| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 44 | Number 8 | Pages 1401-1411 | 2003
© Oxford University Press 2003
Crystal Size Distribution of Plagioclase and Amphibole from Soufrière Hills Volcano, Montserrat: Evidence for Dynamic Crystallization–Textural Coarsening Cycles
SCIENCES DE LA TERRE, UNIVERSITÉ DU QUÉBEC À CHICOUTIMI, CHICOUTIMI, QUE., CANADA, G7H 2B1
* Corresponding author. Telephone: 418 545 5011 ext. 5052. Fax: 418 545 5012. E-mail: mhiggins{at}uqac.ca
RECEIVED MAY 12, 2002; ACCEPTED JANUARY 10, 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONTHE SOUFRIERE HILLS VOLCANO,...CRYSTAL SIZE DISTRIBUTIONSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The crystal size distributions (CSDs) of plagioclase and amphibole were determined from andesites of the Soufrière Hills volcano, Montserrat. Plagioclase occurs as separate crystals and as chadocrysts in large amphibole oikocrysts. The chadocrysts represent an earlier stage of textural development, preserved by growth of the oikocryst. Seventeen rock and eight chadocryst plagioclase CSDs are considered together as a series of samples of textural development. All are curved, concave up, and coincident, differing only in their maximum crystal size. Three amphibole CSDs have a similar shape and behaviour, but at a different position from the plagioclase CSDs. A dynamic model is proposed for the origin of textures in these rocks. Crystallization of plagioclase started following emplacement of andesite magma at a depth of at least 5 km. A steep, straight CSD developed by nucleation and growth. This process was interrupted by the injection of mafic magma into the chamber, or convective overturn of hotter magma. The magma temperature rose until it was buffered, initially by plagioclase solution and later by crystallization. During this period textural coarsening (Ostwald ripening) of plagioclase and amphibole occurred: small crystals dissolved simultaneously with the growth of large crystals. The CSD became less steep and extended to larger crystal sizes. Early stages of this process are preserved in coarsened amphibole oikocrysts. Repetitions of this cycle generated the observed family of CSDs. Textural coarsening followed the ‘Communicating Neighbours’ model. Hence, each crystal has its own, unique growth–solution history, without appealing to mixing of magmas that crystallized in different environments.
KEY WORDS: Ostwald ripening; textural coarsening; oikocryst; CSD; texture
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONTHE SOUFRIERE HILLS VOLCANO,...CRYSTAL SIZE DISTRIBUTIONSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Magma chambers can be examined by looking at samples from lava lakes, at volcanic products or at plutons, which are the fully solidified final product. Such materials are generally studied by quantitative chemical and isotopic analysis and by qualitative petrological observations of texture. However, it is possible to look quantitatively at some aspects of rock textures. The most common technique is by measuring crystal size distributions (CSDs). This subject is still in its infancy, and hence lacks the foundation of comparative data that is available in more conventional analyses. However, it can give a fresh and complementary way of looking at processes in magma chambers (Marsh, 1988, 1998). It has already been used to document many igneous processes, such as nucleation and growth, magma mixing, polybaric crystallization and textural coarsening (Cashman & Marsh, 1988; Marsh, 1988, 1998; Armienti et al., 1994; Higgins, 1996, 1998).
A fundamental problem in igneous petrology is how magmas solidify. Traditionally, solidification has been considered just as crystallization of the liquid components of the magma as it cools until either the rock is completely solid or eruption preserves the liquid as a glass. However, magma in a chamber can be warmed by addition of hotter magma (Sparks et al., 1977; Murphy et al., 2000), or just mixing in the chamber (Couch et al., 2001), and reduction of total or partial pressure during ascent of magma can reduce or increase undercooling. If the undercooling of a magma is reduced, then crystals may dissolve. If the temperature is just kept stable close to a mineral liquidus, then small crystals of that mineral may dissolve and large crystals grow by the process of textural coarsening [also known as Ostwald ripening or textural maturation; see review by Voorhees (1992)]. Hence, solidification can also include destruction of crystals or parts of crystals. This material can go into solution in the liquid part of the magma or be directly converted to other minerals. Solidification can therefore be a complex interplay of crystallization and solution: each crystal may have its own history of solution and growth depending on its size and environment.
