The Seal Island Pluton outcrops only on small islands located on the continental shelf 45 km south of Nova Scotia, although geophysical data indicate that the pluton is part of large granitoid units that cover thousands of square kilometres farther offshore. Based on the island outcrops, the Seal Island Pluton consists of biotite monzogranite and muscovite–biotite monzogranite of uncertain relative age. Metasedimentary xenoliths combined with characteristic magnetic patterns indicate that the pluton intruded the Cambrian–Ordovician Meguma Group. Compared with the biotite monzogranite, the muscovite–biotite monzogranite is higher in SiO2, more peraluminous, and more depleted in heavy rare-earth elements, and also has lower εNd (−1.39 versus +0.82), possibly the result of more contamination by Meguma Group sedimentary rocks. The biotite monzogranite yielded a Late Devonian U–Pb (zircon) age of 362.8 ± 0.7 Ma. Although the relatively minor petrological differences between the two units do not preclude a co-magmatic relationship, the muscovite–biotite monzogranite could be 10–15 Ma older than the biotite monzogranite, based on its petrological similarities to parts of the onshore ca. 376–372 Ma Shelburne and Port Mouton plutons. Comparison with granite samples in offshore drill core indicates that granitoid rocks similar to those exposed on Seal and surrounding islands form part of large plutons farther offshore in the Meguma terrane. The age and petrochemical data from both onshore and offshore plutons indicate that peraluminous granitoid rocks in the Meguma terrane were derived from similar sources over a span of at least 20 million years. Magma genesis may have been related to mantle upwelling and stepping back of the subduction zone to the southeast subsequent to docking of Meguma terrane with adjacent Avalonia.
Meguma, the most outboard of northern Appalachian terranes (e.g., Hibbard et al. 2006), is exposed on land only in southern Nova Scotia, although geophysical data and sparse drill core samples (e.g., Jansa and Wade 1975; Hutchinson et al. 1988; Pe-Piper and Loncarevic 1989; Pe-Piper and Jansa 1999) indicate that the terrane extends far offshore to the east and south (Fig. 1, inset). The Seal Island Pluton is the southernmost subaerial exposure of rocks of the Meguma terrane, outcropping on Seal and other nearby small islands 45 km south of mainland Nova Scotia (Fig. 1). The granite thus occupies an important position in correlation between onshore and offshore parts of the terrane, in particular with the large areas of varied granitoid rocks that apparently dominate the terrane offshore from southern Nova Scotia (Pe-Piper and Loncarevic 1989; Todd et al. 2005; Brake 2006). The current study was undertaken to obtain a reliable age for the Seal Island Pluton and to better characterize its petrological features in comparison with other plutonic units of the Meguma terrane, both onshore and offshore.
The Meguma terrane (Fig. 1) consists mainly of Cambrian to Lower Ordovician clastic metasedimentary rocks of the Meguma Group, disconformably overlain by Early Silurian and Early Devonian metasedimentary and metavolcanic rocks, and intruded by Devonian granitoid rocks dominated by the South Mountain Batholith (Keppie 2000; White et al. 2006). Petrological differences between the South Mountain Batholith and plutons peripheral to the batholith to the southwest (Port Mouton, Shelburne, and Barrington Passage; Fig. 1) have been attributed to different origins—the central plutons derived entirely from the crust (sub-Meguma Group crust combined with contamination from the Meguma Group) and the peripheral plutons from mixed sources (sub- Meguma Group crustal source mixed with mantle-derived mafic magmas) (Clarke et al. 1997; Tate and Clarke 1997). These earlier workers also suggested that the peripheral plutons may have somewhat older crystallization ages than the central plutons (ca. 376 Ma versus ca. 372 Ma). However, the postulated age difference has not been borne out by more recent dating, which has suggested that at least parts of the South Mountain Batholith may be slightly older, having crystallized at about 381 Ma (Reynolds et al. 2004), than the peripheral plutons, which have igneous crystallization ages around 378–372 Ma (Currie et al. 1998; Keppie and Krogh 1999; White 2003 and references therein).
