3 The ca. 630–615 Ma volcanic and sedimentary rocks of the Caledonia terrane, known as the Broad River Group, are exposed mainly in the eastern highlands and in small faulted slivers along the Bay of Fundy coast (Fig. 1). The Broad River Group consists dominantly of intermediate and felsic crystal and lithic-crystal tuff, with less abundant intermediate and felsic flows, mafic tuff and flows, and epiclastic sedimentary rocks. Barr and White (1999a, b) divided the group into eleven lithological units, which were subsequently modified and given formation status (Barr and White 2004). Although internally varied, each formation has distinctive components which distinguish it from other formations in the group, but in general, stratigraphic relations among the formations are not known, except in rare, less deformed, areas where bedding and evidence of younging direction have been preserved. The Broad River Group has undergone pervasive regional metamorphism and contains mineral assemblages typical of the greenschist facies. Locally, the rocks have been affected also by contact metamorphism. In most places the volcanic or sedimentary protolith can be inferred, and the protolith names are, therefore, used in describing and naming the rocks. Based on rock types and geochemistry, Barr and White (1999a, b) inferred that the Broad River Group and related plutons formed in an Andean-type convergent margin subduction zone.

4 Previous direct controls on the age of the Broad River Group were from a rhyolite flow and felsic tuff in the southeastern part of the Caledonia terrane which yielded U– Pb ages of ca. 618 Ma and 613 ± 2 Ma, respectively (Bevier and Barr 1990; Barr et al. 1994). A third sample, felsic tuff from near the northern margin of the terrane (Fig. 1), yielded a younger age of 600 ±1 Ma. Because of this younger age, Barr and White (2004) re-assigned these rocks to the Coldbrook Group, on the basis that the age might reflect inherited zircon rather than the igneous crystallization age.

5 The Point Wolfe River pluton, largest in the Caledonia terrane, extends for nearly 50 km through the central and eastern highlands (Fig. 1). The northwestern margin of the Point Wolfe River pluton, where it is in contact with mainly tuffaceous rocks of the Broad River Group, is irregular and intrusive, whereas the southern margin is faulted, mainly against younger rhyolite and basalt of the Coldbrook Group. The plutonic rocks are typically protomylonitic near the faulted contact, which is the St. Martins-Stewart Mountain shear zone of Park et al. (2008, 2017). Barr and White (1999a, b) divided the Point Wolfe River pluton into six mappable lithologic units. An intrusive sequence from more mafic to more felsic rocks was inferred, although contact relations to confirm this assumption are limited. U–Pb dating from the two largest units of the Point Wolfe River pluton (Pollett River granodiorite and Old Shepody Road granite) gave ages of 625 ± 5 Ma and 616 ± 3 Ma, respectively (Bevier and Barr 1990). Although these ages do not overlap, even considering error limits, the close spatial association and petrological similarities of the units suggest that they are comagmatic (Barr and White 1999a). Granitoid plutons apparently similar to those of the Point Wolfe River suite occur throughout the Broad River Group southeast of the St. Martins-Stewart Mountain shear zone (Fig. 1). Units dated previously by U–Pb (zircon) are the Millican Lake pluton at Cape Spencer (623 ± 2 Ma; Watters 1993), granodiorite of the Kent Hills pluton in the northeastern highlands (615 +1/-2 Ma; Barr et al. 1994), and gabbro of the Caledonia Mountain pluton, also in the northeastern highlands (615 ± 1 Ma; Barr et al. 2000).

6 The Hammondvale Metamorphic Suite occurs in a narrow fault-bounded block along the northwestern margin of the Caledonia terrane (Fig. 1). It consists of albite- and garnet-porphyroblastic mica schist, as well as minor marble, calc-silicate rocks and quartzite. Peak pressure conditions during metamorphism were 12.4 kbar at 430°C, and peak temperature conditions were 580°C at 9.0 kbar (White et al. 2001). Muscovite ⁴⁰Ar/³⁹Ar muscovite ages from three samples range up to 618 ± 15 Ma, a minimum age for the high-P/low-T metamorphism in this unit. Based on detrital zircon ages, the maximum depositional age is about 650 Ma (Satkoski et al. 2010). The Hammondvale Metamorphic Suite has been interpreted to be a fragment of an accretionary complex formed on the margin of the Caledonia terrane during the subduction event that produced the Broad River Group and related plutons, suggesting that subduction was to the southeast in present-day coordinates (White et al. 2001).

