Journal of Petrology | Volume 40 | Number 3 | Pages 423-450 | 1999
© Oxford University Press 1999
The Geochemistry and Petrogenesis of the Agnew Intrusion, Canada: A Product of S-Undersaturated, High-Al and Low-Ti Tholeiitic Magmas
1 Department of Earth Sciences, Laurentian University Sudbury, Ont. P3E 2C6, CANADA
2 School of Earth Sciences, University of Melbourne Parkville, VIC. 3052, AUSTRALIA
Received July 29, 1997; Revised typescript accepted August 10, 1998
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
The 2100 m thick Agnew Intrusion (50 km2) in central Ontario, Canada, is a deformed, Palaeoproterozoic, layered leucogabbronoritic to gabbronoritic pluton that is believed to have intruded as a sub-volcanic sill between Archaean granitic basement of the Superior Province and overlying Palaeoproterozoic flood basalts. Its emplacement was part of a major magmatic event in the region, which included the extensive Hearst–Matachewan dyke swarm, and was followed by rifting and accumulation of the thick Huronian Supergroup succession in the Southern Province. Litho- and chemostratigraphic analyses of the Agnew Intrusion show that it is the product of at least three major magma pulses, giving rise sequentially to a Marginal, Lower, and Upper Series. The final and largest magma pulse produced a closed-system differentiated sequence grading from olivine gabbronorites at the base to ferrosyenites and alkali-feldspar granites at the top. Parental magmas of the Agnew Intrusion were S-undersaturated, high-Al and low-Ti tholeiites, exhibiting some minor and chalcophile element affinities with boninites. These magmas have major element compositions that are very similar to the model parent liquids proposed for the mafic portions of the Stillwater and Bushveld Complexes. Other mafic dyke groups that are spatially and temporally associated with the Agnew Intrusion have strong petrological and geochemical similarities with the Hearst–Matachewan dyke swarm, but are not comagmatic with the intrusion. Possible mantle sources to the Agnew Intrusion include the mantle residue after partial melting to form the Archaean greenstone sequences, and plagioclase-bearing mafic or ultramafic intrusions that have ponded at the crust–mantle boundary during the Archaean. Partial melting in these mantle sources may have been induced by ‘thermal’ plumes.
KEY WORDS: Agnew Intrusion; geochemistry; mantle source; parental magma; petrogenesis
|
Introduction |
|---|
|
TOPABSTRACTIntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
The Agnew Intrusion (
50 km2 in outcrop area) is located 70 km west of Sudbury adjacent to the boundary between the Archaean Superior Province and Palaeoproterozoic Southern Province in central Ontario, Canada (Fig. 1). It is the best exposed member of several leucogabbronoritic to gabbronoritic layered intrusions that belong to the East Bull Lake suite (Bennett et al., 1991
). Other layered intrusions of the suite include the East Bull Lake, River Valley, May Township, Drury Township, Wisner Township, and Falconbridge Township Intrusions (Fig. 1; James & Harris, 1977; James & Born, 1985; Ashwal & Wooden, 1989; McCrank et al., 1989; Peck et al., 1993a, 1995; Prevec, 1993; Vogel et al., 1998a). The best estimate of the age of the Agnew Intrusion is a U–Pb zircon age of 2491 ± 5 Ma from a granophyric rock taken from the stratigraphic top of the intrusion (Krogh et al., 1984). Available age data for the other East Bull Lake suite intrusions (Fig. 1) suggest that all were emplaced coevally, and were part of a major magmatic event that included intrusive emplacement of the laterally extensive Hearst–Matachewan dyke swarm and some small granitic batholiths (Murray and Creighton plutons), as well as extrusion of the Elliot Lake Group bimodal continental flood basalt sequence. This magmatic event coincided with incipient rifting in the region of the Archaean proto-continent and subsequent development of the Palaeoproterozoic Huronian rift zone (James et al., 1994), which is now manifested in the 
12 km thick volcanic–sedimentary rock succession of the Huronian Supergroup in the Southern Province (Zolnai et al., 1984).
|
The Agnew Intrusion and the East Bull Lake Intrusion are known to host significant PGE–Cu–Ni mineralization in marginal rock units (Peck & James, 1990
; Peck et al., 1993a, 1993b, 1995; Vogel et al., 1997). It has been postulated that some of the massive Cu–Ni sulphide ores at the base of the 1850 Ma, meteorite impact-induced Sudbury Igneous Complex may in fact represent proto-ore associated with older crustal rocks such as the East Bull Lake suite (Pattison, 1979; Innes & Colvine, 1984; Peck et al., 1993a; Golightly, 1994; James et al., 1994; Keays et al., 1995). Therefore, characterizing these intrusions may provide a better understanding of the genesis of the Sudbury Igneous Complex and its ores, as well as of other dynamic sytems that exploit pre-existing crustal rocks and their sulphides. This paper presents whole-rock major element, trace element, rare earth element (REE), and platinum-group element (PGE) data for the entire range of exposed rock types within the Agnew Intrusion. In the absence of suitable chilled margin exposures, dykes proximal to the intrusion are assessed on the basis of field relationships and geochemistry as to their possible role as parental magmas to the intrusion. These data are fundamental to understanding the petrogenetic evolution of the intrusion and the nature of its mantle source. The magmatic relationship between the Agnew Intrusion and the temporally associated Hearst–Matachewan dyke swarm is also considered.
|
The Geology of the Agnew Intrusion |
|---|
|
TOPABSTRACTIntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
Comprehensive geological summaries of the central Ontario region have been given by Card (1978)
, Zolnai et al. (1984), Bennett et al. (1991), Dressler et al. (1991) and Roscoe & Card (1992). Preliminary descriptions of the geology of the Agnew Intrusion and its immediate environs were provided by Card & Palonen (1976) and James & Harris (1977). More detailed accounts, based on 8 months of field investigations, have been presented by Vogel (1996) and Vogel et al. (1998a). Figures 2 and 3 illustrate the results of the last two studies, and a brief summary is given below.
|
|
Geological setting
The Agnew Intrusion and the other East Bull Lake suite layered intrusions lie at the base of the Palaeoproterozoic Huronian Supergroup in the Southern Province, and immediately overlie granitic rocks and orthogneisses of the Archaean Ramsey–Algoma granitoid suite (2710–2665 Ma; Prevec, 1993
) in the southern part of the Superior Province (Fig. 1). Intrusive relationships with the granitic rocks are typically obscured by pseudotachylitic breccias associated with the Sudbury impact event (Chubb et al., 1994) and by later mafic dyke–sill(?) emplacement along the intrusion–footwall contact. The top of the intrusion is disconformably and locally unconformably overlain at its eastern margin by a thick Huronian Supergroup sedimentary sequence. At various other locations within the Southern Province, East Bull Lake suite layered intrusions are overlain by the thick (up to 2600 m), bimodal volcanic succession of the Elliot Lake Group (Fig. 1; Card et al., 1977; Innes, 1978; Card & Jackson, 1995). The lower mafic volcanic component of the Elliot Lake Group has been intruded by the 2477 ± 9 Ma (Krogh et al., 1996) Murray granite pluton near Sudbury (Bennett et al., 1991), suggesting that these mafic volcanic rocks are probably of similar age to the 2491–2441 Ma East Bull Lake layered intrusion suite. It has been postulated that the present outcrop distribution of the Elliot Lake Group volcanic rocks is only a remnant of a much more extensive continental flood basalt plain that covered a large portion of the southern Superior Province in central Ontario prior to erosion and deposition of the continental rift-related Huronian Supergroup sedimentary sequence (Vogel et al., 1998a).
The distribution of the stratigraphic units and igneous layering attitudes indicate that the Agnew Intrusion forms a synclinal structure that plunges shallowly at 25° to the ENE (Fig. 2b). It has been argued that this structural geometry is a product of post-emplacement ductile deformation associated with the 1850 Ma Penokean Orogeny, and is not a primary igneous feature (Vogel et al., 1998a). The regional structure and lithological distribution is best explained by a modified ‘dome-and-basin’ structural interference pattern. The predicted pre-deformational geometry of the Agnew Intrusion is believed to have been a near-horizontal sill that probably now extends underneath the Huronian Supergroup sediments to the east. This interpretation is supported by the location of a regional-scale, positive Bouguer gravity anomaly whose centre lies under Agnew Lake immediately east of the intrusion (Fig. 2b; Popelar, 1971; Gupta et al., 1984; Gupta, 1991a). Judging from present exposures, the original horizontal dimension of the Agnew Intrusion layered sill would have exceeded 20 km with a maximum vertical thickness of 2100 m. Excellent stratigraphic correlations between the Agnew and neighbouring East Bull Lake Intrusions suggest that present exposures of the East Bull Lake suite may represent erosional remnants of one or more much larger mafic sills emplaced at the base of the coeval Elliot Lake Group continental flood basalt sequence (Vogel et al., 1998a). Examples of laterally extensive and continuous sills of this kind (>1000 km2) are common in Antarctica (Marsh, 1989).
Igneous stratigraphy
The Agnew Intrusion is stratigraphically subdivided into three series—Marginal, Lower and Upper Series (Fig. 3)—each being separated by a break during which there was no magmatic injection into the chamber. The Marginal Series is 200 m thick and is predominantly composed of vari-textured leucogabbronorites (Marginal Leucogabbronorite Zone). The zone features a broad-scale gradation from gabbronorites and lesser melanogabbronorites at the base (constituting 20% of the zone), through the main leucogabbronorite sequence (79%) with minor anorthosite, to local granophyric bands near its top (<1%). Marginal Leucogabbronorite Zone rocks contain many small inclusions of granite, massive quartz, and ultramafic rocks (generally <50 cm in diameter). The ultramafic inclusions are typically angular, suggesting that they have not travelled far from their source and may represent fragments of unexposed ultramafic rock cogenetically related to the Agnew Intrusion. The granitic and massive quartz inclusions were probably derived from the immediate footwall to the intrusion and commonly have corroded and recrystallized margins, indicating that Marginal Series magmas have undergone localized in situ crustal assimilation.
The Marginal Leucogabbronorite Zone is locally separated from the Archaean granite basement by the Marginal Gabbronorite Zone, which consists largely of massive, medium-grained gabbronorite in the southwestern part of the intrusion, and 10–20 m wide, contact-parallel diabase dykes–sills(?) in other locations (Fig. 2a). These rocks appear to be the crystallized products of later magmas that have intruded along the base of the intrusion and are therefore younger than the overlying Marginal Leucogabbronorite Zone. Magmatic features in the granitic footwall that are probably related to the emplacement of the Agnew Intrusion include back-intrusive felsic net-vein textures and rare felsic magmatic breccias (Vogel et al., 1998a).
The Lower Series has a maximum thickness of 550 m and is dominated by gabbronorites. Lower Series magmas intruded and disrupted the Marginal Leucogabbronorite Zone in the northwest corner of the intrusion where the projected position of the WNW-striking Streich Dyke would intersect this zone (Fig. 2a). At this location, large outcrop-sized remnants of the Marginal Leucogabbronorite Zone are preserved within the Inclusion-bearing Gabbronorite Zone (Fig. 2a), which is a compositionally and texturally heterogeneous unit containing ubiquitous footwall, ultramafic, and leucogabbronoritic inclusions. The heterogeneous nature and inclusion abundance in the zone decrease with increasing stratigraphic height, giving rise to the overlying homogeneous Lower Gabbronorite Zone, which extends laterally away from the intrusive site. Mesoscale, modal layering of gabbronorite and leucogabbronorite in the Agnew Intrusion is first developed in the Lower Layered Unit of this zone; layers are discontinuous laterally and often fade in and out in vertical section. The contact between the Lower Layered Unit and the underlying homogeneous Massive Unit is an irregular, non-planar surface.
