Evidence for an enriched asthenospheric source for coronitic metagabbros in the Adirondack Highlands

Anorthosite-mangerite-charnockite-granite (AMCG) magmatism is widespread throughout the Mesoproterozoic Grenville Province (Fig. 1), and nowhere else on Earth are massif anorthosite complexes, which underlie approximately ∼20% of the area, so abundant (Fig. 1; Corriveau et al., 2007; Morisset et al., 2008). The petrogenesis of massif anorthosite has been the focus of great interest for many decades, exemplified by the field and geochemical studies of Ronald Emslie (Hamilton et al., 2010). Although anorthosite complexes as old as 1.7 Ga are known in the Grenville Province, the last two major pulses of massif anorthosite intrusion range between 1.14–1.16 Ga and 1.05–1.08 Ga (Chiarenzelli and McLelland, 1991; Corrigan and van Breemen, 1997; Higgins and van Breemen, 1996; McLelland et al., 2010a). Both of these intervals coincide with late to post-tectonic environments following Shawinigan (ca. 1210–1160 Ma) and Ottawan (ca. 1090–1045 Ma) orogenesis, respectively (McLelland et al., 2010a, 2010b; Rivers, 2008). Understanding the links between tectonism and mantle geodynamics will provide considerable insight into the formation of anorthosite massifs and associated Fe-Ti deposits exemplified by the giant deposits at Tahawus, New York, and Lac Allard, Quebec (Laznicka, 2010). A fundamental question is the composition and origin of the parental magma of massif anorthosite (Yoder, 1968).

The Adirondack Highlands host the large Marcy anorthosite massif (Fig. 2) and several smaller anorthosite-cored domes that were emplaced immediately after Shawinigan orogenesis, during which the Adirondacks were accreted to Laurentia (Fig. 2; Chiarenzelli et al., 2009; Chiarenzelli et al., 2010b; McLelland et al., 2004, 2010b). Two widely contrasting models for the petrogenesis of anorthosite and related granitic rocks (AMCG suite) have been proposed (McLelland et al., 2004): (1) all members of the suite are comagmatic and fractionated from parental magma intermediate in composition and derived from lower crustal melting (Duchesne et al., 1999; Longhi et al., 1999; and Vander Auwera et al., 1998); and (2) members of AMCG suite are widely coeval, but not comagmatic. In the second scenario, anorthosites were derived from a gabbroic parental magma, while granitic members formed via melting of the lower crust during ponding of vast quantities of this gabbroic magma at the base of the crust (Corrigan and Hanmer, 1997; Emslie, 1978, 1985; Emslie et al., 1994; Hamilton et al., 1994; McLelland et al., 1996, 2004, 2010b; among others). In such a scenario, both components would be expected to have unique chemical and isotopic signatures, and may be influenced by interaction (assimilation or contamination) with crust of a varying age and origin (e.g., Frost et al., 2010; Peck et al., 2010). In addition, the chemistry of the mantle-derived member would depend on the type of mantle sampled (lithosphere versus asthenosphere, enriched versus depleted, etc.).

McLelland et al. (1996) suggested that the AMCG suite represents a bimodal, late- to postorogenic complex that intruded after lithospheric delamination. Corrigan and Hanmer (1997) suggested that AMCG development is related to thermal delamination of the lower lithosphere during oblique extension immediately following orogenesis. Uranium-Pb sensitive high-resolution ion microprobe (SHRIMP) zircon geochronology (McLelland et al., 1988, 2004) has yielded ages for all members of Adirondack AMCG suite and provided geochronological evidence that the complex is widely coeval, but not necessarily comagmatic (Fig. 3).

From a geographic and structural perspective, the Adirondacks consist of two separate terrains, divided by a late, low-angle, normal fault known as the Carthage-Colton mylonite zone (CCMZ; McLelland et al., 1996; Selleck et al., 2005). On the northwestern side of this boundary, the Adirondack Lowlands were metamorphosed to upper amphibolite facies during the Shawinigan orogeny (ca. 1210–1160 Ma), but evidence for Ottawan deformation has yet to be observed. The Highlands, southeast of the Carthage-Colton mylonite zone, were affected by the Shawinigan orogeny and also metamorphosed to granulite facies during the Ottawan orogeny (ca. 1090–1045 Ma). The Lowlands consist primarily of supracrustal gneisses with igneous intrusions of arc affinity (Carl and deLorraine, 1997; Chiarenzelli et al., 2010a, 2010b; McLelland et al., 1996; among others). Conversely, the Adirondack Highlands are primarily underlain by metaigneous rocks, including widespread granitic and charnockitic gneisses of the AMCG suite, which are much less abundant in the Lowlands. The Lowlands are widely thought to have been elevated in the crust above the Highlands during Ottawan orogenesis (McLelland et al., 2010b; orogenic lid of Rivers, 2008) and later juxtaposed adjacent to the Highlands along the CCMZ during post-Ottawan exhumation at ca. 1045 Ma (Selleck et al., 2005).