7 The Coldbrook Group consists of two main groups of lithologies: (1) dacitic and rhyolitic flows and tuffs which extend from the city of Saint John through much of the northwestern part of the Caledonia terrane, and (2) basaltic and rhyolitic units which dominate in the southern part of the group (Fig. 1). The basaltic and rhyolitic units, together with associated epiclastic sedimentary rocks, are interpreted to be the youngest units in the group because they are overlain by Cambrian rocks in several areas (Fig. 1). Furthermore, the gabbroic and granitic intrusions with which they have been inferred to be comagmatic (Barr and White 1999a) intruded the dacitic flows and tuffs but not the basalt and rhyolite formations (Fig. 1).

8 The Coldbrook Group is generally less deformed and metamorphosed than the Broad River Group, and it is likely that the Broad River Group experienced a low-grade metamorphic and deformational event prior to deposition of the Coldbrook Group (Barr and White 1999a; Park et al. 2008). However, at least in part, this difference is related to geographic distribution, as the northwestern part of the Caledonia terrane is generally less deformed than the coastal and eastern parts, and some of that deformation may be as young as Silurian to Carboniferous, although it began in the Neoproterozoic (Park et al. 2008, 2017). Contacts of the Coldbrook Group with the older Broad River Group appear to be everywhere faulted, but an originally unconformable relationship is inferred. Barr and White (1999a) suggested that the Coldbrook Group and related plutons formed during extension and rifting within the former active margin represented by the Broad River Group.

9 Prior to the present study, the age of the Coldbrook Group was constrained by U–Pb (zircon) ages of 554 +14/-1 Ma from a dacitic tuff in the lower Coldbrook Group and 559 ± 5, 555 ± 3, and 548 ± 1 Ma from rhyolite at widely separated outcrops of the upper Coldbrook Group (Bevier and Barr 1990; Barr et al. 1994; Miller et al. 2000), as well as by ages from cross-cutting plutons as shown on Figure 1.

10 The Bonnell Brook and related plutons occur widely throughout the central and western parts of the Caledonia terrane, north and west of the St. Martins-Stewart Mountain and Arnold Lake shear zones (Fig. 1). In contrast to the older plutons, these plutons are bimodal, and consist mainly of syenogranite and monzogranite, with less abundant dioritic to gabbroic rocks. The largest component of the Bonnell Brook and peripheral plutons is relatively homogeneous medium-grained equigranular syenogranite (to monzogranite) composed of perthitic orthoclase, quartz, and plagioclase, and containing less than 3% biotite, amphibole, titanite, and allanite (Guy 1998). Also present locally, especially around the northern margins of the pluton, is fine-grained syenogranite to monzogranite that consists of mainly plagioclase microphenocrysts in a granophyric groundmass. Especially where granophyric or fine-grained, the syenogranite contains miarolitic cavities. Based on these textural features, the Bonnell Brook pluton is interpreted to be a high-level intrusion, with shallower parts exposed towards the northern part of the map area. A dome-like body of spherulitic rhyolite located north of the largest body of the Bonnell Brook pluton is interpreted to be the highest level (sub-volcanic) part of the pluton. It is texturally very similar to the granophyric unit of the main pluton, except it contains subhedral plagioclase microphenocrysts rimmed by spherulites instead of granophyre.

11 Areas of diorite to quartz diorite are present mainly on the southeastern margin of the Bonnell Brook pluton, inferred to be the deeper part, and are intruded by dykes of the syenogranite, clearly showing the relative ages of the two lithologies. The dioritic rocks consist mainly of plagioclase and hornblende with minor interstitial quartz, K-feldspar and relict clinopyroxene in the cores of some hornblende grains. These dioritic rocks are similar to dioritic parts of the mainly gabbroic (to ultramafic) Mechanic Settlement pluton (Grammatikopoulos et al. 1995).

12 Age constraints on these plutons (Fig. 1) include U–Pb (zircon) ages of 550 ± 1 Ma from the main Bonnell Brook syenogranite, 551 ± 5 Ma from the Upham Mountain syenogranite, and 557 ± 3 Ma from quartz diorite associated with the Mechanic Settlement pluton (Grammatikopoulos et al. 1995; Barr and White 1999a).

