Microbially mediated carbonates in the Mesoproterozoic Stoer Group of NW Scotland; earliest evidence of life in Britain? Open Access

The stratigraphy of the Stoer Group is summarized by Figure 1b and c. The Clachtoll Formation onlaps the Lewisian palaeosurface with incised palaeovalleys infilled by alluvial fan and rockfall deposits (Stewart 2002; Killingback et al. 2021). These pass laterally into a muddy flood basin facies with single or multiple fluvial channels and lakes in interchannel areas (Stewart 2002; Ielpi et al. 2016; Valenza et al. 2019).

The overlying Bay of Stoer Formation consists of sandstone with minor shale deposited as fluvial channels on floodplains (Stewart 2002). The upper part of the Bay of Stoer Formation comprises two Members: the Stac Fada Member, interpreted as a meteorite impact breccia by Amor et al. (2008), and the Poll a'Mhuilt Member. At the Bay of Stoer, the Poll a'Mhuilt Member consists of red organic-rich mudstone with sandstone and carbonate intervals. Stewart and Parker (1979), Stewart (2002) and Parnell et al. (2015) suggested that this represents an evaporitic lake; however, Stüeken et al. (2017) proposed that this interval was influenced by marine water. Upfold (1984) identified planar, mammillated and tufted mats in the Poll A'Mhuilt Member, which he interpreted as microbial stromatolites. At Enard Bay, the Poll a'Mhuilt Member comprises alluvial breccio-conglomerates shed from local palaeohills in the Lewisian land surface that pass laterally into lacustrine mudstone with carbonates (Gracie and Stewart 1967; Stewart 2002). The Meall Dearg Formation consists of cross-bedded fluvial sandstone with some aeolian facies (McMahon and Davies 2018).

The two following sections provide a detailed description and interpretation of the sedimentology, carbonate microfabrics and depositional environments of the sections in which the laminated carbonate structures and sediment surface textures are found. An overview of the possible origins of these structures will be presented in the discussion section.

The logged sections are 1.5–2 m thick intervals of laminated sandstone, siltstone and mudstone that are interbedded with structureless fine muddy sandstone some 8–15 m thick (Fig. 2). These intervals consist mainly of muddy siltstone with planar laminated fine to medium sand a few millimetres to centimetres in thickness and carbonate sheets composed of calcite microspar a few millimetres in thickness (Fig. 3a). Occasional interbedded layers of erosionally based fine to medium sandstone with granule- and fine pebble-sized tabular clasts of red mudstone are also present (Fig. 3b). Some of the sand laminae have been cut by structures injected with sand and silt, and some chaotic intervals with convolute lamination are present. Isolated symmetrical ripples of fine sandstone with cross-bedding up to 2 cm thick and wavelengths of 5–10 cm are present at a number of horizons and are often exposed on bedding planes (Fig. 3c). Lenses of fine sand, sometimes showing bimodal trough-bedding, infill channel-like structures with scoured basal surfaces a few centimetres across (Fig. 3d). Many of the isolated ripples and sand-filled cut and fill structures are deformed and have foundered into the underlying siltstone and mudstone (Fig. 3e). Mud- and siltstone layers are often cut by a polygonal network of subvertical cracks infilled by fine sand that die out downwards; some of these fractures are sinuous (Fig. 3f). Bedding plane exposures of carbonate sheets often have a network of polygonal ridges but with no associated downward penetrating fractures (Fig. 3g). Some mudstone and siltstone bedding planes display low-relief laterally linked domes that are symmetrical to slightly elongate structures up to 20 cm across with a relief of a few millimetres to centimetres (Fig. 3h). The troughs between domes contain stringers of silt- and very fine sandstone that onlap the margins of the domes.

Domical structures made up of sub-parallel, planar to undulatory carbonate layers occur in this facies association (Fig. 4a and b). The carbonate laminae show two fabric types: layers of equant calcite microspar and layers composed of calcite plates that have a sub-parallel to irregular alignment that imparts a ‘crinkly’ fabric to the layering (Fig. 4b and c). Both types of carbonate laminae are up to a few millimetres thick and alternate with sub-millimetre thick partings of mudstone and siltstone. The domes show two morphological types, as follows.

Smooth domes are equant to slightly elongate to symmetrical structures with lateral dimensions of 5–20 cm between 1 and 5–10 cm thick with flat sub-parallel bases and tops with margins whose dip varies from 30° to vertical (Fig. 4a and b). Most smooth domes are isolated structures that are interbedded with and onlapped by planar and ripple-laminated siltstone and fine sandstone (Fig. 4a and b). Carbonate layers within the domes mimic the external form of the dome, often thinning towards the edges of domes. Smooth domes often contain intervals of the microspar fabric alternating with the crinkly fabric (Fig. 4a and b) and the margins of some domes consist of the crinkly fabric whereas the dome interior consists of microspar layers. Clastic mudstone layers also alternate with the microspar and crinkly layers. Smooth domes show alternating offlapping and back-stepping forms (Fig. 4b).