Chiarenzelli et al. (2010a) used trace-element chemistry and Nd systematics of early mafic and ultramafic rocks in the Lowlands to demonstrate that Laurentia was underlain by a highly enriched lithospheric mantle due to long-lived subduction and metasomatism of the mantle prior to Shawinigan orogenesis. The Shawinigan orogeny was recently interpreted as the result of the collapse of a backarc basin (Chiarenzelli et al., 2010b; McLelland et al. 2010b). Collapse was preceded by arc magmatism, which evolved through time and ended with syncollisional magmatic rocks including the ca. 1172 Ma Hyde School Gneiss in the Lowlands (Chiarenzelli and McLelland, 1991; Wasteneys et al., 1999), 1170–1180 Ma Rockport Granite in the Frontenac terrane (Marcantonio et al., 1990), and the ca. 1182 Ma Hermon granitic gneiss of the Lowlands (Heumann et al., 2006). Subsequently large volumes of AMCG rocks were intruded in the Adirondack Highlands (ca.1145–1160 Ma). Significant changes in magma composition and source occurred in concert with Shawinigan tectonic evolution, and it is within this context that the significance of the coronitic metagabbros (CMGs) will be evaluated and discussed.

Coronitic metagabbros generally occur within, or in close proximity to, the Adirondack anorthosite complexes, i.e., the Marcy massif, Oregon dome, and Snowy Mountain dome (Fig. 4) and associated with anorthosite massifs throughout the Grenville Province. Adirondack CMGs have SHRIMP U-Pb zircon ages (McLelland et al., 2004) of 1150 ± 14 Ma, coeval with other AMCG members and form dikes or round to irregularly-shaped bodies within the massif and domes, or within a few kilometers of their margins.

Oxygen isotopes from zircon separated from CMG have significantly lower (∼6.4‰) δ¹⁸O values than the anorthosite and related rocks of the Grenville allochthonous monocyclic belt (Marcy and Morin massifs), which yield elevated δ¹⁸O values (>8.0‰; Peck et al., 2010; Peck and Valley, 2000; Valley et al., 1994). These data suggest that the CMGs have some of the most primitive, mantle-like oxygen isotopic ratios reported from Adirondack metaigneous rocks. The higher δ¹⁸O values in zircons separated from anorthosite are consistent with contamination by, or assimilation of, crustal rocks. Unlike whole-rock, oxygen-isotope measurements, zircon oxygen-isotope analysis provides specific insight into the oxygen conditions prior to, and during crystallization, because it locks the ratio of light to heavy oxygen, which is not disturbed unless zircon's blocking temperature is reached.

Following the Shawinigan orogeny, subsequent deformation during the Ottawan orogeny overprinted the original field relationships and most primary igneous textures within the Adirondack Highlands. Our work suggests that the Sm-Nd isotopic systematics were preserved through granulite-facies metamorphism. This, paired with the major- and trace-element chemistry of CMG, which was also largely not impacted by metamorphism, provides important constraints on the source and evolution of anorthosite massifs within the Adirondacks and may provide insight into the origin of other examples within the Grenville Province, particularly those intruded at the end of the Shawinigan orogeny. Whole-rock Sm-Nd systematics and geochemistry presented herein will evaluate the primary signature of the source region and offer tectonic and petrogenic implications for Adirondack anorthosite and related rocks. The CMGs are part of a broader picture of magmatism that accompanied Shawinigan orogenesis. In particular, they coincide with the termination of the event and, as mafic rocks, provide insight into the state of the underlying mantle at that time.

Metagabbroic rocks (CMGs), originally olivine-rich, are common throughout the Grenville Province (Fig. 1). Where protected from pervasive deformation, these anhydrous rocks often develop coronitic texture at high-grade metamorphic conditions. Similarly, but less pronounced, coronitic textures are often developed within metanorthosite as well. Olivine metagabbros, often with coronitic textures, are widespread in the southern Grenville Province of Ontario and western Quebec (Davidson and Grant, 1986; Davidson and van Breeman, 1988). In the Adirondack Highlands, medium- to coarse-grained metagabbros, with remnant ophitic texture, exhibit corona growth around primary olivine, Fe-Ti oxides, and orthopyroxenes, and commonly exhibit spinel-clouded, zoned plagioclase (Whitney and McLelland, 1973; McLelland and Whitney, 1980). Coronas consist of concentric garnet, biotite, hornblende, and/or pyroxene rims form around and replace olivine and/or oxide grains (Fig. 5). Compositionally, these olivine metagabbros (CMGs) are most similar to rocks mapped in the Au Sable Forks quadrangle (Olmstead and Whitney, 1993), and CMGs were interpreted by the authors as slightly younger than other rocks of the AMCG suite. Bodies of CMG in the eastern and central Adirondacks Highlands are often only a few kilometers in diameter and are primarily associated with anorthosite, but also occur within mangeritic and charnockitic members of the AMCG complex and surrounding country rock (Fig. 6). Igneous compositional layering has been observed in a few of the larger bodies (McLelland et al., 1992).

