1. Introduction
The late Mesoproterozoic Torridonian sequences of NW Scotland (Fig. 1a) comprise three groups: the Stoer, Sleat and Torridon, the latter being subdivided into Diabaig and Applecross–Aultbae formations. Current age dates based on detrital zircon geochronology suggest deposition of the Torridonian occurred between c. 1200 Ma and c. 1000 Ma (Rainbird et al. Reference Rainbird, Hamilton and Young2001; Cawood et al. Reference Cawood, Nemchin, Strachan, Prave and Krabbendam2007; Kinnaird et al. Reference Kinnaird, Prave, Kirkland, Horstwood, Parrish and Batchelor2007). These rocks were deposited during the break-up of the supercontinent Rodinia (Cawood et al. Reference Cawood, Nemchin, Strachan, Prave and Krabbendam2007). Until recently, the only known occurrence of volcanism in the Torridonian was the Stac Fada Member of the Stoer Group, considered to be either a mud-flow (Lawson, Reference Lawson1972), a peperite (Sanders & Johnston, Reference Sanders and Johnston1989) or an impact ejecta blanket (Amor et al. Reference Amor, Hesselbo, Porcelli, Thackrey and Parnell2008).
Figure 1. (a) Sketch map of Scotland showing ‘brown bed’ localities (*). HBF – Highland Boundary Fault. (b) Palaeogeographic recontruction of Rodinia supercontinent at around 600 Ma (Cawood et al. Reference Cawood, Nemchin, Strachan, Prave and Krabbendam2007) when Scotland lay close to latitude 70°S. Ornament represents the Grenville/Sveconorwegian/Sunsas orogen, c. 1.30 Ga to 0.95 Ga, which approximates the Torridonian time span.
The Dalradian Supergroup represents deposition during late Neoproterozoic times and forms the central swathe of the Grampian Highlands (Fig. 1a). It ranges in age from c. 900 Ma to c. 520 Ma, while its detrital zircon age spectrum argues for a source from the Laurentia–Greenland region (Cawood et al. Reference Cawood, Nemchin, Smith and Loewy2003). The only firm depositional age for the Dalradian (upper Argyll Group) comes from a keratophyre in the Tayvallich Volcanic Subgroup, dated at 595±4 Ma (Halliday et al. Reference Halliday, Graham, Aftalion and Dykmore1989) or at 601±4 Ma (Dempster et al. Reference Dempster, Rogers, Tanner, Bluck, Muir, Redwood, Ireland and Paterson2002). The upper Argyll and lower Southern Highland groups contain extensive volcaniclastic deposits named Green Beds, and are considered to represent recycled mafic igneous detritus (van de Kamp, Reference van de Kamp1970; Pickett, Hyslop & Petterson, Reference Pickett, Hyslop and Petterson2006).
In late Neoproterozoic times (c. 600 Ma) Scotland lay close to latitude 70°S (Fig. 1b), surrounded by Greenland, Baltica, Laurentia and Amazonia (Soper, Reference Soper1994). Some 400 Ma earlier, in Torridonian times, Scotland was still juxtaposed to these four landmasses at latitudes of about 40°S (Dalziel & Soper, Reference Dalziel and Soper2001).
The tectonic environment in which these successions were deposited is pertinent to any discussion about eruptive magmatic events and is relevant to the subject of this paper. A recent assessment of tectonic environments during the Torridonian (Kinnaird et al. Reference Kinnaird, Prave, Kirkland, Horstwood, Parrish and Batchelor2007) infers that the Sleat Group was deposited in an extensional sedimentary basin, whereas the Diabaig Formation represents a lacustrine margin deposit. The late Dalradian rocks discussed in this paper are considered to represent deposition in an extensional rift basin (Anderton, Reference Anderton1985; Prave, Reference Prave1999). Rift environments generate magmas of various compositions from basalts to silicic fractionates, although they tend to have lower magnesium and higher concentrations of alkalis compared with primary mantle melts or calc-alkaline arc-related magmas (Wilson, Reference Wilson1989).
A suite of ‘brown beds’ from the Upper Southern Highland Group, Dalradian Supergroup, Scotland (Batchelor, Reference Batchelor2004, a, b), occur in the field as thin albite-rich beds (5–100 mm) with sharp upper and lower contacts with their host sediments, suggesting that they represent subaqueous tephra falls. The ‘brown beds’ occur along most of the strike of the Southern Highland Group in Scotland, a distance of about 200 km (Fig. 1a).
