1. Introduction
The first gold mine in Scotland, at Cononish in the Grampian Highlands (Fig. 1) will possibly commence production soon and it is timely to review the nature of this and other Au–Ag deposits in the Caledonides of Scotland and Ireland. They have many features in common, and the aim of this paper is to propose a possible genetic model to explain their origin. The deposits occur in Dalradian rocks of Neoproterozoic–Lower Ordovician age that form the Grampian Terrane (Fig. 1), and in Ordovician and Silurian rocks that lie unconformably upon them, which were folded prior to the injection of the gold-bearing quartz veins. The Dalradian rocks were affected by polyphase deformation (D1–4) during the ~470 Ma Grampian Orogeny (Oliver, Reference Oliver2001; Cooper et al. Reference Cooper, Crowley, Hollis, Noble, Roberts, Chew, Earls, Herrington and Merriman2011; Hollis et al. Reference Hollis, Roberts, Cooper, Earls, Herrington, Condon, Cooper, Archibald and Piercey2012; Tanner, Reference Tanner2012).
Figure 1. Outline map showing the names and locations of the gold-bearing mineral deposits (1–9), related igneous intrusions (10–12, 14) and the Rhynie cherts (13) in the Grampian Terrane (in grey) of Scotland and Ireland. The inset shows the approximate stratigraphic age of the host rock in each case (numbered as on the map).
Gold-bearing deposits with economic potential were first discovered in the Caledonides of Scotland and Ireland during a burst of exploration activity in the bonanza years, 1983–1989. The finding of in situ gold mineralization at Curraghinalt in the Sperrin Mountains of Northern Ireland in 1983 was followed by similar finds at Lecanvey (Croagh Patrick) (1984); Cononish (1985); Cavanacaw (1985); Urlar Burn (1988); Cregganbaun (1988); and Tombuie and Bohaun in the late 1980s (for locations, see Fig. 1). These deposits, hereafter referred to as the ‘Grampian Au–Ag deposits’, were nearly all located by well-tried prospecting techniques such as field mapping combined with the panning of stream sediments, and tracing of mineralized, gold-bearing float back to its source. However, as is commonly the case, the next phase, that of finding new major deposits by geochemical analysis of stream sediments and soils; channel sampling; and diamond core drilling, has since 1990 been largely unsuccessful. An analogous situation occurred in the world-class Charters Towers goldfield in NE Australia where, in 1871, nearly all of the economic gold deposits were located within two weeks of the initial discovery, and no significant lode-gold deposits have since been found (Kreuzer et al. Reference Kreuzer, Blenkinsop, Morrison and Peters2007).
Commercial exploitation of the Grampian Au–Ag deposits has proved difficult. The mineral exploration licence for Lecanvey, in the Irish Republic, was withdrawn in 1990, as it is located on Croagh Patrick Mountain, which is a religious pilgrimage site. In 2007, following a protracted public enquiry, an opencast mine was opened at Cavanacaw in Northern Ireland. In 2012, Scotgold Resources were granted permission to commence mining at Cononish in Perthshire, Scotland. However, the initial planning application was rejected, largely because the deposit lies within the Loch Lomond and the Trossachs National Park, and there were concerns about the environmental impact of mining in the area.
1.a. The stratigraphical setting
One specific feature of the stratigraphical succession merits special attention. The Dalradian Supergroup is divided into five groups (Fig. 1, inset), of which the Argyll Group records the first rifting event to presage the opening of the Iapetus Ocean. This event is recorded in the Easdale Subgroup by a sequence of sulphide-rich sediments, including two sedimentary exhalative sulphide (SedEx) deposits, and serpentine ostioliths and debris (Garson & Plant, Reference Garson and Plant1973; Hall, Reference Hall, Pattrick and Polya1993; Chew, Reference Chew2001). Originally named the Pyrite Belt, this distinctive set of lithologies may be traced from NE Scotland to western Ireland. There are two main mineralized horizons, a lower Ba-rich horizon and an upper Pb–Zn–Cu–Ni-rich horizon (Fortey & Smith, Reference Fortey and Smith1986). The serpentine-bearing unit, characterized by the presence of fuchsite (chrome-rich muscovite), has been interpreted either as an ophiolite body (Garson & Plant, Reference Garson and Plant1973), an intrusion (Hawson & Hall, Reference Hawson and Hall1987) or as a segment of serpentinized upper mantle or O–CT (ocean–continent transition) marking the floor of a transient rift basin (Chew, Reference Chew2001).
The Ag–Au deposits occur in strata across the whole spectrum from Dalradian rocks of upper Proterozoic age, to Silurian rocks of Wenlock age, as shown on Figure 1 (inset). As the orebodies are structurally controlled, this is of lesser importance than the observation that most of them are linked in some way to one particular horizon, the Easdale Subgroup.
2. General features and mineralogy of the Ag–Au deposits
The Grampian Au–Ag mineralization is associated with quartz bodies that range in size from centimetre-thick veins, to quartz lodes up to 20 m thick, and are traceable for up to several kilometres. The thicker veins consist of massive quartz and/or quartz breccia with sulphide-rich, Au–Ag-bearing horizons. Open-space quartz textures have been reported from most of the deposits, and in the area around Tyndrum and Cononish (Tanner, Reference Tanner2012) (Fig. 1), the quartz breccia commonly consists of wallrock fragments encased in sheets of prismatic quartz crystals, forming a cockade texture (nomenclature after Vearncombe, Reference Vearncombe1993). Marginal alteration of the country rock varies from a slight ‘bleaching’, to metre-wide zones of chloritized, sericitized and silicified country rock. In several deposits, the introduction of K-feldspar has caused a pink-red colouration along each vein margin.
The mineralogy of the Grampian Au–Ag deposits in Scotland at Calliachar–Urlar Burn, Tombuie, Tyndrum and Cononish, and at Curraghinalt in Ireland (excluding quartz and gangue minerals), is as follows:
TYPE A: native Au + hessite (Ag2Te) + electrum (Au, Ag) + tetrahedrite (Cu, Sb sulphosalt) + pyrite + chalcopyrite + arsenopyrite + galena + sphalerite.
(In addition, Bi minerals are present at Tombuie (Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997; Corkhill et al. Reference Corkhill, Ixer, Mason, Irving and Pattrick2010) and U minerals at Tyndrum (Pattrick, Reference Pattrick1985).)
This is not a single, consanguineous mineral assemblage, as most of these deposits have had a long and complex history of mineralization. For example, two or three episodes of mineralization have been recognized in several of the deposits (Pattrick, Reference Pattrick1985; Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993; Wilkinson & Johnston, Reference Wilkinson and Johnston1996), as well as up to four generations of quartz growth (Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000).
The deposits in Scotland, at Calliachar–Urlar Burn, Tombuie, Tyndrum and Cononish, formed in Dalradian rocks at shallow levels in the Earth's crust (c. 3–5 km) and over a considerable temperature range (150–450°C) (see Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997 and Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999 in addition to the references above). In Ireland, the mineralogy of the Curraghinalt veins is identical to that of the Scottish veins, but those from Croagh Patrick, Cregganbaun and Bohaun have a more restricted mineralogy, as follows:
TYPE B: native Au (with up to 41 wt% Ag) + pyrite + chalcopyrite ± other minor sulphides.
(In addition, the Bohaun deposit contains sericite, chlorite and haematite.)
The veins at Croagh Patrick, Cregganbaun and Bohaun are hosted by previously folded Ordovician or Silurian rocks. The latter lie unconformably upon Dalradian rocks, or in one instance (Bohaun) unconformably on previously deformed Ordovician rocks. They formed at a slightly higher level in the Earth's crust than the other veins, and at a lower mean temperature (175–320°C).
The Curraghinalt structure in Northern Ireland differs from that of the other deposits, being probably controlled by structures related to late Grampian thrusting of the Dalradian rocks over Palaeozoic rocks to the SE (Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000).
Despite the many fluid inclusion and stable isotope (H, O, S) studies that have been carried out on these deposits, there is a lack of agreement over the source of the fluids which gave rise to them, with Carboniferous basinal fluids (Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011); plutons of ‘Newer Granite’ age (Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993); the upper mantle (Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997); ‘trapped’ regional metamorphic fluids (Craw, Reference Craw1990); and remobilized stratiform mineral deposits (Hall, Reference Hall1985) having been invoked, with meteoric fluids also playing a crucial role (Craw & Chamberlain, Reference Craw and Chamberlain1996).
The aims of this study are to examine the possibility that the structural model proposed for the Cononish deposit (Tanner, Reference Tanner2012) can be applied to other Au–Ag deposits in the Scottish and Irish Caledonides, and, if possible, to determine whether or not these deposits had a common origin.
3. Transcurrent faulting: kinematics and nomenclature
The first step towards understanding the kinematics of faulting and fracturing in rocks was made by Coulomb in 1773, who concluded that when a rock is subjected to compressive stress, a conjugate pair of fractures develop that make an acute angle with the maximum principal stress, σ1. The angle between the shear plane and σ1 is given by θ = 45°±ϕ/2 (the Coulomb criterion), where ϕ is the internal angle of friction of the rock. In this paper, β = ϕ/2, and is close to 30° for the average dry, intact crustal rock (Fig. 2a). Anderson's most important contribution lay in his recognition that the Earth's surface could not support a shear stress and is therefore a principal plane (Anderson, Reference Anderson1951). Faults and fractures that cut this ‘free surface’ are therefore restricted to three possible configurations, with each principal plane being in turn parallel to the Earth's surface.
Figure 2. Pure shear (a) and simple shear (b) models for the development of major fractures in intact rock, according the Coulomb criterion. The two examples are based upon a notional left-lateral ‘Y’ shear (parallel to the margin of the shear zone) striking at 045°. Note that, on (b), the strike of the local secondary principal stress, σ1S, lies 15° anticlockwise to the strike of the regional σ1R (a). (c) A 3-D representation of (b), viewed from the SE.
Anderson (Reference Anderson1951) used the Coulomb criterion to develop a pure shear (or Anderson–Coulomb) model for transcurrent or ‘wrench faulting’ (Fig. 2a), in which a conjugate pair of left- and right-lateral primary shear fractures (LS and RS, respectively), together with a set of extension fractures (XS), develop in response to a far-field stress, σ1R. In explanatory Figure 2a, a hypothetical value for the strike of the left-lateral Y shear of 045° is used as an example, and illustrates that the fracture formed in response to σ1R striking at 015°. Y is used to describe any plane parallel to the margin of the shear zone.
