond it by 5% SiO₂. Such deviations between inclusion compositional trends and whole-rock and groundmass fields suggest the possibility of magma mixing or assimilation. The elevated MgO composition of the Colima whole-rock field is consistent with olivine accumulation.

Trace elements in whole-rock samples and glass inclusions
The Colima and Mascota lamprophyres and their glass inclusions are enriched in incompatible trace elements (Tables 4 and 9, respectively) relative to those from the calc-alkaline basalt and basaltic andesite (and the andesites of the nearby Cántaro–Colima chain). On normalized multi-element diagrams the calc-alkaline whole-rock samples from Colima (Tezontle basalt) and Mascota (La Esperanza basaltic andesite) define the lower edges of the shaded regions of Fig. 6a and b, respectively. For Colima and Mascota, the alkaline whole-rock samples mirror the calc-alkaline patterns [excluding the heavy rare earth elements (HREE) to the right of Ti], but are shifted to higher degrees of enrichment. Such subparallel incompatible-element behavior was predicted by Foley (1992) for magmas related by vein vs wall-rock melting. Although not visible from the shaded regions, the Colima and Mascota minettes are not readily distinguished on such plots. The Mascota hornblende lamprophyre (MAS02-2A) plots with the Colima basanites.

View larger version (36K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 6. Multi-element diagram normalized to primitive mantle (Sun & McDonough, 1989), with elements arranged in order of decreasing incompatibility in oceanic basalts from left to right. (a) Glass inclusions from Colima (Table 9). The shaded area shows the compositions of whole-rock samples from Colima (Table 4). (b) Glass inclusions from Mascota (Table 9). The shaded area shows the compositions of whole-rock samples from Mascota (Table 4).

 
Trace-element abundances for Colima and Mascota glass inclusions (Table 9) are not unlike those of their whole-rock counterparts, producing similar ‘stacked’ patterns for elements to the left of Ti (Fig. 6a and b). Like the whole-rock data, glass inclusions from calc-alkaline basalt SAY-22F and basaltic andesite MAS02-3 exhibit the lowest trace-element abundances, whereas the glass inclusions from minettes and the hornblende lamprophyre exhibit the highest abundances. A similar pattern of enrichment, with respect to Sr (Fig. 7), is evident for glass inclusions with SiO₂ < 52%. Sr concentrations in inclusions from the calc-alkaline basalt are lower than those in inclusions from basanite and hornblende lamprophyre, which, in turn, are lower than those in inclusions from minette.

View larger version (10K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 7. Glass-inclusion variations of SiO2 wt % (Table 5 and Electronic Appendix 2) vs Sr ppm (Table 7).

 
Like the normalized trace element diagrams, chondrite-normalized REE plots (Fig. 8) exhibit subparallel behavior between the calc-alkaline and alkaline rocks. Whole-rock samples from Colima (shaded region of Fig. 8a) display systematic increases of light REE (LREE) and middle REE (MREE) from the calc-alkaline basalt, through basanites, to minettes. The REE patterns cross, however, near Ho–Dy, such that the calc-alkaline basalt has the highest HREE values (0·33 ppm Lu) and a minette the lowest (0·19 ppm Lu). The Mascota whole-rock samples (shaded region of Fig. 8b) show similar LREE and MREE enrichments, but essentially constant HREE values (0·21–0·24 ppm Lu). The less precise LA-ICP-MS data for the glass inclusions show less regular, but similar REE patterns (Fig. 8a and b).

View larger version (29K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 8. Chondrite-normalized REE patterns (Sun & McDonough, 1989). (a) Glass inclusions from Colima (Table 9). The shaded area shows the compositions of whole-rock samples from Colima (Table 4). (b) Glass inclusions from Mascota (Table 9). The shaded area shows the compositions of whole-rock samples from Mascota (Table 4).

 
Volatile elements: H, C, F, S, %S⁶⁺, and Cl
We measured H₂O^total contents in 17 glass inclusions by FTIR (Table 7 and Electronic Appendix 3) and SIMS (Table 8). Our FTIR values are systematically slightly low compared with the H-manometry values (Fig. 9a). A positive correlation (R² = 0·66) exists between H₂O^total determined by FTIR and ‘H₂O by difference,’ which equals 100% minus the EMP major-element total (Fig. 9b). However, many samples with low measured water values by FTIR have unexpectedly high ‘difference’ values, so that the regression line cuts the y-axis at 1·8 wt % ‘H₂O by difference’. Although analytical conditions were selected to minimize volatilization effects, it is possible that Na loss during EMP analysis may account for the discrepancy. An excellent correlation (R² = 0·95) between H₂O^total values determined by FTIR and SIMS is seen in Fig. 9c, with some scatter at the high end.

View larger version (19K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 9. Comparisons of volatile element abundances by different techniques (EMP data in Table 5 and Electronic Appendix 2; FTIR data in Table 7 and Electronic Appendix 3; SIMS data in Table 8): (a) H2OTotal by FTIR vs H2OTotal by H-manometry (Table 7) for four experimental glasses from C. Mandeville (personal communication); (b) H2OTotal by FTIR vs 100 wt % minus EMP total; (c) H2OTotal by FTIR vs by SIMS; (d) CO2 by FTIR vs by SIMS; (e) S by EMP vs by SIMS; (f) Cl by EMP vs by SIMS. Best linear fit shown as continuous line; 1:1 correlation line is dashed. Error bars for SIMS data show ±1 standard deviation for the average of five counts collected for each measurement (errors smaller than the symbol size are not shown). Error bars for CO2 by FTIR show ±1 standard deviation for the average of measurements made at the 1515 cm–1 and 1435 cm–1 peaks.

 
We measured CO₂ contents for 11 of the same glass inclusions by the same two techniques (Fig. 9d); although the overall trend is positive, the correlation is poor (R² = 0·14). One likely source of analytical problems is the unusual mafic alkaline nature of the samples, for which neither technique is well calibrated. Because the SIMS calibration (Hauri et al., 2002) has shown no compositional dependence, we prefer the SIMS CO₂ results. Reduced confidence in the FTIR measurements also stems from our inability to obtain more than a few reliable CO₂ peaks.

We measured S and Cl contents in 21 glass inclusions by a different pair of techniques, EMP (Table 5 and Electronic Appendix 2) and SIMS (Table 8), as shown in Fig. 9e and f, respectively. The technique correlations for S and Cl are excellent and good (R² = 0·96 and 0·80), respectively, but both correlation lines are slightly steeper than 1 : 1.

Variations of SiO₂ vs H₂O, CO₂, S, Cl, and F are shown in Fig. 10. Measured H₂O concentrations as high as 7 wt % support the high primary volatile contents suggested by the presence of hydrous minerals and the absence of plagioclase phenocrysts. Experimental phase-equilibrium studies have firmly established that these mineral assemblages require high water contents (> 4 wt % H₂O: Righter & Carmichael, 1996). The highest values for H₂O (Fig. 10a), CO₂ (Fig. 10b), and S (Fig. 10c) are in glasses with <52% SiO₂; at higher SiO₂ contents each parameter falls to lower values. These declines with increased SiO₂ contrast with the behavior of Cl (Fig. 10d) and F (Fig. 10e), which show constant to slightly increasing values with increased SiO₂ for inclusions from particular cones, reflecting their lower vapor/melt partition coefficients (Carroll & Webster, 1994). Glass inclusions from the Mascota minettes are richer in Cl than all but one of the Colima minette glasses.