Cuspate domes have similar dimensions to smooth domes, but the carbonate component is made up of the crinkly carbonate fabric (Fig. 4c). The bases of cuspate domes are planar to undulatory carbonate layers developing into more pronounced undulations; the upper parts of cuspate domes display irregular asymmetrical upward-projecting cuspate forms (Fig. 4c). Some clastic-rich layers and grains are included with the cuspate domes. The morphology of the internal layering also mimics the external form of the dome.

The basal 14.6 m of the Poll a'Mhuilt Member, including carbonate-bearing units A to C of Stewart (2002, p. 68), was logged from the top of the Stac Fada Member (Fig. 5).

The lowermost 4.0 m (interval 1, Fig. 5) consists of beds 0.1–0.5 m thick of structureless moderately sorted fine to medium sandstone sand-sized grains and granule-sized tabular clasts of red mudstone alternating with thin shale beds. Sandstone beds are ungraded with non-channelized bases. An interbedded package up to 1.0 m in thickness of fine sandstone, siltstone and mudstone with thin carbonate layers is present (Fig. 5). This comprises planar and ripple-bedded fine sandstone interbedded with siltstone and mudstone. Bedding planes with linear to sinuous symmetrical ripples are present (Fig. 6a). Occasional bedding surfaces with laterally linked low-relief symmetrical domical structures occur in this interval. The individual domes are a few centimetres in lateral dimension and a few millimetres in relief (Fig. 6b, yellow arrow). Bedding surfaces penetrated by sand-filled polygonal desiccation fractures also occur in this interval (Fig. 6b, red arrow).

Planar to undulatory carbonate sheets 1–2 mm in thickness composed of microspar are also interbedded with the rippled and laminated sandstone and siltstone with occasional acicular and chevron-shaped single crystal pseudomorphs infilled by carbonate in associated sediment. These carbonate sheets are mainly planar, but some clusters of microspar form upward-branching tuft-like and conical projections (Fig. 6c) forming elongate to equant low-relief domes a few millimetres to centimetres in height and 10–15 cm in wavelength that are onlapped by siltstone and fine sandstone (Fig. 6d). Upfold (1984) showed that the carbonate texture consists of uniform equant microspar. The topmost 0.5 m of interval 1 consists of ripple-bedded fine sandstone, siltstone and red mudstone interbedded with carbonate sheets composed of coarse equant microspar 0.5–1.0 mm in thickness (Fig. 6e). The carbonate sheets increase in frequency upwards to form a laminated carbonate structure comprising layers of equant microspar alternating with packages of the crinkly carbonate fabric described from Clachtoll Bay, locality 1 (Fig. 2). The internal structure consists of planar to slightly domical carbonate laminae with a sharp top overlain by a cross-bedded medium sandstone.

Interval 2 consists of planar bedded laminated siltstone interbedded with fine sandstone with some tufted carbonate sheets overlain by a laminated carbonate unit 0.6 m thick that sits in a sharp, channelized base (Fig. 5). Its internal structure consists of laterally continuous planar and domical millimetre-scale sheets of calcite microspar that form broad overlapping domical structures. The top of the carbonate unit is sharp.

Interval 3 consists of a planar bedded fine sandstone with occasional carbonate sheets that fines upwards into a laminated siltstone and mudstone with planar carbonate sheets that increase in frequency upwards, amalgamating into a laminated carbonate structure up to 2.0 m in thickness (Fig. 5). The lower part of the carbonate structure comprises planar laminations with some low-relief symmetrical domes (Fig. 6f). This culminates in a layer of smaller higher relief domes up to 0.2 m in lateral and vertical dimension (Fig. 6g). The upper part of the carbonate interval shows a series of broad overlapping symmetrical domes that develop on the layer of smaller domes (Fig. 6h). These are 0.5–1.9 m across with a vertical relief of between 0.2 and 0.3 m. The internal morphology of the carbonate laminae mimics the external form of the structures. The top of this carbonate interval is sharp with broad domical structures onlapped by medium to fine sandstone (Fig. 5).