The final texture of a rock can clearly say something about how the rock crystallized, but this texture may be approached by many different paths. Records of intermediate textures can help define these paths, and hence how the rock solidified. Higgins (1998) proposed that different parts of large (30 cm) olivine oikocrysts in a troctolite preserve a sequence of textural development in that rock. Oikocrysts in volcanic rocks provide this same opportunity, but their smaller size means that just one earlier texture will be preserved in a single oikocryst.
Here, we will examine the solidification of magma in a long-lived crystal-rich andesitic system, that of the Soufrière Hills volcano, Montserrat. We want to determine if solidification was dominated by progressive cooling and crystal growth (e.g. Cashman & Marsh, 1988) or if the textures were moulded by repeated cycles of heating and cooling (Devine et al., 1998a; Murphy et al., 2000).
|
THE SOUFRIÈRE HILLS VOLCANO, MONTSERRAT |
|---|
|
TOPABSTRACTINTRODUCTIONTHE SOUFRIERE HILLS VOLCANO,...CRYSTAL SIZE DISTRIBUTIONSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Volcanic activity started on Montserrat about 4·5 Myr ago. Five centres, all broadly andesitic, have been active since that time. The Castle Peak dome was emplaced 350 years ago, but the eruption appears to have been short lived and relatively minor. The current activity cycle of Soufrière Hills volcano began in 1896 with a seismic swarm that lasted a year and was followed by two further swarms at about 30 year intervals. Volcanic activity began on 18 July 1995 and despite a hiatus in 1998–1999 (Robertson et al., 2000) continues in 2002. The current eruption has proceeded by cyclical dome growth and collapse (Young et al., 1998). The pyroclastic flows that resulted have sampled both the solidified dome exterior and the magmatic interior.
Petrology and geochemistry
The current eruption of the Soufrière Hills volcano has produced a relatively uniform crystal-rich andesite (Murphy et al., 2000). It comprises 25–35% plagioclase (counting only crystals larger than 0·05 mm), 5–10% amphibole, and minor (<5%) quantities of orthopyroxene, titanomagnetite, clinopyroxene and quartz (Murphy et al., 2000). The groundmass has a similar mineralogy and comprises 5–25% glass. The only other magmatic component is represented by rare basaltic to basaltic andesite enclaves, which comprise <1% of the magma (Devine et al., 1998a; Murphy et al., 2000). These are thought to represent the remnants of a mafic magma that was injected into the chamber just before the eruption (Murphy et al., 1998, 2000).
Amphibole is noticeable in most samples because it occurs as large crystals up to 12 mm long, but actual crystal numbers are low. Many amphibole crystals have reaction rims, which may consume the whole crystal. These form during the eruption because amphibole is not stable above a depth of 5 km, and imply an ascent rate of 1–12 mm/s (Devine et al., 1998b). Many amphibole crystals contain abundant chadocrysts of plagioclase (5–15%), orthopyroxene and magnetite. The plagioclase chadocrysts are discussed below.
