Geochemical and tectonic evolution of the Ordovician Bronson Hill arc and Silurian and Devonian Connecticut Valley–Gaspé trough: Eastern Vermont and western New Hampshire, USA Open Access

The tectonic evolution of volcanic arcs, arc-continent collision, and subsequent basin formation and post-orogenic extension are often considered separate and unrelated in many geologic studies. Study of any relationship between the two may be further complicated in polydeformed terranes because metamorphism or deformation have changed the mineralogy, obliterated primary textures, and disrupted the original stratigraphic relationships. Despite these complications, interpreting the geochemical relationships of metamorphosed plutonic and volcanic arc rocks, and metavolcanic rocks within post-orogenic basins, can still be useful for understanding the evolution of ancient volcanic arcs and arc-continent collisions. The development and evaluation of geochemical discrimination diagrams using relatively immobile trace elements has proven useful to scientists for determining the tectonic setting, significance, and evolution of altered and metamorphosed magmatic rocks (Shervais, 1982; Pearce et al., 1984; Hildebrand et al., 2018). In New England, USA, debate has centered around the relationship, if any, between igneous rocks of the Oliverian Plutonic Suite and Ammonoosuc Volcanics and how the Bronson Hill arc (BHA) evolved through time as it collided with Laurentia (e.g., Stanley and Ratcliffe, 1985; Robinson et al., 1991; Ratcliffe et al., 1998; Karabinos et al., 1998, 2017; Dorais et al., 2008). Further, the overlying Silurian through Devonian Connecticut Valley–Gaspé trough (CVGT; Hibbard et al., 2006) is generally considered separately from the Ordovician BHA (e.g., Tremblay and Pinet, 2005; Rankin et al., 2007; Dorais et al., 2017). The goal of this paper is to evaluate how these rocks relate to each other and to illustrate that there is a continuous evolution of magmatism in the northern Appalachians from island arc magmatism to arc-continent collision to post-orogenic basin development and slab breakoff following the Taconic orogeny.

The ~600-km-long BHA extends from Long Island Sound to western Maine, and the >1000-km-long CVGT extends from southern Connecticut to the Gaspé Peninsula in Québec. Both form prominent geologic features in the Northern Appalachians of New England and Eastern Canada (e.g., Williams, 1978; Hibbard et al., 2006; Perrot et al., 2018; Valley et al., 2020; Figs. 1 and 2). The rocks of the BHA are exposed in a series of gneiss domes, which are the result of a complex history of thrust faulting and folding that developed during the Devonian Acadian through Carboniferous to Permian Alleghanian orogenies (Walsh et al., 2020b). The BHA consists of metamorphosed Ordovician basalt and basaltic andesite, rhyolite, trondhjemite, and granite, and a metamorphosed cover sequence of graphitic-sulfidic schist, volcanic rocks, and minor quartzite (e.g., Billings, 1956; Zen et al., 1983; Tucker and Robinson, 1990; Lyons et al., 1997; Moench and Aleinikoff, 2003; Hollocher et al., 2002; Ratcliffe et al., 2011). The rocks are interpreted to be part of the larger Shelburne Falls–Bronson Hill arc system that collided with Laurentia and resulted in the Taconic orogeny from ca. 465 Ma to 445 Ma. (e.g., Stanley and Ratcliffe, 1985; Karabinos et al., 1998; Valley et al., 2020).

Based on Nd, Pb, and Sr isotopes (Aleinikoff et al., 2007; Dorais et al., 2008) and detrital zircon data, the BHA is built on Ganderian crust (Macdonald et al., 2014; Karabinos et al., 2017). Exposed within the southern BHA in Massachusetts, the Dry Hill Gneiss (dated at 613 ± 3 Ma) is crust that predates the Ordovician arc (Tucker and Robinson, 1990). The Dry Hill Gneiss represents Ganderian crust beneath the BHA (Aleinikoff et al., 2007). The BHA is one of several Northern Appalachian volcanic arcs that were built on a peri-Gondwanan (Ganderian) crustal fragment in the Iapetus Ocean (Walsh et al., 2021); others include the Penobscot arc-backarc system (513–482 Ma), and the Popelogan-Victoria arc (475–455 Ma), the latter of which is the on-strike and correlative with the Bronson Hill arc in Newfoundland, Maine, and New Brunswick (Fig. 1; Hibbard et al., 2006; van Staal and Barr, 2012; van Staal et al., 2016).

Overlying the BHA, the CVGT forms an unconformable autochthonous to parautochthonous cover sequence on the pre-Silurian rocks (Figs. 2 and 3). The origins of CVGT are debated. Previous publications have proposed that basin development was the result of: (1) inter-arc or intercontinental extension following slab breakoff and upwelling of asthenospheric melts (van Staal and de Roo, 1996; Robinson et al., 1998; Tremblay and Pinet, 2005; Rankin et al., 2007; Dostal et al., 2020; this study), (2) a backarc basin that developed due to westward subduction at the end of the Taconic orogeny (Karabinos et al., 1998, 2017), (3) extension related to slab rollback of the subducted plate (Moench and Aleinikoff, 2003; Rankin et al., 2007), (4) a transpressional rift (Keppie and Dostal, 1994), and (5) a transtensional basin related to the Salinic disturbance in eastern Canada (e.g., Dostal et al., 1993; Tremblay and Pinet, 2016; Perrot et al., 2018). The CVGT transitioned from an intercontinental backarc, or intra-arc extensional setting after the Ordovician Taconic orogeny to a foreland basin setting over a west-dipping subduction zone at the onset of the Devonian Acadian orogeny (Bradley and Tucker, 2002; Tremblay and Pinet, 2005, 2016; McWilliams et al., 2010; Perrot et al., 2018).

Subduction polarity directions and polarity reversals are commonly the subject of debate in ancient orogens (e.g., Pavlis et al., 2019, and references within). Early models for the Taconic orogeny favored eastward subduction beneath the BHA (Rowley and Kidd, 1981; Stanley and Ratcliffe, 1985). This model was subsequently challenged based on U-Pb zircon ages between ca. 454 Ma and 442 Ma from the BHA in Massachusetts (Tucker and Robinson, 1990). These ages apparently postdated the 465 Ma Taconian metamorphic ages (Sutter et al., 1985), implying that the BHA was too young to have participated in the Taconic orogeny. Later interpretations incorporated polarity reversal and development of multiple arcs to account for the age discrepancy (Karabinos et al., 1998). More recent geochronologic studies showed that the BHA is as old as ca. 492 Ma in northern New Hampshire and was active for at least 50 million years from ca. 490 Ma to ca. 440 Ma (Moench and Aleinikoff, 2003; Rankin et al., 2013; Valley et al., 2020; Fig. 4) and that Taconic metamorphism lasted until ca. 445 Ma (Hames et al., 1991; Ratcliffe et al., 1998; Walsh et al., 2004). The older magmatic ages and younger metamorphic ages eliminate the need for subduction polarity reversal. Hildebrand and Whalen (2021) used trace element ratios to suggest that magmatic rocks younger than 450 Ma were the result of slab failure and not island arc generated magmas, which supports the hypothesis that polarity reversal was not needed.

