Journal of Petrology

This Article Abstract FREE Full Text (PDF) 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 Preston, R. J. Articles by Rogers, G. Search for Related Content Social Bookmarking          What's this?

Journal of Petrology | Volume 39 | Number 3 | Pages 519-550 | 1998
© Oxford University Press 1998

The Loch Scridain Xenolithic Sill Complex, Isle of Mull, Scotland: Fractional Crystallization, Assimilation, Magma-Mixing and Crustal Anatexis in Subvolcanic Conduits

R. Jeremy Preston^1,*, Brian R. Bell¹ and Graeme Rogers²

¹ Department of Geology and Applied Geology, University of Glasgow Glasgow G12 8QQ, UK
² Isotope Geosciences Unit, Scottish Universities Research and Reactor Centre East Kilbride, Glasgow G75 0QF, UK

Received October 7, 1996; Revised typescript accepted October 21, 1997

	ABSTRACT

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
The petrogenesis of a suite of high-level minor intrusions which intruded the Palaeogene lava field and older rocks in the vicinity of Loch Scridain, Isle of Mull, NW Scotland, has been investigated using mineral chemistry data together with whole-rock major and trace element and radiogenic Sr–Nd–Pb isotope data. Three distinct magma groups are present. Group I consists of aphyric tholeiitic basalts and basaltic andesites, the most primitive of which have MORB-like chemical affinities. Group I rocks are markedly xenolithic, containing both crustal and cognate gabbroic xenoliths. Group II comprises plagioclase- and pyroxene-phyric andesites and dacites. Group III consists solely of fine-grained rhyolites. Compositional trends displayed by the Group I basic magmas can be explained in terms of relatively low-pressure fractionation of the assemblage olivine + clinopyroxene + plagioclase, followed by plagioclase + clinopyroxene + low-Ca pyroxene. The extreme incompatible-element enrichment, the high initial Sr isotope ratios, and radiogenic Pb isotope ratios of the more evolved members of Group I can, however, only be explained through a process of combined fractional crystallization and crustal assimilation of the underlying Moine pelitic schists. The isotope and trace element data imply that the contaminant was a high-Nd partial melt of Moine metasediments rather than a bulk addition of Moine pelite. The major, trace element and Sr–Nd isotope geochemistry of the Group III rhyolites is consistent with their derivation mainly from the partial melting of a pelitic crustal source. The geochemical characteristics of the Group II intermediate rocks suggest that they represent simple mixtures between basic magma and Group III silicic melts.

KEY WORDS: fractional crystallization; assimilation; subvolcanic; sill complex

	Introduction

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
Assimilation of continental crust by basic magmas is a widely invoked process for explaining the compositional and isotopic characteristics of associated intermediate and silicic igneous rocks. The same geochemical characteristics, however, can be produced via crystal fractionation processes, combined with mixing of coexisting basic and silicic magmas. Nevertheless, direct field and geochemical evidence for all of these processes occurring within one suite of coeval rocks is rarely seen.

Bowen (1922, 1928) provided the first rigorous studies on possible mechanisms for, and the consequences of, crustal contamination. Addressing the question in the context of crystal–liquid equilibria, Bowen showed that the limiting factors on the process of crustal assimilation were the difference between the heat capacity and heat of fusion of the materials involved, noting that magmas at or near the Earth's surface are unlikely to be substantially superheated. As a result, Bowen concluded that the heat required to raise the temperature of the country rock to that of the magma is provided by the latent heat of crystallization of the phases with which the liquid is currently saturated; the compositional trend of the liquid follows the normal crystallization sequence just as it would without assimilation, and only the proportions of the crystalline end-products differ.

Subsequently, many crustal assimilation mechanisms have been suggested, and used to explain the petrogenesis of tholeiitic, alkaline and calc-alkaline magmas from many different tectonic environments. Recent studies of continental basalt provinces have revealed the occurrence of two major contrasting styles of crustal contamination (e.g. Thompson et al., 1982; Thirlwall & Jones, 1983; Fodor et al., 1985a, 1985b; Mantovani et al., 1985; Devey & Cox, 1987):

1. Contamination accompanied by concurrent fractional crystallization within magma chambers or dyke complexes, where the heat released by crystallization allows fusion of the country rocks. This process has been termed AFC (assimilation and fractional crystallization) by DePaolo (1981), and has been widely used to interpret trace element and radiogenic isotope variations within suites of volcanic rocks. In such AFC processes, well-constrained correlations should exist between indices of fractionation (e.g. wt % SiO₂) and incompatible trace element ratios or initial Sr–Nd–Pb isotopic ratios.
   
2. In those provinces where magma chambers develop at the base of the crust, only the hottest picritic magmas will be able to assimilate the refractory wall-rocks. Therefore crystal fractionation may occur without substantial contamination. When these more evolved magmas subsequently ascend to higher levels in the crust, again only the hottest magmas will be capable of assimilating the more easily fusible upper-crustal rocks. Unlike the AFC process, this typically leads to the most primitive magmas being the most contaminated (Huppert & Sparks, 1985; Wilson, 1989; Kerr et al., 1995).
   

Many studies which call upon substantial crustal contamination to explain compositional variations in a suite of co-magmatic rocks have had to make assumptions about the nature of the contaminant, because of the lack of exposure of appropriate basement rocks, or of entrained xenoliths.

The Loch Scridain Sill Complex (LSSC) on the Isle of Mull, NW Scotland, provides an important opportunity to study the interaction of basic magmas with crustal material. The LSSC is a suite of high-level, inclined, basic to acidic sheets that are intruded into a variety of country rocks. Many of the sheets preserve large numbers of crustal xenoliths in all stages of partial fusion, as well as ultrabasic xenoliths of cumulate origin (Preston & Bell, 1997). In this paper we present details of the field relationships, mineralogy, mineral chemistry, and whole-rock major and trace element geochemistry and isotope geochemistry of the LSSC, to fingerprint the petrogenetic processes which occurred to produce the large range of compositions seen within the suite.

	Geological Setting

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
The LSSC is restricted to a relatively small area on the Ross of Mull, between Carsaig and Pennyghael in the east, and Bunessan in the west (Fig. 1). Sheets from the same suite can also be found, in less profusion, to the north of Loch Scridain around Tiroran on the peninsula of Ardmeanach (Thomas, 1922; Bailey et al., 1924). The suite is mainly intruded into an extensive plateau lava field which is of Palaeogene age. Whereas available Ar–Ar ages for the lava sequence are between 60 and 57 Ma (Mussett, 1986), recent palynological data (Bell & Jolley, 1997) suggest that the upper part of the sequence may be as young as 55 Ma. Consequently, all the Sr–Nd–Pb isotope data presented in the study have been age-corrected to 55 Ma. The plateau lavas overlie a thin sequence of Mesozoic (Triassic–Cretaceous) clastic and carbonate sediments, which in turn overlie Precambrian metasedimentary rocks of the Moine Supergroup. Sheets from the LSSC intrude all of these pre-Palaeogene lithologies. Basement rocks are likely to be Archaean–Palaeoproterozoic Lewisian amphibolite and granulite facies gneisses; although these do not crop out on Mull itself, they are found nearby on the island of Iona (Fig. 1).

View larger version (25K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 1. Map showing the geology of the Loch Scridain District, Isle of Mull, and major distribution of the xenolithic sill complex (after Bailey et al., 1924). Number of sheets depicted in suite is greatly reduced for clarity. Inset shows location of Mull in relation to the other major Tertiary igneous complexes; light grey areas are lava fields, black areas central intrusive complexes.

 
The magmatic activity which produced the LSSC clearly post-dates the extrusion of the majority of the Mull Plateau Group lavas as defined by Kerr, (1995a, 1995b), and is thought to be related to the activity of the Mull central volcano, and the production of the Centre 1 of the central igneous complex (Dagley et al., 1987). The latter is one of a number of intrusive centres in the British Tertiary Igneous Province (BTIP), which extends from Skye, south along the west coast of Scotland, into Northern Ireland (Fig. 1). A large portion of the lava field on Mull, and most of the central igneous complex, has been severely affected by hydrothermal alteration. Bailey et al., (1924) showed that on Mull the hydrothermal metamorphism is particularly extreme in a `zone of pneumatolysis' incorporating the central plutonic complex, and parts of the lava field within 5 km of the plutons (Fig. 1). Outside this area the lavas have been affected by zeolite-facies hydrothermal burial metamorphism, resulting in flat-lying metamorphic zones with distinctive zeolite mineral assemblages (Walker, 1970; Morrison, 1979).

The geochemistry and stratigraphy of the Mull lava field have been discussed by Bailey et al., (1924), Beckinsale et al., (1978), and more recently by Kerr, (1995a, 1995b) and Kerr et al., (1995). Kerr and coworkers have shown that three geochemically distinct magma types exist within the Mull lava pile. The bulk of the plateau lavas are transitional to mildly alkalic basalts and their derivatives (the so-called Mull Plateau Group; Kerr, 1995a), the more magnesian of which are often contaminated with small amounts (<5%) of Lewisian crust (Kerr, 1995b). Modelling suggests that the uncontaminated basalts are the result of between 6 and 10% partial melting of a depleted garnet-bearing mantle source over a range of pressures (Kerr, 1995a). Younger lavas (the Coire Gorm type) are slightly more tholeiitic in nature, and appear to be the products of shallower (spinel lherzolite facies), and more extensive (8–12%) melting (Kerr, 1995a). The youngest lavas (the Central Mull Tholeiites) are the result of extensive melting (12–17%) of a depleted spinel lherzolite mantle source (Kerr, 1995a). These types are similar to those seen in the Skye lava pile, with the Skye Main Lava Series (SMLS), the tholeiitic Fairy Bridge magma type, and the mid-ocean ridge basalt (MORB)-like tholeiitic Preshal More (PM) magma type corresponding to the Mull Plateau Group, Coire Gorm magma type, and the Central Mull Tholeiites, respectively (Thompson et al., 1972, 1980; Mattey et al., 1977).

The LSSC is mainly composed of fine-grained, aphyric basalts and basaltic andesites [using the classification of Le Bas et al., (1986)]. A number of the sheets (20%) are intermediate in composition; these can be distinguished from the basic rocks in the field in that they are typically glassy, and commonly contain phenocrysts of plagioclase and pyroxene. Both basic and intermediate rock types occur throughout the Loch Scridain district. A smaller number of sheets from the LSSC are composed of fine-grained rhyolite.

Members of the LSSC show great variation in thickness, from centimetres, up to 14 m; most sheets, however, vary between 0.5 m and 6 m. There were several phases of intrusion, with cross-cutting relationships between two or more sheets being relatively common. Most sheets are gently inclined (5°) and discordant to the bedding of the lavas. Unlike cone-sheets, however, theyshow no tendency to dip towards a central focal-point; dip directions and degree of dip are highly variable (5–60°).

The large majority of the sheets have well-developed chilled margins of several centimetres thickness. Some individual sheets, however, exhibit a local tendency not to chill at one or both margins, which may be attributed to turbulent flow within the conduit (Huppert & Sparks, 1980; Kille et al., 1986). Where the magma has failed to chill, contact metamorphic effects are occasionally evident, particularly in those sheets which invade the Moine metasedimentary and Mesozoic sedimentary sequences.

A small number of sheets are composite in nature. For example, that at Rudh' a' Chromain (Fig. 1) consists of a 10 m thick central portion of rhyolite with 1 m thick upper and lower margins of fine-grained tholeiitic dolerite which are markedly xenolithic. The dolerite is chilled against an earlier micro-monzonite intrusion (bostonite; Bailey et al., 1924), and Jurassic sandstones. The internal contacts between the rhyolite and the dolerite show no evidence of chilling, and instead are characterized by broad zones (up to 1 m) of hybridization. Locally, sheets of intermediate composition may have margins of holocrystalline dolerite, the boundaries generally being very sharp. Composite sheets of this nature can be found throughout the region, and attest to the coexistence of basic, intermediate and silicic magmas.