A second discovery of two ‘brown beds’ in the Sleat Group (Torridonian) on Skye (Batchelor, Reference Batchelor2005) showed that their lithology and chemical composition are almost identical to the Dalradian examples. More ‘brown beds’ have been discovered since in the Sleat Group and Diabaig Formation (Torridonian), and in the Morar Group, Moine Supergroup (Batchelor, unpub. data), implying that this lithology is not a localized phenomenon, either in place or time.
‘Brown beds’ from Torridonian successions occur within slates or siltstones which were subjected to prehnite–pumpellyite-facies metamorphism (Johnson et al. Reference Johnson, Kelley, Oliver and Winter1985; new data, this paper), whereas the Dalradian rocks were subjected to low greenschist-facies metamorphism (Harris et al. Reference Harris, Haselock, Kennedy, Mendum, Gibbons and Harris1994). Their chemical composition is characterized by relatively high levels of Fe2O3 (3–9%) and Na2O (3–6%), and low levels of CaO and MgO (<2%).
Mineralogically the ‘brown beds’ typically comprise albite (60%), chlorite (15%), muscovite (15%) and quartz (10%). In thin-section they contain angular clasts of quartz and plagioclase feldspar (determined to be albite–oligoclase on the basis of maximum extinction angles) set in a cryptocrystalline matrix coated with a brownish-orange material. The macroscopic feldspar grains display strong lamellar twinning with no alteration rims. The brownish-orange material, which coats many of the grains, is assumed to be amorphous hydrated Fe–Mn oxide (?limonite), since it does not feature on X-ray diffractograms (XRD) and there is no other mineral to account for the relatively high levels of Fe2O3.
A literature search for analogous lithologies failed to yield any specific examples, although four references may allude to ‘brown beds’. Gunn, Clough & Hill (Reference Gunn, Clough and Hill1897, pp. 18–19) described ‘. . . many calcareous, ochreous weathering bands and lenticles, rarely exceeding a few inches in breadth and a few yards in length, which show no clear foliation at all. These calcareous bands are great helps [sic] in making out the original bedding’. Harris (Reference Harris1972) mentions ‘thin sandy beds with sharp tops and bottoms’ which occur within the Birnam Slate Formation near Dunkeld in Perthshire. These are probably the ‘brown beds’ described by Batchelor (Reference Batchelor2004a). Some putative ash beds, 5–10 cm thick and weathering to a yellowish-green colour, have been described from the Birnam Slate Formation 7 km northeast of Dunkeld, Perthshire (Crane, Reference Crane2002). Brown-weathering friable lenticular bodies in the Southern Highland Group psammites in the Sma' Glen of Perthshire were described as ‘diagenetic concretions’ (P. T. S. Rose, unpub. Ph.D. thesis, Univ. Liverpool, 1989). These lenticular bodies are in fact incompetent ‘brown beds’ described by Batchelor (Reference Batchelor2004b).
White quartz-albite rocks from the Permian succession of Australia, which had been originally misidentified as cherts or siliceous siltstones, were described as tuffs (Leitch, Reference Leitch1981). A study of their field relationships, mineralogy and chemistry concluded that they had resulted from burial metamorphism of silicic vitric tuffs which altered first to analcite-rich tuffs, then to albite-rich tuffs.
A serendipitous discovery by ASR of a ‘brown bed’ (sample GH/RB/112) with a grey core in the Birnam Grit Formation (Dalradian Supergroup), Glen Shee, Methven, Perthshire, [NN 9755 3495] led to the suspicion that the ‘brown bed’ was in fact an alteration product of another lithology. Further discoveries of similar ‘brown-grey’ assemblages in rocks of the Sleat Group (Torridonian) (sample SK/RB/05), at a viewpoint car park 1 km east of Kyle of Lochalsh [NG 7725 2725] and in the Diabaig Formation (Torridonian) (sample TN/RB/026) at Lower Diabaig [NG 7925 6025], some 25 m east of its contact with the Torridonian Applecross Group, prompted this study, which aims to explain the origins of this unusual lithology in the late Proterozoic rocks of Scotland.
