Constraints on the Palaeoproterozoic tectono‐metamorphic evolution of the Lewisian Gneiss Complex, NW Scotland: Implications for Nuna assembly

Despite extensive investigation, the tectono‐thermal evolution of the Archean crust in the Lewisian Gneiss Complex in NW Scotland (LGC) is debated. Most U–Pb zircon geochronological and metamorphic studies have focused on rocks from the central region of the mainland LGC, where granulite facies assemblages associated with the oldest (Badcallian) tectono‐metamorphic event at c. 2.75 Ga are overprinted by younger amphibolite facies assemblages related to the Inverian (c. 2.5 Ga) and subsequent Laxfordian (c. 1.9–1.65 Ga) tectono‐thermal events. In the southern and northern regions of the mainland LGC, deformation and metamorphism associated with the Laxfordian event are pervasive, although the timing and conditions are poorly constrained. Here, we present new field, petrographic and structural data, U–Pb zircon and titanite geochronology and phase equilibrium modelling of amphibolite samples from the northern and southern regions. Our field observations show that in both regions, pre‐Laxfordian structures are significantly reworked by steep NW‐striking fabrics that are themselves pervasively overprinted by co‐axial deformation and amphibolite facies metamorphism related to the Laxfordian event. In situ U–Pb titanite geochronology yields Laxfordian ages of 1853 ± 20 Ma in the southern region (P = 6–8 kbar and T = 640–690°C) and 1750 ± 20 Ma and 1776 ± 10 Ma in the northern region (P = 6–7.5 kbar and T = 740–760°C). While U–Pb dating of zircon rims from felsic gneisses in the central region shows a dominant Inverian metamorphic overprint at c. 2500 Ma, zircon rims in felsic gneisses from the northern and southern regions commonly yield Laxfordian dates as young as c. 1800 Ma. Combined, the results support the idea that, during the Palaeoproterozoic, the central region of the LGC acted as low‐strain domain, in which intense deformation and metamorphism were restricted to crustal‐scale shear zones. By contrast, in the southern and northern regions, early (c. 1.85 Ga) and late (c. 1.75 Ga) Laxfordian deformation and fluid‐mediated metamorphism were much more pervasive and at higher P–T conditions than previously proposed. The diachronous Laxfordian evolution of the southern and northern regions indicate that they reflect early and late snapshots of collisional to transpressional tectonics in the mainland LGC. The long‐lasting Laxfordian evolution documents the collision of the Rae and North Atlantic cratons during the Palaeoproterozoic amalgamation of the supercontinent Nuna, with implications for the palaeogeographic configuration of NW Scotland during Palaeoproterozoic Nuna.

conditions than previously proposed.The diachronous Laxfordian evolution of the southern and northern regions indicate that they reflect early and late snapshots of collisional to transpressional tectonics in the mainland LGC.The long-lasting Laxfordian evolution documents the collision of the Rae and North Atlantic cratons during the Palaeoproterozoic amalgamation of the supercontinent Nuna, with implications for the palaeogeographic configuration of NW Scotland during Palaeoproterozoic Nuna.
K E Y W O R D S in situ titanite petrochronology, Laxfordian metamorphism, North Atlantic Craton, strain partitioning, supercontinent assembly

| INTRODUCTION
The Lewisian Gneiss Complex in NW Scotland (LGC; Figure 1) is a segment of Archean crust that was variably reworked during the Palaeoproterozoic amalgamation of the supercontinent Nuna (Park et al., 2005; e.g., Goodenough et al., 2013).The LGC crops out on the NW Scottish mainland and on the islands of the Outer Hebrides (Figure 1).In most palaeogeographic reconstructions, the LGC is located between Baltica and the Rae-North Greenland blocks as part of the Palaeoproterozoic Lapland-Kola and Nagssugtoqidian orogenic belts (Buchan et al., 2000;Park, 1995).However, the timing and the tectono-thermal evolution of the mainland LGC during Nuna amalgamation remain poorly constrained.
Like most other examples where deeper levels of Archean continental crust are exposed, the LGC is dominated by magmatic protoliths of the tonalitetrondhjemite-granodiorite (TTG) series and records polyphase magmatism, deformation and metamorphism as the result of several tectono-metamorphic events (Goodenough et al., 2010(Goodenough et al., , 2013;;Sutton & Watson, 1950).The mainland LGC traditionally has a tripartite subdivision (Sutton & Watson, 1950), comprising a granulite facies central region that is separated from the amphibolite facies northern and southern regions by crustal-scale F I G U R E 1 Simplified geological map of the Lewisian Gneiss Complex (LGC), NW Scotland.The central region records dominant granulite facies metamorphism, whereas the northern and southern regions record amphibolite facies conditions.This map is derived from British Geological Survey © UKRI.All rights reserved.
The advent of in situ U-Pb spot analysis of zircon led researchers to propose a model for the LGC that departed from the simple tripartite subdivision, which envisaged the mainland LGC to be composed of up to six crustal blocks that amalgamated during the Laxfordian event (e.g., Friend & Kinny, 2001;Kinny et al., 2005).This terrane model is variably accepted, as well as the Laxfordian and the Gairloch Shear Zones (Figure 1) being major tectonic boundaries (Goodenough et al., 2010(Goodenough et al., , 2013;;Park, 2005), although debates about the number and position of terranes, their tectonic boundaries and their timing of amalgamation remain open (e.g., Fischer et al., 2021;Guice et al., 2022).In the granulite facies central region of the mainland LGC, amphibolite facies retrogression associated with the Inverian and Laxfordian overprinting stages is concentrated along crustal-scale shear zones (e.g., Corfu et al., 1994;Goodenough et al., 2013;Zirkler et al., 2012).Nevertheless, a continuous spread of apparent concordant U-Pb zircon ages from 3.1-2.8Ga down to 2.5 Ga commonly recorded by TTGs from this region has been challenging to interpret (MacDonald et al., 2015;Whitehouse & Kemp, 2010).Some propose that Badcallian and Inverian metamorphism in the central region record the onset and termination of a single protracted hightemperature event (Taylor et al., 2020), whereas others suggest them to be discrete metamorphic episodes (Fischer et al., 2021;MacDonald et al., 2015).
In contrast to the central region, the tectono-thermal evolution of the northern and southern regions during the Palaeoproterozoic has received less attention and its timing and tectono-thermal evolution remain poorly constrained.In this contribution, we combine existing data with new field, petrographic and structural observations, mineral chemistry of key mineral phases, in situ U-Pb titanite and zircon geochronology and phase equilibrium modelling to better constrain the tectono-metamorphic evolution of the mainland LGC and its significance within the Palaeoproterozoic amalgamation of the supercontinent Nuna (or Columbia).