13 Volcanic and non-marine coarse clastic sedimentary rocks of the upper part of the Coldbrook Group are overlain by the Saint John Group, a lower Cambrian through lower Ordovician Paleozoic platformal sedimentary sequence which contain a typical “Avalonian” fauna (e.g., Tanoli and Pickerill 1988; Landing 1996; Landing and Westrop 1998; Palacios et al. 2011). Following the terminology of Tanoli and Pickerill (1988), the lowermost formation, Ratcliffe Brook, consists of siliciclastic sedimentary units including coarse conglomerate, feldspathic quartz sandstone, arkosic sandstone, siltstone and minor mudstone. In contrast to the underlying similar redbeds of the Coldbrook Group, all units of the Saint John Group contain notable quantities of detrital white mica, and the coarser lithic clasts are recognizable rock types derived from the underlying Coldbrook Group and related intrusions as well as mylonite. Tanoli et al. (1985) first suggested that Neoproterozoic redbed units contain little or no detrital muscovite whereas overlying Cambrian rocks contain abundant detrital muscovite, a difference supported by subsequent studies (Reynolds et al. 2009; Satkoski et al. 2010). However, Barr and White (1999a) were not aware of this difference during their mapping and did not distinguish between muscovite-present and muscovite-absent redbeds in their separation of the Ediacaran and Cambrian redbed units, so some areas of muscovite-bearing redbeds were incorrectly assigned to the Seely Beach Formation of the Coldbrook Group on their maps and vice versa (Barr and White 1999b, 2004). Hence, for simplicity at page scale, all Seely Beach Formation of Barr and White (2004) is included with the Saint John Group on Figure 2.

14 The oldest direct radiometric age constraint on the Saint John Group is an age of 530.7 ± 0.9 Ma (U–Pb zircon; Isachsen et al. 1994) later revised to 528.1 ± 0.9 Ma (Compston et al. 2008) for an ash layer from the Ratcliff Brook Formation well above its base (Palacios et al. 2011).

15 Barr and White (2004) assigned the name Grassy Lake Formation to an area of flows and tuffs in the west-central Caledonia terrane (Fig. 1). The rocks are mainly dark-grey, light grey, or pinkish-grey dacitic to rhyolitic flows and tuffs. Flow-banding is present in flows and fine-grained crystal (welded) tuffs. The crystals include feldspar and embayed quartz, in a cryptocrystalline to spherulitic felsic groundmass. Locally, lithic lapilli tuff is abundant. All these rocks are similar to those in the surrounding units of the Coldbrook Group and the Grassy Lake Formation was recognized as a separate formation only because U–Pb (zircon) dating indicated a maximum age of 479 ± 8 Ma (Barr et al. 1994). The extent of the formation and contact relationships with surrounding rocks of the Coldbrook Group remain uncertain because of limited outcrop. However, they were inferred by Barr and White (1999a, b) to be unconformable because of the age contrast.

16 Like the Grassy Lake Formation, the Fairfield Formation (Barr and White 2004) was separated from the older units by Barr and White (1999a, b; their unit D_F) on the basis of radiometric dating which suggested a poorly constrained maximum age of 367 Ma (Barr et al. 1994). The formation consists of pink to red rhyolitic flows and red felsic lithic tuff with tuffaceous siltstone and sandstone. In the dated area these rocks are surrounded by tuffaceous rocks of the Broad River Group, and in postulated faulted contact with rhyolite of the Coldbrook Group (Fig. 1). Outcrops of similar rocks west and north of St. Martins were assigned also to the Fairfield Formation based on lithologic similarity.

17 Basaltic rocks that occur in small fault-bounded areas along the coast of the Bay of Fundy were assumed to be Carboniferous by Barr and White (1999a, b; their unit Cb) based on lithological similarities to the ‘Mispec Formation’ or ‘West Beach Formation’ in the Saint John area (Alcock 1938; Rast et al. 1978; Strong et al. 1979). McLeod (1987) had identified them as Carboniferous or older. Currie (1992) included similar rocks in the Saint John area in the ‘Lorneville beds’ and inferred an “Eo-Cambrian” age. However, Park et al. (2014) demonstrated that these rocks in the Saint John area (Taylors Island Formation) are Late Devonian to Early Carboniferous (Famennian to Tournaisian) and intruded by granitoid plutons of Early Carboniferous age.