Upper Series rocks constitute over half of the entire Agnew stratigraphic sequence and have a maximum thickness of 1350 m. The Upper Series has been subdivided into three zones—Olivine Gabbronorite Zone, Upper Gabbronorite Zone, and an uppermost Fe–Ti Oxide Zone (Fig. 3). The Olivine Gabbronorite Zone at the base of the Upper Series is a poorly exposed, well-layered, 50 m thick interval separating the texturally similar Lower and Upper Layered Units (Fig. 3). Layering in the Olivine Gabbronorite Zone is characterized by alternating, isomodal, 20 cm thick layers of olivine gabbronorite, leucogabbronorite, and minor olivinemelanogabbronorite. This zone has also been recognized at the equivalent stratigraphic level within the neighbouring East Bull Lake Intrusion (Peck et al., 1993a).
Within the Upper Gabbronorite Zone, the Upper Layered Unit is overlain locally by the Mixed Unit, which is a lithologically and texturally chaotic rock interval (locally with mafic dendritic textures; see below) containing irregularly distributed granitic, gabbronoritic, and diabase inclusions. The presence of these inclusions suggests that the Mixed Unit is the product of a separate magma pulse. The main part of the Upper Gabbronorite Zone consists of the Porphyritic Unit (Fig. 3), which is characterized by a variable abundance of plagioclase phenocrysts and glomerophenocrysts. Gabbronorites are dominant over leucogabbronorites in an approximate volume ratio of 70:30. Most of the Porphyritic Unit features diffuse, macrorhythmic decametre-scale layering of gabbronorite and leucogabbronorite; centimetre-scale layering is prominent in its basal and upper parts. The Pod-bearing Unit occurs within the lower stratigraphic parts of the Porphyritic Unit (Fig. 3) and is distinguished by the presence of rounded pods (<1 m in diameter) of porphyritic leucogabbronorite and granophyre set within a porphyritic gabbronorite host rock. The pods are of local derivation and are characterized by diffuse boundaries. They are believed to have formed as a result of late-stage slumping and magmatic deformation within the crystal pile. The Transition Unit at the contact between the Porphyritic Unit and the overlying Fe–Ti Oxide Zone rocks is characterized by large-scale intermingling and interdigitating gabbronorites and leucogabbronorites derived from both the overlying and underlying stratigraphic units (Vogel, 1996; Vogel et al., 1998a). The unit is interpreted to be the product of late-stage, large-scale slumping of the crystal pile.
The uppermost rocks of the Agnew Intrusion belong to the Fe–Ti Oxide Zone (Fig. 3). Field data suggest that parts of the Fe–Ti Oxide Zone and any pre-existing overlying strata were eroded before deposition of the overlying Matinenda Formation conglomerates of the Huronian Supergroup. The Leucogabbro Unit is composed of massive, coarse-grained leucogabbro, often containing large, altered titanomagnetite crystals 2–3 cm in size. The overlying Ferrosyenite Unit features an upward lithological gradation from dark ferrosyenite to light alkali-feldspar granite. Contact relationships between the Leucogabbro and Ferrosyenite Units are not exposed. All rock types recognized in the Upper Gabbronorite and Fe–Ti Oxide Zone succession, except those from the Ferrosyenite Unit, have been observed locally in gradational vertical sequence on an outcrop-scale within the Upper Series, suggesting that they are probably a comagmatic differentiation sequence.
Highly vari-textured and pegmatitic gabbronorites and leucogabbronorites of the Dendrite Unit occur as largely conformable bands of variable thickness (10–75 m) on either side of the Lower Series–Upper Series boundary (Fig. 3). They are interpreted as the products of late intrusive, volatile-bearing magma pulses (Peck et al., 1993a; Vogel, 1996). The most striking feature within this unit is the common presence of large, curved and branching dendrites that are up to 30 cm long (Vogel et al., 1998a). Each individual dendrite is a sheath-like aggregate of smaller amphibole crystals that have replaced original pyroxene. Individual dendrite-bearing bands commonly contain ultramafic and granitic inclusions near their base, and lenses of granophyre along their upper contacts.
Mineralogy and order of crystallization
Upper greenschist to lower amphibolite facies metamorphism associated with the Penokean Orogeny has variably modified the igneous mineralogy of the Agnew Intrusion, but igneous textures are generally preserved. Plagioclase typically preserves igneous compositions (dominantly labradorite to bytownite), except in upper parts of the intrusion where it has often been recrystallized to fine-grained oligoclase. The primary mafic minerals, olivine and pyroxene, have been replaced pseudomorphically by calcic amphibole, and titanomagnetite has been altered to biotite, titanite and leucoxene, often preserving a relict herringbone pattern. The metamorphic mineral assemblage and phase compositions in Agnew Intrusion rocks are strongly influenced by the original igneous whole-rock compositions.
On the basis of detailed petrography, igneous textural relationships, CIPW-normative compositions and tetrahedron projections in the OL–CPX–PLAG–QTZ system (Irvine, 1970), the general primary crystallization order in the Agnew Intrusion is interpreted to be:plagioclase, olivine, orthopyroxene, clinopyroxene, and titanomagnetite. The distributions and textures of the original mineral phases in the Agnew Intrusion [according to the textural terminology of McBirney & Hunter (1995)] are shown in Fig. 3.
Plagioclase
Plagioclase is the only mineral present as a primocryst phase throughout the stratigraphic sequence and is estimated to constitute 60% of the total exposed volume of the intrusion. Vogel (1996) showed that calculated CIPW-normative An compositions provide an excellent estimate of the average primary magmatic An values for rocks from the Agnew Intrusion. As a good stratigraphic coverage of microprobe data for unaltered plagioclase could not be obtained, we have utilized whole-rock CIPW-normative An data for 145 surface samples taken from throughout the intrusion (Fig. 3). Given that plagioclase is the first mineral to crystallize, these data can be effectively used to distinguish whole-rock differentiation trends.
The plagioclase-rich Marginal Series rocks have a wide normative compositional range from An75 to An50. In much of the Lower Series succession, samples have a more uniform normative composition of An70 decreasing to An60 in the Lower Layered Unit. Dendrite Unit rocks also have An60 compositions. The Olivine Gabbronorite Zone at the base of the overlying Upper Series is relatively rich in total primocrysts (up to 90% plagioclase + olivine ± orthopyroxene) and yields significantly higher values (An91–69), indicating the influx of a new magma pulse at this stratigraphic level. Normative plagioclase compositions decrease gradationally to An60 within the Upper Layered Unit and do not generally exceed this composition throughout the remaining Upper Series sequence. A marked decrease in An content to albite compositions occurs above the Porphyritic Unit, consistent with closed-system fractionation.
Analysed plagioclase compositions (using a Cameca SX50 electron microprobe) for the Marginal and Lower Series rocks of the Agnew Intrusion indicate that cores of zoned primocryst plagioclase crystals are characterized by either An79–69 or An64–51 compositions. In examples where the core has an An79–69 composition, the An64–51 composition is most often present as the surrounding rim. Core–rim boundaries are marked by sharp drops of 5–19% An, and are commonly irregular, indicating corrosion and/or resorption of the core before rim crystallization. Such features are interpreted as the product of plagioclase crystallization at varying pressures (Bowen, 1913; Carr, 1954; Vance, 1965; Wiebe, 1968; Smith & Lofgren, 1983), indicating that some parental magmas to the intrusion entered the Agnew chamber containing small, intratelluric plagioclase crystals. The volumetric proportion of plagioclase with An79–69 compositions in gabbronorites of the Lower Series is 10%, but reaches 50% in some leucogabbronorites of the Marginal Series. Limited microprobe data for plagioclase from the Upper Series suggest that they contain very few plagioclase primocrysts with compositions of 
An70. However, analytical data for plagioclase from the Olivine Gabbronorite Zone in the neighbouring East Bull Lake Intrusion indicate that higher values (An77–69; Peck et al., 1995) are fairly common. Preserved igneous plagioclase phenocrysts within the uppermost Ferrosyenite Unit have albite (<An3) compositions, consistent with whole-rock CIPW-normative data.
Olivine
Pseudomorphs of amphibole after olivine occur as primocryst mineral phases in the Marginal and Lower Series rocks of the Agnew Intrusion; they are absent from the Dendrite Unit and much of the Upper Series (Fig. 3). Olivine rarely accounts for more than 10% of any given rock, except in the primocryst-rich Olivine Gabbronorite Zone where it may have reached 25%. Unaltered olivine crystals in parts of the East Bull Lake Intrusion, specifically in the Rhythmically-layered and Olivine Gabbronorite Zones (equivalent to the Lower Layered Unit and Olivine Gabbronorite Zone in the Agnew Intrusion), have unzoned to weakly zoned Fo65–59 and Fo72–65 compositions, respectively (Peck et al., 1995; Vogel et al., 1998a). This upward compositional change in olivine supports the view that the base of the Olivine Gabbronorite Zone is the level of a new magma pulse injection that is correlative across both intrusions.
Pyroxene
Amphibole pseudomorphs after pyroxene are the most common mafic component in Agnew Intrusion rocks, except in rocks from the Fe–Ti Oxide Zone at the top of the intrusion where titanomagnetite and magnetite predominate. In the Marginal and Lower Series, pyroxenes generally occur interstitially and/or as oikocrysts enclosing plagioclase and olivine. Within the Upper Series, pyroxene is a common primocryst mineral phase (Fig. 3). Pyroxene also crystallized as primocrysts in the late-stage magma pulses that produced the Dendrite Unit bands, and occasionally within the Olivine Gabbronorite Zone. Unaltered pyroxene primocrysts in the Olivine Gabbronorite Zone of the East Bull Lake Intrusion are hypersthenes (En70Fs26Wo4–En67Fs29Wo4), whereas interstitial pyroxene crystallized as augite (En43Fs14Wo43–En41Fs15Wo44; Vogel et al., 1998a) that exsolved hypersthene upon cooling. Hypersthene and augite are also recognized within the underlying Rhythmically-layered Zone of the East Bull Lake Intrusion, but with slightly more evolved mineral compositions (Vogel et al., 1998a), consistent again with an injection of new magma at this stratigraphic level. An uncommon metamorphic mineral assemblage in the Massive Unit (Lower Series) of the Agnew Intrusion consists of Ca-rich actinolitic hornblende lamellae within a Ca-poor chlorite host, and suggests that some pyroxenes in the intrusion also crystallized as pigeonites that later inverted upon cooling. The amphibole minerals that replace pyroxenes increase in Fe content with increasing stratigraphic height in the upper half of the Upper Series from Fe/Mg = 0.4 in actinolite to 1.5 in ferro-tschermakite, consistent with a whole-rock Fe-enrichment trend in the Upper Series.