Plagioclase occurs as microlites and phenocrysts up to 5 mm long. Most larger plagioclase crystals are not euhedral, but are irregular in shape. Smaller crystals tend to be more euhedral. Larger crystals are complexly zoned and occasionally form polycrystalline aggregates (Figs 1 and 2). Almost all plagioclase crystals have oscillatory-zoned cores in the range An48–58 (Murphy et al., 2000). Some larger crystals lack a contrasting rim, but many are reversely zoned with rim compositions from An65 to An80. In some crystals the rim is sieve textured, with a composition of An88. Microphenocrysts and groundmass plagioclase cover about the same compositional range as that seen in the larger crystals. The shape of the crystals and the zones indicates that there have been periods of solution as well as growth (Stewart & Fowler, 2001). Many plagioclase crystals are broken, and the fragments have commonly been dispersed. This is seen in both the pyroclastic and dome samples. Zoning patterns do not indicate any significant growth following fracturing. In some parts of rocks the overall grain size has been reduced by this process, but there are no macroscopic fractures visible.
|
|
Plagioclase chadocrysts in the amphibole are smaller than the matrix plagioclase, ranging from 0·05 mm (the limit of detection) to 0·5 mm, and are not generally euhedral (Fig. 3). Crystals are generally isolated, but may form loose aggregates of a few crystals. Compositions are restricted to the sodic end of the compositional range of the matrix plagioclase, An48–77.
|
There is strong evidence that the oikocrysts preserve early textures. The matrix surrounding the oikocrysts is not richer in plagioclase than elsewhere, as might be expected if the growth of the amphibole oikocryst had pushed aside existing crystals. In addition, the zoning pattern of a plagioclase chadocryst shows that when it was engulfed by the amphibole it was not pushed aside, but just ceased to grow on the inboard side (Fig. 4). The outboard side continued to grow in competition with the amphibole. Simple geometry indicates that the growth rate of the amphibole was about three times that of the plagioclase. In some oikocrysts the long axes of the plagioclase chadocrysts have a preferential orientation parallel to the former growth surface (Fig. 3). This appears to be the only textural aspect that was developed during growth of the oikocrysts. Little energy is required to rotate crystals in the magma, and this is what appears to have occurred as the oikocryst engulfed plagioclase crystals.
|
Orthopyroxene is also an important phase with crystals similar in size to plagioclase. It occurs as chadocrysts in the amphibole as well as in the groundmass as individual crystals and in clots. Other minerals include iron–titanium oxides and quartz.
|
CRYSTAL SIZE DISTRIBUTIONS |
|---|
|
TOPABSTRACTINTRODUCTIONTHE SOUFRIERE HILLS VOLCANO,...CRYSTAL SIZE DISTRIBUTIONSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods
Samples were collected from a number of eruptions, spanning the whole eruption period of the volcano (Table 1). The widest variety of material was sampled: dense fragments of the dome, pumice and volcanic bombs. Mafic enclaves were observed, but not examined in this study.
|
Plagioclase CSDs were measured in the following way. Microphotographs were mosaiced to 30 cm by 45 cm and the outline of each crystal was traced by hand onto a digitizing tablet, after verifying the outline under a polarizing microscope. Each individual crystal in crystal clots was measured separately. The raw positional data were reduced to intersection parameters (length, width and area) using a specially written program. The porphyritic texture of the samples necessitated measurements at two different scales. The intersection results were added, with weighting for the different areas measured. Where necessary the areas were adjusted to exclude vesicles, to permit the accurate calculation of CSDs in the magma before vesiculation.
Amphibole is much less abundant than plagioclase; hence a different method was used to determine its CSD. The intersection length and width of crystals were measured directly using the graticule in the ocular. Photographs were used to verify that all crystals had been measured.
The CSD of crystals is a volumetric measure, but crystal sizes and numbers were measured in two dimensions (that is, in thin sections), hence the raw data must be inverted. The modified Saltikov approach and program of Higgins (CSDCorrections 1.2; Higgins, 2000) was used to calculate the true CSDs. Crystal aspect ratio (short:intermediate:long dimensions) and roundness factor (0 = block; 1 = ellipsoid) are needed for the conversion of intersection data (Higgins, 2000). The ratio of the short to intermediate dimensions was determined from the mode of the intersection width/length ratios (Higgins, 1994). The ratio of the long to intermediate dimensions was determined by comparison between the plagioclase volumetric proportion determined from the total area of the intersections measured and the volumetric proportion determined from the CSD (Higgins, 2002). The shapes used were 1:1·5:1·5, roundness = 0·5 for plagioclase and 1:1·5:2·5, roundness factor = 0·2 for amphibole. The crystal fabric was considered to be isotropic.