The CVGT contains Silurian to Devonian metasedimentary rocks and bimodal volcanic rocks. Metasedimentary rocks are predominantly metamorphosed calcareous or siliceous mudstones. U-Pb zircon ages indicate maximum depositional ages between ca. 432 Ma and 407 Ma (Aleinikoff and Karabinos, 1990; Rankin and Tucker, 2009; McWilliams et al., 2010; Dorais et al., 2017; Perrot et al., 2018; Converse et al., 2022). Rare fossils from the CVGT range from late Silurian (Pridoli) to Early Devonian (Emsian; Boucot and Drapeau, 1968; Hueber et al., 1990; Lavoie and Asselin, 2004). Previous studies of the geochemistry of metavolcanic rocks from the CVGT and contemporaneous intrusive rocks from adjacent belts indicate that these Silurian to Devonian rocks contain mixed geochemical signatures from island arc basalt to backarc basin or mid-ocean ridge basalt to within-plate tholeiites (e.g., Dostal et al., 1993, 2017; Rankin et al., 2007). Proposed driving mechanisms for the associated volcanism include backarc basin extension, slab breakoff, slab rollback, transtensional shearing, lithospheric delamination, and extensional collapse. (e.g., Dostal et al., 1993, 2020; Tremblay and Pinet, 2005; Rankin et al., 2007; McWilliams et al., 2010; Dorais et al., 2017).

Some of the older studies that have tried to identify tectonic setting and petrogenesis of the BHA and CVGT have lacked the benefit of precise whole-rock geochemical analytical techniques, a lack of robust geochemical tectonic discrimination diagrams, and a lack of modern geochronology. In this paper, we present new major and trace element geochemistry for 94 samples (data are available in Valley et al., 2023, and Table S1¹). These results were evaluated alongside the existing geochemical data from the BHA (Aleinikoff et al., 2007; Dorais et al., 2008, 2012; Hollocher, 1993; Leo, 1991; Schumacher, 1988) and CVGT (Rankin et al., 2007; Dorais et al., 2017) where possible, to reach a better understanding of the tectonic and magmatic evolution along the length of the BHA and CVGT in New England. The geochemistry is combined with the most recent modern geochronologic ages for the sampled lithologies where possible. In many previous studies, the BHA and CVGT are considered separately. The addition of new modern major and trace element geochemistry and the synthesis of existing data, allow us to create a robust model for the spatial and temporal evolution of the New England Taconic orogeny and the onset of the Acadian orogeny from subduction to collision, and basin formation in this ancient polydeformed mountain belt. Our study benefits from the fact that new geochemical samples were collected as part of an eight quadrangle U.S. Geological Survey, 1:24,000-scale bedrock mapping project in eastern Vermont and western New Hampshire (Fig. 3). This has allowed for geochemical findings to be interpreted in the proper context of their field relations.

Understanding the relationships among the igneous rocks of the BHA and CVGT, their geochemical evolution through space and time, and ultimately the plausible tectonic setting in which they originated, is essential for understanding the tectonic evolution of New England. Evaluating this extensive dataset in the context of existing U-Pb zircon geochronology (Valley et al., 2020) allows for a clearer understanding of the geologic evolution of New England from the Lower Ordovician to the beginning of the Devonian and the onset of the Acadian orogeny.

New 1:24,000 scale geologic mapping by the U.S. Geological Survey in the Connecticut Valley of eastern Vermont and western New Hampshire shows that the geology of the area consists of highly deformed and metamorphosed Mesoproterozoic through Devonian metasedimentary and metaigneous rocks. These rocks were intruded by Devonian to Mesozoic rocks (Fig. 3; Walsh, 2016; Walsh et al., 2020a, 2020b). On the western side of the study area, Mesoproterozoic gneisses of the Mount Holly Complex are exposed Laurentian basement in the Chester dome (Ratcliffe et al., 2011). The Moretown slice structurally overlies the Chester dome along the Keyes Mountain thrust fault (Ratcliffe, 2000a, 2000b; Ratcliffe et al., 2011); the fault represents the Beothuk Lake Line² (BLL), the ancient Ordovician suture between crustal blocks with Laurentian versus Ganderian affinity (Macdonald et al., 2014; Valley et al., 2020; Beranek et al., 2022). The allochthonous Cambrian through Ordovician Moretown slice of Ganderian affinity includes the Moretown and Cram Hill Formations and the North River Igneous Suite. Silurian and Devonian metasedimentary and metavolcanic rocks of the CVGT unconformably overlie the Moretown slice in a post-Taconic successor basin. The easternmost extent of the CVGT is exposed in western New Hampshire. Ordovician to Silurian and Devonian metasedimentary and metaigneous rocks, including the BHA, are collectively referred to as the New Hampshire sequence (NHS; Billings, 1937, 1956; White and Jahns, 1950; Ratcliffe et al., 2011; Rankin et al., 2013) and the entire NHS structurally overlies the CVGT along the Monroe thrust fault (Walsh, 2016; Walsh et al., 2020a, 2020b). The oldest part of the NHS was built on Ganderian crust (Rankin et al., 2013) and consists of Ordovician metamorphosed volcanic, plutonic, and sedimentary rocks, including the Ammonoosuc Volcanics, the Partridge Formation, and the Oliverian Plutonic Suite. The Ammonoosuc Volcanics are the base of the exposed section in the study area, but farther north near Littleton, New Hampshire, the Ammonoosuc Volcanics stratigraphically overlie the Albee Formation on which the arc was built (Rankin et al., 2013). The base of the BHA is partly correlative with rocks of the Cambrian-Ordovician Shelburne Falls arc that is found in the Moretown slice (Coish et al., 2015; Karabinos et al., 2017; Valley et al., 2020). The BHA rocks are exposed in fault-bounded structural belts in the study area (Fig. 3). In the NHS, Silurian to Devonian metasedimentary rocks of the Clough Quartzite and the Fitch and Littleton Formations unconformably overlie the BHA rocks. Though similar in age, the upper part of the NHS is not considered part of the CVGT due to the structural discontinuity along the Monroe thrust fault. All the rocks in the study area are crosscut by Devonian granitoids of the New Hampshire Plutonic Suite and the post-tectonic Mesozoic White Mountain Igneous Suite (Walsh et al., 2020a, 2020b).

From oldest to youngest, the stratigraphy of the NHS consists of the Albee Formation (Late Cambrian or older), Ammonoosuc Volcanics (Middle to Late Ordovician), Partridge Formation (Late Ordovician), Clough Quartzite (early Silurian), Fitch Formation (late Silurian–Early Devonian), and Littleton Formation (Early Devonian). A regionally significant Silurian unconformity occurs at the base of the Clough Quartzite (Billings, 1937), and this unconformity is attributed to the Salinic disturbance (e.g., Boucot, 1962; Perrot et al., 2018). According to Rankin et al. (2013), the stratigraphy of the Ammonoosuc Volcanics (from oldest to youngest) includes: (1) rusty sulfidic slate, other metasedimentary rocks, and felsic tuff; (2) metasiltstone, phyllite, and volcaniclastic rocks; (3) metadolostone and siltstone; (4) metarhyolite tuff and siltstone; (5) metaandesite, basaltic tuff, and pillow lavas; (6) metarhyolite tuff, lapilli tuff, and lava; and (7) metafelsic and mafic volcanics, volcaniclastic rocks, and metasedimentary rocks. The upper part of the Ammonoosuc Volcanics is age correlative to the lower part of the Partridge Formation (Rankin et al., 2013). The lower part of the Partridge Formation consists of interbedded metavolcanic rocks, rusty sulfidic schist, and slate interlayered with the upper part of the Ammonoosuc Volcanics. The Partridge Formation is overlain by the Quimby Formation in northern New Hampshire and western Maine. Felsic metatuff in the Quimby Formation yields an age of 443 ± 4 Ma (Moench and Aleinikoff, 2003), but is not present in the study area.