22 A quartz-feldspar porphyry dyke is exposed in a roadside outcrop of mafic volcanic rocks along the new Fundy Trail Parkway east of St. Martins. These deformed and metamorphosed mafic rocks form the headland between Big Salmon River and Long Beach and are separated from Cambrian and Coldbrook Group rocks by a south-dipping reverse fault (Park et al. 2017). The mafic rocks previously had no direct age constraints but were assigned to the West Beach Formation (Barr and White 2004), one of several such enclaves of fault-bounded mafic rocks along the coast of the Bay of Fundy between Saint John and Alma. Based on new mapping (Park et al. 2017) and the age reported here, these rocks have been assigned to a new unit, the Long Beach Formation (Fig. 2).

23 Dated porphyry dyke sample NB14-335 consists of plagioclase and quartz phenocrysts in a fine-grained groundmass of feldspar, quartz, and sericite. The quartz phenocrysts are embayed and contain prominent deformation lamellae. Plagioclase phenocrysts are broken and deformed. The sample yielded abundant euhedral zircon grains with igneous zoning (e.g., Fig. 3a). Four zircon fractions were analysed via CA-TIMS, two of which are concordant with ²⁰⁶Pb/²³⁸U ages of 700 ± 15 Ma (Z2, 1 small low-U prism) and 694 ± 5 Ma (Z3). Together these data yield a weighted average ²⁰⁶Pb/²³⁸U age of 695 ± 5 Ma (MSWD = 0.69). One analysis (Z4) of a clearly inherited zircon grain is at 828 ± 5 Ma and not shown on the diagram, and the fourth analysis (Z1) is discordant (Fig. 3b). Because of the importance of these unexpectedly old ages, zircon grains from this sample were also analysed by LA-ICP-MS to test for the potential presence of younger grains missed in selecting grains for TIMS work. The zircon grains are of high quality and euhedral to subhedral (Fig. 3a). Obvious bright-CL overgrowths on darker, but euhedral cores, are of the same age generation of zircon, and are likely due to influx of new magma of different chemistry into the magma chamber, growing new layers on the same crystals, a common feature in felsic igneous rocks. One exception in the CL images is in the grain (upper left in Fig. 3a) with an anomalously older 797 Ma rim (analysis 116), which seems to overgrow a corroded core. One grain (lower right, Fig. 3a) gave a young age of 626 ± 26 Ma (analysis 130) but another part of the same growth zone gave an age of 649 ± 27 Ma (analysis 131). A weighted average ²⁰⁶Pb/²³⁸U age using 11 analyses of clear igneous growth-zoned areas gives 679 ± 9 Ma with an excellent MSWD of 0.46 (Fig. 3c). However, only 6 grains overlap tightly enough to calculate an acceptable concordia age which is 685 ± 10 Ma with MSWD = 0.97 (Fig. 3d).

24 The igneous crystallization of the porphyry is best reported as 685 ± 10 Ma, as this age covers the range of the 3 different ages calculated from the different combinations of data described above. This age makes this porphyry the oldest rock yet dated in the Caledonia terrane, and provides a minimum age for the even older mafic volcanic rocks of the Long Beach Formation that it intruded.

25 Sample NB14-334 was collected from the same outcrop as the sample dated by Barr et al. (1994) which yielded a poorly constrained Devonian age and was assigned to the Fairfield Formation (Fig. 1). Like the original sample, it consists of micro- to cryptocrystalline flow-banded rhyolite with microspherulites in a sericitic quartz and feldspar groundmass. It contains scattered microphenocrysts of plagioclase. The sample yielded abundant clear euhedral prisms which display simple igneous growth zoning, with some very luminescent cores or rims (Fig. 4a).

26 Six zircon fractions were analysed via CA-TIMS, most consisting of 1 or 2 grains. These six analyses yielded 2 discrete ages: two grains yielded ²⁰⁶Pb/²³⁸U ages of 632 ± 4 and 634 ± 6 Ma and are likely inherited. Four overlapping younger concordant analyses yield a weighted average ²⁰⁶Pb/²³⁸U age of 622 ± 1.9 Ma with MSWD = 1.18 (Fig. 5a), interpreted to be the igneous crystallization age of the rhyolite.

27 This age is consistent with U–Pb ages from the Broad River Group and related plutons (Figs. 1, 2). Given the uncertainty of the previously reported age, the name Fairfield Formation is abandoned here, and the rocks formerly included in that unit are considered part of the surrounding East Branch Black River Formation of the Broad River Group (Fig. 2). The new age is consistent with the observation that the East Branch Black River Formation is intruded by the Millican Lake plutons, one of which yielded an age of 623 ± 2 Ma (Watters 1993) and supports the interpretation that the plutons are comagmatic with associated volcanic rocks of the Broad River Group (Barr and White 1999a). However, this sample also demonstrates the difficulty of correlating rock units in the Caledonia terrane based on rock type or chemistry – the dated flow-banded rhyolite sample is to all appearances identical to flow-banded rhyolite in the Coldbrook Group which yields ages of ca. 555 Ma, and the samples are chemically similar as well (Barr and White 1999a).

