The Brazil Lake pegmatite (BLP), southwestern Nova Scotia, is a Li-Cs-Ta (LCT)-type ( erný 1991a, b) as manifested by the presence of abundant, coarse (i.e., m-scale) spodumene, rubidium-rich K-feldspar and mica, and Nb-Ta oxide mineralization (Hutchinson 1982; Corey 1993, 1995; Hughes 1995; Kontak 2004). Although the pegmatite was first discovered in 1960 (Taylor 1967) and has since been the focus of exploration and petrological studies (Corey 1993, 1995; Hughes 1995), the age of emplacement is constrained only to less than that of the host rocks (i.e., ca. 440 Ma White Rock Formation; MacDonald et al. 2002), with possible genetic links to magmatic or meta-morphic events. For example, the BLP may have a magmatic association, with either the 440 Ma Brenton Pluton (Keppie & Krogh 2000) or 380 Ma South Mountain Batholith (Kontak et al. 2003) as progenitors, or a metamorphic origin relating to the ca. 400– 410 Ma Acadian Orogeny. The ambiguous genetic affiliation of LCT-type pegmatite has been noted and discussed by erný (1991a) and continues to be a subject of debate (e.g., London 1992, 1996). Given that the age of the BLP will strongly influence any genetic interpretation and by inference have clear implications for the petrogenesis of the BLP and, more importantly, the metallogenetic potential of the southwest Nova Scotia Sn-base metal domain (Chatterjee 1983), a project was initiated to resolve the time of pegmatite formation and subsequent thermal history. A multi-disciplinary geochronological approach was used that involved U-Pb dating of tantalite, Re-Os dating of molybdenite, and ⁴⁰Ar/³⁹Ar dating of muscovite. The results of these analyses are presented and discussed below.

The study area occurs in the southwestern part of the Meguma terrane of the Canadian Appalachian Orogen and is dominated by Lower Paleozoic metasedimentary and metavolcanic rocks and Devonian granitic intrusions (Fig. 1a). The area of interest is located just past the western edge of the 380 Ma (Reynolds et al. 2004; Kontak et al. 2003), peraluminous South Mountain Batholith and some 40 km from the past producing (1985– 1992) East Kemptville tin deposit (Fig. 1b). Importantly, many other lithophile-element mineralized centres occur in this region, informally referred to as the southwest Nova Scotia tin-base metal domain (Chatterjee 1983) and discussed in detail by Kontak et al. (1989). The Brazil Lake area is underlain by a mixed package of metasedimentary and metavolcanic rocks that are cut by small pegmatite bodies and the Brenton Pluton, the latter of which is also transected by pegmatite (Fig. 2a).

The oldest rocks surrounding the study area (Fig. 1b) are the Cambrian-Ordovician Meguma Group that consists of an older, sandstone-dominant Goldenville Formation and younger overlying siltstone- and shale-dominant rocks of the Halifax Formation. Overlying the Meguma Group are mixed metasedimentary and metavolcanic rocks of the Silurian White Rock Formation. The metasedimentary rocks are dominated by massive quartzite with minor pelitic beds, whereas the metavolcanic rocks are basaltic, although southwest of the study area around Yarmouth minor felsic rocks also occur (White et al. 2001; MacDonald et al. 2002). A U-Pb zircon age of 438 ± 3 Ma was obtained for the felsic rocks (MacDonald et al. 2002). The Brenton pluton, a probable subvolcanic intrusion related to the White Rock Formation based on its chemistry (MacDonald et al. 2002), has a U-Pb zircon age of 439 ± 4 Ma (Keppie & Krogh 2000), and is in fault contact with both the White Rock and Halifax formations (Fig. 1b). The Brenton pluton consists of biotite-muscovite syenogranite to monzogranite and is highly deformed with a penetrative Acadian fabric. Within the intrusion are two mica-bearing, garnet-tourmaline pegmatite bodies (Fig. 2a) that commonly cut and retrograde the fabric of the granite (Fig. 2b).

Acadian deformation in the area produced northeast-trending, upright folds with a shallow southwest plunge and a related penetrative fabric; regional scale shear zones are present throughout the study area (Fig. 1b). An overprinting fabric of Alleghanian age (Culshaw & Liesa 1997; Culshaw & Reynolds 1997; Moynihan 2003) has crenulated the S1 fabric with the orientations of related fabrics and lineations depending on geographic location (Culshaw & Liesa 1997; White et al. 2001; Moynihan 2003). Metamorphic grade varies from greenschist- to amphibolite-facies, with the higher grade rocks localized to the contact areas between the White Rock and Goldenville formations on the margins of the Yarmouth syncline (MacDonald, 2000; MacDonald et al. 2002). Peak meta-morphic conditions were estimated at 4 kbars and 550– 600°C by Moynihan (2003). This change in metamorphic grade also coincides with a boundary between two distinct trends of vertical gradient magnetic anomalies, which OReilly et al. (1992) termed the Deerfield Shear Zone and which White et al. (2001) later mapped in as the Chebogue Point Shear Zone, following the terminology of Culshaw & Liesa (1997). These two zones merge (CP-DSZ in Fig. 1b), but a formal name has yet to be used for the laterally extensive shear zone (C. White, personal communication 2004).

