Protoliths

These were diagnosed from the chemical and mineralogical affinities of Lewisian rocks, but also from the shapes and relationships of bodies of rock in the field. A ‘Fundamental Complex’ was intruded by a ‘great series’ of dykes and sills. All the major rock types that we know in the Lewisian today were recognized. In the Fundamental Complex, ‘Rocks of probably Plutonic Origin’ (Peach et al. 1907, p. 33) comprise: ultrabasic rocks, such as pyroxenites; pyroxene gneisses with some feldspar and quartz; hornblende rocks with feldspar almost always present; quartz-feldspar-biotite gneisses; and muscovite-bearing gneisses. Basic rocks are usually older than acidic ones in outcrops showing diagnostic relationships (Peach et al. 1907, p. 73). Still in the Fundamental Complex, ‘Altered sediments’ include garnet schists, graphitic schists and calcareous rocks. The ‘Later Intrusions (Dykes and Sills)’ include ultrabasic through acidic material. In some regions of the Lewisian, the Survey found ‘ … the ultrabasic and basic intrusions preserve their dyke-like character, showing clearly that they rose along vertical or highly inclined fissures, after the development of early banding or foliation in the gneisses of the Fundamental Complex’ (Peach et al. 1907, p. 36). Thus, the petrology and shape of these bodies together pointed to their igneous origin. Cross-cutting relationships again give timing, and the ultrabasic dykes are usually later than the basic ones (e.g. Peach et al. 1907, map on p. 163). These ‘Later Intrusions’ are now referred to as the Scourie dykes.

Foliations and lineations

In the memoir banding was commonly interpreted as the result of deformation, though sometimes in a partially molten state, motivated by comparisons with plutons such as Criffel in the Southern Uplands of Scotland. ‘Basic inclusions are here extremely common, and they have been drawn out into lenticles and bands by differential movement which has affected also the later acid veins’ (Peach et al. 1907, p. 73) – the Lewisian commonly looks similar.

The memoir contains many references to lineations. For example, in a granite the lineation ‘is often more or less parallel to the foliation of the gneiss’ (Peach et al. 1907, p. 108); in Scourie dykes ‘a linear foliation or rod-structure replaces the plane-parallel foliation’ (Peach et al. 1907, p. 120); in gneisses in the Laxford area, The minerals on the foliation planes of the granulitic gneisses are often arranged with their long axes parallel. The direction of elongation, or stretching, varies in different localities, but near the same thrust is persistent for long distances. Over the greater part of the district, when the observer looks north towards the foliation planes, the lines of elongation appear diagonal, about half way between the directions of strike and dip of the planes, and with their lower ends on his right hand. (Peach et al. 1907, p. 138; our italics)

The last quote shows a sharp awareness of the three-dimensional (3D) nature of lineation (though the lack of a succinct way of expressing it), and also implies that the origin of such lineations is through stretching – a modern interpretation, though at the time there was no precise way in which to pursue the implications. In places on the published one-inch maps, the horizontal projections of the lineations are marked, with a note indicating that these should be used together with marked dips to determine the 3D orientation (Peach et al. 1907, p. 208).

A final quote shows that the possibility of solid-state deformation was appreciated. In relation to rocks near Gairloch (Peach et al. 1907, p. 98), These facts prove that the rock was either intruded during movement and consolidated in its present form, or that a previously formed rock was entirely recrystallised under the influence of the stresses which produced the ‘rodded’ or ‘mullion-structure’ of the district.

Quantitative structural geology, as we know it, was not available to assist in interpreting such data – but the next two sections show how deep insight was gained from the general patterns of deformation.

Dykes as time markers: small-scale relationships

Perhaps the most significant contribution that the memoir makes in regard of the Lewisian relates to timing of deformation. A single sentence captures that contribution. The detailed mapping [of the Lewisian] has shown that, after the eruption of the ultra-basic, basic and intermediate dykes, the whole area was in pre-Torridonian time subjected to earth movements which, affecting both the various members of the Fundamental Complex and also these later intrusions, gave rise to rapid plication [folding] of these strata and to lines of disruption or shear-zones, accompanied or followed by re-crystallisation of the original constituents, development of foliation, and occasional mylonitisation of the rocks. (Peach et al. 1907, p. 36)

There are several perfectly recognizable modern concepts here. It is crucial to be aware that this reworking is heterogeneous on all scales, and therefore is commonly visible within a single dyke. This theme appears many times in the memoir. For example, ‘Transitions from massive epidiorite [hornblende-plagioclase rock] to foliated hornblende-schist may take place within the limits of a hand specimen or even a microscopic section’ (Peach et al. 1907, p. 89; see also Teall 1885). On a still local, but larger, scale, we have (Peach et al. 1907, p. 150), A mile and a third E.S.E. of Ben Auskaird [itself 7 km SE of Scourie], the gradual bending, thinning and production of foliation in a broad dyke near a line of disruption can be clearly seen. About 20 to 30 yards on either side of the thrust, the dyke averages 50 yards in width, but in the deflected portions it is only three or four feet.

Note that 1 yard=0.9 m, 1 foot=0.3 m. Such deformation also affects the country rock: ‘gneisses next to the intrusive rock have been involved in the changes of dip and structure’ in some places (Peach et al. 1907, p. 150). Elsewhere, for example in what we now call the Canisp Shear Zone, reworking within gneisses is pictured (Fig. 3), and ‘ … between the lines of disruption, the folia of the older gneiss are sharply contorted and dragged into parallelism with the disruption planes’ (Peach et al. 1907, p. 167).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Sketch of fabrics from near Achmelvich, from the memoir. The original caption reads ‘Relation of newer to older planes of foliation in pyroxenic and hornblendic gneiss’. (Peach et al. 1907, fig. 8 on p. 1670).