28 The Gordon Falls Formation of Barr and White (2004) occurs in a small fault-bounded sliver on Pollett River near the northern margin of the Caledonia terrane, sandwiched between the ca. 625 Ma Pollett River granodiorite on the south and Carboniferous sedimentary rocks to the north (Figs. 1, 6). It consists mainly of dacitic to rhyolitic crystal tuff, with minor lithic-crystal tuff and siliceous felsite with abundant pyrite (Barr and White 1999a). Bevier and Barr (1990) reported a U–Pb date of 600 ± 1 Ma for a dacitic tuff sample from Gordon Falls, which seems incompatible with the age of 625 Ma obtained from the adjacent Pollett River granodiorite, although the contact is faulted. This line-ofreasoning led Barr and White (2004) to re-assign the unit to the Coldbrook Group on the basis that the age might reflect zircon inheritance.

29 Sample SH15-137 was collected from the same outcrop as the previously dated sample to try to resolve the uncertainty. Like the sample dated by Barr et al. (1994), it is a deformed felsic crystal tuff that consists of relict plagioclase augen in a strongly foliated groundmass of polycrystalline quartz, feldspar lenses, epidote, and sericite. The sample contained abundant small euhedral to subhedral zircon prisms (Fig. 4b). Two analyses of 8 and 2 prisms yielded overlapping concordant points with ages of 620.5 ±5.5, and 617.5 ± 2.5 Ma, indicating an age of 620 ± 5 Ma based on the higher quality analysis (Fig. 5b). This age is interpreted to be the eruption age of the tuff, superseding the previously reported younger age by Barr et al. (1994). The younger age may be the result of minor lead-loss from a multiple-grain fraction that dragged that point down along concordia from 620 Ma. The new ca. 620 Ma is more consistent with the geological setting of the Gordon Falls Formation.

30 The McBrien Lake Formation is interpreted to be the oldest unit in the Coldbrook Group (Barr and White 2004). It consists mainly of grey-green dacitic flows best exposed in the western part of the Caledonia terrane, where dated sample NB16-01 was collected (Fig. 2). The unit extends into the city of Saint John where a flow-banded, welded, dacitic tuff sample previously yielded a U–Pb (zircon) age of 554 +14/-1 Ma (Barr et al. 1994; location shown on Figure 1, west of the area of Figure 2). Similar flows occur in places in the overlying Ben Lomond Formation, a mainly tuffaceous unit in which many of the clasts resemble the flows of the McBrien Lake Formation (Barr and White 1999a).

31 Dated sample NB16-01 is from a dacite flow, and consists of skeletal, needle-like plagioclase in a spherulitic groundmass with scattered flakes of altered biotite. Secondary minerals include abundant sericite, epidote, chlorite, and carbonate minerals. The sample yielded abundant small euhedral zircon prisms (Fig. 4c) and 5 analyses were carried out (Table A1). Analysis Z1 plots below the rest, interpreted to be due to lead loss. The other 4 (Z2–Z5) are all concordant and overlap, and these yield a weighted average ²⁰⁶Pb/²³⁸U age of 557.4 ± 2.5 Ma (MSWD = 1.04, Fig. 5c), interpreted to be the igneous crystallization age of the flow.

32 The mainly tuffaceous Fletcher Brook Formation is interpreted to underlie the Silver Hill Formation and overlie the Ben Lomond Formation in the north-central part of the Coldbrook Group (Fig. 2). The formation contains a wide variety of rock types, including amygdaloidal basalt, andesitic to basaltic lapilli tuff, pyroclastic breccia and debris flows, laminated felsic ash tuff, vari-coloured dacitic lithic lapilli tuff, andesitic to dacitic flows, and locally spherulitic felsic tuff with prominent eutaxitic banding. Dated felsic tuff sample NB15-003 from the latter rock type contains embayed quartz, feldspar, and felsic rock fragments in a cryptocrystalline groundmass. It yielded small zircon grains that look like simple igneous prisms of one morphology and age. Some display very luminescent outer growth zones in CL (Fig. 4d). Four fractions were analyzed, each consisting of 1, 2, or 3 grains and these tightly overlap on concordia and provide a weighted average ²⁰⁶Pb/²³⁸U age of 555.5 ± 1.9 Ma (MSWD = 0.081, Fig. 5d). This age is interpreted to be the igneous crystallization age of the tuff.