View larger version (15K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 10. SiO2 wt % (olivine-corrected and normalized to 100%) in glass inclusions vs: (a) H2O by FTIR; (b) CO2 by SIMS, (c) SO3Total by EMP; (d) Cl by EMP, and (e) F by SIMS. Error bars for SIMS data show ± 1 standard deviation for the average of five counts collected for each measurement (errors smaller than the symbol size are not shown).

 
These trends in volatile concentrations are consistent with entrapment of the inclusions with >52 wt % SiO₂ at late stages of magma ascent, when crystallization and degassing were considerably advanced. All but one of the 19 glass inclusions with unusually large vapor bubbles (Table 2 and Electronic Appendix 1: >8 vol. %) and all of the orange-colored glass inclusions with multiple vapor bubbles from Mascota minette cone Novillero have >52 wt % SiO₂ and associated low H, C, and S contents. We speculate that the unusual Novillero glass inclusions (Fig. 2b) may reflect entrapment and rapid quenching of vesiculated melt. Glass inclusions trapped in this important domain of upper crustal magma ascent, degassing, and crystallization have been discussed by many others, including Blundy & Cashman (2005), Blundy et al. (2006), and Spilliaert et al. (2006). As our focus in this paper is on deeper source-region systematics, most of the following discussion concerns glass inclusions with <52% SiO₂.

	DISCUSSION

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
Magmatic temperature and fO₂ estimates
Sugawara (2000) evaluated a large set of experimental olivine-saturated silicate melt compositions and devised algorithms for estimation of temperature in anhydrous and water-bearing melts. For the glass inclusions in this study, the corrections for post-entrapment crystallization of host olivine lead to T increases of 17°C for every 2 wt % olivine added. The average correction for inclusions in this study is 3% olivine added. Because we were unable to correct for potential diffusive loss of Fe from glass to olivine (Danyushevsky et al., 2000), it is possible that the corrected MgO values are still too low, resulting in underestimates of temperature. However, the water content of the melt has a greater effect on the calculation, with temperature decreasing by 35°C for every 2 wt % H₂O. We used glass water contents derived by SIMS (Table 8) when available and FTIR (Table 7 and Electronic Appendix 3) for the calculation. Results for the hydrous T estimates are given in Table 5 (and Electronic Appendix 2), and the hydrous T estimates are plotted versus glass K₂O content in Fig. 11a. Most of the inclusions fall within the estimated temperature range of 1100–1180°C. Inclusions from Colima basalt cone Tezontle form the only tight cluster of data points (Fig. 11a), yielding the highest average calculated temperature (1165 ± 6°C: n = 13), largely because of consistently low water contents. Inclusions from the other cones yield wider ranges of estimated temperature, largely because of variable water contents. The greatest temperature range is exhibited by Colima minette cone La Erita (108°C, n = 21). It is not clear whether this temperature range is accurate or perhaps an artifact of the algorithm at low water contents (Vigouroux et al., 2008), but it surely reflects degassing on ascent. Calculated temperatures for the other samples show smaller total ranges starting below 1180°C and reaching as low as 1059°C.

View larger version (11K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 11. (a) Glass K2O wt % vs Thydrous (after Sugawara, 2000) calculated for the olivine-corrected glass compositions using melt water contents determined by SIMS and FTIR. (b) Glass-inclusion K2O vs %S6+ for 52 glass inclusions with typical 1 errors.

 
One approach that has been used to estimate magmatic oxygen fugacity for whole-rock samples is based on Fe³⁺/Fe²⁺ and major-element composition (Kress & Carmichael, 1991). Results of this calculation for the 11 scoria-fall samples of this study are listed in Table 4, expressed as NNO (= log fO₂ in sample – log fO₂ for the NNO oxygen buffer of Huebner & Sato, 1970), and range from +1·3 for the Colima basanite from Telcampana to +4·7 for Mascota minette from Novillero. Similarly high values (+1 to +6) have been reported by the same technique for other western Mexican lamprophyres from Los Volcanes and Mascota by Wallace & Carmichael (1992) and Carmichael et al. (1996), marking the extreme high end of the known fO₂ ranges for terrestrial magmas (Carmichael, 1991).

An alternative approach to estimation of oxygen fugacity is based on EMP measurements of %S⁶⁺ (Table 5 and Electronic Appendix 2) reflected in SK peak-position shifts compared with end-member sulfate (S⁶⁺: anhydrite) and sulfide (S^2–: troilite). Glass-inclusion variations of K₂O vs %S⁶⁺ are shown in Fig. 11b, which includes all data from Electronic Appendix 2. Although some %S⁶⁺ values have large 1 repeatability errors (Table 5 and Electronic Appendix 2), the patterns on this plot are virtually unchanged if only the best analyses are included (%S⁶⁺ 1 values <10%). The low-K₂O group includes the calc-alkaline basalt and basaltic andesite. For the Colima basalt, 13 glass inclusions have %S⁶⁺ values of 2–40%, with a single high outlier at 69%. That outlier falls in the lower part of the range for Mascota basaltic andesite glass inclusions (67–87%, n = 8), which in turn span most of the range of variation found in the high-K₂O group. For glass inclusions from basanites, hornblende-lamprophyres, and minettes, comprising the high-K₂O group, the total range in %S⁶⁺ is 34–90%. These results indicate that with the exception of the more-reduced Colima basalt, all other Colima and Mascota magmas, calc-alkaline and lamprophyric alike, shared the same general range of oxygen fugacity.

Three proposed relationships between %S⁶⁺ values and oxygen fugacity are shown in Fig. 12. All three studies define a similar S-shaped relationship, showing that a huge shift in %S⁶⁺ occurs over a narrow increase in oxygen fugacity at NNO = 0 to + 1·5. The polynomial fit by Matthews et al. (1999) is based on the largest set of data, particularly for the high %S⁶⁺ values of interest in this study. Accordingly, that equation was used to calculate oxygen fugacity for each glass inclusion (Table 5 and Electronic Appendix 2), expressed as NNO. The flattening of the Matthews et al. (1999) curve in Fig. 12 at %S⁶⁺ >80 indicates that the errors on conversion to NNO values increase significantly. The highest %S⁶⁺ value measured is 90, for glass inclusion MAS02-2B-D.a from Mascota hornblende-lamprophyre cone Molcajete Galope, which equates to NNO = 2·7. The lowest %S⁶⁺ values of 2–40% are for 13 inclusions from the Colima basalt, yielding NNO = –2 to 0·7 (average = 0·1 ± 0·8). Glass inclusions from all other cones (n = 93), with %S⁶⁺ values of 34–90%, yield NNO = 0·5 to 2·7 (average = 1·1 ± 0·4).

View larger version (11K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 12. Three different formulations of the relationship between oxygen fugacity (expressed as NNO) and %S6+: Carroll & Rutherford (1988), Wallace & Carmichael (1994), and Matthews et al. (1999).