Accessory minerals
Titanomagnetite, quartz, and apatite usually occur as fine-grained accessory and interstitial igneous mineral phases within the Agnew Intrusion (Fig. 3). In the Leucogabbro Unit of the Fe–Ti Oxide Zone, the uniform grain size (2–3 cm) and high modal abundance (up to 40 vol. %) of original titanomagnetite suggest that it probably became a primocryst mineral phase at this stratigraphic level. Fine-grained magnetite begins to crystallize in abundance at the base of the overlying Ferrosyenite Unit. The magnetite occurs as parallel curvilinear lamellae that are interpreted as a post-emplacement deformational fabric. The main surface exposures of the Ferrosyenite Unit on the western shore of Agnew Lake (Fig. 2a) are characterized by marked positive magnetic anomalies (Gupta, 1991b).
The intrusion is host to fine-grained, disseminated, but irregularly distributed sulphide minerals in the form of chalcopyrite with lesser amounts of pyrrhotite. In general, total sulphide abundances are <1%, but rock units in contact with the footwall during emplacement (i.e. the Marginal Leucogabbronorite and Inclusion-bearing Gabbronorite Zones) locally contain coarse-grained sulphide minerals in greater amounts of up to 5%. Many of these latter occurrences are spatially associated with quartz blebs and felsic footwall inclusions (compare the East Bull Lake Intrusion; Peck et al., 1993a), perhaps indicating the influence of local crustal contamination on sulphide precipitation along the margins of the intrusion (Vogel et al., 1997). In other parts of the intrusion, fine-grained, disseminated sulphide minerals tend to occur within the interstitial component of the rocks in association with quartz, apatite and sometimes titanomagnetite, suggesting that sulphides crystallized late in the Agnew magmas.
|
The Geochemistry of The Agnew Intrusion |
|---|
|
TOPABSTRACTIntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
Sampling and analytical techniques
Sixty-seven representative samples from the various zones and units of the Agnew Intrusion, its granitic footwall, and a variety of spatially associated mafic dykes were selected from a total suite of
300 surface samples. Most samples collected in the field exceeded a total weight of 2 kg. Weathered surfaces on all samples were removed by a rock saw. Jaw crushers with hardened steel plates were used to reduce the sample size to small chips, and a representative sample aliquot of 100 g was removed for milling in an alumina ball mill.
Whole-rock major and trace element compositions were determined at the University of Melbourne by X-ray fluorescence (XRF) on fused glass discs using 2 g of sample powder. Accuracy and precision are better than 0.5% for all major elements, and better than 10% for all trace elements. Rare earth elements (REE: La–Lu) and low abundance high-field strength elements (HFSE: Nb, Ta, Th, Pb, U) were analysed using 0.1 g of sample powder by inductively coupled plasma-mass spectrometry (ICP-MS) on a VG PlasmaQuad PQ2+ instrument at Monash University, Melbourne. Accuracy and precision are better than 5% for the REE, and better than 15% for the analysed HFSE. Platinum-group element (PGE: Ru, Pd, Ir, Pt) and Au concentrations were determined for several samples by radiochemical neutron activation (RNAA) at the University of Melbourne using the technique described by Hoatson & Keays (1989). The precious metals were pre-concentrated from 25 g of sample powder into a 2 g nickel sulphide button by fire assay (16 h at 1050°C). The buttons were subsequently irradiated for 24 h at the ‘HIFAR’ reactor in Lucas Heights, Sydney. The accuracy and precision of the reported precious metal concentrations are better than 10% for Ru, Pd, Pt and Au, and better than 20% for Ir. Quoted accuracy and precision levels for all utilized analytical methods are based on geological standard and replicate analyses.
Evaluation of element mobility
Whole-rock major element, trace element, REE and PGE data for 40 Agnew rocks are presented in Table 1. As these rocks have undergone at least upper greenschist facies metamorphism, the effects of element mobility on the geochemistry of the rocks must be considered. The HFSE, such as Zr, are generally thought to be immobile during low grades of metamorphism (Pearce & Norry, 1979; Lesher et al., 1991; Jenner, 1996), and are strongly concentrated in the residual liquid during fractionation of basaltic melts. Strong covariations between Zr and other incompatible elements, such as Ce (Fig. 4a), Yb and Y, therefore, provide good evidence that these elements and all REE concentrations have not undergone significant post-solidus modification, and can be considered with relative confidence from an igneous perspective. However, the incompatible large-ion lithophile elements (LILE), such as Rb (Fig. 4b), K and Ba, are widely scattered in plots against Zr and cannot be treated as immobile elements.
|
|
The PGE are strongly chalcophile (e.g. Peach et al., 1994
), and hence, their mobility is influenced by the extent of sulphide remobilization in the rocks during metamorphism. Although field and petrographic observations identify microscale migration of sulphides within the interstitial component of the rocks, there is no evidence for large-scale sulphide remobilization and PGE transport in the Agnew Intrusion. However, precedence does exist for sulphide remobilization adjacent to major shear zones as documented at the East Bull Lake Intrusion (Peck et al., 1993a).
Major element variations
Selected major elements for our entire sample suite are plotted against stratigraphic height in Fig. 5. A notable change in rocks with predominantly <50 wt % SiO2 to those with >50 wt % SiO2 occurs above the Olivine Gabbronorite Zone in the Upper Series. The base of the Porphyritic Unit (at a stratigraphic height of 1050 m) corresponds to a general change from olivine-normative to quartz-normative compositions. The lowest SiO2 concentrations are in the primocryst-rich Olivine Gabbronorite Zone (45 wt %) and are consistent with its high modal olivine abundance relative to the rest of the stratigraphic sequence. In the Ferrosyenite Unit at the top of the intrusion (>2000 m), SiO2 concentrations increase rapidly, but gradationally from 50 to 73 wt %. High SiO2 contents (51–55 wt %) in the basal Marginal Gabbronorite Zone are confined to late-intrusive fine-grained diabase dykes or sills whose chemical relationship with the main stratigraphic sequence is uncertain. Sample 438-DV1078 (Table 1), which was collected from a 2 m wide dyke in the Marginal Gabbronorite Zone, has a distinct whole-rock chemical composition similar to that of siliceous high-magnesian basalts or modern boninitic rocks as described, for example, by Crawford et al. (1989). Compositionally similar dykes are locally observed in the vicinity of 2450 Ma layered intrusions in Finland (Vuollo et al., 1995; Vogel et al., 1998b).
|
The TiO2 concentration is controlled by the abundance of Fe–Ti oxides. Below the Transition Unit (<1700 m), TiO2 is low, generally <0.5 wt %. The Dendrite Unit has the highest TiO2 contents in the lower half of the intrusion (excepting marginal dykes or sills). Significantly higher TiO2 concentrations in the Fe–Ti Oxide Zone (>1800 m) reflect the onset of early and abundant titanomagnetite crystallization. A maximum of 2.25 wt % TiO2 is reached at the contact between the Leucogabbro Unit and the overlying Ferrosyenite Unit, followed by a significant drop in TiO2 concentrations to 0.37 wt %. Fe2O3* follows a similar path, increasing to almost 23 wt % at the same contact and then decreasing to
4 wt % at the top of the intrusion (Table 1). Al2O3 concentrations are principally governed by the amount of plagioclase within individual samples. A negative correlation is observed between Fe2O3* (or MgO) and Al2O3, and reflects an increasing amount of plagioclase at the expense of mafic minerals. It is not necessarily an indicator of magmatic differentiation. Values of Al2O3 generally range between 15 and 20 wt %, but show no systematic variation with stratigraphic height. Rocks defined as leucogabbronorites, as distinct from gabbronorites, typically contain in excess of 20 wt % Al2O3. Leucogabbros in the Transition Unit and the Leucogabbro Unit near the top of the intrusion contain less Al2O3 (18–20 wt %), and this is attributed to the more sodic composition of the plagioclase at this level relative to stratigraphically lower leucogabbronorites.
Marginal and Lower Series rocks appear to have gradually decreasing whole-rock mg-numbers from 0.7 to 0.6 (Fig. 5), despite the fact that it is clear from field relationships that Lower Series magmas were intrusive into the plagioclase-rich rocks of the Marginal Series. A marked shift to higher mg-numbers occurs at the Lower Series–Upper Series boundary. The mg-numbers of Upper Series rocks then decreases gradationally with stratigraphic height from 0.72 in the Olivine Gabbronorite Zone to <0.30 in the uppermost Ferrosyenite Unit. Fe2O3* concentrations tend to increase with stratigraphic height in the Upper Series until the top of the Leucogabbro Unit; this is typical of a tholeiitic Fe-enrichment trend. The increase in mg-number within the Ferrosyenite Unit is due to the large amounts of fractionation of titanomagnetite and magnetite in the lower parts of the unit. The Upper Series data are consistent with a closed-system fractionation process. The highest mg-number in the Agnew Intrusion (0.75) is in the siliceous high-magnesian basalt-like dyke in the Marginal Gabbronorite Zone.
Trace element and REE variations
Selected trace elements for our entire sample suite (except La) are plotted against stratigraphic height in Fig. 6. The analysed samples lack visible sulphide minerals and chromite, so that variations in Ni and Cr abundances are principally governed by the different relative amounts of mafic primocryst and interstitial material they contain, as well as the relative degree of fractionation of the crystallizing liquid.
|
The Marginal and Lower Series rocks are highly variable in their Ni and Cr abundances (106–367 ppm and 95–506 ppm, respectively; Table 1), reflecting differences in the proportions of plagioclase and mafic minerals in the rocks. These modal differences may in large part be due to variations in the efficiency of separating crystals from liquid, as well as open-system behaviour involving the influx of multiple magma pulses into the chamber. In contrast, the Upper Series exhibits a systematic decrease in Ni and Cr with increasing stratigraphic height (Fig. 6), which is more consistent with the effects of closed-system fractionation. The Olivine Gabbronorite Zone at the base of the Upper Series contains anomalously high Ni abundances (up to 725 ppm) that can be attributed to the presence of abundant olivine primocrysts (up to 25%). Apart from a single melanogabbronorite sample in the Upper Layered Unit (428-DV967), Cr abundances in the Upper Series are very low, i.e.
179 ppm Cr in the Olivine Gabbronorite Zone grading to 33 ppm Cr above the Mixed Unit. This suggests that the parental magma giving rise to the Upper Series was probably more evolved than those that produced the underlying Marginal and Lower Series. Therefore, the primitive composition of the Olivine Gabbronorite Zone, in terms of whole-rock mg-number, normative An content, and Ni abundances, as well as their high primocryst and mafic mineral content, may indicate that these rocks formed as cumulates sensu stricto (Wager et al., 1960) from a much larger magma pulse that produced the Upper Series, rather than as products of a small, discrete primitive magma pulse as suggested by Peck et al. (1995) for the Olivine Gabbronorite Zone in the East Bull Lake Intrusion. Incompatible trace elements include Zr, Y, Nb and La (Fig. 6). Their relative abundances are primarily governed by the proportions of primocrysts to interstitial components in the sample and by their relative degree of differentiation. Concentration profiles through the stratigraphic sequence are crudely asymptotic with very low incompatible trace element abundances in the Marginal and Lower Series and gradually increasing abundances in the Upper Series. In detail, there is a recognizable discontinuity at the Lower Series–Upper Series boundary, manifested by a sudden decrease in incompatible trace element abundances from the Lower Layered Unit to the overlying Olivine Gabbronorite Zone. This decrease reflects the influx of a new magma pulse, which resulted in a sudden change to a less differentiated bulk rock composition characterized by a lower amount of interstitial material. The highest incompatible trace element abundances occur within the Ferrosyenite Unit (e.g. Zr = 422 ppm, Y = 61 ppm, Nb = 23 ppm, La = 34 ppm) and reflect the highly evolved nature of these rocks.