The Wedgeport Pluton (Fig. 1) was excluded from the central-peripheral designation (e.g., Clarke et al. 1997; Tate and Clarke 1997) because it had yielded significantly younger U–Pb and Rb–Sr ages of ca. 316 Ma (Cormier et al. 1988). However, more recent U–Pb (zircon) dating has shown that these dates did not reflect the igneous crystallization age of the pluton, now known to be 357 ± 1 Ma (MacLean et al. 2003), about 15 million years younger than the nearby Barrington Passage and Shelburne plutons.
In the offshore Meguma terrane, granitic samples from drill holes located about 40–50 km west and south of Seal Island (Fig. 1, inset) were divided into two groups by Pe-Piper and Loncarevic (1989): an ilmenite-bearing peraluminous group (represented by cores 14 and 17) and a slightly less peraluminous magnetite-bearing group (represented by cores 20 and 21). The ilmenite-bearing rocks were correlated tentatively with granite of the Seal Island Pluton, based on the limited petrologic information of Rogers (1988), and with the South Mountain Batholith, whereas no correlative of the magnetite-bearing unit appears to occur in the onshore (Pe-Piper and Loncarevic 1989).
Field relations and petrography
Outcrop is relatively sparse on Seal and surrounding islands, as the shorelines are mainly covered by granitic boulders that are lithologically similar to and likely derived from the underlying pluton. However, the presence of less abundant metasedimentary boulders of andalusite hornfels and migmatite, typical lithologies of the Meguma Group, as well as the rare occurrence of similar metasedimentary xenoliths in the Seal Island granite, suggest that the pluton intruded the Meguma Group. This interpretation is supported further by aeromagnetic data that show typical Meguma Group magnetic patterns extending into the area from the offshore (White and King 2002; White 2003; Brake 2006).
The geographical extent of the islands on which the Seal Island Pluton outcrops (Fig. 2) suggests that the minimum area of the pluton is 100 km2. However, Watts (1974) interpreted a large low gravity anomaly, including the Seal Island area, to be caused by a granitic body that extends as much as 100 km to the south toward the continental slope. This interpretation was corroborated by Brake (2006), who interpreted the Seal Island Pluton to extend at least 100 km to the southwest, with an area of several thousand square kilometres. Her geophysical modeling also suggested that the pluton is composite, with physical properties different from those of the granitic rocks sampled in the drill holes studied by Pe-Piper and Loncarevic (1989).
Based on mapping and sampling during the current project, combined with the field observations and sample collection of Rogers (1988), the Seal Island Pluton is divided into two units (Fig. 2). The northern unit, outcropping on Round, Mud, and Noddy islands, is coarse-grained, inequigranular to porphyritic, biotite monzogranite with phenocrysts of K-feldspar. The southern unit, which occurs on Seal Island and adjacent small islands to the west, is also coarse-grained but more equigranular, and contains muscovite as well as biotite. The southern unit tends to have more quartz and K-feldspar, and grades to syenogranite. Biotite forms about 15% of the biotite monzogranite, but is less abundant (about 10%) and finer grained in the muscovite–biotite monzogranite, where it occurs in clusters with muscovite. Zircon, titanite, apatite, and ilmenite are common, mostly as inclusions in biotite, and are more abundant in the biotite monzogranite than in the muscovite–biotite monzogranite. Tourmaline layers and patches locally compose up to 2% of the muscovite–biotite monzogranite. The biotite monzogranite and muscovite–biotite monzogranite units are not seen in contact, and hence their relative age is not known.
Weak foliation resulting from preferred orientation of biotite ± muscovite is present in most outcrops of both units, and generally trends north–south. Steep, post-magmatic shear zones (<50 cm wide) occur throughout the pluton but are more abundant along the western side of Seal Island. These zones trend north to north-northwest and contain subhorizontal mineral lineations, locally well developed C–S fabrics, and asymmetric porphyroclasts indicating dextral strike-slip sense of shear. The shear zones may be related to offshore extensions of major shear zones recognized onshore, such as the Chebogue Point Shear Zone of Culshaw and Leisha (1997) and White (2003).