In at least one location, the CMGs truncate a foliation in pelitic gneisses defined by a sillimanite-garnet assemblage, displaying the region's complex polymetamorphic history (McLelland and Chiarenzelli, 1989). This observation was crucial in determining that the region had undergone at least one previous high-grade (sillimanite-grade) metamorphic event prior to the Ottawan orogeny. Contacts between CMG and members of the AMCG suite are rarely preserved, but on Prospect Mountain, mangerite contains orthopyroxenes defining a linear fabric parallel to contact between the two lithologies (Fig. 6). The contact is sharp and irregular. No xenoliths or baked contacts are observed. The CMGs themselves often lack any mesoscale foliation or lineation, but occasionally grade into strongly foliated garnetiferous amphibolites at their margins.

In the central Adirondacks just east of Indian Lake, two CMG samples (SRG09-10a and 10b; Table 1) from the same outcrop vary in grain-size and the degree of fabric development. Here a foliated mangeritic dike cuts across the coarse-grained, foliated CMG (SRG09-10a), and contains xenoliths of gabbro, while the finer-grained CMG (SRG09-10b) lacks foliation. Field relationships at Humphrey and Sawyer mountains, both within ten kilometers of Indian Lake, indicate that strongly foliated charnockitic and amphibolitic gneisses envelope central cores of CMG. Here the CMG bodies are most likely rootless and may have been injected as sills and were subsequently broken into discrete blocks as the charnockite deformed around the competent gabbro. Furthermore, the CMG often grades into garnetiferous amphibolite along the margins containing xenoliths of, or layered with, coarse-grained pyroxenite. The pyroxenite likely represents restitic remnants of melting or metamorphic segregation along the hydrated margins of the bodies.

Mantle-like δ¹⁸O zircon values of the CMG suggest limited amounts of crustal contamination, assimilation, or interaction of meteoric waters (Peck et al., 2010; Peck and Valley, 2000; Valley et al., 1994; Valley, 2003), and their modal mineralogy indicates an olivine gabbronorite protolith (Fig. 7). The composition of the CMG suggests a mantle-derived origin (McLelland et al., 2004) and U-Pb zircon ages of ca.1145–1160 Ma suggest that the CMGs are contemporaneous with, or slightly younger than, Adirondack anorthosite. Several bodies of CMG appear to represent small, satellite plutons on the margin of the Marcy massif near Tahawus suggesting they were either sheltered from subsequent deformation, or are somewhat younger in age.

Although underlying a small percentage of the Adirondacks, the CMGs are crucial in understanding the petrogenesis of massif anorthosite (Table 1). The monomineralic nature of anorthosite restricts trace-element geochemical interpretation. However, related CMGs provide primary major- and trace-element geochemical compositions more amenable to interpretation (Table 2). Compositionally, the CMGs are primarily olivine gabbronorites (Fig. 7) with silica ranging from 45.51 to 48.10 wt%. Shown for comparison in Table 2 is the mean of 113 CMG major-element analyses provided by J. McLelland (Supplemental Table 1¹). These data together with new data presented herein are plotted on histograms showing normal distribution of MgO ranging from 5.15 to 10.13 wt%. For comparison purposes, an ultramafic sill in marble from outcrops along the Hudson River at Warrensburg (Moon Mountain) is shown, and occurs as an outlier with 20 wt% MgO (Fig. 8).

Normative mineral compositions of CMGs range from 18.52% to 31.49% (CIPW) olivine. Igneous olivine reacted with surrounding plagioclase to form much of the observed orthopyroxene observed in thin section (Table 3). The Moon Mountain ultramafic rock contains 48.87% olivine (CIPW), consistent with a cumulate origin. On AFM diagrams, the CMGs display a strong tholeiitic trend, although using the expanded data set of McLelland (Supplemental Table 1 [see ^{footnote 1}]), a minority of these analyses has calc-alkaline characteristics, most likely due to the lack of iron enrichment in these samples (Fig. 9). A classification diagram for basalts based on normalized clinopyroxene versus anorthite (Irvine and Baragar, 1971) constrains the composition of the CMGs to mostly alkaline basalts (Fig. 10; CIPW normative). Other tectonic discrimination diagrams, such as Zr versus Ti and Y versus Cr, conflict somewhat in detail, but consistently show enriched, within-plate composition, ranging from tholeiitic to alkaline for the CMG.