Selected bivariate plots for all samples in Table 1 are illustrated in Fig. 7. A regression line through Y vs Zr data intercepts very close to the origin (Fig. 7a), confirming their incompatible behaviour during crystallization of Agnew magmas. Therefore, the Zr/Y value for individual samples should be equivalent to that of their source material, where any variation in the ratio reflects heterogeneity in the source, magma mixing, and/or crustal contamination. The excellent linear correlation between Y and Zr suggests that all Agnew rocks were derived from a similar mantle source material. Exceptions include the Archaean granite, an ultramafic inclusion from the Marginal Leucogabbronorite Zone, and the siliceous high-magnesian basalt-like dyke in the Marginal Gabbronorite Zone (438-DV1078), all of which deviate from the linear array defined by the main intrusion. The average Zr/Y value of 3.7 ± 1.2 for Agnew samples is not sufficiently distinctive to categorize its source in terms of modern mantle reservoirs or their respective tectonic environments (e.g. Pearce & Norry, 1979; Meschede, 1986; Kerrich & Wyman, 1996). Figure 7b illustrates a continuous variation trend between Y and mg-number, particularly in the Upper Series with the primocryst-rich Olivine Gabbronorite Zone samples plotting at high mg-number and low Y, and Ferrosyenite Unit samples at low mg-number and high Y, supporting a closed-system fractionation process. As in Fig. 7a, the ultramafic inclusion and sample 438-DV1078 fall off the main data trend, indicating that these rocks are not comagmatic with the Agnew Intrusion.
|
Average chondrite-normalized REE patterns for each of the major stratigraphic subdivisions are illustrated in Fig. 8a and b; the salient features of these patterns are given in Table 2. All Agnew rocks have patterns of light REE (LREE) enrichment with a narrow range in (La/Yb)n values from 2.0 to 6.5. Each stratigraphic subdivision yields an average REE pattern that is subparallel to the others. The only exceptions are samples 303-DV554 and 438-DV1078 from the Marginal Gabbronorite Zone, which have slopes and patterns that are significantly different from all other rocks of the intrusion. Slight variations in the REE slope between samples can generally be attributed to modal differences, such that samples with lower (La/Yb)n values have higher relative mafic mineral to plagioclase contents and vice versa, which is consistent with known REE partitioning in the common rock-forming minerals (Henderson, 1982
). This is best illustrated by rocks from the Olivine Gabbronorite Zone. An olivine melanogabbronorite (392-DV823) comprising 70% mafic minerals has an (La/Yb)n value of 2.0, whereas an immediately overlying olivine gabbronorite (391-DV823) with a lower mafic mineral content of 45% has a more LREE-enriched (La/Yb)n value of 4.0.
|
|
The Marginal and Lower Series (excluding Marginal Gabbronorite Zone samples) exhibit a gradual increase in
REE abundances with increasing stratigraphic height from 5* to 11* chondritic abundances and a concomitant decrease in their positive Eu anomaly (Eu/Eu*) from 2.04 to 1.34 (Fig. 8a; Table 2). Although these data are compatible with closed-system magmatic differentiation dominated by plagioclase fractionation, they are not consistent with field data that indicate an intrusive relationship and temporal break between the Lower and Marginal Series. The large positive Eu anomaly in the Marginal Leucogabbronorite Zone reflects the high abundance of primocrysts of calcic plagioclase.
The Upper Series is similarly characterized by a gradual increase in REE with increasing stratigraphic height from 6 x to 50 x chondritic abundances (Fig. 8b; Table 2), lending further support to the hypothesis that the Upper Series succession is the product of closed-system fractionation. Most Upper Series rocks have a positive Eu anomaly, but there is no systematic change in its magnitude with stratigraphic height, indicating that fractionation involved significant amounts of other mineral phases in addition to plagioclase, probably olivine and later orthopyroxene and titanomagnetite.
The Ferrosyenite Unit at the top of the Agnew Intrusion is the only stratigraphic subdivision with a negative Eu anomaly, and is therefore consistent with representing the final product of plagioclase-dominated closed-system fractional crystallization within the Upper Series. This is supported by the three- to four-fold enrichment in REE and most other incompatible trace elements, and severe depletion in Cr, Ni, Sr and Co relative to rocks from the main underlying Upper Gabbronorite Zone. The strong Fe enrichment at the base of the Ferrosyenite Unit (23 wt % Fe2O3* at a SiO2 content of 50 wt %; Table 1) is the same as that calculated for late-stage melts proposed for the Skaergaard and Kiglapait Intrusions ( 22 wt %; Wager & Brown, 1968; Morse, 1981). Toplis & Carroll (1995) have shown that immediately following the onset of magnetite crystallization in a magma, the residual melt may proceed along a trend of Fe depletion and SiO2 enrichment. Such a trend accounts very well for the rapid, but gradual upward mineralogical and chemical change within the Ferrosyenite Unit from Fe-rich ferrosyenite to Fe-poor and SiO2-rich alkali-feldspar granite. An alternative origin for the Ferrosyenite Unit sequence by increasing contamination of Agnew magmas by felsic material appears unlikely given that all recognized felsic rocks in the region are either too low in REE or have REE slopes that are too steep to account for the Ferrosyenite Unit REE pattern (Fig. 8c).
Figure 8d presents average REE patterns for the Dendrite, Mixed and Lower and Upper Layered Units, to compare their chemical characteristics and determine the compositional relationships between them. Marked similarities exist between the REE characteristics of the Dendrite and Mixed Units (Table 2). Combined with the presence of various inclusions and mafic dendritic textures in both units, the data suggest that these rock units may be products of the same volatile-rich magma pulse that intruded at a late magmatic stage. The greater lithological variability and poorer preservation of dendritic textures within the Mixed Unit compared with the Dendrite Unit may be related to its emplacement site. The Mixed Unit occurs more centrally within the intrusion (Fig. 2a) and was perhaps emplaced into less crystallized and consolidated material at the time of intrusion of the Dendrite Unit magma. Comparisons of the REE patterns with the adjacent Lower and Upper Layered Units suggest that the Dendrite Unit rocks either (1) more closely approach liquid compositions relative to the adjacent units or (2) crystallized from a magma that differed in composition only by being slightly more evolved than those which formed the Lower and Upper Layered Units.
The bell-shaped REE pattern of the ultramafic inclusion occurring within the Marginal Leucogabbronorite Zone (Fig. 8d) is consistent with a rock formed by hornblende crystallization (Rollinson, 1993). Coupled with the observation that it falls off all geochemical trends defined by other Agnew rocks (Fig. 7), such features suggest that these inclusions are not cogenetic with the intrusion, but are xenoliths entrained in Agnew magmas during their ascent through the crust.
Figure 9a and b presents multi-element diagrams that characterize and summarize the average incompatible element distribution for each of the major stratigraphic subdivisions. The patterns are subparallel and show the same general step-wise increases in concentrations with stratigraphic height within the Marginal–Lower Series and Upper Series as the REE patterns in Fig. 8a and b. LILE (Ba, Rb, Th, K, Sr) abundances, which may have been subjected to some remobilization during metamorphism and weathering (Fig. 4b), are typically 10–100 x primitive mantle values, whereas the HFSE (Nb, Ta, P, Zr, Ti, Y) are generally 1–10x primitive mantle. As a result, distinct features of all of these multi-element patterns are negative HFSE anomalies, particularly for Nb–Ta, P, and Ti. These data are relevant to the petrogenetic evolution of the Agnew Intrusion, and are considered later. Pronounced positive Sr anomalies, principally in the Marginal and Lower Series, correlate with high calcic plagioclase abundances, whereas conversely, the negative Sr anomalies in the upper parts of the Upper Series reflect decreasing amounts of calcic plagioclase coupled with increasing proportions of sodic plagioclase.
|
Parallel trace element and REE characteristics are consistent with all Agnew Intrusion rocks (excluding some parts of the Marginal Gabbronorite Zone) being cogenetic. The data do not support differential amounts of crustal contamination between stratigraphic subdivisions. Proposed parental magmas of the intrusion must have trace element and REE patterns that are identical in slope and anomaly characteristics to those of the Agnew rocks. The predominance of rocks with positive Eu anomalies and absence of volumetrically significant rock types with negative Eu anomalies in the intrusion suggests that either: (1) large amounts of residual liquid with negative Eu anomalies were lost from the chamber during crystallization, or (2) the parental magmas already exhibited positive Eu anomalies and entered the chamber carrying significant amounts of plagioclase crystals. The latter alternative is supported by previous interpretations that the presence of irregular and compositionally discontinuous zonation patterns within many plagioclase primocrysts of the Marginal and Lower Series reflects crystallization of plagioclase under different pressure regimes.
Chalcophile element (PGE, Cu, Ni) variations in the Agnew Intrusion
Finely disseminated and locally coarse-grained sulphide mineralization occurs within the Agnew Intrusion, particularly near its margins in the Marginal Leucogabbronorite and Inclusion-bearing Gabbronorite Zones. Similar sulphide occurrences in marginal rock types of the East Bull Lake Intrusion have been described in detail by Peck et al., (1993a). Given that the PGE have extremely high sulphide–silicate partition coefficients (e.g. Peach et al., 1990), they are very sensitive indicators of sulphide ore-forming processes. Their abundance and distribution in a given rock provide a measure of the S-saturation status of the magma from which the rock crystallized (Hamlyn & Keays, 1986; Peck & Keays, 1990; Vogel & Keays, 1997). This information is relevant to establishing the general metallogenic potential of the Agnew Intrusion in terms of PGE, Cu and Ni, as well as further characterizing its petrogenesis.
PGE (Pd, Pt, Ir, Ru) and Au concentrations in 22 sulphide-poor (<<1%) rocks from various zones and units within the Agnew Intrusion vary widely, but are relatively high (Table 1), namely: Pd (range 0.2–58 ppb; most data lie between 10 and 40 ppb), Pt (range 0.7–62 ppb; most data lie between 10 and 30 ppb), Ir (range 0.003–1.68 ppb; most data lie between 0.02 and 0.50 ppb), Ru (range 0.01–5.9 ppb; most data lie between 0.05 and 0.50 ppb), Au (range 0.17–7.76 ppb; most data lie between 0.5 and 3.0 ppb). For example, these concentrations are generally greater than those reported for unmineralized mafic rocks from below the J-M Reef in the Stillwater Complex (7 ppb Pd, 13 ppb Pt, 0.17 ppb Ir, 0.4 ppb Au; Peck & Keays, 1990), but are similar to PGE abundances within equivalent rock types of the adjacent East Bull Lake Intrusion (Peck et al., 1995). Whole-rock S concentrations in most Agnew Intrusion rocks are very low (typically <150 ppm; Table 1), and are believed to closely approximate the original S tenor of the magma.