A typical example of a ‘brown bed’ formed from its grey precursor is shown in Figure 2. This example (sample SK/RB/15) comes from the Kinloch Formation, Sleat Group (Torridonian) in the Sleat Peninsula of Skye [NG 695 185].
Figure 2. Photograph of a ‘brown bed’ BB (sample SK/RB/15) from the Sleat Group grading into its grey (‘GY’) equivalent and showing banding in the alteration fronts. The view shows the alteration zone sandwiched between two siltstone layers. Specimen size is 120 mm × 35 mm.
2. Analytical results
2.a. Petrography
A thin-section of the grey material from the Dalradian locality (sample GH/RB/112GY) (Fig. 3a) shows an assemblage of angular to sub-angular quartz and albitic feldspar (mostly lamellar-twinned), clusters of chlorite and some flakes of muscovite with no evidence of preferred orientation. Calcite generally forms the interstitial phase but in places it is found inside grains of albite, which suggests its formation during metamorphic recrystallization in the presence of CO2. A curved grain of lamellar-twinned albite suggests the feldspar grain pre-dates the regional deformation events. The cognate brown material (sample GH/RB/112BN) contains a similar assemblage of quartz, albite, chlorite and muscovite, but without calcite. It has a porous texture and the minerals are coated with an amorphous brown material, assumed to be hydrated Fe–Mn oxides (?limonite). In contrast, a thin-section of the host slate contains predominantly quartz, muscovite and chlorite orientated along a well-developed slaty cleavage.
Figure 3. (a) Photomicrograph of sample GH/RB/112GY, the grey material from the Birnam Grit, Dalradian at Little Glen Shee, Perthshire. (b) Photomicrograph of sample SK/RB/05GY, grey material from the Sleat Group, Torridonian, Isle of Skye. (c) Photomicrograph of SK/RB/05BN (upper)–mudstone (lower) interface (marked by the black line). Sleat Group Group, Isle of Skye. (d) Photomicrograph of TN/RB/026GY, grey material from the Diabaig Group, Torridonian. Image width: 1 mm for all photomicrographs. ab – albite, cc – calcite, ep – epidote, qz – quartz.
The grey material from the Sleat Group sample SK/RB/05GY (Fig. 3b) similarly contains angular to sub-angular quartz and albitic feldspar set in a matrix of calcite. Other quartz grains display embayments, which have the appearance of corrosion textures, and plagioclase feldspars show irregular surfaces. In addition, untwinned K-feldspar grains are present. The brown equivalent (sample SK/RB/05BN) contains the same mineral assemblage but with the absence of calcite. In contrast, the host rock (Fig. 3c) is a finely laminated mudstone with angular to sub-angular quartz grains <50 μm set in a fine chloritic matrix. Sparse muscovite laths lie parallel to bedding. No feldspar was identified optically. The contact between the mudstone and the ‘grey bed’ is sharp.
The grey material of sample TN/RB/026GY (Fig. 3d) from the Diabaig Formation contains angular quartz grains set in a matrix of calcite. Two quartz grains in the figure show euhedral pyramid terminations, suggestive of broken quartz bipyramids, a common form of quartz in volcanic rocks. Angular grains of lamellar-twinned plagioclase feldspar have low extinction angles suggesting albite–oligoclase compositions. There are sporadic anhedral grains of epidote. The brown equivalent of this sample (TN/RB/026BN) has exactly the same mineral composition, except that calcite is absent and the voids are coated with an orangey-brown to dark-brown material, assumed to be an amorphous hydrated iron oxide.
2.b. X-ray diffraction (XRD)
A portion of the brown material and grey material from all three samples was powdered and subjected to XRD analysis at St Andrews University using a Philips powder diffractometer using Co Kα radiation. Semi-quantitative mineral concentrations were calculated from the XRD data using SIROQUANT software (CSIRO Energy Technology, Private Mail Bag 7, Bangor, NSW 2234, Australia).
Dalradian (sample GH/RB/112GY): albite (50.9%), chlorite (12.7%), muscovite (12.7%), quartz (8.5%), calcite (15.2%).
Diabaig (sample TN/RB/026GY): albite (31%), chlorite (4%), illite (3%), quartz (26%), calcite (36%).