Because of the importance of potential field modeling in determining the extent of offshore plutons (e.g., Brake 2006), magnetic susceptibility was measured in outcrops of the Seal Island Pluton using a KT-9 magnetic susceptibility meter manufactured by Exploranium G.S. Systems Limited, Mississauga, Ont. Values are low in both the biotite monzogranite and the muscovite–biotite monzogranite, ranging from 0.00 to 0.35 × 10−3 SI units (Moran 2005); the biotite monzogranite tends to have slightly higher values (average 0.12) than the muscovite–biotite monzogranite (average 0.09). These values are similar to those measured in granite of the Shelburne and Wedgeport plutons (both average 0.07 × 10−3 SI units; C. White, unpublished data, 2003), and also to data reported by Pe-Piper and Loncarevic (1989) for cores 14 and 17 from south and west of Seal Island (Fig. 1), but are much lower than those reported from cores 20 and 21 (6.5 and 10.1 × 10−3 SI units, respectively).
A 20 kg biotite monzogranite sample (P5-W02-81) was collected for dating from outcrop on the eastern shore of Mud Island (Fig. 2), and processed in the Jack Satterly Geochronology Laboratory at the University of Toronto, Ont., using standard techniques (Table 1, footnote). Highest quality zircon grains were selected from the least paramagnetic separates by hand picking under a binocular microscope and were then air-abraded according to the method of Krogh (1982). Weights of each final, selected fraction (consisting of one to three grains) were determined by making a volume estimate from digital photomicrographs and using the density of zircon. Uncertainties in the calculated weights are estimated to be <20% in most cases. This uncertainty affects only the calculation of Pb and U concentrations and has no influence on age data.
Sample P5-W02-81 yielded abundant high-quality zircon, dominated by moderately flat euhedra and elongate (up to 6:1) prismatic morphological types. Length-parallel cracks were observed in the latter population, which therefore was not selected for further analysis. Instead, best quality, waterclear, colourless flat euhedra (and fragments thereof) were chosen for dating and the results for five fractions, comprising between 1 and 3 grains each, are presented in Table 1. Uranium concentrations in these zircon grains are moderately high, ranging from ∼700 to 1000 ppm. Calculated Th/U ratios cluster mostly between 0.30 and 0.36, falling on the low side of typical igneous compositions; and the flattest grain analyzed (B1) shows an even lower Th/U ratio of 0.22.
The data are clustered on or very near concordia (3 out of 5 have error ellipses that straddle concordia significantly), and have 206Pb/238 U ages that fall tightly between 362.2 and 363.4 Ma (Fig. 3). A weighted mean 206Pb/238U age calculated on the basis of all five fractions is 362.8 ± 0.7 Ma and, although the 2σ errors for each fraction are overlapping, the associated mean square of weighted deviates (MSWD) for this average is relatively high (3.1). By comparison, the 207Pb/206Pb ages for these fractions range from 365.0 to 368.6 Ma, and by assuming that the fractions are displaced along a zero-aged Pb-loss trajectory, they imply an upper intercept age of Ma with a lower MSWD (0.7) and somewhat better probability of fit. However, the fact that the data are clustered and mostly straddling concordia leads us to favour the mean 206Pb/238U age of 362.8 ± 0.7 Ma as the most accurate age for emplacement and crystallization of the biotite monzogranite unit of the Seal Island Pluton.
We acquired major and trace element data for 21 samples, in addition to the 5 samples from Rogers (1988) (Table 2), for a combined total of 12 samples from the biotite monzogranite unit and 14 samples from the muscovite– biotite monzogranite unit (Fig. 2). Rare-earth element (REE) data were obtained for five of these samples (Table 3), and Sm–Nd isotopic data were obtained for the dated biotite monzogranite sample and a muscovite–biotite monzogranite sample (Table 4). Analytical methods are summarized in the table footnotes.
Silica concentrations are lower in the biotite monzogranite samples (average 70%) than in the biotite–muscovite monzogranite samples (average 73.7%). The samples from the two units form more or less continuous negative correlation trends on silica variation diagrams for most major elements, with an overlap in SiO2 contents at about 72% (e.g., Figs. 4a–4c). Exceptions are MgO, which is lower in the muscovite–biotite monzogranite (Fig. 4d); K2O, which is essentially the same in both units (Fig. 4e); and P2O5, which is higher in the muscovite–biotite monzogranite (Fig. 4f). Trace elements also generally show continuous trends between the two units (e.g., Figs. 4g–4k). Crystal fractionation involving plagioclase, K-feldspar, and biotite could account for much of the chemical variation within and between the two units, although the lack of continuous variation in some components between the two units suggests that they may not be gradational.