The BLP outcrops on the eastern side of the Yarmouth syncline within rocks of the White Rock Formation (Fig. 1b). Rock types in the area include fine- to coarse-grained amphibolite, minor mafic tuff, dark green-black pelite, psammite, and adjacent to the pegmatite lenses, fine-grained tourmaline-bearing, silica-rich rock. The pelitic unit may contain staurolite, garnet, and andalusite, whereas tourmaline and rare holmquistite (Hutchinson 1982) occur in the amphibolite adjacent to the pegmatite lenses. The quartzite adjacent to the pegmatite lenses is mainly buff, but locally the buff color gives way to white which may in part be an exomorphic alteration (i.e., silicification) related to pegmatite emplacement (D. Black, personal communication 2003). In addition, thin (cm-scale) white seams of very pure quartzite (perhaps representing re-crystallized quartz veins?) cut the buff quartzite.

The BLP consists of two separate pegmatite lens, each with a subcrop extent of a few hundred m length and ≤20– 25 m width, that strike northeasterly with near vertical to steeply southeast dips (Fig. 1c). The pegmatite lenses have complex internal zonation and are dominated by areas of megacrystic (i.e., 1– 2 m) blocky K-feldspar and spodumene with an inter-granular matrix with variable proportions of quartz-feldsparspodumene-muscovite. Formation of secondary albite after K-feldspar is manifest as either fine-grained sacchroidal albite or euhedral platy cleavelandite. These zones of secondary al-bite, previously been referred to as albitic aplite (Hutchinson 1982; Hughes 1995), are locally pervasive and such areas are preferentially enriched in Ta-Nb and phosphate minerals (Kontak 2004).

The metasedimentary and mafic metavolcanic (amphibolite) rocks of the area are strongly deformed and contain a penetrative fabric oriented northeast with near-vertical dip. Within the amphibolite, small (≤1 cm), elongate lenses of plagioclase have a vertical plunge. Although the pegmatite is generally iso-tropic and the megacrysts of K-feldspar and spodumene in fact have the long axes of crystals oriented perpendicular to the wall rock fabric, the pegmatite locally contains a fabric, as manifest by : (1) small shears through muscovite-rich zones (Fig. 2c, d); (2) flattened quartz megacrysts with 10– 20:1 aspect ratios (Fig. 2e, f); (3) a penetrative micro-fabric in the albitic aplite zones (Fig. 2g); and (4) rare cases of kinked spodumene crystals in the intergranular areas adjacent to large megacrysts (Fig. 2h). In these cases microfolds have subvertical fold axes, thus similar in orientation to the plunge of plagioclase lenses in the amphibolite unit. In addition, we note that the boundary between the pegmatite and wall rock is not sheared, thus no preferential partitioning of strain occurred along this horizon.

The mineral assemblages in the metamorphic rocks in the area of the BLP were used by Moynihan (2003) to estimate peak metamorphic conditions at near 550– 600°C and 4 kbars. Interestingly, these conditions are close to or exceed the solidus of a Li-rich felsic melt at such pressures (London 1992).

The timing of magmatic and tectono-metamorphic events in the East Kemptville-Yarmouth area of southwestern Nova Scotia is constrained by U-Pb zircon and monazite, Rb-Sr whole rock, Re-Os molybdenite, and ⁴⁰Ar/³⁹Ar whole rock and mineral ages. Relevant to this study are the following: (1) 439 ± 4 Ma mafic-felsic volcanic rocks of the White Rock Formation and the co-magmatic Brenton Pluton (U-Pb zircon; Keppie & Krogh 2000; MacDonald et al. 2002); (2) intrusion at 375 Ma of the Davis Lake Pluton and related mineralization at East Kemptville (Rb-Sr whole rock and Re-Os molybdenite, respectively; Dostal & Chatterjee 1995; Kontak et al. 2003); (3) regional Acadian deformation and metamorphism at ca. 400 Ma (⁴⁰Ar/³⁹Ar; Muecke et al. 1988) and overprinting Alleghanian deformation at 320– 300 Ma (Rb-Sr and ⁴⁰Ar/³⁹Ar dating: Kontak & Cormier 1991; Kontak et al. 1995; Culshaw & Reynolds 1997); and (4) emplacement of the Wedgeport Pluton at 357 Ma (U-Pb zircon; MacLean et al. 2003).