| REGIONAL GEOLOGY
The LGC (Figure 1) consists of voluminous TTG (grey) gneisses with metamorphosed mafic-ultramafic rocks and subordinate garnet-biotite schists of possible sedimentary origin (e.g., Wheeler et al., 2010).In the original model of Sutton and Watson (1950), the granulite facies central region of the mainland LGC was interpreted to represent the lower crustal levels of a once contiguous fragment of Archean crust, whereas the amphibolite facies southern and northern regions were interpreted as shallower (mid-crustal) levels (Park & Tarney, 1987).Geochronological campaign-type studies involving U-Pb geochronology of zircon from TTG gneisses proposed a terrane model in which the mainland LGC is divided into six crustal blocks including the Rhiconich Terrane in the northern region, the Assynt and Gruinard terranes in the central region and the Rona, Gairloch and Ialltaig terranes in the southern region (Figure 1).These terranes were defined on varying crystallization and metamorphic ages and they were interpreted to have amalgamated by accretion and stacking during the Palaeoproterozoic Laxfordian event (Friend & Kinny, 2001;Goodenough et al., 2010;Kinny et al., 2005;Love et al., 2010;Park et al., 2005).Between the Inverian and Laxfordian tectono-metamorphic events, the LGC was exposed to a period of extension (Figure 2).Swarms of mafic dykes (i.e., Scourie Dykes) were emplaced into the Archean crust during major and minor magmatic events at c. 2.4-2.2Ma and c. 2.0-1.9Ma (Figure 2), respectively (Baker et al., 2019;Davies & Heaman, 2014;Heaman & Tarney, 1989;Waters et al., 1990).This extensional stage was associated with the deposition of a Palaeoproterozoic volcano-sedimentary sequence, the Loch Maree Group (2.2-1.9Ga; Park et al., 2001;Whitehouse et al., 1997) in the southern region of the mainland LGC and the Leverburgh metasediments in the Outer Hebrides (Cliff et al., 1983(Cliff et al., , 1998)).Active convergence was suggested to occur after 2 Ga, and by 1.91 Ga, part of the continental margin was buried, producing the first metamorphic components in the Harris Granulite Belt (Mason, 2015).The Loch Maree Group accreted as a mélange during ongoing subduction of the continental margin in the early Laxfordian (c.1.9 Ga) synchronously with arc magmatism reflected by granitic intrusions such as the Ard Gneiss in the mainland LGC and metadioritic and anorthositic rocks in the Harris Granulite Belt (e.g., Mason, 2015).During this stage, local granulite facies metamorphism was recorded by the Ialltaig gneiss at c. 1.88 Ga in the southern region of the mainland LGC (Love et al., 2010;Mason, 2015).After 1.87 Ga, subduction terminated with the collision of continental blocks during which some crustal segments were buried at high pressure (HP) conditions typical of collisional-type metamorphism (e.g., Harris Granulite Belt; Cliff et al., 1998;Hollis et al., 2006), while others remained at higher crustal levels (e.g., Loch Maree Group) and escaped HP metamorphism (Mason, 2015).The late-Laxfordian phase between 1.85 Ga and 1.65 Ga was characterized by ductile deformation, collisional-type metamorphism, and strike-slip shearing accommodating the post-collisional stage (e.g., Mason, 2015).During this prolonged Laxfordian evolution, granitic sheets and potassic pegmatitic dykes were emplaced (Figure 2) along the major tectonic boundaries such as the LSZ and across the northern and southern regions of the mainland LGC (e.g., Goodenough et al., 2013;Wheeler et al., 2010).
The Laxfordian event was first dated using Rb-Sr and K-Ar geochronology on biotite and hornblende in gneisses from the central region at c. 1.85 Ga to 1.75 Ga (Lambert & Holland, 1972).Deformed granitic intrusions yielded a LA-ICPMS U-Pb zircon age of c. 1.88 Ga (Goodenough et al., 2013) and a SHRIMP U-Pb zircon age of c. 1.85 Ga (Friend & Kinny, 2001), whereas undeformed pegmatites yielded U-Pb ID-TIMS ages as young as c. 1.77 Ga (Goodenough et al., 2013).Additionally, U-Pb geochronology of titanite in biotite-hornblende gneisses southwest of the Laxford bridge yield dates of c. 1.75 Ga (Corfu et al., 1994), coeval with titanite dates from felsic gneisses within the LSZ (Goodenough et al., 2013).In the southern region of the mainland LGC, deformation and metamorphism associated with the latest Laxfordian stages occurred at c. 1.67 Ga in the Rona Terrane (Love et al., 2010).
The Laxfordian deformation (D3) within the central region increases in intensity northwards, becoming pervasive in the southern part of the LSZ (e.g., Coward, 1990), but is localized in centimetre-to metre-wide high-strain zones in the core of the central region (Figures 1 and 3).In these deformed domains, the S2 fabric is refolded, and the Scourie Dykes are foliated (e.g., Goodenough et al., 2010).The steep NW-striking Laxfordian D3 fabrics were previously inferred to be associated with a dextral shear component (see also, e.g., Coward & Park, 1987;Park et al., 1987) and have recorded amphibolite facies metamorphism (e.g., Goodenough et al., 2013).
In the northern region (Figure 4), shallow pre-Laxfordian fabrics have been previously interpreted as potential Badcallian structural relicts (e.g., Beach et al., 1974;Coward & Park, 1987).Beach et al. (1974) F I G U R E 3 Simplified geological map of the central region (Assynt and Gruinard terranes) with structural measurements, stereographic diagrams (equal area, lower hemisphere projections) and foliation trajectories of the interpreted Badcallian (red), Inverian (orange) and Laxfordian (green) fabrics.Dashed grey foliation trajectory and associated measurements for the pre-Laxfordian fabrics are shown in the upper-right side of the map.The map is derived from 50k scale BGS Digital Data under Licence No. 2021/018 British Geological Survey © UKRI.All rights reserved.
described Laxfordian folds as metre-scale, open structures that deform the Scourie Dykes and the gneissic layering in the TTG gneisses and that are associated with syn-to post-Laxfordian pegmatite and microgranite sheets and dykes oriented parallel to S3.
In the southern region (Figure 5), heterogeneously reworked zones of Archean crust that contain mafic lenses, boudins, and intrafolial folds have been interpreted as Badcallian structural relicts preserved within Inverian/ Laxfordian low-strain domains that were rotated into the dominant, steep NW-striking Inverian to Laxfordian foliations in the Torridon area (e.g., Coward, 1974;Wheeler et al., 1987).Nonetheless, the southern region is dominated by steep NW-striking Laxfordian fabrics (e.g., Coward, 1974;Wheeler et al., 1987) that are associated with a SE-plunging lineation (Cresswell, 1973).The Gairloch Shear Zone (Figure 5) has been interpreted to have formed late during the Laxfordian event through folding and dextral strike-slip shearing related to the development of a re-oriented NW-plunging linear fabric (Park, 2010).In the Gairloch area, a Laxfordian pervasive foliation and a dominant SE-plunging lineation developed during thickening at amphibolite facies conditions (Park, 2010;Wheeler et al., 2010).In the Torridon area, the Diabaig Shear Zone (Figure 5) was regarded by Beach (1976) to record a dominant simple shear component associated with the Inverian event, whereas Wheeler et al. (1987) and Wheeler (2007) considered the Inverian and the Laxfordian deformation to be coaxial, forming dominant NW-SE-striking shear zones and a mineral lineation that plunges at moderate angles towards the SE.Wheeler et al. (1987) suggested that the Inverian structures controlled the orientation of emplacement of the Scourie Dykes, which, in turn, controlled fluid circulation and strain localization during the Laxfordian (Attfield, 1987;Cartwright, 1989;Coward & Park, 1987).

| New field, structural and petrographic observations
In this study, petrographic and structural (i.e., petrostructural after Volante, Collins, et al., 2020) investigations were carried out on felsic and mafic lithologies across the central (Figure 3), northern (Figure 4) and southern (Figure 5) regions (Table 1) to better constrain the Laxfordian tectonometamorphic evolution of the mainland LGC.Three felsic gneisses (samples LC1903, LC1803 and LC1930) were collected from high-strain zones within the central region where Badcallian fabrics (Figure 6a,b) are pervasively reworked by Inverian and/or Laxfordian deformation (Figure 3).In the northern region, three mafic (amphibolerich) gneisses (samples LC1909, LG2039 and LG2031) and two felsic gneisses (samples LC1914 and LC1808) were collected (Figures 1 and 6c-h).The amphibolites contain leucocratic veinlets and stretched pockets that are aligned parallel to, and/or have been folded and deformed parallel to, the dominant Laxfordian foliation (Figure 6d,e).Sample LC1909 was collected from a beach near Durness (Figure 4), where pervasively foliated, subvertical up to 2 m thick amphibolitic layers alternate with felsic gneisses and pegmatite sheets.Millimetre-thick leucocratic veinlets occur parallel to the dominant Laxfordian foliation.Sample LG2039 is a garnet-free amphibolite collected from near the main road (Figure 4), where amphibolitic lenses occur within felsic gneisses.Sample LG2031 is an amphibolitic lens hosted by TTG gneisses at Sandwood beach that exhibits millimetre-thick small leucocratic veins (Figure 6d).At this locality, pre-Laxfordian structures in the amphibolitic and felsic gneiss layers were refolded and reworked during the Laxfordian deformation event (Figures 4 and 6c).Felsic TTG gneisses from the northern region (LC1914 and LC1808) are foliated biotite-bearing TTG gneisses (bt-TTG) collected north and south of Kinlochbervie (Figure 4).In the southern region, one garnet-bearing amphibolite (LC1923) was collected from a Laxfordian high-strain zone, where a dense swarm of Scourie Dykes and their host felsic gneisses are intensely deformed.At this locality, centimetre-to millimetre-thick leucocratic layers and veinlets are elongated or tightly folded along the dominant axial planar fabric (Figure 6f).Samples LC1948 and LC1926 are foliated bt-TTG gneisses collected a short distance north of the Laxfordian Gairloch Shear Zone (Figure 5).A complete list of the investigated samples can be found in Table 1.

| Central region
Lower hemisphere equal area pole-to-plane projection shows the Badcallian shallow W-dipping gneissic layering (Figure 6a) and associated shallowly to moderately W-plunging lineation (Figure 3).The latter is defined by hornblende or biotite in TTG gneisses or by amphibolitic to hornblenditic pods that exhibit a top-to-the-west sense of shear (Figure 6b).Within the felsic gneisses, F1 folds overprint an already differentiated foliation (pre/early D1?) that formed under high-temperature conditions (Attfield, 1987;Sheraton et al., 1973).The high-strain and complex nature of this D1 pervasive gneissic foliation is evidenced by parallel, stretched, thinned and boudinaged abundant mafic lenses and pods as well as subparallel, stretched and/or boudinaged syn-to late-kinematic leucocratic veins (Figure 6a).Considering that the pre/early D1 foliation and S1 both developed during syn-kinematic and continuous melt segregation (Johnson & T A B L E 1 List of samples investigated in this study.2010) in the Mineralogy column, except for white mica (wm).White, 2011;Zirkler et al., 2012;e.g., Feisel et al., 2018), the folded, differentiated fabric could be either related to an earlier deformation event or, more likely, reflects early stages of the D1 event (early-S1) during which progressive folding and fabric development occurred under hightemperature conditions (e.g., Volante, Collins, et al., 2020).Therefore, the simple nature of the 'Badcallian' layering may obscure a complex and progressive deformation history that is commonly homogenized and obliterated under granulite facies conditions.The Inverian overprint is commonly characterized by coronitic textures of amphibolite facies mineral phases partially replacing the granulite facies assemblages.Closer to tectonitic and mylonitic higher strain domains, such as the CSZ and the LSZ (Figure 3), Badcallian structures are refolded by open-to-tight upright folds, thinned, steepened and transposed into a SW-dipping, WNWstriking axial planar fabric defined by amphibolite facies mineral assemblages and associated with a SE-plunging mineral lineation (Figure 3; LC1803, LC1903, LC1930).The Laxfordian fabrics were distinguished from pre-Laxfordian structures where superimposed Badcallian, Inverian and Laxfordian deformation structures were identified or, more often, where mafic Scourie Dykes were present.In the central region, pervasive development of Laxfordian fabrics in the mafic dykes and felsic gneisses have been commonly observed within highstrain zones related to the development of a moderate to steep NW-ESE-striking foliation (S3) associated with a moderately SE-plunging lineation (Figure 3).