The CSD data were plotted on an ln(population density) vs size diagram following Marsh (1988). Logarithmic length intervals were used such that each bin is 100·25 times the size of the previous bin (Table 2). This bin width ensures that there are adequate numbers of crystals in the larger size bins. It also ensures that the total number of bins is low and hence that the errors introduced during the data conversion are also low. There were no gaps in the CSDs, i.e. size bins with no crystals. The left truncation (lower size limit) of the CSD, typically 0·05 mm for plagioclase, is the smallest crystal that could be measured and not the smallest crystal that exists in the rock. Large size intervals with only one or two intersections, and small size intervals with very large uncertainties, were removed from the CSD diagrams as they are imprecise. The errors in other points are small, and have been omitted for clarity from Figs 5 and 6.
|
|
|
Plagioclase
Plagioclase occurs both as chadocrysts within oikocrysts of amphibole and orthopyroxene, and in the matrix between the oikocrysts. Some amphibole oikocrysts contained up to 200 plagioclase crystal intersections, allowing the construction of a precise CSD. In contrast, the smaller orthopyroxene oikocrysts yielded far fewer plagioclase intersections, and it was not possible to measure a CSD. Plagioclase CSDs have been examined separately for the amphibole oikocrysts and the matrix.
The CSDs of plagioclase chadocrysts in amphibole were determined from a number of large amphibole oikocrysts (Fig. 5a). The left truncation (lower size limit) of these CSDs is the smallest size that could be measured (generally 0·05 mm). The total number of plagioclase chadocrysts in each oikocryst is relatively low, hence the CSD data from individual oikocrysts are less precise than those from the matrix. Where CSDs could be determined from several oikocrysts in a sample there was little variation between oikocrysts. In other samples it was necessary to amalgamate data from several oikocrysts, to give better precision. All CSDs were coincident both within a single sample and between samples. The only variation was in the maximum crystal size: it is this that accounts for the visual difference between phenocrysts.
Seventeen CSDs were measured for plagioclase in the rock matrix. Regions were selected where the abundance of broken crystals was minimal. The left truncation of the CSD is the smallest crystal that could be measured at the resolution of the observations (0·1 mm). The CSDs are strongly curved, concave up (Fig. 5b). All CSDs are coincident and there does not appear to be any progression or change in the CSD from the earliest sample (erupted in December 1995) to the latest measured here (erupted in February 2000). The only difference was in the maximum crystal size. This does not seem to correlate with eruption time or other volcanological parameters.
One CSD was determined from a sample of the Castle Peak dome, which is the predecessor of the current eruption. Its CSD was coincident within experimental error with those of samples from the current eruption, except for a slightly higher abundance of large crystals (Fig. 5b).
Amphibole
Amphibole in many samples has been altered during the eruption process, with the production of pyroxene-rich rims (Murphy et al., 2000). Amphibole CSDs could only be measured in three samples that lack this alteration (Table 1). It was not possible to measure the CSD of amphibole in rocks from the earliest part of the eruption, as defined by Harford & Sparks (2001). The left truncation of the CSD is the smallest crystal that could be measured at the resolution of the observations (0·2 mm). The maximum crystal sizes are variable, but as few large crystals were counted there is a significant uncertainty in this value. All amphibole CSDs are curved, concave up (Fig. 6). If the largest and least well-determined parts of the CSDs are ignored (points at 12 mm) then there is a fanning of the CSDs about a size of 2 mm.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONTHE SOUFRIERE HILLS VOLCANO,...CRYSTAL SIZE DISTRIBUTIONSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
CSD theory
Crystal size distribution (CSD) theory was initially developed for industrial crystallization and was applied to geological systems by Marsh (1988). Work continues in this subject and the most recent summary is that by Marsh (1998). The model of Marsh showed that in an open, steady-state system, with continuous crystallization, the CSD is a straight line on a graph of ln(population density) vs size. Marsh (1988, 1998) further showed that under certain conditions closed systems crystallizing in response to linearly increasing undercooling could also produce straight CSDs. Such CSDs were subsequently found in many volcanic systems (e.g. Cashman & Marsh, 1988).