Moody & Hill (Reference Moody and Hill1956) extended this model to include second, third and higher order sets of fractures. They invoked several possible causes for the formation of these fractures, including ‘stress reorientation in one fault block or a block between two parallel faults’ (Moody & Hill, Reference Moody and Hill1956, p. 1212), so anticipating the alternative model. During the same period, laboratory experiments by Riedel (Reference Riedel1929), Cloos (Reference Cloos1955) and later by Tchalenko (Reference Tchalenko1970) provided the basis for a simple shear (or Riedel–Coulomb) model for transcurrent faulting (Fig. 2b, c), which differs from the pure shear model in that σ1R gives rise to a secondary σ1S at 45° to Y (the principal shear plane), which in turn is accompanied (potentially) by a further set of conjugate fractures. Other workers quantified different aspects of these two models: Hubbert & Rubey (Reference Hubbert and Rubey1959) recognized the importance of pore fluids in reducing frictional resistance on a fracture surface; Griffith (Reference Griffith1920) stressed the importance of microcracks in decreasing the strength of rock materials; and Secor (Reference Secor1965) demonstrated that overpressure in pore fluids could, in theory, permit extensional fractures to develop several kilometres below the Earth's surface.
Sibson (Reference Sibson2000) incorporated some of these concepts into the Sibson–Coulomb model (as named here), and demonstrated that β is reduced from 30° to 27° for fractures developed in the presence of a pore fluid. In this model, a pair of Riedel shear fractures, R (synthetic) & R′ (antithetic), form initially at β = 27–30° to either side of σ1S. They are accompanied, in some cases, by extensional fractures (T) parallel to σ1S (Fig. 2b). β is used here for the angle seen in plan view, and β* denotes the calculated maximum angle between the two structural elements. When applying one or other of these models to the Grampian Au–Ag deposits (see Section 7), it is important to appreciate that the regional or far-field stress, σ1R, lies at an angle γ = 15° anticlockwise to the secondary maximum principal stress, σ1S (Fig. 2a, b). Because of a lack of published dip information at individual sites, strike-line diagrams are used in Figure 4 instead of stereograms. Faults and fractures are tacitly assumed, following Anderson (Reference Anderson1951), to be vertical or nearly so. This is supported by the finding that 80% of the faults in the Dalmally–Tyndrum area have a dip of >65° (Tanner, Reference Tanner2012), and most of the large strike-slip faults are, on average, vertical, with the dip direction changing about the vertical along their length.
4. Geodynamic setting
The Grampian orogenic belt, which trends NE–SW across Scotland and the northern part of Ireland (Fig. 1), resulted, in Early Ordovician time, from either an arc–continent collision at the SE margin of Laurentia (Ryan & Dewey, Reference Ryan, Dewey, Brown and Ryan2011 and references therein), or from NW-directed subduction beneath this margin (Rose & Harris, Reference Rose and Harris2000; Tanner, Reference Tanner2007, Reference Tanner, Corfu, Gasser and Chew2013). Details of the structural and regional metamorphic events that occurred in the short interval between 480 and 465 Ma ago (Oliver, Reference Oliver2001; Dewey, Reference Dewey2005) are not relevant here, but are to be found in Stephenson et al. (Reference Stephenson, Mendum, Fettes and Leslie2013) and Tanner et al. (Reference Tanner, Bendall, Pickett, Roberts, Treagus and Stephenson2013a ,Reference Tanner, Thomas, Harris, Gould, Harte, Treagus and Stephenson b ).
Most of the previous models for the plate tectonic evolution of the Grampian Orogeny require that the leading edge of the Laurentian plate was being subducted beneath the Midland Valley Terrane to the SE during the Caradocian (~455 Ma), at which time subduction-flip took place (Atherton & Ghani, Reference Atherton and Ghani2002 and references therein). Whichever model is favoured, it is generally agreed that once subduction-flip had occurred, NW-directed subduction continued until around 426 Ma (Atherton & Ghani, Reference Atherton and Ghani2002; Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009), whence it was halted by slab break-off and the emplacement of large volumes of intermediate and acid magmas, forming the ‘Newer Granites’.
This 426 Ma event is highly significant in the evolution of the Grampian Terrane as it was (1) the time at which large-scale fracturing of the Earth's crust in the area being discussed changed from left-lateral transpression to orogen-parallel left-lateral transtension (Dewey & Strachan, Reference Dewey and Strachan2003); (2) at this time that the Grampian Terrane in Scotland was divided into a series of compartments by NE-trending left-lateral transcurrent faults; (3) coeval with the post-Wenlock/pre-Early Devonian folding event, now recognized as being Scandian (Leake, in press).
The Garabal, Tyndrum, Ericht–Laidon and Etive–Laggan faults (Fig. 3) are broadly parallel to the Great Glen Fault and were probably contemporaneous with it for part of the period between 428 and 400 Ma (Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009). Major faults that developed during the latter period, including the Great Glen Fault, had an extended and complex history, with some movement taking place prior to 426 Ma, followed by sticking of faults, stitching by plutons and subsequent rejuvenation (Treagus, Reference Treagus1991). A number of workers followed Watson (Reference Watson1984) in recognizing the causal relationship between these major faults and the emplacement of the Newer Granite suite (Hutton, Reference Hutton1987; Jacques & Reavy, Reference Jacques and Reavy1994; Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009).
Figure 3. Map of part of the Central Highlands of Scotland showing the major, orogen-parallel, left-lateral faults of the ‘Great Glen’ set, and the negative gravity anomaly (palest grey), probably resulting from a buried extension of the Etive Granite at shallow depth beneath the Cononish area. An important feature of this map is that the long axis of the Lochaber Batholith lies parallel to the strike of the major transcurrent faults. A and B are localities mentioned in the text. The location of the Tyndrum–Dalmally area is shown boxed.
New, high-precision radiometric dating, mainly by Re–Os on molybdenite (Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009; Conliffe et al. Reference Conliffe, Selby, Porter and Feely2010), has shown that the Lochaber Batholith (Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009) (Fig. 3) was emplaced ~430–408 Ma ago. This coincided with the switch from transpression to transtension (Dewey & Strachan, Reference Dewey and Strachan2003), when the compartments bounded by transcurrent faults would have been subjected to ‘pure’ simple shear and developed primary fractures with the angular relationships predicted by the Coulomb–Sibson model. Tanner (Reference Tanner2012) concluded that the formation of the major sets of quartz veins described from the Tyndrum–Dalmally area took place at around 428–426 Ma, following the peak of the regional metamorphism in the Grampian Terrane at 470–465 Ma (Baxter, Ague & Depaolo, Reference Baxter, Ague and Depaolo2002), and the uplift and erosion of several tens of kilometres of Dalradian rocks. This estimate must now be upgraded to c. 410 Ma following recent radiometric dating (see Section 7.b), which shows that Si-rich fluids were being introduced into the Dalradian rocks at the same time as the youngest component of the Lochaber Batholith, the Starav dyke swarm, was being emplaced (Rice et al. Reference Rice, Mark, Selby and Hill2013).
The Connemara Terrane represents part of the western continuation of the Grampian Terrane in western Ireland that has been moved southwards and translated to its present position by left-lateral transcurrent faulting late in the orogeny (Hutton & Dewey, Reference Hutton and Dewey1986) (Fig. 1).
5. Salient features and structural controls of individual gold deposits
In the following accounts, the number in brackets is that assigned to each deposit on Figure 1.
5.a. Calliachar–Urlar Burn (1) and Tombuie (2)
These two small Type A Au–Ag deposits in Perthshire, Scotland, are considered together as they are only 4 km apart and both are hosted by psammites and semi-pelites of the Southern Highland Group (Dalradian) (Fig. 1. inset). They occur as mineralized quartz veins and fault breccias, with an almost identical mineralogy. The Loch Tay Fault is a major left-lateral transcurrent fault (Fig. 3) (Treagus, Reference Treagus1991), which passes 10 km west of the two deposits and trends at ~031° (Figs 3, 4).
Figure 4. Strike-line diagrams showing the orientations of major structures at (a) Calliachar–Urlar Burn and Tombuie; (b) Tyndrum and Cononish; (c) Croagh Patrick; (d) Bohaun. The diagrams show how the mineralized veins (open circles) are located preferentially along secondary Riedel shears (R, R′) and some extensional fractures (T).
The Calliachar–Urlar Burn mineral deposit occurs within the outcrop of the Southern Highland Group. Vertical cross-sections drawn close to the prospect (Treagus, Reference Treagus2000, figs 14, 18) show that the Easdale Subgroup is folded around a stack of recumbent D2 folds, the lowest of which, the Meall Tairneachan Fold, lies above the Boundary Slide. As a result of this folding, the Easdale Subgroup lies some 5 km below the surface in the Callichar Burn area, and a similar situation is found at Tombuie. The stratigraphical context is very similar to that at Cononish, where the Boundary Slide marking the structural top of the inverted Easdale Subgroup is 2–2.5 km below the surface.
The Calliachar–Urlar Burn deposit comprises 14 steeply dipping quartz veins up to 2 m thick with a mean trend of 152° (Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997, fig. 1). The veins are accompanied by extensive alteration of the country rock, seen in the field as sericitization, chloritization and ‘bleaching’. The Urlar Burn Fault trends at 016°, has a left-lateral displacement of 1.5 km and is occupied by a mineralized fault breccia up to 3 m thick. It postdates a felsite dyke and was later intruded by a microdiorite dyke (Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997).
The gold mineralization at Tombuie (Corkhill et al. Reference Corkhill, Ixer, Mason, Irving and Pattrick2010) is contained within quartz-carbonate veins that trend at 130–150°, and are associated with the NNE-trending Tombuie Fault, which shows both lateral and vertical displacements. Mineralization is of a type typically associated with a granitoid-derived hydrothermal fluid (Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997), and shows many geochemical features characteristic of the gold-bearing quartz vein system at Tyndrum.
Mineralization episodes. Only one episode of mineralization has been reported from each deposit.
Interpretation. Faults and fractures at Calliachar–Urlar Burn and Tombuie together form a conjugate pair (Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997) of Riedel shears with β = 22–28°, and R making an angle of 15° with the Loch Tay Fault (Fig. 4a), giving σ1S = 351°. It is deduced that the mineralized veins at these two localities formed in a secondary stress field caused by left-lateral movement on the Loch Tay Fault (Fig. 4a).