The major reasoning behind characterizing the LSSC as a co-magmatic suite is the fact that the intrusions contain a ubiquitous and relatively homogeneous suite of crustal xenoliths. These consist of quartzites, schists and refractory residues from partial melting of metasediments (mullite–buchite plagioclase, spinel and corundum xenoliths) probably derived from the Moine Complex (Thompson et al., 1986; Preston, 1996), along with sandstones and conglomerates sourced from the local Triassic-Cretaceous clastic lithologies.

	Analytical Techniques

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
All mineral analyses were carried out on a Cameca SX-50 electron probe microanalyser fitted with four wavelength spectrometers, at the Department of Geology and Applied Geology, University of Glasgow. Typical working conditions were 15 kV and 20 nA, with a beam size of between 1 and 3 µm. Standards comprised a series of pure elements and compound standards supplied by Cameca and Micro Analysis Consultants Ltd. Whole-rock major and trace element analyses were carried out using a Philips PW 1480 automatic X-ray fluorescence spectrometer in the Department of Geology and Geophysics, University of Edinburgh. Sample preparation, and accuracy and precision of the analyses have been described by Fitton & Dunlop, (1985). Whole-rock rare earth element (REE) analyses were performed at the Department of Geology, Royal Holloway University, London, by inductively coupled plasma atomic emission spectroscopy (ICP-AES). Details of the ion-exchange separation technique of the REE from whole-rock samples, together with running procedures, accuracy, and precision of the ICP-AES have been described by Walsh et al., (1981).

Rb–Sr, Sm–Nd and Pb isotopic data were obtained at the Scottish Universities Research and Reactor Centre (SURRC), East Kilbride. Details of the separation techniques for Rb–Sr and Sm–Nd have been reported by Janousek et al., (1995), and by Barbero et al., (1995). Sr, Nd and Sm, isotope analyses were performed on a VG Sector 54-30 thermal ionization mass spectrometer, whereas Rb analyses were carried out using a VG MM30 thermal ionization mass spectrometer. During the course of this study, the JM Nd standard gave ¹⁴³Nd/¹⁴⁴Nd = 0.511500 ± 10 (2 SD), and repeat analyses of NBS 987 Sr standard gave ⁸⁷Sr/⁸⁶Sr = 0.710236 ± 19 (2 SD). Pb was separated from whole-rock powders using standard HBr–HCl anion exchange techniques, and analysed on a VG 54E thermal ionization mass spectrometer. The Pb isotopic data were corrected for mass fractionation of 0.15% per atomic mass unit (a.m.u.) based on replicate analysis of NBS981. External reproducibility of Pb isotopic ratios is 0.2% (2 SD). Analytical blanks were <1 ng.

	Petrography and Mineral Chemistry

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
Subdivision of the LSSC
During the initial mapping of the island, Bailey et al. (1924) coined numerous terms to subdivide the basic, intermediate and silicic intrusions, mainly on the basis of field and thin-section characteristics. For the purpose of clarifying the terminology used in this paper it is necessary to subdivide the members of the LSSC into three major rock groups. As the basic and certain intermediate rock-types tend to grade into each other petrographically, this subdivision is based on whole-rock geochemistry rather than petrography:

1. Group I-tholeiitic basalts and basaltic andesites;
   
2. Group II-tholeiitic andesites and porphyritic dacites;
   
3. Group III-rhyolites.
   

This subdivision can be most clearly illustrated on a plot of wt % TiO₂ vs mg-number [= wt % MgO/(MgO + Fe₂O₃*)], which shows three distinct lineages (Fig. 2). In the field there is no spatial distribution of the various rock types, with sheets of all types being found throughout the area. The mineralogy and petrography of the LSSC have been described in detail by Thomas (1922) and Bailey et al. (1924), and will only be summarized here.

View larger version (10K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 2. TiO2–mg-number plot, upon which the subdivision of the LSSC is defined.

 
Group I basalts and basaltic andesites
Rocks of Group I make up the bulk of the LSSC. These are olivine-poor or olivine-free tholeiites, primarily composed of plagioclase and augite, and are generally devoid of phenocrysts (<1 vol. %). Groundmass plagioclase (1 mm) shows a typical compositional range from core to rim of An₆₆ to An₅₆. Oscillatory zoned microphenocrysts of plagioclase up to 2 mm in length also occur, displaying either euhedral or resorbed crystal shapes. Microphenocryst compositions are strongly bimodal (Fig. 3). The more common, euhedral microphenocrysts have compositions similar to those of the groundmass crystals. The more strongly resorbed crystals, which often have cloudy interiors, are more calcic (An_80–70), and typically have an overgrowth of composition An_50–60. These calcic crystals are believed, therefore, to be xenocrystic. The clinopyroxene (augite) is a colourless to pale yellow, non-pleochroic variety and occurs as ophitic plates surrounding the plagioclase, as isolated grains, or as greatly elongated feathery crystals up to 3 mm in length. The groundmass augite (Fig. 4) shows a wide range of compositions, both within the group and within individual specimens, which may be due either to local small-scale variations in melt composition, or to disequilibrium crystal-liquid partitioning (Smith & Lindsley, 1971). Fe–Ti oxides form irregular masses or single small octahedra, always in association with the pyroxene. The oxide phase is typically a highly titaniferous magnetite.

View larger version (10K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 3. Plagioclase groundmass, phenocryst and xenocryst compositions in Group I and II sheets projected into the ternary anorthite–albite–orthoclase at 2 kbar water pressure (after Tuttle & Bowen, 1958).

 

View larger version (19K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 4. Pyroxene groundmass and phenocryst compositions in Group I and II sheets. Trend for pyroxenes from Skaergaard intrusion from Brown (1957); trend for pyroxenes from the Shiant Isles from Gibb (1973). Field for Preshal More from Bell et al. (1994); data for Mull Plateau Group pyroxenes supplied by A. C. Kerr (unpublished data, 1996).

 
Group II andesites and porphyritic dacites
Rocks of Group II can be split into those which are broadly similar in mineralogy and texture to the basaltic andesites of Group I and which are not porphyritic, and those which are porphyritic in nature. Although most rocks contain at least a few phenocrysts of plagioclase, the term porphyritic, as used here, relates to rocks with >5% by volume of phenocrysts, which are visible in hand specimen.

Non-porphyritic rocks (mainly andesites)
Compared with Group I, the rocks of Group II are more typically hypocrystalline, with small needles of plagioclase and grains of pyroxene set in a glassy groundmass. They differ from the basaltic andesites in having abundant pigeonite and occasional orthopyroxene in the groundmass, along with occasional phenocrysts of plagioclase and orthopyroxene. The plagioclase occurs as small needles (<1 mm) and as larger, slightly zoned laths, and varies from An₅₈ to An₆₂ (Fig. 3). Rare phenocrysts and xenocrysts of plagioclase are found within these rocks, and have morphologies and compositions similar to those seen in Group I. The groundmass pyroxenes of the Group II rocks typically have elongate feathery and skeletal shapes, indicating rapid growth histories. They show a very large compositional range (Fig. 4) often in individual samples, which may be due to small-scale, local inhomogeneities in magma composition during rapid crystallization. Orthopyroxene (Wo_1.7En_60.6Fs_37.7) phenocrysts occasionally occur in Group II andesites, often associated with plagioclase and magnetite in glomeroporphyritic clots. Groundmass titanomagnetite forms small, isolated grains which show no evidence of exsolution.

Porphyritic rocks (dacites)
The porphyritic dacites have completely glassy groundmasses (pitchstones), with patches and streaks of darker glass being intermingled with paler glass. However, within the larger sheets, the interior portions, which most likely cooled more slowly, have crystallized to a fine-grained intergrowth of plagioclase, K-feldspar, clinopyroxene and quartz. The plagioclase phenocrysts form stumpy prisms up to 3 or 4 mm in length. Many crystals preserve evidence of earlier resorption surfaces, with later oscillatory growth. Compositions range from An₄₀ to An₆₀ from core to rim (Fig. 3). The plagioclase often has a resorbed appearance, with embayed crystal edges. Phenocrystal pyroxene is an unzoned pigeonite (Wo_7.5En_42.5Fs₅₀) (Fig. 4), and comprises rounded crystals, generally <2 mm in diameter. Within the finely crystalline centres of the larger dacite sheets, groundmass clinopyroxene forms small grains of composition Wo₃₂En₃₆Fs₃₂. Titanomagnetite forms small, euhedral octahedra, often completely enclosed in pigeonite phenocrysts. Microphenocrysts of apatite and rare zircon occur as inclusions within phenocrysts of plagioclase, and occasionally within the groundmass.

Group III rhyolites
Compact, light grey rhyolites, often showing rusty-weathered surfaces, make up 10% of the LSSC. They are generally aphyric, and are composed of an intergrowth of oligoclase (An₂₅), orthoclase, and quartz; acicular pyroxene is generally completely pseudomorphed by chlorite. Sporadic phenocrysts of oligoclase form laths up to 2 mm long; alteration of these phenocrysts, and the groundmass feldspar, to sericite, calcite and epidote is characteristic. Occasional microphenocrysts of apatite and zircon can also be found. Porphyritic rhyolites contain phenocrysts of oligoclase 3–4 mm in length, and pyroxene (pigeonite; Buist, 1961), the latter generally pseudomorphed by chlorite and calcite. Aggregates of magnetite are associated with the pyroxene, and microphenocrysts of apatite and zircon occur within the groundmass.

	Whole-Rock Major Element Geochemistry

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
Normative mineralogy and phase relationships
Representative whole-rock major and trace element data for the entire compositional range displayed by the LSSC are presented in Table 1. In terms of their normative mineralogy, all rocks are hypersthene (hy) normative (4–28%), with most being silica oversaturated (Q in the norm) (0–34%). Five samples (all Group I basalts) are silica saturated with normative ol (3–10%).

View this table: [in this window] [in a new window]   Table 1: Whole-rock major and trace element analyses of Group I basalts and basaltic andesites, Group II dacites, Group III rhyolites and Moine pelitic schists

 
It is important when evaluating the validity of petrologic conclusions based on normative mineralogy to eliminate the possibility that any trends might be the result of hydrothermal alteration. Tilley & Muir, (1962) suggested that, with progressive alteration and leaching as a result of hydrothermal activity, the contents of H₂O⁺ and Fe₂O₃ in basaltic rocks will increase, thus increasing the apparent silica saturation as a result of the excessive production of magnetite, releasing silica to produce other phases. They concluded that the Hebridean plateau magma types were all originally ne-normative, and that the hy-normative character suggested by much of the published data was due to the analysis of hydrothermally altered samples. However, subsequent work by Thompson et al., (1972) and Ridley, (1973) showed that fresh basalts from the BTIP range from ne- to hy-normative, and preserve petrographic features of alkali olivine basalts, such as groundmass olivine. Thompson et al., (1972, 1980) also showed that the chemical diversity was not a result of variable hydrothermal metamorphism by demonstrating a strong correlation between silica saturation and elements thought to be essentially immobile during this type of alteration, such as Ti, P and Zr. Further work by Morrison, (1978) on hydrothermally metamorphosed basalts from Mull confirmed the relative immobility of elements such as Ti, Nb, Y and Zr. However, recent work by Bouch et al., (1995) has shown that even the REE can be mobile within low-temperature fluids. The LSSC shows no correlation between H₂O⁺ [measured as loss on ignition (LOI)] and silica saturation (Preston, 1996). A small number of the Group I basaltic andesites have high H₂O⁺ values (>1.5%), and therefore may be slightly affected by hydrothermal alteration. Plateau lavas from outside the zone of pneumatolysis have been reported to have H₂O⁺ or LOI values in excess of 3.7% (Beckinsale et al., 1978; Kerr, 1994). This suggests that, in general, the LSSC has not been as pervasively hydrothermally altered as the plateau lavas, a fact which is very often evident in the field. It is therefore concluded that the normative mineralogy of the LSSC can be used to formulate valid petrologic conclusions.