In the NHS, the base of the Silurian and Devonian cover sequence is the Clough Quartzite, which mostly consists of quartzite and quartz pebble conglomerate, with minor schist, calc-silicate, and amphibolite. The upper Silurian to Lower Devonian Fitch Formation overlies the Clough Quartzite and consists of calc-silicate granofels and schist, with minor quartzite and greenstone. The Devonian Littleton Formation overlies the Fitch Formation and is interpreted as a metaturbidite (Bradley, 1983) dominated by biotite-muscovite schist, with minor conglomerate and quartzite.

Mafic greenstone, amphibolite, and metadiabase are rare in the lower part of the cover sequence but are present within the Clough Quartzite and Fitch Formation. Here they have been interpreted as flows with rare pillow lavas, as non-distinct layers, or as dikes (Walsh, 2016; Walsh et al., 2020a). The intrusive rocks are correlated with the Comerford Intrusive Complex (Walsh, 2016; Walsh et al., 2020a, 2020b).

Following the Taconic orogeny, a period of extension created the CVGT and Central Maine trough. The CVGT lies unconformably or in local fault contact (McWilliams et al., 2010; Karabinos et al., 2010; Ratcliffe et al., 2011) atop remnants of the early Paleozoic Shelburne Falls arc to the west (Karabinos et al., 1998; Karabinos and Hepburn, 2001) and the BHA to the east (Stanley and Ratcliffe, 1985; Ratcliffe et al., 1998; Tucker and Robinson, 1990; Leo, 1985, 1991; Dorais et al., 2008, 2017; Valley et al., 2020). From oldest to youngest the CVGT is composed of the Silurian to Devonian Shaw Mountain, Northfield, Waits River, and Gile Mountain Formations. The Waits River and Gile Mountain Formations represent the majority of the CVGT in Vermont, and are correlative with the Ayers Cliff and Compton Formations, respectively, in Québec (e.g., Perrot et al., 2018). The Waits River Formation consists of variably calcareous and graphitic schist and phyllite, impure marble, and calcareous to non-calcareous quartzite (Walsh, 2016; Walsh et al., 2020a, 2020b). The overlying Gile Mountain Formation marks a transition from calcareous to more siliceous sediment input. The unit is composed of feldspathic quartzite, quartzite, metapelite, conglomerate, graphitic slate and phyllite, and minor marble (Walsh, 2016; Walsh et al., 2020a, 2020b).

In Vermont and New Hampshire, the metavolcanic rocks in the CVGT consist of a heterogeneous assemblage of mafic and intermediate to felsic volcanic rocks interbedded with volcaniclastic and sedimentary rocks. The rocks were previously assigned different names, including the Standing Pond Amphibolite Member of Memphremagog Formation (Doll, 1944), the Amphibolite Member of the Waits River Formation (Lyons, 1955), hornblende schist within the Gile Mountain Formation (Lyons, 1955), the Standing Pond Volcanic Member of the Waits River Formation (Doll et al., 1961), the Putney Volcanics (Trask, 1980), and the Standing Pond Volcanics (Hepburn et al., 1984). Metavolcanic and meta-volcaniclastic rocks are interbedded at multiple horizons within the metasedimentary rocks, and mostly occur within the Waits River Formation but also occur in the Gile Mountain Formation (e.g., McWilliams et al., 2010; Ratcliffe et al., 2011). The metavolcanic rocks cannot be mapped continuously as a single stratigraphic horizon between the Waits River and Gile Mountain Formation as previously thought (Doll et al., 1961), and thus the volcanic rocks are no longer assigned separate formation names (Ratcliffe et al., 2011).

The Frontenac Formation is laterally correlative with the Waits River Formation and is exposed in northeast Vermont and northwest New Hampshire, where it continues into Québec (Ratcliffe et al., 2011; Dorais et al., 2017; Perrot et al., 2018). The Littleton Formation is mapped at the top of the CVGT and the NHS and is laterally correlative with the Gile Mountain Formation (Rankin and Tucker, 2009; Ratcliffe et al., 2011).

The CVGT and the BHA are separated by the Monroe fault (Fig. 3; Hatch et al., 1988; Lyons et al., 1997; Ratcliffe et al., 2011; Spear et al., 2003, 2008; Rankin et al., 2013; Walsh, 2016; Walsh et al., 2020a, 2020b). Early workers considered this contact to be an unconformity (Billings, 1956; Thompson et al., 1968, 1997). Rocks of the BHA occur in a thrust sheet floored by the Monroe fault (Fig. 3), which carried the deformed section of the Oliverian Plutonic Suite, Ammonoosuc Volcanics, Partridge Formation, Clough Quartzite, and the Fitch and Littleton Formations (Walsh, 2016; Walsh et al., 2020a, 2020b). The Monroe thrust sheet placed the BHA rocks over the CVGT during an Acadian F1 nappe-stage event prior to peak metamorphism at lower-amphibolite-facies conditions. The Monroe fault in this in this area is characterized by upper- and lower-plate truncations, mylonite, and local mélange (Walsh, 2016; Walsh et al., 2020a, 2020b). Above the Monroe thrust sheet are sillimanite grade rocks of the Fall Mountain slice. Historically, these thrust sheets were interpreted as fold nappes (Thompson et al., 1968), and later as thrust nappes (Robinson et al., 1991). Rocks of the BHA and the CVGT were deformed and metamorphosed from greenschist to upper-amphibolite grade during the Devonian Acadian orogeny and to a lesser extent in the Carboniferous to Permian Alleghanian orogeny (Laird et al., 1984; Sutter et al., 1985; Harrison et al., 1989; Spear and Harrison, 1989; Spear et al., 2008; McAleer et al., 2017). Thrust sheets and early isograds are deformed and folded by an F2 doming event. Lower greenschist-facies (Acadian to Alleghanian) faults truncated peak-metamorphic assemblages, isograds, and older F1 folds and faults (Spear et al., 2008; McWilliams et al., 2013; McAleer et al., 2017). Late-stage F3 folds show preferred left-lateral rotation sense and were probably related to late dome-stage Alleghanian deformation or motion along lower-greenschist-facies faults (Walsh, 2016; Walsh et al., 2020a, 2020b). Muscovite dating by ⁴⁰Ar/³⁹Ar records Alleghanian normal faulting at ca. 245 Ma (McAleer et al., 2017, 2021). Subsequent Mesozoic brittle deformation along with the Ammonoosuc and Grantham faults, and many smaller unnamed brittle faults, had sufficient vertical or oblique-slip components to place sillimanite-grade rocks adjacent to greenschist-facies rocks and further offset the isograds (Walsh et al., 2012; McAleer et al., 2017). Apatite fission-track data support Cretaceous fault displacement and reactivation of older Paleozoic faults (Roden-Tice et al., 2009).