| Northern region
In the southernmost part of the northern region (i.e., north of the LSZ), TTG gneisses record a pervasive, NNE-striking and shallowly SE-dipping gneissic layering (Figure 4) and hornblenditic to amphibolitic lenses are elongated and aligned parallel to it.This gneissic layering has a similar strike and shallow dip as the dominant Badcallian gneissic fabric in the adjacent central region (Figure 4b), but no granulite facies mineral assemblages were observed.These field observations suggest that this shallow fabric could either be a structural relict of the retrogressed Badcallian foliation recorded in the central region or another pre-Laxfordian foliation that is unrelated to the fabrics preserved in the central region.
North and south of Loch Inchard between Rhiconich and Kinlochbervie (Figure 4) gneissic rocks record a dominant WNW-striking and moderately SW-to NEdipping axial planar fabric (S2), which is folded by asymmetric D3 open folds (F3; Figure 4b; samples LC1914 and LC1808).This folded fabric in the TTG gneisses is crosscut by irregular, undeformed, Kfs-rich and coarse-grained to pegmatitic sheets and veins (Goodenough et al., 2013).These granitic to pegmatitic sheets and veins also crosscut foliated amphibolites that were previously interpreted as deformed Scourie Dykes (Goodenough et al., 2010(Goodenough et al., , 2013;;Kinny & Friend, 1997).Hence, this folded fabric is here interpreted as Inverian, and the undeformed granitic pegmatite sheets were interpreted to intrude relatively late during the Laxfordian event (Goodenough et al., 2013).The Inverian strain rate increases towards the LSZ, where the related foliation becomes intensely reworked during the Laxfordian (i.e., sample LC1803).On the west coast at Sandwood beach (Figure 4a), TTG and mafic gneisses are folded by NW-striking open to tight, upright to inclined folds (Figure 6c).The mafic gneisses are metre-wide Scourie Dykes that are stretched, locally boudinaged and folded together with the hosting TTG gneisses.Both lithologies record a pervasive NWstriking Laxfordian axial planar foliation (S3), moderately to steeply SW-dipping (Figure 4b).Locally, amphibolitic lenses exhibit centimetre-thick, leucocratic lenses and pockets that are interpreted as leucosomes (Figure 6d).Centimetre-to metre-thick, undeformed, coarse-grained and Kfs-rich veins and dykes are either developed parallel to the dominant axial planar foliation (S3) or at an angle to it.The dominant generation of folds and the associated axial planar fabric at Sandwood beach are ascribed as Laxfordian structures, superimposed on the Inverian fabrics (Figure 6c).Near Durness (Figure 4a), mafic layers (i.e., LC1909) are elongated and boudinaged parallel to the dominant, steeply NE-dipping axial planar fabric.Metre-thick amphibolite layers interlayered and wrapped by the TTG gneisses are possibly Scourie Dykes deformed during the Laxfordian.

| Southern region
In the southern region, the widespread presence of mafic Scourie Dykes enable the discrimination of Laxfordian and older structures (Figure 5).In the Gairloch area, Laxfordian fabrics are characterized by a moderate to subvertical NW-SE-striking foliation that is pervasively recorded in the Scourie Dykes (Figure 5; sample LC1923).In the Torridon/Diabaig areas, the Laxfordian deformation heterogeneously reworked zones where older relict fabrics are preserved (Figures 5 and 6g).NNE-striking and shallowly ESE-dipping fabrics are identified in the Meall Ceann na Creige and the An Ruadh-Mheallan low-strain domains (SE and ENE of Diabaig, highlighted in blue in Figure 5) and have been previously interpreted as Badcallian structural relicts re-oriented during the Inverian-Laxfordian shearing (Figure 6h; e.g., Coward, 1974;Wheeler et al., 1987).In the northernmost part of the Diabaig area (Figure 5), the dominant foliation is consistent with NW-SE-striking Laxfordian structures previously described by Wheeler et al. (1987).This foliation dips steeply to the NE and it is associated with a NW-plunging lineation (Figure 5).Previously documented SE plunging lineation in this region (Cresswell, 1973) could possibly represent part of a set of doubly plunging linear structures or be the result of several superimposed deformation events.

| Electron microprobe analysis
Major element compositions of pyroxene, amphibole, garnet, biotite, plagioclase, titanite and ilmenite (Table S1) in selected amphibolitic samples (Figure 7) were measured using a Cameca SX-50 EPMA fitted with four wavelength-dispersive spectrometers at the Institute of Geology, Mineralogy and Geophysics at Ruhr-Universität Bochum.The operational conditions for spot analyses were 15 kV accelerating voltage and 8 nA beam current.The analytical spot diameter was set at $5 μm (or focused), keeping the same current conditions.Onand off-peak counting times were 20 s, and dead-time corrections were applied to all analyses.Natural and synthetic mineral standards were used for calibration.Structural formulae of mineral phases were calculated stoichiometrically based on the following number of oxygen: 6 for pyroxene, 12 for garnet, 8 for feldspar, 11 for biotite, 23 for amphibole, 5 for titanite and 3 for ilmenite.Mineral formulae of amphibole were calculated following Locock (2014).Fe 3+ was recalculated for biotite based on charge balance and for garnet was calculated following Droop (1987).Mineral abbreviations used throughout the text, tables, and figures follow those suggested by Whitney and Evans (2010).A complete list of the acquired analysis can be found in Table S1.For sample LC1909, X-ray compositional maps of garnet (Figure S1) for major elements were also produced by measuring a grid of spots over the grains, with 15 kV accelerating voltage, 55 nA beam current and 35 ms dwell time.Elemental concentrations in garnet were determined for Na, Si, Ti, Mn, K, Mg, Al, Cr, Fe and Ca.

| Whole-rock analysis
The whole-rock major element data used for the phase equilibrium calculations were determined at Actlabs (Ontario, Canada) using X-Ray fluorescence (XRF) analysis and titration for ferrous iron contents.For sample preparation, the amphibolitic samples were dried, crushed and subsequently pulverized in a vibrating disc pulveriser.Samples were cast using a 66:34 flux with 4% lithium nitrate added to form a glass bead, which was then analysed for Al 2 O 3 , As, BaO, CaO, Cl, Co, Cr 2 O 3 , Cu, Fe 2 O 3 , K 2 O, MgO, MnO, Na 2 O, Ni, P 2 O 5 , Pb, SiO 2 , SO 3 , SrO, TiO 2 , V 2 O 5 and Zn via XF spectrometry on oven-dry (105 C) samples unless otherwise stated.Losson-ignition (LOI) values were determined using a robotic thermogravimetric analysis (TGA) system.Furnaces in the system were set to 110 and 1000 C. Bulk-rock chemical data are reported in Table 2.

| U-Pb zircon geochronology
U-Pb geochronology was performed on the zircon rims in seven samples, including four biotite-bearing felsic gneisses from the southern and northern regions and three hornblende-bearing felsic gneisses from the central region.All samples were processed at the mineral separation facility at the Ruhr-University Bochum by standard jaw crushing, milling and Wilfley table to concentrate mineral constituents.Zircon fractions were separated using a Frantz for magnetic separation and lithium heteropolytungstate (LST) heavy liquids.Zircon grains were mounted on a double-sided tape in epoxy resin and polished to reveal the approximate centre of the zircon grains.The resin discs were carbon-coated for cathodoluminescence (CL) imaging on a Tescan Mira3 variable pressure field emission gun scanning electron microscope (VP-FEG-SEM) at the John de Laeter Centre (JdLC) at Curtin University, Perth, Western Australia.U-Pb analyses in zircon rims were acquired using a LA-ICP-MS on a RESolution/M50 with a 193 nm wavelength Ar-F excimer laser coupled to an Agilent 8900 triple quadrupole ICPMS at the GeoHistory Facility, JdLC.Sample energy density of $3 J cm À2 and a repetition rate of 5 Hz were used to sputter the target zircon for 25 seconds of analysis time and 45 s of total baseline capture.All analyses were preceded by two cleaning pulses and undertaken at a spot diameter of 23 μm.The sample cell was flushed by ultrahigh purity He (320 mL min À1 ) and N 2 (1.2 mL min À1 ).U-Pb analyses of unknowns were bracketed by OG1 (n = 17; 3465.4 ± 0.6 Ma; Stern et al., 2009) and 91500 (n = 21; 1062.4 ± 0.4 Ma; Wiedenbeck et al., 1995) as the primary reference materials used for U-Pb geochronology to monitor and correct for mass fractionation and instrumental drift, and secondary zircon reference materials GJ1 (n = 22; 601.86 ± 0.37 Ma; Jackson et al., 2004), Plesovice (n = 17; 337.13 ± 0.37 Ma;Sl ama et al., 2008) and Maniitsoq (n = 21; 3008.9 ± 0.7 Ma; Marsh et al., 2019) are used to monitor data accuracy and precision.The time-resolved mass spectra were reduced using the U_Pb_Geochronology4 data reduction scheme in Iolite (Paton et al., 2011).During U-Pb data reduction, OG1 and 91500 were used as the primary reference materials to calculate 207 Pb/ 206 Pb and 206 Pb/ 238 U ratios, respectively, dates and their respective 2σ propagated uncertainties.Within the text and figures, individual 207 Pb/ 206 Pb dates of the analysed samples are reported with 1s uncertainty as recalculated using IsoplotR (Vermeesch, 2018).During the analytical sessions, the secondary reference materials yielded statistically reliable dates compared to the published values (see Table S2).Discordancy depends on the Concordia distance, and it was calculated by using IsoplotR (Vermeesch, 2018).The latter was also used to generate Concordia Wetherill diagrams for visualization of the acquired data (Figure 8).S3 and summarized here.U-Pb analyses of unknowns were bracketed by GJ-1 zircon (n = 28; 601.86 ± 0.37 Ma; Jackson et al., 2004) as the primary reference material and secondary titanite reference materials BLR-1 (n = 14; 1047.1 ± 0.4 Ma; Aleinikoff et al., 2007) and NAMA1 (n = 14; 1023 ± 2 Ma; in-house reference material Namaqualand, personal communication from Wolfgang Dörr).Mass fractionation induced by differences in matrix matching between zircon and titanite were corrected using BLR-1, whereas NAMA1 was used for validation.Data were processed using an in-house Visual-Basic MS Excel © spreadsheet program (Gerdes & Zeh, 2006, 2009).The matrix offset for 206 Pb/ 238 U between GJ-1 zircon and titanite was calculated as 7.5%.The results are shown in Figure 9, where they are plotted on a Tera-Wasserburg diagram with the associated lower intercept age for each analysed sample.All the diagrams were plotted using IsoplotR (Vermeesch, 2018).The uncertainty of individual dates is reported as 95% CI.For full methods, including uncertainty propagation, see Table S3.No common Pb correction is assumed for pooled dates; instead, the initial Pb composition is determined as the upper intercept of the regression line for common-Pb uncorrected U-Pb titanite data.Individual spot dates in Table S3 are calculated using the 208 Pbbased common Pb correction method.BLR-1 and NAMA1 titanite yielded common-Pb corrected weighted mean dates of 1049.7 ± 3.5 Ma (n = 14, MSWD = 1) and 1018.9 ± 5.8 Ma (n = 14, MSWD = 1.56), respectively, consistent within uncertainty of the reported ages.
Trace element analyses were collected adjacent to U-Pb ablation pits in a second analytical session.For internal standardization, 29Si (EPMA values were used, Table S1) was utilized and analytical uncertainties were propagated using the internal standard with the software LADR (Norris & Danyushevsky, 2018).Analyses of unknowns were bracketed by those of NIST610 as a primary reference material and NIST612 and GSE-1G as secondary reference materials (Pearce et al., 1997;Table S4;Jochum et al., 2005Jochum et al., , 2011)).Spot sizes of 50, 33 and 23 μm were used for NIST612, NIST610 and GSE-1G and for unknowns, respectively (Table S3) with repetition rate of 6 Hz, and sample energy density of $3 J cm À2 .A complete list of titanite U-Pb and trace element analyses for reference material and unknowns is reported in Table S3.