Many rocks have CSDs that are strongly curved concave up. There are many possible processes that can produce such a CSD. Higgins (1996) proposed that curved CSDs of plagioclase in dacites from the Kameni islands, Greece, were produced by physical mixing of two magmas. Indeed, mixing of almost any two crystal populations with straight CSDs will produce such a curve. Armienti et al. (1994) proposed that curved CSDs from Etna magmas were produced by three sequential periods of cooling during the ascent and emplacement of the magma.
Another important process is textural coarsening, also known as annealing, Ostwald ripening and textural maturation [see review by Voorhees (1992)]. This is the process by which the total surface energy of a phase is minimized by solution of grains smaller than a critical size and simultaneous growth of larger grains. There are many solutions to this problem, but most are based on the Lifshitz–Slyozov–Wagner theory (Lifshitz & Slyozov, 1961; Voorhees, 1992). These were initially developed for infinitely dilute systems in which all crystals communicate instantly with each other, so that diffusion is not a limiting factor. Much work has been done to extend this model to real situations; however, although such solutions give good results for temporal development of textures, they do not yield CSDs that correspond to those of natural materials.
DeHoff (1991) proposed a completely different model, which he called ‘Communicating Neighbours’ (CN). This model assumes that diffusion does play a role and that each crystal ‘sees’ only its neighbours. That is, if a large crystal is surrounded by small crystals then during textural coarsening the small crystals will dissolve, increasing the local concentration of the mineral component and allowing the large grain to grow. If another large grain is surrounded by other large grains, then during textural coarsening there will be no solution of grains and hence no extra material to crystallize. In addition, this process is not necessarily symmetric around each crystal. Therefore, this model allows for the commonly observed local variations in the growth rate of crystals in experimental systems (termed ‘growth rate dispersion’), which also accords with the observation that in many volcanic rocks each crystal appears to have its own, complex growth history (e.g. Murphy et al., 2000). Higgins (1998) showed that the CN model successfully modelled the solidification of a plutonic rock. However, the question remains as to whether textural coarsening is also important in volcanic rocks. Higgins (1998) interpreted the CSDs of samples from the Makaopuhi lava lake to show the effects of textural coarsening, but other interpretations are also possible (Cashman & Marsh, 1988).
Plagioclase
The CSDs of plagioclase in the amphibole oikocrysts and the rock matrix can be considered to be a series of textural samples, linked by some combination of petrological processes. The earliest textures are those that have the least maximum crystal size. Later, or more developed textures, continue the CSD out to greater maximum crystal sizes, but always along a coincident path. This pattern is seen within both the chadocryst CSDs and the matrix CSDs, and in both combined (Fig. 5). The challenge of this study is to find a model that can generate such a progression.
Linearly increasing undercooling will produce a family of CSDs that all have the same slope, but increasing intercepts (Fig. 7a). This is not observed: all CSDs have similar abundances of small crystals. CN textural coarsening produces a family of CSDs that rotate around a small size, with a turndown for small sizes (Fig. 7b). Again, this is not the behaviour shown by the Montserrat CSDs. However, it is possible to model conceptually and qualitatively the observed CSDs using a combination of these two processes.
|
The model starts with the emplacement of a batch of andesite magma into the crust at a depth of at least 5 km, the stability limit of amphibole. Heat is withdrawn at a constant rate, resulting in nucleation and growth of phases, particularly plagioclase (Fig. 8). The total amount of solid phases increases approximately linearly. At the end of this cycle the CSD is straight and steep, terminating at a small size.