Dykes as time markers: reworking and large scale structure

These small-scale observations must have helped promote the idea that bodies made entirely of hornblende schist, perhaps with gneissic banding parallel to their margins, were once intrusions. This interpretation leads, then, to a picture of heterogeneous deformation of the whole Lewisian outcrop. In terms of large-scale structure, this modification is not seen everywhere – thus a central district, from near Scourie to Loch Broom, is described as unmodified, with easily recognizable dykes. In the northern district between Laxford and Cape Wrath, ‘The basic dykes … appear in the form of bands of hornblende-gneiss, which can be followed for only short distances’ (Peach et al. 1907, p. 37). In the southern district, from Gruinard Bay to Loch Torridon and Raasay, basic dykes appear ‘either in the massive form or as hornblende-schists’ (Peach et al. 1907, p. 38). There seems to have been little doubt that sheets of hornblende schist could be correlated with dykes of the central district. These three districts remain a simple and useful division of the mainland and we will use them, in capitals to indicate they are the original definitions.

Large scale structures were also recognized, particularly in the Southern District, by tracing sheets of metasediment and/or hornblende schists around folds. Since those schists themselves represent deformed dykes, and those dykes themselves intruded sometimes banded rock of the Fundamental Complex, there are effectively at least three stages of deformation documented in the memoir.

Metamorphism

There was, at the time of the memoir, no theoretical framework for interpreting metamorphic mineral assemblages. Although the theory of thermodynamics had been developed in the 1870s (collected works in Gibbs 1906) it would be decades before it was applied in geology and even the facies concept was yet to appear (Eskola 1915). The account shows that the metamorphism was understood in terms of the rearrangement of material within individual rocks and, in some cases, bulk addition and removal. Note that the term ‘granulitic’ is used in a purely textural sense in the memoir, so when read in the quotes below it should not be confused with the modern concept of granulite facies.

Many types of gneiss that are characterized by pyroxene or hornblende correspond very roughly to granulite and amphibolite facies respectively. Quartz, when present in the granulite facies rocks, shows a distinctive blue or opalescent colour and in thin section has inclusions as ‘rows of minute dots’ and ‘extremely thin hairs’ (Peach et al. 1907, p. 54). Teall, doubtless prompted by his earlier observations (Teall 1885) notes that The rocks with hornblende occurring in fibrous or other aggregates … are common in the central zone in which the pyroxene-gneisses abound, and they are connected with these in such a manner as to suggest that the hornblende is in many, if not in all cases a secondary product after pyroxene.

This theme is reiterated in many places and is a description of amphibolite facies overprinting granulite facies minerals. So, on p. 138 (Peach et al. 1907) we find … the less altered gneisses near most of the pre-Torridonian thrusts are chiefly those with pyroxene or with aggregates of hornblende and biotite, which have probably replaced the pyroxene, and from these most of the granulitic biotite gneisses have evidently been derived. The biotite must have been chiefly formed from pyroxene or hornblende, and the white mica from feldspar.

Such changes may be associated with shearing: in the Canisp shear zone (Peach et al. 1907, p. 167) ‘… the granules of quartz lose their opalescence, are elongated and become granulitic, biotite is developed from the hornblende, and the feldspars are granulitised.’ Changes are documented in gneisses and dykes, both basic and ultrabasic. Thus, regarding a picrite dyke near Lochinver, ‘A series of specimens … show progressive alteration into a perfect schist composed of chlorite, pale hornblende and talc’ (Peach et al. 1907, p. 170).

Many observations made in passing are clear enough to be interpreted today. For example, in basic rocks north of Loch Laxford ‘Garnets … frequently occur in small aggregates surrounded by a thin white rim of feldspar granules’ (Peach et al. 1907, p. 120, see also p. 210 etc.). These are reaction rims and could relate to decompression. In the authors' experiences such textures are widespread in the Scourie dykes, though a detailed study is yet to be made. In gneiss from Ard Ialltaig, near Gairloch, a dark fine-grained rock shows ‘white spots, about the size of peas, with curved tail-like processes’ (Peach et al. 1907, p. 198): a tantalizing note given the general absence of kinematic indicators in the Lewisian. Even grain shape and crystallographic fabrics are documented: in a Scourie dyke from near Ben Stack, hornblendes are ‘arranged with their longest diameters parallel to the direction of stretching’ (Peach et al. 1907, p. 94). Thin sections were cut in three perpendicular orientations. The preparation at right angles to the direction of stretching abounds in sections [through individual hornblende grains] which show the characteristic cleavages meeting at angles of 124°, whereas such sections are rare in the other two preparations.

The hornblendes thus have a crystallographic preferred orientation with their c-axes (which are parallel to the cleavage intersections) aligned parallel to the stretching direction.

In terms of the large-scale distribution of minerals, in the district between Laxford and Cape Wrath, pyroxene-gneiss with blue quartz is absent (Peach et al. 1907, p. 37). In the district stretching from Gruinard Bay to Loch Torridon and Raasay, there is an ‘absence or comparative absence of pyroxene-gneiss with blue quartz’. Granulite facies assemblages, then, are only well preserved in the Central District. Together with the previous section we see that the memoir documents a Central District with granulite facies and comparatively little post-dyke deformation, with Northern and Southern Districts characterized by more intense post-dyke deformation and amphibolite facies assemblages.

Summary of memoir view of Lewisian evolution

The Fundamental Complex was made dominantly of igneous rocks, with much banding related perhaps to magmatic flow or solid state deformation. Some patches of basic rocks were intruded by the more acid varieties which form the bulk of the Lewisian; there were also some garnet mica schists interpreted as altered sediments.

After much deformation, a suite of basic dykes was intruded into the Fundamental Complex, cross-cutting earlier fabrics. Ultrabasic dykes are less common than basic ones and crosscut them.

Later, the dykes and their surroundings were deformed, but much less so in the Central District than in those to the north and south. Deformation involved the ‘drawing out’ of patches of rock, or ‘stretching’ to produce linear structures. Bending, thinning and foliation of dykes were all correlated, as was the metamorphism of the original igneous minerals to hornblende-plagioclase mineralogies. Consequently, foliated basic sheets with entirely that mineralogy, and parallel to foliation in the country rocks, were interpreted as original dykes. Such ‘hornblende schist’ sheets were folded, together with layers of metasediment, around large late folds. The Central District was characterized by pyroxene gneisses and these were transformed into hornblende bearing rocks, much as the dykes were. In this metamorphism biotite was commonly formed, and the blueish quartz often found in the pyroxene gneisses lost its distinctive colour. Ultrabasic rocks were metamorphosed to amphibole-talc assemblages. The Northern and Southern Districts have pervasive post-dyke deformation, and also generally lacked pyroxene and blue quartz in the country rocks.