Normalized to chondritic values, coronitic metagabbros show slightly sloping trends (slightly LREE enriched) with small anomalies and rare-earth element (REE) enrichment up to roughly 100 times chondritic values (Fig. 11). For comparison purposes, also shown in Figure 11, is a suite of Paleoproterozoic gabbroic sills from the Central Hearne Domain (Griffin gabbros; Aspler et al., 2002) which were derived from an enriched asthenospheric mantle source (Aspler et al., 2002). Figure 12 presents the CMG data normalized to three common basaltic reservoirs—normal mid-ocean ridge basalt (NMORB), enriched mid-ocean ridge basalt (EMORB), and ocean-island basalt (OIB). The CMGs have patterns most similar to those of EMORBs and OIBs. La/Sm ratios, normalized to chondritic values, range from 1.42 to 1.98. Large-ion lithophile and high field strength elements concentrations in CMGs normalized to primitive mantle (Fig. 13; Sun and McDonough, 1989) show a positive Cs anomaly and a very slight negative Nb and Ti anomaly (small, and variable calc-alkaline component), but otherwise contain few distinguishable anomalies, notably lacking the pronounced subduction-related signatures seen in older arc-derived plutonic rocks of the Adirondacks (Chiarenzelli et al., 2010a; 2010b; McLelland and Chiarenzelli, 1990).

Also shown in Figure 13 is the incompatible element trend of CMGs compared to trends of pre-Shawinigan mafic and ultramafic rocks found in the Adirondack Lowlands (Chiarenzelli et al., 2010a). Notice the relative lack of Pb, Nb, P, and Ti anomalies in CMGs. This difference is crucial in determining the source of CMGs in the context of regional evolution. These trends suggest a source different from that of a subduction-altered lithospheric mantle for the CMGs even though such a source is known to have underlain much of the south-central Grenville Province prior to Shawinigan orogenesis (Chiarenzelli et al., 2010a, 2010b).

Samarium-Nd isotopic ratios have proven to be very useful in determining the source region and contamination of mafic igneous rocks in the Grenville Province and elsewhere (Chiarenzelli et al., 2010a; Cousens et al., 2001; Dickin and McNutt, 2007; McLelland et al., 1993; among others). Samarium and Nd concentrations were measured in nine CMG samples and one anorthosite sample, SRG09-03 (Table 4). Neodymium concentrations span from 11.95 to 39.87 ppm. Samarium ranges from 4.14 to 9.78 ppm. Epsilon Nd (ε_Nd(1150)) ranges from +3.13 to +3.69 and plots well below depleted mantle at time of crystallization (ε_Nd(1100) ∼+6, Hollings et al., 2007). The, CMGs have a very limited range in initial ε_Nd values compared to earlier magmatic suites like the Antwerp-Rossie Suite in the Lowlands (ca. 1203 Ma; Chiarenzelli et al., 2010b).

Coronitic metagabbros have higher Sm/Nd ratios and slightly shallower Nd evolution pathways than other members of the AMCG suite. However, the shallow slopes yield unrealistically old T_DM ages (mantle separation ages based on depleted mantle), which are older than any other suite yet measured in the Adirondacks, and are therefore unlikely to have been derived from a depleted mantle source. Depleted mantle model ages (T_DM) are consistent and range from 1546 to 1574 Ma. One outlier, the sample from the Mount Colden dike, which is perhaps the most primitive sample with the shallowest Nd evolution slope, yields a 1662 Ma T_DM. The T_DM calculated for the anorthosite sample analyzed is within the range measured for the CMGs (1570 Ma, Fig. 14). An errorchron age based on the Sm-Nd systematics of the CMGs yields a poorly constrained, and unrealistically old, age of 1521 +150/–160 Ma.

Twelve samples of the CMG and one anorthosite were analyzed to characterize their Rb-Sr isotopic systematics (Table 5). Measured ⁸⁷Sr/⁸⁶Sr ratios range from 0.704712 (SRG09-1) to 0.708498 (SRG09-7). One outlier, SRG09-10b, provides a ⁸⁷Sr/⁸⁶Sr ratio of 0.711463. This sample was taken from the outcrop near Indian Lake with two texturally distinct varieties of CMG and provides evidence of more than one episode of gabbroic intrusion. The ⁸⁷Sr/⁸⁶Sr versus ⁸⁷Rb/⁸⁶Sr errorchron yields a poorly constrained age of 976 ± 220 Ma, which overlaps zircon crystallization and metamorphic ages in the Highlands, and is too poorly constrained to be useful. It provides evidence that the Rb-Sr system was open during the Ottawan and, perhaps, Rigolet phases of orogenesis. When initial ε_Nd is plotted versus initial ⁸⁷Sr/⁸⁶Sr, a horizontal line is obtained, displaying the little variability in the Nd data, but a considerable spread in Sr data. This is also indicative of widespread Sr systematic disturbance during metamorphism and residence time within the deep crust, while Nd apparently remained relatively immobile.

In a geochemical study of mafic rocks within and up to 50 km away from the Adirondack anorthosite massifs, Olson (1992) considered the most primitive rocks of the mafic series to be silica-undersaturated gabbroic troctolites. In addition, she concluded that the more evolved rocks are basaltic in composition and derived from the same primitive silica-undersaturated gabbroic troctolites, further noting that the mafic series as a whole has high abundance of A1₂O₃, FeO, the light rare-earth elements (LREEs), other incompatible trace elements, and that even the most geochemically primitive compositions have high FeO contents. Together the mafic series and the silica-saturated anorthosite series were grouped into an anorthosite-norite-troctolite (ANT) suite.