Selected bivariate plots of chalcophile elements are presented in Fig. 10a–d. In general, correlations between the chalcophile elements are poor. A good positive correlation exists between Pd and Pt, indicating that these elements behaved sympathetically during magmatic crystallization. Pd and Ir show no obvious systematic covariation, although both elements are generally elevated in the Marginal and Lower Series relative to the Upper Series. Ni and Cu show positive and negative correlations with whole-rock mg-number, respectively, suggesting that Ni partitioned into olivine primocrysts during magmatic differentiation, whereas Cu preferentially partitioned into the residual liquid. Ni abundances are generally higher and Cu abundances lower in the Marginal and Lower Series than in the Upper Series. An exception is the high Ni content in the Olivine Gabbronorite Zone, which probably reflects the higher relative abundances of olivine primocrysts.
|
Most Agnew Intrusion rocks are characterized by a step pattern on a mantle-normalized PGE and chalcophile element diagram and cover the range 0.01–10x mantle values (Fig. 11). For all samples, the Pt–Pd–Au–Cu portion of the diagram is elevated with respect to the Ni–Ir–Ru portion. Ni tends to be weakly enriched relative to Ir (Nin/Irn is typically 1–7), and Cu is generally depleted relative to Pd (Cun/Pdn = 0.9 ± 0.6). Samples from the Inclusion-bearing Gabbronorite Zone have elevated Ir and Ru abundances relative to the rest of the intrusion, and are less fractionated in terms of Pd/Ir. Both Ir and Ru probably occur as metal alloys that are preferentially incorporated within early crystallizing mineral phases (Agiorgitis & Wolf, 1978
; Keays, 1982; Crocket & MacRae, 1986). The increased concentrations of Ir and Ru in the two Inclusion-bearing Gabbronorite Zone samples may be accounted for by their relatively high modal abundances (20%) of early-crystallizing plagioclase primocrysts with An79–69 core compositions. Rocks from the Marginal Leucogabbronorite and Olivine Gabbronorite Zones also have large proportions of primocrysts and correspondingly elevated Ir concentrations relative to other samples (Fig. 11). These observations suggest that the highly variable PGE abundances in Agnew rocks and the poorly developed evolutionary PGE trends through the intrusion as a whole may in part be explained by different proportions of primocrysts vs later-crystallizing interstitial material in the samples. Multiple magma pulses and variable degrees of magmatic differentiation have undoubtedly also influenced PGE concentrations.
|
The chalcophile element data presented above are entirely consistent with the view that the stratigraphic sequence crystallized from magmas that were S undersaturated when first introduced into the Agnew chamber. These data include: (1) high PGE concentrations in sulphide-poor rocks; (2) a positive correlation between Ni and mg-number highlighting the low importance of sulphides during magmatic crystallization; (3) the incompatible behaviour of Cu during crystallization; (4) the very low S contents of all rocks, i.e. well below the abundances expected for mantle-derived magmas that have been saturated with S (>800 ppm; Peach et al., 1990
). Figure 12 presents Pd and Cu data for various sulphide-poor mafic intrusive and volcanic rocks, and discriminates between rocks that formed from S-undersaturated and S-saturated magmas (Vogel & Keays, 1997). Most Agnew samples plot well within the S-undersaturated field, supporting the data presented above. Therefore, the parental magmas of the Agnew Intrusion entered the chamber as metal-fertile magmas that had not previously segregated any sulphides. Field evidence suggests that in lower parts of the intrusion (specifically the Marginal Leucogabbronorite and Inclusion-bearing Gabbronorite Zones), S saturation was induced locally through assimilation of siliceous country rock, a mechanism proposed byIrvine (1975) and Naldrett (1989) to operate in other mafic intrusions. In one location, melanogabbronorites that provide both field and geochemical evidence for in situ crustal contamination (Vogel et al., 1997), contain abundant blebs of chalcopyrite with highly enriched PGE concentrations (PGE = 10 ppm; BP Resources Canada Ltd, unpublished assay data, 1991). However, judging from exposed rock types, crustal contamination was not pervasive within any part of the Agnew Intrusion and its magmas appear to have remained largely S undersaturated until the very latest stages of their crystallization history.
|
The two analysed Ferrosyenite Unit samples from near the top of the intrusion are extremely depleted in Ni, Ir, and Pd, and yield positively sloping metal patterns (Fig. 11) that are similar to those of ocean-floor basalts (Barnes et al., 1988
). Ocean-floor basalts are believed to be the products of S-saturated magmas from which the metal-bearing sulphides were segregated in the mantle source or en route to the surface (Hamlyn et al., 1985). If the Upper Series succession is a product of closed-system magmatic differentiation, then the decrease in chalcophile element abundances in the Ferrosyenite Unit may have coincided with an S-saturation event that depleted the residual melt in these elements. Pd and Cu data in Fig. 12 are consistent with the Ferrosyenite Unit rocks crystallizing from S-saturated magmas. S saturation may have been induced by the combined effects of progressive differentiation, decreasing FeO content related to titanomagnetite and magnetite crystallization, and quartz entering onto the liquidus (Haughton et al., 1974; Wendlandt, 1982). |
Parental Magmas of The Agnew Intrusion |
|---|
|
TOPABSTRACTIntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
A variety of mafic dykes that are spatially associated with the Agnew Intrusion have been investigated in an effort to identify a suitable parental magma composition(s) for the intrusion.
Subdivision, characterization and distribution of mafic dykes
Mafic dykes in the Agnew Intrusion area can be subdivided into four groups on the basis of their petrographic and geochemical characteristics. More than 200 separate dykes are exposed at the present erosional level. They intrude into both the Archaean granite footwall and the Agnew Intrusion itself, but do not extend into the overlying Huronian Supergroup metasediments immediately to the east of the intrusion. Dyke density is highest to the northwest of the intrusion in the vicinity of the Streich Dyke (Fig. 2a), which is by far the longest and widest of the dykes in the area. All four dyke groups are typically oriented along a 110–120° trend parallel to the Streich Dyke. Crosscutting relationships between different dyke groups have not been observed. Regional greenschist facies metamorphism associated with the 1850 Ma Penokean Orogeny has resulted in pervasive recrystallization of dyke mineralogy. The main physical characteristics of each dyke group are outlined below.
Group I—Streich-type gabbronorite dykes
This group consists essentially of the Streich Dyke, which crops out along a prominent 4 km ridge that almost links the coeval Agnew and East Bull Lake Intrusions. It varies in width from 50 to 300 m and has well-developed, weakly plagioclase-phyric chilled margins. A rare unaltered plagioclase phenocryst has a composition of An78. At its easternmost end, the Streich Dyke is characterized by a coarse-grained and locally vari-textured interior zone. Other dykes with Group I-type chemical compositions are uncommon; one occurs adjacent to the contact between the Archaean footwall and the Agnew Intrusion as part of the Marginal Gabbronorite Zone, and another occurs as a chemically distinct component within a Group III composite dyke.
Group II—aphyric gabbronorite dykes
Dykes belonging to Group II are typically fine- to medium-grained and massive, ranging in width from 50 cm to 50 m. They constitute the most abundant dyke variety in the Agnew area and are distinguished geochemically from Group I dykes. Their distribution is widespread in the Archaean granitic rocks and throughout the intrusion.
Group III—plagioclase-phyric gabbronorite dykes
These dykes are common and are distinguished in the field by differing amounts (2–50%) of small plagioclase glomerophenocrysts set in a fine- to medium-grained gabbronoritic matrix. The glomerophenocrysts have been recrystallized to oligoclase compositions. The dykes generally range in width from 2 to 20 m, and some appear to be composite. Group III dykes are most common to the northwest of the Agnew Intrusion, where some are truncated by the footwall–intrusion contact, but they are also abundant with chilled margins within the intrusion, indicating that some of these dykes post-date the crystallization of the Agnew Intrusion. Their abundance is lowest within the Upper Series.
Group IV—coarse-grained gabbronorite dykes
This group is characterized by coarse-grained, ‘spotty’ interiors consisting of 50% equant mafic crystals (originally pyroxene?) set within a plagioclase matrix. The dyke interiors grade evenly over 50 cm into well-developed chilled margins. The dykes are generally 5–10 m wide. Large megacrysts of saussuritized and locally chloritized plagioclase (5 cm) may be present in amounts of up to 60%. The presence of such large plagioclase megacrysts has also been reported from many Hearst—Matachewan dykes (Phinney et al., 1988; Halls & Bates, 1990; Nelson et al., 1990; Ashwal, 1993). Group IV dykes are geographically confined to northern areas both within and outside the Agnew Intrusion, and do not appear to intrude rocks above the approximate stratigraphic base of the Upper Series.
Geochemistry of the mafic dykes
Table 3 presents average whole-rock geochemical data for Group I–IV and Hearst–Matachewan mafic dykes. Average data for the latter have been calculated from 16 analyses taken from Condie et al. (1987) and Nelson et al. (1990). These were chosen from a larger set of 32 analyses on the basis of their similar REE patterns and are considered to be the best estimate of the composition of the dyke swarm. All Group I–IV and Hearst–Matachewan dykes are sub-alkaline and are high-Fe tholeiitic basalts (Fig. 13). Excepting Group I, which are olivine-normative, most dykes are quartz-normative; Group IV and Hearst–Matachewan dykes comprise both varieties. Group I dykes are also distinguished by their substantially lower TiO2 (0.45 wt %) and higher Al2O3 (>17 wt %) and mg-number (0.64). The other dyke groups have in excess of 1.0 wt % TiO2, 13–15 wt % Al2O3 and a more compositionally evolved mg-number of 0.32–0.59 (typically 0.45). In various tectonic discrimination diagrams, the dyke compositions have the geochemical characteristics of both destructive plate margins and continental flood basalts (Fig. 14a and b). This suggests that, despite their continental setting, these dykes share geochemical characteristics with modern volcanic arc tholeiitic basalts. Data for Group I dykes are always displaced from the other dyke groups and plot within the compositional field of boninites, indicating a distinct minor and chalcophile element chemistry that may represent a fundamental difference in their mantle sources. However, although they have some significant chemical similarities with modern boninites, Group I dykes are too depleted in SiO2 and MgO, and too enriched in Al2O3 to serve as boninitic analogues.
|
|
|
Average REE and incompatible element characteristics for the mafic dykes, including the Hearst–Matachewan dyke swarm, are shown in Fig. 15a and b, and important REE ratios are given in Table 3. All dykes are characterized by LREE enrichment relative to chondrite, with Groups I and II showing steeper patterns [(La/Yb)n = 3.7] similar to those of the Agnew Intrusion, compared with Groups III and IV [(La/Yb)n = 2.4–2.6]. These slope differences cannot be accounted for by differing degrees of fractionation, and therefore dykes of Groups I and II are probably not comagmatic with Groups III and IV. Group III and IV dykes may, however, be comagmatic with the Hearst–Matachewan dykes, at least in terms of their REE patterns. The Group I Streich-type gabbronorite dykes are the only group with a positive Eu anomaly (Eu/Eu* = 1.34). Assuming that normal mantle-derived liquids do not have positive Eu anomalies, and using partition coefficients for Eu into plagioclase of basaltic liquids (Arth, 1976
; Fujimaki et al., 1984), it is calculated that these dykes contain up to 10% intratelluric plagioclase. The other dyke groups have slightly negative Eu anomalies, indicating that they have probably fractionated small amounts of plagioclase at deeper crustal levels. All analysed mafic dykes in the Agnew Intrusion area, as well as the Hearst–Matachewan dykes, have negative anomalies in Nb–Ta, P, and Ti relative to LILE and REE (Fig. 15b), identical to those shown by the Agnew Intrusion. A positive Sr anomaly in Group I dykes correlates with the positive Eu anomaly and high Al2O3 contents, and is consistent with elevated plagioclase concentrations.