5.b. Tyndrum Mine (3)
The Tyndrum base metal deposit lies on the Tyndrum Fault (Figs 3, 5) (Pattrick, Reference Pattrick1985; Pattrick, Boyce & MacIntyre, Reference Pattrick, Boyce and MacIntyre1988; Craw, Reference Craw1990; Craw & Chamberlain, Reference Craw and Chamberlain1996; Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999; Tanner, Reference Tanner2012). It is marked by a shear zone 50–80 m thick of hydrothermally altered wallrock containing metre-scale mineralized quartz veins and the 1–5 m thick ‘Hard Vein’ of massive galena–sphalerite (Pattrick, Reference Pattrick1985; Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993; Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999), and is essentially a Pb–Zn deposit. The fault zone dips at 90±15° SE, trends at 042° and has a left-lateral strike-slip displacement of > 4 km and a dip-slip displacement of 2 km down to the east (Treagus, Reference Treagus1991). Although it is structurally linked with the Cononish Au–Ag deposit (Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999; Tanner & Thomas, Reference Tanner and Thomas2010; Tanner, Reference Tanner2012) (Fig. 4b), it generally yields low Au values.
Figure 5. Geological map of the Tyndrum–Dalmally area, Scotland, showing the outcrop pattern of the Dalradian Coire Daimh Pelite (in grey), which is controlled by early major isoclinal folds, re-folded by the E–W-trending Orchy Dome, and cut by sets of major faults. The figure is based on figure 6 of Tanner (Reference Tanner2012). The earliest faults and veins originated as ‘Y’ shears parallel to the Tyndrum Fault. They were followed by the development of secondary left-lateral Riedel and anti-Riedel shears, which form a conjugate set bisected by a N–S set of extensional fractures (T). The rose diagram (inset) shows the relationship between the total length of the mapped faults versus their strike direction, using a 2° interval. The bold half-arrows give the sense of displacement across the Dalmally–Tyndrum compartment. σ1R is the calculated regional maximum principal stress.
The gold mineralization at Tyndrum is contained within Halliday's Vein that consists of a mineralized quartz breccia that crops out in a stream a short distance NNE of the Tyndrum Pb–Zn mine. It is possibly related, both structurally and temporally, to the Cononish Vein, but no physical connection between them has been found.
Mineralization episodes. Three main episodes have been recognized, namely (1) quartz-bearing Au–Ag–Te mineralization (Pattrick, Boyce & MacIntyre Reference Pattrick, Boyce and MacIntyre1988); (2) base metal Pb–Zn mineralization, which forms the major ore body in the Tyndrum Fault zone (Pattrick, Reference Pattrick1985); and (3) rare U- and Ba-bearing veins (Pattrick, Reference Pattrick1985).
Interpretation. The Tyndrum Pb–Zn deposit lies on a NE-trending primary strike-slip fault (Y) that forms the SE margin to the Tyndrum–Dalmally compartment (Fig. 5) (Tanner & Thomas, Reference Tanner and Thomas2010) and is equated with the Lower Adit Vein at Cononish. Similarly, the Cononish Au-bearing vein is equated with Halliday's Vein at Tyndrum.
5.c. Cononish mineral deposit (4)
The Cononish deposit consists of a steeply dipping to vertical, mineralized quartz vein, up to 6 m thick (Parker, Clifford & Meldrum, Reference Parker, Clifford and Meldrum1989; Earls, Clifford & Meldrum, Reference Earls, Clifford, Meldrum, Bowden, Earls, O'Connor and Pyne1992; Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993; Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999; Dominy et al. Reference Dominy, Platten, Xie and Sangster2009; Hill et al. Reference Hill, Jenkin, Boyce, Sangster, Catterall, Holwell, Naden, Gunn and Rice2011; Tanner, Reference Tanner2012) that cuts the main foliation in the Dalradian rocks at a high angle, and can be traced for >1.5 km.
A late, regional-scale fold, the Orchy Dome (Bailey & Macgregor, Reference Bailey and Macgregor1912; Tanner & Thomas, Reference Tanner and Thomas2010), which is probably of Scandian age (426 Ma, see below), dominates the structure of Dalradian rocks in this area (Fig. 5). This dome spatially controlled the emplacement of a suite of lamprophyre–appinite intrusions, which in the absence of the radiometric ages, were inferred to be 430–428 Ma old (Tanner, Reference Tanner2012). These intrusive bodies, which are associated with explosion breccia pipes, are cut by giant mineralized and non-mineralized quartz-breccia veins (Tanner, Reference Tanner2012) (Fig. 5).
The network of quartz-breccia veins in the Dalmally–Tyndrum area comprises a single, parallel set (Cononish set) and two separate conjugate sets (Figs 4b, 5a). The earlier of these conjugate sets, the proto-Mother Vein and River Vein pair, has β = 30.5° (β* = 31°), close to the value predicted by the Coulomb model for dry rock fractures, and γ = 14°, giving σ1S = 356°. The later conjugate pair (Mother Vein – Barren Vein) has β = 23° (β* = 25°), close to the value predicted by the Coulomb–Sibson model for hydraulic fracturing, and γ = 14°, giving σ1S = 004° (Fig. 4b). Further details of the structural geometry of the mineralized and non-mineralized veins from the Tyndrum–Dalmally area may be found in Tanner & Thomas (Reference Tanner and Thomas2010) and Tanner (Reference Tanner2012).
The major Au–Ag deposit at Cononish is the mineralized portion of the Cononish Vein (Fig. 5), and is parallel to the Tyndrum Fault. Tanner (Reference Tanner2012) concluded that the Cononish Vein set cuts the Orchy Dome, and that both the dome and the two sets of conjugate veins had formed in response to N–S shortening at around 428–426 Ma. The Easdale Subgroup is exposed in the south of the area; it is right-way-up, occupies the top limb of the southward-dipping Beinn Chuirn Anticline and rests upon the Boundary Slide. The now eroded continuation of the subgroup lies above the gold mine, but beneath the surface it passes around the hinge zones of three D2 folds to form a gently inclined, southward-dipping sheet, which lies c. 2.5 km vertically below the mine (Tanner, Reference Tanner2012, fig. 3f; Hill et al. Reference Hill, Jenkin, Holwell, Catterall, Boyce, Mark, Naden, Rice, Jenkin, Lusty, McDonald, Smith, Boyce and Wilkinson2013). As mentioned earlier, this structure is directly analogous in geometry and scale to that of the Calliachar–Urlar Burn deposit.
The giant quartz veins in the Cononish area preserve cockscomb, crustiform and vuggy textures that are characteristic of open-space crystallization at a high level in the Earth's crust (Fig. 6). The minor quartz veins at Cononish are closely similar to those found in the external part of the Bohaun Vein (Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011, fig. 4c, d). Cockade structures, in which country rock xenoliths are mantled by successive layers of quartz prisms orthogonal to the contact (Tanner, Reference Tanner2012, fig. 12) (Fig. 6b), are found in the interior of these veins, and resemble closely the texture found in the outer part of the quartz veins at Rhynie (Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995), with the difference that the layers there are composed of chert. The cores of these major veins commonly consist of a quartz-cemented breccia, as at Cononish (Fig. 6c), Bohaun (Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011, fig. 4b) and Rhynie (Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995, fig. 12).
Figure 6. Open-space quartz fabrics in the giant Mother Vein from the Tyndrum–Dalmally area. They include: (a) breccia fragments (WR) encased within multiple generations of crack-fill (vuggy) veins, from the damage zone of the original fault, giving a crustiform texture (CF); (b) cockade texture (CC) from the interior part of the major vein; and (c) apparently massive quartz (with traces of cockades) containing angular fragments of the wallrock (WR). The scale bar is 5 cm long, ringed in (c).
Mineralization episodes. There are two generally recognized mineralization episodes: (1) Au–Ag mineralization (the ‘A-min’) of Earls, Clifford & Meldrum (Reference Earls, Clifford, Meldrum, Bowden, Earls, O'Connor and Pyne1992) at Cononish, which also occurs in the Beinn Udlaidh Vein; and (2) Pb–Zn mineralization (‘B-min’ or ‘Lower Adit Vein’), which is generally associated with (1). The sequence of mineralizing events is undoubtedly more complex than this, but detailed petrographic, petrological or geochemical studies have yet to be carried out.
Interpretation. The quartz veins occupy fault planes that are identified as secondary R, R′ and T fractures, resulting from the shear stress imposed on a 15 km-wide crustal compartment by displacements on the Tyndrum and Ericht–Laidon major left-lateral transcurrent faults (Tanner, Reference Tanner2012) (Figs 3, 4b). The mineralized Cononish Vein probably formed first, as a Y-fracture parallel to the Tyndrum Fault (Figs 3, 4b), and was succeeded by the development of two sets of conjugate veins.
5.d. Curraghinalt (5)
Au–Ag mineral deposits hosted by quartz veins occur at Curraghinalt and Cavanacaw, in the Sperrin Mountains of County Tyrone, Northern Ireland (Fig. 1). They are occur 10 km apart in small, adjacent inliers of heavily faulted Dalradian rocks belonging to the Argyll Group (Fig. 1, inset), surrounded by strata of Lower Palaeozoic age. They are described separately as they differ in mineralogy, gold type and structure.
At Curraghinalt, the ‘mineral deposit’ occurs in Dalradian rocks correlated with the Easdale Subgroup and thrust SE onto Ordovician pillow lavas, and metasediments belonging to the Tyrone Igneous Complex (Alsop & Hutton, Reference Alsop and Hutton1993). The ‘deposit’ comprises a number of parallel, steeply dipping to vertical, mineralized quartz veins that trend at 080°, are up to 2.7 m thick and may be traced for several kilometres (Earls, Clifford & Meldrum, Reference Earls, Clifford, Meldrum, Bowden, Earls, O'Connor and Pyne1992; Clifford, Reference Clifford, Bowden, Earls, O'Connor and Pyne1992; Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000). They are parallel to near vertical E–W shear zones that are displaced by normal faults (Clifford, Reference Clifford, Bowden, Earls, O'Connor and Pyne1992; Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000).
Mineralization episodes. Detailed studies of the Curraghinalt deposit by Wilkinson et al. (Reference Wilkinson, Boyce, Earls and Fallick1999) and Parnell et al. (Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000) showed that the vein complex developed in four stages, Q1–Q4. The earliest phase (equated with the Grampian event at 470 Ma) and Q3 and Q4 were thought to be caused by the circulation of brines associated with the inversion of the nearby Carboniferous Newtownstewart basin. The timing of the inferred ‘470 Ma event’ has recently been modified by new 40Ar–39Ar ages of 461–449 Ma (Rice et al. Reference Rice, Mark, Selby and Hill2013) from the Curraghinalt veins. They show that the veins formed after the peak of the regional metamorphism at 460 Ma, during the period when the orogen was beginning to collapse (Cooper et al. Reference Cooper, Crowley, Hollis, Noble, Roberts, Chew, Earls, Herrington and Merriman2011).