|
This cycle is interrupted by injection of mafic magma into the chamber (Devine et al., 1998a), or perhaps by convective mixing of magma already warmed by mafic magma injections (Fig. 8; Couch et al., 2001). As the undercooling of the andesite magma decreases, plagioclase crystals smaller than the critical size will dissolve, and larger crystals will continue to grow. The critical size starts very small, but increases as the temperature of the magma approaches the plagioclase liquidus. Hence, the size range of crystals that dissolve will grow. The temperature of the magma will rise until it is buffered near the liquidus, initially by the solution of plagioclase and later by the crystallization of plagioclase. It is during this period of buffering that most textural coarsening will occur. This process will yield a new CSD that has a shallower shape, a turndown for small sizes (low abundance of small crystals) and a greater maximum crystal size. As the system cools, undercooling will increase until textural coarsening ceases and nucleation starts. The cycle then recommences with renewed nucleation and growth. The CSD will now have a new steep, straight segment for small sizes. The CSD will extend out to greater sizes, as larger crystals have grown during textural coarsening. Repeated action of this cycle will progressively extend the CSD to larger sizes, whilst maintaining the population density of the smaller crystals. If the rock is erupted after a nucleation and growth event (as shown in Fig. 8), then there will be a high population density of small crystals. The similarity of the CSDs from the Castle Peak eruption 350 years ago to those of the current eruption suggests that the cyclic processes proposed here been active since that time.
Amphiboles
It is always difficult to explain the origin of large, but sparse crystals, such as the amphibole oikocrysts seen here. Arguments based on original sparse nucleation seem difficult to reconcile with observations of abundant nuclei and crystals in similar rocks (Higgins, 1998). Perhaps a more likely explanation is that the oikocryst phase, amphibole, originally formed as many small crystals. A period of textural coarsening reduced the crystal number density and enlarged the largest crystals by solution of the smallest crystals (Voorhees, 1992).
The dynamic crystallization–textural coarsening model developed above to account for the plagioclase CSDs would also have affected the amphibole crystals. Hence, during periods of textural coarsening amphibole would have grown fast enough to engulf the plagioclase crystals and preserve the rock texture. The rounded shape and complex zoning in some plagioclase chadocrysts indicates that there must have been cycles of growth and solution before these crystals were incorporated into the amphibole oikocrysts (Fig. 8).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONTHE SOUFRIERE HILLS VOLCANO,...CRYSTAL SIZE DISTRIBUTIONSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This is the first published study of volcanic rocks in which the CSDs of plagioclase were measured at two different stages of textural development. This is possible only in rocks that contain large oikocrysts with abundant chadocrysts, restricting application of the method to a few other volcanoes.
CSDs of plagioclase from amphibole oikocrysts and the rock matrix are all curved, concave up, and coincident. There is no lower size limit for crystals. The maximum crystal size, as expressed by the upper end of the CSD, is the only major difference between the CSDs. This type of behaviour has not been reported from other volcanoes. Amphibole CSDs are also curved, with no lower size limit. However, they are not completely coincident, but appear to fan around a small size.
The oikocryst and rock matrix plagioclase CSDs, and hence textures, cannot be related either by simple nucleation and growth of plagioclase, or by ‘Communicating Neighbours’ model textural coarsening. Both processes appear to be necessary and a conceptual model is proposed that involves repeated cycles of nucleation and growth alternating with periods of textural coarsening. Nucleation and growth occur when undercooling is increasing. Heating events, such as injection of mafic magma, or convective mixing, terminate these periods and initiate textural coarsening. Nucleation and growth resumes when release of latent heat cannot buffer the system near the plagioclase liquidus.
An important feature of this model is that Communicating Neighbours model textural coarsening leads to a unique growth and solution for each crystal, and even each face of a crystal. It does this without aid of complex mixing of magmas with very different histories.