This broad picture remains correct today.

Dykes as time markers: reworking and large scale structure

Sutton & Watson (1951) formalized the structural history of the Lewisian with the aid of new nomenclature. On the assumption that the dykes were of a single age they distinguished Scourian (pre-dyke) and Laxfordian (post-dyke) metamorphism and deformation, with a Central District relatively unaffected by Laxfordian reworking (Fig. 1). In that paper, the word ‘Scourian’ is defined as a metamorphic episode but is also used to describe protoliths which are older than the Scourie dykes: thus the term covers pre-dyke events and protoliths (Fig. 4). Dykes were proposed to have a single age or, more precisely, there were no deformation events during the period of dyke emplacement. In contrast, Bowes (1969) suggested four stages of dyke emplacement separated by deformation. Neither idea can be proved conclusively without absolute ages for deformation and intrusion. The first hypothesis has been generally adopted, yet in a modified form, recognizing that there was some ductile deformation during dyke emplacement (Tarney 1973). In the 1960s there was much development of structural ideas as the relationship of foliations and lineations to strain became clearer (Flinn 1965) and multiple stages of deformation were recognized.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Regional cross sections Modified from previous work (Coward 1984; Coward et al. 1970; Coward & Park 1987; Graham 1980; Park 2005), these are perpendicular to the general Laxfordian lineation, so any shear movements related to that lineation will be into or out from the page. The ‘faint’ colours and lines indicate extrapolation above present surface. (a) Outer Hebrides. The possibility that the South Harris Igneous Complex is in a synform is indicated, as is the hypothesis that the Scourian granulites of South Uist were once above the amphibolite facies rocks; (b) Mainland. The geometry as shown is well established, except for the northernmost continuation of the Loch Maree Group. This is shown arching over an antiform and passing beneath the Scourian block of the Central District as suggested by Park (2005); (c) and (d) Speculative sketches showing geometry along lines (a) and (b) after removing effects of late Laxfordian folding. The speculative Laxfordian D2 suture is shown, which may or may not have had metasediments and metaigneous rocks caught up along it. The sites of some ‘future’ D3 structures are shown – see text for details.

It is hazardous to correlate structures on the basis of style and orientation but we still find it useful for general description to use a sequence of D numbers for the Laxfordian deformation episodes as first described by Coward (1973) for the southern Outer Hebrides and subsequently by Coward & Park (1987) for the whole Lewisian. In this scheme, D2 is the main fabric-forming event, producing fabrics in the dykes and transposing them parallel to foliations in the country rock, although D2 does not appear everywhere. D3 is the event during which (usually) NW–SE-trending folds and shear-zones were formed that deformed the D2 fabrics, and D4 is a localized deformation producing minor structures associated with retrogressive metamorphism.

The Outer Hebrides were recognized as being dominated by Laxfordian deformation (Coward et al. 1970), with amphibolite sheets parallel to gneissic foliations almost everywhere (D2), and those foliations were affected by later stages of D3 folding around NW–SE axes (Fig. 4a) (Coward 1973). The effects of superimposed deformations were invoked to explain complicated structural patterns, using the (then) new idea that rock fabrics were the finite results of superimposed strains. For example, dominantly linear fabrics could be explained as the result of two superimposed plane strain deformations (Coward 1973).

Similar structures exist on the mainland in the Northern and Southern Districts (Fig. 4b); in particular NW–SE folds affect an earlier fabric and will be referred to as D3. In our account we make the simplest (but unproven) assumption that D3 folds are of roughly similar age throughout the Lewisian. On the mainland it was recognized that, where they are best preserved, the Scourian rocks themselves had a complex history prior to dyke intrusion. In particular amphibolite facies fabrics can be shown to be pre-dyke in parts of the Central District, despite similarities to Laxfordian tectonites in grade and NW–SE orientation. This retrogressive event, which must have involved a large input of water (of unknown origin), is defined as the ‘Inverian’ (Park 1964; Evans 1965). The older granulite-facies metamorphism and deformation in the Scourian of the Central District was named as ‘Badcallian’ (Park 1970). Field identification of Inverian structure is difficult, especially in the Southern District where the Lewisian is entirely amphibolite facies. The earlier Scourian though, can be identified in the field in terms of fabrics which are demonstrably early and do not trend NW–SE (Fig. 5a, b). Systematic mapping identified a substantial component of deformation as Inverian, based on, for example, subtle low angle discordances between Scourie dykes and wall rock fabrics (Park & Cresswell 1973; Fig. 5c). The dyke intrusion was argued to be controlled by these pre-existing NW–SE fabrics, not just in terms of orientation but also in terms of more abundant and thinner dykes in zones of high Inverian strain (Park & Cresswell 1973). The Inverian is now widely accepted as an important stage of Lewisian deformation (Coward & Park 1987; Wheeler 2007). In zones lacking Inverian fabrics, dykes are subvertical and the simplest explanation is that those zones have not rotated around horizontal axes since dyke emplacement (this assumption is used in the discussion of Laxfordian events, below).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Photographs illustrating key aspects of the Lewisian Complex: (a) Migmatitic early Scourian gneiss, south shore of Loch Torridon; (b) Early Scourian banding cut by thin undeformed Scourie dyke (a larger dyke forms the top of the hillock). North side of Loch Torridon; (c) Scourie dyke cross-cutting systematic foliation in gneisses at a low angle. The dyke is deformed but the angular relationship suggests that the country rock fabric is pre-dyke (Inverian). North side of Loch Torridon; (d) View NW of thick Scourie dyke on right (north) side of headland is parallel to (likely Inverian) foliation in country rock on left. North of Scourie, approaching the highest strain part of the Laxford Shear Zone; (e) Shear zone (Laxfordian) in amphibolitized Scourie dyke. NW coast of North Uist; (f) Scourie dyke, pervasively amphibolitized, transposed into parallelism with foliation in country rocks and folded during late Laxfordian deformation. NW coast of North Uist.