33 The Dolan Brook Formation is located at the southwestern margin of the Bonnell Brook pluton, and in a smaller area to the west. Like the Fletcher Brook Formation, it is a lithologically varied unit that includes interlayered basalt and rhyolite as well as a variety of mafic to felsic lithic lapilli tuffs and laminated ash tuff. Some of the dacitic tuffs are similar to those in stratigraphically lower formations of the Coldbrook Group such as Ben Lomond Formation, and the amygdaloidal basalt and rhyolite components resemble those in the Burley Lake, Fletcher Brook, and Silver Hill formations. Hence the stratigraphic significance of this heterogeneous unit is uncertain.

34 Dated sample NB16-02 is a felsic lithic crystal tuff with scattered crystals of embayed quartz and feldspar in a cryptocrystalline quartz and feldspar (ash) groundmass with abundant sericite. Lithic clasts are mainly rhyolitic, some with flow banding. The sample yielded abundant small euhedral, high-quality igneous prisms (Fig 4e) and 4 analyses were carried out of small fractions of 3 or 4 crystals each. These yielded a simple data set of four overlapping concordant points with a range of ²⁰⁶Pb/²³⁸U ages from 551.8 to 557.0 (Fig. 5e). One analysis, Z3, is of distinctly poorer quality and shifted slightly to the right; it was not used in the age calculation. Analyses Z1, Z2, and Z4 yield a weighted average ²⁰⁶Pb/²³⁸U age of 552.8 ± 1.6 Ma (MSWD = 1.07), interpreted as the crystallization age of the tuff.

35 Sample NB16-03 is fine-grained granite from the chilled margin of the Bonnell Brook syenogranite, very close to its contact with the Ben Lomond Formation of the Coldbrook Group (Fig. 2). It consists of large subhedral crystals of quartz, K-feldspar, and rare altered biotite in a groundmass that varies from granophyric to spherulitic. Secondary minerals include sericite, chlorite, and epidote.

36 This sample yielded abundant simple prisms with excellent igneous growth zoning (Fig. 4f) similar in size and morphology to those in the sample from the Dolan Brook Formation (Fig. 4e). The zircon grains are smaller than those typical of granite, consistent with the fact that this part of the Bonnell Brook pluton is high-level. Four analyses, of 3 to 6 crystals per fraction, yielded tightly overlapping concordant points, with ²⁰⁶Pb/²³⁸U ages between only 551.9 and 553.1 Ma (Fig. 5f). The weighted average ²⁰⁶Pb/²³⁸U age using all four analyses is 552.5 +/-1.5 Ma (MSWD = 0.094). This age corroborates the age of 550 ± 1 Ma obtained previously from the interior part of the pluton (Barr et al. 1994) and supports the interpretation of Barr and White (1999a) that the Bonnell Brook pluton intruded comagmatic rocks of its own volcanic carapace.

37 The Burley Lake Formation north of St. Martins (Fig. 2) consists of basalt, rhyolite, and interlayered clastic sedimentary rocks, overlain by sedimentary rocks of the Saint John Group. Sample NB16-04 is cryptocrystalline to microcrystalline flow-banded rhyolite from that formation. Texture varies from eutaxitic to spherulitic, with lithophysae filled with larger quartz crystals.

38 The sample yielded small clear euhedral zircon prisms and a small number of larger elongate zircon grains with hematitized cores (Fig. 4g). Only the small clear prisms were analysed and three analyses yield a weighted average ²⁰⁶Pb/²³⁸U age of 553.3 ± 3 Ma (MSWD = 0.053, Fig. 5g), interpreted to be the igneous crystallization age of the rhyolite.

39 New roadcuts along the Fundy Trail Parkway east of Big Salmon River display complex structural relationships between deformed rhyolite of the Silver Hill Formation (Coldbrook Group) and quartz arenite, sandstone, siltstone, and shale of the Ratcliffe Brook Formation (Saint John Group) (Park et al. 2017). Sample NB15-344 consists of quartz-rich phyllite composed mainly of quartz, chlorite, and sericite. Its close association with deformed and recrystallized rhyolite suggests that, like the rhyolite, it is part of the Silver Hill Formation (Park et al. 2017).