During the course of 1:24,000 scale geologic mapping, 59 samples from the BHA (Ammonoosuc Volcanics, Partridge Formation, and Oliverian Plutonic Suite), and 35 igneous rock samples from the CVGT (Waits River Formation, Gile Mountain Formation, and Comerford Intrusive Complex) and time-equivalent Bronson Hill cover sequence (Clough and Fitch Formations) were analyzed for major and trace element geochemistry (Valley et al., 2023; Table S1). Major and trace element geochemistry for whole-rock samples were determined by Activation Laboratories Ltd. (Actlabs), Ancaster, Ontario, Canada. Powdered samples were fused with lithium metaborate/tetraborate, and major elements were acquired by inductively coupled plasma–optical emission spectrometry (ICP-OES) and trace elements by ICP–mass spectrometry (ICP-MS). Laboratory standards were run periodically. Previous studies that are referenced in this paper report major element data, but trace element analyses vary in the elements that were measured and in the method used for obtaining the data (Aleinikoff et al., 2007; Dorais et al., 2008, 2012, 2017; Hollocher, 1993; Leo, 1991; Schumacher, 1988; Rankin et al., 2007; Table S1). Samples with loss on ignition (LOI) greater than 4% were not included in the dataset, as LOI can be used as a proxy for post-crystallization chemical alteration due to hydrothermal seafloor alteration, metamorphism, or weathering (Lechler and Desilets, 1987). In general alteration may affect the original concentration of major element oxides and highly mobile large ion lithophiles (Rollinson, 1993). Therefore, plots relying on rare earth elements (REEs), transition metals, and high field strength elements were used for most diagrams and interpretations in this paper as they are much less mobile during alteration and metamorphism (e.g., Rollinson, 1993).

The intrusive and extrusive volcanic rocks from the Ammonoosuc Volcanics, Partridge Formation, and Oliverian Plutonic Suite are shown in Figure 5. Rock types of the BHA are dominated by rhyolite/granitoids, and basalt to basaltic andesite, along with a smaller percentage of dacite, and minor trachyte, trachyandesite, and andesite (Fig. 5A). Only a few samples from this study are of intermediate (andesite) compositions. One intermediate sample (HV-1001) is from the Lebanon dome quartz diorite gneiss of the Oliverian Plutonic Suite, which is the border phase of the core Lebanon dome granite. This quartz diorite is in contact with the Partridge Formation. The other sample (A910) is an amphibolite within the Partridge Formation (Table S1). The classification of arc rock types based on K₂O versus SiO₂, confirms a bimodal grouping of basalt to basaltic andesite mafic rocks and felsic granitoids that have a wide range of K concentrations (Fig. 5B). Mafic rocks from the Ammonoosuc Volcanics and Partridge Formation have similar compositions. Mafic rocks are rare in the Oliverian Plutonic Suite and were not observed in the study area. Felsic rocks from the Ammonoosuc Volcanics and Oliverian Plutonic Suite overlap as tonalitic or trondhjemitic, but the Oliverian Plutonic Suite has a subset of high-K granites. The four granite samples with the highest K values were previously dated and all crystallized between ca. 455 Ma and ca. 445 Ma (Valley et al., 2020).

Compositional trends for seven major elements (Al₂O₃, TiO₂, Fe₂O₃, MgO, CaO, and total alkalis [Na₂O and K₂O], normalized to 100% water free) are shown in Figure 6. Some samples may have been altered regarding their major element geochemistry following crystallization, but in general, the geochemistry shows that the Ammonoosuc Volcanics, Oliverian Plutonic Suite, and Partridge Formation form a bimodal volcanic suite. Felsic rocks within the three units share similar compositional space in the geochemical diagrams in Figure 6. Some granitic rocks in the Oliverian Plutonic Suite form a subgroup with higher alkali concentrations than their Ammonoosuc Volcanic counterparts. Mafic rocks in the Ammonoosuc Volcanics and Partridge Formation show similar geochemical trends. One sample from the upper part of the Partridge Formation has a composition with 14.57 wt% MgO and elevated light REE (LREE; Table S1). The same sample has relatively high concentrations of Cr (550 ppm) and Ni (160 ppm). This is 3–8 times the concentration in the other mafic samples.

A comparison of P₂O₅ versus Zr in samples of the Ammonoosuc Volcanics and the Partridge Formation is shown in Figure 7. This diagram shows that most mafic rocks from the BHA are tholeiitic. Only two samples from the Partridge Formation plot in the alkali basalt field. One of these two samples (SP2221) has very low Al₂O₃ (8.36 wt%), and relatively high MgO compared to the rest of the mafic samples. The chemistry of this sample suggests that it is a mafic cumulate and is not appropriate for tectonic discrimination diagrams intended for basaltic liquids and will not be considered further. Trace element spider diagrams for the BHA are presented in Figure 8. Mantle-normalized basaltic samples in the Ammonoosuc Volcanics (Fig. 8A) and Partridge Formation (Fig. 8B) have similar geochemistry with some variability in the more mobile large ion lithophiles, especially Cs, Rb, and K. All samples are slightly enriched relative to the primitive mantle. There are prominent positive Pb and U anomalies and negative Nb and Ta anomalies in the majority of samples. Two Partridge samples show higher concentrations of Nb and Ta relative to other mafic samples in the BHA. Mafic samples that appear to lack a positive Pb anomaly are artificial and result from Pb concentrations being below the detection limit. Felsic samples in the Ammonoosuc Volcanics and Oliverian Plutonic Suite show nearly identical trace element patterns with pronounced positive Th and negative Nb and Ta anomalies (Figs. 8C and 8D). Three samples from the Oliverian Plutonic Suite have very high Th (~500–1000× chondrite) relative to the rest of the felsic samples. These samples correspond to samples from the Croydon dome granodiorite (453.8 + 3.4 Ma), Mascoma dome granite (450.1 ± 4.1 Ma), and Lebanon dome granite (448.0 ± 5.1 Ma) which are younger intrusive rocks relative to the span of ages in the BHA (ca. 493–435 Ma; Valley et al., 2020, and references therein).

Chondrite-normalized REE plots for the felsic Oliverian Plutonic Suite are plotted in Figure 9A. REE values typically fall between 10 and 100 times chondrite. Rare earth element patterns are slightly to moderately concave up, with all samples showing variable enrichment in LREEs and flat to slightly enriched heavy REEs (HREEs). Negative Eu anomalies are present in most of the samples but are variable in magnitude depending on rock type. Samples with the greatest LREE enrichment have the largest negative Eu anomalies. Three of these samples show a pronounced enrichment in LREE, up to 260 times chondrite. Two of these enriched samples are granite. The two granite samples are the same as the young high-Th samples mentioned in the previous paragraph on trace element concentrations: the Mascoma dome granite (450.1 ± 4.1 Ma) and the Lebanon dome granite (448.0 ± 5.1 Ma). The third high LREE sample is a tonalite of unknown age.

Felsic rocks in the Ammonoosuc Volcanics show similar REE trends to felsic rocks of the Oliverian Plutonic Suite (Fig. 9B). The REE profiles range from moderately to slightly concave up, and negative Eu anomalies vary from minor to pronounced. A few samples show enrichment in HREE suggesting garnet was incorporated into the melt. All the felsic samples from the Ammonoosuc are tonalitic or trondhjemitic. No granite samples were found within the Ammonoosuc Volcanics.

Chondrite-normalized REE plots for basaltic rocks in the Ammonoosuc Volcanics are mostly flat with one sample showing increased LREE relative to HREE and a second subset depleted in La, Ce, and Pr relative to middle and HREE (Fig. 9C). Most mafic samples lack negative Eu anomalies. REE concentrations are 10–100 times chondrite. Mafic samples from the Partridge Formation have relatively flat REE profiles, but with a subset of samples enriched in LREE relative to HREE (Fig. 9D).