In these works, the various deformations were not attributed to an overall tectonic driving force or regime – but there would soon be renewed impetus for ideas to evolve. First, plate tectonics arrived. Second, Ramsay & Graham (1970) pointed out the importance of shear zones as a style of deformation (Fig. 5e). The Lewisian provided examples of outcrop scale shear zones (from a heterogeneously deformed Scourie dyke in North Uist) to fuel the discussion. Simple shear quickly became established as a model for large scale, as well as, outcrop scale structure, partly because it readily explains heterogeneous strain. One clear example of a strain gradient is the northern boundary of the Central District on the mainland, the Laxford Shear Zone (Fig. 1). This was interpreted in a model involving simple shear, with Inverian south side up reverse sense shear explaining the juxtaposition of the Central District granulites with the Northern District which appears never to have been affected by granulite facies (Beach et al. 1974). Later Laxfordian discrete sinistral normal sense shear zones formed just to the south. The curved shapes of these shears suggest they may be rotated listric structures (Wynn 1995). The evolution of the Laxford Shear Zone is discussed in detail elsewhere in this volume (Goodenough et al. 2010). Simple shear with dextral sense was also invoked to explain the kilometre-scale pattern of foliations in South Harris (Graham & Coward 1973), and is a major component of deformation in the Gairloch (Odling 1984) and Canisp (Attfield 1987) Shear Zones.

Laxfordian strain gradients on the kilometre scale are seen in the Southern District, for example, in the Gruinard area the strain gradually increases southwards (Figs 1 & 4b) in a manner broadly analogous to the Laxford Shear Zone. There is more complexity here though because of the folded metasediments of the Loch Maree Group and also because of the low-strain block, several kilometres across, in the Torridon area (the Ruadh Mheallan block, Fig. 4b). Passing south from this block, Laxfordian strain again increases but with a quite different style. Instead of a rather smooth increase in strain as the transition region is traversed, an anastomosing set of shear zones separating low-strain lenses is present (Wheeler et al. 1987; Wheeler 2007). Passing SW the abundance of low-strain lenses decreases, until the deformation is uniformly high strain and resembles that of the Northern District and the Outer Hebrides.

Simple shear on a still larger scale is a possible explanation for those high-strain fabrics which are folded but have flat enveloping surfaces in Laxfordian areas such as the Outer Hebrides (Coward 1975); such an idea was developed simultaneously in other Precambrian regions (Bridgwater et al. 1973). Coward alluded to the possibility that the lineation indicated overall transport direction, and that this could be related to plate movements: though he was cautious about both, this remains a ‘modern’ interpretation of this area (Park 2005) and is a widely applicable tectonic model (Shackleton & Ries 1984). The hinges of the D3 folds of the mainland and Outer Hebrides (Fig. 4) are broadly parallel to the NW–SE lineation, but they were formed in a later event that may relate to a separate collision, much younger than D2, described in the synthesis section later.

The Outer Hebrides exhibit, running along their eastern sides, a complicated zone of greenschist facies ductile fabrics, brittle fault rocks and pseudotachylites: the Outer Hebrides Fault Zone. These SE dipping fabrics and faults crosscut Laxfordian amphibolite facies structures and are later. They show a spectrum of plastically and brittle deformed rocks with somewhat similar kinematics, leading Sibson (1977) to propose that they are snapshots of a kinematically linked brittle fault – plastic shear zone system. This model underlines the role of shear zones as the plastic equivalent of faults on crustal scale, an idea widely applied since. Although the Outer Hebrides Fault Zone is not Laxfordian, the idea of linked brittle-ductile shear zone systems is as likely to have applied to the Laxfordian as to younger belts where it can be more easily demonstrated – there are no known remnants of major Laxfordian faults to which the shear zones coupled, although there are brittle structures which might, in principle, be Laxfordian. By the 1980s knowledge of thrust tectonics had evolved to include the geometries and consequences of ramp-flat geometries and fault-related folds (Boyer & Elliott 1982). These concepts, together with the idea that ductile shear zones are kinematically equivalent to faults, led to a vision of lower crustal deformation in which linked shear zones could form flats and ramps and could provide detachments for folds above them: a vision elegantly summarized by Coward (1984).

Current understanding of the largest scale Laxfordian features then is as follows. In the southern Outer Hebrides there is no doubt that later folds have affected what may well have been a subhorizontal shear zone, which was many kilometres thick since its upper and lower margins are not revealed in the cores of later antiforms or synforms (Fig. 4a). These D3 folds have distinctive form, with open antiforms and tight synforms and will have detached on a shear zone at depth. There are only small low-strain zones. On the mainland, the geometry of the Loch Maree Group leaves no doubt that there has been folding of a large shear zone. Here, though, the orientation of dykes in low strain zones can, if they are assumed vertical prior to the Laxfordian, provide a guide to possible ramp geometries. Thus, on Figure 4d the early Laxfordian fabrics are shown as ramping down beneath the low-strain zone near Gruinard. A band of metasediments, too thin to illustrate, occurs in this ramp region, and is possibly correlated with the Loch Maree Group. Ramps are also suggested beneath the Ruadh Mheallan low-strain zone at Torridon and beneath the Central District in the Laxford Shear Zone.

Concepts of crustal-scale shear zones led to a reassessment of the Inverian structural evolution, which was interpreted in terms of dextral transpression on discrete steep shear zones such as Canisp as well as on subhorizontal structures (Coward & Park 1987). The Inverian foliations and lineations are often coplanar with the later Laxfordian fabrics and, in places, collinear. One explanation is that intense Laxfordian strain will inevitably rotate earlier structures into parallelism. In addition the Scourie dykes appear to have been, at some times, relatively weak during the Laxfordian. At Torridon, Inverian shear zones anastomose and, to a first approximation, they share a common lineation. Scourie dyke intrusion was controlled by these shear zones. If slip is localized on these dykes, then, movement is ‘guided’ by dyke intersections and is thus parallel to the Inverian lineation (Wheeler et al. 1987).