|
Average S concentrations for the four dyke groups are given in Table 3. Although S is considered relatively mobile during metamorphism and weathering, the observation that very low S abundances (60 ppm) in a thin,
1 m wide Group I dyke were preserved immediately adjacent to high S abundances (800 and 1100 ppm) within a Group III composite dyke, suggests that the reported S values are probably close to their original magmatic concentrations, as both values are typical of other dykes in their respective groups. The S data are plotted on a Poulson–Ohmoto diagram (Fig. 16), which discriminates, on the basis of FeO content, between magmas that have S abundances in excess of their S capacity and those that are undersaturated in S (Poulson & Ohmoto, 1990). Group I dykes are clearly the most S undersaturated, and given their high PGE concentrations (e.g. Pd = 19 ppb), they were metal-fertile magmas. Group II, III and IV dykes have significantly higher S abundances and generally plot near their estimated S capacity. Their Pd abundances are highly varied (between 0.45 and 28 ppb), suggesting that some dykes of Groups II, III and IV crystallized from S-saturated magmas that may or may not have segregated sulphide, whereas others probably remained S undersaturated.
|
Relationship between mafic dykes and Agnew Intrusion
There are several factors that indicate that Group I dykes of Streich-type gabbronoritic composition were parental to the Agnew Intrusion. These include:
- field relationships, showing that the trend of the Streich Dyke intersects the intrusion at the site where the Inclusion-bearing Gabbronorite Zone magmas intruded and disrupted the Marginal Leucogabbronorite Zone;
- high Al2O3, consistent with plagioclase crystallizing as the earliest and principal liquidus mineral phase within the intrusion, and low TiO2, which is consistent with the absence of ilmenite and the late crystallization of titanomagnetite;
- the presence of positive Eu anomalies, suggesting the presence of up to 10% intratelluric plagioclase in the parental magma, which supports interpretations based on plagioclase zonation patterns, as well as accounting for the general absence of rocks with negative Eu anomalies;
- olivine-normative compositions, capable of giving rise to the entirely olivine-normative Marginal and Lower Series in which pseudomorphed olivine crystals show no evidence of having been resorbed;
- similar REE and multi-element patterns characterized by LREE enrichment and distinctly negative HFSE anomalies;
- PGE-rich and S-undersaturated compositions, in accordance with the high background PGE concentrations and low S abundances within the Agnew Intrusion.
Can Group I Streich-type gabbronorite dykes be parental to all three series in the Agnew Intrusion? This can be evaluated on the comparative basis of REE abundance data and (La/Yb)n values for the intrusion and Group I dykes (Tables 2 and 3). Average compositions for the Lower Series and Group I dykes have very similar REE abundances and similar major and trace element characteristics as a whole (Tables 1 and 3). Some differentiation from the Inclusion-bearing Gabbronorite Zone to the Lower Layered Unit did occur, but in general, the Lower Series appears to have crystallized without substantial modification from a single pulse of Group I magma.
Average compositions for the Marginal Series (excepting those from the Marginal Gabbronorite Zone) are depleted in REE, but have greater positive Eu anomalies relative to Group I dykes. Therefore, if Group I dykes were parental to the Marginal Series, then the Marginal Series rocks must have crystallized under conditions of open-system fractionation, producing an evolved, mafic residual liquid with relatively elevated REE abundances and a negative Eu anomaly. This residual liquid was lost from the system and may have erupted as part of the thick Elliot Lake Group volcanic sequence.
In the Upper Series, all rocks except those of the Olivine Gabbronorite Zone have elevated REE abundances relative to Group I dykes. Given that the Upper Series succession is consistent with being a product of closed-system fractionation, Group I dykes are too primitive and too low in their REE abundances to account for the large proportional volume of Upper Series rocks with enhanced REE concentrations. Therefore, Group I dyke compositions were not parental to the Upper Series. In contrast, Group II dykes are probably too evolved in terms of their REE concentrations to be parental to the Upper Series. The data are most consistent with a parental magma for the Upper Series that had an intermediate whole-rock composition between Group I and II dykes.
Dykes of Groups III and IV and the Hearst–Matachewan swarm have relatively flat REE characteristics that effectively preclude them as important parental magma compositions to the Agnew Intrusion. It is possible that these dykes were feeders to the Elliot Lake Group volcanic sequence at different stages during the emplacement of the Agnew Intrusion sub-volcanic sill. This would explain both the local truncation of some dykes at the footwall–intrusion contact and the sudden decline in dyke numbers at series boundaries, particularly the Lower–Upper Series boundary. Group III and Hearst–Matachewan dykes are chemically almost identical (Fig. 15; Table 3), whereas Group IV dykes may be related comagmatically by lower relative degrees of fractional crystallization of a basaltic mineral assemblage. Therefore, Group III and IV dykes may represent a southerly extension of the Hearst–Matachewan dyke swarm, confirming that the swarm does indeed rotate by 30° from a southeast to ESE direction towards a focal point near Sudbury, as predicted by Halls & Bates (1990). It also suggests that the Hearst–Matachewan dyke swarm is not comagmatic with the Agnew Intrusion, but may instead have fed parts of the Elliot Lake Group volcanic sequence.
Comparison with parental magmas of other layered intrusions
Table 4 presents major element data for proposed parental magmas of the gabbroic portions of some of the best known layered intrusions in the world. All have high-Al basaltic compositions (>17 wt % Al2O3) and similar mg-numbers. The proposed parental magmas of the Agnew, Stillwater and Bushveld Complexes are also characterized by very low TiO2 concentrations, whereas the much younger Skaergaard Intrusion was fed by a magma containing almost three times as much TiO2. Column 5 in Table 4 is a model parent liquid composition (Ao) for the Stillwater and Bushveld Complexes calculated through the addition of 17% plagioclase (An70) to a chilled Bushveld rock composition (Irvine et al., 1983). The major element data indicate that the parental magmas proposed for the mafic portions of the Stillwater and Bushveld Complexes have almost identical compositions to Group I dykes that formed the Agnew Intrusion.
|
An average composition for modern boninites is also provided in Table 4. Boninites are thought to be analogous in composition to the parental magmas that gave rise to the ultramafic portions of the Stillwater and Bushveld Complexes (e.g. Hamlyn & Keays, 1986
). Clearly, such magmas are very different in composition and appear very difficult to relate by fractional crystallization to parental magmas of gabbroic portions of layered intrusions. |
The Source of Agnew Intrusion Parental Magmas |
|---|
|
TOPABSTRACTIntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
The
2475–2450 Ma East Bull Lake layered intrusion suite, Hearst–Matachewan dyke swarm and Elliot Lake Group volcanic sequence in the southern Superior Province are believed to be remnants of a large Palaeoproterozoic continental flood basalt province. Vogel et al. (1998a) argued on the basis of the large volume and early appearance of mafic magma that rifting, which led to the accumulation of the Huronian Supergroup, was caused by mantle plume-induced magmatism rather than tectonically induced adiabatic melting of the subcontinental lithospheric mantle. However, Agnew Intrusion parental magmas, i.e. Group I dykes, do not exhibit typical mantle plume signatures with respect to major or trace element characteristics. Their low TiO2 contents, negative HFSE anomalies, and S-undersaturated nature are geochemical features that are more typical of modern magmas derived from depletedlithospheric mantle sources that have been modifiedby LREE-enriched, subduction-related fluids, or, alternatively, derived through lower-crustal contamination of melts from deeper mantle sources. Given the strong geochemical similarity between several successive and discrete magma pulses to the Agnew Intrusion, crustal contamination of the Group I feeder dykes is considered unlikely, unless the contamination process was remarkably uniform in both space and time. Therefore, magmas associated with the Agnew Intrusion are believed to have inherited their low TiO2, HFSE and S abundances from a previously melted mantle source enriched by LREE-bearing fluids. The extensive production of greenstones during the Archaean would have given rise to a thick mantle residue that could have provided both the volume and a suitable source composition for the early Palaeoproterozoic magmatism in the southern Superior Province. As Group I dyke compositions are probably too evolved to be primary mantle-derived melts, their high Al2O3 concentrations could have been generated through high-pressure magma fractionation at depth (Lee & Ripley, 1996). More Fe-rich tholeiitic magmas, in the form of Group II, III and IV dykes, and the Hearst–Matachewan dyke swarm, may then represent fractionated partial melt products of deeper, less depleted mantle (Ohta et al., 1996).
An alternative, but perhaps less likely source for the Agnew Intrusion parental magmas is plagioclase-bearing mafic or ultramafic material that has ponded and crystallized at the crust–mantle boundary during the Archaean. A review of chemical compositions of Archaean mafic–ultramafic rocks [e.g. that by Kerrich & Wyman (1996)] indicates that these are generally low in TiO2 and HFSE, and S undersaturated (Lesher & Groves, 1986). Relative to the Abitibi greenstone belt 300 km to the north, the southern Superior Province in the vicinity of the Agnew Intrusion lacks substantial amounts of exposed Archaean greenstone, such that much of it in this area may have ponded at depth. Partial melting of this material could produce aluminous melts that crystallize plagioclase upon intrusion at shallow depths (Ashwal &Seifert, 1979).
Both source models presented above still require the influence of a plume if our interpretations regarding the evolution of the rift zone, and the volume and relative timing of magmatism are correct. Perhaps, the Archaean–Palaeoproterozoic era was more typically characterized by ‘thermal’ plumes that induced partial melting in upper-mantle rocks, but did not themselves contribute much in the form of magma, which would have imparted plume-like chemical signatures to all of the magmatic rocks. Alternatively, Archaean-Palaeoproterozoic mantle source compositions may have been significantly different from those giving rise to modern magmas, such that the modern subduction-like chemical signature in Agnew rocks (e.g. depleted HFSE relative to LILE and LREE) could have been a more pervasive feature in the Archaean–Palaeoproterozoic mantle than at present (Vogel et al., 1998a, 1998b).
|
Summary and Conclusion |
|---|
|
TOPABSTRACTIntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
The Palaeoproterozoic Agnew Intrusion is a 2100 m thick leucogabbronoritic to gabbronoritic pluton that is believed to have been emplaced as a sub-volcanic sill between Archaean granitic rocks and a thick, overlying sequence of flood basalts. It has been subdivided into three major and distinct stratigraphic series, each of which was separated by a time interval in which there was no magmatic injection into the chamber.
The lowermost Marginal Series is composed mainly of 200 m of leucogabbronorites formed predominantly by plagioclase fractionation from a tholeiitic parental magma composition equivalent to that of an S-undersaturated, high-Al and low-Ti Group I dyke. Group I dykes are similar in composition to the model parent liquid proposed for the mafic portions of the Stillwater and Bushveld Complexes. An evolved residual liquid was removed from the system during Marginal Series crystallization, indicating open-system fractionation conditions. This liquid possibly erupted as part of the Elliot Lake Group volcanic sequence.
Another Group I parental magma subsequently intruded through the semi-consolidated Marginal Series at the southeastern end of the Streich Dyke and extended laterally within the chamber forming the 550 m thick Lower Series sequence. The Lower Series crystallized producing compositions very near those of the original parental magma.