The time of emplacement of the Brazil Lake pegmatite is constrained by the following field relationships. As noted above, deformation of the pegmatite is manifested by thin shear bands within muscovite-rich zones, flattened quartz megacrysts, a penetrative fabric within albitic pegmatite, and kinked spodumene crystals. In addition, the fabric in the wall rock records peak metamorphic conditions of ca. 4 kbars and 600°C (Moynihan 2003) and adjacent to the pegmatite the mineral assemblages in the amphibolite are not retrograded. Collectively, these observations suggest the pegmatite may have been emplaced before or during the time of regional deformation. Important additional observations are boulder trains of pegmatite that have been traced from the Brazil Lake area to the Brenton Pluton (Figs. 1b, c) where, in places, un-deformed pegmatite is observed to cross cut and retrograde the penetrative biotite fabric in the granite (Figs. 2a, b). Thus, locally some pegmatite post-dated the formation of the biotite fabric, which is presumably of Acadian age.

The most recent and relevant geochronological study in this part of the Meguma Terrane is that of Moynihan (2003) who used the spot laser⁴⁰⁴⁰Ar/³⁹Ar technique to date amphibole and mica. Included in this study were muscovite and amphibole from the BLP and wall rock. Maximum ages obtained were ca. 350 Ma, but younger ages to ca. 325 Ma were also obtained for muscovite. These data are discussed in more detail below in the context of this work.

Three samples of material were selected for dating in the present study: muscovite and tantalite from within the pegmatite and molybdenite from quartzite adjacent to the pegmatite (i.e., <1 m away). The three samples are from the southern pegmatite lens of the BLP (Fig. 1c), and are from within a few metres of each other.

The tantalite sample for U-Pb geochronology (Fig. 3a) consisted of coarse radiating tantalite crystals from albite-rich intergranular material within a spodumene-rich part of the pegmatite. Although not used routinely for dating, tantalite is ideally suited for dating of rare-element (e.g., LCT-type) pegmatite and has been successfully applied in, for example, Archean terranes of Canada (Smith et al. 2000) and Australia (Kinny 2000), the Proterozoic of northern Sweden (Romer & Wright 1992) and younger, Cretaceous pegmatite of the Northwest Territories (Mauthner et al. 1995). At Brazil Lake, the tantalite crystals are up to 1– 2 cm long and 2– 5 mm thick and are dark brown with a tabular habit (Fig. 3a). The occur-rence of tantalite associated with secondary albite, in part of cleavelandite habit, is typical of Ta-Nb mineralization within the pegmatite. A single large (<~1 cm), tantalite crystal was extracted from whole rock pegmatite using metal tweezers. Five fragments from this single tantalite crystal were analyzed at the University of Alberta Radiogenic Isotope Facility, Edmonton, Alberta. The five fragments weighing between 16 and 495 micrograms were dissolved in Teflon TFE digestion vessels housed in monel jackets using a 48%HF:7N HNO3 acid mixture at 220°C for 72 hours. A measured amount of ²⁰⁵Pb/²³⁵U tracer solution was added to the acid mixture in each vessel prior to digestion. Purification of Pb and U from the dissolved tantalite fragments was accomplished using anion exchange chromatography with an HBr/HCl procedure similar to that reported for Fe-bearing minerals such as titanite or rutile (e.g., Heaman et al. 2002). The purified uranium and lead aliquots were loaded together onto out-gassed Re filaments in a mixture of phosphoric acid and silica gel. Their isotopic compositions were determined on a VG354 thermal ionisation mass spectrometer using both Faraday and Daly photomultiplier (analogue) detectors in single detector peak-hopping mode. The Pb and U isotopic data were corrected for mass discrimination (0.12 and 0.14%/amu, respectively), blank (2 and 0.5 pg, respectively), and spike addition. The isotopic composition of common Pb present in the tantalite analyses above that estimated to be blank contribution was calculated using the two-stage model of Stacey & Kramers (1975). The uranium decay constants used in this study (²³⁸U = 1.55125 × 10^-10 yr^-1 and ²³⁵U = 9.8485 × 10^-10 yr^-1) and the isotopic composition of present day uranium of 137.88 are the values recommended by Jaffey et al. (1971).