The Outer Hebrides Fault Zone was itself thought to have a thrust sense of movement, partly because it brings granulite facies rocks above amphibolite facies in South Uist, but there is much evidence for normal movement in addition (White & Glasser 1987). Consequently, it appears that the same zone has accommodated different movements and is a persistent weak zone: several episodes of reactivation have been proposed (Butler et al. 1995).

Nomenclature

In this contribution we use the term ‘Scourian’ for any pre-Scourie dyke protolith or metamorphism (with or without deformation), and the term ‘Laxfordian’ for any post-Scourie dyke protolith or metamorphism (with or without deformation), ‘Badcallian’ means Scourian granulite facies metamorphism (with or without deformation) and ‘Inverian’ means Scourian amphibolite facies metamorphism (with or without deformation) which, in the Central District, is seen to postdate the Badcallian. In the Southern District, most fabrics are amphibolite facies but have a variety of ages and orientations. Here, the latest Scourian fabrics run NW–SE and are referred to as Inverian because they are relatively late in the Scourian history and are parallel to Inverian structures in the Central District. Our terms are based on field relationships. Kinny et al. (2005) propose a large number of new names for Lewisian events, based on zircon dating coupled to a proposal that the Lewisian is divided into numerous terranes (an idea discussed later). Currently we cannot relate this nomenclature scheme to field relationships or to the Lewisian literature that we review here, so we adhere to the traditional terms.

Protoliths and geochemistry

Geochemistry provides a precise basis for diagnosing the origins and/or metasomatic history of protoliths. The bulk of the Lewisian was formed in the late Archaean; we split this section into three according to the ages of protoliths as established in the field, though for one or two we deploy geochronological data which are discussed in detail in a later section.

Metamorphism

There are three developments we highlight here. First, the advancement of metamorphic techniques allowing precise characterization of conditions; second the link between metamorphism and large-scale tectonics and third the recognition of complicated feedbacks between deformation and metamorphism.

Geochronology

Field based and petrographic observations can, at best, constrain the relative order of events. Absolute dating is essential to understand the Lewisian, but we need to ask: what exactly does a radiometric age mean in a metamorphic region (Cliff 1985) – the formation of the protolith (igneous, in the case of the quartzofeldspathic gneisses and Scourie dykes), the metamorphism, the deformation, a fluid influx event, or the cooling? Indeed in crust, which has grown through magmatic accretion, can we distinguish completely between igneous and metamorphic events, especially in the case of magmas added directly to the deeper crust? This issue arises when interpreting not just the Badcallian but also the earliest recrystallization of the Scourie dykes, and the South Harris Igneous Complex metamorphism. We need to appreciate that age data are most valuable when tightly linked to field observations and the petrographic setting of individual dated minerals. At the time of writing, geochronology is the most controversial aspect of Lewisian research. In the account we give, note that the early work may have been re-interpreted or superceded: and this may eventually happen to recently obtained data too. We will not review all dates but give a basic outline, with examples selected to highlight current debates in interpretation.

Metamorphic ages

Igneous ages may be used to date, or constrain dates for, metamorphism if it is demonstrable in the field that the igneous body is contemporaneous with, or has a definite time relationship to, the metamorphism. For example a trondhjemitic sheet near Scourie dated at 2720 Ma (Corfu et al. 1994) is interpreted as syn-granulite facies (Cartwright 1988).

Of more general applicability is the idea that if a mineral shows a growth stage which is interpreted to be below its closure temperature and can be linked to a particular metamorphic event (via textures) then that event can be dated. Zircons commonly show distinctive growth textures related to metamorphic events, and have a very high closure temperature. U–Pb concordia diagrams may also show discordant ages which are related to lead loss due to heating during metamorphism. Thus, zircon may be used, with care, to date metamorphism (Parrish 2001) but a key issue is the ambiguity in relating growth and/or lead loss to specific metamorphic events (Parrish & Noble 2003).

Scourian metamorphic ages include recrystallized zircon rims from Gruinard Bay showing c. 2730 Ma, interpreted as granulite facies, and monazites (enclosed in garnets) showing 2760 Ma near Scourie (Table 1). Monazites from less than 1.5 km from the 2760 Ma sample show 2526 Ma; new growth of zircon in a quite fresh granulite is dated at 2490 Ma; and zircon from a pyroxenite grew at 2470±30 Ma (Kinny & Friend 1997). The latter age is referred to as ‘Inverian’ but in its original usage this was intended to refer to an amphibolite facies overprint. The last two dates are from quite fresh granulites and so it is argued must relate to a granulite facies event. However it is certainly the case that zircon can grow at amphibolite facies (Parrish & Noble 2003). The granulite facies region as a whole exhibits much heterogeneous amphibolite facies overprinting, that heterogeneity being determined by the supply of water not by variations in temperature. How sure can one be that zircon growth was not triggered at amphibolite facies temperatures (the rock as a whole retaining its granulite facies assemblage due to lack of water)? Note also that the c. 2500 Ma ages are not registered in all rock types, even those which are so close together that they must have shared the same P–T history (an example being the trondhjemitic sheet from Badcall dated at 2720 Ma). Thus, in terms of minerals dating metamorphic events, ‘absence of evidence is not evidence for absence’.

Laxfordian metamorphic ages cover a considerable span. Cooling after granulite facies metamorphism in the South Harris Igneous Complex anorthosites was dated by Sm–Nd as 1870± 40 Ma; this agrees with zircon metamorphic overgrowths of 1880 Ma (Friend & Kinny 2001). Because the metamorphic ages are close to the intrusive ages here, it is plausible that the UHT conditions recorded represent residual magmatic heat (Baba 2004). The granite/migmatite complex is much later (1675 Ma).