If plagioclase developed by such cycles of nucleation and growth followed by textural coarsening then the amphibole oikocrysts must have grown in the same way. Further examination may reveal similar complex zoning patterns, as already suggested by the isotopic studies of Harford & Sparks (2001).
|
ACKNOWLEDGEMENTS |
|---|
We would like to thank the Montserrat Volcanological Observatory staff, Steve Sparks, Georg Zellmer, Chloe Harford, Frederic Villemure, Hugh Tuffen and Tim Druitt, as well as Dougal Jerram and an anonymous reader for perceptive reviews. This research was funded by a grant from NSERC (Canada) to M.D.H. It was started while M.D.H. was on sabbatical at Université Blaise-Pascal, France.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONTHE SOUFRIERE HILLS VOLCANO,...CRYSTAL SIZE DISTRIBUTIONSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Armienti, P., Pareschi, M. T., Innocenti, F. & Pompilio, M. (1994). Effects of magma storage and ascent on the kinetics of crystal growth. Contributions to Mineralogy and Petrology 115, 402–414.[CrossRef][Web of Science]
Cashman, K. V. & Marsh, B. D. (1988). Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallisation II. Makaopuhi lava lake. Contributions to Mineralogy and Petrology 99, 292–305.[CrossRef][Web of Science]
Couch, S., Sparks, R. S. J. & Carroll, M. R. (2001). Mineral disequilibrium in lavas explained by convective self-mixing in open magma chambers. Nature 411, 1037–1039.[CrossRef][Medline]
DeHoff, R. T. (1991). A geometrically general theory of diffusion controlled coarsening. Acta Metallurgica et Materialia 39, 2349–2360.[CrossRef][Web of Science]
Devine, J. D., Murphy, M. D., Rutherford, M. J., Barclay, J., Sparks, R. S. J., Carroll, M. R., Young, S. R. & Gardner, J. E. (1998a). Petrologic evidence for pre-eruptive pressure–temperature conditions, and recent reheating of andesitic magma erupting at the Soufrière Hills Volcano, Montserrat, W.I. Geophysical Research Letters 25, 3669–3672.[CrossRef][Web of Science]
Devine, J. D., Rutherford, M. J. & Gardner, J. E. (1998b). Petrologic determination of ascent rates for the 1995–1997 Soufrière Hills Volcano andesitic magma. Geophysical Research Letters 25, 3673–3676.
Harford, C. L. & Sparks, R. S. J. (2001). Recent remobilisation of shallow-level intrusions on Montserrat revealed by hydrogen isotope composition of amphiboles. Earth and Planetary Science Letters 185, 285–297.[CrossRef][Web of Science]
Higgins, M. D. (1994). Determination of crystal morphology and size from bulk measurements on thin sections: numerical modelling. American Mineralogist 79, 113–119.[Abstract]
Higgins, M. D. (1996). Magma dynamics beneath Kameni volcano, Greece, as revealed by crystal size and shape measurements. Journal of Volcanology and Geothermal Research 70, 37–48.[CrossRef][Web of Science]
Higgins, M. D. (1998). Origin of anorthosite by textural coarsening: quantitative measurements of a natural sequence of textural development. Journal of Petrology 39, 1307–1325.[CrossRef][Web of Science]
Higgins, M. D. (2000). Measurement of crystal size distributions. American Mineralogist 85, 1105–1116.
Higgins, M. D. (2002). Closure in crystal size distributions (CSD), verification of CSD calculations, and the significance of CSD fans. American Mineralogist 87, 171–175.