The 1350 m thick Upper Series was conformably emplaced onto the Lower Series succession, crystallizing largely with closed-system fractionation and producing a lithologically and geochemically graded sequence from primitive and cumulus olivine gabbronorites at the base to highly evolved ferrosyenites and alkali-feldspar granites at the top. The parental magma of the Upper Series probably had a composition intermediate between Group I and II dykes. The Agnew sequence was later locally intruded by small-volume, volatile-rich magma pulses that gave rise to conformable Dendrite Unit bands.
Group III and IV dykes, which are spatially associated with the Agnew Intrusion, are unrelated to the intrusion sequence, but show strong petrographic and geochemical similarities with the extensive Hearst–Matachewan dyke swarm to the north. This suggests that magmas giving rise to the Hearst–Matachewan dyke swarm were not comagmatic with the Agnew Intrusion.
Possible mantle sources to the Agnew Intrusion include: (1) the thick mantle residue after partial melting in the Archaean, which produced the widespread greenstone sequences; or (2) plagioclase-bearing mafic or ultramafic material that ponded and crystallized at the crust–mantle boundary during the Archaean. To satisfy magma volume, as well as temporal constraints on the evolution of the rift zone, ‘thermal’ plumes may be required to induce partial melting in these sources.
|
Acknowledgements |
|---|
This study formed part of a Ph.D. dissertation undertaken by D.C.V. at the University of Melbourne, Australia, while in receipt of an Australian Post-graduate Research Award. We are indebted to David Peck, who first introduced D.C.V. to the Agnew Intrusion. I. Campbell and R. Eckstrand are thanked for their helpful reviews of the original Ph.D. thesis on which much of this paper is based. We also appreciate subsequent discussions and comments by P. Golightly and S. Prevec on earlier versions of this manuscript, and those of the journal reviewers, D. D. Lambert, A. R. McBirney and D. C. Peck. Financial support for fieldwork and chemical analyses was provided by Inco Exploration and Technical Services Ltd, research funds from the University of Melbourne, Australia, and Laurentian University, Sudbury, Canada, and an NSERC grant to R.R.K. and R.S.J., all of which are gratefully acknowledged.
* Corresponding author. Present address: 39 Helendale Drive, MS 2131, Toowoomba, Qld. 4352, Australia. Fax: +61-746-976-827. e-mail: dcfvogel{at}hotmail.com
|
References |
|---|
|
TOPABSTRACTIntroductionThe Geology of the...The Geochemistry of The...Parental Magmas of The...The Source of Agnew...Summary and ConclusionReferences |
|---|
Agiorgitis G., Wolf R. Aspects of osmium, ruthenium and iridium contents in some Greek chromites. Chemical Geology (1978) 23:267–272.[Web of Science]
Arth J. G. Behaviour of trace elements during magmatic processes–a summary of theoretical models and their applications. Journal of Research of the US Geological Survey (1976) 4:41–47.[Web of Science]
Ashwal L. D. Anorthosites. Minerals and Rocks 21 (1993) Berlin: Springer-Verlag. 422.
Ashwal L. D., Seifert K. E. Rare earth element evidence for mafic enriched residual liquids of the massif–type anorthosite suite, Adirondacks, New York. Abstracts with Programs, Geological Society of America (1979) 11:2.
Ashwal L. D., Wooden J. L. River Valley pluton, Ontario: a late Archean/Early Proterozoic anorthositic intrusion in theGrenville Province. Geochimica et Cosmochimica Acta (1989) 53:633–641.[Web of Science]
Barnes S.-J., Boyd R., Korneliussen A., Nilson L.-P., Often M., Pedersen R. B., Robins B. The use of mantle normalization and metal ratios in discriminating between the effects of partial melting, crystal fractionation and sulphide segregation on platinum–group elements, gold, nickel and copper: examples from Norway. In: Geo–Platinum '87 Symposium Volume—Prichard H. M., Potts P. J., Bowles J. F. W., Cribb S. J., eds. (1988) Amsterdam: Elsevier. 113–143.
Bennett G. B., Dressler B. O., Robertson J. A. The Huronian Supergroup and associated intrusive rocks. In: Geology of Ontario. Ontario Geological Survey, Special Volume—Thurston P. C., Williams H. R., Sutcliffe R. H., Stott G. M., eds. (1991) 549–591.
Bowen N. L. The melting phenomena of the plagioclase feldspars. American Journal of Science (1913) 35:577–599.[Web of Science]
Card K. D. Geology of the Sudbury–Manitoulin area. Ontario Geological Survey, Report (1978) 166:238.
Card K. D., Jackson S. L. Tectonics and metallogeny of the Early Proterozoic Huronian foldbelt and the Sudbury Structure of the Canadian Shield. Field Trip Guidebook, Precambrian '95. Geological Survey of Canada, Open File Report (1995) 3139:55.
Card K. D., Palonen P. A. Geology of the Dunlop–Shakespeare Area, District of Sudbury. Ontario Division of Mines, Geoscience Report (1976) 139:52.
Card K. D., Innes D. G., Debicki R. L. Stratigraphy, sedimentology, and petrology of the Huronian Supergroup in the Sudbury–Espanola Area. Ontario Division of Mines, Geoscience Study (1977) 16:99.
Carr J. M. Zoned plagioclases in layered gabbros of the Skaergaard intrusion, East Greenland. Mineralogical Magazine (1954) 30:367–375.
Chai G., Eckstrand R. Rare–earth element characteristics and origin of the Sudbury Igneous Complex, Ontario, Canada. Chemical Geology (1994) 113:221–244.[Web of Science]
Chubb P. T., Vogel D. C., Peck D. C., James R. S., Keays R. R. Occurrences of pseudotachylyte at the East Bull Lake and Shakespeare–Dunlop Intrusions, Ontario, Canada. Canadian Journal of Earth Sciences (1994) 31:1744–1748.
Condie K. C., Bobrow D. J., Card K. D. Geochemistry of Precambrian mafic dykes from the southern Superior Province of the Canadian Shield. In: Mafic Dyke Swarms—Halls H. C., Fahrig W. F., eds. (1987) 34. Geological Association of Canada, Special Paper. 95–108.
Crawford A. J., Falloon T. J., Green D. H. Classification petrogenesis and tectonic setting of boninites. In: Boninites—Crawford A. J., ed. (1989) London: Unwin Hyman. 1–49.
Crocket J. H., MacRae W. E. Platinum–group element distribution in komatiitic and tholeiitic volcanic rocks from Munro Township, Ontario. Economic Geology (1986) 81:1242–1252.
Dressler B. O., Gupta V. K., Muir T. L. The Sudbury Structure. In: Geology of Ontario. Ontario Geological Survey, Special Volume—Thurston P. C., Williams H. R., Sutcliffe R. H., Stott G. M., eds. (1991) 4(1):593–625.
Fujimaki H., Tatsumoto M., Aoki K. Partition coefficients of Hf, Zr and REE between phenocrysts and groundmasses. Proceedings of the 14th Lunar and Planetary Science Conference, Part 2. Journal of Geophysical Research Supplement (1984) 89:B662–672.
Golightly J. P. The Sudbury Igneous Complex as an impact melt: evolution and ore genesis. Proceedings of the Sudbury–Noril'sk Symposium, Sudbury. Ontario Geological Survey, Special Volume (1994) 5:105–117.
Gupta V. K. Bouguer gravity of Ontario, east–central sheet. Ontario Geological Survey, Map 2594, scale 1:1 000 000 (1991a).
Gupta V. K. Vertical magnetic gradient of Ontario, east–central sheet. Ontario Geological Survey, Map 2590, scale 1:1 000 000 (1991b).
Gupta V. K., Grant F. S., Card K. D. Gravity and magnetic characteristics of the Sudbury Structure. In: The Geology and Ore Deposits of the Sudbury Structure. Ontario Geological Survey, Special Volume—Pye E. G., Naldrett A. J., Giblin P. E., eds. (1984) 1:381–410.
Halls H. C., Bates M. P. Evolution of the 2.45 Ga Matachewan Dyke swarm, Canada. In: Mafic Dikes and Emplacement Mechanisms—Parker A. J., Rickwood P. C., Tucker D. H., eds. (1990) Rotterdam: A. A, Balkema. 237–249.
Hamlyn P. R., Keays R. R. Sulfur saturation and second–stage melts: application to the Bushveld Platinum Metal Deposits. Economic Geology (1986) 81:1431–1445.
Hamlyn P. R., Keays R. R., Cameron W. E., Crawford A. J., Waldron H. M. Precious metals in magnesian low–Ti lavas: implications for metallogenesis and sulfur saturation in primary magmas. Geochimica et Cosmochimica Acta (1985) 9:1797–1811.
Haughton D. R., Roeder P. L., Skinner B. J. Solubility of sulfur in mafic magmas. Economic Geology (1974) 69:451–467.
Heaman L. M. U–Pb dating of mafic rocks: past, present, future. Abstract Volume of Geological Association of Canada–Mineralogical Association of Canada Annual Meeting, Victoria (1995) A43.
Henderson P. Inorganic Geochemistry (1982) Oxford: Pergamon. 368.
Hess H. H. Stillwater Igneous Complex, Montana—a quantitative mineralogical study. Geological Society of America, Memoir (1960) 80:230.
Hoatson D. M., Keays R. R. Formation of platiniferous sulfide horizons by crystal fractionation and magma mixing in the Munni Munni Layered Intrusion, West Pilbara Block, Western Australia. Economic Geology (1989) 84:1775–1804.
Innes D. G. Proterozoic volcanism in the Southern Province of the Canadian Shield. (1978) Sudbury, Ont.: Laurentian University. 161. M.Sc. Thesis.
Innes D. G., Colvine A. C. The regional metallogenetic setting of Sudbury. In: The Geology and Ore Deposits of the Sudbury Structure. Ontario Geological Survey, Special Volume—Pye E. G., Naldrett A. J., Giblin P. E., eds. (1984) 1:431–446.
Irvine T. N. Crystallization sequences in the Muskox intrusion and other layered intrusion, I. Olivine–pyroxene–plagioclase relations. Geological Society of South Africa, Special Publication (1970) 1:441–476.
Irvine T. N. Crystallization sequences of the Muskox intrusion and other layered intrusions—II. Origin of chromitite layers and similar deposits of other magmatic ores. Geochimica et Cosmochimica Acta (1975) 39:991–1020.[Web of Science]
Irvine T. N., Keith D. W., Dodd S. G. The J–M platinum palladium reef: II. Origin by double–diffusive convective magma mixing and implications for the Bushveld Complex. Economic Geology (1983) 78:1287–1334.
James R. S., Born P. Geology and geochemistry of the East Bull Lake intrusion, District of Algoma, Ontario. Canadian Journal of Earth Sciences (1985) 22:968–979.
James R. S., Harris B. J. Geology of the Shakespear–Dunlop layered gabbro–anorthosite intrusion, Ontario. In: Report of Activities, Part A (1977) 77–1A. Geological Survey of Canada, Paper. 411–414.
James R. S., Peck D. C., Keays R. R., Gibson H. L. The Huronian–Nipissing magmatic belt: product of a global, Early Proterozoic tectonic and metallogenic episode. Abstract Volume of Geological Association of Canada–Mineralogical Association of Canada Annual Meeting, Waterloo (1994) A53.
Jenner G. A. Trace element geochemistry of igneous rocks: geochemical nomenclature and analytical geochemistry. In: Trace Element Geochemistry of Volcanic Rocks: Applications for Massive Sulphide Exploration. Geological Association of Canada, Short Course Notes—Wyman D. A., ed. (1996) 12:1–77.