Mafic samples from the Ammonoosuc Volcanics and Partridge Formation are plotted using immobile high field strength elements (Zr, Nb, Y, La) and transition metals (V, Ti) on the sequence of tectonic discrimination diagrams shown in Figures 10A–10F. This sequence of figures illustrates the fact that a single tectonic discrimination diagram may not be enough to discriminate certain groups of rocks. Figure 10A simply separates within-plate basalt (WPB) from all other basalts. All the samples from the BHA plot as “other” on this diagram, with two samples plotting very close to the WPB field. In Figure 10B, the tectonic environment cannot be determined for samples that plot in field “B.” Two samples from the Partridge Formation plot in the WPB in Figure 10B. Figures 10C and 10D are subsequently used for further separation but some of the discriminating fields still overlap. These tectonic diagrams illustrate that most basaltic samples from the Ammonoosuc Volcanics and the Partridge Formation are a mixture of both island arc tholeiites and mid-ocean ridge basalt (MORB; Figs. 10B–10E). Because of the similarities between MORB and backarc basin basalt (BAB; Fig. 10E), they are difficult to separate. Figure 10F is a ternary plot of La-Y-Nb. This diagram does separate MORB from BAB and continental basalts. The majority of our samples plot as island arc tholeiite, with minor calc-alkaline basalt, or transitional between the two. A few samples plot close to the BAB field, but the data are inconclusive and two samples from the Partridge Formation plot in the continental basalt field.

All felsic samples from the Ammonoosuc Volcanics and Oliverian Plutonic Suite are plotted on the granite discrimination diagrams shown in Figure 11. Both Rb versus Y+Nb and Ta-Yb plots show that most of these rocks plot as volcanic arc granites that trend toward syn-orogenic granites. A sub-group of Oliverian rocks plot closer to the syn-orogenic field than those of the Ammonoosuc Volcanics (Fig. 11). When rock type is taken into consideration, granite and granodiorite plot closer to the syn-orogenic field than do the trondhjemite and tonalites.

Most samples of metaigneous rocks from the CVGT are mafic. The total-alkali silica diagram shown in Figure 12 shows most of the igneous rocks in the CVGT are basalt or basaltic andesite. Mafic rocks range from 43% to 56% SiO₂ (normalized to 100% water free)_. There is a small subset of felsic rocks in the Waits River Formation and one sample in the Gile Mountain Formation. Most felsic rocks are rhyolitic with greater than 70% SiO₂ (Table S1). Only one sample, from the Waits River Formation, falls within the dacite field.

Compositional trends for seven major elements (Al₂O₃, TiO₂, Fe₂O₃, MgO, CaO, and total alkalis (Na₂O and K₂O), normalized to 100% LOI-free, are shown in Figure 13. In general, the geochemistry shows that there is considerable overlap between mafic rock samples from the Clough, Fitch, Waits River, and Gile Mountain Formations, and the Comerford Intrusive Complex. The greatest geochemical variations are associated with the Waits River Formation. Three felsic rock samples within the Waits River and Gile Mountain Formations share similar compositional space apart from the one Waits River Formation dacite sample.

Most basaltic samples from the CVGT are tholeiitic, but there is a small group of samples in the Waits River Formation and one sample in the Gile Mountain Formation that plot as alkali basalt or transitional between tholeiitic and alkali basalt (Fig. 14). Primitive-mantle normalized multi-element diagrams for selected elements are presented in Figure 15. Most basaltic samples have similar trace element characteristics between formations. Some variability in the mobile large ion lithophiles is present throughout and positive U, Th, and Pb anomalies, and negative Ta and Nb anomalies are common. The Waits River Formation shows two distinct trends of samples with both positive and negative Nb and Ta anomalies, and only one sample from the Gile Mountain Formation (Fig. 15D) shows a similar positive Nb and Ta anomaly. Samples from the Comerford Intrusive Complex are more consistent in their geochemistry profile and similar to many of the other CVGT samples, with negative Nb and Ta anomalies and positive Pb anomaly (Fig. 15E). Felsic samples in the CVGT share similar geochemical characteristics with negative Nb and Ta anomalies and positive Th, La, and Zr anomalies (Fig. 15F).

Chondrite-normalized REE plots for mafic rocks in the CVGT are broadly similar and have flat REE profiles (Figs. 16A–16D). The Waits River Formation shows two distinct geochemical populations, one of which is flat and the other with increased LREE concentrations (Fig. 16B). The increase in LREE is consistent with a within-plate source for the basalt with an alkali basalt composition as shown in Figure 13. Most mafic rock samples in the CVGT are 10–100 times chondrite and have similar REE profiles and concentrations to MORB and enriched MORB (EMORB; Fig. 16F). REE element profiles for felsic rocks in the CVGT are shown in Figure 16E. Samples are concave up and enriched in LREE. Three of four samples have little to no negative Eu anomaly with a slight enrichment in HREE. The dacitic sample from the Waits River Formation has a pronounced negative Eu anomaly and the highest concentrations of REE (Fig. 16E).

Figures 17 and 18 are tectonic discrimination diagrams for the CVGT rocks. Mafic rock geochemistry indicates that these rocks formed in two different tectonic settings. The samples are split between WPB and MORB/BAB or mixed MORB and volcanic arc basalt (Figs. 16A and 16C). All the WPB are in the Waits River Formation except for one in the Gile Mountain Formation. These WPB correspond to the samples with positive Nb and Ta anomalies and LREE enrichment (Figs. 16B and 16C). Similarly, only the Waits River and Gile Mountain Formations contained felsic rocks in the CVGT. Three of the felsic samples plot as volcanic arc granites and one as a within-plate granite (Fig. 18).

The data presented in the Results section indicate that the BHA evolved from tholeiitic basalts and island arc tonalite and trondhjemite magmas to more calc-alkaline or alkali basalts and granites (Figs. 10 and 11). Basaltic rocks from the Ammonoosuc Volcanics and Partridge Formation are chemically indistinguishable with only a few minor exceptions (Figs. 5–8). Basaltic magmas are dominantly tholeiitic with only one sample from the top of the Partridge Formation being alkaline. The data indicate that these are dominantly island arc tholeiite (IAT) and minor MORB or BAB (Fig. 10). This is consistent with prior work in the area (Aleinikoff, 1977; Leo, 1985, 1991; Schumacher, 1988; Hollocher, 1993; Dorais et al., 2012). Some of the previous research concluded that samples with MORB-like geochemistry were likely BAB. This was based on the presence of bimodal volcanism, variable Ti/V, and the observation that some samples were transitional between island arc and ocean floor basalts (Hollocher, 1993; Dorais et al., 2012). Figure 10F is one of the few basalt discrimination diagrams that separate BAB from the rest of the basalts using La-Y-Nb. A few samples plot in or near the BAB field on the La-Y-Nb diagram, and two samples plot within the continental basalt field, but the majority are in the IAT field. Additionally, the negative Ta and Nb anomalies and positive Pb anomalies observed in many Ammonoosuc Volcanics and Partridge Formation samples (Figs. 8A and 8B), are consistent with volcanic arc magmas that have interacted with subduction zone fluids in an island arc setting (Briqueu et al., 1984; Pearce and Peate, 1995). An island arc origin is further supported by field relationships. Our mapping shows that most contacts between the plutons of the Oliverian Plutonic Suite, and the Ammonoosuc Volcanics and Partridge Formation are intrusive, not tectonic (Valley et al., 2020; Walsh et al., 2020a, 2020b). If the plutons represent the arc, which we think they do, then it is not possible for the majority of Ammonoosuc Volcanics and Partridge Formation basalts to have their origin in a backarc basin without being tectonically transported. Therefore, we conclude, based on field relations and geochemistry, that the majority of mafic samples from the Partridge Formation and Ammonoosuc Volcanics were generated in an island arc setting.