This suggestion is further supported by differences in chondrite-normalized REE patterns (Fig. 5a). Although the light REE patterns for the two units are similar, the biotite monzogranite samples have flat heavy REE patterns whereas the muscovite–biotite monzogranite samples show a steady decrease in heavy REE. That decrease may be the result of removal of accessory phases such as zircon with high concentrations of heavy REE, possibly through more biotite fractionation during the evolution of that unit. Both units show similar negative Eu values, consistent with feldspar fractionation.
The biotite monzogranite has a higher εNd value (+0.82) than the muscovite–biotite monzogranite (−1.39), both calculated at 363 Ma (Table 4). These values suggest that the source region was relatively juvenile (model ages are 821 and 1189 Ma; Table 4), and do not allow for extensive contamination from the metasedimentary host rocks of the Meguma Group, which have very negative εNd values of ca. −7 to −12 (Clarke and Halliday 1985; Currie et al. 1998). The more negative εNd value for the muscovite–biotite monzogranite sample is consistent with more contamination by metasedimentary material, consistent with its more peraluminous composition (average molar Al2O3/(CaO + Na2O + K2O) = 1.27 compared with 1.07 in the biotite monzogranite; Fig. 6a).
Overall, the petrological data do not resolve with certainty the question of whether or not the two rock types in the Seal Island Pluton are comagmatic. However, mineralogical and chemical differences between them are minor, and do not preclude that relationship.
Comparison with other granitoid rocks of the Meguma terrane
The U–Pb (zircon) age of 362.8 ± 0.7 Ma shows that the biotite monzogranite unit of the Seal Island Pluton is only a few million years older than biotite monzogranite of the Wedgeport Pluton, located about 40 km to the north (Fig. 1), which yielded a U–Pb (zircon) age of 357 ± 1 Ma (MacLean et al. 2003). However, the Wedgeport biotite monzogranite differs mineralogically from the Seal Island biotite monzogranite by having proportionately more quartz and less biotite, and in the presence of accessory garnet, epidote, titanite, and fluorite (MacLean et al. 2003). Chemically, it has higher SiO2 content than the Seal Island biotite monzogranite, and related other chemical differences, most notably higher Rb, Th, Ta, Y, and Yb, and lower Ba and P (Figs. 4, 7a). Based mainly on its mineralogical and chemical features (e.g., Fig. 6b), MacLean et al. (2003) suggested that the Wedgeport Pluton has more A-type and within-plate characteristics than other Meguma terrane granites.
The age obtained from the Seal Island biotite monzogranite is about 15 million years younger than the ca. 376–372 Ma ages of the Barrington Passage, Shelburne, and Port Mouton plutons (Currie et al. 1998; Keppie and Krogh 1999). The Barrington Passage Pluton is composed of tonalite with lower SiO2 than the Seal Island units (e.g., Fig. 4). However, muscovite–biotite granodiorite (gradational to monzogranite) similar to the Seal Island Pluton muscovite–biotite monzogranite is a major component in both the Shelburne and Port Mouton plutons (e.g., Rogers and Barr 1988; Douma 1988; Currie et al. 1998). Similarities in the average chemical compositions of these muscovite-bearing units to those of the Seal Island Pluton (e.g., Figs. 4, 5b, 7a) suggest the possibility that the Seal Island muscovite–biotite monzogranite could be related to those older units and hence might not be of the same age as the geographically associated biotite monzogranite. Resolution of this uncertainty will require more geochronology.
Whatever the age of the muscovite–biotite monzogranite, the 363 Ma age from the Seal Island biotite monzogranite demonstrates that Meguma terrane granites, including South Mountain Batholith, with their characteristic peraluminous signatures (Fig. 6a) and εNd values that range from slightly positive to slightly negative (Fig. 8), were generated over a time span of at least 15–20 Ma, an important constraint on tectonic models as it suggests that the source area and magma genesis processes did not change significantly during that time. The new data from the Seal Island monzogranite units are consistent with the trend toward higher εNd values in the onshore peripheral plutons previously noted and attributed to more interaction with mantle-derived magma compared with the South Mountain Batholith (e.g., Clarke et al. 1997; Tate and Clarke 1997).