At Curraghinalt, the major fractures appear to be controlled by a lateral ramp in the footwall to the E–W-trending Omagh Thrust, which underlies the area at a fairly shallow depth (Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000) and is thought to have been active from around 460 Ma.
5.e. Cavanacaw (now Omagh Mine) (6)
The main feature of the Cavanacaw Au–Ag deposit in the Lack Inlier (Dalradian) is the N–S-trending quartz vein complex, the Kearney Structure (Cliff & Wolfenden, Reference Cliff, Wolfenden, Bowden, Earls, O'Connor and Pyne1992). It is a steeply dipping ‘structure’ ~20 m across and 0.85 km long in which 16 named vein structures strike at 140° and dip steeply eastwards (Parker & Pearson, Reference Parker and Pearson2012). The veins locally reach 6.6 m in width and are accompanied by a selvedge several metres thick in which the wallrock has been sericitized and bleached. The Kearney vein structure may be interpreted as a T-fracture associated with the inferred N–S σ1R. No mineralogical or geochemical studies have been carried out but the gold contains up to 2.5 parts of Ag to 1 part of Au and is associated with pyrite, chalcopyrite, arsenopyrite and galena (Parker & Pearson, Reference Parker and Pearson2012).
The Cavanacaw deposit is located in rocks originally thought to belong to the Southern Highland Group but have recently been correlated with the Easdale Subgroup (McFarlane, Cooper & Chew, Reference McFarlane, Cooper and Chew2009) that includes detrital serpentinite containing fuchsite, as in Scotland. The mineralogy of this deposit is not known in sufficient detail for it to be definitely assigned to Type A.
5.f. Croagh Patrick (7), South Mayo trough
This deposit, previously referred to as the Lecanvey Au–Ag prospect, is located 1–2 km west of Croagh Patrick (Fig. 1) (Aherne, Reynolds & Burke, Reference Aherne, Reynolds, Burke, Bowden, Earls, O'Connor and Pyne1992; Wilkinson & Johnston, Reference Wilkinson and Johnston1996). Mineralized quartz veins cut the polyphase-deformed Silurian rocks that lie unconformably on the Ordovician Deer Park ophiolite in the footwall to the E–W-trending Omagh Thrust. In the Brocaigh zone (northern part of the prospect), gold mineralization is found in quartz veins that may be divided into three main sets: NNW–SSE- and NE–SW-trending veins that occupy shear zones, and E–W veins orientated parallel to S-directed thrusts.
The veins are generally 0.2–2.0 m thick (Johnston & McCaffrey, Reference Johnston and McCaffrey1996) and where they occupy shear zones, they develop a set of en échelon ‘feather veins’ at ~30° to the strike of the zone (Fig. 4c) cut by a through-going thicker quartz vein (Fig. 4c). The large quartz veins have sigmoidal shapes, with the NE-trending set passing into the E–W-trending veins, suggesting contemporaneity. The quartz veins cut the Croagh Patrick Anticline, which is an isoclinal fold that verges and faces south, and is considered to have formed significantly earlier than the faulting and vein formation at this locality (Aherne, Reynolds & Burke, Reference Aherne, Reynolds, Burke, Bowden, Earls, O'Connor and Pyne1992).
Mineralization episodes. Multiple growth episodes were reported by Johnston & McCaffrey (Reference Johnston and McCaffrey1996), with five phases of quartz growth and fabric development being recognized.
Interpretation. The NNW- and NE-trending sets of quartz veins have an orientation, shear sense and β = 30° (Fig. 4c), compatible with their being R and R′ shears (Aherne, Reynolds & Burke, Reference Aherne, Reynolds, Burke, Bowden, Earls, O'Connor and Pyne1992) that formed in response to σ1N trending at 011°. Johnston (Reference Johnston, Bowden, Earls, O'Connor and Pyne1992) attributed this system of shear fractures to late Caledonian left-lateral transpression across the Clew Bay – Fair Head Line (Fig. 1), but they could equally well be primary fractures, with the RS–LS fractures having formed synchronously with the E–W-trending, S-directed thrusts, as shown on Figure 4c.
5.g. Cregganbaun (8), South Mayo trough
This vein deposit is located in County Mayo (Fig. 1). It features a 10–14 m thick mineralized zone with quartz veins that trend E–W, but there is little published information available for this prospect. The Au-bearing quartz lode lies within the Cregganbaun Shear Zone that cuts tuffaceous Ordovician rocks belonging to the Sheeffry Formation. It can be traced along strike for 33 km and contains the assemblage quartz, gold, pyrite, tetrahedrite, sphalerite and galena (Morris et al. Reference Morris, Long, McConnell and Archer1995).
5.h. Bohaun (9), Connemara
This Au–Ag deposit occurs in NE Connemara and lies on the Bohaun Fault (Figs 1, 7), which cuts folded Silurian rocks (Leake & Tanner, Reference Leake and Tanner1994; Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011). It belongs to a network of conjugate transcurrent faults that may be separated into two sets trending at approximately 030° and 330°. The Bohaun Fault displaces the Lugnabrick–Benlugmore Syncline, by a left-lateral displacement, accompanied by a small downthrow to the west (B. E. Leake, pers. comm).
Figure 7. Geology of the Bohaun area, Connemara, showing sets of conjugate fractures (R and R′) formed as secondary structures related to left-lateral movement on the Maam Valley Fault Zone (Y). This figure is based on Leake & Tanner (Reference Leake and Tanner1994, fig. 13).
The Silurian rocks lie unconformably across the faulted contact between deformed Ordovician rocks belonging to the South Mayo Trough and the Dalradian basement (Leake & Tanner, Reference Leake and Tanner1994, fig. 13). This contact is a continuation of the Doon Rock Fault (Ryan & Archer, Reference Ryan and Archer1980), and represents the terrane boundary along which the Connemara block had moved laterally to its present position south of the Grampian Terrane (Figs 1, 7). At Lough Nafooey, it has a downthrow to the north of ~15 km but is concealed beneath Silurian rocks in the study area. The Maam Valley Fault Zone is a major right-lateral transcurrent fault (Fig. 7) that may be traced for 40 km to the NW across part of the South Mayo Trough (Aherne, Reynolds & Burke, Reference Aherne, Reynolds, Burke, Bowden, Earls, O'Connor and Pyne1992, fig. 1).
The Bohaun mineralized zone is located just south of Benlugmore (Fig. 7), where it is marked by a unit of brecciated and silicified country rock up to 50 m wide that can be traced for 1.6 km. The fault zone trends at 015° and comprises a network of metre-scale veins and breccia units up to 50 m thick with a mean N–S orientation (Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011, fig. 3). The range of quartz fabrics seen at outcrop, mainly open-space (Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011, figs 4, 5), matches that in the Tyndrum–Dalmally area (Tanner, Reference Tanner2012, figs 6, 10, 12).
Mineralization episodes. The mineralization is rather sparse and quite different to that of the Scottish deposits. Quartz shows multiple growth stages, with evidence that two fluids of different salinity were involved.
Interpretation. The Bohaun deposit has many of the structural features seen in the Cononish area, and the structure may be interpreted as follows. The Maam Valley Fault Zone is a primary right-lateral transcurrent fault (Y on Fig. 4d) (e.g. the mirror image of the Tyndrum Fault), with the conjugate faults being secondary Riedel shears (R, R′) that cut and displace E–W-trending folds. The Maam Valley Fault has a sinuous profile in plan view, and the orientation shown in Figure 4d is a mean value. The Bohaun Fault is interpreted as a pure tensional fracture (T) containing σ1S = 000°; the upright folds in the Dalradian rocks could have formed first, locked up and been transected by the sets of transcurrent faults (?rejuvenated basement faults) that formed in the same stress field. The Bohaun deposit was considered ‘atypical’ when compared with other Grampian Au–Ag deposits (Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011); however, it is seen here as the high-level equivalent to the Tyndrum and Cononish deposits, and is a member of the same continuum.
Conjugate sets of faults are developed throughout the Connemara Dalradian inlier (Leake & Tanner, Reference Leake and Tanner1994, pl. 1) and are related to a N–S σ1 direction. Quartz breccias associated with these faults are almost unknown, but a striking exception is seen 25 km west of Bohaun where a 30 m thick fault breccia occupying the Bengower Fault (Leake & Tanner, Reference Leake and Tanner1994, pl. 1) forms a wall-like feature close to the summit of Bengower at [L 784 507]. The structural position of this open-space quartz breccia is significant as it occurs on a releasing-bend where the NW-trending, right-lateral transcurrent fault turns northwards to become N–S-trending. Minor quartz veins and contacts associated with the major breccia strike between 177 and 189° and dip 75–86°W. The breccia body has the geometry of a T-fracture, and is seen as an exploration target.
6. Related igneous features, including a putative hot-spring
6.a. Mineralized igneous intrusions
In the Grampian Terrane, there are two examples of a possible link between plutons emplaced at ~425 Ma and the mineralizing fluids:
(1) The Lagalochan ((10) on Fig. 1) and Kilmelford (11) granitoids acted as feeders for the Lorn Lavas and were subsequently truncated by the Ericht–Laidon Fault. The Kilmelford Igneous Complex gave a Re–Os (molybdenite) age of 425.8±1.7 Ma (Conliffe et al. Reference Conliffe, Selby, Porter and Feely2010), compared with a U–Pb (zircon) age of 425±0.7 Ma for the Lorn Lavas (Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009). The hydrothermally altered rocks in the sub-volcanic centre at Lagalochan contain pyrite, chalcopyrite, arsenopyrite, sphalerite, galena and molybdenite, and show high Au and Ag values (Zhou, Reference Zhou1988). The complex was affected by left-lateral transcurrent fault movement on the Ericht–Laidon Fault (Fig. 1), with some sulphide mineralization developed on subsidiary faults.
The porphyry-stock-related and plutonic-hosted mineralization associated with these two granitoids was studied by Lowry et al. (Reference Lowry, Boyce, Fallick and Stephens1995), who concluded that it took place at a depth of ~3 km, similar to that postulated for the Tyndrum and Cononish vein deposits (Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993; Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999). Although the mineralizing fluid did not generally include Au or Ag, it is significant that it had a composition closely resembling that responsible for the formation of the vein systems in the Dalradian metasediments of the Tyndrum–Dalmally area (Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993).