Figure 5 shows the normative nepheline–olivine–diopside–hypersthene–quartz projection (after Thompson, 1982) derived from the basalt tetrahedron (Yoder & Tilley, 1962). The low-pressure (1 atm) and high-pressure (9 kbar), anhydrous cotectics for olivine + plagioclase + Ca-pyroxene in equilibrium with basaltic liquid are shown. The Group I rocks define a trend which falls just below, but generally parallels, the 1 atm cotectic, although several samples fall well below this cotectic. This equilibrium boundary migrates away from the diopside apex with rising pressure (Morrison et al., 1985) and, judging from experimental data (Thompson, 1982), such samples were probably in pre-emplacement cotectic equilibrium at 2–3 kbar, equivalent to a depth of 7–10 km. The low-pressure cotectic terminates with the crystallization of Ca-poor clinopyroxene (±Fe–Ti oxides) from a tholeiitic basaltic andesite magma (Morrison et al., 1985), as exemplified by the more evolved Group I rocks of the LSSC.

View larger version (12K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 5. Normative di, hy, ol, ne and Q for the LSSC. The field for the Skye Preshal More Basalt type, together with the anhydrous cotectic curves at 1 atm and 9 kbar for olivine + plagioclase + clinopyroxene + basaltic liquid, are from Thompson, (1982). Field for Skye Main Lava Series and Mull Plateau Group lavas from Bell et al., (1994) and Kerr, (1995a).

 
Grove et al., (1982) published a clinopyroxene–olivine–quartz phase diagram based on low-pressure melting experiments on a basalt–andesite–dacite–rhyolite suite of lavas from Medicine Lake, California. This diagram (Fig. 6a) has the advantage over most normative projections in that both parental basic compositions and residual silicic liquids can be plotted on the same diagram, and the entire fractionation history of a suite of volcanic rocks can be examined. Figure 6b shows the data from the LSSC projected onto this phase diagram. As stated above, it is believed that the LSSC basic magmas may have last equilibrated at between 2 and 3 kbar. The effect of increasing pressure under anhydrous conditions in the system clinopyroxene–olivine–quartz and in natural basalts is to decrease the size of the olivine primary phase volume and to expand the orthopyroxene and augite volumes. The opposite effect is seen under conditions of increasing H₂O saturation (Grove & Baker, 1984). However, from the data presented by Grove & Baker, (1984) for the positions of the cotectics at 5 kbar, it is evident that at a pressure of 2–3 kbar, the shifting of the cotectics will be minimal when compared with the 1 atm cotectics.

View larger version (13K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 6. (a) System Cpx–Ol–Qtz (after Grove et al., 1982) at 1 atm, showing schematic fractionation, AFC, assimilation and mixing vectors, and fields occupied by the Skye Preshal More Basalt type and the Mull Plateau Group lavas (data from Bell et al., 1994; Kerr, 1995a). (b) System Cpx–Ol–Qtz at 1 atm, showing data for LSSC Groups I, II and III, and field occupied by Central Mull Tholeiites (data from Kerr, 1995a)

 
As a result of their low Mg contents, none of the projected samples approaches compositions which may represent mantle-derived parental basic liquids. However, those samples which lie near the olivine-augite join could represent liquids derived via the fractional crystallization of olivine + plagioclase from a parental magma similar to the Central Mull Tholeiite and Preshal More magma types. Many Group I samples fall on the augite–pigeonite–plagioclase cotectic, and therefore could be derived via fractional crystallization of the less evolved liquids. Those samples which fall well within the augite + plagioclase primary phase volume could be produced by the mixing of less evolved and evolved liquids from Group I. Similarly, those samples which fall within the pigeonite + plagioclase primary phase volume (many Group I samples and all of Group II) may possibly have been derived from mixing of basaltic melts and more silicic liquids.

Whole-rock major element data
Whole-rock major element geochemical data for the LSSC are presented in Fig. 7 as Harker variation diagrams. SiO₂contents vary from 48.3 to 71.6 wt %, although specimens with SiO₂ between 48.3 and 57.0 wt % are by far the most common (60% of the data presented). The most magnesian specimen of Group I (PGB9) contains 7.75 wt % MgO at mg-number = 0.41, whereas the most evolved Group I specimen (PGB10) contains 4.52 wt % MgO at mg-number = 0.32. Group II rocks range from 2.64 to 0.91 wt % MgO (mg-number = 0.24–0.10). The rhyolites of Group III have the lowest MgO contents, varying between 0.85 and 0.39 wt %, at mg-number = 0.14–0.07. In Group I, CaO varies from 5.49 to 11.80 wt %, whereas Groups II and III have a narrower range, varying from 1.5 to 5.0 wt %. The LSSC as a whole is enriched in the alkali elements when compared with other continental flood basalt suites and with similar tholeiitic suites from the BTIP (Wilson, 1989; Bell et al., 1994), with K₂O rising steadily from 0.2 to 4.3 wt % through Group I to Group III, and Na₂O generally lying within a more constant, but scattered, range of 2–4 wt %. Despite this relative enrichment in alkalis, the LSSC define a sub-alkaline or tholeiitic trend on the K₂O vs SiO₂ plot (Fig. 7g). However, in terms of iron-enrichment trends the LSSC is perhaps atypical when compared with other tholeiitic suites. For example, in the Skaergaard intrusion, the highly reduced state of the initial liquids is thought to have caused a delay in the appearance of magnetite as a fractionating phase, thus contributing to the extreme iron enrichment in the early stages of magma evolution (Wager & Brown, 1968; see Hunter & Sparks, 1987). A similar explanation was put forward for the iron-enrichment trend of the Thingmuli volcano (Carmichael, 1964). The calc-alkaline trend of iron depletion is thought to result from the early removal of magnetite from more oxidized magmas. However, Grove & Baker, (1984) have shown that the difference between the two trends is controlled mainly by the proportions of olivine, plagioclase and pyroxene that crystallize from the parental basaltic melt. The tholeiitic trend is produced by fractional crystallization of a basaltic magma at low pressures, with plagioclase being the dominant fractionating phase, whereas the calc-alkaline trend is produced at moderate to high pressures, where olivine and augite precipitate in slightly greater proportions than plagioclase. The assimilation of silicic crustal material can also assist in the production of the calc-alkaline trend (Grove & Baker, 1984). The LSSC falls between the extremes of tholeiitic and calc-alkaline differentiation trends in terms of iron enrichment (Preston, 1996). From the previous discussion on the normative mineralogy and phase relationships of the LSSC, it was concluded that the magmas equilibrated at upper-crustal levels (2–3 kbar), and that mixing of basic magmas with silicic melts, combined with fractional crystallization, contributed to the formation of some of the intermediate liquids. This combination of magmatic processes at slightly elevated pressures probably contributed to the Loch Scridain sills defining neither a strictly tholeiitic nor a calc-alkaline fractionation trend.

View larger version (17K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 7. (a–h) Major element variation (in wt % oxides) for the LSSC plotted against wt % SiO2. Fields in Fig. 7g after Middlemost, (1975). Symbols as in Fig. 2.

 
Coherent trends on Harker diagrams for related volcanic rocks can be produced through processes of fractional crystallization, differential partial melting, magma mixing and crustal assimilation, or any combination thereof. The variations of MgO, CaO and K₂O vs SiO₂ are generally well correlated (R², the correlation coefficient, for Groups I and II is typically >0.8.) Fe₂O₃* and MnO are well correlated with SiO₂, taking the suite as a whole, although there is marked inter-group scatter. P₂O₅ shows a relative enrichment with increasing SiO₂ content through Group I (0.07–0.20 wt %), although the correlation coefficient is <0.8. For the majority of Groups II and III, P₂O₅ has a relatively constant value of 0.2 wt % with increasing SiO₂ content. However, several samples of Group II fall well off this general trend at slightly elevated P₂O₅ values (0.32 wt %). These are the porphyritic dacites, and, as described above, they contain occasional microphenocrysts of apatite which could give rise to the elevated P₂O₅ contents of these rocks. Na₂O, TiO₂, and Al₂O₃ do not correlate well with SiO₂.

	Whole-Rock Trace Element Geochemistry

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
All specimens were analysed for the trace elements Ba, Ce, Cr, Cu, La, Nb, Nd, Pb, Rb, Th, Sc, Sr, V, Y, Zn and Zr, and selected samples for the full range of the REE. Representative trace element analyses are presented in Table 1. Figure 8 shows the variation of selected trace elements with wt % SiO₂.

View larger version (15K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 8. (a–h) Variation of selected trace elements (in ppm) for the LSSC plotted against wt % SiO2. Symbols as in Fig. 2.

 
Whole-rock trace element data
The concentrations of Ni and Cr show strongly bimodal distributions within Group I (Fig. 8a), although overall the least evolved members of Group I have the highest concentrations which decrease with increasing degree of fractionation (78–42 ppm and 350–180 ppm for Ni and Cr, respectively). The more evolved members of Group I contain consistently low concentrations of Ni (10–24 ppm) and Cr (30–90 ppm). Concentrations of Ni and Cr are generally low in Groups II and III (<10 ppm and 69 ppm, respectively). Both Sc and V show a steady decrease in concentration through Group I (52–32 ppm Sc; 360–250 ppm V). Sc contents in Groups II and III are low (<30 ppm), whereas V concentrations are highly variable in Group II (50–200 ppm), and low in Group III (<40 ppm). Systematic variations in the large ion lithophile elements (LILE; Rb, Ba and Sr) are relatively well constrained. Ba contents in Group I show a steady increase (180–580 ppm), rising to higher values in Groups II and III (>600 ppm); one sample in Group I has an anomalously high Ba content at 780 ppm. Rb shows a similar distribution (2–60 ppm in Group I; >100 ppm in Group II; >120 ppm in Group III). However, Sr shows a much more complex behaviour; concentrations in Group I fall within a relatively constant range of 200–300 ppm, although one sample has 125 ppm, and one 425 ppm. In Group II, Sr concentrations decrease from 310 ppm to 150 ppm with increasing SiO₂, and are at sub-150 ppm values in the Group III rhyolites. This complex behaviour of Sr suggests that the bulk distribution coefficient of Sr in Group I was close to or less than unity, implying plagioclase did not make up more than 60% of the fractionating assemblage. Plagioclase fractionation may have been more important in Group II. Concentrations of the high field strength elements (HFSE; Nb, Zr and Y) are relatively well constrained, with all showing approximately two-fold increases with degree of fractionation.

Rare earth elements
Concentrations of the REE within the LSSC are presented in chondrite-normalized (Nakamura, 1974) Masuda-Coryell plots in Fig. 9. The profiles range from being light REE (LREE) depleted (Ce/Yb_N= 0.61), with the most basic member of Group I (ORB2) having LREE concentrations between 10 and 20 times chondritic values, to LREE-enriched profiles [(Ce/Yb)_N= 5.65], with values reaching 200 times chondritic values in the more evolved members of Groups I and II. The HREE profile (Gd–Lu) is generally flat in both Groups I and II. More evolved members of Groups I and II have negative Eu anomalies (Eu/Eu* = 0.72), indicative of plagioclase fractionation. The profile for the least evolved member of Group I (ORB2) shows a striking similarity to the profiles shown by the Preshal More magmas (Fig. 9a), and is somewhat different from the slightly ‘humped’ profiles shown by the Central Mull Tholeiites (Mattey et al., 1977; Bell et al., 1994; Kerr, 1995a). The profiles for the rhyolites of Group III are strongly LREE enriched [(Ce/Yb)_N= 6.29], with deep negative Eu anomalies (Eu/Eu* = 0.51), and are very similar to those shown by local Moine pelitic schists (Fig. 9d).