The molybdenite-bearing sample is quartz-rich material adjacent the BLP (Fig. 3b). The clear, white color of the molybdenite-bearing, mineralized sample contrasts with the buff color of the White Rock Formation quartzite in the area, thereby suggesting that it may be a deformed and recrystallized quartz vein. However, field relationships are equivocal and, hence, preclude absolute certainty regarding its classification as either White Rock Formation quartzite or deformed quartz vein material. Similar multi-cm wide zones of white quartzite oriented at high angles to transposed bedding, which may also represent recrystallized quartz vein material, occur in other parts of the White Rock Formation. In the sample of interest, one large (1 × 1 cm) grain of molybdenite (BRZ-2002-1) was removed and the material homogenized by use of a diamond-tipped dental drill. A trial analyses established the Re abundance of the molybdenite concentrate, followed by two full Re-Os analyses (BRZ-2002-1a, 1b; Table 2). Duplicate analysis of this material was done because of the problematic nature of dating coarse molybdenite grains due to internal decoupling of Re and Os (Selby & Creaser 2004). This effect was monitored by the analysis of two sample aliquots with vastly different weights (94 and 256 mg), i.e., if sample homogenization has not been achieved the two different aliquots will return different ages. In addition, the remaining quartz material was pulverized and all remaining molybdenite extracted to comprise another analysis (BRZ-2002-2; Table 2). Re and Os concentrations in the molybdenite were determined using isotope dilution mass spectrometry at the University of Alberta Radiogenic Isotope Facility, Edmonton, Alberta using Carius tube, solvent extraction, and chromatographic techniques. Complete analytical procedures and error calculations are given in Selby et al. (2003).

A single sample of coarse mica of ca. 2 cm width, part of a book of muscovite 1 cm thick, was collected from an inter-granular mixture of quartz-feldspar-mica between spodumene - K-feldspar megacrysts of ca. 1 m size. This muscovite book was separated parallel to its cleavage into thin sheets and a single, cylindrical-shaped grain prepared for irradiation. Core versus rim analysis of the single muscovite sample by laser fusion was done at the Geochronology Laboratory, Queen's University, Kingston, Ontario, following the procedures outlined in Kontak et al. (1999). Dates and errors (Table 3) are calculated using formulae of Dalrymple et al. (1981), and the constants recommended by Steiger & Jäger (1977). Errors calculated represent analytical precision at 2𝛔, assuming that the errors in the ages of the flux monitors are zero. This technique is suitable for comparing within-spectrum variation and determining which steps form a plateau (McDougall & Harrison 1988, p. 89). A conservative estimate of this error in the J-value is 0.5 % and can be added for inter-sample comparison.

The isotopic results obtained for tantalite, molybdenite, and muscovite are given in Tables 1, 2 and 3, respectively, and we discuss the data starting with the oldest ages. The U-Pb results for five fragments from one tantalite crystal are presented in Table 1 and displayed on a concordia diagram in Fig. 4. All three fragments display moderate uranium contents (214– 567 ppm), low model thorium contents (1.4– 3.2 ppm) and correspondingly low Th/U ratios (0.004– 0.009). The Brazil Lake tantalite contains a modest amount of common lead (9– 22%), which is reflected in the relatively low ²⁰⁶Pb/²⁰⁴Pb ratios (265– 737) reported in Table 1, and this component contributes significantly to the overall uncertainty in the analyses. All five tantalite fractions yield nearly concordant U-Pb ages with ²⁰⁷Pb/²³⁵U and ²⁰⁶Pb/²³⁸U ages that vary between 366.1 and 394.6 Ma (Table 1).

The results for the three Re-Os molybdenite analyses are identical within error at 352– 354 Ma. Most importantly, the two ages for the coarse molybdenite grain (BRZ-2002-1a, 1b) are identical for the two weight fractions (i.e., 94 and 236 mg) which are identical to the finer grain fraction (BRZ-2002-2), thus indicating that decoupling of Re and Os is not a problem for this material. Incorporating the error in the decay constant of 0.31% (Smoliar et al. 1996) and propagating it with the analytical uncertainly gives errors of 2.5 Ma, we thus interpret the Re-Os molybdenite age for this sample to be 353.0 ± 2.5 Ma.

The muscovite grain was analyzed along the margin area (5 analyses) and the core (4 analyses). For the margin, ages range from 318 to 342 Ma, but three fall in the narrower range of 332.2 to 334.3 Ma. For the core area, ages range from 328 to 347 Ma. The integrated age for the sample is 333.7 ± 1.5 Ma.

The results using three different geochronometers to date the BLP have resulted in ages spanning some 77 Ma, ranging from 395 Ma for U-Pb tantalite (²⁰⁶Pb/²³⁸U age for fraction 3) to 318 Ma for laser ⁴⁰Ar/³⁹Ar spot analysis on a single grain of coarse muscovite. In order to assign an age to pegmatite formation and unravel its subsequent thermal history, it is important to assess the nature of the data for each of these chronometers in the context of the method and their reliability and then to interpret the data in the context of the geological history of the region.