(2) The Ratagain Pluton ((12) on Fig. 1) cuts Moine rocks belonging to the Northern Highland Terrane (Fig. 1), but is included here as it has a petrological link with the Grampian deposits. It consists mainly of quartz-monzonites and diorites (see Alderton, Reference Alderton1988 and references therein). Prior to its complete crystallization it was deformed by left-lateral shear on the Strathconon Fault, a major NE-trending transcurrent fault belonging to the Great Glen set (Fig. 1) (Hutton & McErlean, Reference Hutton and McErlean1991). The granite is of the same age (425±3 Ma on baddeleyite (ZrO2); Rogers & Dunning, Reference Rogers and Dunning1991) as the Lagalochan & Kilmelford granitoids, and is cut by Au- and Ag-bearing quartz veins that have a full suite of minerals characteristic of the Au–Ag deposits, namely native gold, hessite, electrum, tetrahedrite, pyrite, chalcopyrite, arsenopyrite, galena, sphalerite and molybdenite (Alderton, Reference Alderton1988, fig. 2). These late veins at Ratagain have the same distinctive red (K-feldspar) selvedges as found associated with the Cononish and Island Vein sets in the Dalmally–Tyndrum area; they may be segregated from the granite magma or be infills of veins fed by the same fluids as are found in the Grampian Terrane.
(3) Gold mineralization also occurs within the shear zone marking the course of the Tyndrum Fault at Inverchorachan south of the mine (Fig. 3) (Scotgold, 2007), and may have been deposited from fluids generated by the emplacement of the Garabal Hill – Glen Fyne Complex (Rogers & Dunning, Reference Rogers and Dunning1991), which is cut by the fault. This complex is the oldest member of the ‘Newer Granites’ and gave a (revised) U–Pb (zircon) age of 428±9.8 Ma (Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009).
These examples provide some evidence to suggest that igneous fluids were involved to a minor extent in the production of the mineralizing fluids, but they were probably more valuable for providing thermal energy to the system rather than for their gold content.
6.b. Relationship of the Rhynie cherts (13) to the Cononish Vein
The Rhynie cherts occur within a small outlier of Lower Devonian rocks in NE Scotland, west of Aberdeen (Fig. 1). They were laid down in a half-graben floored by Dalradian rocks belonging to the Southern Highland Group, together with igneous rocks (mainly norite) and are part of a sedimentary sequence that includes andesitic lavas and tuffs (Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995). The cherts are famous for their excellent preservation of the earliest land plants and arthropods, and were the product of the oldest known Au-bearing hot-springs (Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995). Together with the fault breccias, they are also of industrial interest, as they contain abnormally high concentrations of Au, As, Sb, W, Mo and Hg (Rice & Trewin, Reference Rice and Trewin1988; Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995). The mineralization is concentrated in fault breccias associated with steep to vertical N–NE-trending basin-bounding faults that delimit the west side of the basin. These breccias show open-space quartz textures and other features indicating that the growth of quartz took place at, or very close to, the surface of the Earth (Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995).
The most striking feature common to both the Cononish Vein and the Rhynie cherts is the presence of red, often vividly coloured, K-feldspar-rich selvedges to the Au-bearing chert veins. This form of hydrothermal wallrock alteration is also associated with a number of quartz veins, including the River Vein set 7 km NW of Cononish (Fig. 5) and the Au-bearing veins that cut the Ratagain Granite.
It was suggested by Rice, Trewin & Anderson (Reference Rice, Trewin and Anderson2002) that Rhynie-type cherts may have been deposited from the same fluids that, deeper in the crust, were responsible for the formation of mineral deposits similar to those at Cononish and Tyndrum. Initially, dating of the cherts by 40Ar–39Ar at 396±12 Ma (Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995) and of K-feldspar from the metasomatized country rock adjacent to the Cononish Vein by K–Ar at 410±14 Ma, made this unlikely (Tanner, Reference Tanner2012).
Further work led to an ongoing dispute over the timing of Rhynie hot-spring activity, which has arisen mainly from the use of different chronometers, between Mark, Rice & Trewin (Reference Mark, Rice and Trewin2013), who favour 407.6±2.2 Ma (40Ar–39Ar), and Parry et al. (Reference Parry, Noble, Crowley and Wellman2011), who determined an age of 411.5±1.3 Ma (206Pb–238U, zircon). 40Ar–39Ar dating of hydrothermal K-feldspar from a feeder vein to the Rhynie system gave a (recalculated) age of 407.1±2.1 Ma (Mark et al. Reference Mark, Rice, Fallick, Trewin, Lee, Boyce and Lee2011). This latest age determination is within error of new 40Ar–39Ar determinations (on metasomatic K-feldspar from the altered wallrock at Cononish) of 407±1 and 408±2 Ma (Rice et al. Reference Rice, Mark, Selby and Hill2013) (Table 1).
Table 1. Radiometric ages of selected igneous intrusions, lavas and Au-bearing mineral deposits in the Grampian Terrane
These results show that at least some of the quartz veins in the Cononish–Tyndrum area were contemporaneous with the final magma pulse in the Lochaber Batholith (Table 1), and could have been linked to Rhynie-type hot-springs at the extant surface.
7. Structural synthesis
At six of the nine localities studied here (Fig. 4a, b, c, d), the attitudes of the mineralized veins and fractures, and in some cases their geometric relationship to folds and thrusts, can be provisionally interpreted in terms of a N–S secondary maximum principal palaeostress direction (σ1S) that trends at 348–011°. The geometry of these features at one of the remaining localities, Curraghinalt, appears to have developed as a result of space problems caused by an oblique footwall ramp on the S-directed Omagh Thrust (Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000), but there is insufficient published information with which to analyse the structures at Cavanacaw and Cregganbaun.
At most of the sites (Fig. 4a, b, d), the fractures are conjugate Riedel shears, having a conventional angular relationship to a master transcurrent fault (simple shear model). Of the remaining sites, only at Croagh Patrick (Fig. 4c) may the pattern be explained in terms of primary fractures (the pure shear model). In this model, the secondary σ1S is 15° anticlockwise to σ1R. When this correction is applied to the results in Figure 4a, b and d, the strike of σ1R for Croagh Patrick at 011° compares well with the corrected values of 006° for Calliachar–Urlar Burn and Tombuie, 019° for Tyndrum and Cononish, and 015° for Bohaun, giving a mean value for these sites of σ1R = 013°.
The preliminary conclusion drawn from this analysis is that the regional maximum principal stress (σ1N) was consistently orientated approximately N–S in all of the locations described here in Scotland and Ireland for which adequate structural data are available. This stress pattern relates to the final collision of Avalonia with Laurentia.
7.a. Connectivity of the fault system
In models such as the one presented here, connectivity between the master fault (Y), which potentially scavenges fluids from mantle depths, and second-order faults (R, R′, T), in which they crystallized, is an essential requirement, as recognized by Neumayr, Hagemann & Couture (Reference Neumayr, Hagemann and Couture2000) and Groves et al. (Reference Groves, Goldfarb, Robert and Hart2003). Indeed, the latter stated that ‘Gold deposits are commonly, although not exclusively, in second- or third-order structures that were probably connected to the first-order structures at the time of gold mineralization’ (Groves et al. Reference Groves, Goldfarb, Robert and Hart2003, p. 9). In the Cononish area, the main mineral deposit and its satellites lie on the LS set of Y shears (Cononish Vein) parallel to the Tyndrum Fault, which is host to the other economic Au–Ag deposit in the area (Figs 3, 4b). The conjugate sets of Riedel (R, R′) and extensional (T) faults are interconnected, but are not seen to be joined to the Y structures at the present level of erosion.
7.b. Dating of events
The previous estimate for the age of the quartz veins in the Dalmally–Tyndrum area (Tanner, Reference Tanner2012) was based on the field relationships between the Ericht–Laidon Fault and the lavas and plutons of the Lochaber Batholith, recently dated by Neilson, Kokelaar & Crowley (Reference Neilson, Kokelaar and Crowley2009). If the thesis that the veins formed in Riedel shears in response to strike-slip movements on the Tyndrum and Ericht–Laidon fault-pair (Tanner, Reference Tanner2012) is correct, then the timing of the displacements on these faults will constrain the age of the fractures, and hence of the veins. As there is no indisputable field evidence that these fault movements continued after 426 Ma, the age of the veins was concluded to be between 428 Ma (the inferred age of the appinite–lamprophyre suite) and 426 Ma.
However, this suggestion has now to be reconsidered in the light of recent radiometric dating. The K–Ar age of 410±14 Ma on K–feldspar from metasomatized country rock adjacent to the Cononish Vein (Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999) has recently been supported by Rice et al. (Reference Rice, Mark, Selby and Hill2013), who obtained an 40Ar–39Ar age of 407.1±2.1 Ma by dating of similar material from Cononish. Whilst caution has to be taken in interpreting this result, as there are at least two major mineralizing events in the Cononish mineral deposit, the main vein-forming event in this area would therefore seem to be coincident with the final magmatic events in the Lochaber Batholith. This new age determination, together with the post-Early Devonian (< 417 Ma) age of the fold structures cut by the veins at Croagh Patrick and Bohaun, suggests that the whole suite of Au–Ag deposits in the Grampian Terrane, apart from those at Curraghinalt and Cavanacaw, may be of Lower Devonian (c. 408 Ma) age.
8. Discussion
This section presents a critical assessment of progress to date in determining the sources of the fluids that gave rise to the Grampian Au–Ag deposits and a critique of the proposal that Riedel fractures (R, R′) are an important feature that controls the distribution and geometry of the mineralized quartz veins.
8.a. Origin of the fluids
The Grampian Au–Ag deposits form a coherent group, both in composition and in the local and regional structural controls on their emplacement, having a spatial distribution that is independent of stratigraphy (Fig. 1, inset). The deposits in Scotland, together with that at Curraghinalt in Northern Ireland, have overall an identical mineralogy (albeit the product of several mineralization episodes), including unusual minerals such as hessite, electrum and tetrahedrite. The mineralized quartz veins crystallized at moderate temperatures (< 450°C), and display open-space quartz textures betokening their shallow depth of formation (5–3 km). Microchemical characterization of detrital gold grains, sampled from small streams or burns in the vicinity of each deposit, supports the strong inter-relationship between the grains types from the Cregganbaun, Croagh Patrick, Cononish and Calliachar–Urlar Burn deposits (Chapman et al. Reference Chapman, Leake, Moles, Earls, Cooper, Harrington and Berzin2000). Gold from the Grampian Terrane commonly contains >20% Ag, in addition to inclusions of silver and bismuth tellurides that distinguish it from gold derived from deposits in neighbouring terranes. Despite their similarities in composition and field occurrence, no single model has been proposed that applies to all of the Grampian deposits, and many fundamental questions remain unanswered, such as:
(1) What was the nature of the source that: (a) supplied the extremely large volume of silica-rich fluid required to form the quartz veins, the largest of which are up to 7 km long and up to 20 m thick (Tanner, Reference Tanner, Corfu, Gasser and Chew2013)? Was this fluid injected into active fractures as countless pulses, or as mobile hydrofractures (Bons, Reference Bons2001)? (b) generated sufficient fluid to scavenge, and transport, in solution, an estimated Joint Ore Reserves Committee (JORC) reserve of 169200 oz Au, at a cut-off grade of 3.1g/t and 631300 oz Ag, to form the Cononish deposit? (c) provided fluid hot enough to dissolve pre-existing Au, Ag and metal sulphides and redeposit them at a shallower depth (3–5 km) in the Earth's crust, at temperatures of 300–400°C? (d) supplied an additional aliquot of fluid rich in base metals, to form a separate Pb–Zn horizon spatially linked to the gold-bearing veins?