View larger version (22K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 9. Chondrite-normalized (Nakamura, 1974) REE diagrams for: (a) and (b) the Group I basalts and basaltic andesites; (c) Group II andesites and porphyritic dacites; (d) Group III rhyolites. Comparative data for Preshal More basalts from Mattey et al., (1977) and Bell et al., (1994). Data for Mull Plateau Group and Central Mull Tholeiites from Kerr, (1995a). Comparative data for Moine pelitic schists (this study), and Lewisian gneisses from Whitehouse, (1990) and Whitehouse & Robertson, (1995).

 

	Whole-Rock Isotope Geochemistry

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
Whole-rock Rb–Sr, Sm–Nd and Pb isotope data have been determined for a number of samples covering the entire compositional and petrographical variations of the suite. Two local Moine pelitic schists from Traigh Bhan na Sgurra (Fig. 1) were also analysed. These data are presented in Table 2, and as stated previously, have been age corrected to 55 Ma.

View this table: [in this window] [in a new window]   Table 2: Whole-rock Sr-, Nd- and Pb-isotope data for the Loch Scridain Sill Complex, and two local Moine pelites

 
Figure 10 shows that there is a good negative correlation between (⁸⁷Sr/⁸⁶Sr)₅₅ and (¹⁴³Nd/¹⁴⁴Nd)₅₅. Only one Group I specimen plots near the Mantle Array [ORB2: (⁸⁷Sr/⁸⁶Sr)₅₅ = 0.70376, (¹⁴³Nd/¹⁴⁴Nd)₅₅ = 0.51293], with the remainder of Group I falling on a curved trend at elevated (⁸⁷Sr/⁸⁶Sr)₅₅ (0.70769-0.71545) and low (¹⁴³Nd/¹⁴⁴Nd)₅₅ values (0.51217–0.51188). Members of Group II and III have relatively constant (¹⁴³Nd/¹⁴⁴Nd)₅₅ values (0.51181–0.51187), but variable (⁸⁷Sr/⁸⁶Sr)₅₅ (0.71419–0.72048). (⁸⁷Sr/⁸⁶Sr)₅₅ shows a well-constrained curved trend with SiO₂ (Preston, 1996), and coherent relationships with trace element ratios such as Zr/Nb. The Moine samples have (⁸⁷Sr/⁸⁶Sr)₅₅ and (¹⁴³Nd/¹⁴⁴Nd)₅₅ = 0.72841–0.72942 and 0.51175–0.51176, respectively.

Figures 11a and 11b are plots of ²⁰⁷Pb/²⁰⁴Pb vs ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb vs ²⁰⁶Pb/²⁰⁴Pb for the LSSC. Fields for the Mull Plateau Group and the SMLS are plotted for comparison (Dickin, 1981; Kerr et al., 1994), along with data from Lewisian amphibolites and granulites (Whitehouse & Robertson, 1995), and Moine pelitic schists (Thompson et al., 1986; this study). The LSSC as a whole has a restricted range of Pb isotope ratios when compared with the Mull Plateau Group lavas or the SMLS. ²⁰⁶Pb/²⁰⁴Pb ratios for Groups I, II and III vary from 17.953 to 18.993. The ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb ratios show even less variation (15.467–15.619 and 38.002–38.796, respectively). The Moine pelites have ²⁰⁶Pb/²⁰⁴Pb, ²⁰⁷Pb/²⁰⁴Pb and ²⁰⁸Pb/²⁰⁴Pb ratios of 19.240–19.990, 15.679–15.697 and 38.932–38.996, respectively.

View larger version (22K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 11. (a) and (b) Pb–Pb isotope variations for the LSSC and local Moine pelitic schists. Data for Skye and Mull lavas from Dickin et al., (1984) and Kerr et al., (1995); for Lewisian granulite and amphibolite facies gneisses from Whitehouse, (1990) and Whitehouse & Robertson, (1995). Northern Hemisphere Reference Line (NHRL) from Hart, (1988).

 
These data and relationships reveal that the basic magmas of the LSSC did not evolve via closed system fractional crystallization processes. The high (⁸⁷Sr/⁸⁶Sr)₅₅ values of the more evolved members of Group II and the Group III rhyolites are consistent with their having a greater crustal component than the other groups.

	The Petrogenesis of the LSSC

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
The geochemical evolution of the individual (magma) groups of the LSSC can be shown to be dominated by distinctly different petrogenetic processes. The following discussion draws together the evidence from the major and trace element and isotopic data to examine the causes for the intra-group compositional variations.

Geochemical evolution of the Group I basic magmas
The major element variation of the Group I basic magmas suggests that fractional crystallization may have played an important role in producing the observed geochemical trends. It is evident from the major element data that the observed compositional trends of the LSSC (Fig. 7) require the fractionation of an Mg-rich phase throughout Group I. Olivine fractionation is one possibility, although this is absent as a phenocryst (and groundmass) phase in Group I. Mass-balance fractionation modelling (Preston, 1996) suggests that olivine made up no more than 10% of the crystal extract during the early stages of fractionation. The Group I compositional trends also call for a calcic phase to be fractionating. As Group I rocks contain occasional microphenocrystic and glomeroporphyritic plagioclase, this phase would seem an obvious candidate. The marked scatter in the Al₂O₃ data does not negate this possibility, because between 40 and 50% plagioclase must be in the fractionate to allow Al₂O₃ to diminish noticeably (Albarède, 1992). Whereas plagioclase phenocryst accumulation may be the cause of some of the scatter about the Group I trend, olivine + plagioclase removal alone cannot explain the observed trends. The role of a third phase during the evolution of the Group I magmas is confirmed when the CaO/Al₂O₃ data are examined. From a consideration of the data displayed in Fig. 12, it is evident that the fractionation of a phase, or combination of phases, with CaO > Al₂O₃ is required. Combining these observations and those obtained from the low-pressure phase relationships, a calcium-rich clinopyroxene (diopside or augite) would be the obvious choice. However, diopside and augite, like olivine, do not occur as phenocryst phases in Group I.

View larger version (7K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 12. Plot of CaO/Al2O2 vs wt % SiO2, showing importance of clinopyroxene fractionation in the evolution of the LSSC basic magmas. Symbols as in Fig. 2.

 
The relatively low Ni contents of Group I rocks suggest that they do not represent primary magma compositions. The trend of decreasing Ni with increasing SiO₂ is consistent with the removal of olivine from the less evolved specimens as concluded above for the major elements. The Ni concentrations of the more evolved members of Group I are consistent with highly evolved tholeiitic magmas which have already fractionated olivine. The decreasing Cr contents with increasing SiO₂ can be explained through the removal of Cr-spinel, the removal of clinopyroxene, or through the mixing of coeval unevolved and fractionated magmas. However, as Cr-spinel appears very early on the liquidus during crystallization of basaltic magmas (Irvine, 1967), the higher concentrations of Cr in the more basic liquids would suggest that the Cr depletion through Group I is not a consequence of Cr-spinel crystallization, but was controlled primarily by the removal of late-crystallizing clinopyroxene (see Bell et al., 1994). The compatible behaviour of Sc and V is also consistent with clinopyroxene fractionation.

The need to suggest the role of the three-phase mineral assemblage olivine + clinopyroxene + plagioclase during the evolution of basaltic magmas, even though the members of a suite are devoid of one or more of these minerals (especially clinopyroxene) within the phenocryst assemblage, has provided a paradoxical problem for many studies of tholeiitic basalt suites from both the continental and oceanic environments. Within the BTIP, the cone-sheet complexes of Skye (Bell et al., 1994) and Ardnamurchan (Holland & Brown, 1972) highlight the so-called pyroxene paradox. The work of Cox & Mitchell, (1988) on plagioclase-phyric and aphyric basalts from the Deccan Traps, India, suggests that only the denser liquids (Fe-rich tholeiites) are capable of transporting suspended crystals to the site of emplacement, thus discriminating strongly in favour of less dense plagioclase over denser olivine and clinopyroxene. The aphyric rocks, which are generally more primitive (Fe poor), were therefore considered to have lost all the fractionating phases via a process of crystal settling. This is identical to the case of the Skye tholeiitic cone-sheets, where the least evolved specimens (mg-number 0.43) are the most Fe poor (wt % Fe₂O₃* = 11.42) there being a steady increase in Fe₂O₃* with progressive crystal fractionation (Bell et al., 1994). Although the majority of the Loch Scridain Group I rocks are not as Fe enriched as those studied by Cox & Mitchell, (1988), the least evolved specimen does have Fe₂O₃* at a comparable 16 wt %. A similar process may have been in operation in the Loch Scridain magma chamber(s), with the cognate xenoliths being testament to the efficient separation of liquids and crystals (Preston & Bell, 1997).

Francis, (1986) argued that the pyroxene paradox may actually be illusory. His work on Tertiary picrites from Baffin Bay, West Greenland, suggests that the compositional variation that is responsible for the problem of ‘occult’ pyroxene fractionation may be attributed to the dispersion of olivine fractionation trends between the limits of equilibrium and fractional crystallization. However, the evolved nature of the Loch Scridain magmas suggests that olivine fractionation played only a minor role in the production of the intra-suite variation recorded here.

This problem was also addressed by Walker et al., (1979), who explained it in terms of mixing processes between unevolved and fractionated compositions on a common curved cotectic so that the resultant liquids plot within the high-Ca pyroxene saturated field on relevant phase diagrams. Shibata et al., (1979) found that such mixed magmas often contained significantly greater abundances of trace elements (particularly Ni) than would be expected from fractionation calculations; this is seen in the LSSC, with some of the most evolved members of Group I (e.g. ARB1) having relatively high Ni and Cr contents (34.7 ppm and 152.7 ppm, respectively). Those LSSC specimens with high Ni and Cr plot within the augite primary phase volume on the ol–cpx–qtz phase diagram (Fig. 6b), and can therefore be interpreted as mixtures between coeval unevolved and fractionated liquids (Walker et al., 1979). Further evidence for such mixing events is found in the highly calcic plagioclase xenocrysts which these samples contain.

A closed-system crystal fractionation scheme was tested by Preston, (1996) using standard least-squares modelling and mass-balance approaches (e.g. Bryan et al., 1969). In general, these models were found to be unable to account for the SiO₂ and TiO₂ enrichment of the complete Group I trend. However, it was found possible to model the evolution of those Group I samples which fall on the plagioclase + augite + pigeonite cotectic in Fig. 6b. The results of these calculations are shown in Table 3. However, closed system crystal fractionation is unable to account for the extreme enrichment of the incompatible elements seen in Group I.