| Phase equilibrium modelling
All calculations were performed using THERMOCALC v3.50beta1 (Powell & Holland, 1988, update December 2019) with an updated version of the internally consistent data set ds62 of Holland and Powell (2011), created on 6 February 2012.Calculations were undertaken in the , which offers the most realistic investigation of phase equilibria in mafic rocks at present.The following activity-composition (a-x) relations were used: tonalitic melt (L), augitic (aug) clinopyroxene, clinoamphibole (hb; Green et al., 2016), garnet (g), orthopyroxene (opx), biotite (bi), ilmenite-hematite (ilmmhemm), chlorite (chl) and muscovite (mu; White et al., 2014); olivine (ol) and epidote (ep; Holland & Powell, 2011); plagioclase (plc-pli) and K-feldspar (ksp; Holland & Powell, 2003) and magnetite-spinel (mt, sp; White et al., 2002).Pure phases included quartz (q), albite (ab), titanite (sph), rutile (ru) and aqueous fluid (H 2 O).Bulk-rock compositions used for calculations were obtained by X-ray fluorescence (XRF) analysis.The CaO contents of these bulk compositions were adjusted according to the measured P 2 O 5 contents (i.e., Ca correction = À33 mol of P 2 O 5 ) to account for the presence of apatite, which was observed to be the sole P-bearing phase in all samples.The ratio of ferrous to ferric iron in each sample was determined by standard titration methods.Field and microscale evidence for incipient partial melting of the modelled amphibolitic samples was observed (Figures 6 and 7).For this reason, samples were modelled at suprasolidus conditions, and a H 2 O content that just saturated the solidus was used to stabilize a high proportion of hornblende as observed in the modelled samples.The normalized molar bulk compositions used for phase equilibrium modelling are given in Table 2.Although the Green et al. (2016) model for the mafic system does not include MnO in amphibole and clinopyroxene, to stabilize garnet in garnetbearing amphibolite samples (LC1923 and LC1909), MnO was added to the NCKFMASHTO system.Therefore, Mnbearing phases from the metapelitic system (White et al., 2014) were used to replace the following Mn-free phases from the mafic system: orthopyroxene (opx), garnet (g), ilmenite (ilmm), biotite (bi) and chlorite (chl).
Samples LC1914 and LC1808 are felsic, foliated biotite-bearing TTG gneisses (bt-TTG) that contain plagioclase, quartz, biotite, minor hornblende, K-feldspar, titanite and accessory minerals such as apatite and zircon.Sample LC1914 exhibits a spaced foliation defined by dark brown biotite and minor dark green hornblende associated with elongate to prismatic titanite grains and round-shaped apatite.Anhedral to subhedral quartz occurs as internally deformed grains and recrystallized strain-free crystals, respectively.Plagioclase is partially replaced by sericite.Sample LC1808 exhibits a foliation defined by biotite, epidote and ilmenite exsolutions along the dominant cleavage alternating with quartz and plagioclase.Anhedral quartz grains show undulous extinction and subgrain boundaries.Quartz also occurs as round-shaped inclusions in plagioclase and as symplectitic intergrowths with biotite and plagioclase.Locally, quartz films around plagioclase are observed.Minor microcline shares straight contacts with quartz and plagioclase.

| Southern region
The garnet-bearing amphibolite sample LC1923 consists of hornblende, plagioclase, quartz, garnet, titanite, ilmenite and epidote.The foliation is defined by compositional alternation of amphibole-and plagioclase-rich layers.Subhedral amphibole crystals with granoblastic texture commonly form triple junctions and have straight contacts with titanite, plagioclase and quartz.Euhedral to rounded titanite is commonly included in amphibole crystals (Figure 7e) and is partially overgrown by symplectitic ilmenite and quartz aggregates.Titanite in this sample can exhibit granular textures with euhedral grains forming titanite aggregates.Sub-centimetre sized garnet porphyroblasts are partially replaced by a fine-grained assemblage of plagioclase-epidote-green amphibole (Figure 7f).Quartz has an anhedral shape, commonly occurs as strings or films along grain boundaries and exhibits vermicular texture when hosted by amphibole (Figure 7d).Plagioclase is subhedral and less abundant compared to quartz.Samples LC1948 and LC1926 are foliated bt-TTG gneisses with the foliation defined by dark biotite alternated with plagioclase-and quartz-rich layers.These bt-TTG gneisses consist of plagioclase, quartz, biotite, K-feldspar, minor hornblende and Fe-Ti oxide and contain accessory phases such as apatite and zircon.In sample LC1948 minor white mica is present.

| Northern region
Bt-TTG samples LC1808 and LC1914 contain euhedral and prismatic zircon grains that range between 100 and 500 μm (Figure 8a,b).These zircon grains contain cores with oscillatory zoning typical of magmatic zircon (Corfu et al., 2003), usually an inner CL-dark grey rim and un-zoned outer, very thin, CL-bright rim.Sixteen and 13 rim analyses were acquired in samples LC1808 and LC1914, respectively.In sample LC1808, 9 out of 16 analyses are concordant or near-concordant (≤5% discordance) and show a spread of dates between c. 2700 Ma and c. 1800 Ma (Figure 8a).The two oldest 207 Pb/ 206 Pb dates of 2712 ± 14 Ma and 2701 ± 13 Ma were measured in darker-CL inner rim with oscillatory zoning and are characterized by Th/U ratios of 1.45 and 1.75, respectively (Table S2).Three analyses, reflecting mixed domains within darker and brighter CL rims, yield 207 Pb/ 206 Pb dates of 2588 ± 12 Ma, 2522 ± 12 Ma and 2560 ± 13 Ma with Th/U ratios of 0.06, 0.13 and 0.01, respectively.The remaining four, CL-bright rim analyses yield dates between c. 2000 Ma and c. 1800 Ma (1958 ± 29 Ma, 1878 ± 14 Ma, 1809 ± 25 Ma and 1805 ± 12 Ma) and Th/U ratios between 0.01 and 0.05 (Table S2).For sample LC1914, only three analyses were concordant or near-concordant.For two analyses of very thin CL-bright rims, a significant portion of a darker CL domain was also ablated, resulting in potentially mixed analysis with 207 Pb/ 206 Pb dates of 2550 ± 8 Ma and 2528 ± 18 Ma and Th/U ratios of 0.25 and 0.29 (Figure 8b).The third concordant date is located on a thicker, CL-bright zircon rim which more confidently yields a rim 207 Pb/ 206 Pb date of 1828 ± 70 Ma with Th/U ratio of 0.01.

| Southern region
Zircon in bt-TTG samples LC1948 and LC1926 from the southern region is euhedral, with grain sizes ranging between 70 and 400 μm.The shape of the grains is variable, from subrounded to acicular and prismatic.In CL image, they often exhibit a core with oscillatory zoning, and when present, rims are bright-CL either very thin (LC1926) or up to 30 μm thick (LC1948).For samples LC1948 and LC1926, 20 and 14 analyses were acquired, respectively.In sample LC1948, eight analyses yield concordant 207 Pb/ 206 Pb dates (Figure 8f).Of those, three yield 207 Pb/ 206 Pb dates older than 2700 Ma (2788 ± 14 Ma, 2728 ± 17 Ma and 2713 ± 28 Ma with Th/U ratio of 0.99, 0.84 and 0.88, respectively), two CL-bright rims yield dates of 2693 ± 21 Ma and 2688 ± 22 Ma with Th/U ratios of 0.93 and 0.86, whereas other three thick CL-bright rims yield dates at 2508 ± 26 Ma, 2435 ± 35 Ma and 2424 ± 40 Ma (Figure 8f) and have Th/U ratios of 0.87, 0.56 and 0.3, respectively.For sample LC1926, only five analyses yield concordant dates (Figure 8d).The oldest 207 Pb/ 206 Pb date of 2776 ± 15 Ma was obtained from a round-shaped zircon grain with a CL-bright rim and Th/U ratio of 0.72.Two slightly younger 207 Pb/ 206 Pb dates of 2667 ± 12 Ma and 2639 ± 20 Ma were obtained from a CL-bright zircon rim and a mixed analysis that ablated part of a CL-bright rim and of CLdarker inner rim.Those two analyses yield Th/U ratios of 0.2 and 1.7.A CL-bright zircon rim yielded a 207 Pb/ 206 Pb date of 2518 ± 25 Ma and Th/U of 0.59, whereas the youngest CL-bright rim yielded a 207 Pb/ 206 Pb date of 2130 ± 35 Ma and Th/U of 0.1 (Figure 8g).
A summary of the zircon dates for all three regions is shown in a Kernel Density estimation diagram, where the dates are colour-coded by region (Figure 8h).Zircon with darker grey, oscillatory zoning domains from the central region record the oldest dates between c. 2850 and 2700 Ma, commonly with often high Th/U ratios (>0.5) which have been previously suggested to be characteristic of magmatic zircon and/or zircon grown at UHT conditions (Kirkland et al., 2015).From these older ages, a smear of apparent concordant dates is formed in all three samples down to c. 2400 Ma (an age interpreted to reflect the Inverian metamorphic event).Dates younger than c. 2400 Ma are not recorded in the central region.By contrast, samples from the northern region record dates with low Th/U ratios (<0.1) as young as 1800 Ma, whereas zircon rims from the southern region are rare and only sample LC1926 yields dates as young as c. 2000 Ma (Figure 8h).