 
Although both methods of fO₂ estimation indicate high values for the Colima and Mascota lamprophyres (NNO plus several log units), correlation between the two techniques is poor. For the whole-rock approach, an issue of concern is the possibility that the samples have undergone post-eruptive oxidation related to weathering. As one evaluation of that possibility we plotted loss on ignition values (LOI) versus calculated NNO for the 11 whole-rock samples from this study (Table 4) and 34 other lamprophyres from Colima and Mascota reported by Luhr & Carmichael (1981) and Carmichael et al. (1996), respectively. The plot shows a poor positive correlation R² = 0·47, with many samples diverging markedly from correlation [e.g. SAY-5A has among the highest NNO values (+5·3) but the lowest LOI in the set (0·29 wt %)]. Based on this evaluation, post-eruptive oxidation is not an adequate explanation for the unusually high values found in the Colima and Mascota lamprophyres overall. Nevertheless, it is probable that some form of oxidation (post- or even syn-eruptive) may have raised the calculated NNO in some individual samples. Likewise, several factors might have modified glass-inclusion %S⁶⁺ values since quenching, including loss of hydrogen from the melt by diffusion through the host olivine (Qin et al., 1992; Metrich & Clocchiatti, 1996). Recent estimates of oxygen fugacity made by Vigouroux et al. (2008) for samples from some of these same cones are slightly lower (NNO + 0·9 to NNO + 1·3 for the alkaline magmas) but consistent with our estimates from glass-inclusion %S⁶⁺.

Pressures of glass-inclusion entrapment and degassing paths during ascent
Gas-saturation pressures (Table 8) calculated from H₂O and CO₂ values in glass inclusions, after Newman & Lowenstern (2002), are shown in Fig. 13 along with gas-saturation isobars and representative degassing paths. The four lowest calculated entrapment pressures (<260 bars, 1 km depth) are for inclusions with >52% SiO₂. The highest calculated pressure is 6600 bars or 24 km depth (pressures above 5000 bar calculated by VolatileCalc may not be highly accurate) for inclusion VF99-08L-C from Colima minette cone La Erita. It has high H₂O (7 wt %) and CO₂ (1460 ppm) and can be considered as a parental volatile mixture to most other glass inclusions only under closed-system degassing with unrealistically high weight per cent of exsolved vapor (25 wt % in Fig. 13). Failure of this degassing model suggests either that even the most pristine inclusions in this study have undergone degassing or that degassing was more complicated than modeled by VolatileCalc. Nevertheless, closed-system degassing with high weight per cent of exsolved vapor is consistent with the curious glass inclusions with multiple bubbles from Mascota minette cone Novillero (Fig. 2b). A distinct parental volatile mixture at considerably lower H₂O (2 wt %) but similar CO₂ (1740 ppm) is represented by inclusion VF99-07A-J.a from Colima minette cone San Isidro, which yields a pressure of 3725 bars (13·6 km). The markedly different water contents of these inclusions from Colima minettes again suggests that they have undergone degassing.

View larger version (21K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 13. H2O and CO2 (by SIMS, Table 8) values in 22 glass inclusions along with saturation isobars at 500–5000 bars pressure calculated at T = 1100°C using the VolatileCalc routine (Newman & Lowenstern, 2002). Degassing paths for three different conditions are shown: open, and closed with 1 wt % and 25 wt % exsolved vapor. A fictive La Erita parent at 5·8 wt % H2O and 950 ppm CO2 was used as a starting composition instead of the values for glass inclusion VF99-08L-C because its calculated pressure (6·6 kbar) exceeds the 5 kbar pressure limits of the VolatileCalc model.

 
The signatures of phlogopite, garnet, and apatite in mantle source regions
K2O in phlogopite-saturated melts
Once phlogopite in a mantle source rock begins to melt incongruently (Yoder & Kushiro, 1969; Wyllie & Sekine, 1982), the resultant liquid will be effectively buffered at K₂O levels dictated by liquid/crystal partitioning relationships. Therefore, as melting proceeds, with residual phlogopite in the source, we expect K₂O to be essentially unchanged, whereas concentration of a highly incompatible element such as La declines (Sun & Hanson, 1975); melt K₂O should begin to fall once phlogopite is exhausted. Sets of melting experiments further demonstrate the above relationship: melts coexisting with residual phlogopite in the experimental charges of Righter & Carmichael (1996) were effectively buffered in K₂O until the phlogopite was exhausted; continued melting resulted in declines in melt K₂O.

Whole-rock data for La vs K₂O are shown in Fig. 14a for the 11 samples from this study, together with a larger set of samples from the same two volcanic fields and two other western Mexican volcanic fields where lamprophyres and calc-alkaline volcanic rocks were also coeval. Below 50 ppm La, the data form a steep linear array that projects to the origin, a trend that is consistent with progressive partial melting in the absence of phlogopite. To the right of this array and above 3 wt % K₂O, many of the alkaline rocks from these four volcanic fields drift to higher La values, which is the predicted signature of residual phlogopite in the source; with progressive partial melting in the presence of residual phlogopite, La values should decline at roughly constant K₂O.

View larger version (19K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 14. La ppm vs K2O wt %. (a) The 11 Colima and Mascota whole-rock samples of this study (Table 4) plus 70 other whole-rock analyses for calc-alkaline and lamprophyric volcanic rocks from Colima (Luhr & Carmichael, 1981; Calanchi et al., 1995), Mascota (Lange & Carmichael, 1990; Carmichael et al., 1996), San Sebastián (Lange & Carmichael, 1991), and Los Volcanes (Wallace & Carmichael, 1989, 1992). (b) The 14 glass inclusions from this study analyzed by LA-ICP-MS (Table 9).

 
The set of glass-inclusion data (Fig. 14b) is small and does not exhibit kinks or inflection points to mark the shift from melting with residual phlogopite to melting without phlogopite (Feldstein & Lange, 1999). Nevertheless, the compositions of glass inclusions from La Erita (Colima minette) and Molcajete Galope (Mascota hornblende lamprophyre) with high La values are consistent with progressive partial melting in the presence of phlogopite. The calc-alkaline rocks from both fields plot close to the origin.

Phlogopite and apatite as hosts for H, Cl, and F
Most natural phlogopites are rich in both K₂O (10 wt %) and H₂O (4 wt % for hydroxy phlogopite). In K-rich volcanic rocks such as lamprophyres, lamproites, and kamafugites, phlogopites are also commonly rich in F, which competes for the same site as OH^–, averaging 1·7, 2·2, and 3·5 wt %, respectively (Edgar et al., 1994). Phlogopites in mantle xenoliths have considerably lower F contents (0·3–0·8 wt %: Aoki & Kanisawa, 1979; Edgar et al., 1994). Thus, phlogopite has long been considered a major mantle host for not only K and H during mantle melting, but also F (Aoki & Kanisawa, 1979; Edgar et al., 1996). On diagrams of K₂O vs F, Aoki & Kanisawa and Edgar et al. have shown roughly linear arrays of volcanic whole-rock compositions between the origin and points representing xenolithic and volcanic phlogopites. For the glass inclusions from this study (Fig. 15a), those from calc-alkaline Colima and Mascota samples lie close to the origin, whereas those from basanite and lamprophyre samples fan out at upper values along a regression line that predicts 1·25 wt % F in pure phlogopite at 10 wt % K₂O; this lies between the F values mentioned above for phlogopites from lamprophyres (1·7 wt % F) and mantle xenoliths (0·3–0·8 wt % F).

View larger version (9K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 15. F correlations for glass inclusions. (a) K2O wt % (Table 5 and Electronic Appendix 2: olivine-corrected and normalized) vs F ppm by SIMS (Table 8). (b) P2O5 wt % (Table 5) vs F by SIMS.