The exclusively amphibolite facies Laxfordian on the mainland appears much younger than the granulite facies in the South Harris Igneous Complex, despite the existence of an 1855 Ma granite north of the Laxford Shear Zone. Monazites are interpreted to have grown at amphibolite facies at 1752 Ma. Titanites grew at 1760 Ma and 1670 Ma in the Scourie–Laxford area (U–Pb) and were linked to hydrothermal activity related to granites (Corfu et al. 1994) although no comparable dates are known from nearby granites. Rutile from a Scourie dyke shows a formation or reset U–Pb age of 1689 Ma. Note that the closure temperatures of these minerals are not well known so there is a chance that these are closure not crystallization ages. Pegmatites in the Gairloch area synchronous with Laxfordian D3 deformation are 1700 Ma (Park et al. 2001): this is the only Laxfordian date that can be directly linked to structures.

Disregarding low closure temperature cooling ages, these are the youngest metamorphic events in the main part of the Lewisian, though pseudotachylites are common in the Lewisian & Sherlock et al. (2008) find Grenville ⁴⁰Ar/³⁹Ar ages from the Gairloch region (1024–980 Ma). These are actually younger than the oldest age for unmetamorphosed Torridonian sediments, 1200 Ma (Kinnaird et al. 2007). The eclogites of the eastern part of the Glenelg-Attadale inlier are Grenville in age (Sanders et al. 1984; Storey et al. 2005). Remarkably, in view of their proximity and similar grade, the eclogites in the western part of the Glenelg-Attadale inlier are much older, c. 1700 Ma (Storey 2002).

We noted in the metamorphic discussion that the Lewisian has been described as undergoing slow cooling through its history. However the Loch Maree Group metasediments undoubtedly show a burial cycle in the Laxfordian – the adjacent gneisses may have experienced this too. Moreover, the younger Scourian metamorphic ages described in this section have been attributed to a discrete heating event (Corfu et al. 1994). We cannot yet be sure whether the gaps between recorded events relate to a smooth or cyclic P–T evolution. Figure 6 relates metamorphic conditions to dates for three parts of the Lewisian.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Metamorphic P–T histories and dates in Ga for three parts of the Lewisian. It is important to understand that these paths are based on just a few points, and potentially have as yet unrecognized P–T excursions and reburial episodes. Scourie path from Cartwright & Barnicoat (1989) except Scourie dyke ages span 2.0–2.4 Ga and exhumation to the surface was complete by 1.2 Ga; other dates are not adjusted from the published figure in view of current uncertainty about possible multiple granulite facies events in the Scourian. South Harris path from Baba (1998) with the metamorphic age marked by us (speculatively) as relating to peak pressure. Loch Maree Group shown as a single PT estimate (Droop et al. 1999) with a speculative age, plus an indication that those sediments were deposited later than 2.2 Ga, to emphasize the difference to Scourie.

Geophysics

Large scale movements and plate tectonic setting

Palaeomagnetism may in principle give information on crustal movements if contrasting remanence vectors of equivalent ages are found. For example Piper (1992) shows how such data, together with matching of aeromagnetic anomalies, support c. 95 km of (presumed Caledonian) sinistral strike slip in the Minches. Scourie dyke trends provide a comparable estimate (Lisle 1993). Remanence is often acquired during cooling so the directions recorded relate mainly to cooling after the Laxfordian event: they vary regionally but, apart from the above result, variations relate to the complex details of magnetic carrier mineralogy.

On a much larger scale, palaeomagnetism contributes towards reconstructions of the positions of major cratons. A plausible palaeomagnetic reconstruction at c. 1350 Ma (Buchan et al. 2000) places the Proterozoic continent of Baltica against the eastern margin of Laurentia, close to the position of the Lewisian fragment (Fig. 7) but in an orientation c. 90° anticlockwise from its mid-Phanerozoic position. Thus Palaeoproterozoic events and structures in Greenland, Labrador and Baltica have a direct relevance to the Lewisian (Table 2). In particular there are deformed belts formed at c. 1.91–1.84 Ga in Greenland (the western and eastern Nagssugtoqidian) and in Baltica (the Lapland-Kola) as shown on Figure 7, and in Labrador (the Torngat), It has in the past been proposed that these belts were ‘intracratonic’ in that they formed by localized shortening within a contiguous region. It is now recognized that they were collisional orogens hosting subduction-related calc-alkaline plutons (Kalsbeek et al. 1993; Park 1994). During the same period, major subduction-related crustal growth occurred in the Svecofennian belt on the southeastern side of Baltica.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Palaeogeographic evolution of North Atlantic area during the Laxfordian. Arrows indicate convergence directions at each stage. Abbreviations: Lew, Lewisian; NS, north Scotland; NI, north Ireland; for others see Table 2. Colours relate to the age of last main pervasive deformation (Table 2), except for green which indicates material in island arcs etc. which is juvenile in relation to each time illustrated. Dots indicate calc-alkaline igneous activity. Modified after Park (2005) and references therein.

View this table:

• View inline
• View popup

Table 2.

Tectonic units in the North Atlantic area (see also Fig. 6) with broad ages of Palaeoproterozoic deformation and magmatism (Park 1995, 2005)

Following on from the period of cratonic amalgamation, a major phase of subduction-related crustal growth took place in both Laurentia and Baltica at the margins of the combined continental mass from c. 1830 Ma to c. 1600 Ma in the Ketilidian & Gothian belts (Fig. 7) and also in the Labradorian and Makkovik belts (Park 1991, 1994; Gower & Krogh 2002).

We should expect the Lewisian to fall into this general framework. The Loch Maree Group and South Harris Igneous Complex then represent igneous activity and sedimentation related to a closing ocean, with igneous intrusion and metamorphism close together at c. 1900 Ma. The ‘main’ Laxfordian deformation, D2, associated with flat-lying fabrics and a NW–SE lineation has not been dated but plausibly that relates to the same oceanic closure. Material both above and below the suture is heterogeneously deformed during the Laxfordian. The Laxford Shear Zone is not likely to have accommodated huge movements during the Laxfordian, so does not represent a Laxfordian suture – it is probably Inverian (Goodenough et al. 2010) – consequently the difference between the Laxfordian of the Central and Northern Districts is due to heterogeneous strain in what was a single body of continental material. Similarly, the strain gradient at Loch Torridon does not, in the field, appear to include a major suture, and the area exemplifies strain heterogeneity on all scales [e.g. fig. 2 of Wheeler (2007)].