(2) How could regional metamorphic dehydration reactions provide the fluids necessary to create the Cononish orebody, as well as other gold-bearing veins, in the Grampian Terrane, when there is a mismatch between the time at which the fluid was produced (460–400 Ma ago) and the actual age of the quartz veins (467–408 Ma ago), a discrepancy of up to 60 Ma? i.e. how to reconcile the inferred sequence of events with the available radiometric ages.
Before addressing these questions, the various types of fluid that could be involved are listed, as follows:
(1) Sedimentary: pore fluid in chemical equilibrium with the host rock minerals, mobilized by the first deformation to affect the sediments. (In rocks subsequently affected by orogenesis, it is not possible at present to separate this fluid from that resulting from the early dehydration reactions, so it is not mentioned again in this account.)
(2) Basinal: brines arising from the development of an overlying or adjoining sedimentary basin.
(3) Metamorphic: fluids resulting from dehydration reactions that took place during regional metamorphism of the sedimentary prism, particularly at the greenschist to amphibolite-facies transition.
(4) Igneous: fluids arising from buried granitoid batholiths or plutons undergoing crystallization.
(5) Mantle: fluids from the upper mantle.
(6) Meteoric: surface waters driven by the hydrostatic head resulting from a juvenile topography developed during post-orogenic uplift of the orogen.
8.a.1. Granitoid intrusions as a source of fluids
A widely supported hypothesis for the origin of the Au–Ag mineral deposits in the Grampian Terrane is that the Si-rich mineralizing fluids originated from the crystallization of one or more igneous plutons at a shallow depth beneath the ground surface at that time (Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993; Lowry et al. Reference Lowry, Boyce, Fallick and Stephens1995; Tanner, Reference Tanner, Corfu, Gasser and Chew2013). The negative gravity anomaly in the Tyndrum–Dalmally area (Fig. 3) (Cononish deposit) might be taken to indicate the presence of a buried granitoid pluton 5–10 km below the contemporary ground surface. If confirmed, it would be a rare example of the preservation of the fractured and mineralized country rocks forming the roof of a pluton (Leake, Reference Leake1990), a prospective but very rare situation. However, the Cononish–Tyndrum example may be a special case, as granitoid or similar plutons are not obviously associated with the other Au–Ag deposits in the Grampian Terrane, and some other fluid source must have been involved. More generally, it has been found that although circumstantial evidence supports the notion that buried igneous intrusions are responsible for supplying the mineralizing fluids in such cases, it is difficult to find incontrovertible evidence that this has occurred (Groves et al. Reference Groves, Goldfarb, Robert and Hart2003, p. 10).
8.a.2. A mantle contribution?
The inferred compositions of fluids responsible for forming the Tombuie, Tyndrum–Cononish and Curraghinalt deposits may indicate (Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997) that members of the lamprophyre–appinite suite made a minor contribution to the Au-bearing fluids in the Grampian Terrane. A minor intrusion belonging to this suite, in the Grampian Terrane of County Tyrone, has yielded a U–Pb (zircon) age of 426.7±0.8 Ma (Cooper et al. Reference Cooper, Crowley, Hollis, Noble and Henney2013) (Table 1), with the magma being derived from the upper mantle. Other evidence for possible mantle involvement includes:
(1) The presence of Pt-bearing lamprophyres and picrites in the Tyndrum–Dalmally area that were derived from a depth of ~100 km (Atherton & Ghani, Reference Atherton and Ghani2002); and the occurrence of picrite in the Appinite Complex in Glen Orchy (Kynaston & Hill, Reference Kynaston and Hill1908; Atherton & Ghani, Reference Atherton and Ghani2002), together with Pt-group minerals (D. Catterall, pers. comm., 2011);
(2) The occurrence of Pt-group minerals in the Sron Garbh Appinite Complex north of Tyndrum (Scotgold, 2012; Tanner, Reference Tanner2012).
8.a.3. Basinal brines and the base metal mineralization
There is evidence from fluid inclusion studies that basinal brines were responsible for introducing Pb–Zn mineralization into some of the deposits. For example, the latest generation of quartz-rich (Q4) fluids in the Curraghinalt deposit were thought to have been basinal brines possibly derived from the Carboniferous Newtownstewart basin (Wilkinson et al. Reference Wilkinson, Boyce, Earls and Fallick1999; Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000). The isotopic signature of the Bohaun deposit was considered by Lusty et al. (Reference Lusty, Naden, Bouch, McKervey and McFarlane2011) to reflect the importance at Bohaun of basinal fluid (Pb–Zn) of probable Carboniferous age.
8.a.4. Significance of meteoric fluids
These surface-derived fluids have an important role in the formation of Au–Ag deposits, not so much for their contribution to the metal content of the ore, but in reacting with warmer, mineralized and Si-rich fluids from below and establishing oxidizing conditions in which gold compounds are unstable, leading to precipitation of the gold. Evidence of this is seen in the presence of haematite in Au-bearing veins described by Craw & Chamberlain (Reference Craw and Chamberlain1996), and in the Bohaun deposit by Lusty et al. (Reference Lusty, Naden, Bouch, McKervey and McFarlane2011). It is significant that most of the Grampian Au–Ag deposits crystallized at depths of 3–5 km, equal to the maximum depth to which meteoric fluid can penetrate.
8.a.5. A regional (Barrovian) metamorphic origin?
Craw (Reference Craw1990) first recognized that leaching of certain elements had accompanied the formation of late (D3/D4) quartz veins in the rocks of the Southern Highland Group (Dalradian) exposed in road cuttings and natural exposures in an area 25–35 km south of Tyndrum and west of Loch Lomond (Fig. 3). The quartz veins crystallized at c. 350°C at a depth of 8–10 km and contain the assemblage: chlorite + pyrite + pyrrhotite + chalcopyrite ± sphalerite ± galena ± cubanite ± native gold. The assemblage was thought to have formed at a late stage in the uplift of the Grampian Orogenic Belt. It closely resembles that of the Type B deposits described in this paper, but it seems unlikely that the same fluid that sourced these veins was responsible for the development of the major Tyndrum and Cononish veins, as the regional metamorphism reached a peak at 475–465 Ma and the veins formed between 430–407 Ma (see Tanner, Reference Tanner, Corfu, Gasser and Chew2013, p. 71).
At least some of the veins studied by Craw (Reference Craw1990) are more likely to be of D2 rather than D4 age (Tanner & Thomas, Reference Tanner and Thomas2010), but this does not resolve the universal problem of how such fluids may be retained in the crust for tens of millions of years (for review, see Yardley & Cleverley, Reference Yardley, Cleverley, Jenkin, Lusty, McDonald, Smith, Boyce and Wilkinson2013). To the contrary, rapid uplift in the later stages of an orogenic event, as documented by Craw et al. (Reference Craw, Upton, Yu, Horton and Chen2010) from the Taiwan tectonic belt, favours the production of metal-bearing metamorphic fluids and the consequent formation of orebodies. Of relevance to the present problem, Yardley & Cleverley (Reference Yardley, Cleverley, Jenkin, Lusty, McDonald, Smith, Boyce and Wilkinson2013) make the prescient statement that ‘perhaps late orogenic metamorphism can be driven by fluids from late orogenic mineralization . . .’
8.a.6. A new approach: the Otago model
Pitcairn et al. (Reference Pitcairn, Teagle, Craw, Olivo, Kerrich and Brewer2006) tested the hypothesis that intergranular fluids resulting from metamorphic dehydration reactions could, while migrating upwards to the free surface, leach sufficient noble and other elements from the country rocks to form an economically viable ore deposit at a higher level in the crust. This study was based on the Otago and Alpine Schists on South Island, New Zealand, a group of monotonous metagreywackes and metabasalts of Mesozoic age that had been subjected to deformation and progressive regional metamorphism. The same group of schists was sampled (a) in their non-metamorphosed state; (b) where they had been affected by greenschist metamorphism; and (c) by amphibolite-facies metamorphism. From a statistical analysis of trace- and major-element analyses of these three groups of rocks, Pitcairn et al. (Reference Pitcairn, Teagle, Craw, Olivo, Kerrich and Brewer2006) demonstrated that certain elements were largely, if not wholly, removed from the greenschists during their conversion, by devolitization, to amphibolite-facies rocks. In order of decreasing depletion, they were: Au, Ag, As, Sb, Hg, Mo and W. By the time that the regional metamorphism had reached the garnet zone, pyrite, cobaltite, galena and sphalerite had also been removed from the system and, of the sulphur-bearing minerals, only pyrrhotite remained. Their most significant finding was that the minerals that constitute the Macraes orebody (the largest Au-bearing deposit in the Otago Schists) were identical to those that had been extracted from the greenschist-facies rocks and were present in the same relative proportions as in the host rocks. Peak metamorphic conditions in the amphibolite facies of the Otago Schists (Pitcairn et al. Reference Pitcairn, Teagle, Craw, Olivo, Kerrich and Brewer2006) were T = 650°C and P = 10 kb, identical to those in the Grampian Terrane (Vorhies & Ague, Reference Vorhies and Ague2011).
Individual isotopic studies may be selected from work done on deposits in the Grampian gold province to show that crystallizing granitoid intrusions had probably contributed to the mineralizing fluids. However, the remarkable similarities in mineralogy and other features shown by the widely scattered Au–Ag Grampian deposits are an indication that the scale of the source region is greater than that of an individual hidden pluton.