Previous workers concluded that felsic rocks from the Ammonoosuc Volcanics are different than the Oliverian Plutonic Suite based on variable K content, a volcanic arc versus a more continental geochemical signature, and Pb isotopes which suggested different crustal sources between these units (Hollocher, 1993; Hollocher et al., 2002; Moench and Aleinikoff, 2003; Aleinikoff et al., 2007; Dorais et al., 2008, 2012). Jacobi and Mitchell (2018) suggested this change in composition was due to the BHA overriding an aseismic ridge causing melting in the overriding plate due to flat-slab subduction. Felsic intrusive rocks from the Ammonoosuc Volcanics were thought to be mostly trondhjemites and tonalites while the magmas from the Oliverian Plutonic Suite were distinctly more granitic. The apparent difference in coeval composition of the Oliverian and Ammonoosuc rocks was based in part on a supposed time gap between older trondhjemitic and tonalitic magmas and the younger granites (Moench and Aleinikoff, 2003). We now know this time gap does not exist (Fig. 4; Rankin et al., 2013; Valley et al., 2020). Our research shows that felsic rocks from both the Oliverian Plutonic Suite and Ammonoosuc Volcanics are trondhjemitic and tonalitic in composition, with the Oliverian Plutonic Suite having a subset of high-K granites (Fig. 5B). Felsic samples from both overlap chemically, and consistently plot as island arc magmas that evolve toward the syn-collisional field on tectonic discrimination diagrams (Fig. 11). Pearce and Peate (1995) showed that in the granite discrimination diagram of Rb versus Y+Nb that volcanic arc granites that plot close to, or trend toward the syn-collisional granite field, are the result of arc continent collision. In Figures 4 and 11, there is a distinct shift from trondhjemite to granite as the arc evolves.

When the same samples that were used for geochemistry are correlated with existing U-Pb zircon ages (Valley et al., 2020, and references within), we can observe that most BHA granites are younger than 455 Ma, while the majority of trondhjemitic and tonalitic rocks are older (500–455 Ma), regardless of whether they belong to the Ammonoosuc Volcanics or Oliverian Plutonic Suite (Figs. 11 and 18). This trend is present in both the BHA and the Shelburne Falls arc (SFA; Fig. 4). We attribute this geochemical transition from trondhjemite to granite and the transition of tholeiite to calc-alkaline/alkali basalt in mafic rocks to the arrival of the arc to the Laurentian margin or delivery of Laurentian continental material to the subduction zone. This conclusion is supported by Pb, Nd, and Sr isotope ratios which show a mixed continental (Laurentian) and island arc source (peri-Gondwanan) for the Killingworth complex (ca. 459–456 Ma), Highlandcroft Plutonic Suite (ca. 452 Ma), and Quimby Formation (ca. 443 Ma; Aleinikoff et al., 2007; Dorais et al., 2008). REE patterns are a result of minerals that were present during partial melting and crystal fractionation, the source region, and the tectonic setting. Two granite samples show the greatest degree of LREE enrichment when compared to their trondhjemitic counterparts. These two samples are the Mascoma and Lebanon dome granites and are dated at ca. 450 Ma and ca. 448 Ma, respectively. The LREE enrichment is consistent with greater continental input as the BHA collided with Laurentia. A plot of Sr versus Y shows that the felsic rocks of Oliverian Plutonic Suite have two samples with high Sr (Fig. 19A). Yttrium can be subdivided into two groups of high and low Y. Both Y groups have samples from the Ammonoosuc Volcanics and Oliverian Plutonic Suite (Fig. 19A). The two samples with the highest Sr concentrations are the Croydon dome granodiorite (ca. 455 Ma) and the Mascoma dome granite (ca. 450 Ma, mentioned in the Results section; Valley et al., 2020). Hollocher et al. (2002) identified similar high-Sr rocks in the Fourmile Gneiss (ca. 451 Ma) and Monson Gneiss (ca. 442 Ma; included in Figs. 19A and 19B). However, the source of Sr was attributed to the magmatic source region with the high-Sr magmas being from a deeper source than the low-Sr magmas. When Sr/Y ratios are plotted against K, we see that higher Sr ratios are positively correlated with higher K concentrations (Fig. 19B). This correlation provides additional evidence for continental contamination of younger BHA magmas by the proximity of the arc to Laurentia. When the following evidence is considered together, it is clear that the BHA represents an evolving arc that collided with Laurentia ca. 450 Ma: (1) the transition from trondhjemite to granite, (2) the switch from tholeiitic to calc-alkaline basalt, (3) an increase in LREE in both felsic rocks and alkali basaltic samples, (4) the increase in Rb, and (5) the increase in Laurentian derived Pb isotopes and high-Sr concentrations in younger granitic rocks.

A modern-day equivalent to the idea that continental material can change the chemistry of island arc lavas is in the Lesser Antilles. The southern Lesser Antilles have a distinct continental signature. This is due to the Oronoco River delivering Amazonian sediment into the trench, while lavas in the northern Lesser Antilles have a volcanic arc signature (e.g., Burke 1988).

Geochemical results from the CVGT (Waits River Formation, Gile Mountain Formation) and related Silurian rocks (Clough Quartzite, Fitch Formation) show that most samples are tholeiitic basalt with a subset of alkali basalt in the Waits River Formation (Fig. 14) and rare felsic rocks (Fig. 12). The tectonic discrimination diagrams in Figure 16 show that the tholeiitic rocks in the CVGT are a mix between IAT and MORB or BAB (Fig. 17C). This mixed signature is supported by the variability in Nb-Ta anomalies. The subset of alkali basalts in the Waits River Formation plot as within-plate magmas. All the WPB are from the Waits River Formation except for one sample in the Gile Mountain Formation. When samples from the Frontenac Formation (Waits River Formation equivalent), which are geochemically transitional between WPB and the mixed IAT-MORB magmas, are compared to our samples, we see a trend in the Waits River Formation from continental to WPB (Fig. 17C). The Frontenac Formation samples are tholeiitic but plot close to the line discriminating alkaline from tholeiitic basalts (see Dorais et al., 2017). Of the four felsic samples, three are in the Waits River Formation and one is in the Gile Mountain Formation. These four samples plot as volcanic arc granite and one Waits River Formation sample plots as a within-plate granite (Fig. 18).

Ce/Yb and La/Ta ratios of basaltic rocks may infer the source and depth of melting. The Ce/Yb ratio is controlled by the presence of garnet in the source region and the partition coefficients of Ce (<1) and Yb (>1) in garnet (McKenzie and O'Nions, 1991; Johnson, 1994; Gaetani et al., 2003). This means that partial melting of garnet peridotite will have a higher Ce/Yb ratio than spinel peridotite (Ellam, 1992). The La/Ta ratio of basalt can distinguish between the magma originating in the asthenosphere or the lithosphere (Fitton et al., 1988; Leat et al., 1988; Thompson and Morrison, 1988). Figure 19C is a plot of Ce/Yb versus La/Ta of CVGT samples (Dorais et al., 2017). The alkaline, WPB of the Waits River Formation and tholeiitic WPB of Frontenac Formation show that they both originate from a spinel peridotite source in the asthenosphere, while the strictly tholeiitic MORB/volcanic arc samples throughout the CVGT have variable La/Ta values that are nevertheless all consistent with a magma source within the crust or lithospheric mantle. In Figure 19C, the samples from the Frontenac Formation form a trend from asthenospheric melts toward samples which contain a degree of lithospheric contamination.