Lifshitz, I. M. & Slyozov, V. V. (1961). The kinetics of precipitation from supersaturated solid solutions. Journal of Physics and Chemistry of Solids 19, 35–50.[CrossRef][Web of Science]
Marsh, B. D. (1988). Crystal size distribution (CSD) in rocks and the kinetics and dynamics of crystallization I. Theory. Contributions to Mineralogy and Petrology 99, 277–291.[CrossRef][Web of Science]
Marsh, B. D. (1998). On the interpretation of crystal size distributions in magmatic systems. Journal of Petrology 39, 553–600.[CrossRef][Web of Science]
Murphy, M. D., Sparks, R. S. J., Barclay, J., Carroll, M. R., Lejeune, A. M., Brewer, T. S., Macdonald, R., Black, S. & Young, S. R. (1998). The role of magma mixing in triggering the current eruption at the Soufrière Hills Volcano, Montserrat, West Indies. Geophysical Research Letters 25, 3433–3436.[CrossRef][Web of Science]
Murphy, M. D., Sparks, R. S. J., Barclay, J., Carroll, M. R. & Brewer, T. S. (2000). Remobilization of andesitic magma by intrusion of mafic magma at the Soufrière Hills volcano, Montserrat, West Indies. Journal of Petrology 41, 21–42.
Robertson, R. E. A., Aspinall, W. P., Herd, R. A., Norton, G. E., Sparks, R. S. J. & Young, S. R. (2000). The 1995–1998 eruption of the Soufrière Hills Volcano, Montserrat, WI. Philosophical Transactions of the Royal Society of London 358, 1619–1637.[CrossRef]
Sparks, R., Sigurdsson, H. & Wilson, L. (1977). Magma mixing: a mechanism for triggering acid explosive eruptions. Nature 267, 315–318.[CrossRef]
Stewart, M. L. & Fowler, A. D. (2001). The nature and occurrence of discrete zoning in plagioclase from recently erupted andesitic volcanic rocks, Montserrat. Journal of Volcanology and Geothermal Research 106, 243–253.[CrossRef][Web of Science]
Voorhees, P. W. (1992). Ostwald ripening of two-phase mixtures. Annual Review of Materials Science 22, 197–215.[CrossRef][Web of Science]
Young, S. R., Sparks, R. S. J., Aspinall, W. P., Lynch, L. L., Miller, A. D., Robertson, R. E. A. & Shepherd, J. B. (1998). Overview of the eruption of Soufrière Hills Volcano, Montserrat, 18 July 1995 to December 1997. Geophysical Research Letters 25, 3389–3392.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Gagnevin, J. S. Daly, and G. Poli Insights into granite petrogenesis from quantitative assessment of the field distribution of enclaves, xenoliths and K-feldspar megacrysts in the Monte Capanne pluton, Italy Mineralogical Magazine, December 23, 2008; 72(4): 925 - 940. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Hammer Experimental Studies of the Kinetics and Energetics of Magma Crystallization Reviews in Mineralogy and Geochemistry, January 1, 2008; 69(1): 9 - 59. [Full Text] [PDF]
|
|
|
|
 |
|
 M. B. Holness, A. T. Anderson, V. M. Martin, J. Maclennan, E. Passmore, and K. Schwindinger Textures in Partially Solidified Crystalline Nodules: a Window into the Pore Structure of Slowly Cooled Mafic Intrusions J. Petrology, July 1, 2007; 48(7): 1243 - 1264. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. ANNEN, J. D. BLUNDY, and R. S. J. SPARKS The Genesis of Intermediate and Silicic Magmas in Deep Crustal Hot Zones J. Petrology, March 1, 2006; 47(3): 505 - 539. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. MOCK and D. A. JERRAM Crystal Size Distributions (CSD) in Three Dimensions: Insights from the 3D Reconstruction of a Highly Porphyritic Rhyolite J. Petrology, August 1, 2005; 46(8): 1525 - 1541. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. RUTHERFORD and J. D. DEVINE Magmatic Conditions and Magma Ascent as Indicated by Hornblende Phase Equilibria and Reactions in the 1995-2002 Soufriere Hills Magma J. Petrology, August 1, 2003; 44(8): 1433 - 1453. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||