Sleat (sample SK/RB/05GY): albite (19%), chlorite (10%), illite (11%), quartz (33%), microcline (6%), calcite (21%).
The profiles for all three pairs of samples are similar, containing variable amounts of albite, quartz, chlorite and muscovite, except that the grey material contains calcite in addition, confirming the petrographic analysis (Fig. 4). The Sleat sample SK/RB/05 contains additional K-feldspar (identified as microcline) and it contains more quartz than albitic feldspar, but the overall mineral assemblage is the same as the other two samples.
Figure 4. X-ray diffractogram of a typical grey–brown pair (sample GH/RB/112). M – muscovite, Chl – chlorite, Q – quartz, Ab – albite, Cc – calcite.
In order to constrain the degree of metamorphism which the Torridonian samples had experienced, illite crystallinity parameters Hbrel and b0 were measured, as described by Weber (Reference Weber1972) and Sassi & Scolari (Reference Sassi and Scolari1974) respectively, on cognate shales from the Diabaig Formation and Sleat Group, and muddy siltstones from the Stoer Group of the Torridonian. The Weber Hbrel value (Fig. 5a) is calculated as the half-peak width of the illite [100] reflection, divided by the half-peak width of quartz [100], multiplied by 100. Sleat and Diabaig rocks reached transitional prehnite–pumpellyite to greenschist facies (anchizone–epizone), ~350°C. The Stoer Group (Lower Torridonian) rocks reached transitional prehnite–pumpellyite to zeolite facies (diagenesis–anchizone), ~200°C (Blenkinsop, Reference Blenkinsop1988). The b0 parameter (Fig. 5b), based on the absolute position of the illite [060] reflection internally standarized to the nearby [211] quartz reflection, gives an indication of the relative pressure regime: low values correlate with low pressure (zeolite facies), while increasing pressure leads to higher values (Sassi & Scolari, Reference Sassi and Scolari1974). Diabaig Formation illites give b0 values of 9.005–9.010, indicating relatively low pressure conditions (Johnson et al. Reference Johnson, Kelley, Oliver and Winter1985). Sleat and Stoer Group illites give a mean value of 9.03, which also indicates low pressures consistent with the transition from zeolite to prehnite–pumpellyite-facies metamorphism. These values are akin to those found in slates from Silurian inliers in the Midland Valley of Scotland (Oliver et al. Reference Oliver, Smellie, Thomas, Casey, Kemp, Evans, Baldwin and Hepworth1984).
Figure 5. Illite crystallinity data: (a) Hbrel values based on the illite [100] peak. Asterisks represent data for Stoer rocks from Johnson et al. (Reference Johnson, Kelley, Oliver and Winter1985); (b) b0 values based on the illite [060] peak. Diagenesis – zeolite facies; Anchizone – prehnite–pumpellyite facies; Epizone – greenschist facies.
2.c. Geochemistry
Major and trace elements were determined by X-ray fluorescence, using traditional fused disc and pressed powder methods, on the grey and brown material from the three samples (Table 1). Analyses were carried out at St Andrews University using a Spectrolab EDPXRF system. The high levels of CaO and loss-on-ignition (reflecting the abundance of volatile components) in the grey lithologies reflect the high concentrations of calcite. Notably MgO is consistently. ≤1.0%.
A theoretical composition for the Dalradian brown-bed sample (GH/RB/112BN) was calculated by removing excess CaO (9.88–0.28=9.60%) and loss-on-ignition (8.8–2.1=6.70%, assumed to represent mostly CO2) from the grey sample (GH/RB/112GY), and recalculated to the original total for the grey sample ‘112GY’. The correspondence between the calculated versus the actual ‘brown bed’ composition is close. A calculated value for CaCO3 of 16.3% in the grey rock agrees well with the calculated semi-quantitative calcite concentration of 15.2% obtained from the XRD data using SIROQUANT software. Sample SK/RB/05 has lost 9.66% CaO and 7.6% volatiles in the weathering process, values similar to the Dalradian sample. Levels of K2O are higher, which reflects the presence of K-feldspar as identified by XRD and petrography. Sample TN/RB/026 has lost 14.78% CaO and 10% volatiles during the weathering process. These values suggest a slightly higher calcite concentration, in the region of 23%.