| Northern region
In sample LC1909 titanite grains are between 100 and 500 μm in size and BSE images show a homogeneous internal texture.However, subordinate, thin, BSE-bright rims are observed in association with thin BSE-brightfilms that permeate through the titanite grains (Figure 9a).BSE-bright rims and linear features cutting grain cores were avoided during analysis.A total of 45 analyses were acquired, and one younger outlier was excluded (Table S3).The remaining 44 analyses yield an intercept age of 1750 ± 18 Ma, with an initial 207 Pb/ 206 Pb = 0.769 ± 0.0.069 and MSWD = 3.7 (Figure 9a).Titanite grains from sample LC2039 are between 100 and 300 μm.Titanite crystals are euhedral to subhedral with a round or ellipsoidal shape (Figure 9b).BSE images show a darker core surrounded by a very thin, BSE-bright annulus which in turn is surrounded by a BSE-dark rim (Figure 9b).For this sample, 55 analyses (Table S3) yield an intercept age of 1776 ± 10 Ma with initial 207 Pb/ 206 Pb = 1.076 ± 0.054 and MSWD = 0.92 (Figure 9b).While cores and rims were systematically targeted, uncertainties in the isotopic ratios are sufficiently large such that the timing of specific growth events cannot be resolved (Figure 9b).

| Southern region
In sample LC1923, titanite grains are from 50 to 200 μm and form aggregates of up to 600 μm in size (Figure 9c,d).Titanite is euhedral, commonly forms granular textured assemblages (multi-grain) and is present as inclusion in amphibole.Titanite is often associated with vermicular quartz and locally is partially replaced by skeletal to vermicular ilmenite intergrown with quartz (Figure 9d).BSE images of titanite show a homogeneous internal texture (Figure 9c,d).A total of 32 analyses yield an intercept age of 1853 ± 20 Ma with initial 207 Pb/ 206 Pb = 1.0012 ± 0.0159 and MSWD = 4 (Figure 9c).

| Trace elements in titanite
Trace element analyses show that titanite in the garnetbearing amphibolite from the southern region (LC1923) is relatively depleted in REE compared to the two samples from the northern region (Figure 10a).Sample LC1923 exhibits an arch-shaped REE pattern with a light-REE (LREE) positive slope reflecting LREE depletion, followed by a relatively flat MREE and a negative heavy-REE (HREE) slope (Figure 10a).In contrast, samples from the northern region show REE patterns with a negative slope, with the garnet-free amphibolite (LC2039) being slightly more LREE-enriched than the garnet-bearing amphibolite sample (LC1909).Titanite from the garnet-bearing amphibolite LC1909 exhibits a weak positive Eu anomaly, whereas sample LG2039 shows a weak negative Eu anomaly (Figure 10a).
The zirconium-in-titanite (Zr-in-titanite) content ranges between 92 and 355 ppm in sample LC1909, 120 and 574 ppm in sample LG2039 and 19 and 85 ppm in sample LG1923 (Figure 10b).Titanite in the garnetbearing amphibolite LC1909 (northern region) displays Y contents between $238 ppm and 993 ppm, with a median value of 392 ppm (Figure 10c,d).By contrast, titanite in the garnet-free amphibolite (LG2039) shows a spread in Y contents from 350 ppm to 793 ppm with a median value of 677 ppm (Figure 10c,d).This spread reflects core, and rim domains identified in the BSE images of titanite grains (Figures 7d and 9).Titanite grains from sample LC1923 display an inverse correlation between Y and Sr contents (Figure 10c).The relatively low Y versus Al contents (Figure 10d) reflect the depletion of HREE in titanite in the presence of garnet and amphibole.

| Phase equilibrium modelling and thermobarometry
Pseudosections were constructed for three amphibolite samples from the northern region (LG2031, LG2039 and LC1909) and one (LC1923) from the southern region (Figure 11a-d).In the P-T pseudosection for sample LC1909 (Figure 11a), the calculated solidus is located at 625-710 C. The stability field of the Laxfordian peak assemblage in the presence of melt (amphibole + garnet + plagioclase + quartz + titanite + augite + biotite + ilmenite) occurs immediately above the solidus, between $680 C and $745 C and $4-8.6 kbar.This stability field is bounded by the loss of ilmenite and biotite towards higher pressures and temperatures, as well as by the loss of quartz at lower temperatures and pressures.The lower temperatures boundary is defined by the introduction of epidote and the hydrated solidus.The calculated melt modes for this sample range between 5 and 9 mol.%, which is consistent with the presence of quartz films and temperature estimates obtained from the Zrin-titanite thermometry (725-770 C at 5-9 kbar; Hayden et al., 2008;Figure 12; Table S3), suggesting incipient partial melting at suprasolidus conditions.Additionally, the Ti in amphibole thermometer (Otten, 1984) yields median temperatures of $795 ± 30 C, which are somewhat higher than the ones predicted by phase equilibrium calculations.Amphibole and plagioclase thermometry also yields a range in temperatures between 750 C and 780 C (Holland & Blundy, 1994).Additionally, the Al in amphibole barometer yields pressures of 7.6 ± 0.3 kbar (Hollister et al., 1987) and 7.1 ± 0.2 kbar (Hammarstrom & Zen, 1986), in agreement with the amphibole-plagioclase barometer that estimates pressures between 6 and 7 kbar (Molina et al., 2015).
In the P-T pseudosection for sample LG2039 (Figure 11b), the calculated solidus is located at 630 C to 710 C, depending on the pressure.The stability field of the Laxfordian peak assemblage in the presence of melt (amphibole + plagioclase + quartz + titanite + augite + biotite) extends from $640 C to $760 C and $3.3-9.7 kbar.It is bounded by the ilmenite-in field at lower pressures and higher temperatures.The epidote-in and plagioclase-out fields confine the peak assemblage at higher pressures, whereas the hydrated solidus defines the lower temperatures boundary.The calculated melt modes for this sample range between 1 and 3 mol.%,which is consistent with the microstructural observations and temperature estimates obtained from the Zrin-titanite thermometry (735-785 C at 5-9 kbar; Hayden et al., 2008;Figure 12; Table S3) suggesting incipient partial melting and high-temperature conditions, respectively.Additionally, Ti in amphibole (Otten, 1984) and amphibole-plagioclase (Holland & Blundy, 1994) thermometers yield median temperatures of $716 ± 41 C and a range between 700 C to 850 C, which are within the range of temperatures predicted by the phase equilibrium modelling.Also, Al in amphibole barometer yields pressures of 6.5 ± 1.3 kbar (Hollister et al., 1987) and 6.1 ± 1.0 kbar (Hammarstrom & Zen, 1986), which agree with the values obtained by the amphibole-plagioclase barometer between 5 and 7 kbar (Molina et al., 2015).
In the P-T pseudosection of sample LG2031 (Figure 11c), the calculated solidus is located at 625-710 C. The stability field of the inferred peak assemblage defining the dominant foliation at suprasolidus conditions (amphibole + plagioclase + quartz + ilmenite F I G U R E 1 1 Calculated P-T pseudosections for (a-c) three samples from the northern region and (d) one sample from the southern region.The stability fields of the metamorphic peak assemblage defining the Laxfordian fabrics is highlighted in red.
+ augite + biotite + melt) is bounded by the introduction of titanite towards higher pressures, the loss of quartz and biotite at higher temperatures and by the solidus at lower temperatures.The range of temperatures and pressures for the peak assemblage is between 650 and 760 C and 3.8-10 kbar, respectively.The calculated melt modes for this sample range between 10 and 14 mol.%, which are somewhat higher than the melt mode estimates based on field and microscopic observations.The Ti in amphibole thermometer (Otten, 1984) yields a median temperature of $815 ± 28 C, higher than that predicted by the phase equilibrium calculation, but towards the higher end of the peak assemblage P-T field.
For the garnet-bearing amphibolite sample LC1923 from the southern region (Figure 11d), the calculated solidus is located at 640-740 C, depending on pressure.The stability field of the peak assemblage defining the dominant Laxfordian foliation (melt + amphibole + quartz + garnet + plagioclase + titanite ± biotite) is predicted to be between $640-690 C and $5.9-7.75 kbar (Figure 11d).It is confined by the introduction of ilmenite at lower pressures and augite at higher pressures and temperatures.The lower temperature boundary is F I G U R E 1 2 Synoptic P-T diagrams for (a-c) three samples from the northern region and (d) one sample from the southern region with representative microscale textures.The P-T stability field of the Laxfordian metamorphic peak assemblage is reflected by the (a-c) purple and (d) green filled polygons.Additional empirical thermobarometers are used to refine the P and T conditions for the investigated samples.The red arrow in (a) and (c) starts from the lowest temperature end, increasing to the right, estimated using Ti-in-amphibole thermometer.
constrained by the solidus.The calculated melt modes for this sample range between 1 and 2 mol.%, which agree with both microstructural observations and temperature estimates obtained using Zr-in-titanite thermometry (655-680 C at 6-8 kbar; Hayden et al., 2008;Figure 12; Table S3).