View this table: [in this window] [in a new window]   Table 3: Results from least-squared modelling

 
The Sr, Nd and Pb isotope data also imply that the Group I basic magmas assimilated crustal material during their evolution. These magmas trend towards the field for Moine schists, both in terms of their trace element and Sr–Nd–Pb isotope geochemistry (Figs 10, 11 and 13). Unlike some of the basal Staffa Magma Type lavas from Mull, which have been contaminated with both Lewisian leucogneiss and Moine metasediments (Morrison et al., 1985; Thompson et al., 1986), the Pb isotope data for the LSSC show no hint of the unradiogenic ²⁰⁶Pb/²⁰⁴Pb ratios that are characteristic of the interaction of mantle-derived magmas and Lewisian gneisses (Dickin, 1981). Therefore, it is a reasonable assumption at this point to assume that the basic magmas have been contaminated predominantly with material derived from the Moine. The diversity of near-monomineralic cognate xenoliths found in many of the Group I sheets also implies that the fractionating assemblage was highly variable (Preston & Bell, 1997). Even given a constant rate of assimilation (r = 0.4), the range of possible AFC trends is potentially very large, because of the variations in the fractionating assemblage resulting in highly variable K_D for Sr and Nd, as shown in Fig. 13. However, Fig. 13 also demonstrates that the contaminant cannot have been a bulk Moine pelite, but is more likely to have been a silicic melt derived from the Moine. Such melts are likely to have the high-Nd characteristics necessary to give the strong curvature seen in the best-fit AFC curve (e.g. Fowler, 1988, 1992). The composition of the Moine melt was calculated from the analysis of a local pelite (SOM1) using an equilibrium melting model (Wood & Fraser, 1976) and mineral-melt distribution coefficients from Bender et al., (1984); evidence from partly digested pelitic xenoliths suggests that the restite was dominated by quartz and plagioclase (75%), with lesser biotite (20%) and minor garnet (5%) at between 20 and 30% melting. The fractionation scheme of olivine + plagioclase + augite in their approximate cotectic proportions (10:45:45), followed by plagioclase + augite + pigeonite in the proportions 45:40:15 as suggested by modelling of the major element variations (Preston, 1996), can also account for the trace element and isotope characteristics of the Group I magmas. However, the general trend of increasing TiO₂ with continuing fractionation in Group I, and the lack of evidence for the accumulation of titanomagnetite, suggests that certain batches of basic magma possibly interacted with a Ti-enriched contaminant. Preston, (1996) has shown that the basic magmas of the LSSC interacted extensively with highly aluminous (and Ti-rich) liquids generated by high degrees of melting of Moine metasedimentary rocks. These liquids are now represented by the mullite-buchite xenoliths present in the vast majority of the Group I sheets. These glassy xenoliths often have reaction rims of plagioclase, corundum and spinel, and document a very complex series of magma-xenolith interactions. The exact mode of contamination probably involved the extraction of silicic melts from the Moine country rocks, accompanied by interaction with the resulting residue.

View larger version (16K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 10. (87Sr/86Sr)55 vs (143Nd/144Nd)55 isotope diagram for the LSSC, along with fields for the SMLS, Mull Plateau Group lavas, Preshal More Basalt type, BTIP silicic intrusives, Lewisian granulite and amphibolite facies gneisses, and Moine metasediments. LSSC and Moine data points, this study. Additional data from Walsh et al., (1979), Dickin et al., (1984), Holden et al., (1987), Clayburn, (1988), Whitehouse, (1990), Bell et al., (1994), Kerr et al., (1995), Whitehouse & Robertson, (1995). Symbols as in Fig. 2.

 

View larger version (13K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 13. AFC modelling for the LSSC. Crystal extracts are those suggested by the least-squares modelling, and those seen within the gabbroic cognate xenoliths (Preston & Bell, 1997). (a) Plot of Zr/Nb vs (87Sr/86Sr)55 showing modelled AFC trend using gabbroic crystal extract with 10% olivine, 45% plagioclase and 45% clinopyroxene: KD Sr = 0.85, KD Zr = 0.068, KD Nb = 0.0077; r = 0.4; Co, ORB2; Ca, modelled Moine melt with 308 ppm Sr, 135 ppm Nd, 612 ppm Zr, 57 ppm Nb. (b) (87Sr/86Sr)55 vs (143Nd/144Nd)55 isotope diagram for the LSSC with modelled AFC trends. Parameters for AFC trends are as follows: (1) represents an extreme clinopyroxene cumulate: KD Sr = 0.06, KD Nd = 0.23; r = 0.4; Co, ORB2; Ca, SOM1; (2) represents gabbroic extract as in (a): KD Sr = 0.85, KD Nd = 0.12; r = 0.4; Co, ORB2; Ca, SOM1; (3) represents extreme plagioclase cumulate: KD Sr = 1.83, KD Nd = 0.055; r = 0.4; Co, ORB2; Ca, SOM1; (4) KD Sr = 0.85, KD Nd = 0.12; r= 0.4; Co, ORB2; Ca, modelled Moine melt as in (a). Divisions on AFC curves are percent of magma crystallized in 10% steps. KD values taken from compilation by Rollinson, (1993). r, rate of assimilation to crystallization; Co, composition of parent magma; Ca, composition of assimilant.

 
The isotopic data suggest that 30% crystallization is required to produce the Group I trend using the above parameters and fractionating assemblages (Fig. 13). However, the AFC calculations also show that there is substantial decoupling not only between the major and trace elements, but also between individual incompatible elements. For example, elements such as Zr, Nb, Ba and the LREE require between 40% and 60% crystallization to achieve the required enrichment, whereas elements such as Y and the heavy REE (HREE) only require between 5% and 50% crystallization. This decoupled behaviour between major and trace elements during crystal-liquid fractionation could be explained in terms of the magma evolving in a chamber undergoing periodic replenishment, tapping, and fractionation (RTF processes) (O'Hara, 1977; O'Hara & Mathews, 1981; O'Cox, 1988). However, as the more evolved members of Group I contain a substantial amount of crustal material, it would seem more reasonable to suggest that the concentrations of these more incompatible elements are controlled predominantly by the contamination process. The AFC calculations were not extended to the Group II rocks, as textural features suggest they originated via simple magma mixing processes (see below).

Geochemical evolution of the Group II intermediate magmas
The major element evolution of the Group II intermediate rocks is more difficult to evaluate. The plot of TiO₂ vs mg-number (Fig. 2) shows that Groups I and II are not related via simple fractional crystallization. The slope on the CaO/Al₂O₃ vs SiO₂ diagram (Fig. 12) becomes slightly less negative, suggesting that a Ca-poor pyroxene started to fractionate, along with plagioclase. This is supported by the presence of plagioclase and pigeonite phenocrysts in the Group II dacites. The dacites also contain glass of two distinct compositions, one silicic, and one more basic (Table 4). The basic glass is a highly evolved ferrodiorite, as shown by its elevated SiO₂, K₂O and TiO₂ contents (e.g. Bell, 1983), and the silicic glass has a near-minimum melt normative composition. The plagioclase phenocrysts in the Group II dacites have crystal cores that would be in equilibrium with the silicic glass, and a more calcic overgrowth which would be in equilibrium with the more basic glass (Preston, 1996). This textural, mineralogical and major element evidence suggests that the Group II intermediate rocks are the result of simple mixing between coexisting basic and silicic magmas. Such a process is also supported by the major elements, with whole-rock compositions falling on relatively well-constrained mixing lines between the basic and silicic glass involving between 70 and 80% of the silicic end-member (Preston, 1996). If this is to be confirmed, both the trace element and Sr–Nd isotope characteristics must also be consistent with this hypothesis (e.g. Langmuir et al., 1978; Stamatelopoulous-Seymour & Vlassopoulos, 1992). Figures. 14a and 14b are plots of (La/Nd)_N vs (Sm/Yb)_N, and (⁸⁷Sr/⁸⁶Sr)₅₅ vs (¹⁴³Nd/¹⁴⁴Nd)₅₅, respectively, showing calculated mixing trends. As no trace element or isotope analyses of the glasses could be performed, the two end-members chosen were the most SiO₂-poor member of Group I (ORB1), which also has the least contaminated Sr–Nd isotopic signature, and the most SiO₂-enriched member of Group III (PGF1), which also has the most ‘crustal’ Sr&Nd isotope values. The fit of the Group II data to these calculated mixing curves is very good, and suggests that between 75 and 90% of the silicic end-member was involved in the mixing process, consistent with the proportions obtained from the major elements. Although many of the Group I samples also appear to fall on these mixing lines, mixing relationships are not consistent throughout the full range of major and trace elements (Preston, 1996), confirming that fractional crystallization and assimilation processes were also important in the petrogenesis of Group I as shown above.

View this table: [in this window] [in a new window]   Table 4: Glass compositions in a Group II porphyritic dacite (CUPI1)

 

View larger version (10K): [in this window] [in a new window] [Download PowerPoint slide]   Fig. 14. Calculated mixing trends showing the probable mixed-magma nature of the Group II intermediate rocks , shown by (a) (La/Nd)N vs (Sm/Yb)N, and (b) (87Sr/86Sr)55 vs (143Nd/144Nd)55.

 
Similar glasses and phenocryst assemblages can be found in dacitic and rhyolitic intrusions elsewhere in the BTIP. For example, the Loch Ba ring dyke on Mull, dominantly a banded rhyolitic tuff, contains dark glasses with phenocrysts of plagioclase (An_65–30), augite and pigeonite, associated with light rhyolitic glasses with phenocrysts of plagioclase (An_32–21), sanidine, hedenbergite and fayalite (Sparks, 1988). Sparks, (1988) interpreted the mineralogical and textural relationships within the Loch Ba ring dyke as evidence for the intimate mixing of basic and silicic magmas during emplacement.

Geochemical evolution of the Group III silicic magmas
As there is no evidence for extensive titanomagnetite or ilmenite fractionation within the LSSC magmas, the low-TiO₂ nature of the Group III rhyolites suggests that they are not primarily the products of extensive fractional crystallization of a basic magma; rather, the phase relationships and major element geochemistry are consistent with their having been derived mainly via crustal anatexis of basement rocks. It is suggested that the Group III rhyolites represent partial melts extracted from the more pelitic units of the Moine metasedimentary complex. Partial melts extracted from pelitic metasediments will be peraluminous [cationic Al > (K + Na + 2Ca)], corundum normative (e.g. Gardien et al., 1995; Patiño-Douce, 1995), and they will also have high K/Na ratios (Whitney, 1988). Partial melting experiments on common, natural metasedimentary rocks (MacRae & Nesbitt, 1980; Thompson, 1981; Vielzeuf & Holloway, 1988) show that liquid compositions similar to the rhyolitic rocks of the LSSC can be produced at temperatures of between 700 and 950°C under both dry and H₂O-saturated conditions. The Group III rhyolites fulfil all these geochemical requirements, being peraluminous, and having near ‘minimum" normative compositions (Preston, 1996).

The trace element geochemistry of silicic melts derived from crustal sources will be controlled by the phases involved in the melting process. For example, if muscovite and biotite are consumed in the melting process, the melt is likely to be enriched in trace elements such as Rb, Ba and Li. It is therefore likely that major local variations in melt trace element geochemistry will occur because of the typically heterogeneous nature of the crustal rock sources, and because of the local physical conditions (pressure, temperature and H₂O content) of melting.

However, given these obvious limitations, it is believed that the trace element and Sr–Nd isotopic characteristics of the Group III rhyolites are consistent with their derivation from crustal sources. The rhyolites are characterized by high K₂O contents (>4 wt %), low Na₂O (<3 wt %), low Sr (<132 ppm), high Ba (>816 ppm) and high Rb (>130 ppm). This is confirmed when the Sr–Nd isotope data are examined (Fig. 10). All the rhyolites have high (⁸⁷Sr/⁸⁶Sr)₅₅(0.71730-0.72048) and low (¹⁴³Nd/¹⁴⁴Nd)₅₅ (0.51185), which are typical for S-type granites derived from old continental crust, as defined by Chappell & White, (1992). All the Group III rhyolites are LREE enriched, with (Ce/Yb)_N in excess of seven, and they all have negative Eu anomalies, consistent with crustal melting leaving a plagioclase-rich restite. Figure 9d shows the chondrite-normalized REE diagrams for the Group III rhyolites along with profiles from local Moine pelitic schists (this study), and the range shown by Lewisian amphibolite and granulite facies gneisses from Iona and elsewhere in NW Scotland (Weaver & Tarney, 1980, 1981; Kerr, 1995b). These data strongly suggest that Moine pelitic schists were the source for the Group III rhyolites. However, depending upon the accessory phases remaining in the restite, silicic partial melts are likely to have higher trace element concentrations than their source, a fact not observed in the Group III rhyolites. Mineralogical and geochemical evidence from partially digested pelitic xenoliths and the glassy mullite-buchite xenoliths suggests that garnet was a residual phase during the partial melting process (Preston, 1996). As such, silicic partial melts are likely to be much more HREE depleted that the rhyolites as seen now. This suggests that small (<10%) amounts of HREE-enriched basic magma mixed with the silicic crustal melts to produce the Group III rhyolites. The Sr–Nd isotope diagram (Fig. 10) plots the fields for Lewisian granulite and amphibolite facies gneisses, and Moine pelitic schists, and it is evident that the Group III rhyolites cannot have been derived via the partial melting of Lewisian rocks, and that Moine pelitic schists are the most likely source, a fact reinforced by the radiogenic Pb isotope ratios (Fig. 11). This contrasts strongly with other silicic intrusions from the Mull central complex, which formed either by a combination of partial melting of Lewisian basement together with some melts derived by the fractional crystallization of basaltic magma, or dominantly by the fractional crystallization of basaltic magma, along with a small amount of crustal contamination (Pankhurst et al., 1978; Walsh et al., 1979).