 
Apatite can also host significant F, at roughly the same abundance levels as coexisting phlogopite (Aoki & Kanisawa, 1979; Edgar et al., 1994). Although Edgar et al. (1996) downplayed the importance of apatite as an important F host during mantle melting because the correlation between P₂O₅ and F was ‘not as pronounced’ as that between K₂O and F, such is not the case with the glass inclusions analyzed in this study (Fig. 15b). If the glass inclusions with >52 wt % SiO₂ are excluded (four of 21) the correlation coefficient for P₂O₅ vs F (R² = 0·89) is only slightly lower than for K₂O vs F (R² = 0·92). On this basis, phlogopite and apatite both could have been major contributors to F in melts derived from the mantle source. Apatite probably had a shorter (if any) melting interval as a residual mineral. The experimental work by Watson (1980) indicated that at typical P–T conditions, mantle-derived basaltic magmas can dissolve 3–4 wt % P₂O₅, considerably higher than the highest P₂O₅ value found in this study (1·4 wt % for whole-rock MAS02-1A, -6A, and -6B). As Watson (1980) stated, this ‘virtually precludes the occurrence of residual apatite in mantle source regions’.

Residual garnet and HREE
The HREE-depleted signature of partial melting involving residual garnet was established 40 years ago by Gast (1968), who recognized that garnet/melt partition coefficients are significantly higher for the HREE than for the MREE. Lower concentrations of MREE to HREE in most of the alkaline Colima samples relative to the Colima calc-alkaline basalt (Table 4) are consistent with a greater role for residual garnet in formation of the HREE-depleted alkaline samples (Luhr, 1997). Using the larger Colima and Mascota whole-rock datasets discussed for Fig. 14 (Colima and Mascota data from this study as well as from Luhr & Carmichael, 1981; Calanchi et al., 1995; Lange & Carmichael, 1990; Carmichael et al., 1996), enrichment in the LREE La vs the HREE Yb (Fig. 16a) is consistent with melting within the garnet stability field for the Colima and Mascota alkaline rocks. In the Colima alkaline suite (Fig. 16b), as La falls from 80 ppm to 30 ppm (again we infer that decreasing La abundances track with increasing degree of source partial melting), Yb values remain below 1·6 ppm. For La <30 ppm, however, Yb rises to values as high as 2·1 ppm in all basanites from the youngest cone Apaxtepec and in the calc-alkaline basalt from Tezontle. This transition to higher Yb may mark the point in the melting regime when garnet began to contribute significantly to the melt, thus flooding it with HREE. No such trends are apparent for the Mascota samples, which show relatively constant Yb values higher than those for the Colima suite for La >30 ppm, and a slight decline in Yb for La <30 ppm.

View larger version (16K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 16. REE data for the larger Colima and Mascota whole-rock datasets plotted in Fig. 14: (a) La vs La/Yb, (b) La vs Yb.

 
Lamprophyre–basanite–basalt geochemical trends and vein–wallrock melting models
The formation of phlogopite in the mantle wedge above subduction zones through interactions between rising fluids or melts and the ambient peridotite has been explored experimentally by Wyllie & Sekine (1982) and Sekine & Wyllie (1982a, 1982b, 1983). The hybridization products from these experiments include orthopyroxene, clinopyroxene, garnet, and phlogopite, assemblages corresponding to garnet–phlogopite pyroxenites or websterites. Phlogopite-bearing pyroxenite assemblages have also been studied in veined ultramafic xenoliths (Lloyd, 1987; Wilshire et al., 1988; Grégoire et al., 2001; Franz et al., 2002; Downes et al., 2004; Bell et al., 2005); the most recent of these studies attributed the vein assemblages to formation by metasomatic reactions involving subduction-related fluids. Foley (1992) explored the behavior of peridotite veined by phlogopite pyroxenites during subsequent partial melting, and concluded that the veins would melt first, with more-advanced partial melting spreading to the peridotite wall-rock. He concluded that potassic alkaline magmas are a hybrid of vein and wall-rock components.

One goal of this study is to understand the partial-melting relationships in the mantle that gave rise to the mafic lamprophyre–basanite–basalt volcanic rocks in the Colima and Mascota areas. Our data are consistent with a scenario in which nearly all K₂O (in whole-rock or glass-inclusion analyses: Tables 4 and 5) was contributed by melted phlogopite, and nearly all P₂O₅ was contributed by melted apatite. Both minerals presumably reside in veins in the upper-mantle source region above the subducted Rivera plate, the top of which lies 100 km beneath the Colima and Mascota areas (Pardo & Suárez, 1993). The smallest degree partial melts are represented by the lamprophyres with the highest La contents in Fig. 14a. These vein melts formed in the presence of residual phlogopite and garnet. With progressive partial melting, exhaustion of residual phlogopite in the veins is signaled by K₂O values falling. The fading signature of residual garnet, expressed by increases in the concentrations of the HREE, probably relates to exhaustion of garnet in the veined source, but could mark a shallowing of the focus of mantle melting into the spinel stability field (Luhr, 1997). The steep linear array of La–K₂O data (Fig. 14a), which extends from lamprophyres, through basanites, and to calc-alkaline basalts and basaltic andesites heading toward the origin, may reflect dilution of the vein component by the surrounding wall-rock peridotite. Accordingly, the geochemical signatures of the lamprophyre whole-rock and glass compositions (Fig. 6) are mirrored by the calc-alkaline samples but at considerably lower abundance levels.

Sample MAS02-1 from Mascota minette cone Novillero is the best whole-rock representative of vein-melt from this study, as it falls among the highest K₂O samples along the linear array (Fig. 14a). However, a Los Volcanes minette (MN100, Wallace & Carmichael, 1989) is the most K₂O-rich minette from the Colima, Los Volcanes, and Mascota regions and might be more representative. If we take MN100 as an example of a pure vein melt (8·55 wt % K₂O and 1·94 wt % P₂O₅) and assume that pure wall-rock peridotite melt has the composition of enriched mid-ocean ridge basalt (E-MORB; based on trace element data of Vigouroux et al., 2008, fig. 13a; average E-MORB from Sun & McDonough, 1989), a simple mixing calculation suggests that Mascota sample MAS02-1 (5·3 wt % K₂O, 1·35 wt % P₂O₅) represents approximately 65% vein melt and 35% wall-rock melt (Fig. 17a). Closest to the MORB end-member of the mixing line are the calc-alkaline mafic rocks, including Colima basalt SAY-22F (Tezontle) and Mascota basaltic andesite MAS02-3 (La Esperanza). The K₂O (0·69 wt %) and P₂O₅ (0·20 wt %) concentrations in the Colima basalt suggest that this sample represents only 5% vein melt. The same mixing calculation applied to K₂O vs Al₂O₃ (Fig. 17c; 17 wt % Al₂O₃ for the MORB end-member is based on data from the PetDB database (http://www.petdbWeb/index.jsp) for samples from the East Pacific Rise) and Ba vs Sr (Fig. 17e) yields similar estimates of vein contribution for these two samples (Novillero and Tezontle), although there is more scatter in the data. Feasibility of the vein–wallrock melting model is also illustrated by glass inclusion compositions (Fig. 17b, d and f), which fall along the mixing line at positions corresponding to per cent vein melt values that are generally consistent with the estimates based on the whole-rock compositions.