A spider plot (Fig. 6) of the three ‘grey’ samples shows good consistency between them. They are compared with their respective host sediment. While the patterns are roughly similar in shape they differ in absolute values. This may reflect a measure of clastic input into the volcanic deposits. It is notable that Sr is generally lost during the alteration process, and given its affinity for Ca, this implies it is lost when calcite was dissolved and removed by weathering. The prominent Nb and Ti negative anomalies are a common feature of magmas generated in subduction zones or active continental margins (Rollinson, Reference Rollinson1993; Wilson, Reference Wilson1989).
Figure 6. Spider diagram showing the three grey rocks and their respective host sediments normalized to MORB (data from Saunders & Tarney, Reference Saunders, Tarney, Kokelaar and Howells1984; Sun, Reference Sun1980). Open symbols represent the host sediment of the related grey rock (filled symbol).
Multi-elemental parameters can be applied to petrogenetic problems. Their advantage is in maximizing the use of data. One such technique was developed by de la Roche et al. (Reference de la Roche, Leterrier, Grand Claude and Marchal1980), who devised two parameters, R1 and R2, developed from combinations of millication values for the major elements: R1 represents 4Si − 11(Na+K) − 2(Fe+Ti) and R2 represents 6Ca + 2Mg + Al. The advantage of this approach is that mineral compositions can be readily plotted on the same diagram and can indicate their influence on rock chemical variation. These parameters for rocks from this study are presented in Figure 7, together with compositions for related minerals and for host sediments for the three rocks described here. The most obvious feature of this diagram is the orientation of sample pairs SK/RB/05GY-BN. Extending a tie line beyond the GY sample leads to the carbonates (cc and dol) mineral pole, strongly suggesting that carbonate removal directly influenced the conversion of 05GY to 05BN. The more vertical orientations for samples GH/RB/112 and TN/RN/026 suggest that removal of a combination of carbonate and quartz may have been involved in the conversion. Assuming that the grey lithology approximates the original rock composition before metamorphism, the three GY samples plot in the field represented by basaltic compositions, as defined by de la Roche et al. (Reference de la Roche, Leterrier, Grand Claude and Marchal1980), although low Mg and high Fe in the GY samples suggests that the original magma was tholeiitic, a feature of active continental margins and continental rift zone ferrobasalts (Wilson, Reference Wilson1989). The position of the associated sediments reflect combinations of illite, chlorite and quartz, in order of abundance.
Figure 7. Multicationic factors R1 and R2 (de la Roche et al. Reference de la Roche, Leterrier, Grand Claude and Marchal1980) for brown beds (bn), grey beds (gy) and host sediments. Mineral composition data from Deer, Howie & Zussman (Reference Deer, Howie and Zussman1992). Host sediment compositions are circled.
2.d. Scanning electron microscopy (SEM)
A sample of a cognate Sleat ‘brown bed’ SK/RB/02 (Batchelor, Reference Batchelor2005) was viewed under SEM at the School of Biological Sciences, St Andrews University. Quartz, feldspar and chlorite were identified, which confirmed the petrographic analysis. At the sub-10 μm level, clusters of honeycomb-textured grains were seen (Fig. 8) which resemble smectite. Elemental analysis of the surface indicated Si, Al, Mg, K, Fe and Mn (Table 2). This assemblage, when converted into oxide values, does not represent the common Fe-bearing clay minerals chlorite or nontronite, nor illite or smectite (Deer, Howie & Zussman, Reference Deer, Howie and Zussman1992). In particular, Fe and Mn are anomalously high. Assuming that the Fe and Mn components form a thin coating (as determined petrographically) and assuming the mineral is an illitic clay, the analysis was re-adjusted on the basis of 10% H2O. This produces a composition which resembles a mixed-layer illite–smectite. This possibility is supported by the slight asymmetry seen in the 10 Å XRD peak. This suggests that the honeycomb structure represents smectite which subsequently altered to an illite–smectite mixed layer. Authigenic Fe–Mn oxides formed as part of the metamorphic process (Fisher & Schmincke, Reference Fisher and Schmincke1984) or were released by the breakdown of calcite. These oxides give the rock its variable brown colour.
Figure 8. SEM image of clusters of illite–smectite coated with Fe–Mn oxides. (Sleat Group ‘brown bed’ sample SK/RB/02).