	Discussion

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
Parental magmas, and contamination processes
From the previous discussion, it is evident that the basic magmas of the LSSC have suffered extensive fractional crystallization and contamination with crustal material. It is therefore difficult to assign parental magma status to any of the analysed samples. Although similar to the Central Mull Tholeiites (CMT), the least evolved, least contaminated sample (ORB2) has trace element and isotopic characteristics which most closely resemble the Preshal More magma type from Skye (Mattey et al., 1977; Bell et al., 1994; Kerr, 1995a).

Because of fractional crystallization and assimilation processes, the more evolved Group I magmas show greater incompatible-element enrichment, and more enriched Sr–Nd isotopic signatures when compared with similarly evolved basic magmas in the BTIP and other continental flood basalt provinces. This is unusual for a suite of particularly ‘normal’ basic rocks. No other basic magmas from the BTIP have such high ⁸⁷Sr/⁸⁶Sr initial ratios (e.g. Moorbath & Thompson, 1980; Thirlwall & Jones, 1983; Thompson et al., 1986; Bell et al., 1994; Kerr et al., 1995), and the values from the LSSC basic magmas are perhaps only matched by certain low-Ti basalts from the Paraná, southern Brazil (e.g. Bellieni et al., 1984; Mantovani et al., 1985; Gibson et al., 1995). Gibson et al., (1995) suggested that most of the geochemical variation within the Paraná continental flood basalt suite can be explained by the mixing of depleted, asthenospheric basalts with small quantities of variably enriched melts derived from the sub-continental mantle lithosphere (SCML). Nevertheless, it was still necessary to call upon crustal contamination to account for the very high initial Sr isotope ratios of some of the low-Ti basalts (Gibson et al., 1995). In the BTIP, Kerr, (1993) has suggested that some of the lavas towards the base of the Mull lava pile were produced by the addition of an enriched small-fraction-melt from the SCML to an uncontaminated magma of the Mull Plateau Group. Such an addition to a CMT or PM-type parent magma cannot account for the major and trace element characteristics of the more evolved members of Group I. Therefore, the incompatible element enrichment and Sr–Nd isotope signatures of the LSSC can only be the result of crustal contamination.

It has been shown that turbulent flow of hot basic magma within conduits is a very effective way to assimilate crustal material (Huppert & Sparks, 1980; Kille et al., 1986; Kerr et al., 1995). The process of assimilation during turbulent ascent (ATA) (Moorbath & Thompson, 1980; Thompson et al., 1982, 1986; Huppert & Sparks, 1985) has been suggested as a mechanism to explain the contamination of the Mull Plateau Group lavas (Kerr et al., 1995). Significantly, such a process will lead to the hottest, most magnesian magmas becoming the most contaminated. Unlike most of the Skye and Mull lavas, which fractionated first at high pressure at the base of the crust and then assimilated Lewisian granulite facies material (Thompson et al., 1982; Thirlwall & Jones, 1983; Kerr et al., 1995), the PM-type magmas underwent fractional crystallization and assimilation at higher crustal levels and typically display AFC-type contamination signatures (Thompson et al., 1986; Bell et al., 1994). The bulk addition of the appropriate crustal contaminant to a PM-type magma is often able to account for incompatible-element abundances and Sr–Nd isotope signatures, but commonly at the expense of being unable to model major element abundances (Thompson et al., 1986; Bell et al., 1994; Preston, 1996). In the BTIP, AFC-style crustal contamination is exemplified by the Skye cone-sheet magmas (Bell et al., 1994), and by the picrites and picrodolerites of the Trotternish sills, northern Skye (Gibson, 1990).

However, in a detailed study of the crustal xenoliths of the LSSC, Preston, (1996) has shown that the small-scale interactions between basic magma and the crustal xenoliths may leave complex trace element and isotope fingerprints on individual batches of magma that are difficult to model effectively through simple AFC calculations.

The magmatic plumbing system for the LSSC
The geochemistry of the LSSC (i.e. phase relationships, major and trace element and isotopic characteristics) requires that the initial parental melts were stored within the upper crust, where they underwent extensive AFC processes. The presence of gabbroic cumulate xenoliths within many of the sheets suggests that discrete magma reservoirs developed within the crust. This is also a requirement for the process of combined assimilation and fractional crystallization, by which the majority of the Group I magmas evolved.

However, what is far less well understood is the form of these magma reservoirs. There is no geophysical evidence for large solidified magma chambers under the Loch Scridain district (Bott & Tandrigoda, 1987). One possibility is that the main magma storage reservoir was situated under the central igneous complex (for which there is geophysical evidence; Bott & Tandrigoda, 1987), and that basic magmas were contaminated and picked up crustal xenoliths through their turbulent migration to the site of emplacement. However, such a process is more likely to result in the hottest, most magnesian magmas becoming the most contaminated (e.g. Thompson et al., 1982, 1986; Huppert & Sparks, 1985; Kerr et al., 1995). However, if the magmas underwent fractional crystallization, or AFC, before emplacement, these latter processes are likely to have the dominant control on the final magma composition.

The very high initial Sr isotope and trace element enriched signatures of the basalts and basaltic andesites may suggest that a two-stage AFC process, similar to that proposed by Reiners et al., (1995, 1996) was in operation. Van der Laan & Wyllie, (1993) have shown that the interaction of mafic melts with silicic melts rich in alkalis and H₂O results in the lowering of the liquidus temperature of the melt, and Sisson & Grove, (1993) suggested that this process will also enlarge the stability field of olivine at the expense of plagioclase and clinopyroxene. In this way, the early stages of combined crustal assimilation and fractional crystallization, when the magmas are at their hottest, are characterized by high rates of assimilation, without much crystallization (perhaps olivine alone). This may result in large changes in the isotopic and trace element characteristics of the magmas with little major element differentiation. The second stage of AFC, beginning with the crystallization of plagioclase and pyroxene, is characterized by lower rates of assimilation, and the magma evolves more rapidly (Reiners et al., 1995). However, this style of contamination depends upon the ability of the basic melts to buffer the system against any individual component of the contaminant. For example, the high Sr concentrations of many basic magmas compared with those of the crustal contaminants buffer ⁸⁷Sr/⁸⁶Sr of the melt at low degrees of contamination. The early stages of contamination within the LSSC resulted in rapid changes in the ¹⁴³Nd/¹⁴⁴Nd of the magma with little change in ⁸⁷Sr/⁸⁶Sr (Fig. 10). This corroborates the contention that the contaminant was a high-Nd, low-Sr silicic Moine melt rather than bulk Moine, and suggests that the ¹⁴³Nd/¹⁴⁴Nd of the magma was buffered at a relatively early stage. Later crystallization of abundant plagioclase, as is suggested by the presence of numerous cognate gabbro and anorthosite cumulate xenoliths (Preston & Bell, 1997), would allow the magmas to attain the higher ⁸⁷Sr/⁸⁶Sr values seen in the more evolved members of Group I.

The favoured model is one whereby basic magmas pond, fractionate and assimilate crustal material in a network of poorly connected, high-level, sheet-like bodies (e.g. Cox, 1980). The large surface area-to-volume ratio of sills enables the magmas to interact extensively with the surrounding crust (e.g. Kille et al., 1986; Thompson et al., 1986). Turbulent flow of magma in sill- or dyke-like conduits will also be an efficient mechanism for the breaking-off of wall-rock xenoliths, for their transport into connected sheet-like reservoirs, and for the continuous mixing of silicic partial melts from the wall-rocks into the body of the magma. Once a magma reservoir is filled with hot basic magma, both wall-rocks and xenoliths can be melted, and fractional crystallization can occur. A similarly complex magmatic plumbing system has been proposed for the evolution of the lava successions on Skye and Mull (Dickin et al., 1984; Morrison et al., 1985).

	Summary and Conclusions

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
The LSSC provides perhaps a unique window into studying the interaction of basic magmas with continental crust. The suite consists of basic, through intermediate to silicic rocks, all of which evolved by separate petrogenetic processes. Depleted, MORB-like Central Mull Tholeiite or Preshal More type parental magmas ponded within the upper crust, and underwent combined assimilation and fractional crystallization, to give rise to what are now the Group I basic magmas. The basic magmas assimilated large quantities of Moine metasedimentary material in the form of a silicic melt during fractional crystallization, resulting in extreme incompatible-element enrichments and imparting very high (⁸⁷Sr/⁸⁶Sr)₅₅ and low (¹⁴³Nd/¹⁴⁴Nd)₅₅ values on what were ‘normal’ basic magmas. The basic magmas may also have interacted extensively with highly aluminous liquids generated during the partial fusion of the subjacent Moine metasedimentary rocks. These unusual crustal liquids are now represented by numerous glassy xenoliths, which possess reaction rims of plagioclase, corundum and spinel. Strong decoupling between the major elements and incompatible elements, and even between individual incompatible-elements, during fractional crystallization, suggests that magma recharge processes (RTF) may have operated in the LSSC magma storage reservoirs. The effect of magma ponding in the crust was to allow partial fusion of the Moine metasedimentary basement rocks to produce the Group III rhyolitic magmas. These mixed with small amounts of basic magma, resulting in rocks of intermediate composition (Group II). Similar mixed-magma intrusions exist elsewhere on Mull (e.g. Loch Ba; Sparks, 1988), attesting to the common coexistence of basic and silicic magmas.

It is envisaged that the magma storage reservoir system for the LSSC was not a large chamber, but rather a network of poorly connected sill-like bodies and conduits. Such a system would provide a large surface area-to-volume ratio, allowing extensive partial melting of the upper crust to occur. Turbulent flow within the sill complex would also facilitate crustal assimilation processes.

The LSSC provides an unusual opportunity to study the products and mechanisms behind a wide range of petrogenetic processes, confined to a relatively small geographical area. The magma chemistry, tectonic setting, and thermal conditions of the LSSC are not perceived to have been unusual in any way, and therefore the processes which operated are probably very common. The very fact that a particularly normal suite of basaltic rocks has interacted with the crust in this way should encourage caution when interpreting the contamination history of other continental basaltic magmas where the preservation of crustal xenoliths is not as good, or indeed absent.

	Acknowledgements

 
Thanks go to Malcolm Hole for helpful and constructive comments on a earlier draft of this paper. Ann Morrison and Andrew Kerr are thanked for their thoughtful reviews. Andrew Kerr is also thanked for making his extensive data set on the Mull lavas readily available. We are most grateful to Godfrey Fitton and Dodie James, Department of Geology and Geophysics, University of Edinburgh, for provision of major and trace element data. Nick Walsh of Royal Holloway University, London, kindly made available analytical facilities for the REE. Robert MacDonald provided the technical assistance for the electron probe analysis at the Department of Geology and Applied Geology, University of Glasgow. Anne Kelly and Vincent Gallagher are thanked for their sterling efforts in the radiogenic isotope laboratories at the SURRC. R.J.P. acknowledges with gratitude the receipt of an NERC training award (1992–1995). The isotopic analyses at SURRC were supported by the Scottish Universities.