Phillips & Powell (Reference Phillips and Powell2010) presented a genetic model for the metamorphic devolatilization process that occurs at the greenschist to amphibolite-facies transition, referred to here as the ‘dehydration front’. These reactions take place at 440–520°C and yield 2–5 wt% aqueous fluid, and Phillips & Powell (Reference Phillips and Powell2010) emphasized the importance of tectonic strain in liberating the intergranular fluid from the reaction site (Cox, Reference Cox2005), enabling it to migrate towards the ground surface. Cu, Zn and Pb were not leached out during this process and remained in situ, available to be scavenged by a later hotter fluid from an (?)igneous source. This observation may explain the dichotomy between Au–Ag–Te deposits (which formed first) and Pb–Zn–Cu deposits, which generally accompany them and invariably developed at a later stage. Although there are problems with the Otago model, as recognized by Pitcairn et al. (Reference Pitcairn, Teagle, Craw, Olivo, Kerrich and Brewer2006, p. 1541) and later by Phillips & Powell (Reference Phillips and Powell2010), it provides a quantitative testable framework for explaining the origin of the Grampian deposits.
Hill et al. (Reference Hill, Jenkin, Holwell, Catterall, Boyce, Mark, Naden, Rice, Jenkin, Lusty, McDonald, Smith, Boyce and Wilkinson2013) made a comprehensive study of the S-isotopes in the Cononish Au–Ag deposit, to determine its origin. Like Gallacher (in Hall, Reference Hall, Pattrick and Polya1993, p. 89), they speculated that the ore minerals in the Cononish deposit were probably derived by remobilization of elements from an underlying SedEx deposit in the Easdale Subgroup, which is associated with most of the deposits. Hill et al. (Reference Hill, Jenkin, Holwell, Catterall, Boyce, Mark, Naden, Rice, Jenkin, Lusty, McDonald, Smith, Boyce and Wilkinson2013) found that much of the local Dalradian succession (Fortey & Smith, Reference Fortey and Smith1986) that lies structurally beneath the Cononish mineral deposit was deficient in sulphides and concluded that only the constituent Easdale Subgroup could have contributed sulphides in sufficient quantity to form the overlying Cononish deposit. They concluded that, on balance, the S-isotope data showed that the Cononish deposit was predominantly of sedimentary origin, but had a significant igneous-derived content. Although it does not affect their conclusion, it is worth noting that SedEx deposits can be derived from reduced seawater sulphates, rather than from magmatic or metamorphic S (Chang, Large & Maslennikov, Reference Chang, Large and Maslennikov2008).
The paucity of Au, Ag and Te in the local Easdale Subgroup could be that, because they are at garnet grade (amphibolite facies), their content of Au and other metals will, according to the Otago model (Pitcairn et al. Reference Pitcairn, Teagle, Craw, Olivo, Kerrich and Brewer2006), have already been depleted of ore minerals by devolatilization. This possibility carries the implication that the pyrite δ34S values may have been modified during the scavenging process, before the second leaching by hotter igneous fluids. These provisos highlight the problem of determining the source of a sulphide deposit using a single isotope system (Phillips & Powell, Reference Phillips and Powell2010, p. 690), especially where there is a large variation in δ34S values.
In an attempt to explain how Grampian deposits of differing ages, but with similar characteristics, formed, the early history of the terrane is now outlined, with two hypothetical examples to indicate briefly the different routes that could have been taken by the fluids depositing (a) the Curraghinalt orebody at 460 Ma, and (b) the Cononish orebody at 408 Ma.
8.a.7. General model for the Grampian gold deposits
During the early stages of the Grampian deformation, post-D1, pre-D2 planar, bedding-parallel quartz veins formed under greenschist-facies conditions, and the sedimentary fluids sensu stricto held in the pore space began to migrate towards the surface as the temperature of the fluid increased. The early veins were then folded and boudinaged, and a new generation of veins formed during D2 (Tanner & Thomas, Reference Tanner and Thomas2010), accompanied by the widespread development of spaced pressure-solution cleavages. Much of the quartz was probably redistributed within local closed systems, but some may have been incorporated into the intergranular fluid and mixed with metamorphic fluids arising from the dehydration reactions. Noble metals, and elements such as As, were scavenged in increasing amounts as the temperature of the fluid increased. Some of the veins examined by Craw (Reference Craw1990) possibly formed at this time. Before the onset of D3, the D1/D2 fabric was flat-lying and the dehydration front probably had a similar orientation, with the greenschist-facies rocks overlying those of amphibolite facies (Tanner, Reference Tanner, Corfu, Gasser and Chew2013, fig. 5). Quartz veins of all sizes developed in the vanguard of the dehydration front.
8.a.7.a. Formation of the Curraghinalt deposit
At Curraghinalt, the Omagh Thrust carried Dalradian rocks SE over the Tyrone Igneous Complex (Alsop & Hutton, Reference Alsop and Hutton1993); quartz veins developed in active fractures that developed to accommodate a ramp in the footwall of the major thrust. The regional metamorphic peak occurred at 475–465 Ma and the Curraghinalt veins formed at 467–461 Ma (Rice et al. Reference Rice, Mark, Selby and Hill2013). The sulphides in this deposit are predicted to be largely of metamorphic origin, being supplied by fluids from the devolatilization process in rocks beneath the Tyrone Igneous Complex, but may have been contaminated with igneous and basinal fluids and reacted with meteoric fluid to form the deposit.
8.a.7.b. Formation of the Cononish deposit
The time line for this deposit is longer and more complex than that for Curraghinalt. A dehydration front would have been established 470–465 Ma ago, at a depth of some 30 km in the crust, and risen towards the Earth's surface as uplift and erosion took place. There are as yet no known gold-bearing deposits between 460–410 Ma in age in the Grampian Terrane in Scotland (Rice et al. Reference Rice, Mark, Selby and Hill2013). The elements scoured from the country rocks by this fluid must either have been lost at the Earth's surface, as Rhynie-type hot springs, or crystallized as sub-economic quartz veins at a shallow depth, to be later recycled as erosion penetrated deeper into the greenschist-facies rocks. It is proposed that, during its waning stages, the Lochaber Batholith produced hot, reactive, Si-rich fluids, which scavenged the Dalradian rocks, and garnered the remaining noble metals to produce a mineral deposit (at 408 Ma) of mixed metamorphic–igneous parentage ‘contaminated’ (as in the case of Croagh Patrick) by additions of basinal, meteoric and mantle-derived fluids, to form the Cononish deposit.
These cameos may be discarded in the future as further radiometric age determinations and detailed stable isotope and geochemical studies reveal the true complexity of these processes. One aspect of the geology that has not been explained is why the widespread magmatic events at 426 Ma are not represented by, or did not give rise to, any mineral deposits between 430 and 408 Ma. If confirmed, the youngest gold deposits at 408 Ma can without doubt be related to the final magmatism in the batholith represented by the Inner Starav Granite at 408 Ma (Table 1) and probably occurred during a period of rapid uplift analogous to that reported from the Taiwan Orogen by Craw et al. (Reference Craw, Upton, Yu, Horton and Chen2010).
8.b. Structural controls on gold mineralization
The aim of this section is, in the light of the results presented in this paper, to review the hypothesis that Riedel fractures (R, R′) are an important factor in controlling the distribution and geometry of Au–Ag-bearing quartz lodes and veins in the upper crust.
Firstly, there are a number of well-documented, natural examples of simple shear fracture patterns that support the theoretical models of, for example, Coulomb (Reference Coulomb1773), Anderson (Reference Anderson1951), Moody & Hill (Reference Moody and Hill1956) and Sibson (Reference Sibson2000), and the experimental results of Riedel (Reference Riedel1929), Cloos (Reference Cloos1955) and Tchalenko (Reference Tchalenko1970). The resulting Coulomb–Sibson model has been successfully applied to Au-bearing quartz lode and vein deposits (Mueller, Harris & Lungan, Reference Mueller, Harris and Lungan1988; Vearncombe, Reference Vearncombe1998; Mares, Reference Mares1998; Neumayr, Hagemann & Couture, Reference Neumayr, Hagemann and Couture2000; Kolb & Hagemann, Reference Kolb and Hagemann2009; Dietrich et al. Reference Dietrich, Gutierrez, Nelson and Layer2011), but is hampered in many cases by the lack of quantitative structural data, and the past tendency of mine geologists to record the strike but not the dip of planar surfaces. However, this is not such a serious omission when dealing with strike-slip faults, as they are usually steeply dipping to vertical.
Secondly, there are a number of case studies in which conjugate strike-slip faults, together with penecontemporaneous folds, and thrust, reverse and domino-style faults, form a two-fold symmetrical pattern about σ1. Classical examples of this configuration have been described from the Devonian and Carboniferous rocks affected by the Variscan Orogeny in SW England and South Wales, UK. Anderson (Reference Anderson1951; see also Hancock, Reference Hancock1965, fig. 1) described conjugate sets of mineralized and non-mineralized strike-slip faults from south Pembrokeshire, which were grouped as follows: dextral, trend 345°; sinistral, trend 055°; and associated fold traces, trend 280°; giving β = 25° and σ1 = 020°.
A similar pattern is seen farther west in Cornwall and South Devon where Dearman (Reference Dearman1963) recognized that both mineralized and barren quartz veins occupy conjugate sets of strike-slip faults, but that, over much of the area, the major dextral faults are dominant, to the virtual exclusion of the sinistral set. The effects of this suppression of one of the two sets of faults forming the conjugate pattern was analysed in a detailed study of a small-scale strike-slip network by Nixon, Sanderson & Bull (Reference Nixon, Sanderson and Bull2011) at Westward Ho!, North Devon. They examined how strain compatibility had been achieved within the network, between areas controlled by pure shear conjugate patterns and those dominated by rotational domino-style faults. A similar dichotomy, on a much larger scale, exists between the conjugate patterns of strike-slip faults with their small displacements, seen in the Dalmally–Tyndrum area (inset on Fig. 5), and the larger (by an order-of-magnitude), dominant Great Glen set of NE-trending, left-lateral strike-slip faults (Fig. 3) that caused their development.
On a smaller scale, Wright, Woodcock & Dixon (Reference Wright, Woodcock and Dickson2009) described a strike-slip system from the Gower Peninsula, South Wales, where the morphology of individual fault-induced fissures and voids that had developed along a fault plane in the upper few kilometres of the Earth's crust could be examined in detail. The fissures are from 2–25 m wide and contain matrix-supported breccias and calcite veins with cockade textures. These primary (pure shear) fractures are seen as precursors to the type of vein found on Croagh Patrick, and the open fissures with a partial fill of different types of breccia are a proxy for the original nature of the quartz-breccia veins in the Cononish–Dalmally area (Tanner, Reference Tanner2012), before they were injected by silica-rich fluids. Thus, the geometrical relationships between faults and folds seen at a high crustal level in the frontal part of the Variscan orogen are replicated by some of the Grampian faults and quartz veins.