Scoates and Mitchell (2000) studied the composition of troctolitic intrusions of 1.34–1.29 Ga Nain Plutonic Suite associated with the Voisey's Bay massive Ni-Cu sulfide deposit in coastal Labrador, Canada. Many of the plutons studied had fine-grained, olivine-rich marginal rocks whose composition approaches that of the parental magma. These olivine-normative, high Al basaltic magmas are interpreted by Scoates and Mitchell (2000) to represent the least evolved compositions in Proterozoic anorthosite plutonic suites (Mg# = 0.43–0.64; An = 45–60). They have high Al₂O₃ (15–19 wt%) and FeO_total (9–15 wt%) contents, and typically show the least contaminated Sr and Nd isotope compositions in a given plutonic suite. These traits are shared by the CMG analyzed and discussed here.

Seifert et al. (2010) recently summarized the geochemical data from over 700 samples of the Adirondack AMCG suite. Among their findings is that all rocks types, including anorthosites, leucogabbros, gabbros, OAGNs, OGNs, jotunites, monzodiorites, mangerites, and charnockites, display a mixture of tholeiitic and calc-alkaline compositional characteristics with iron-rich, major-element compositions and negative Nb and Ta anomalies. Of special relevance to this study are the compositions of oxide-apatite gabbronorites (OAGNs) and oxide-gabbronorites (OGN), separated on the basis of cumulate apatite, which can be directly compared to the CMGs (olivine gabbronorites) studied in this contribution.

Coronitic metagabbros have a much more limited composition range than either the OAGNs or OGNs. The oxide-rich gabbroic rocks contain 27.16 to 45.34 wt % SiO₂, while the range for the CMG samples investigated in this study is only 45.51 to 48.10 wt%. Data from J. M. McLelland for 113 coronitic gabbros in Supplemental Table 1 (see ^{footnote 1}) indicate they contain 47.38 ± 2.56 wt% SiO₂. Numerous other chemical differences occur between the oxide-rich gabbros and the CMGs, and this is likely due to the accumulation of oxides and, occasionally, apatite. One possibility, favored here, is that the CMGs represent nearly unfractionated parental magmas of the OAGNs, OGNs, leucogabbros, and anorthosite.

The CMG suite has lower TiO₂ than OGN and OAGN, and lower P₂O₅ than OAGN. This provides evidence for lack of significant cumulate oxides and apatite within the CMG. Figure 15 shows the subdivision of OAGNs and OGNs by plotting wt% of TiO₂ versus P₂O₅ (Seifert et al., 2010). On this diagram the CMGs plot below the OGNs and to the lower left of the OAGNs, indicating that they are indeed of a distinct composition and lack cumulate apatite and contain relatively small amounts of oxides. Note that gabbros with intermediate P₂O₅ values (0.75–1.25 wt%) were not observed by Seifert et al. (2010) or in this study.

As noted by Seifert et al. (2010), AMCG igneous rocks without significant alkali feldspar and quartz have REE values that can be almost entirely explained as a mixture of plagioclase and mafic minerals. In these diagrams, select rare-earth elements and ratios are normalized to chondritic values and plotted against Sm_N. As can be seen in the diagram plotting La_N/Sm_N versus Sm_N (Fig. 16), the CMGs plot in a near linear, restricted field between anorthosites, gabbros, leucogabbros, and OAGNs. On a similar diagram in the same figure where Sm_N/Eu_N is plotted against Sm_N, the same relationship holds. Together these diagrams suggest that mafic and intermediate rocks of the AMCG suite can be derived from magmatic compositions similar to CMG by the accumulation of plagioclase (anorthosite and leucogabbros) or extraction of cumulate oxides ± apatite (OGNs and OAGNs), and that the range in REE composition of the CMG is determined by such a process on a more limited scale.

Plotting of a compatible element (Cr) versus an incompatible element (Sm) reveals that anorthosites, gabbros, and leucogabbros plot near the origin and extend along both axes toward higher concentrations (Fig. 17). All OAGNs and OGNs have relatively high Sm and low Cr content and thus plot along the Sm axis. Coronitic metagabbros define a field that extends from the Sm axis toward higher Cr values. In general, the CMGs have more Sm than anorthosites and more Cr than most OAGNs and OGNs, likely as a function of their olivine content.

The relationships shown on these diagrams for whole-rock major- and trace-element compositions indicate that the CMGs have a distinct, and fairly homogeneous chemical composition among AMCG rocks and Adirondack mafic rocks in general (Olson, 1992). In addition, they generally plot in compositional fields that are distinct and consistent with the formation of other members of the anorthosite-mafic suite by fractional crystallization. Specifically, CMGs plot between gabbroic rocks enriched in oxides (OAGNs and OGNs) and rocks enriched in plagioclase feldspar (anorthosite and leucogabbros). These trends, particularly the Sm versus Cr relationships (Seifert et al., 2010), suggest differentiation of the CMG into other members of the suite can be readily explained by the accumulation of mafic minerals, particularly oxides and olivine, and plagioclase. Thus the composition of CMG is permissible as the parental magma of the anorthosite and related mafic rocks and may explain the origin of oxide-rich gabbroic ores, as has been suggested previously by numerous workers.