View larger version (15K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 17. Simple mixing calculation applied to whole-rock data (a) K2O vs P2O5, (c) K2O vs Al2O3, and (e) Ba vs Sr, and to glass-inclusion data (b) K2O vs P2O5, (d) K2O vs Al2O3, and (f) Ba vs Sr, illustrates the feasibility of the vein–wallrock melting model. Only glass inclusions with SiO2 < 52 wt % (corrected for olivine crystallization) are plotted in (b), (d), and (f), in an effort to screen crustal effects. The pure vein melt end-member is based on minette MN100 from Wallace & Carmichael (1989), which has 8·55% K2O and 1·94% P2O5. The peridotite wall-rock end-member is based on E-MORB average values of K2O, P2O5, Ba, and Sr from Sun & McDonough (1989); 17% Al2O3 is based on data from the PetDB database [http://www.petdbWeb/index.jsp] for samples from the East Pacific Rise. Based on this calculation, the amount of vein contribution varies from 65% in the most enriched samples (from Mascota) to 5% in the Colima and Mascota calc-alkaline samples.

 
Insight into the volatile contents of the vein melts is best gained by considering the measured concentrations in glass inclusion VF99-08L-C from Colima minette cone La Erita. Inclusion VF99-08L-C contains 6·0 wt % K₂O and 1·2 wt % P₂O₅ (corrected for olivine crystallization) and has among the highest values in the investigated glass inclusions for H₂O (7 wt %), CO₂ (1460 ppm), calculated pressure (6660 bars), calculated depth (24·1 km), SO₃^Total (2·0 wt %), Cl (0·24 wt %), and F (1 wt %). However, failure of the VolatileCalc degassing model (Fig. 13) and estimates of vein contribution (variable, but possibly as high as 75%, Fig. 17b, d and f) suggest that primary vein melts have higher H₂O and CO₂ contents.

Insight into the primary volatile contents of the mafic calc-alkaline magmas is best gained by considering those with the highest measured values. However, once again, the most volatile-rich melt inclusions in our study probably underwent degassing before they were trapped. For Colima, glass-inclusion SAY-22F-L has the highest value for H₂O in the suite, yet contains only 0·9 wt % (along with 970 ppm CO₂, calculated pressure of 2050 bars, calculated depth of 7·5 km, 0·34 wt % SO₃^Total, 600 ppm Cl, and 850 ppm F). For Mascota, glass-inclusion MAS02-03B-L has 2·8 wt % H₂O, 300 ppm CO₂, calculated pressure of 1425 bars, calculated depth of 5·2 km, 0·77 wt % SO₃^Total, 800 ppm Cl, and 720 ppm F. Despite the possibility of degassing, these calc-alkaline samples are the most representative in our study of the melts that are parental to the volumetrically dominant andesites of the Cántaro–Colima volcanic chain, and the smaller volume of andesites erupted in the Mascota area.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
The western end of the Mexican Volcanic Belt is characterized by a diverse assemblage of rocks that reflects a tectonic environment complicated by simultaneous convergence and divergence. Large central calc-alkaline volcanoes are found in close association with Quaternary scoria cones composed of alkaline basic magmas not found elsewhere in Mexico. These alkaline rocks, generally characterized by high concentrations of Ba, Sr, K, and P, form a transitional series marked by an increase in the abundance of hydrous minerals (hornblende or phlogopite) from basanite to lamprophyre. Elemental-abundance patterns that mimic those of the calc-alkaline rocks (although at much higher abundance levels) suggest a genetic relationship between these alkaline rocks and the volumetrically dominant andesites.

The lamprophyres from Colima and Mascota represent melts from an enriched mantle source, apparently allowed to erupt with little modification in this region only because of its unusual upper crustal extensional regime. The compositions of the whole-rock samples and 124 glass inclusions analyzed in this study are consistent with a model in which these lamprophyres are formed at a depth of 90–100 km in the mantle wedge from partial melting of phlogopite-pyroxenite veins. Whole-rock and glass-inclusion compositions show that phlogopite, and possibly apatite and garnet were residual during the earliest stages of the vein melting. Dilution of vein melts with partial melts from the surrounding peridotite wall rock results in basanites, a transition marked by coincident decreases in incompatible element concentrations, Mg#, and phlogopite abundance. Continued dilution by wall-rock melts results in the mafic calc-alkaline magmas that characterize the large volcanic centers. If so, these lamprophyre vein melts represent the "essence" of subduction-zone geochemistry.

High concentrations of water and other volatile elements in glass inclusions indicate a volatile-rich environment, consistent with the hydrous mineral assemblages that characterize these rocks. Estimates of magmatic fO₂ from measured whole-rock Fe ³⁺/Fe²⁺ and glass-inclusion %S⁶⁺ reveal that all of these magmas were oxidized, up to several log units above the Ni-NiO buffer. The highest concentrations of water and most other volatile elements (7% H₂O, 1,460 ppm CO₂, 2% SO₃^Total, 2,400 ppm Cl, and 1% F) were recorded for a glass inclusion (VF99-08L-C) from a Colima minette, and are consistent with a depth of entrapment of 24.1 km (calculated pressure of 6,660 bars). Simple open- and closed-system degassing models indicate that this inclusion's volatile makeup is not representative of a truly primary parental melt. Nevertheless, of the samples in our study, we consider the chemical properties of this glass inclusion to be most representative of a phlogopite-pyroxenite vein melt. A simple mixing model yields variable estimates for this sample, but suggests that it may contain as much as 75% of the vein component.

It is notable that many of the glass inclusions from minettes yielded surprisingly low water contents of 2% or less. Given the wide variations in water content and relatively poor correlation with K, P, S, Cl and other incompatible elements, diffusion of H through the olivine host may be partially responsible. However, it is reasonable to assume that glass inclusions get trapped at different stages along the ascent path, resulting in a range of compositions, with respect to volatile content. Published phase-equilibrium experiments on an olivine-bearing minette (Righter and Carmichael, 1996) imply that phlogopite is stable down to pressures of 200–400 bars, allowing the possibility that these low-water inclusions were trapped at shallow depth. The curious orange-brown glass inclusions from Novillero with multiple bubbles might represent entrapment and quenching of a low-pressure foaming melt.

The mafic calc-alkaline whole-rock and glass-inclusion compositions are consistent with formation through dilution of the lamprophyre-type vein melts with partial melts from the surrounding peridotite wall-rock. Among the calc-alkaline glass inclusions, the Mascota basaltic andesite has the highest concentrations of water and most other volatile elements with 2·8% H₂O, 296 ppm CO₂ (1425 bars pressure and 5·2 km depth of entrapment), 0·8% SO₃^Total, 870 ppm Cl, and 720 ppm F. Based on a simple mixing calculation, this sample represents a mixture of approximately 5–10% vein-melt component with 90–95% melt from peridotite wall-rock. Such mafic calc-alkaline melts are envisioned as parental to the volumetrically dominant andesites of western Mexico.

	SUPPLEMENTARY DATA

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
Supplementary data for this paper are available at Journal of Petrology online.