40 The sample yielded some rounded zircon grains but a high proportion of clear euhedral prisms. Most grains are sharp and euhedral and have continuous growth zoning (Fig. 4h). Four zircon fractions were dated, three of which (Z1, Z3, Z4) are excellent pristine analyses of 1 or 2 grains each. The fourth grain (Z2) is significantly down a line towards 347 Ma, with large uncertainty. Its position on concordia is interpreted to be due to lead-loss not removed by chemical abrasion. The remaining 3 analyses from euhedral grains yield a weighted average ²⁰⁶Pb/²³⁸U age of 556 ± 2 Ma (MSWD = 0.22) (Fig. 5h).

41 The age of 556 ± 2 Ma is interpreted to be the maximum age of deposition of the phyllite protolith, and is similar to ages of volcanic rocks from the older parts of the Coldbrook Group (Fig. 9). These rocks are likely the main source of detritus in the phyllite. A quartz arenite sample from the overlying Ratcliffe Brook Formation but from a location farther east in the Caledonia terrane yielded a similar weighted mean ²⁰⁶Pb/²³⁸U date of 553.3 ± 5.3 Ma (MSWD = 0.33) (Satkoski et al. 2010).

42 The Grassy Lake Formation was identified as separate from the Coldbrook Group based on a poorly constrained age of less than 479 ± 8 Ma (Barr et al. 1994). Otherwise, field evidence did not provide any reason to separate this daciticrhyolitic tuffaceous unit from the surrounding Ben Lomond Formation of the Coldbrook Group. At that time the age of the Ben Lomond Formation was already constrained by the ca. 550 Ma U–Pb age (Bevier and Barr 1990) of the Bonnell Brook pluton which intruded it. Attempts to better constrain the age of the Grassy Lake Formation by additional sampling at that time were thwarted by absence or scarcity of zircon grains, and additional field work over the years has not provided any additional constraints on relations with the surrounding Coldbrook Group rocks.

43 To further pursue the unresolved problem of the age of the Grassy Lake Formation, sample NB08-268 was collected from an outcrop 200 m north of the sample dated previously by Barr et al. (1994) but similar in rock type to the previously dated sample. It consists of grey microcrystalline to cryptocrystalline rhyolite with flow-banding and spherulitic layering. Scattered phenocrysts are mainly of plagioclase. Secondary minerals include sericite, chlorite, and epidote.

44 Eight zircon fractions were analysed by TIMS from sample NB08-268 (Fig. 7a). The two youngest fractions (Z2, Z6) are overlapping and concordant at 442 ± 3 and 446 ± 3.5 Ma. A line can be calculated (60% probability of fit) with an upper intercept of ca. 560 and lower intercept at 404 Ma, but the lower intercept is too imprecise (± 65 Ma) to be useful. It could be that the ca. 442–446 Ma age of duplicated concordant analyses represents the age of the rock, and that is why there is no lower point. In any case, the Grassy Lake Formation cannot be older than the igneous growth-zoned rims that are a significant part of the grains in the CL image (Fig. 7b). If the grains that have yielded ²⁰⁶Pb/²³⁸U ages of 446 ± 3.5 and 442 ± 3 Ma are devoid of older cores, then this is the true age of eruption of this rock. However, if they contain small older cores which we cannot know from our data, then the true age of the rims (and the rock) is younger than 443 Ma.

45 To attempt to better constrain the age of this rock, zircon grains from this sample were also analysed by LA-ICP-MS (Fig. 7b) and yielded multiple concordant ages (Fig. 7c). One subhedral grain is concordant at 1779 ± 68 Ma, whereas two analyses on another grain gave ²⁰⁶Pb/²³⁸U ages of 1262 ± 51 and 1270 ± 54 Ma (Fig. 7c). Three igneous growth-zoned grains gave spot ages of 602 ± 30 Ma, 609 ± 25, 617.5 ± 25, and 614.7 ± 25 Ma, and combined with the TIMS analysis at 609 ± 16 Ma, this group of grains gives a ²⁰⁶Pb/²³⁸U age of 611 ± 10 Ma (Fig. 7d) with MSWD = 0.20. Five laser spot analyses (#73, 74, 79, 80, 95) on 3 entire growth-zoned igneous crystals gave ages between 573 – 588 Ma and yielded a weighted average age of 578 ± 11 Ma (Fig. 7d) with MSWD = 0.33. Because the two Ediacaran age clusters are from well-preserved igneous grains that have been chemically abraded it is possible to say that the Ediacaran clusters of data do not significantly overlap, and the data set shows no evidence of zircon grains with ages of ca. 555 Ma, the age of the Coldbrook Group.