As summarized above, the Adirondack CMGs: (1) have mantle-like zircon δ¹⁸O isotope ratios; (2) contain petrographic evidence of, and normative, olivine; (3) have incompatible trace-element patterns most similar to EMORBS and other gabbroic rocks derived from an enriched asthenospheric source; (4) lack a strong subduction-related signature; (5) are dominantly tholeiitic; and (6) have consistent, but limited range in composition and in initial ε_Nd values compared to other metaigneous rocks in the Adirondacks. In particular, they yield depleted mantle ages 400 million years older than their crystallization age. Any model for their origin must take these and other characteristics into account.

Recent isotopic and geochemical studies suggest that much of the south-central part of the Grenville Province in Ontario, Quebec, and New York was underlain by metasomatized lithospheric mantle and lower crust (Chiarenzelli et al., 2010a; Bickford et al., 2010). This is compatible with the suggestion that the southeastern margin of Laurentia underwent several hundred million years of subduction beneath an Andean-type margin (Hanmer et al., 2000; McLelland et al., 2010a; 2010b) prior to the Shawinigan orogeny. This metasomatized mantle is reflected in the geochemistry and Nd systematics of several pre-Shawinigan intrusive suites within the Adirondack Lowlands, Highlands, and Frontenac Terrane (Chiarenzelli et al., 2010a, 2010b). Anorthosite and felsic members of the AMCG suite also display this calc-alkaline geochemical trend observed in older igneous rocks, likely derived from their source or via incorporation and assimilation of crustal material in the case of anorthosites. Any plutonic suite to have been derived from this source would, without a doubt, share its hallmarks.

Coronitic metagabbros have different incompatible element patterns and lack subduction-related anomalies shown in other members of the AMCG suite and older metaplutonic rocks throughout the Adirondacks. However, they are enriched when compared to typical NMORBs and in fact, EMORBs are a better match for their geochemical characteristics. As shown by older mafic rocks in the Adirondacks Lowlands (Chiarenzelli et al., 2010a), they also have relatively low ε_Nd (T_DM) values, unrealistically old T_DM ages, but high Sm/Nd ratios, and a Nd evolution pathway with a shallow slope. The CMGs have ε_Nd(1150) values that range from +3.13 to +3.69, and these are higher than that of the anorthosite (+2.19). This, along with mantle δ¹⁸O values in zircon, likely reflects crustal contamination or interaction with meteoric waters in the roof zone experienced by the anorthosite that is not observed in the CMG. The geochemical trends shown by the CMG are very similar to that of a series of Paleoproterozoic gabbro sills, intruding the base of the Hurwitz Group in Nunavut (Fig. 13). These gabbros, from the central Hurwitz basin, show no crustal contamination, an elevated and flat incompatible element trend, a sloping REE element trend, and are indicative of a peridotite garnet-bearing asthenospheric source, which formed in response to opening of the Maniwekan Ocean between the Rae-Hearne and Superior cratons (Aspler et al., 2002). Different mantle sources, dependent on tectonic history, age, and location are possible for other anorthosite massifs, and a depleted mantle source has been widely proposed (e.g., Frost et al., 2010; Morisset, 2008) for other massif anorthosites.

For these reasons we believe the CMGs were derived from asthenospheric mantle and that its characteristics noted above are primary and reflect this source. An intriguing question is: how was this asthenospheric source tapped and eventually intruded into the crust? In addition, how did this relate to the tectonic evolution of the Grenville Province during this time frame?

Given the strong influence of metasomatized enriched mantle on igneous suites that intruded before and during the Shawinigan orogeny (Chiarenzelli et al., 2010a), an intriguing question is why this signature is not present in late to post-Shawinigan igneous rocks, particularly the CMGs? If they too were derived from the same lithospheric source, the CMGs would display similar enrichments in large-ion lithophile elements and high field-strength element depletions relative to primitive mantle, which they do not. We propose, as numerous others have done, that delamination or considerable removal of the underlying lithosphere is required for asthenospheric mantle to reach the base of the crust and pond without significant interaction with the lithospheric mantle. Delamination in this case is due to physical detachment of the subducting plate, which descends into the mantle after an oceanic plate is consumed. Generally, this process is associated with a crustal block overriding and locking with the last bit of the subducting plate, which then detaches and descends. Suites like the Hermon granitic gneiss and Rockport granite in the Lowlands, intruded during mid- to late-Shawinigan orogenesis, have a mild incompatible element signature. Thus delamination may have already begun by ca. 1182 Ma in the Adirondack Lowlands and Frontenac Terrane due to the increase in the A-type character of igneous suites after this time (Heumann et al., 2006). The CMGs then represent the rise and melting of a hot, asthenospheric mantle that moved in as space was created during delamination and sinking of the dense lithospheric plate (McLelland et al., 2010a).