	ACKNOWLEDGEMENTS

 
I am indebted to Jim Luhr for introducing me to the Colima and Mascota volcanic fields and the science of glass-inclusion analysis, and for guiding my efforts throughout this project. Because of Jim's sudden death before completion of the manuscript, I am especially grateful to Sally Gibson, Sandro Conticelli, Robert Trumbull, and Paul Wallace for their thorough reviews and helpful comments. In particular, this paper has benefited tremendously from numerous conversations with Paul Wallace. Assistance with analytical techniques was offered by many members of the Smithsonian's Department of Mineral Sciences, including Tim Gooding, Marc Lipella, Amelia Logan, Tim McCoy, Tim Rose, and Ed Vicenzi. Charles Mandeville, of the American Museum of Natural History, shared advice and glass standards for FTIR. Erik Hauri of the Department of Terrestrial Magnetism, Carnegie Institution provided assistance with ion microprobe analyses. Bill McDonough of the University of Maryland provided assistance with LA-ICP-MS analyses. Carlos Navarro-Ochoa helped with fieldwork at Mascota. Lee Siebert and Paul Kimberly helped track down missing data, samples, and files following Jim's death. The opportunity to engage in this study with Jim Luhr was made possible by a Smithsonian Fellowship award. Additional support came in the form of an Excellence through Engagement Summer Research Fellowship provided by the Lilly Endowment, Inc.

	FOOTNOTES

 
Deceased

---

*Corresponding author. Telephone: (812) 461-5326. Fax: (812) 465-1052. E-mail: ahmaria{at}usi.edu

	REFERENCES

TOP ABSTRACT INTRODUCTION GEOLOGICAl SETTING SAMPLES STUDIED ANALYTICAL TECHNIQUES RESULTS DISCUSSION CONCLUSIONS SUPPLEMENTARY DATA REFERENCES

 
Allan JF. Geology of the Northern Colima and Zacoalco Grabens, southwest Mexico: late Cenozoic rifting in the Mexican Volcanic Belt. Geological Society of America Bulletin (1986) 97:473–485.[Abstract/Free Full Text]

Allan JF, Nelson SA, Luhr JF, Carmichael ISE, Wopat M, Wallace PJ. Pliocene–Recent rifting in SW Mexico and associated volcanism: An exotic terrane in the making. The Gulf and Peninsular Province of the Californias. American Association of Petroleum Geologists, Memoirs—Dauphin JP, Simoneit BAH, eds. (1991) 47:425–445.

Aoki K-I, Kanisawa S. Fluorine contents of some hydrous minerals derived from upper mantle and lower crust. Lithos (1979) 12:167–171.[CrossRef][Web of Science]

Bell DR, Grégoire M, Grove TL, Chatterjee N, Carlson RW, Buseck PR. Silica and volatile-element metasomatism of Archean mantle: a xenolith-scale example from the Kaapvaal Craton. Contributions to Mineralogy and Petrology (2005) 150:251–267.[CrossRef][Web of Science]

Bloomer SH, Stern RJ, Fisk E, Geschwind CH. Shoshonitic volcanism in the Northern Mariana arc. I. Mineralogic and major and trace element characteristics. Journal of Geophysical Research (1989) 94:4469–4496.

Blundy J, Cashman K. Rapid decompression-driven crystallization by melt inclusions from Mount St. Helens volcano. Geology (2005) 33:793–796.[Abstract/Free Full Text]

Blundy J, Cashman K, Humphreys M. Magma heating by decompression-driven crystallization beneath andesite volcanoes. Nature (2006) 443:76–80.[CrossRef][Medline]

Calanchi N, Lucchini F, Navarro-Ochoa C, Rossi PL, Sanchez-Perez J. The magmatic evolution of the modern activity of the Nevado de Colima volcano in relation to the Colima Volcanic Complex activity (Mexico). Journal of the National Volcanology Group of Italy (1995) 7:75–84.

Carmichael ISE. The redox states of basic and silicic magmas: a reflection of their source regions? Contributions to Mineralogy and Petrology (1991) 106:129–141.[CrossRef][Web of Science]

Carmichael ISE, Lange RA, Luhr JF. Quaternary minettes and associated volcanic rocks of Mascota, western Mexico: a consequence of plate extension above a subduction-modified mantle wedge. Contributions to Mineralogy and Petrology (1996) 124:302–333.[CrossRef][Web of Science]

Carmichael ISE, Frey HM, Lange RA, Hall CM. The Pleistocene cinder cones surrounding Volcán Colima, Mexico re-visited: eruption ages and volumes, oxidation states, and sulfur content. Bulletin of Volcanology (2006) 68:407–419.[CrossRef][Web of Science]

Carroll MR, Rutherford MJ. Sulfur speciation in hydrous experimental glasses of varying oxidation state: Results from measured wavelength shifts of sulfur X-rays. American Mineralogist (1988) 73:845–849.[Abstract]

Carroll MR, Webster JD. Solubilities of sulfur, noble gases, nitrogen, chlorine, and fluorine in magmas. Volatiles in Magmas. Mineralogical Society of America, Reviews in Mineralogy—Carroll MR, Holloway JR, eds. (1994) 30:231–279.

Cortés-Cortés A, Garduño-Monroy VH, Navarro-Ochoa C, Komorowski J-C, Saucedo R, Macías JL, Gavilanes JC. Carta Geológica del Complejo Volcánico de Colima; Geología del Complejo Volcánico de Colima. Universidad Nacional Autónoma de México, Instituto de Geología, Cartas Geológicas y Mineras (2005) 10.

Danyushevsky LV, Della-Pasqua FN, Sokolov S. Re-equilibration of melt inclusions trapped by magnesian olivine phenocrysts from subduction-related magmas: petrological implications. Contributions to Mineralogy and Petrology (2000) 138:68–83.[CrossRef][Web of Science]

Dixon JE, Pan V. Determination of the molar absorptivity of dissolved carbonate in basanitic glass. American Mineralogist (1995) 80:1339–1342.[Abstract]

Dixon JE, Stolper EM, Holloway JR. An experimental study of water and carbon dioxide solubilities in mid-ocean ridge basaltic liquids. Part I: Calibration and solubility models. Journal of Petrology (1995) 36:1607–1631.[Abstract/Free Full Text]

Downes H, MacDonald R, Upton BGJ, Cox KG, Bodinier JL, Mason PRD, James D, Hill PG, Hearn BC. Ultramafic xenoliths from the Bearpaw Mountains, Montana, USA: Evidence for multiple metasomatic events in the lithospheric mantle beneath the Wyoming craton. Journal of Petrology (2004) 45:1631–1662.[Abstract/Free Full Text]

Edgar AD, Lloyd FE, Vukadinovic D. The role of fluorine in the evolution of ultrapotassic magmas. Mineralogy and Petrology (1994) 51:173–193.[CrossRef][Web of Science]

Edgar AD, Pizzolato LA, Sheen J. Fluorine in igneous rocks and minerals with emphasis on ultrapotassic mafic and ultramafic magmas and their mantle source regions. Mineralogical Magazine (1996) 60:243–257.[Abstract]

Ellam RM, Menzies M, Hawkesworth CJ, Leeman WP, Rosi M, Serri G. The transition from calc-alkaline to potassic orogenic magmatism in the Aeolian Islands, Southern Italy. Bulletin of Volcanology (1988) 50:386–398.[CrossRef][Web of Science]