| DISCUSSION
5.1 | Meso-to Neoarchean Badcallian event in the LGC Our field, petrographic, and structural observations in the mainland LGC indicate the preservation of a shallowly E-dipping fabric in the southernmost part of the northern region that was refolded by pervasive Inverian (D2) and/or Laxfordian (D3) overprinting structures (Figure 4).In the Rhiconich area (Figure 4), one of the investigated bt-TTGs (LC1808) contains two zircon grains with oscillatory zoning yielding 207 Pb/ 206 Pb dates of 2712 ± 14 Ma and 2701 ± 13 Ma and high Th/U ratios (Figure 8a).These ages are within the range identified as the Badcallian event in the adjacent granulite facies central region.However, the oscillatory zoning suggests a magmatic origin for these older domains, consistent with previous zircon dating of TTGs in the northern region indicating relatively young (c.2700 Ma) magmatic protolith ages (Friend & Kinny, 1995;Love et al., 2010).
In the central region, the Badcallian, shallowly, W-dipping S1 foliation is a composite fabric associated with a moderately W-plunging stretching lineation (Figure 3), subparallel extensional structures with a topto-the-west shear component (Figure 6a,b) and subparallel leucocratic sheets and leucosomes.These petrological and structural elements indicate that these Badcallian structures progressively developed during hightemperature metamorphism and continuous melt segregation at deep crustal levels.Previous P-T estimates (Johnson & White, 2011;Zirkler et al., 2012;e.g., Feisel et al., 2018;Gopon et al., 2022), point to high dT/dP (>775 C/GPa) metamorphism (Brown & Johnson, 2018) interpreted to coincide with initial development of paired metamorphic belts during the amalgamation of continental blocks in the Neoarchean (Feisel et al., 2018).Zircon grains from the three investigated hbl-TTGs (Figure 8c-e) exhibit CL-bright rims that yield ages between c. 2850 and 2700 Ma, with relatively high Th/U ratios (>0.1; Figure 8).These ages $2800 Ma are commonly associated with mixed analyses including core domains with oscillatory zoning reflecting magmatic crystallization ages (Gutieva et al., 2021), whereas ages of c. 2700 Ma of mixed analyses including an inner darker-CL rim domain between the core and the bright-CL outer rim represent metamorphic ages (Table S2).These older ages also record higher Th/U than the younger Laxfordian ages.High (and a wide range of) Th/U values are commonly recorded in zircon from terrains that underwent UHT conditions, where partial melting and continuous melt segregation occurred for a relatively prolonged time (Kirkland et al., 2015;Kunz et al., 2018;Yakymchuk et al., 2018).Nevertheless, Th and U variations and availability are also controlled by the abundance, presence, or absence of other major and accessory phases, particularly phosphates such as monazite and allanite, and their partitioning with zircon (Kunz et al., 2018).Therefore, Th/U ratios should be considered cautiously for the interpretation of ages in UHT terrains as they commonly do not reflect the typical metamorphic versus magmatic Th/U values as in lower temperature rocks (Kunz et al., 2018;Yakymchuk et al., 2018).
In the southern region, a shallow, west-dipping fabric associated with a moderate, west-plunging stretching mineral lineation in TTG gneisses from the Diabaig area (Figures 5 and 6c) were also identified.The shallow foliation may reflect a Badcallian relict (S1) preserved within Inverian and Laxfordian low strain domains (Love et al., 2010;Wheeler, 2007;Wheeler et al., 1987) or another pre-Laxfordian foliation that is not related to the Badcallian tectono-metamorphic event in the central region.The oldest zircon rims from both bt-TTG samples from the southern region yield metamorphic ages of c. 2750 Ma and high Th/U ratios (>0.1 and >0.5, Figure 8f,g).Combined, this data may indicate that while the central region of the mainland LGC was undergoing granulite facies conditions and partial melting (Badcallian event), the southern region was also exposed to thermal conditions favourable to zircon growth.Nonetheless, in this region, no granulite facies mineral assemblages have been reported and there is no evidence for the melt loss-related chemical depletion reported for the central region (Rollinson & Windley, 1980).It remains therefore unclear if both regions shared a common early metamorphic history.

| Neoarchean to early Palaeoproterozoic Inverian event
In all three regions, the presence of mafic Scourie Dykes is the main discriminator to distinguish Laxfordian from older fabrics, although the Laxfordian tectono-thermal evolution is known to include a range of magmatic and metamorphic events associated with different stages of deformation and stress regimes, challenging the identification of the sequence of superimposed structures (e.g., Coward & Park, 1987;Goodenough et al., 2013;Park & Tarney, 1987).In the northern region, NWstriking, pre-Laxfordian structures (interpreted here as Inverian) are associated with a SE-plunging mineral lineation and dominate the coastal part of the Kinlochbervie area (Figure 4).The Inverian S2 foliation is moderately to steeply SSW and NNE dipping, likely due to the Laxfordian deformation refolding and transposing the Inverian fabrics (Figure 4b).Both bt-TTG samples from the northern region contain very thin, CL-bright rims that yield 207 Pb/ 206 Pb dates of 2588 ± 12 Ma and 2522 ± 12 Ma and Th/U ratios <0.5, values that are typically interpreted as HT metamorphic zircon (Rubatto, 2017;Yakymchuk et al., 2018).These rim ages are consistent with the timing of the c. 2500 Ma Inverian metamorphic event (Fischer et al., 2021;Goodenough et al., 2013;Taylor et al., 2020;Zirkler et al., 2012).Although no P-T constraints are available for the Inverian event in the northern region, the formation of zircon rim domains with high Th/U values indicates that this region underwent at least amphibolite facies conditions (Rubatto, 2002) during this time.
The Inverian D2 structures in the central region are dominantly confined within the CSZ and LSZ (Figure 3).The Inverian structures are characterized by moderately to steeply SW-dipping foliation associated with a SEplunging lineation; however, they are mostly refolded during the D3 Laxfordian event, resulting in a more chaotic orientation of the S2 structures (Figure 3).These structural observations are consistent with previous work that identified a dominant SE-plunging Inverian lineation associated with a NW-striking axial planar S2 fabric in the central region (Cresswell, 1973).Previous P-T estimates suggest that the development of the D2 Inverian structures occurred in the mid crust at amphibolite facies at 500-625 C and 5.0-6.5 kbar (Cartwright et al., 1985;O'Hara, 1977;Sills, 1982;Zirkler et al., 2012).Nonetheless, hbl-TTGs from the central region contain zircon grains with thick, CL-bright rims in elongated, prismatic grains and slightly darker thick rims in round shaped zircon (Figure 8c-e).In all three samples, the analyses record a smear of concordant 207 Pb/ 206 Pb dates from c. 2850 Ma to c. 2450 Ma and Th/U ratios between 0.1 and 0.8 (Figure 8c-e), with a dominant population of bright CL rims at 2500 Ma.This is consistent with the results of previous geochronological studies, where a similar smear of apparent concordant zircon dates was interpreted to reflect either a prolonged permanence of the crust of the central region at high temperature conditions between c. 2750 Ma and c. 2500 Ma (Taylor et al., 2020) or, two, discrete, upper-amphibolite to granulite facies metamorphic events during which the U-Pb system was disturbed by ancient Pb loss (e.g., Fischer et al., 2021;MacDonald et al., 2015;Whitehouse & Kemp, 2010).The thick zircon rims in the hbl-TTG samples indicate that at c. 2500 Ma favourable thermal conditions for significant zircon growth were present, likely higher ($600-800 C; Rubatto et al., 2001) than previously envisaged (Zirkler et al., 2012).
In the southern region, the steep NW-striking, Inverian axial planar fabric is subvertical to moderately NE-and SW-dipping and is associated with a pervasive NW-plunging lineation (Figure 5).This lineation is likely the same SE-plunging lineation of the Inverian fabrics observed in the other regions and its scatter could be related to the Laxfordian overprinting event (Figure 5).Thick, CL-bright rims yield dates between 2528 ± 25 Ma and 2424 ± 40 Ma and high Th/U ratios between 0.3 and 0.9 (Figure 8f,g).In this region, no P-T modelling has been previously done, challenged by the pervasive Laxfordian overprinting metamorphism and scarce preservation of Inverian mineral assemblages.