The late Laxfordian D3 deformation seems to relate to NE–SW shortening and is substantially younger (c. 1700 Ma). The structures may be explained by an approximately north–south-orientated compressive stress. Park (1994, 2005) relates this tectonothermal event to the formation and subsequent collision of a marginal volcanic arc co-extensive with the contemporary Gothian belt of Baltica (Fig. 6) and the Labrador belt in Laurentia. The Laxfordian D3 event did not form new fabrics on the large scale, hence in Figure 6 the Lewisian is shown as outside the ‘main’ part of the Gothian-Ketilidian belt.

Whatever the precise assignment of subareas to the lower and upper Laxfordian plates, it seems that both experienced the Inverian, both experienced intrusion of a dyke swarm, and where undeformed those dykes in the final assembly have a rather consistent NW–SE trend.

Early Scourian

The earliest rocks are metabasites with associated metasediments – some of which could be fragments of Archaean oceanic crust. This crust was metasomatized during subduction, losing U, Th and K. It melted at c. 20 kb, creating voluminous TTG magmas (at 3000–2900 Ma), which themselves had low U, Rb, Th and K and are found in the Central District and southern Outer Hebrides. At other times and/or in other places, crust was subducted with a variety of dip angles, and was not always depleted prior to melting. Similar magmas, though not depleted in those LIL elements, intruded at 3100 Ma in the Outer Hebrides and Southern Districts and at c. 2800 Ma in the Northern District, though all districts have a span of TTG ages. Associated with this early crust-forming event was the formation of the subcontinental lithospheric mantle. Structural arguments that the crust deformed by subhorizontal shear in the Laxfordian, and general analogies with Phanerozoic orogens, indicate the possibility of large horizontal movements which can displace crust sideways from the lithospheric mantle it was originally above. Nevertheless it seems that the Scourie dykes (which provide isotopic evidence for the age of their source) formed above lithospheric mantle which has the same age as the earliest Scourian crust.

Deformation forming early (Badcallian) banding and fabrics accompanied the peak metamorphic granulite facies metamorphism, which is substantially later than TTG formation in all parts of the Lewisian where it is seen. In the Gruinard area mafic rocks crystallized at 2800 Ma or earlier but granulite facies metamorphism occurred at 2700 Ma. In the Central District (Scourie region) a 2760 Ma granulite facies event is documented in monazites. There are several metamorphic dates near 2500 Ma from the Central District, which are suggested to relate to a later granulite facies event. In the Torridon area and further south, there is no firm evidence for Scourian granulite facies but it is present in the Outer Hebrides above the Outer Hebrides Fault Zone.

Pressure estimates indicate that the granulite facies rocks we see now at the surface were once at >30 km depth, and there is another 30 km or so of crust beneath them now, so it is possible that there was a 70 km-thick crust in the Badcallian, with the present exposure level in the middle. As these rocks are mostly very low in heat-producing elements, radiogenic heating is unlikely to account for the metamorphism. Magma addition could produce not only thickening but also heat – however it is worth remembering the time gap between the youngest TTG ages and the granulite-facies ages in all Scourian regions. Alternatively, as in Phanerozoic orogens, the high grade rocks found at the surface are the result of thrust-related uplift (which could equally be referred to as tectonic underplating) coupled to unroofing by erosion and/or tectonic extension. The widespread occurrence of granulites in Archaean cratons led O'Hara (1977) to prefer ‘non-tectonic’ (i.e. igneous) underplating. He argued that the granulites were initially not far above the Moho: as a consequence most of the present thickness of the crust would have been added later. We do not dismiss the possibility of magmatic underplating but as discussed in the metamorphic section, there are limits to what it can explain. It cannot produce the juxtaposition of rocks with different P–T–t histories, nor can it explain repeated uplift and reburial; we therefore prefer the tectonic explanation.

Later Scourian

There is limited petrographic or field evidence for two granulite facies events in the Central District. Instead, there is abundant evidence for amphibolite facies overprinting which preceded the Scourie dykes – the ‘Inverian’ event as diagnosed from field criteria. Whether the term should be redefined as a time period, including the postulated second granulite facies event, or restricted to refer to amphibolite facies episode, is a matter of nomenclature.

The major question is how c. 2500 Ma dates in granulite-facies rocks can be reconciled with field and petrographic information that suggest an amphibolite-facies event prior to Scourie Dyke intrusion, which started at 2400 Ma. As defined on field criteria, the Inverian has a systematic structural style: steep NW–SE shear zones formed at amphibolite facies, retrogressing anhydrous rocks if they were the local precursors. Associated is the static retrogression of granulites to amphibolites in the Central District, so large amounts of water were added to the present exposure level at this time. Pressure is substantially less than 10 kb, a result of the first stage of gradual unroofing of the Badcallian complex. The Inverian itself could plausibly have caused the Badcallian granulites to be thrust over lower-grade rocks on a ductile ramp-flat system (Coward 1984): tectonic underplating assisting the uplift and unroofing of the granulites. Inverian shear zones include the Canisp Shear Zone which is within what is usually considered a coherent piece of crust, and should be contrasted with the Laxford Shear Zone, which separates quite different pieces of crust and may relate to docking of those different terranes late in the Scourian (Goodenough et al. 2010).

The Inverian may be as important as the Laxfordian in terms of deformation, but its effects are partly masked by that later stage and so are difficult to unravel. Where fabrics in quartzofeldspathic gneisses are parallel to completely foliated Scourie dykes, as in the Northern District and Outer Hebrides, there are no field criteria allowing the two events to be separated.

Scourie dykes

The Scourie dyke swarm is NW–SE trending and dominated by quartz tholeiites derived by melting of enriched subcontinental lithospheric mantle. It represents NE–SW crustal extension and span 2400–2000 Ma; ultrabasic dykes arrived late in the sequence according to field criteria but yield an early date. Intrusion geometry was commonly controlled by NW–SE Inverian fabrics where those were present: dykes appear to have been thinner, and to anastomose more, in such areas, but roughly the same dyke trends exist regardless of wall rock structures. In areas of control by steep Inverian structures, the dykes are subvertical (i.e. sill-like) but are also subvertical when seen far from areas of Inverian deformation.