46 The five youngest analyses, from 4 crystals, are 406 ± 18 to 376 ± 16 Ma. These ages all overlap, and an age of 389 ±14 Ma can be calculated from all 5 analyses. However, the youngest 3 analyses (#78, 87, 90), from 3 excellent whole zircon grains, are tightly clustered, have smaller than average errors and a calculated concordia age for these is 382.8 ± 8.3 Ma (Fig. 7e). This is interpreted to be the best estimate of the age of the youngest igneous zircon in the rock, and hence is a maximum age for the Grassy Lake Formation. The combined U–Pb zircon TIMS and laser data set are consistent, with TIMS ages matching or falling in between the clusters of ages identified by laser.

47 The Bloomsbury Mountain Formation is a dome of brown to pink felsite that was assigned to the Coldbrook Group by Barr and White (1999a, b, 2004). It was sampled for dating because of the possibility that this dome-like undeformed rock might be younger than other units in the Coldbrook Group, and also because gold anomalies had been recognized by prospectors in the area. Viewed in thin section, sample NB15-348 is homogeneous and consists mainly of spherulitic material, with rare feldspar crystals, and is interpreted to be a shallow intrusive rock. It yielded not-perfectly-euhedral zircon grains and every grain imaged has a corroded core, with a significant part of the grain composed of younger overgrowths with sharp igneous growth zones in the overgrowth (Fig. 8a). The less luminescent cores also have igneous growth zones. The linear data array for the 4 analyses from ca. 561 to 443 Ma intersects concordia at 417 Ma (with large uncertainties of ca. 100 myrs), indicating a likely Silurian–Devonian age and that the magma may have formed as a melt of a single-age Proterozoic igneous rock. The different distances that the analyses plot down the discordia line (Fig. 8b) likely correlate with the proportion of those growth-zoned igneous overgrowths. The age of sample NB15-348 is not really pinned down by the TIMS analyses; it is younger than (or equal to) 443 ± 5 Ma (the youngest ²⁰⁶Pb/²³⁸U age). The most conservative interpretation is that 443 ± 5 Ma is the oldest possible crystallization/eruption age, but a younger age, like the projected 417 Ma lower intercept, cannot be ruled out.

48 To further investigate the age of the felsite, zircon grains were analysed by LA-ICP-MS. Relatively few analyses are concordant, and most are ca. 555 Ma (Fig. 8c), as are 2 of the TIMS analyses (in grey). The 19 best analyses give a weighted mean age of 554.7 ± 2.6 Ma (Fig. 8d, e). These include both core and rim analyses (see CL image) so many of these overgrowths are likely due to magma chamber complexity in refilling, corrosion, and new growth in cycles.

49 While the 555 Ma age cluster is dominant, one grain has 2 spot ages of 574 ± 26 and 574.9 ± 23 Ma (analyses 14, 15), which overlap with the older age cluster in sample NB08-268, so perhaps these 2 rocks have sampled most of the same ages of rocks, in different proportions, although we are wary of subdividing too much with analyses of such low precision compared to TIMS (all uncertainties are propagated at 2 standard error of the mean). Two analyses are 655 ± 35 Ma (rim) and 641 ± 27 Ma (core). That core has an igneous rim at 610 ± 25 Ma, an interesting age match with 4 zircon analyses from NB08-268 and not surprising given that these rocks unconformably overlie or intruded the ca. 625–615 Ma Broad River Group and associated plutons.

50 The five youngest analyses, from 4 grains, range from 436 to 398 Ma (Fig. 8f). They all overlap (barely) and each looks like a simple one-age zircon crystal. Three of these grains give 4 spot analyses (#19, 33, 34, 54) from which an acceptable age of 427 ± 9 Ma (MSWD = 0.39) can be calculated. This is likely the true age (within uncertainties) of the youngest group of igneous zircon grains in the rock and hence provides a maximum age of emplacement for the felsite. The single 398 ± 21 Ma analysis (#62) is 24 million years younger than the next youngest. Again, the combined TIMS and laser ICP-MS data set makes sense and they are consistent, with TIMS ages matching or falling in between the clusters identified by laser, and the calculated TIMS lower intercept age of ca. 417 (albeit ± 100 Ma) in agreement with the youngest age group identified by laser analysis.