Table 2. Chemical analysis of ‘smectite’ clusters
2.e. Stable isotopes of carbon and oxygen
Samples of the grey material from the three localities were analysed for their carbonate C and O isotope signature, using a Finnigan Delta Plus XL mass spectrometer at St Andrews University, in order to determine the source of the carbonate. Duplicate samples were flushed through with He gas, then treated with 100% phosphoric acid. The isotopic compositions of C and O were determined on the liberated CO2 gas. Standardization was based on carbonate standard NBS18 and standard deviation in all cases was ≤0.03 (1 S.D.). Results are listed in Table 3. The estimated CaCO3 content for the Dalradian sample (GH/RB/112GY), based on the yield of CO2 gas, was 18% by weight, which is close to the 16.3% calculated as excess CaO and LOI from the whole rock analysis (Table 1). These data are displayed in Figure 9, together with data for other rocks.
Table 3. Carbon and oxygen isotope values for the ‘grey beds’ and analytical standard
Figure 9. Diagram of δ13C v. δ18O for the three grey beds. Data for other rocks types from Rollinson (Reference Rollinson1993).
3. Petrogenesis
Two processes could account for the production of the grey facies described above: either albite forms via analcite, or albite is derived from the breakdown of Ca-plagioclase feldspar.
3.a. Scenario 1
Coombs (Reference Coombs1954) first recognized progressive burial metamorphism and zeolitization of glassy tuffs in Triassic sediments of New Zealand. This mechanism was invoked for altered Jurassic tuffs, rich in analcite and albite, from Idaho and Wyoming, USA (Gulbrandsen & Cressman, Reference Gulbrandsen and Cressman1960). Iijima (Reference Iijima, Sand and Mumpton1978) subsequently recognized four stages of burial metamorphism for volcanic glass, which involved the sequential formation of smectite, alkali zeolites, analcite and ultimately albite. The albite zone is characterized by authigenic albite and co-existing quartz, chlorite and white mica (Utada, Reference Utada, Bish and Ming2001).
This mechanism was also proposed for the formation of quartz–albite rocks from the Permian succession of Australia, described by Leitch (Reference Leitch1981), which had been originally misidentified as cherts or siliceous siltstones. A study of their field relationships, mineralogy and chemistry concluded that they had resulted from burial metamorphism of silicic vitric tuffs which had altered first to analcite, then to albite.
3.b. Scenario 2
The alteration of basaltic glass produces palagonite in the first instance by a series of hydration and oxidation processes. Burial metamorphism generates smectite and chlorite, followed by chlorite + zeolite + opal CT (cryptocrystalline silica) + Fe/Mn oxides or carbonates (Fisher & Schmincke, Reference Fisher and Schmincke1984). Metamorphism to prehnite–pumpellyite facies will produce albite + chlorite + pumpellyite + prehnite + quartz, while greenschist facies will generate chlorite + albite + epidote ± actinolite + quartz. The latter assemblage is found in the Dalradian ‘Green Beds’, which occur in the Tayvallich Volcanic Subgroup (van de Kamp, Reference van de Kamp1970; Pickett, Hyslop & Petterson, Reference Pickett, Hyslop and Petterson2006), a sequence of basaltic volcaniclastic sediments generated in a rifting environment and placed at the base of the Southern Highland Group (Graham, Reference Graham1986). Under burial metamorphism at zeolite facies, Ca-plagioclase loses Ca to form albite, and the Ca goes to form calcite if there is sufficient CO2 (Leichmann, Broska & Zachovalová, Reference Leichmann, Broska and Zachovalová2003) or Ca-zeolite if there is not. A glassy phase would initially produce smectite as described above. The presence of calcite in the ‘grey bed’ suggests that the system was externally buffered by CO2 (Thompson, Reference Thompson, Mackenzie and Zussman1974). It is assumed that the total Ca budget in the GY rocks approximates the original tuff composition. Ca is also released by the decomposition of amphibole and pyroxene.