Feldstein SN, Lange RA. Pliocene potassic magmas from the Kings River region, Sierra Nevada, California: evidence for melting of a subduction-modified mantle. Journal of Petrology (1999) 40(8):1301–1320.[CrossRef][Web of Science]

Foley S. Vein-plus-wall-rock melting mechanisms in the lithosphere and the origin of potassic alkaline magmas. Lithos (1992) 28:435–453.[CrossRef][Web of Science]

Franz L, Becker KP, Kramer W, Herzig PM. Metasomatic mantle xenoliths from the Bismarck microplate (Papua New Guinea)—Thermal evolution, geochemistry and extent of slab-induced metasomatism. Journal of Petrology (2002) 43:315–343.[Abstract/Free Full Text]

Gast PW. Trace element fractionation and the origin of tholeiitic and alkaline magma types. Geochimica et Cosmochimica Acta (1968) 32:1057–1086.[CrossRef][Web of Science]

Grégoire M, McInnes BIA, O’Reilly SY. Hydrous metasomatism of oceanic sub-arc mantle, Lihir, Papua New Guinea—Part 2. Trace element characteristics of slab-derived fluids. Lithos (2001) 59:91–108.[CrossRef][Web of Science]

Hauri E, Wang J, Dixon JE, King PL, Mandeville C, Newman S. SIMS analysis of volatiles in silicate glasses 1. Calibration, matrix effects, and comparisons with FTIR. Chemical Geology (2002) 183:99–114.[CrossRef][Web of Science]

Hooper DM. Computer-simulation models of scoria cone degradation in the Colima and Michoacán–Guanajuato volcanic fields, Mexico. Geofísica Internacional (1995) 24:321–340.

Huebner JS, Sato M. The oxygen fugacity–temperature relationships of manganese oxide and nickel oxide buffers. American Mineralogist (1970) 55:934–952.[Web of Science]

Jarosewich E, Nelen JA, Norberg JA. Reference samples for electron microprobe analysis. Geostandards Newsletter (1980) 4:43–47.

Koloskov AV, Flerov GB, Seliverstov VA, Dorendorf F, Churikova TG. Potassic volcanics of Central Kamchatka and the Late Cretaceous–Paleogene Kuril–Kamchatka alkaline province. Petrologiya (1999) 7:527–543.

Komorowski JC, Navarro C, Cortes A, Saucedo R, Gavilanes JC, Siebe C, Espindola JM, Rodriguez S. The Colima volcanic complex: Part I: Quaternary multiple debris-avalanche deposits. Part II: Historical pyroclastic sequences (1913, 1991, 1994). In: IAVCEI General Assembly, Puerto Vallarta, Mexico, January 19–24, 1997, Fieldtrip Guidebook, Excursion No. 3 (1997).

Kress VC, Carmichael ISE. The compressibility of silicate liquids containing Fe₂O₃ and the effect of composition, temperature, oxygen fugacity, and pressure on their redox states. Contributions to Mineralogy and Petrology (1991) 108:82–92.[CrossRef][Web of Science]

Lange RA, Carmichael ISE. Hydrous basaltic andesites associated with minette and related lavas in western Mexico. Journal of Petrology (1990) 31:1225–1259.[Abstract/Free Full Text]

Lange RA, Carmichael ISE. A potassic volcanic front in western Mexico: the lamprophyric and related lavas of San Sebastian. Geological Society of America Bulletin (1991) 103:928–940.[Abstract/Free Full Text]

Le Bas MJ, Le Maitre RW, Streckeisen A, Zanettin B. A chemical classification of volcanic rocks based on the total alkali–silica diagram. Journal of Petrology (1986) 27:745–750.[Abstract/Free Full Text]

Lloyd FE. Characterization of mantle metasomatic fluids in spinel lherzolites and alkali clinopyroxenites from the West Eifel and South West Uganda. (1987) In: Menzies, M. A. & Hawkesworth, C. J. (eds) Mantle Metasomatism. London: Academic Press, pp. 91–123.

Luhr JF. Extensional tectonics and the diverse primitive volcanic rocks in the western Mexican Volcanic Belt. Canadian Mineralogist (1997) 35:473–500.[Web of Science]

Luhr JF. Glass inclusions and melt volatile contents at Parícutin Volcano, México. Contributions to Mineralogy and Petrology (2001) 142:261–283.[Web of Science]

Luhr JF. Petrology and geochemistry of the 1991 and 1998–1999 lava flows from Volcán Colima, México. Journal of Volcanology and Geothermal Research (2002) 117:169–194.[Web of Science]

Luhr JF, Carmichael ISE. The Colima Volcanic Complex, México: Part I. Post-caldera andesites from Volcán Colima. Contributions to Mineralogy and Petrology (1980) 71:343–372.[CrossRef][Web of Science]

Luhr JF, Carmichael ISE. The Colima Volcanic Complex, Mexico: part II. Late-Quaternary cinder cones. Contributions to Mineralogy and Petrology (1981) 76:127–147.[CrossRef][Web of Science]

Luhr JF, Carmichael ISE. The Colima Volcanic Complex, México: Part III. Ash- and scoria-fall deposits from the upper slopes of Volcán Colima. Contributions to Mineralogy and Petrology (1982) 80:262–275.[CrossRef][Web of Science]

Luhr JF, Carmichael ISE. Petrological monitoring of cyclical eruptive activity at Volcán Colima, Mexico. Journal of Volcanology and Geothermal Research (1990) 42:235–260.[CrossRef][Web of Science]

Luhr JF, Haldar D. Barren Island Volcano (NE Indian Ocean): Island-arc high-alumina basalts produced by troctolite contamination. Journal of Volcanology and Geothermal Research (2006) 149:177–212.[CrossRef][Web of Science]

Luhr JF, Prestegaard KL. Caldera formation at Volcán Colima, México, by a large Holocene volcanic debris avalanche. Journal of Volcanology and Geothermal Research (1988) 35:335–348.[CrossRef][Web of Science]

Luhr JF, Nelson SA, Allan JF, Carmichael ISE. Active rifting in southwestern Mexico: Manifestations of an incipient spreading-ridge jump. Geology (1985) 13(1):54–57.[Abstract/Free Full Text]

Luhr JF, Allan JF, Carmichael ISE, Nelson SA, Hasenaka T. Primitive calc-alkaline and alkaline volcanic rock types from the western Mexican Volcanic Belt. Journal of Geophysical Research (1989) 94(B4):4515–4530.

Luhr JF, Kimberly P, Siebert L, Aranda-Gómez J, Housh TB, Kysar Mattieti G. México's Quaternary volcanic rocks: Insights from the MEXPET petrological and geochemical database. (2006) In: Siebe, C. Macías, J. L. & Aguirre-Díaz, G. J. (eds) Neogene–Quaternary Continental Margin Volcanism: A Perspective from Mexico. Geological Society of America, Special Papers 401, 1–44, doi:10.1130/2006.2402(01).

Matthews SJ, Moncrieff DHS, Carroll MR. Empirical calibration of the sulfur valence oxygen barometer from natural and experimental glasses: method and applications. Mineralogical Magazine (1999) 63:421–431.[Abstract]

Metrich N, Clocchiatti R. Sulfur abundance and speciation in oxidized alkaline melts. Geochimica et Cosmochimica Acta (1996) 60:4151–4160.[CrossRef][Web of Science]

Michael PJ, McDonough WF, Nielsen RL, Cornell WC. Depleted melt inclusions in MORB plagioclase: messages from the mantle or mirages from the magma chamber? Chemical Geology (2002) 183:43–61.[CrossRef][Web of Science]

Nakamoto K. Infrared and Raman Spectra of Inorganic and Coordination Compounds (1978) 448 pp: 3rd edn. New York: Wiley.