The importance of second- and third-order faults in localizing Au-deposits is becoming increasingly recognized, particularly in greenstone belts, such as the Yilgarn craton, Australia, and the Abitibi greenstone belt, Canada. Previous workers interpreted the major shear zones in these rocks as crustal-scale conduits for transporting mineralizing fluids to a higher level in the crust, as well as providing a host for the ore body. It is now recognized that whilst these shear zones enabled fluids to reach the upper crust, deposition of the mineral content took place largely in the network of smaller, second-order Riedel fractures that accompanied them (Mueller, Harris & Lungan, Reference Mueller, Harris and Lungan1988; Vearncombe, Reference Vearncombe1998; Neumayr, Hagemann & Couture, Reference Neumayr, Hagemann and Couture2000). The mineralized quartz veins that occupy these Riedel shears are disproportionately long compared to their width. Excellent examples of this style of deformation are provided by Katz, Weinberger & Aydin (Reference Katz, Weinberger and Aydin2004) and in particular, by Davis et al. (Reference Davis, Bump, Garcia and Ahlgren1999) (compare their small-scale map of the Sheets Gulch area (fig. 16) with Fig. 4 of this paper). Complications in interpretation arise when vein formation is initiated before polyphase deformation affecting the host rocks has ceased (Mares, Reference Mares1998), or there is a change in shear direction (Kolb & Hagemann, Reference Kolb and Hagemann2009).
The clearest demonstration that mineralization took place on second-order Riedel fractures is given by the analysis of the 2 km long Huevos Verdes vein system in the Deseado massif, Argentina, by Dietrich et al. (Reference Dietrich, Gutierrez, Nelson and Layer2011). In the San José district of the massif, a pair of sinistral strike-slip faults trending 018–028° bracket a compartment in which R (350°) and R′ (280°) faults and veins are developed. The main mineralized (extensional) fractures (T) trend at around 315° and β = 35°, identical to that predicted for Coulomb fractures in wet rock.
9. Conclusions
Most of the Au–Ag mineral deposits described here from the Grampian Terrane of Scotland and Ireland are hosted by primary and secondary faults that cut and displace folded and metamorphosed rocks belonging to the Neoproterozoic–Lower Ordovician Dalradian Supergroup (Tanner & Sutherland, Reference Tanner and Sutherland2007 and references therein). There is neither stratigraphical nor local lithological control on the development of these deposits and they are confined spatially to a 50 × 350 km area (Fig. 1) designated the Grampian orogenic gold province. In addition, several deposits (Bohaun, Cregganbaun & Croagh Patrick) occur in younger, folded Ordovician and Silurian rocks. This folding was previously thought to be an end-Silurian event, but has now been shown, through a combination of stratigraphical and radiometric dating in the west of Ireland, to be of Scandian (426 Ma) age (Table 1) (Leake, in press).
The Dalradian rocks were strongly deformed in the c. 470 Ma Grampian Orogeny that resulted from NW-directed subduction of part of the floor and of the contents of the Dalradian basin, beneath the SE margin of Laurentia (Tanner, Reference Tanner, Corfu, Gasser and Chew2013). This process operated throughout the duration of the orogeny, and supplied heat and fluids to the orogen until slab break-off occurred at 426 Ma (Atherton & Ghani, Reference Atherton and Ghani2002). It resulted in the emplacement of magma and hot fluids into the metasediments, and culminated in the formation of the Lochaber Batholith. The ‘Great Glen’ set of NE-trending strike-slip faults in Scotland is thought to have been intimately involved in the emplacement of the sub-crustal and crustal melts that formed the ‘Newer Granite’ suite, and either triggered this process (Watson, Reference Watson1984) or arose from it (Neilson, Kokelaar & Crowley, Reference Neilson, Kokelaar and Crowley2009). This involvement forged links between the strike-slip faults, the Newer Granites (Lochaber Batholith) and the gold-bearing deposits.
Au–Ag deposits are found at Calliachar–Urlar Burn, Tombuie, Tyndrum and Cononish in Scotland and Curraghinalt, Cavanacaw (Omagh mine), Croagh Patrick, Cregganbaun and Bohaun in Ireland. An important feature of these deposits is that most of them share the same overall mineralogy, although they are dispersed over a distance of >300 km, parallel to the strike of the Caledonide belt (Fig. 1). The general assemblage is: native gold, electrum (Au, Ag), hessite (Ag2Te), tetrahedrite, pyrite, chalcopyrite, arsenopyrite, galena and sphalerite ± bornite. Published isotopic and fluid inclusion data (Curtis et al. Reference Curtis, Pattrick, Jenkin, Fallick, Boyce and Treagus1993; Wilkinson & Johnston, Reference Wilkinson and Johnston1996; Ixer, Pattrick & Stanley, Reference Ixer, Pattrick and Stanley1997; Treagus, Pattrick & Curtis, Reference Treagus, Pattrick and Curtis1999; Wilkinson et al. Reference Wilkinson, Boyce, Earls and Fallick1999; Parnell et al. Reference Parnell, Earls, Wilkinson, Hutton, Boyce, Fallick, Ellam, Gleeson, Moles, Carey and Legg2000; Lusty et al. Reference Lusty, Naden, Bouch, McKervey and McFarlane2011) showed that the veins, both regionally and in individual deposits, crystallized under broadly similar P–T conditions with minimum temperatures of 140–290°C and maximum temperatures of 240–350°C at a burial depth of c. 3–5 km.
The Au–Ag vein complexes in Ireland, at Cregganbaun and Bohaun, occur in deformed Lower Palaeozoic rocks within the confines of the Grampian Terrane. These deposits show a more restricted mineralogy than the group described above, and only occur in western Ireland. The quartz veins at Curraghinalt and Cavanacaw occur in Dalradian host rocks and have an almost identical mineralogy to those in Scotland.
Repetition of the same structural style is a significant, unifying aspect of the Au–Ag deposits in the Grampian Terrane. All of these deposits, for which some structural data are available, have their development controlled by conjugate pairs of Andersonian faults. The mineralization occurred in second-order Riedel fractures or in the primary left-lateral transcurrent faults that gave rise to them (Fig. 4a, b, c, d). The calculated regional maximum principle stress (σ1R) direction is surprisingly constant across the Grampian Terrane at 348–011°, and β (the angle between σ1R and the Riedel fracture, R) varies from 25–30°, compared with the theoretical value of 27° for fractures developed in the presence of a pore fluid (Sibson, Reference Sibson2000). At Cononish and Bohaun the primary faults (Y) are accompanied by folds, the Orchy Dome at Cononish and the Lugnabrick–Benlugmore Syncline at Bohaun, and by major thrusts, such as the Omagh Thrust at Curraghinalt, that appear to have formed under the same kinematic regime. It is concluded that, during Late Silurian to Early Devonian times, the Au-bearing quartz vein systems described here from the Grampian Terrane developed in response to a N–S-trending σ1R of Scandian age, linked to the final closure of the Iapetus Ocean to the south.
Research into the origin of the fluids that gave rise to the Au-bearing deposits has so far proven inconclusive. The Grampian Au–Ag deposits have many features in common, such as a highly specific mineralogy and a similar gold microchemical fingerprint, that can only be explained in the context of a province-wide mechanism or process. A large volume of hot Si-rich fluid is necessary to produce a mineral deposit, but as there was a dearth of igneous intrusions of an appropriate age and volume to supply this, the sole remaining option would appear to be the metamorphic fluid. The basic tenet of the Otago model (Pitcairn et al. Reference Pitcairn, Teagle, Craw, Olivo, Kerrich and Brewer2006) adopted here is that this fluid was released in situ by regional metamorphic dehydration reactions. This ‘devolatization’ process (Phillips & Powell, Reference Phillips and Powell2010) was most active at the greenschist/amphibolite-facies boundary where, for example, hydrous minerals like chlorite were converted to biotite with the consequent liberation of a few per cent water. Owing to the large surface area presented by porous clastic rocks in particular, this fluid was able to scavenge Au and other noble metals, together with some base element sulphides, from the metasediments and transport them towards the Earth's surface. A ‘dehydration front’ developed that separated amphibolite-facies rocks at depth, which had been depleted of much of their metal content, from the overlying greenschist-facies rocks. Thus, early in the development of the orogenic belt, rocks metamorphosed at a higher grade than greenschist facies will have had most of their ore minerals removed by leaching. It is inferred that the overlying greenschist rocks were traversed by small bodies of silica-rich fluid that crystallized as metalliferous quartz veins. Alternatively, these narrow passageways could have developed into conduits capable of draining a large volume of fluid with the potential of developing economically viable Au-bearing quartz veins in these upper crustal greenschists.
It is thought that major transcurrent faults developed in the upper crust, propagated downwards and connected with conduits fed from below to supply over-pressured fluids to the breccia-filled cores and damage zones of a network of second-order R, R′ and T fractures (Tanner, Reference Tanner2012). These once-open voids were now seen as crystalline open-space quartz veins and quartz breccias. Hot fluids from crystallizing acid igneous magmas possibly provided the driving force for this process, and as any residual fluids approached the contemporary ground surface, boiling took place, and the remaining gold was precipitated during hot-spring activity (Rice et al. Reference Rice, Ashcroft, Batten, Boyce, Caulfield, Fallick, Hole, Jones, Pearson, Rogers, Saxton, Stuart, Trewin and Turner1995).
The application of the Otago model (Pitcairn et al. Reference Pitcairn, Teagle, Craw, Olivo, Kerrich and Brewer2006) to the problems posed by the Grampian deposits, primarily that of finding a source for the mineralizing fluids, has introduced a new layer of complexity to the search but does offer the possibility of constructing a quantitative framework that can be tested using a number of different techniques.
Acknowledgements
I am indebted to Judith Tanner for her probing questions and painstaking attention to detail whilst helping to prepare this manuscript, and to Mike Shand (University of Glasgow) for invaluable assistance with the draughting. I am grateful to Chris Sangster and Jeff Smith (both of Scotgold Resources); David Catterall (Farscape Exploration); Nyree Hill and Gawen Jenkin (both of Leicester University); and Clive Rice (University of Aberdeen) for many useful discussions on the geology of the Cononish gold–silver deposit and its surroundings; also to Mark Cooper (Geological Survey of Northern Ireland) for a constructive critique of an early draft of the manuscript; to two anonymous referees for their advice, and for alerting me to some key references; and to Dave Craw (University of Otago) for taking a critical look at the final manuscript.