---

^* Present address: Department of Geology and Petroleum Geology, Meston Building, Kings College, University of Aberdeen, Old Aberdeen AB24 3UE, UK. Telephone:+44 (0)1224 273467 e-mail: j.preston{at}abdn.ac.uk

	References

TOP ABSTRACT Introduction Geological Setting Analytical Techniques Petrography and Mineral... Whole-Rock Major Element... Whole-Rock Trace Element... Whole-Rock Isotope Geochemistry The Petrogenesis of the... Discussion Summary and Conclusions References

 
Albarède F. How deep do common basaltic magmas form and differentiate? Journal of Geophysical Research (1992) 97(B7):10997–11009.

Bailey E. B., Clough C. T., Wright W. B., Richey J. E., Wilson G. V. Tertiary and Post-Tertiary Geology of Mull, Loch Aline, and Oban. Memoirs of the Geological Survey of Scotland (1924) Edinburgh: HMSO.

Barbero L., Villaseca C., Rogers G., Brown P. E. Geochemical and isotopic disequilibrium in crustal melting: an insight from the anatectic granitoids from Toledo, Spain. Journal of Geophysical Research (1995) 100(B8):15745–15765.

Beckinsale R. D., Pankhurst R. J., Skelhorn R. R., Walsh J. N. Geochemistry and petrogenesis of the Early Tertiary lava pile of the Isle of Mull, Scotland. Contributions to Mineralogy and Petrology (1978) 66:415–427.[Web of Science]

Bell B. R. Significance of ferrodioritic liquids in magma mixing processes. Nature (1983) 306:323–327.

Bell B. R., Jolley D. W. Application of palynological data to the chronology of the Paleogene lava fields of the British Province: implications for magmatic stratigraphy. Journal of the Geological Society, London (1997) 154:701–708.[Abstract/Free Full Text]

Bell B. R., Claydon R. V., Rogers G. The petrology and geochemistry of cone-sheets from the Cuillin igneous complex, Isle of Skye: evidence for combined assimilation and fractional crystallization during lithospheric extension. Journal of Petrology (1994) 35:1055–1094.[Abstract/Free Full Text]

Bellieni G., Brotzu P., Comin-Chiaramonti P., Ernesto M., Melfi A., Pacca I. G., Piccirillo E. M. Flood basalt to rhyolite suites in the Southern Paraná Plateau (Brazil): palaeomagnetism, petrogenesis and geodynamic implications. Journal of Petrology (1984) 25:579–618.[Abstract/Free Full Text]

Bender J. F., Hanson G. N., Bence A. E. Cortlandt complex: differentiation and contamination in plutons of alkali basalt affinity. American Journal of Science (1984) 284:1–57.[Abstract/Free Full Text]

Bott M. H. P., Tandrigoda T. A. Interpretation of the gravity and magnetic anomalies over the Mull Tertiary intrusive complex, NW Scotland. Journal of the Geological Society, London (1987) 91:17–28.

Bouch J. E., Hole M. J., Trewin N. H., Morton A. C. Low-temperature aqueous mobility of the rare-earth elements during sandstone diagenesis. Journal of the Geological Society, London (1995) 152:895–898.[Abstract/Free Full Text]

Bowen N. L. The behaviour of inclusions in igneous magmas. Journal of Geology (1922) 30:513–570.[Web of Science]

Bowen N. L. The Evolution of the Igneous Rocks. (1928) Princeton, NJ:: Princeton University Press.

Brooks C. K. The Fe₂O₃/FeO ratio of basalt analyses: an appeal for a standardised procedure. Bulletin of the Geological Society of Denmark (1976) 25:117–120.

Brown G. M. Pyroxenes from the early and middle stages of fractionation of the Skaergaard intrusion, East Greenland. Mineralogical Magazine (1957) 31:511–543.

Bryan W. B., Finger L. W., Chayes F. Estimating proportions in petrologic-mixing equations by least squares approximation. Science (1969) 163:926–927.[Abstract/Free Full Text]

Buist D. S. The composite sill of Rudh' a' Chromain, Carsaig, Mull. Geological Magazine (1961) 98:67–76.[Abstract]

Carmichael I. S. E. The petrology of Thingmuli, a Tertiary volcano in eastern Iceland. Journal of Petrology (1964) 5:435–460.[Abstract/Free Full Text]

Chappell B. W., White A. J. R. I- and S-type granites in the Lachlan Fold Belt. Transactions of the Royal Society of Edinburgh: Earth Sciences (1992) 83:1–26.[Web of Science]

Clayburn J. A. P. The crustal evolution of Central Scotland and the nature of the lower crust: Pb, Nd and Sr isotope evidence from Caledonian granites. Earth and Planetary Science Letters (1988) 90:41–51.[Web of Science]

Cox K. G. A model for flood basalt volcanism. Journal of Petrology (1980) 21:629–650.[Abstract/Free Full Text]

Cox K. G. Numerical modelling of a randomized RTF magma chamber: a comparison with continental flood basalt sequences. Journal of Petrology (1988) 29:681–697.[Abstract/Free Full Text]

Cox K. G., Mitchell C. Importance of crystal settling in the differentiation of Deccan Trap basaltic magmas. Nature (1988) 333:447–449.

Dagley P., Musset A. E., Skelhorn R. R. Polarity, stratigraphy and duration of the Tertiary igneous activity of Mull, Scotland. Journal of the Geological Society, London (1987) 144:985–996.[Abstract/Free Full Text]

DePaolo D. J. Trace element and isotopic effects of combined wallrock assimilation and fractional crystallization. Earth and Planetary Science Letters (1981) 53:189–202.[Web of Science]

Devey C. W., Cox K. G. Relationships between crustal contamination and crystallization in continental flood basalt magmas with special reference to the Deccan Traps of the Western Ghats, India. Earth and Planetary Science Letters (1987) 84:59–68.[Web of Science]

Dickin A. P. Isotope geochemistry of Tertiary igneous rocks from the Isle of Skye, N.W. Scotland. Journal of Petrology (1981) 22:155–189.[Abstract/Free Full Text]

Dickin A. P., Brown J. L., Thompson R. N., Halliday A. N., Morrison M. A. Crustal contamination and the granite problem in the British Tertiary Igneous Province. Philosophical Transactions of the Royal Society of London, Series A (1984) 310:755–780.

Fitton J. G., Dunlop H. M. The Cameroon line, West Africa, and its bearing on the origin of oceanic and continental alkali basalts. Earth and Planetary Science Letters (1985) 72:23–38.[Web of Science]

Fodor R. V., Corwin C., Roisenberg A. Petrology of Serra Geral (Paraná) continental flood basalts, Southern Brazil: crustal contamination, source material, and South Atlantic magmatism. Contributions to Mineralogy and Petrology (1985a) 91:54–65.[Web of Science]

Fodor R. V., Corwin C., Sial A. N. Crustal signatures in the Serra Geral flood-basalt province, southern Brazil: O- and Sr- isotope evidence. Geology (1985b) 13:763–765.[Abstract/Free Full Text]

Fowler M. B. Elemental evidence for crustal contamination of mantle-derived Caledonian syenite by metasediment anatexis and magma mixing. Chemical Geology (1988) 69:1–16.[Web of Science]

Fowler M. B. Elemental and O-Sr-Nd isotope geochemistry of the Glen Dessary syenite, NW Scotland. Journal of the Geological Society, London (1992) 149:209–220.[Abstract/Free Full Text]

Francis D. The pyroxene paradox in MORB glasses—a signature of picritic parental magmas? Nature (1986) 319:586–589.

Gardien V., Thompson A. B., Grujic D, Ulmer P. Experimental melting of biotite + plagioclase + quartz ± muscovite assemblages and the implication for crustal melting. Journal of Geophysical Research (1995) 100(B8):15581–15591.

Gibb F. G. F. The zoned clinopyroxenes of the Shiant Isles sill, Scotland. Journal of Petrology (1973) 14:203–230.[Abstract/Free Full Text]

Gibson S. A. The geochemistry of the Trotternish sills, Isle of Skye: crustal contamination in the British Tertiary Igneous Province. Journal of the Geological Society, London (1990) 147:1071–1081.[Abstract/Free Full Text]

Gibson S. A., Thompson R. N., Dickin A. P., Leonardos O. H. High-Ti and low-Ti mafic potassic magmas: key to plume–lithosphere interactions and continental flood-basalt genesis. Earth and Planetary Science Letters (1995) 136:149–165.[Web of Science]

Grove T. L., Baker M. B. Phase equilibrium controls on the tholeiitic versus calc-alkaline differentiation trends. Journal of Geophysical Research (1984) 89(B5):3253–3274.

Grove T. L., Gerlach D. C., Sando T. W. Origin of calc-alkaline series lavas at Medicine Lake volcano by fractionation, assimilation and mixing. Contributions to Mineralogy and Petrology (1982) 80:160–182.[Web of Science]

Hart S. R. Heterogeneous mantle domains: signatures, genesis and mixing chronologies. Earth and Planetary Science Letters (1988) 90:523–528.

Holden P., Halliday A. N., Stephens W. E. Neodymium and strontium isotope content of microdiorite enclaves points to mantle input into granitoid production. Nature (1987) 330:53–56.

Holland J. G., Brown G. M. Hebridean tholeiitic magmas: a geochemical study of the Ardnamurchan cone-sheets. Contributions to Mineralogy and Petrology (1972) 37:139–160.[Web of Science]

Hunter R. H., Sparks R. S. J. The differentiation of the Skaergaard Intrusion. Contributions to Mineralogy and Petrology (1987) 451:461.

Huppert H. E., Sparks R. S. J. Chilled margins in igneous rocks. Earth and Planetary Science Letters (1980) 92:397–405.

Huppert H. E., Sparks R. S. J. Cooling and contamination of mafic and ultramafic magmas during ascent through continental crust. Earth and Planetary Science Letters (1985) 74:371–386.[Web of Science]

Irvine T. N. Chromian spinels as a petrological indicator. Part 2. Petrologic applications. Canadian Journal of Earth Sciences (1967) 4:71–103.

Janouek V., Rogers G., Bowes D. R. Sr–Nd isotopic constraints on the petrogenesis of the Central Bohemian Pluton, Czech Republic. Geologische Rundschau (1995) 84:520–534.

Kerr A. C. Elemental evidence for an enriched small-fraction melt input into Tertiary Mull basalts, Western Scotland. Journal of the Geological Society, London (1993) 150:763–769.[Abstract/Free Full Text]

Kerr A. C. Lithospheric thinning during the evolution of a continental large igneous province: a case study from the North Atlantic Tertiary province. Geology (1994) 22:1027–1030.[Abstract/Free Full Text]

Kerr A. C. The geochemistry of the Mull-Morvern Tertiary lava succession, NW Scotland: an assessment of mantle sources during plume-related volcanism. Chemical Geology (1995a) 122:975–978.