Turning to the offshore plutons, of which the Seal Island Pluton is a part based on geophysical data (e.g., Pe-Piper and Loncarevic 1989; Brake 2006), the similarities between the muscovite–biotite monzogranite of cores 14 and 17 (Fig. 1, inset) and the Seal Island Pluton, as suggested by Pe-Piper and Loncarevic (1989), are borne out by the now larger database from the Seal Island Pluton. However, the chemical match is best between the core samples and the Seal Island biotite monzogranite, rather than the muscovite–biotite monzogranite. The core samples have generally lower SiO2 contents than the Seal Island muscovite–biotite monzogranite (Fig. 4), and they do not show the same depletion in heavy REE (Fig. 5c). Samples from the other two granite cores, 20 and 21 (Fig. 1, inset) have less similarity to the Seal Island units. Two of the three samples from core 20 are tonalitic, with SiO2 contents of 63%–65% (Pe-Piper and Loncarevic 1989). Although similar in SiO2 content to tonalitic units from the Barrington Passage Pluton, they show other chemical differences that do not support correlation with that unit (Fig. 4). The third core 20 sample is biotite monzogranite, with SiO2 content of 73%, in the range of the Seal Island muscovite–biotite monzogranite samples (Fig. 4). However, the core 20 biotite monzogranite sample differs in its lower TiO2 and P2O5, and higher K2O (Figs. 4a, 4e, 4f), as well as other chemical differences (Fig. 7b). The samples from core 21 are also granitic, with SiO2 contents similar to the Seal Island biotite monzogranite, but they are characterized by higher Ba, Sr, Y, and Zr (Figs. 4g, 4i–4k). Furthermore, as reported by Pe-Piper and Jansa (1999), the core 20 and 21 samples have negative εNd values (Fig. 8), and higher magnetic susceptibilities than the Seal Island monzogranite or onshore units. They were classified as magnetite–granite by Pe-Piper and Loncarevic (1989), who also suggested that they may be younger (<326 Ma) based on limited U–Pb (zircon) data.
Overall, the available data indicate that at least some of the voluminous granitoid rocks that underlie the offshore Meguma terrane are similar in composition and presumably age to the Seal Island Pluton, but granitoid bodies of different composition and possibly different age are also located there.
Tectonic setting of the Seal Island Pluton
The Seal Island monzogranite units, like other Meguma terrane granites, have chemical features that suggest that they formed in a subduction zone setting. For example, they plot in the volcanic-arc field on the Rb–Y + Nb discrimination diagram (Fig. 6b), and show depletions in Nb, P, and Ti typical of magmas generated in subduction zone settings (Figs. 7a, 7b). However, the significance of these chemical features in felsic rocks is equivocal and the tectonic setting of the Meguma terrane in the Devonian is not yet well understood. For example, Pe-Piper and Jansa (1999) suggested that the subduction-related chemical characteristics of Meguma terrane rocks do not reflect a direct link to subduction but were inherited from a mantle source previously influenced by Late Proterozoic Pan-African subduction. They proposed that Meguma terrane plutons formed in an extensional regime, resulting from transtensional shear between the docking Laurentian and Gondwanan plates, and that decompression melting of the upper mantle produced mafic magmas, which mixed with crustal material and also induced crustal melting. In this scenario, the Seal Island Pluton would have been a manifestation of the same general processes that produced granitoid and volcanic rocks of similar age, although different petrological character in Avalonia north of the Cobequid–Chedabucto fault system in northern mainland Nova Scotia (e.g., Dunning et al. 2002). Our data do not disprove the Pe-Piper and Jansa (1999) interpretation, but it seems unlikely that voluminous magmatism would have occurred so far from the Cobequid–Chedabucto fault system, the locus of transtension between Meguma and Avalonia.