| Palaeoproterozoic Laxfordian event
The northern region reflects a Laxfordian high-strain domain.Pervasive Laxfordian reworking and transposition of pre-Laxfordian structures by major upright folds (D3) and subvertical dextral shear zones indicate that the Laxfordian was dominated by NNE-SSW compression and significant NW-directed shearing, resulting in a dominant transpressional regime (Coward & Park, 1987;Goodenough et al., 2013).Commonly observed thin, millimetre-thick leucocratic veinlets and pockets in the amphibolites indicate that these rocks underwent incipient partial melting (Figure 6d,e,f).In the bt-TTGs from the northern region, the pervasive Palaeoproterozoic overprint is preserved by CL-bright zircon rims yielding 207 Pb/ 206 Pb dates between 1958 ± 29 Ma and 1805 ± 12 Ma and low Th/U ratios between 0.01 and 0.05, typical of metamorphic zircon (Rubatto, 2017;Yakymchuk et al., 2018).The Laxfordian foliation in the amphibolites is defined by zoned amphibole, where the Ti content (Otten, 1984) reflects decreasing temperatures from core-to-rim (Figure S1).The Ti-rich core commonly hosts euhedral to subhedral titanite grains and rounded to vermicular quartz inclusions, which indicate equilibrium between these mineral phases at high temperatures.The cuspate-lobate shape of quartz inclusions and interstitial quartz together with the presence of quartz films along amphibole grains are interpreted to indicate hydrous melt-rock interactions during cooling (Prent et al., 2019;Putnis et al., 2021) and incipient partial melting (Ferrero et al., 2012;Tacchetto et al., 2019;Vernon, 2004).The garnet-amphibolite LC1909 contains garnet crystals elongated parallel to the amphiboledefined foliation with Mn content increasing from coreto-rim (Figure S1).This reverse zoning indicates growth of garnet under temperatures high enough to allow internal Mn diffusion (Lopez Ruiz, 1976).In situ U-Pb titanite dating yields ages of 1776 ± 10 Ma and 1750 ± 18 Ma (Figure 9), which constrain the timing of the mineral assemblage defining the Laxfordian foliation for these two samples to c. 1760 Ma.These ages are within uncertainty of previous titanite ages (c.1750 Ma) obtained from Laxfordian granitic sheets in the northern region (Corfu et al., 1994;Goodenough et al., 2013).Titanite U-Pb dates cannot resolve the ages of the identified core and rim domains in sample LG2039 that overlap within analytical uncertainty, but we interpret the zoning to reflect titanite growth during different stages of the Laxfordian event in the presence of hydrothermal fluids (Goodenough et al., 2013).This interpretation is consistent with the observed spread in Zr and Y contents in titanite for sample LG2039, which may be related to the progressive growth and cooling of titanite.Titanite in sample LG2039 also records a negative Eu anomaly reflecting the presence of plagioclase, whereas the LREE enrichment indicates lack of other LREE-bearing phases during its growth (Figure 10).By contrast, titanite in the garnet-bearing amphibolite (LC1909) is slightly more depleted in LREE, and it is characterized by a weak positive Eu anomaly, suggesting either lower plagioclase content or a progressive breakdown of plagioclase due to water-rock interaction (e.g., Horie et al., 2008).The HREE-depletion in this sample is consistent with the presence of garnet and/or amphibole (Figure 10), whereas LREE depletion indicates the presence of other mineral phases such as epidote/allanite and apatite (Garber et al., 2017;Walters et al., 2022).The P-T constraints for the peak metamorphic assemblages of the investigated amphibolites indicate suprasolidus conditions at $5-7 kbar and $750 C for the northern region and at $6-8 kbar and $670 C for the southern region (Figure 12a-d).
In the central region, the Laxfordian structures are recorded within crustal-scale shear zones such as the LSZ (e.g., Goodenough et al., 2013) and in narrow, centimetre-to metre-wide high-strain zones (Figure 3).The moderate to steep WNW-ESE-striking Laxfordian fabrics are associated with a moderately SE-plunging lineation (Figure 3) and a dextral shear component (see also e.g., Coward & Park, 1987;Park et al., 1987).In contrast to the northern and the southern regions, the investigated hbl-TTGs from the central region do not record dates younger than 2400 Ma, although rare Laxfordian zircon overgrowth within the LSZ was reported (e.g., Goodenough et al., 2013).Previous P-T constraints for the Laxfordian event in the central region are limited and point towards amphibolite facies (Cartwright, 1990).
The southern region reflects another Laxfordian highstrain domain, recording major NE-directed compression and significant NW-directed shearing in a dominant transpressional regime.Laxfordian structures are associated with the development of a subvertical axial planar fabric pervasively recorded in bt-TTG gneisses and amphibolites.Laxfordian high-strain domains and steep belts extend across the southern region with subordinate low-strain domains in the Diabaig area (Figure 5).In this area, the strong reworking and transposition of pre-Laxfordian structures by major upright folds (D3) is associated with both dextral and sinistral shear zones (Wheeler et al., 1987).Isoclinally folded centimetre-thick leucocratic layers and veinlets elongated parallel to the dominant Laxfordian foliation in mafic Scourie Dykes indicate incipient and small degree of partial melting during the Laxfordian deformation event (Figure 6f).Micro-textural evidence such as vermicular and interstitial quartz films are also consistent with a mineral assemblage developed at suprasolidus conditions (Ferrero et al., 2012;Prent et al., 2019;Vernon, 2004;Waters, 2001).In the bt-TTGs investigated in this study, the Palaeoproterozoic overprint is recorded by CL-bright zircon rims yielding a 207 Pb/ 206 Pb date of 2130 ± 35 Ma and Th/U of 0.1, which is however slightly older than the Laxfordian zircon rim ages in the northern region.This older zircon age may be associated with the initiation of convergence in the southern region earlier than in the northern region (e.g., Mason, 2015).
In situ U-Pb dating of the titanite grains yields a discordia age of 1853 ± 20 Ma (Figure 9), constraining the timing of the peak mineral assemblage for sample LC1923.Titanite in this sample exhibits a granular texture which has been attributed to various microscale deformation mechanisms including dynamic recrystallization during shearing (Papapavlou et al., 2017), nucleation of neoblastic titanite along twin boundaries (e.g., McGregor et al., 2021;Papapavlou et al., 2017) or static recrystallization via heterogeneous nucleation and growth of new neoblasts due to influx of fluids (Cavosie et al., 2022).In this study, we interpret the titanite aggregates with granular textures to have formed by the growth of neoblasts as a result of prolonged deformation and titanite crystallization.The grain aggregates are texturally well-equilibrated with the peak metamorphic minerals, meaning that their age reflects the final stages of Laxfordian deformation.Additionally, the low-HREE pattern and the decrease in Y content are consistent with abundant garnet and amphibole in the sample, whereas the LREE depletion and the negative correlation in Y versus Sr indicate the presence of epidote/allanite and/or apatite (Garber et al., 2017;Walters et al., 2022).A relatively flat MREE pattern for titanite in sample LC1923 (Figure 10a) combined with a positive Eu anomaly indicate low plagioclase abundance, which is consistent with the scarcity of plagioclase in this sample.
Collectively, these results suggest that the metamorphic assemblage of amphibole, garnet, quartz, plagioclase, titanite, apatite and melt formed at c. 1850 Ma, which is $100 Ma earlier than in the northern region.This titanite age is only slightly younger than the granulite facies metamorphism recorded by the nearby Ialltaig Gneiss (1877 ± 13 Ma; Love et al., 2010).The P-T constraints for the Laxfordian peak assemblages reveal conditions of $6-8 kbar and $640-690 C, indicating distinctly lower temperatures than those recorded in the northern region (Figure 12d).The P-T conditions for the southern region are consistent with previous estimates of $6 ± 1.5 kbar and $630 ± 30 C for the rocks of the Loch Maree Group (Droop et al., 1999).

| The Palaeoproterozoic tectonothermal evolution the LGC
Previous studies have shown that the LGC records a composite (e.g., Coward & Park, 1987;Mason, 2015;Park & Tarney, 1987) and long-lasting (Goodenough et al., 2013) Laxfordian tectono-thermal evolution (Figure 13).Prior to the development of the convergent regime, the LGC was dominated by an extensional phase characterized by the intrusion of c. 2400-2200 Ma mafic Scourie Dykes (Heaman & Tarney, 1989;Waters et al., 1990) and by the deposition of volcano-sedimentary sequences (e.g., the Loch Maree Group) in the southern region of the mainland LGC, which were either associated with an extensional intracontinental basin environment (O'Nions et al., 1983;Whitehouse et al., 1997) or with an oceanic plateau basalt (Park et al., 2001).Palaeoproterozoic magmatic ages of c. 2000 Ma for the Ialltaig gneiss protoliths in the southern region of the mainland LGC (Love et al., 2014) and detrital zircon populations of c. 2000 Ma for the Loch Maree Group (Park et al., 2001;Whitehouse et al., 1997) indicate that an active margin of the convergent system initiated by that time, although earlier components may have been already recycled providing juvenile material to the volcano-clastic sequence of the Loch Maree Group in the mainland LGC and the Leverburgh in the Outer Hebrides (Cliff et al., 1998).Within this context, the c. 2130 Ma zircon rims found in the bt-TTG gneisses from the southern region of the mainland LGC may reflect part of this early convergence history recorded by mid-crustal levels.Subsequently, ongoing accretion between c. 1900 and 1800 Ma was associated with the emplacement of c. 1900-1850 Ma alkaline granite sheets such as the Ard gneiss intruding the Loch Maree Group in the southern region (Friend & Kinny, 2001;Goodenough et al., 2013;Kinny et al., 2005).In addition, high-pressure (≤14 kbar) granulite facies metamorphism at c. 1880 Ma recorded by the Harris Granulite Belt (Outer Hebrides) indicates deep F I G U R E 1 3 Schematic tectonic evolution of the LGC (modified after Park & Tarney, 1987) as part of the neighbouring Nagssugtoqidian Orogen system in Greenland (modified after Garde & Hollis, 2010).The stars indicate the inferred location of the LGC crust within the Nagssugtoqidian Orogen.Younger titanite ages at c. 1700-1650 Ma from the LSZ were previously interpreted as cooling ages coinciding with the post-collisional transpressional stage related to exhumation of the middle and deeper crustal levels of the LGC (Love et al., 2010).
subduction of crustal material to mantle depth (Baba, 2002;Cliff et al., 1983Cliff et al., , 1998;;Love et al., 2010).During the latter stage portions of the accreted continental margin were thus buried to greater depth than others, including the Loch Maree Group and most of the Archean blocks of the LGC (Mason, 2015).Our data show that by c. 1850 Ma, rocks of the southern region equilibrated at upper amphibolite facies conditions (6-8 kbar and $640-690 C), pointing to significant tectonic thickening and a Barrovian, collisional-type metamorphism.These new results combined with lack of HP relicts within the Ialltaig gneisses could well indicate that the Ialltaig gneisses recorded the same upper amphibolite facies conditions as the investigated rocks.If the latter is true, it would imply that the Harris Granulite Belt crustal block was buried to greater depths than the rocks of the southern region which, in contrast, recorded Barrovian, collisional-type metamorphism (Figure 13).In the northern region, upper amphibolite facies metamorphism reached slightly higher temperatures ($6-7.5 kbar and 730-760 C) at c. 1780-1750 Ma (Figure 9).This event is consistent with the collisional-type metamorphism of a thickened crust that may have been related to the late Laxfordian stage associated with dominant crustal thickening and strike-slip shearing.The diachronous metamorphic overprint between the southern and northern regions is here interpreted to indicate that the northern region was originally located further away from the active subduction zone and was only incorporated into the orogen during the later stages of the Laxfordian deformation.Collectively, our data for the northern region suggest a slightly younger and hotter tectonothermal evolution during the Laxfordian collisional event.This late-stage deformation was coeval with the intrusion of granite sheets between 1790 and 1750 Ma (Corfu et al., 1994;Goodenough et al., 2013;Kinny & Friend, 1997).Titanite cooling ages of c. 1670 Ma from granite sheets of the northern region (Goodenough et al., 2013) are broadly coeval with c. 1690 Ma pegmatitic veins from the southern region (Park et al., 2001).During the long-lasting Laxfordian tectono-thermal evolution, the granulite facies central region reflected a fragment of dry lower crust that behaved as a rigid low-strain domain within the hydrated thickened Laxfordian crust.In this scenario, hydrated retrogression to amphibolite facies conditions localized along crustal-scale shear zones in the central region.By contrast, the southern and northern regions represent Laxfordian high-strain domains that were fully hydrated during the Palaeoproterozoic reworking and crustal amalgamation.Potential, minor low-strain domain slivers preserved in those regions may have represented granulite facies Badcallian relicts that escaped the fluid-bearing Laxfordian high-strain deformation.Alternatively, these relict slivers may represent amphibolite facies low-strain domains that preserved pre-Laxfordian structures related to an earlier event in the region.The first scenario would imply that the Archean Lewisian crust was once contiguous and that the three regions reflect different crustal levels juxtaposed along crustal-scale shear zones.By contrast, the second scenario would imply that the structural relics identified in the northern and southern regions were related to another pre-Laxfordian deformation event.