Dykes often have a static amphibolite facies overprint, but those may be cut by fresh dykes, as if the Inverian episode of water addition was continuing. Other dykes have autometamorphic garnet or two pyroxene metamorphic assemblages indicating intrusion into hot country rock: overall then we are looking at considerable depths and wall rock temperatures despite the obvious brittle failure involved in intrusion. There is no evidence for any ambient pressure change from the earliest to the latest dykes, so there must be a limit on the amount of erosion that occurred during this 400 Ma time interval.

Early Laxfordian

Just after the youngest known Scourie dyke, the Loch Maree Group metasediments and calc-alkaline rocks of South Harris formed and were incorporated into a Laxfordian orogen. Figure 4c, d show a simple (two plate) model for the early Laxfordian structures. Park (2005) argues that a SE plate was subducted under a NW plate. NW–SE directed ductile overthrusting created crustal-scale shear zones throughout much of the Lewisian.

The Loch Maree Group lies beneath basement rocks that must have been thrust over it during the Laxfordian and the simplest explanation for its metamorphism is burial and conductive/radioactive heating. Thus, the early Laxfordian involved burial and progradation in the style of a Phanerozoic orogeny. It is rather hard to envisage the footwall rocks to the Loch Maree Group being at amphibolite facies when the Loch Maree Group was emplaced into its present position. The Loch Maree Group, an accretionary complex, is likely to have been structurally higher than the middle-lower crust of the Lewisian continental margin over which it was emplaced. Instead, it is more likely that the footwall had been exhumed to high crustal levels before the Laxfordian, and was then tectonically buried, heated and deformed just as was the Loch Maree Group above it. Figure 4c shows most deformation occurring in the footwall to this main suture; deformation occurred at amphibolite facies throughout the levels exposed today on the mainland, and Scourie dykes were transposed parallel to Laxfordian wall rock foliations. A single granite sheet north of the Laxford Shear Zone is known to be of this age.

The main Laxfordian fabrics on the mainland are on a large scale closely related to the Inverian ones. This is due to the transposition of Inverian fabrics during Laxfordian strain, and to the behaviour of dykes as easy-slip horizons leading to a coincidence in lineation directions (Wheeler et al. 1987).

In the Outer Hebrides, the calc-alkaline nature of the South Harris Igneous Complex suggests that it belongs to the upper plate, regardless of whether it was an island arc or was intruded into adjacent rocks. It experienced granulite facies metamorphism at 1880 Ma; there is no evidence for this in the TTG gneisses and, if they really did not experience it, there must be large relative movements between the South Harris Igneous Complex and the TTG gneisses. Figure 4c interprets the South Harris Igneous Complex as part of a larger nappe, thrust over pervasively deformed lower plate rocks [but carrying the South Uist granulites above it (Park 2005)]. The remaining outcrop is interpreted as synformal as a result of D3 (the antiformal closure of the anorthosite body may have been established earlier, and hence is not diagnostic of the overall South Harris Igneous Complex geometry). Coward (1984) linked the overthrusting of a hot granulite facies nappe to the formation of the migmatite complex in its footwall. New dates show that those migmatites are 200 Ma younger and cannot be related in that fashion. However emplacement of such a nappe would undoubtedly have led to temperature rise in its footwall, by downwards heat conduction and by providing thermal insulation. Pursuing that argument, the footwall rocks would have been colder than amphibolite facies prior to the overthrusting, and experienced a prograde path, though there is no direct evidence for this. Moreover, the footwall rocks did not experience granulite facies so the South Harris Igneous Complex nappe must have been partly unroofed prior to emplacement: a common phenomenon in Phanerozoic orogens.

Later Laxfordian

Late Laxfordian NW–SE trending folds and dextral shear zones (D3) that overprint the main Laxfordian fabrics (D2) are associated with syntectonic pegmatites, dated at 1700 Ma, considerably post-dating South Harris Igneous Complex metamorphism (at c. 1850 Ma) suggesting a separate event. The D3 structures involved roughly north–south shortening, in places partitioned into steep dextral shear zones. There is no D3 suture in the main outcrop of the Lewisian. However, the scale of folding shown in D3 (Fig. 4) is quite considerable, and there must have been ductile shear zones at depth on which these folds detached, so significant thickening of the crust is implied. It has been suggested by Park (1994) that these D3 structures may be related to collision with an arc or small terrane further south (Fig. 7). Monazites record a 1750 Ma amphibolite facies event near Scourie and migmatization occurred in the Uig Hills – Harris Granite Complex. The causes of amphibolites facies metamorphism and migmatization remain to be clarified but significant crustal thickening is expected to have had thermal consequences.

On the mainland, NW–SE folds are present in the Northern and Southern Districts. The low strain zone at Torridon presents a problem in interpretation. The Loch Maree Group is clearly folded (Fig. 4b) as is its immediate footwall. The Gairloch geometry suggests an isoclinal fold. If the original base of the Loch Maree Group nappe was subhorizontal then it, and the adjacent rocks to the south, would have been rotated 90°. However the Scourie dykes just to the south are still subvertical. For this reason we propose a D3 shear zone juxtaposing the Ruadh Mheallan block (in its original attitude) next to the Loch Maree Group (tightly folded). This is just one of several options for the structural evolution of this area.

These events established the main Lewisian features in the form we see them today, apart from the removal of another 15–20 km of material to exhume the Laxfordian amphibolite facies rocks, a process complete when the earliest Torridonian rocks were deposited prior to 1200 Ma (Kinnaird et al. 2007). Younger than that, isolated Grenville ages relate to pseudotachylites along certain faults, and to eclogites in the Glenelg-Attadale Inliers. Caledonian and later deformation affected the Outer Hebrides Fault Zone and led to 80–90 km of sinistral offset along the Minches.