Kerr A. C. The geochemical stratigraphy, field relations and temporal variation of the Mull-Morvern Tertiary lava succession, NW Scotland. Transactions of the Royal Society of Edinburgh, Earth Sciences (1995b) 86:35–47.[Web of Science]

Kerr A. C., Kempton P. D., Thompson R. N. Crustal assimilation during turbulent magma ascent (ATA); new isotopic evidence from the Mull Tertiary lava succession, NW Scotland. Contributions to Mineralogy and Petrology (1994) 119:142–154.[Web of Science]

Kille I. C., Thompson R. N., Morrison M. A., Thompson R. F. Field evidence for turbulence during flow of basaltic magma through conduits from South-West Mull. Geological Magazine (1986) 123:693–697.[Abstract]

Langmuir C. H., Vocke Jr, R. D., Hanson G. N. A general mixing equation with applications to Icelandic basalts. Earth and Planetary Science Letters (1978) 37:393–400.[Web of Science]

Le Bas M. J., Le Maitre R. E., 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]

MacRae N. D., Nesbitt H. W. Partial melting of common metasedimentary rocks: a mass balance approach. Contributions to Mineralogy and Petrology (1980) 75:21–26.[Web of Science]

Mantovani M. S. M., Marques L. S., De Sousa M. A., Civetta L., Atalla L., Innocenti F. Trace element and strontium isotope constraints on the origin and evolution of Paraná continental flood basalts of Santa Catarina State (Southern Brazil). Contributions to Mineralogy and Petrology (1985) 26:187–209.

Mattey D. P., Gibson I. L., Marriner C. F., Thompson R. N. The diagnostic geochemistry, relative abundance, and spatial distribution of high-calcium, low-alkali olivine tholeiite dykes in the Lower Tertiary regional swarm of the Isle of Skye, NW Scotland. Mineralogical Magazine (1977) 41:273–285.[Web of Science]

Middlemost E. A. K. The basalt clan. Earth-Science Reviews (1975) 11:337–364.

Moorbath S., Thompson R. N. Strontium isotope geochemistry and petrogenesis of the Early Tertiary lava pile of the Isle of Skye, Scotland, and other basic rocks of the British Tertiary Province: an example of magma–crust interaction. Journal of Petrology (1980) 21:295–321.[Abstract/Free Full Text]

Morrison M. A. The use of immobile trace elements to distinguish palaeotectonic affinities of metabasalts: applications to the Palaeocene basalts of Mull and Skye, NW Scotland. Earth and Planetary Science Letters (1978) 39:407–416.[Web of Science]

Morrison M. A. Igneous and metamorphic geochemistry of the Mull lavas (1979) University of London. Ph.D. Thesis.

Morrison M. A., Thompson R. N., Dickin A. P. Geochemical evidence for complex magmatic plumbing during development of a continental volcanic centre. Geology (1985) 13:581–584.[Abstract/Free Full Text]

Mussett A. E. ⁴⁰Ar–³⁹Ar step-heating ages of the Tertiary igneous rocks of Mull, Scotland. Journal of the Geological Society, London (1986) 143:887–896.[Abstract/Free Full Text]

Nakamura N. Determination of REE, Ba, Fe, Mg, Na and K in carbonaceous and ordinary chondrites. Geochimica et Cosmochimica Acta (1974) 38:757–773.[Web of Science]

O'Hara M. J. Geochemical evolution during fractional crystallization of a periodically refilled magma chamber. Nature (1977) 266:503–507.

O'Hara M. J., Mathews R. E. Geochemical evolution in an advancing, periodically replenished, periodically tapped, continuously fractionating magma chamber. Journal of the Geological Society, London (1981) 138:237–277.[Abstract/Free Full Text]

Pankhurst R. J., Walsh J. N., Beckinsale R. D., Skelhorn R. R. Isotopic and other geochemical evidence for the origin of the Loch Uisg granophyre, Isle of Mull, Scotland. Earth and Planetary Science Letters (1978) 38:355–363.[Web of Science]

Patiño-Douce A. E. P. Experimental generation of hybrid silicic melts by reaction of high-Al basalt with metamorphic rocks. Journal of Geophysical Research (1995) 100(B8):15623–15639.

Preston R. J. The petrogenesis of the Loch Scridain Xenolithic Sill Complex, Isle of Mull. (1996) University of Glasgow. Ph.D. Thesis.

Preston R. J., Bell B. R. Cognate gabbroic xenoliths from a tholeiitic subvolcanic sill complex: implications for fractional crystallization and crustal contamination processes. Mineralogical Magazine (1997) 61:329–349.[Abstract]

Reiners P. W., Nelson B. K., Ghiorso M. S. Assimilation of felsic crust by basaltic magma: thermal limits and extents of crustal contamination of mantle-derived magmas. Geology (1995) 23:563–566.[Abstract/Free Full Text]

Reiners P. W., Nelson B. K., Nelson S. W. Evidence for multiple mechanisms of crustal contamination of magma from compositionally zoned plutons and associated ultramafic intrusions of the Alaska Range. Journal of Petrology (1996) 37:261–292.[Abstract/Free Full Text]

Ridley W. I. The petrology of volcanic rocks from the Small Isles of Inverness-shire. Report of the Institute of Geological Sciences, 73/10. (1973).

Rollinson H. R. Using Geochemical Data: Evaluation, Presentation, Interpretation (1993) New York: John Wiley.

Shibata T., DeLong S. E., Walker D. Abyssal tholeiites from the Oceanographer fracture zone I. Petrology and fractionation. Contributions to Mineralogy and Petrology (1979) 70:89–102.[Web of Science]

Sisson T. W., Grove T. L. Experimental investigations of the role of H₂O in calc-alkaline diVerentiation and subduction zone magmatism. Contributions to Mineralogy and Petrology (1993) 113:143–166.[Web of Science]

Smith D. G. W., Lindsley D. H. Stable and metastable augite crystallization trends in a single basalt flow. American Mineralogist (1971) 56:225–233.[Web of Science]

Sparks R. S. J. Petrology and geochemistry of the Loch Ba ring dyke, Mull (NW Scotland): an example of the extreme differentiation of tholeiitic magmas. Contributions to Mineralogy and Petrology (1988) 100:446–461.[Web of Science]

Stamatelopoulous-Seymour K., Vlassopoulos D. Magma mixing at Nisyros volcano, as inferred from incompatible trace-element systematics. Journal of Volcanology and Geothermal Research (1992) 50:273–299.[Web of Science]

Thirlwall M. F., Jones N. W. Isotope geochemistry and contamination mechanics of Tertiary lavas from Skye, Northwest Scotland. In: Continental Basalts and Mantle Xenoliths—Hawkesworth C. J., Norry M. J., eds. (1983) Nantwich, UK: Shiva. 186–250.

Thomas H. H. On certain xenolithic Tertiary minor intrusions in the Island of Mull (Argyllshire). Quaternary Journal of the Geological Society (London) (1922) 78:229–260.

Thompson R. N. Thermal aspects of the origin of the Hebridean Tertiary acid magmas. 1. An experimental study of partial fusion of Lewisian gneisses and Torridonian sediments. Mineralogical Magazine (1981) 44:161–170.[Web of Science]

Thompson R. N. Geochemistry and magma genesis. In: The Igneous Rocks of the British Isles—Sutherland D. S., ed. (1982) London: John Wiley. 461–477.

Thompson R. N., Esson J., Dunham A. C. Major element chemical variation in the Eocene lavas of the Isle of Skye, Scotland. Journal of Petrology (1972) 13:219–253.[Abstract/Free Full Text]

Thompson R. N., Gibson I. L., Marriner G. F., Mattey D. P., Morrison M. A. Trace-element evidence of multi-stage fusion and polybaric fractional crystallization in the Palaeocene lavas of Skye, NW Scotland. Journal of Petrology (1980) 21:265–293.[Abstract/Free Full Text]

Thompson R. N., Dickin A. P., Gibson I. L., Morrison M. A. Elemental fingerprints of isotopic contamination of Hebridean Palaeocene mantle-derived magmas by Archaean sial. Contributions to Mineralogy and Petrology (1982) 79:159–168.[Web of Science]

Thompson R. N., Morrison M. A., Dickin A. P., Gibson I. L., Harmon R. S. Two contrasting styles of interaction between basic magmas and continental crust in the British Tertiary Igneous Province. Journal of Geophysical Research (1986) 91(B6):5985–5997.

Tilley C. E., Muir I. D. The Hebridean plateau magma type. Transactions of the Edinburgh Geological Society (1962) 19:208–215.

Tuttle O. F., Bowen N. L. Origin of granite in the light of experimental studies in the system NaAlSi₃O₈–KAlSi₃O₈–SiO₂–H₂O. Geological Society of America, Memoirs (1958) 74.

Van der Laan S. R., Wyllie P. J. Experimental interaction of granitic and basaltic magmas and implications for mafic enclaves. Journal of Petrology (1993) 34:491–517.[Abstract/Free Full Text]

Vielzeuf D., Holloway R. J. Experimental determination of the fluid-absent melting relations in the pelitic system—consequences for crustal differentiation. Contributions to Mineralogy and Petrology (1988) 98:257–276.[Web of Science]

Wager L. R., Brown G. M. Layered Igneous Rocks (1968) Edinburgh: Oliver & Boyd.

Walker D., Shibata T., DeLong S. E. Abyssal tholeiites from the Oceanographer fracture zone II. Phase equilibria and mixing. Contributions to Mineralogy and Petrology (1979) 70:111–125.[Web of Science]

Walker G. P. L. The distribution of amygdale minerals in Mull and Morvern (western Scotland). In: Studies in Earth Sciences—Murty T. V., Rao S. S., eds. (1970) Faridabad: University of Saugar. 181–194. West Commemorative Volume.

Walsh J. N., Beckinsale R. D., Skelhorn R. R., Thorpe R. S. Geochemistry and petrogenesis of Tertiary granitic rocks from the Isle of Mull, Northwest Scotland. Contributions to Mineralogy and Petrology (1979) 71:99–116.[Web of Science]

Walsh J. N., Buckley F., Barker J. The simultaneous determination of the rare earth elements in rocks using inductively coupled plasma source spectrometry. Chemical Geology (1981) 33:141–153.[Web of Science]

Weaver B. L., Tarney J. Rare earth geochemistry of Lewisian granulite-facies gneisses, northwest Scotland: implications for the petrogenesis of the Archaean lower crust. Earth and Planetary Science Letters (1980) 51:279–296.[Web of Science]

Weaver B. L., Tarney J. Lewisian gneiss geochemistry and Archaean crustal development models. Earth and Planetary Science Letters (1981) 55:171–180.[Web of Science]

Whitehouse M. J. Isotopic evolution of the southern Outer Hebridean Lewisian gneiss complex: constraints on Late Archaean source regions and the generation of transposed Pb–Pb palaeoisochrons. Chemical Geology (1990) 86:1–20.[Web of Science]

Whitehouse M. J., Robertson C. J. Isotopic evolution of the Lewisian Complex of Tiree, Inner Hebrides and correlation with the mainland. Scottish Journal of Geology (1995) 31:131–137.[Web of Science]

Whitney J. A. The origin of granite: the role and source of water in the evolution of granitic magmas. Geological Society of America Bulletin (1988) 100:1886–1897.[Abstract/Free Full Text]

Wilson M. Igneous Petrogenesis (1989) London: Unwin Hyman.

Wood B. J., Fraser D. G. Elementary Thermodynamics for Geologists (1976) Oxford: Oxford University Press.

Yoder H. S., Tilley C. E. Origin of basalt magmas: an experimental study of natural and synthetic rock systems. Journal of Petrology (1962) 3:342–532.[Abstract/Free Full Text]

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

This article has been cited by other articles:

				A. C. Kerr, R. W. Kent, B. A. Thomson, J. K. Seedhouse, and C. H. Donaldson Geochemical Evolution of the Tertiary Mull Volcano, Western Scotland J. Petrology, June 1, 1999; 40(6): 873 - 908. [Abstract] [Full Text] [PDF]

				R. J. Preston, T. J. Dempster, B. R. Bell, and G. Rogers The Petrology of Mullite-bearing Peraluminous Xenoliths: Implications for Contamination Processes in Basaltic Magmas J. Petrology, April 1, 1999; 40(4): 549 - 573. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) 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 Preston, R. J. Articles by Rogers, G. Search for Related Content Social Bookmarking          What's this?

Online ISSN 1460-2415 - Print ISSN 0022-3530

Copyright © 2009 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