| The LGC during Nuna amalgamation
Palaeogeographic reconstructions correlate the LGC to the Archean and Palaeoproterozoic evolution of the North Atlantic and Rae cratons prior to and during the amalgamation of the Earth's first supercontinent Nuna (Baba, 2002;Hughes et al., 2014;Park, 2022;Park & Tarney, 1987).Between 2000 Ma and 1600 Ma, major collisional orogenic belts formed across the globe as a result of convergence, accretion and collision between different major cratons and smaller crustal blocks to form the supercontinent Nuna (Figure 14; e.g., Kirscher et al., 2021;Nordsvan et al., 2022;Pourteau et al., 2018;Volante, Collins, et al., 2020;Volante, Pourteau, et al., 2020;Volante et al., 2022;Wan et al., 2020).Several active margins and convergence directions between adjacent cratonic blocks forming Laurentia and Baltica generated major Palaeoproterozoic (1900-1700 Ma) orogenic belts (Figure 14) such as the Nagssugtoqidian Orogen, following collision between the Rae and the North Atlantic cratons, and the Lapland-Kola Orogen after the collision between the Karelia and the Kola cratons (Buchan et al., 2000;St-Onge et al., 2009).At c. 2000-1900 Ma, the Rae Craton was subducting underneath the North Atlantic Craton (NAC) (Garde & Hollis, 2010;Kolb, 2014;Müller et al., 2018;Park, 2005;Park & Tarney, 1987).It was previously suggested that the LGC reflects crustal portions of this system (Park, 2022).The pre-Laxfordian continental mafic dykes in NW Scotland were also identified in the Nagssugtoqidian Orogen, where they are also associated with an extensional regime preceding the early Palaeoproterozoic compressional deformation and related to the contemporaneous deposition of volcanic-sedimentary sequences much like the Loch Maree Group (e.g., Nutman et al., 1999).The latter was interpreted as a marine sedimentary sequence deposited on a basaltic oceanic plateau (Park et al., 2001).Thus, the Loch Maree Group may represent remnants of a tectonic melange trapped in a suture zone between the Rae Craton and the NAC, reflecting a lower and upper plate environment, respectively (Park, 2022).
During the accretional stage, arc-type alkaline magmatism occurred at c. 1900 Ma in the LGC much like juvenile magmatism recorded at c. 2000-1900 Ma in the Lapland-Kola Belt (Daly et al., 2006) and at c. 1940-1870 Ma in the Nagssugtoqidian Orogen (Nutman et al., 2008;e.g., Lebrun et al., 2018).This magmatic activity was followed by c. 1890-1870 Ma eclogite and HP granulite facies metamorphism in Nagssugtoqidian Orogen in Greenland (Müller et al., 2018), in the Outer Hebrides (Mason, 2015), the Trans-Hudson Orogen in North America (St-Onge et al., 2009) and the Lapland-Kola Belt in Baltica (Lahtinen et al., 2009).In the Nagssugtoqidian Orogen, the timing of collision between c. 1870 and 1820 Ma correlates with medium-pressure amphibolite facies conditions of 5-7 kbar and 600-700 C (Garde & Hollis, 2010;Müller et al., 2018), which is indistinguishable from the conditions and timing established for the southern region and the Loch Maree Group.By contrast, the northern region of the mainland LGC records the final stages of transpressional deformation and higher temperature metamorphism ($6-7.5 kbar and 730-760 C) at 1780-1750 Ma (Figure 13).The associated pervasive NW-striking fabrics reflect dominant NW-directed shearing during NNE-compression, likely associated with the LGC being sandwiched between Baltica and the NAC during Nuna assembly (Figure 14).

F
I G U R E 4 (a) Simplified geological map of the northern region (Rhiconich terrane) with structural measurements, sample localities and foliation trajectory of the shallow pre-Laxfordian foliation (in grey).The map is derived from 50k scale BGS Digital Data under Licence No. 2021/018 British Geological Survey © UKRI.All rights reserved.(b) Stereographic diagrams (equal area, lower hemisphere projections) from top to bottom: density diagram of the pre-Laxfordian S1 poles; density diagram of the inferred Inverian S2 poles and related linear structures (i.e., L2); density diagram of the Laxfordian S3 poles, overlain by linear structures (e.g., L3).

F
I G U R E 6 Representative lithologies and fabrics in the central, northern and southern regions of the LGC.The green arrow indicates the north.Central region: (a) Felsic gneisses at Badcall with subhorizontal Badcallian axial planar fabric (S1) of isoclinal recumbent folds (F1), related to extensional structures such as boudinage of mafic layer parallel to S1.(b) Felsic gneisses at Badcall with hornblenditic pods on the Badcallian, shallowly W-dipping S1 plane recording dextral (top-to-the-east) shearing and associated with a west-plunging stretching lineation.Northern region: (c) S2 Inverian fabrics and mafic dykes folded during the Laxfordian deformation event in the western part of the northern region.(d) Foliated amphibolitic lens in felsic TTG gneisses with mm-thick veins and pockets of leucocratic material.(e) Foliated mafic Scourie Dyke with leucocratic veinlets elongated parallel to the dominant Laxfordian foliation.Southern region: (f) Foliated mafic Scourie Dyke with folded leucocratic veins and veinlets elongated parallel to the dominant Laxfordian foliation.(g) Pre-Laxfordian shallowly W-dipping fabric in an Inverian-Laxfordian low-strain domain in the Diabaig locality.(h) Steep, NW-striking, Inverian-Laxfordian highstrain zone.Felsic and mafic gneisses folded by the Laxfordian deformation event.

F
I G U R E 7 Plane polarized light (a-c, e, f), and backscattered electron (BSE, d) images of representative mineral assemblages and textures in the studied samples.(a) Garnet-bearing amphibolite from the northern region with titanite grains aligned parallel to the foliation.Relicts of clinopyroxene are preserved.(b) Garnet-free amphibolite with relict clinopyroxene.Quartz forms interstitial and vermicular textures.(c) Garnet-free amphibolite from the northern region with zoned amphibole crystals exhibiting darker brown cores with brighter green rims.Presence of quartz films and vermicular intergrowths of quartz and amphibole and titanite inclusions in amphibole crystals.(d) Subhedral biotite grains in amphibolite from the northern region.(e) Vermicular quartz and titanite grains in amphibole crystal from a garnet-bearing amphibolite (southern region).(f) Garnet-bearing amphibolite from the southern region.

|
Petrography and mineral chemistry4.1.1 | Central regionThree foliated, hornblende-bearing TTG (hbl-TTGs) samples (LC1903, LC1930 and LC1803) consist of plagioclase, quartz, hornblende, minor biotite, ilmenite, titanite (LC1803) and accessory minerals such as apatite and zircon.The foliation is defined by dark green hornblende alternating with quartz-and plagioclase-rich layers.Quartz forms either recrystallized neoblasts or coarser grains with subgrain boundaries.Ilmenite and rare apatite are also aligned along the foliation.Zircon occurs as inclusions in plagioclase.Sample 1930 is a foliated hbl-TTG gneiss from the Gruinard terrane with subhorizontal foliation defined by hornblende alternated with quartzand plagioclase-rich layers.Hornblende forms prismatic grains as well as symplectitic intergrowth with quartz.The latter are surrounded by a solid hornblende rim.Locally, hornblende cores exhibit a spongy texture, where they are replaced by late quartz and calcite aggregates, which are in turn surrounded by a green amphibole corona.F I G U R E 9 Tera-Wasserburg diagrams for LA-ICP-MS analyses of titanite and representative BSE images of the investigated titanite grains for (a) a garnet-bearing amphibolite (sample LC1909) and (b) a garnet-absent amphibolite (sample LG2039) from the northern region and (c and d) a garnet-bearing amphibolite from the southern region (sample LC1923).

F
I G U R E 1 0 (a) Rare earth element diagrams of titanite LA-ICP-MS analyses for sample LG2039 (in green), LC1909 (in blue) and LC1923 (in yellow).Concentration on the y-axis are normalized relative to the chrondrite values of Anders and Grevesse (1989).Binary diagrams of (b) Zr versus Al, (c) Y versus Sr and (d) Y versus Al for the analysed titanite-bearing samples.
The spread of apparent U-Pb concordant dates along the Concordia diagrams challenges calculation of weighted mean ages for individual samples.For this reason, the acquired data are referred to as individual 207 Pb/ 206 Pb date and no weighted means are calculated for individual samples (see below).