The presence of both asthenospheric and lithospheric melts in the Waits River Formation (and Frontenac Formation) may be the result of slab rollback and slab breakoff after the Taconic orogeny during the Salinic disturbance (e.g., Rankin et al., 2007; Perrot et al., 2018). The production of felsic magmas is controlled by crustal anatexis from underplating by asthenosphere melts. The mixed island-arc/MORB or BAB lithospheric melts in the CVGT may be the result of the melting of existing rock types that form the basement under the CVGT, either through host rock assimilation or partial melting of the BHA-SFA rocks, the leading edge of Laurentia, or both, by rising asthenospheric magmas. It has been reported that the tectonic setting of felsic magmas may be inherited from the source rock and may not represent the tectonic setting of the sampled rock (Förster et al., 1997). This may explain the observed volcanic arc granite signature of the felsic rocks in the CVGT (Fig. 18) and the variable La/Ta values which suggest lithospheric contamination in most mafic rocks throughout the CVGT (Fig. 19C). The LREE/HREE ratios show that most of these magmas have a variable but a significant lithospheric component.

To better understand the importance of only finding bimodal, within-plate asthenospheric magmas in the Waits River and Frontenac Formations, we need to understand the timing of events in the CVGT. The age of fossils from the CVGT range from late Silurian (Pridoli) to Early Devonian (Emsian; Boucot and Drapeau, 1968; Hueber et al., 1990; Lavoie and Asselin, 2004). Volcanic and metasedimentary rocks within the CVGT have U-Pb zircon ages indicating maximum depositional ages between ca. 434 ± 4 Ma and 407 ± 3 Ma (Aleinikoff and Karabinos, 1990; Rankin and Tucker, 2009; McWilliams et al., 2010; Dorais et al., 2017; Perrot et al., 2018; Converse et al., 2022). Primary U-Pb zircon ages are rare and include the following in descending stratigraphic order:

• 407 ± 3 Ma (thermal ionization mass spectrometer [TIMS]): Metarhyolite in the Littleton Formation (Rankin and Tucker, 2009)

• 407 ± 3 Ma (TIMS): Metarhyolite in the Meetinghouse Slate Member of the Gile Mountain Formation (Rankin and Tucker, 2009)

• 409 ± 5 Ma (sensitive high-resolution ion microprobe [SHRIMP]): Volcanic zircons in quartzite of the Gile Mountain Formation (McWilliams et al., 2010)

• 416 ± 4 Ma (laser ablation ICP-MS [ LA-ICP-MS]): East Inlet Pluton and felsic volcanic rock in the Frontenac Formation (Converse et al., 2022)

• 423 ± 4 Ma (TIMS): Metafelsite in the Waits River Formation (Aleinikoff and Karabinos, 1990; Hueber et al., 1990)

• 424 ± 4 Ma (LA-ICP-MS): Biotite metagranite sill in the Frontenac Formation (Converse et al., 2022)

• 432 ± 8 Ma (LA-ICP-MS): Metafelsite in the Frontenac Formation in New Hampshire (Dorais et al., 2017)

Rocks of the late Silurian Lake Memphremagog Intrusive Suite cut the pre-Silurian rocks on both the west and east sides of the CVGT and yield somewhat overlapping U-Pb zircon ages with the dated volcanic rocks of the CVGT (Ratcliffe et al., 2011). Age determination of rocks located just outside the CVGT include the following:

West of the CVGT:

• 418 ± 1 Ma (TIMS): Felsic dike near Bridgewater, Vermont (Aleinikoff and Karabinos, 1990)

• 419 ± 0.39 Ma (TIMS): Metadiorite of the Braintree Intrusive Complex (Black et al., 2004)

• 421 ± 7Ma (SHRIMP): Granite in the Braintree Intrusive Complex (Aleinikoff et al., 2011)

• 425 ± 3Ma (SHRIMP): Trondhjemite of the Newport Intrusive Complex (Aleinikoff et al., 2011)

• East of the CVGT:

• 419.8 ± 2.6 Ma (TIMS): Comerford Intrusive Complex at Comerford quarry (Rankin et al., 2007)

• 419.3 ± 1.3 Ma (TIMS): Comerford Intrusive Complex at Leighton Hill (Rankin et al., 2007)

• 418.5 ± 2.0 Ma (TIMS): Comerford Intrusive Complex at Peaked Mountain (Rankin et al., 2007)

Dated samples from the margins of the CVGT are limited from ca. 425 Ma to ca. 418 Ma while samples within the CVGT are 423 Ma or older (see Dorais et al., 2017) and ca. 409 Ma or younger. However, age data are sparse, and mostly from felsic lithologies, whereas the majority of geochemistry samples are basaltic. The dated mafic samples are limited to the Braintree and Comerford Intrusive Complexes. Neither of these complexes intrudes the CVGT. However, undated basaltic samples within the BHA Silurian cover sequence and CVGT, geochemically overlap with the Comerford Intrusive Complex (Figs. 12–17). We interpret these samples as being related. The Early Devonian ages from the Gile Mountain Formation (detrital) and it's Meetinghouse Slate Member (metarhyolite) and the Devonian Littleton Formation (metarhyolite), coincide with age of the Piscataquis magmatic belt and the onset of the Acadian orogeny (Bradley, 1983; Bradley et al., 1998; Bradley and Tucker, 2002) and mark the end of basin formation (McWilliams et al., 2010).

The BHA evolved from trondhjemitic island arc magmas to mixed calc-alkaline island arc/syn-collisional granites. Geochemically the basalts are dominantly tholeiitic volcanic arc basalt with a few BAB and two alkali basalts that have a continental signature. When examined in the context of existing geochronology and geochemistry, we attribute this geochemical fingerprint as evidence of an island arc, built on peri-Gondwanan crust, that incorporated more continental material into the magma as it approached and collided with the Laurentian margin. In modern arc environments, the contribution of subducted sediments can have a significant impact on the isotopic compositions of generated melts (e.g., Labanieh et al., 2012). In the BHA, magmas are dominantly tonalitic or trondhjemitic from >475 Ma to ca. 455 Ma, after which there is a significant shift in magma composition to calc-alkaline granites (Figs. 4 and 11). This shift coincides with the approximate time of terminal crustal thickening (455–445 Ma, during which the plutons are more evolved (granites and granodiorites). At ca. 445 Ma a small number of mafic intrusions appear, in addition to continued granite magmatism. The mafic plutons may be related to slab delamination or sublithospheric tears and mantle upwelling (Hames et al., 1991; Ratcliffe et al., 2012; Hildebrand and Whalen, 2021; Valley et al., 2020) instead of island arc or continental arc magmatism. Evidence of slab breakoff <445 Ma is supported by the presence of alkali WPB samples from the top of the Partridge Formation (Fig. 10), the Frontenac Formation, and in the Waits River Formation near the bottom of the CVGT (Fig. 17). Figure 19C shows that melt for these same WPB samples originated in the asthenosphere and have little evidence of lithospheric contamination.

The transition from Taconic arc magmatism (BHA) to basin formation (CVGT) shows a significant overlap in geochemistry. In addition to the Partridge and Waits River Formations having the subset of mafic, within-plate asthenospheric rocks, the felsic rocks in the Waits River Formation plot as syn-collisional or within-plate granites (Fig. 18), which are similar geochemically to the granites in the BHA. There are other minor felsic rocks at the top of the CVGT, but these are likely related to the onset of the Acadian orogeny (Rankin and Tucker 2009; McWilliams et al., 2010). Mafic rocks in the CVGT are dominantly tholeiitic basalt that shows significant crustal or lithospheric contamination with a mixed volcanic arc/MORB signature (Figs. 17 and 19C). We interpret these mixed geochemical signatures to result in part from the contamination of CVGT magmas by incorporation of BHA/SFA or Laurentian material during extension and the formation of the CVGT above or on the rocks of the arc, and thus inheriting its geochemistry from the arc rocks as CVGT magmas.