These data support the second and complementary part of this model following delamination of lithospheric mantle, namely movement of large volumes of hot, gabbroic magma to the crust-mantle boundary. As outlined by McLelland et al. (2004), fractional crystallization and sinking of olivine and pyroxene would decrease magma density, elevating isotherms, and melt the lower crust (Bickford et al., 2010; Emslie, 1978; Longhi and Ashwal, 1985). At the base of crust the magma then would crystallize abundant plagioclase (massif anorthosite), melt large volumes of continental crust (mangerite, charnockite, and granites), and incorporate a great deal of crustal contamination (McLelland et al., 2004; Bickford et al., 2010) in the felsic and intermediate rocks. Accumulation of oxides and apatite in residual gabbroic magmas, from which plagioclase was separated, would give rise to the OGN and OAGNs. The small group of CMGs with calc-alkaline signature on the AFM diagram may represent samples that have had increased interaction with the thinned, but remnant, lithospheric mantle. However, because the lower crust in this region is juvenile and only slightly older than the intruded magma, it experienced only limited isotopic evolution, and the difference in Nd systematics between mafic and felsic members of the AMCG suite in the Adirondack Highlands is relatively minor. The same is true for the small differences between massif anorthosite and CMG, even though massif anorthosite commonly shows mineralogical, geochemical, or isotopic evidence of variable crustal contamination (e.g., Frost et al., 2010; Peck et al., 2010; Seifert et al., 2010).

However, the CMGs are mutually intrusive with the AMCG suite and may have been entrained within the plagioclase crystal mush resulting in small, chimney-like diapirs. Relative minor volumes of CMG, generally within close proximity of coeval massif anorthosite, indicate that their intrusion was not a separate and unrelated event. Therefore, CMGs would have escaped major contamination (supported by oxygen-isotope and Sm/Nd ratios), providing a pristine sample of the unfractionated parental magma (Valley et al., 1994; Valley, 2003).

Corrigan and Hanmer (1997) proposed thinning of the lithospheric mantle during the late- to post-orogenic stage of mildly extensional regimes to account for the intrusion of massif anorthosite in the Grenville Province. McLelland et al. (2010a) appeals to various possible mechanisms for thinning, or removal, of the lithosphere and ascent of asthenospheric mantle (including flat-slab subduction, backarc extension, slab breakoff, and hotspots), which is considered necessary for the production of massif anorthosites. This is consistent with our geochemical observations, which suggest the presence of lithospheric mantle is very limited and/or nearly absent during AMCG magmatism in the Adirondacks. A thinned or missing metasomatized lithosphere known to have once underlain the Adirondack region (Chiarenzelli et al., 2010a) readily explains the lack of a well-defined subduction signature in the CMG (Cs and Nb anomaly).

(1) CMGs display textures (coronas) indicative of upper granulite-facies reactions associated with Ottawan metamorphism. However, primary igneous features, such as ophitic texture, are still partially preserved. In general, they are thought to have behaved as closed geochemical systems; however, there is evidence for the disturbance of the strontium isotopic signature. However, Nd isotopes display closed system behavior and indicate a homogenous source.

(2) Uranium-Pb zircon geochronology points to a comagmatic origin of CMG with massif anorthosite (McLelland et al., 2004), whereas δ¹⁸O in zircons shows highly elevated values within the anorthosite and other felsic AMCG plutonic rocks and mantle-like values for CMG. This suggests that the CMGs underwent less contamination or assimilation than nearby massif anorthosite, and may represent parental magmas to the Adirondack anorthosite (Valley et al., 1994; McLelland et al., 2004).

(3) Several lines of evidence suggest CMGs from the Adirondack Highlands represent a melt of a primitive asthenospheric mantle source. Incompatible elements normalized to primitive mantle (Sun and McDonough, 1989) and La/Sm ratios support an asthenospheric source. Conversely, derivation from lithosphere mantle enriched by crustal contamination or other processes is thought unlikely because of mantle-like, oxygen-isotope values of zircon separated from CMG and distinct differences in their trace-element composition when compared to samples of the lithospheric mantle known to underlie the region prior to Shawinigan orogenesis (Chiarenzelli et al., 2010a).

(4) Comparison with massif anorthosite and related metaigneous rocks and oxide-apatite gabbronorites and oxide-gabbronorites from the Adirondacks (Seifert et al. 2010) suggests that the CMGs plot in between fields encompassing gabbros and the anorthosites and leucogabbros. Their composition is appropriate and permissible for the derivation of these plagioclase- and oxide-rich daughter products by fractional crystallization, residual concentration, and Fe enrichment.

(5) Isotopic results point to an enriched and homogeneous source with ε_Nd(1150) values for CMGs between +3.13 to +3.69, plotting well below contemporaneous depleted mantle, and yielding T_DM ages in excess of 400 m.y. older than their crystallization. However, this departure from the depleted mantle curve is not likely a function of crustal contamination, as observed in other members of the AMCG suite, including massif anorthosites given the limited compositional range, homogeneous Nd isotopic systematics, and mantle-like, oxygen-isotope values for zircon from CMGs.

(6) Older (pre-Shawinigan) igneous suites within the Adirondacks are derived from highly altered and metasomatized lithospheric mantle (Chiarenzelli et al., 2010a), but such a source is incompatible with the geochemistry of the coronitic gabbros. An asthenospheric source is compatible with models that invoke lithosphere thinning and/or delamination as a triggering mechanism for massif anorthosite production (Corrigan and Hanmer, 1997; McLelland et al., 2010a).

(7) Delamination occurs after orogenic events when the dense, altered lithospheric mantle thins and peels away from the base of the crust or the subducting plate detaches and sinks. These processes would allow the asthenosphere to rise, melt via decompression, and pond at the base of the crust setting the stage for the production of the massif anorthosite (Fig 18). These melts (parental magmas), when unfractionated, have the composition of olivine-bearing alkali basalts. As noted by McLelland et al. (2010b), alternative settings for anorthosite and AMCG suite genesis that can account for asthenospheric rise, melting, gabbro underplating, and fractionation are numerous. In the southern Grenville Province, a strong case for delamination at the end of the Shawinigan orogeny has been made and provides a mechanism for ca. 1155 Ma massif anorthosite genesis.

(8) The gabbroic magma produced by decompressional asthenospheric melting would crystallize abundant plagioclase of compositions consistent with the production of Adirondack anorthosite. Positive buoyancy would allow it to float to the roof of the magma chamber. As dense mafic minerals crystallized synchronously with plagioclase feldspar, the temperature within the magma chamber would rise due to the heat of crystallization, raising crustal isotherms (McLelland et al., 2004). Driven by density differences, the upper plagioclase-rich, less dense portions of the magma chamber would quickly rise into the middle continental crust, melting surrounding wall rock. Evidence for meteoric water suggests relatively shallow depths for ultimate emplacement. The plagioclase crystal mush would likely incorporate a great deal of contamination from meteoric water interactions within the roof zone. Later batches of unfractionated, gabbroic parental magma would rise and cool within proximity to these anorthosite bodies. Some may have been trapped and buoyed up by large plagioclase rafts. They would lack crustal contamination and have geochemical and isotopic characteristics compatible with their asthenospheric source.

Fourteen samples of coronitic metagabbros were collected during the summer of 2009 (one anorthosite and one ultramafic), and were acquired from well-exposed, fresh, blasted road outcrop. The sample from Mount Colden is an exposure on the East side of Avalanche Pass from fresh outcrop of the Colden trap dike. Twelve samples were crushed into small chips using a jaw crusher and then powdered using a tungsten ball mill. Tantalum contributions to the sample were suspect and were disregarded. Analyses were completed at Acme Labs in Vancouver using inductively coupled plasma–mass spectrometry (ICP-MS) for 72-element analysis, including major oxides and trace elements.

Eleven samples were run for Sm-Nd and Rb-Sr isotopic analysis at Carleton University in Ottawa, Ontario, at the end of July. 0.07927 to 0.47564 grams of sample were dissolved in a Teflon dish after the ¹⁴⁹Nd/¹⁴⁸Sm spike was added (0.099 grams of spike) to the samples that were going to be run for Nd isotopes. The samples were dissolved in HNO₃-HF solution and further with HNO₃ and HCl until there was no visible residue. Rare-earth elements, Rb, and Sr were separated using a Dowers 50-XS. This solution was then dissolved in 0.26 N HCl and loaded onto an Eichron chromatograph column. These columns contain Teflon powder coated with HDEHP (di[2-ethylehexyl] orthophosphoric acid); Richard et al., 1976). A 0.26 N and 0.5 N HCl acid was used for Nd and Sm collection, respectively.

Rb-Sr collection was done on a double tungsten filament using the Thermo Finnigan Triton T1 multicollector mass spectrometer. Error reported at 2 sigma (or with 95% confidence) ranges from 0.000008 to 0.000011.

We would like to thank the St. Lawrence University Fellowship Program, namely the Donald K. Rose Fund for providing monetary support for the sample acquisition and preparation. Secondly, we would like to thank William B. Bradbury, Jr. Faculty Award from St. Lawrence University for providing funding for geochemical and isotopic analysis. Karl Seifert graciously allowed us to use the data in his paper on the chemistry of the AMCG suite in this volume and provided an excellent review of the geochemical implications of our study. We would also like to thank Drs. William Peck from Colgate University, Graham Baird at the University of Northern Colorado, and Carol Frost at the University of Wyoming for extremely insightful and helpful reviews. We would also like to acknowledge the fine work of Dr. Ronald Emslie and his career of inquiry into these fascinating rocks.