4. Discussion
Metamorphic albite tends to form either simple twins or untwinned crystals, whereas complex twinning is more common in plagioclase feldspars formed in igneous rocks. Albite formed from the alteration of analcite (cubic system) tends to form featureless aggregates or mosaics (Smith, Reference Smith1974, p. 375). Macroscopic albite showing multiple lamellar twinning is assumed to be a primary feature. The presence of a curved lamellar-twinned albite crystal in the Dalradian sample supports this idea. Macroscopic quartz and albitic grains probably represent broken original xenocrysts and/or broken phenocrysts in a tephra deposit, suggesting that the source material was crystal tuff, rather than purely an aphyric glass. Geochemical data suggests that there may have been variable intermixing with the host sediment during deposition, a point which will have a bearing on the nature and integrity of the zircon population when considering geochronological work.
Under SEM at the sub-10 μm level in a ‘brown bed’ from the Sleat Group, clusters of honeycomb-textured assemblages resemble smectite, and elemental analysis suggests they are illite/smectite. This would support their origin as altered volcanic glass which underwent alteration to smectite before being subjected to low-grade metamorphism. This process supports Scenario 2 above as the mechanism for the production of the observed lithology. The brownish-orange material in the ‘brown beds’ is assumed to be amorphous hydrated Fe–Mn oxides formed by the oxidation and hydration of the carbonate phase by percolating acidic meteoric waters.
The isotopic data for carbon (δ13C −19.4) in the Dalradian sample is consistent with the calcite being formed from CO2 derived from contemporary graphite-rich pelites during metamorphism and suggests a biogenic origin for the carbon (Graham et al. Reference Graham, Greig, Sheppard and Turi1983). In contrast, the δ13C signatures in the Sleat Group (−9.6) and Diabaig Formation (−3.0) groups suggest that the sources of CO2 were less likely to be biogenic. Seawater values for δ13C have varied over time from 0 to +6 to +13‰, so more depleted sources are required to explain these data. Metamorphic fluids in granulites have a compositional range of −5 to −8‰, while mantle-derived carbon ranges from −3 to −8‰. Decarbonation of marine carbonates can yield δ13C values of +2 to −10‰ (Rollinson, Reference Rollinson1993). Given the paucity of limestones in Torridonian rocks, the source of the carbon is more likely to be crustal, released during prehnite–pumpellyite-facies metamorphic devolatilization. A notable feature of Neoproterozoic glacial deposits is their association with negative δ13C excursions down to −10‰ (Halverson et al. Reference Halverson, Hoffman, Schrag, Maloof and Rice2005), a point which may have a bearing on the rocks described here. The negative values for δ18O in all three samples imply interaction with meteoric water (Rollinson, Reference Rollinson1993).
5. Conclusions
This investigation has shown that the ‘brown bed’ lithology is in fact the weathered residuum of an albite–chlorite–muscovite (or illite)–quartz–calcite assemblage after removal of calcite by dissolution. Such an assemblage is highly unlikely to represent any clastic rock but rather is typical of the sub-greenschist (Torridonian) to low-greenschist (Dalradian) metamorphism of an intermediate to mafic igneous rock. The low-Mg and high-Fe signature of the grey beds, assuming these are characteristic of the original igneous rock, suggests a tholeiitic source in an extensional tectonic regime. Crucially, the discovery of this ‘brown bed’ lithology throughout the late Proterozic rocks of Scotland has led to the identification of its ‘grey’ precursor, assumed to be metamorphosed tephra-fall deposits of intermediate to mafic crystal tuff composition. The grey colour probably derives from the combination of fine-grained illitic clays and quartz. Were it not for the sub-aerial weathering effect which produces the distinctive brown, saccharoidal texture, the grey rock would blend in with its host sediment and remain unnoticed. These rocks should be targeted in any attempts to determine geochronological constraints on the Highland rocks, although an assumed basaltic precursor and a variable input of clastic sediment could compromise the use of zircon geochronology in determining the age of deposition.
Acknowledgements
This work was initiated by a Large Carnegie Trust grant awarded to Grahame Oliver which allowed RAB to search the Scottish Highlands for potential volcanic rocks. Rosalind Garton supported much of the fieldwork. Further fieldwork in the Torridonian was funded by a Leverhulme Research Grant awarded to Anthony Prave. Andrew Raeburn gratefully acknowledges support for fieldwork from his parents, Douglas and Wendy Raeburn. Sincere thanks go to Angus Calder for providing the X-ray analytical data, and to Andy Mackie for producing thin-sections. Jonathan Wynn and Alix Cage kindly provided the C and O isotope data.