models have invoked northwesterly subduction of the intervening Rheic Ocean as Meguma terrane converged obliquely with Avalonia, which by mid- Devonian formed the leading edge of Laurentia (e.g., Clarke et al. 1997; Murphy et al. 1999; van Staal 2007, in press). In their model, Clarke et al. (1997) emphasized the subductionrelated characteristics of minor mafic rocks associated with the peripheral plutons, and suggested that those plutons formed by crustal melting caused by intrusion of the subduction-related mafic magmas prior to the final emplacement of the Meguma terrane against Avalonia. They attributed the South Mountain Batholith, which at that time was thought to be younger than the peripheral plutons, to subsequent crustal thickening. However, this overall scenario is less well supported by the younger ages now known for the plutons of the southern Meguma terrane, especially with the ca. 363 Ma Seal Island and ca. 357 Ma Wedgeport plutons included, and with the large volume of likely comagmatic plutons farther offshore.
We prefer a model that links ongoing plutonism in the southern part of the Meguma terrane to renewed subduction from the southeast after Meguma–Avalonia collision (Fig. 9). This model incorporates flat-slab subduction of Rheic Ocean crust and subsequent mantle upwelling and Rheic crust delamination, as proposed by Murphy et al. (1999) and van Staal (in press). The combination of oblique collision (Neoacadian orogeny of van Staal, in press) and flat-slab subduction can explain the relative lack of magmatism in much of Avalonia at this time, and is analogous to the earlier Avalonia–Ganderia relationship during and after the Acadian orogeny, which resulted in voluminous granitoid plutons in more inboard parts of the Appalachian orogen (van Staal 2007, in press). Decreasing age of magmatism in the Meguma terrane could be linked to the stepping back of the locus of subduction to a position now under the continental margin well outboard of exposed Meguma terrane rock, prior to Carboniferous accretion of Gondwana (Fig. 9). The key to confirming or disproving such a speculative model lies in knowing the age(s) and more detailed petrochemical characteristics of the granitoid rocks that apparently form much of the offshore part of Meguma terrane south of the Seal Island Pluton. Related questions include whether or not co-genetic volcanic rocks are present in the offshore parts of Meguma terrane, or even the relict suture with Gondwana.
The Seal Island Pluton, located 45 km offshore from southwestern Nova Scotia, consists of biotite monzogranite and muscovite–biotite monzogranite units. These rocks outcrop on separate small islands and were not observed in contact. Petrological similarities indicate that they are most likely closely related, although perhaps not linked directly by crystal fractionation processes. U–Pb dating of zircon yielded an igneous crystallization age of 362.8 ± 0.7 Ma for the biotite monzogranite unit. The undated muscovite–biotite monzogranite could be as much as 10–15 Ma older, based on its petrological similarities to parts of the onshore Shelburne and Port Mouton plutons, dated at 376–372 Ma. Geophysical data and samples from drill core suggest that granitoid rocks similar to those exposed on Seal and adjacent small islands form large plutons farther offshore in the Meguma terrane. The data suggest that even larger bodies of younger(?), more magnetic granitoid rocks are also present in the area. In combination, the age and petrochemical data from both onshore and offshore plutons indicate that peraluminous granitoid magmas were derived from similar sources over a span of at least 20 million years, subsequent to the docking of Meguma terrane with adjacent Avalonia. Magma genesis may have been related to lithospheric delamination, mantle upwelling, and stepping back of the subduction zone to the southeast. This study draws further attention to the importance of data from the offshore in understanding the tectonic setting in which these granitoid rocks were formed. Although much progress has been made in understanding the granitoid plutons of southwestern Nova Scotia since the days when even their ages were a matter of speculation (e.g., Rogers and Barr 1988), the aerially extensive and relatively little-known plutons under the continental shelf and slope no doubt hold important clues to the rest of the story.
This project was funded by the Nova Scotia Department of Natural Resources through the South Nova Mapping Project, and by a Discovery Grant to S.M. Barr from the Natural Sciences and Engineering Research Council of Canada. It incorporates data from the B.Sc. honours thesis of P.C. Moran. We thank the journal reviewers, G. Pe-Piper and P. Gromet, and journal Associate Editor Louise Corriveau for their helpful comments, which led to improvements in the manuscript.
- Received September 27, 2006.
- Accepted April 19, 2007.
- Published on the NRC Research Press Web site at http://cjes.nrc.ca on October 27, 2007.
- © 2007 NRC Canada