The CVGT evolved from an extensional tectonic setting after the Ordovician Taconic orogeny and Silurian Salinic disturbance to a foreland basin setting during the Devonian Acadian orogeny (Bradley et al., 1998; Black et al., 2004; Tremblay and Pinet, 2005, 2016; McWilliams et al., 2010). The development and location of the CVGT has been assigned to various extensional regimes. Study of the Comerford Intrusive Complex, which intruded the Ammonoosuc Volcanics and other pre-Silurian rocks (Albee Formation) near the CVGT, showed considerable overlap between within-plate, ocean floor, and backarc basin basalts (Rankin et al., 2007). Trace element ratios suggest that the source region was mostly asthenospheric mantle, with a minor contribution from the lithospheric mantle and no crustal contamination. This conclusion was based on the field evidence of crustal extension and the Ce/Yb ratios of dikes in the Comerford Intrusive Complex. However, when the Ce/Yb ratios are plotted against La/Ta in Figure 19C, the majority of samples show evidence of a lithospheric or a crustal component in the magma. Rankin et al. (2007) concluded that magmatism and CVGT basin formation were probably due to hinge migration and slab delamination, but proposed a second tentative model where the CVGT and its volcanics were the result of left lateral, strike-slip faulting and pull-apart basins development.

Study of the Frontenac Formation (Waits Field Formation equivalent) showed that metabasalts and diabases in the unit are mostly within-plate tholeiites. Ce/Yb ratios from the Frontenac Formation suggests that the magma source was the asthenospheric mantle (Dorais et al., 2017). Further analysis of La/Ta, Ti/Yb, and K/P ratios indicates there was little contamination from the lithospheric mantle or the crust. Dorais et al. (2017) concluded that slab rollback may have been the mechanism for basalt magma formation, but the younger within-plate Comerford Intrusive Complex may have been the result of slab breakoff. Felsic magma production was the result of crustal anatexis due to the influx of mafic magma (Dorais et al., 2017).

Slab delamination or slab failure at the end of the Taconic orogeny has also been suggested by Hildebrand and Whalen (2021). Their study uses trace elements from rocks of the BHA to suggest that magmatism younger than ca. 450 Ma was the result of slab breakoff. The timing of slab breakoff in their research coincides with the switch to granitic magmatism. Hildebrand and Whalen (2021) attribute this change in magmatism to Laurentian cratonic underplating. We have modified three figures from Hildebrand and Whalen (2017, 2021) and Hildebrand et al. (2018) that separate arc magmas from melts that are the result of slab breakoff (or failure). Figure 20A is a plot of Nb/Y against Gd/Yb. Only one of our samples, the ca. 450 Ma Mascoma dome granite, plots in the slab failure field. The ca. 448 Ma Lebanon dome granite plots directly on the discrimination line. Figure 20B, a plot of La/Yb versus Gd/Yb, shows similar results, with the Mascoma and Lebanon dome granites both plotting in the slab failure field. Figure 20C is the Pearce plot of Ta versus Yb with the discrimination field for slab failure from Hildebrand et al. (2018) drawn in. In this figure, three samples from the Oliverian Plutonic Suite plot in the slab failure field. One sample is the ca. 454 Ma Croydon dome granodiorite, with the other two samples being the same Mascoma and Lebanon dome granites. The rest of the data in Figure 20C do not make sense if slab breakoff or failure happened ca. 450 Ma as suggested by Hildebrand and Whalen (2021). At least three presumably older island arc trondhjemitic samples plot in the slab failure field, while our youngest sample, from the ca. 447 Ma Alstead dome, plots in the island arc field. The results of these diagrams are at best inconclusive and in the case of Figure 21C may not work. It is possible that slab breakoff occurred ca. 450 Ma based on the Mascoma and Lebanon dome samples, but this is at odds with the geochemical, structural, and metamorphic data presented previously and in this study, which suggests slab breakoff is younger than 445 Ma. Furthermore, the vast majority of the data plot in the arc field and not the slab failure field (Fig. 20), as defined by Hildebrand and Whalen (2017, 2021).

The results of these previous studies, when taken in combination with the new research presented here, indicate that slab breakoff was at the end of the Taconic orogeny, or at the onset of CVGT basin formation. Slab breakoff allowed upwelling of the asthenosphere under the BHA-CVGT, which produced the geochemical results reported. This is the best way to explain asthenospheric melts at the top of the Partridge Formation and the bottom of the CVGT and mixed lithosphere/asthenosphere geochemical fingerprint throughout the CVGT (Figure 19C). High asthenospheric heat flow resulted in the assimilation of existing volcanic arc crust by asthenospheric magmas or partial melting of the overriding plate between ca. 434 Ma and 419 Ma, followed by a 10-m.y. hiatus before the onset of Acadian magmatism and deformation. Figure 21 is a series of cross sections which show the evolution of the BHA and subsequent CVGT basin formation and magmatism. Prior to 455 Ma, the BHA was built on Ganderian crust above an east dipping subduction zone producing predominantly trondhjemitic and tonalitic magmas and tholeiitic basalt. Between 455 Ma and 445 Ma, there is a shift in magma composition from trondhjemitic to granitic felsic magmas as the BHA incorporated increasing amounts of Laurentian-derived sediment and collided with Laurentia. Circa 445–440 Ma, slab breakoff may have been initiated as evidenced by the presence of one WPB sample at the top of the Partridge Formation. This is followed by erosion of the BHA-SFA and deposition of the Clough Quartzite and Fitch Formation which are intruded by ongoing magmatism related to slab breakoff. And asthenospheric upwelling. By 425 Ma, extension, and deposition of sediments in the CVGT and Central Maine basin was taking place in an intra-arc or backarc basin setting. Basin formation coincided with continued magmatism before transitioning to a foreland basin at the onset of the Acadian orogeny ca. 409 Ma.

Whole-rock geochemical data show that the Ordovician Bronson Hill arc evolved from tholeiitic basalts and island arc tonalites and trondhjemites, to calc-alkaline basalts and granites. Most mafic and felsic samples in the Ammonoosuc Volcanics, Partridge Formation, and Oliverian Plutonic Suite are geochemically consistent with being from a single volcanic arc that became more evolved through the Ordovician as the Bronson Hill arc approached Laurentia and eventually collided. Subsequently, in the Silurian, bimodal alkali and tholeiitic within-plate magmas and MORB/volcanic arc rocks were generated in an extensional basin and deposited with marine carbonate and clastic rocks during formation of the Connecticut Valley–Gaspé trough. The geochemical signature of the Connecticut Valley–Gaspé trough is consistent with a post-collisional backarc, or intra-arc basin formation where slab breakoff played a key role in the Salinic disturbance before transitioning to a foreland basin at the beginning of the Acadian orogeny.

We thank Mike Dorais and Nick Ratcliffe for helpful discussions. This manuscript benefited from U.S. Geological Survey reviews by Ryan Deasy, Mark Carter, Alan Pitts, and John Counts. We thank the associate editor, Robert Hildebrand, and Mike Dorais for constructive review comments for this journal. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. government.