Jensen L. S. A new cation plot for classifying subalkalic volcanic rocks. Ontario Division of Mines, Miscellaneous Paper (1976) 66:22.
Keays R. R. Palladium and iridium in komatiites and associated rocks: application to petrogenetic problems. In: Komatiites—Arndt N. T., Nisbet E. G., eds. (1982) London: George Allen & Unwin. 435–458.
Keays R. R., Vogel D. C., James R. S., Peck D. C., Lightfoot P. C., Prevec S. A. Metallogenic potential of the Huronian–Nipissing magmatic province. Canadian Mineralogist (Program and Abstracts) (1995) 33:932–933.
Kerrich R., Wyman D. A. The trace element systematics of igneous rocks in mineral exploration: an overview. In: Trace Element Geochemistry of Volcanic Rocks: Applications for Massive Sulphide Exploration. Geological Association of Canada, Short Course Notes—Wyman D. A., ed. (1996) 12:1–50.
Krogh T. E., Davis D. W., Corfu F. Precise U–Pb zircon and baddeleyite ages for the Sudbury area. In: The Geology and Ore Deposits of the Sudbury Structure. Ontario Geological Survey, Special Volume—Pye E. G., Naldrett A. J., Giblin P. E., eds. (1984) 1:431–446.
Krogh T. E., Kamo S. L., Bohor B. F. Shock metamorphosed zircons with correlated U–Pb discordance and melt rocks with concordant protolith ages indicate an impact origin for the Sudbury Structure. In: Earth Processes: Reading the Isotopic Code. Geophysical Monograph, American Geophysical Union (1996) 95:343–352.
Lee I., Ripley E. M. Mineralogic and oxygen isotopic studies of open system magmatic processes in the South Kawishiwi Intrusion, Spruce Road area, Duluth Complex, Minnesota. Journal of Petrology (1996) 37:1437–1461.
Lesher C. M., Groves D. I. Controls on the formation of komatiite–associated nickel–copper sulfide deposits. In: Geology and Metallogeny of Copper Deposits—Friedrich G., Genkin A. D., Naldrett A. J., Ridge J. D., Sillitoe R. H., Vokes F. M., eds. (1986) Berlin: Springer-Verlag. 43–62.
Lesher C. M., Phillips G. N., Groves D. I., Campbell I. H. Immobility of REE and most high field–strength elements and first transition series metals during Archaean gold–related hydrothermal alteration of metabasalts at the Hunt Mine, Western Australia. In: The Economics, Geology, Geochemistry and Genesis of Gold Deposits—Ladeira E. A., ed. (1991) Rotterdam: A. A, Balkema. 327–334.
Marsh B. D. On convective style and vigor in sheet–like magma chambers. Journal of Petrology (1989) 30:479–530.
McBirney A. R., Hunter R. H. The cumulate paradigm reconsidered. Journal of Geology (1995) 103:114–122.[Web of Science]
McCrank G. F. D., Kamineni D. C., Ejeckam R. B., Sikorsky R. Geology of the East Bull Lake gabbro–anorthosite pluton, Algoma District, Ontario. Canadian Journal of Earth Sciences (1989) 26:357–375.
Meschede M. A method of discriminating between different types of mid–ocean ridge basalts and continental tholeiites with the Nb–Zr–Y diagram. Chemical Geology (1986) 56:207–218.[Web of Science]
Morse S. A. Kiglapait geochemistry IV: the major elements. Geochimica et Cosmochimica Acta (1981) 45:461–479.[Web of Science]
Mullen E. D. MnO/TiO2/P2O5: a minor element discriminant for basaltic rocks of oceanic environments and its implications for petrogenesis. Earth and Planetary Science Letters (1983) 62:53–62.[Web of Science]
Nakamura N. Determination of REE, Ba, Fe, Mg, Na and K in carbonaceous and ordinary chondrites. Geochimica et Cosmochimica Acta (1974) 38:757–775.[Web of Science]
Naldrett A. J. Magmatic Sulfide Deposits (1989) New York: Clarendon Press, 186 pp.
Nelson D. O., Morrison D. A., Phinney W. C. Open–system evolution versus source control in basaltic magmas: Matachewan–Hearst dike swarm, Superior Province, Canada. Canadian Journal of Earth Sciences (1990) 27:767–783.
Ohta H., Maruyama S., Takahashi E., Watanabe Y., Kato Y. Field occurrence, geochemistry and petrogenesis of the Archaean Mid–Ocean Ridge Basalts (AMORBs) of the Cleaverville area, Pilbara Craton, Western Australia. Lithos (1996) 37:199–221.[Web of Science]
Pattison E. F. The Sudbury Sub–layer: its characteristics and relationships with the Main Mass of the Sudbury Irruptive. Canadian Mineralogist (1979) 17:257–274.
Peach C. L., Mathez E. A., Keays R. R. Sulphide melt–silicate melt distribution coefficients for the noble metals and other chalcophile metals as deduced from MORB: implications for partial melting. Geochimica et Cosmochimica Acta (1990) 54:3379–3389.[Web of Science]
Peach C. L., Mathez E. A., Keays R. R., Reeves S. J. Experimentally determined sulfide melt–silicate melt partition coefficients for iridium and palladium. Chemical Geology (1994) 117:361–377.[Web of Science]
Pearce J. A., Norry M. J. Petrogenetic implications of Ti, Zr, Y, and Nb variations in volcanic rocks. Contributions to Mineralogy and Petrology (1979) 69:33–47.[Web of Science]
Peck D. C., James R. S. Mineral potential studies of mafic and ultramafic intrusive rocks in the Elliot Lake area. Ontario Geological Survey, Miscellaneous Paper (1990) 151:62–67.
Peck D. C., Keays R. R. Insights into the behaviour of precious metals in primitive, S–undersaturated magmas: evidence from the Heazlewood River Complex, Tasmania. Canadian Mineralogist (1990) 28:553–577.[Web of Science]
Peck D. C., James R. S., Chubb P. T. Geological environments for PGE–Cu–Ni mineralization in the East Bull Lake gabbro–anorthosite intrusion, Ontario. Journal of Exploration and Mining Geology (1993a) 2:85–104.
Peck D. C., James R. S., Chubb P. T., Keays R. R., Reeves S. J., Lightfoot P. C., Kamineni D. C. Precious–metal, chalcophile–element and rare–earth element geochemistry of the Bull Lake area, Districts of Algoma and Sudbury, Ontario. Ontario Geological Survey, Open File Report (1993b) 5849:94.
Peck D. C., James R. S., Chubb P. T., Prevec S. A., Keays R. R. Geology, metallogeny and petrogenesis of the East Bull Lake Intrusion, Ontario. Ontario Geological Survey, Open File Report (1995) 5923:117.
Phinney W. C., Morrison D. A., Maczuga D. E. Anorthosites and related megacrystic units in the evolution of Archean crust. Journal of Petrology (1988) 29:1283–1323.
Popelar J. Gravity interpretation of the Sudbury area. In: New Developments in Sudbury Geology—Guy–Bray J. V., ed. (1971) 11. Geological Association of Canada, Special Paper. 103–115.
Poulson S. R., Ohmoto H. An evaluation of the solubility of sulfide sulfur in silicate melts from experimental data and natural samples. Chemical Geology (1990) 85:57–75.[Web of Science]
Prevec S. A. An isotopic, geochemical and petrographic investigation of the genesis of Early Proterozoic mafic intrusions and associated volcanism near Sudbury, Ontario. (1993) Edmonton: University of Alberta. 223. Ph.D. Thesis.
Rollinson H. Using Geochemical Data: Evaluation, Presentation, Interpretation (1993) Harlow, UK: Longman. 352.
Roscoe S. M., Card K. D. Early Proterozoic tectonics and metallogeny of the Lake Huron region of the Canadian Shield. Precambrian Research (1992) 58:99–119.[Web of Science]
Seitz H.-M., Keays R. R. Platinum group element segregation and mineralization in a Noritic Ring Complex formed from Proterozoic siliceous high magnesian basalt magmas in the Vestfold Hills, Antarctica. Journal of Petrology (1997) 38:703–725.
Smith R. K., Lofgren G. E. An analytical and experimental study of zoning in plagioclase. Lithos (1983) 16:153–168.[Web of Science]
Taylor S. R., McLennan S. M. The Continental Crust: its Composition and Evolution (1985) Oxford: Blackwell. 312.
Tilley C. E., Yoder H. T., Schairer J. F. Melting relations of igneous rock series. Carnegie Institution of Washington, Yearbook (1967) 66:260–269.
Toplis M. J., Carroll M. R. An experimental study of the influence of oxygen fugacity on Fe–Ti oxide stability, phase relations, and mineral–melt equilibria in ferro–basaltic systems. Journal of Petrology (1995) 36:1137–1170.
Vance J. A. Zoning in igneous plagioclase: patchy zoning. Journal of Geology (1965) 73:636–651.
Vogel D. C. The geology and geochemistry of the Agnew Intrusion: implications for the petrogenesis of early Huronian mafic igneous rocks in central Ontario, Canada. (1996) University of Melbourne, Vic. 455. Ph.D. Thesis.
Vogel D. C., Keays R. R. The petrogenesis and platinum–group element geochemistry of the Newer Volcanic Province, Victoria, Australia. Chemical Geology (1997) 136:181–204.[Web of Science]
Vogel D. C., James R. S., Keays R. R., Turcott M. B., Reeves S. J. PGE–Cu–Ni mineralization and potential of the Agnew Intrusion Ontario, Canada. In: Proceedings of the 4th Biennial SGA Meeting, Turku, Finland—Papunen H., ed. (1997) Rotterdam: A. A, Balkema. 513–516.
Vogel D. C., James R. S., Keays R. R. The early tectonomagmatic evolution of the Southern Province: implications from the Agnew Intrusion, central Ontario, Canada. In: Canadian Journal of Earth Sciences (1998a) in press.
Vogel D. C., Vuollo J. I., Alapieti T. T., James R. S. Tectonic, stratigraphic, and geochemical comparisons between ca. 2500–2400 Ma mafic igneous events in the Canadian and Fennoscandian Shields. Precambrian Research (1998b) 92:89–116.[Web of Science]
Vuollo J. I., Nykänen V. M., Liipo J. P., Piirainen T. A. Palaeoproterozoic mafic dyke swarms in the Eastern Fennoscandian shield, Finland: a review. In: Physics and Chemistry of Dyke—Baer G., Heimann A., eds. (1995) Rotterdam: A. A.: Balkema. 179–192.
Wager L. R., Brown G. M. Layered Igneous Rocks (1968) Edinburgh: Oliver and Boyd, 588 pp.
Wager L. R., Brown G. M., Wadsworth W. J. Types of igneous cumulates. Journal of Petrology (1960) 1:73–85.
Wendlandt R. F. Sulfide saturation of basalt and andesite melts at high pressures and temperatures. American Mineralogist (1982) 67:877–885.[Abstract]
Wiebe R. A. Plagioclase stratigraphy: a record of magmatic conditions and events in a granite stock. American Journal of Science (1968) 266:690–703.[Abstract]
Zolnai A. I., Price R. A., Helmstaedt H. Regional cross section of the southern province adjacent to Lake Huron, Ontario: implications for tectonic significance of the Murray Fault Zone. Canadian Journal of Earth Sciences (1984) 21:447–456.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||