Newman S, Lowenstern JB. VolatileCalc: a silicate melt–H₂O–CO₂ solution model written in Visual Basic for excel. Computers and Geosciences (2002) 28:597–604.[CrossRef]

Nicholls IA, Whitford DJ. Potassium-rich volcanic rocks of the Muriah Complex, Java, Indonesia—products of multiple magma sources. Journal of Volcanology and Geothermal Research (1983) 18:337–359.[CrossRef][Web of Science]

Pardo M, Suárez G. Steep subduction geometry of the Rivera Plate beneath the Jalisco Block in western Mexico. Geophysical Research Letters (1993) 20–21:2391–2394.

Peck LC. Systematic analysis of silicates. US Geological Survey Bulletin (1964) 1170:89.

Qin Z, Lu F, Anderson A. T. Jr. Diffusive reequilibration of melt and fluid inclusions. American Mineralogist (1992) 77:565–576.[Abstract]

Righter K, Carmichael ISE. Phase equilibria of phlogopite lamprophyres from western Mexico: biotite–liquid equilibria and P–T estimates for biotite-bearing igneous rocks. Contributions to Mineralogy and Petrology (1996) 123:1–21.[CrossRef][Web of Science]

Robin C, Camus G, Cantagrel JM, Gourgaud A, Mossand P, Vincent PM, Aubert M, Dorel J, Murray JB. Les Volcans de Colima (Mexique); programme interdisciplinaire de recherche sur la prevision et la surveillance des eruptions volcaniques: France. PIRSEV Bulletin (1984) 87:1–98.

Robin C, Mossand P, Camus G, Cantagrel J, Gourgaud A, Vincent PM. Eruptive history of the Colima Volcanic Complex (Mexico). Journal of Volcanology and Geothermal Research (1987) 31:99–113.[CrossRef][Web of Science]

Robin C, Komorowski J-C, Boudal C, Mossand P. Mixed-magma pyroclastic-surge deposits associated with debris-avalanche deposits at Colima volcanoes, Mexico. Bulletin of Volcanology (1990) 52:391–403.[CrossRef][Web of Science]

Sekine T, Wyllie PJ. Phase relationships in the system KAlSiO₄–Mg₂SiO₄–SiO₂–H₂O as a model for hybridization between hydrous siliceous melts and peridotite. Contributions to Mineralogy and Petrology (1982a) 79:368–374.[CrossRef][Web of Science]

Sekine T, Wyllie PJ. The system granite–peridotite–H₂O at 30 kbar, with applications to hybridization in subduction zone magmatism. Contributions to Mineralogy and Petrology (1982b) 81:190–202.[CrossRef][Web of Science]

Sekine T, Wyllie PJ. Experimental simulation of mantle hybridization in subduction zones. Journal of Geology (1983) 91:511–528.[Web of Science]

Spilliaert N, Allard P, Métrich N, Sobolev AV. Melt inclusion record of the conditions of ascent, degassing, and extrusion of volatile-rich alkali basalt during the powerful 2002 flank eruption of Mount Etna (Italy). In: Journal of Geophysical Research (2006) 111. B04203, 19 pp.

Stolper E. Water in silicate glasses: An infrared spectroscopic study. Contributions to Mineralogy and Petrology (1982) 81:1–17.[CrossRef][Web of Science]

Sugawara T. Empirical relationships between temperature, pressure, and MgO content in olivine and pyroxene saturated liquid. Journal of Geophysical Research (2000) 105:8457–8472.[CrossRef]

Sun SS, Hanson GN. Origin of Ross Island basanitoids and limitations upon the heterogeneity of mantle sources for alkali basalts and nephelinites. Contributions to Mineralogy and Petrology (1975) 52:77–106.[CrossRef][Web of Science]

Sun SS, McDonough WF. Chemical and isotopic systematics of oceanic basalts; implications for mantle composition and processes. (1989) In: Saunders, A. D. & Norry, M. J. (eds) Magmatism in the Ocean Basins. Geological Society, London, Special Publications 42, 313–345.

Ulmer P. The dependence of Fe²⁺–Mg cation-partitioning between olivine and basaltic liquid on pressure, temperature, and composition: An experimental study to 30 kbars. Contributions to Mineralogy and Petrology (1989) 101:261–273.[CrossRef][Web of Science]

Vigouroux N, Wallace P, Kent A. Volatiles in high-K magmas from the western Trans-Mexican Volcanic Belt: evidence for fluid-fluxing and extreme enrichment of the mantle wedge by subduction processes. Journal of Petrology (2008).

Wallace P, Carmichael ISE. Minette lavas and associated leucitites from the western front of the Mexican Volcanic Belt: petrology, chemistry, and origin. Contributions to Mineralogy and Petrology (1989) 103:470–492.[CrossRef][Web of Science]

Wallace P, Carmichael ISE. Alkaline and calc-alkaline lavas near Los Volcanes, Jalisco, Mexico: Geochemical diversity and its significance in volcanic arcs. Contributions to Mineralogy and Petrology (1992) 111:423–439.[CrossRef][Web of Science]

Wallace P, Carmichael ISE. S speciation in submarine basaltic glasses as determined by measurements of SK X-ray wavelength shifts. American Mineralogist (1994) 79:161–167.[Abstract]

Watson EB. Apatite and phosphorus in mantle source regions: An experimental study of apatite/melt equilibria at pressures to 25 kbar. Earth and Planetary Science Letters (1980) 51:322–335.[CrossRef][Web of Science]

Wilshire HG, Meyer CE, Nakata JK, Calk LC, Shervais JW, Nielson JE, Schwarzman EC. Mafic and ultramafic xenoliths from volcanic rocks of the western United States. US Geological Survey, Professional Papers (1988) 1443:179.

Wyllie PJ, Sekine T. The formation of mantle phlogopite in subduction zone hybridization. Contributions to Mineralogy and Petrology (1982) 79:375–380.[CrossRef][Web of Science]

Yoder HS Jr, Kushiro I. Melting of a hydrous phase: Phlogopite. American Journal of Science, Shairer volume (1969) 267-A:558–582.

Zobin VM, Luhr JF, Taran YA, Bretón M, Cortés A, De La Cruz-Reyna S, Domínguez T, Gavilanes JC, Galindo II, Muñíz JJ, Navarro C, Ramírez JJ, Reyes GA, Ursúa M, Velasco J, Alatorre E, Santiago H. Overview of the 1997–2000 Activity of Volcán de Colima, México. Journal of Volcanology and Geothermal Research (2002) 117:1–19.[CrossRef][Web of Science]

CiteULike    Connotea    Del.icio.us    What's this?

This Article Abstract FREE Full Text (PDF) Supplementary Data All Versions of this Article: 49/12/2123    most recent egn060v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Maria, A. H. Articles by Luhr, J. F. Search for Related Content GeoRef GeoRef Citation Social Bookmarking          What's this?

Online ISSN 1460-2415 - Print ISSN 0022-3530

Copyright © 2010 Oxford University Press

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics