The Caledonia terrane (Fig. 2a) includes mainly Neoproterozoic and Cambrian rocks located south of the Caledonia-Clover Hill fault, which has been interpreted to mark a cryptic suture between the Caledonia and Brookville terranes (White et al. 2001). Based on differences in lithology and age, Neoproterozoic volcanic and sedimentary rocks of the Caledonia terrane have been divided into the ca. 620 Ma Broad River Group, exposed mainly in the northeastern and southern parts of the terrane, and the ca. 560– 550 Ma Coldbrook Group which forms most of the western and northern parts (Bevier and Barr 1990; Barr et al. 1994; Barr and White 1999). Ca. 620 Ma compositionally expanded gabbroic to granitic plutons and ca. 560– 550 Ma bimodal gabbroic and granitic plutons are interpreted to be cogenetic with the Broad River Group and basalt and rhyolite of the upper part of the Coldbrook Group, respectively (Barr and White 1999). On the basis of petrological characteristics, the ca. 620 Ma rocks are interpreted to have formed in a continental margin magmatic arc, whereas the ca. 560– 550 Ma rocks formed during subsequent extension (Barr and White 1999). The Coldbrook Group is overlain by a Lower Paleozoic platformal sedimentary sequence containing an "Avalonian" fauna (e.g., Tanoli and Pickerill 1988; Landing 1996; Landing and Westrop 1998).

A fault-bounded area of high-pressure and low-temperature metamorphic rocks located near the northwestern margin of the Caledonia terrane has been interpreted to represent an accretionary complex formed at ca. 620 Ma in association with a southeast-dipping subduction zone that generated the Broad River Group and related plutons (White et al. 2001). This unit is evidence for a terrane boundary between the Caledonia and now-adjacent Brookville terrane.

In contrast to the Caledonia terrane, the Brookville terrane is characterized by mainly metasedimentary rocks (Green Head Group) and abundant plutons (White and Barr 1996; White et al. 2002). The terrane-bounding brittle faults, Caledonia - Clover Hill on the southeast and Kennebecasis on the northwest (Fig. 1), are Carboniferous or younger faults interpreted to mark the approximate locations of older cryptic sutures. The Green Head Group consists of the Ashburn (dominantly marble with minor metaclastic rocks) and Martinon (dominantly metasiltstone with minor calc-silicate rocks, quartzite, conglomerate, and marble) formations, interpreted to be lateral facies equivalents. The early Neoproterozoic or late Mesoproterozoic age proposed for the Green Head Group based on locally preserved stromatolite occurrences (Hofmann 1974) is supported by a minimum U-Pb age of 1230 Ma for detrital zircon grains in a quartzite sample (Barr et al. 2003). The Green Head Group is in tectonic contact along the MacKay Highway shear zone (Nance and Dallmeyer 1994; White 1996) with the Brookville Gneiss, a locally migmatitic paragneiss with sheets of granodioritic to tonalitic orthogneiss, minor calc-silicate and marble layers, and rare quartzite and amphibolite. The paragneiss comprises about 75% of the Brookville Gneiss, and contains detrital zircon indicating a maximum depositional age of ca. 640 Ma (Bevier et al. 1990). The orthogneiss has an igneous crystallization age of 605 ± 3 Ma, and was metamorphosed to amphibolite facies at 564 ± 6 Ma (Bevier et al. 1990; Dallmeyer et al. 1990). These ages indicate that the Brookville Gneiss is younger than the Green Head Group and, hence, does not represent its basement, although the original relationship between the two units remains problematic (White and Barr 1996).

The Green Head Group and Brookville Gneiss have been intruded by plutons of the redefined Golden Grove Plutonic Suite (White et al. 2002). The plutons range in composition from gabbroic to granitic, and have yielded U-Pb and ⁴⁰Ar/³⁹Ar dates of ca. 550 Ma to 527 Ma (Dallmeyer and Nance 1992; White 1996; Currie and Hunt 1991; Currie and McNicoll 1999; White et al. 2002). Petrological characteristics suggest that they constitute a calc-alkalic I-type granitoid suite, probably emplaced in a continental margin subduction zone (White 1996; Eby and Currie 1996), although the older components of the suite (ca. 550 Ma) have characteristics indicative of an early extensional environment that evolved to subduction (White et al. 2002). Eby and Currie (1996) and Currie and McNicoll (1999) suggested that the plutons of the Brookville terrane are related to those in the Caledonia and New River terranes, but Barr and White (1996) argued that they are mainly younger and part of a separate and unrelated tectonic regime. Volcanic rocks of the Dipper Harbour Formation, locally preserved in faulted blocks in the southwestern part of the Brookville terrane (Fig. 2), are chemically similar to the early felsic granitoid components of the terrane (White et al. 2002). However, other workers have suggested that the Dipper Harbour Formation is related to volcanic rocks of broadly similar age in the Coldbrook Group of the Caledonia terrane, although the ages reported for samples from the Dipper Harbour Formation and associated plutons had large errors associated with them (Zain Eldeen 1991; Currie and Hunt 1991).

The most northwesterly exposures of Precambrian rocks known in southern New Brunswick occur in the New River terrane. Rock units in the New River terrane have been described by Johnson and McLeod (1996) and Johnson (2001; 2003), and the interpretations of these authors are generally followed here (Fig. 2).

The New River terrane is cut by numerous faults, and hence relationships among units are uncertain in many cases. The southern part of the belt is dominated by the Blacks Harbour Granite and Ragged Falls suite (Fig. 2). The Blacks Harbour Granite occurs with minor felsic volcanic rocks in a fault-bounded strip on the southern margin of the terrane. Prior to the new age reported here, it was undated. The Ragged Falls suite, as defined by Johnson (2001, 2003), includes syenogranite, monzogranite, tonalite, and gabbro. Ages of 555 ± 2 Ma and 555 ± 10 Ma have been reported for granodiorite and granite, respectively, from the suite (Johnson and McLeod 1996; Currie and Hunt 1991). The Ragged Falls Suite is closely associated with volcanic rocks (Leavitts Head Formation of Johnson 2001, 2003), a felsic component of which has yielded a U-Pb age of ca. 550 Ma (Johnson and McLeod 1996). Also in the area are numerous faulted blocks of volcanic and sedimentary rocks that were assigned to the Ordovician Simpsons Island volcanics and Goss Point Formation by Johnson and McLeod (1996). Based on the U-Pb (zircon) age reported here, Johnson (2003) separated the Simpsons Island volcanics (Simpsons Island Formation) from the fossiliferous Late Ordovician rocks of the Goss Point Formation (Nowlan et al. 1997).

A fault-bounded belt of Early Cambrian and Silurian volcanic and sedimentary rocks separates the Ragged Falls suite and Leavitts Head Formation from the Goose Lake Pluton (undated granodiorite and granite) to the north (Fig. 2a). To the northeast, Johnson (2001, 2003) interpreted the Ragged Falls suite to pass gradationally into similar granitoid rocks of the undated Rocky Lakes Suite, which she described as granite, granodiorite , tonalite, and quartz diorite, cut by numerous mafic dykes. Farther to the northeast, Johnson (2001, 2003) terminated the Rocky Lakes suite at the major northeast-trending Robin Hood Lake Fault (Fig. 2a). Northeast of this fault, ca. 625 Ma granitoid rocks are in complexly faulted relationships with ca. 555 Ma felsic tuff and porphyry, Neoproterozoic to Cambrian mafic and felsic volcanic and sedimentary rocks, and Cambrian rocks assigned to the Saint John Group (Tanoli and Pickerill 1988; Johnson and McLeod 1996; Johnson 2001, 2003). The ca. 625 Ma granitoid rocks were termed the Lingley and Brittain Creek plutons by Currie and McNicoll (1999). However, chemical data from granitic samples reported by Eby and Currie (1996) to be from the Lingley Pluton are at least in part from younger quartz-eye granitic porphyry that Johnson (2001) included with the Lobster Brook Formation. A sample from the Lobster Brook Formation which yielded a U-Pb age of ca. 550 Ma (McLeod et al. in press).

As a result of her work, Johnson (2001, 2003) suggested that the New River terrane is composite. She proposed that units northeast of the Robin Hood Lake Fault, as well as the Blacks Harbour Granite and spatially associated units in the Beaver Harbour area (Fig. 2b), are part of the Avalon zone, and hence comparable to rocks in the Caledonia terrane. In contrast, units southwest of the fault and north of the Blacks Harbour Granite may be related to the Brookville terrane and/or the St. Croix terrane and related units in Maine, which have been linked to the Gander terrane (van Staal et al. 1998; Fyffe et al. 1999).

Samples from the Simpsons Island Formation and Blacks Harbour Granite in the New River terrane and the Dipper Harbour Formation in the Brookville terrane (Fig. 2) were dated by the U-Pb (zircon) method as part of the present study. The analyses were done at the University of North Carolina (Chapel Hill) Isotope Geochemistry Laboratory. Zircon grains were separated from samples weighing ca. 25 kg using standard techniques, and hand picked under a binocular microscope in order to select groups of grains that were similar in morphology and as clear, crack-free, and inclusion-free as possible. All zircon fractions were highly abraded to remove the outermost portions of grains that are most likely to have been affected by Pb-loss. Detailed analytical procedures were similar to those described by Miller et al. (2001).

Rhyolite sample NB00-129 was collected from the Simpsons Island Formation on Adam Island in the southwestern part of the New River terrane (Fig. 2). Johnson and McLeod (1996) interpreted this unit to be Ordovician, because of close spatial association with faulted slivers of sedimentary rocks that contain Late Ordovician fossils. The dated sample is a fine-grained flow-banded rhyolitic flow with well-developed eutaxitic and locally spherulitic texture.

Zircon grains separated from sample NB00-129 were mostly prismatic, and large enough (~150– 200 µm x 30– 40 µm) for analysis of single grains and fractions of a small number of grains (Table 1). Eleven fractions were analyzed and four appear to contain an inherited component (Fig. 3a). Three fractions have relatively large errors due to low radiogenic to common Pb ratios (Table 1). Two of the highest-precision analyses are nearly concordant and one relatively lower precision analysis is concordant within error. The seven analyses define a concordant-to-discordant trend with an upper intercept of 538.8 ± 4.4 Ma and a lower intercept suggestive of recent Pb-loss. We consider the upper-intercept age as the time of zircon crystallization in the magma, and within error of the time of rhyolite extrusion.

This age indicates that the Simpsons Island Formation is earliest Cambrian in age (time scale of Okulitch 2002), consistent with the earlier interpretation of Currie (1997). The new age suggests that volcanic rocks in the area, assigned previously to the Simpsons Island and Goss Point formations, are of early Cambrian age (S. Johnson, personal communication 2003).

The Blacks Harbour Granite is located in a complexly faulted area, associated with unnamed and undated volcanic and sedimentary rocks, in faulted contact with Middle Cambrian mafic volcanic and sedimentary rocks. The volcanic and sedimentary unit, termed the Buckmans Creek beds (Currie 1988; Johnson 2001), has been interpreted to unconformably over-lie the Blacks Harbour Granite (Greenough et al. 1985; Currie 1988; Johnson 2001). Helmstaedt (1968) described the granite as mainly alaskite, described intrusive contacts with adjacent units, and considered the pluton to be Devonian. The dated sample NB99-2 is coarse-grained granodiorite collected in the village of Blacks Harbour. It consists of plagioclase, quartz, and less abundant orthoclase in allotriomorphic granular texture with less than 5% chlorite (after biotite).

Zircon grains separated from sample NB99-2 were of a wide variety of morphologies and sizes (Table 1). Eight zircon fractions ranging from a large single grain to 12 small prismatic grains were analyzed (Fig. 3b). Three fractions have 207Pb/²⁰⁶Pb ages of >685 Ma, which we interpret to contain an inherited component. One fraction is concordant at ~622 Ma and three others form a discordant line from that age. The four analyses yield an upper intercept age of 622 ± 2 Ma and a lower intercept suggesting recent Pb loss. One highly discordant fraction of partly metamict zircon grains plots to the left of this discordant trend, likely due to ancient Pb-loss, the age of which cannot be estimated reasonably from the data at hand. We interpret the upper intercept age of 622 ± 2 Ma to be the time of crystallization of the Blacks Harbour Granite.

A previous U-Pb (zircon) age for the Dipper Harbour Formation had a large error (Zain Eldeen 1991). Leucogranite of the Musquash Harbour Pluton, which appears to have a gradational relationship with the volcanic rocks (Currie and Hunt 1991), also yielded an age with a very wide error (555 ± 15 Ma). In order to better constrain the ages of both of these important units, felsic tuff sample NB01-136 was collected for dating. The sample consists of scattered embayed quartz and less abundant feldspar clasts in a fine-grained groundmass of quartz and feldspar. Zircon grains separated from sample NB01-136 were mainly prisms of sufficient size (70– 150 µm x 20– 40 µm) to do single-grain analyses or analyses of a small number of grains (Table 1). One analysis was a single pyramidal grain tip broken from a prismatic grain during sample processing. Of the eight analyzed zircon fractions, two appear to contain an inherited component (Fig. 3c). The remaining six form a discordant trend with an upper intercept age of 553 ± 3 Ma. One fraction in this discordant trend lies slightly to the right of the others and may contain a small inherited component. It was, nevertheless, included in the regression, which causes a high mean square of weighted deviates (MSWD). By excluding this fraction, the upper intercept age remains essentially the same, but the MSWD is reduced to <1. We consider the upper-intercept age of 553 ± 3 Ma as the time of zircon crystallization in the magma, and within error of the age of the volcanic rock.

Previous U-Pb crystallization ages for igneous units in the Caledonia, Brookville, and New River terranes are compiled on Fig. 4, together with the three new ages reported here. Ages in the Caledonia terrane fall into two main groups - older ages in the range 625– 615 Ma representing the Broad River Group and co-genetic plutons, and younger ages in the range 560– 550 Ma representing the Coldbrook Group and co-genetic plutons. A crystallization age of ca. 600 Ma reported for a tuffaceous sample previously assigned to the Broad River Group (Bevier and Barr 1990) may represent inherited zircon in a ca. 560– 550 Ma unit of the Coldbrook Group. Ash beds in the overlying Cambrian Saint John Group yielded zircon ages of 533 ± 2 Ma (Samson 1995); 531 ± 1 Ma (Isachsen et al. 1994), and ca. 511 Ma (Landing et al. 1998), as well as a detrital zircon population at ca. 523 Ma (Landing et al. 1998).

In the Brookville terrane, orthogneiss in the Brookville Gneiss yielded an age of ca. 605 Ma, interpreted to represent its igneous crystallization age (Bevier et al. 1990). However, plutons in the terrane yielded ages between ca. 553 Ma and 527 Ma, reflecting a long-lived magmatic event. The new age presented here for the Dipper Harbour Formation also constrains the age of the Musquash Harbour and similar plutons in the area because of the gradational relationship between the volcanic and plutonic units in the Dipper Harbour area.

In the New River terrane, ca. 625 Ma ages were reported for the Lingley and Brittain Creek plutons by Currie and McNicoll (1999); both plutons were included in the Lingley Suite by Johnson (2001, 2003). A similar age (622 Ma) was obtained in the present study for the Blacks Harbour Granite in the southwestern part of the New River terrane. In contrast, ca. 555 Ma ages have been obtained throughout the terrane, from the Lobster Brook Formation in the northeast and both the Ragged Falls suite and Leavitts Head Formation in the southwest. The new age reported here for the Simpsons Island Formation is significantly younger at ca. 539 Ma. The youngest dated volcanic unit in the terrane is the ca. 515 Ma Mosquito Lake Road Formation.

Close spatial relationships and chemical similarities support a co-magmatic relationship between volcanic and plutonic units of similar age in each of the Caledonia and Brookville terranes, with a range in U-Pb ages from each suite that likely reflects protracted magmatic activity (Barr and White 1999; White et al. 2002). However this relationship has yet to be tested in the New River terrane. Furthermore, the relationship between igneous units of similar age in different terranes is problematic. Currie and McNicoll (1999) and Johnson (2001) emphasized that the presence of ca. 625 Ma and 550 Ma igneous rocks, combined with the presence of the Cambrian Saint John Group or equivalent units containing Avalonian fauna, in both terranes strongly suggests that at least parts of the New River terrane have close links with the Caledonia terrane. Ca. 550 Ma igneous rocks also occur in the Brookville terrane, and ca. 531 Ma zircon grains in ash layers in the Saint John Group of the Caledonia terrane have similarities to the ages of some plutonic units in the Brookville terrane.

One way to assess whether or not units of similar age are likely to be related to one another is to compare their chemical characteristics, as done in the subsequent section for igneous units in the Caledonia, Brookville, and New River terranes.

A large database of whole-rock chemical analyses is available for the Caledonia (Barr and White 1999) and Brookville (White 1996; White et al. 2002) terranes, but relatively few analyses have been published previously for volcanic and plutonic units in the New River terrane (Johnson and McLeod 1996; Eby and Currie 1996). To enable a preliminary comparison with units of similar ages in adjacent terranes, additional analyses were obtained for 32 granitoid and volcanic samples from units in the New River terrane (Table 2). However, the data are mainly from the southwestern part of the terrane, and units northeast of the Robin Hood Lake Fault are poorly represented. Most of the analysed samples have high SiO₂ content, over 74%, reflecting the highly felsic compositions of the sampled units. Because the ages of mafic units are not well constrained, mafic samples were not included in the present study.

To illustrate compositional similarities and differences among the sample suites, major element oxides and selected trace elements are plotted against SiO₂(Figs. 5, 6). For visual clarity, suites with large numbers of analyses are represented by fields on the diagrams. Individual data points are shown for samples from the New River terrane and the Dipper Harbour Formation and related plutons of the Brookville terrane, as the smaller number of samples makes drawing fields less appropriate.

The ca. 620 Ma granitoid suite from the Caledonia terrane and the ca. 545– 525 Ma granitoid suite of the Brookville terrane show similar compositional trends in major element oxides, with overlapping fields for TiO₂, Al₂O₃, Fe₂O₃ ^t, MgO, CaO, Na₂O, K₂O, and P₂O₅ on plots against SiO₂ (Fig. 5). This similarity is consistent with the interpretations of Barr and White (1999) and White et al. (2002) that both suites formed in continental margin subduction zones. In contrast, the ca. 560– 550 Ma volcanic and granitoid rocks of the Caledonia terrane show similar fields for MgO, CaO, Na₂O, and K₂O but have higher TiO₂ and Fe₂O₃ ^t, and lower Al₂O₃ (Fig. 5), interpreted to be indicative of their origin in an extensional setting. The ca. 555– 545 Ma volcanic and granitoid rocks of the Brookville terrane (Dipper Harbour Formation and related plutons) also likely formed in an extensional setting, but differ from the Caledonia terrane units of similar age in that they show limited range in SiO₂, lower Na₂O, and higher K₂O (Fig. 5). Although Na₂O and K₂O can be mobile during alteration and weathering, petrographic observations showed no signifi-cant differences in these rocks that might be expected to result in the systematic differences in the oxide abundances, so the observed differences are considered to be significant. The ca. 555 Ma volcanic and plutonic units of the New River terrane also show most similarity in major element compositions to the ca. 560– 550 Ma Caledonia terrane units, but have higher K₂O contents in the most felsic samples (Fig. 5g). The number of samples (5) from ca. 625 Ma plutonic units in the New River terrane is too small to make any generalizations or comparisons with plutonic units of similar age in the Caledonia terrane, although the overall compositional variations are similar to those in the ca. 620 Ma Caledonia terrane suite.

In terms of trace elements, the ca. 620 Ma Caledonia terrane and ca. 545– 525 Ma Brookville terrane suites have similarities in Rb, Zr, Nb, and Y (Fig. 6), and hence plot in similar positions of chemical affinity and tectonic setting discrimination diagrams, consistent with their similar tectonic settings. However, the Brookville suite throughout the full range of SiO₂ contents has lower Ba and Sr than the Caledonia suite. Samples from ca. 625 Ma plutonic units in the New River terrane have trace element signatures similar to those in the ca. 620 Ma Caledonia suite, especially the low Y and Zr values in intermediate samples (Fig. 6e, f), and a continental margin subduction zone setting is suggested (Fig. 7).

The ca. 560– 550 Ma Caledonia volcanic and plutonic units have lower Ba (similar to the Brookville plutons in this case) and Sr, but higher Zr, Nb, and Y than the ca. 620 Ma suite from the same terrane. The difference is consistent with the inferred within-plate setting and enhanced "A-type" character of the ca. 560– 550 Ma suite compared to the volcanic-arc signature and more I-type features of ca. 620 Ma Caledonia terrane plutons. Barr and White (1996, 1999) interpreted the change to be related to later extension within a former subduction environment. The opposite pattern is true for the Brookville terrane, in which the younger ca. 545– 525 Ma plutons have a clear subduction-related signature, whereas the older ca. 555– 545 Ma Dipper Harbour Formation and related plutons have more extensional characteristics. The change was recognized by White et al. (2002), who suggested that magmatism in the Brookville terrane began with A-type magmatism, presumably related to crustal extension, but quickly evolved into a regime of continental margin-type subduction. Subduction continued to about 525 Ma and generated the voluminous gabbroic to monzogranitic (but dominantly tonalitic and granodioritic) plutons that comprise most of the Golden Grove Plutonic Suite.

Trace elements show some differences between the ca. 560– 550 Ma units of the Caledonia terrane and the ca. 555– 545 Ma units of the Brookville terrane, in spite of their similar ages and inferred tectonic settings, in that the Brookville terrane units tend to have higher Ba, Rb, and Nb (Fig. 6b, d, e). Comparing both of these units to the ca. 555 Ma volcanic and plutonic rocks of the New River terrane, some differences are also apparent, such as higher Rb in New River terrane units (Fig. 6b), although again the inferred tectonic setting is similar. It is interesting that the ca. 555– 545 Ma felsic units of the Brookville terrane and the ca. 555 Ma units of the New River terrane show more similarity to one another than to units of similar age in the Caledonia terrane. However, some subtle differences are apparent between the New River and Brookville terrane units, such as a tendency for lower Y and Zr in the New River terrane samples (Fig. 6e, f), and lower Ga/Al ratios (Fig. 7c).

Too few data are available from the ca. 555 Ma Lobster Brook and ca. 540 Ma Simpsons Island formations of the New River terrane to make any generalizations, except that they are broadly comparable to the other units and probably also formed in extensional tectonic settings. It appears that wherever these units were located relative to one another in the Late Neoproterozoic, all were forming in an extensional regime. However, it should be noted that Johnson and McLeod (1996) considered that the Simpsons Island Formation formed in a subduction environment, based on analyses from mafic samples in the unit. It is apparent that more detailed studies of these units are needed in order to make more meaningful interpretations.

It is important to emphasize that the lithological components of the Coldbrook Group and associated plutons also serve to distinguish them from units of overlapping age in the Brookville and New River terranes. The Coldbrook Group and associated plutons range in composition from basalt/gabbro to rhyolite/granite, with abundant dacitic volcanic units. None of these rocks match the almost entirely rhyolitic/granitic Dipper Harbour Formation and related ca. 555– 545 Ma plutons. Although not yet described in detail, it appears that lithological similarities are more apparent between the Dipper Harbour Formation and related plutons and the ca. 555 Ma plutonic and volcanic rocks of the Leavitts Head Formation and Ragged Falls suite in the New River terrane, which are also mainly rhyolitic and granitic.

Previous studies (e.g., Whalen et al. 1994, 1996; Samson et al. 2000) suggested that Sm-Nd isotopic data indicate differences among terranes in southern New Brunswick. As part of the present study, two additional samples from the Brookville terrane and one from the New River terrane were analysed for Sm and Nd isotopes (Table 3). Samples from the Lutes Mountain and McCarthy Point plutons (Brookville terrane) have small positive epsilon Nd values of 0.87 and 0.89, whereas the Blacks Harbour granite sample (New River terrane) has a lower epsilon Nd of – 3.13. These new data have been plotted together with pre-existing data against age on Fig. 8, and fit in with the trend for samples from the Brookville and New River terranes which have generally lower epsilon Nd, and more samples with negative values, than the Caledonia terrane (Fig. 8).

The data generally support the suggestion that magma generation was different in the different terranes, although the ranges show some overlap. Too few chemical data are available to provide a characterization of the ca. 620 Ma plutons in the New River terrane relative to ca. 625– 615 Ma plutons in the Caledonia terrane, but the negative signature (epsilon Nd – 3.13) of the Blacks Harbour granite is a major contrast compared to the signatures of ca. 625– 615 Ma units in Caledonia terranes which average 1.67. Similarly, the younger suite also differ on average, with ca. 560– 550 Ma plutons in the Caledonia terrane having average epsilon Nd of 1.0, units of ca. 555 to 525 Ma age in the Brookville terrane an average value of – 0.57, and in the New River terrane of – 1.28.

Kerr et al. (1995) showed that mainly positive epsilon Nd signatures characterize the Avalon terrane sensu stricto in its type area in Newfoundland, whereas negative values are typical of the adjacent Gander zone. Whalen et al. (1996) showed that in southern New Brunswick, Silurian and Devonian plutons have more transitional values in the vicinity of the boundary between the Mascarene and St. Croix terranes, which they considered to be the Avalon - Gander boundary, and that Silurian and Devonian rocks of the Saint George Batholith and the Kingston terranes have high values typical of "Avalon". However, these values are too high to have been derived from the crust of Avalon terrane sensu stricto, which would have had mainly negative values in the Silurian-Devonian. It is likely that the juvenile signatures of those rocks are related to the mainly mantle source, with little crustal contamination, and the change to lower values in plutons to the northwest could be related to increasing crustal contamination in areas of less voluminous magma generation. Our new isotopic data from Neoproterozoic - Cambrian igneous units support previous studies (e.g., Samson et al. 2000) which suggested that the boundary of the Avalon terrane coincides with the Caledonia - Clover Hill fault. The data further suggest that the Brookville and New River terranes have different crustal components, but more complete study of units is needed in the New River terrane to further assess the significance of the difference.