Breccio-conglomerates form a series of tabular bodies 5–10 m in thickness that offlap the Lewisian basement. Some breccio-conglomerate bodies are dominated by pebble-sized conglomerates whereas other are dominated by coarser cobble- and boulder-sized clasts. The breccio-conglomerates comprise poorly sorted rounded to sub-angular, pebble- to cobble- and boulder-sized clasts of Lewisian material with a matrix of finer clasts and medium to very coarse sand-sized quartz grains (Fig. 7a–d). Coarse-grained cobble- and boulder-sized breccio-conglomerates tend to be very poorly stratified with very crudely developed grading (Fig. 7d and e), whereas finer-grained pebble conglomerates and pebbly sandstone tend to show better developed grading and stratification that is picked out by alignment, layering and imbrication of clasts. Individual layers can be distinguished by their sharp bases, contrasting clast angularity and abrupt upward increase of clast size. Each of the tabular breccio-conglomeratic bodies terminates at a low-angle oblique clinoform surface (Fig. 7b and e).

Clasts in pebble-sized conglomerates have a red mudstone matrix that forms a coating of millimetre-thick undulatory mudstone laminae that wraps or drapes rounded pebble-sized clasts and surfaces but does not form coatings of individual clasts (Fig. 7a–c). Some clinoform surfaces within breccio-conglomeratic bodies have also been coated by the red mudstone (Fig. 7b). Mudstone coatings sometimes include centimetre-thick layers of moderately sorted, sub-angular to subrounded coarse sand-sized grains (Fig. 7c). These sand layers are often undulatory with sharp, occasional scoured bases and transitional tops. No desiccation features or any rip-up clasts of the red mudstone have been found.

Calcareous laminated red mudstone interfingers with and overlies the breccio-conglomerate bodies; in some places, it lies directly on the Lewisian palaeosurface (Fig. 7e). It forms a thick drape of clinoform surfaces at the termination of breccio-conglomerate bodies, thickening away to form a body of laminated calcareous mudstone several metres thick (Fig. 7e and f). The total thickness, lateral extent and the nature of the upper surface of the calcareous red mudstone is not known because it has been truncated by the unconformity at the base of the Diabaig Formation (Fig. 7f).

The large-scale structure of the calcareous red mudstone comprises numerous broad, symmetrical low-relief, overlapping convex-upward domical structures, up to 5–10 m across and 0.25–0.50 m in height (Fig. 7f–i). The internal carbonate layers tend to mimic the external form of these domical structures (Fig. 7g–i). Domical structures tend to be centred around individual cobbles and boulders forming smaller, higher relief domes that project upwards from the breccio-conglomerate layer (Fig. 7g). The carbonate and mudstone layers coating some clasts are subvertical to overhanging in attitude (Fig. 7g). These domes expanded beyond the initial clast, becoming much broader and of lower relief, and overgrow any adjacent domical structures (Fig. 7g–i). Much of the upper surface of the calcareous red mudstone comprises numerous low-relief overlapping domes several tens of metres in lateral extent and between 1.5 and 2.0 m in thickness (Fig. 7h and i). No desiccation structures were found in the laminated red mudstone.

The laminated calcareous mudstone comprises millimetre- to centimetre-thick layers of red mudstone alternating with irregular, elongate to equant-shaped calcite bodies concentrated in layers (Fig. 8a and b). The larger irregular carbonate bodies are up to 3 cm in length and 2 cm in thickness; their bases are flat to slightly convex-down, whereas the upper surfaces tend to have low-relief convex-upward forms (Fig. 8b). The larger carbonate bodies are surrounded by smaller equant and tabular-shaped bodies that appear to be fragments of the larger bodies (Fig. 8a and b). Some of the latter show a preferential sub-parallel alignment (Fig. 8e and f). The matrix consists of semi-opaque siliciclastic red mudstone with scattered silt- and fine sand-sized sub-angular detrital quartz and feldspar grains. It is non-luminescent and occurs as pervasive wispy microstylolitic pressure dissolution seams at the margins of the carbonate bodies, sometimes truncating their internal fabric (Fig. 8c and d).

The carbonate bodies are composed of two microfabrics (Fig. 8e and f), as follows.

• Clotted micrite: these are dense to dispersed aggregates of rounded to irregular peloids up to 500 µm in size composed of homogeneous micrite that has a mottled to speckly moderate yellowish orange luminescence (Figs 8e and 9). The pore space between the clotted micrite is infilled by zoned calcite spar. The clotted micrite forms irregular-shaped rounded bodies up to 5–10 mm in size.

• Reticulate micrite: these form a reticulate network of planar to curved micrite bridges with zoned calcite spar-filled pores between them (Figs 8f and 10). These are often broken down into plate-shaped fragments of micrite. The micrite elements often have a preferential vertical alignment (Fig. 8f), corresponding to upright projections in the upper surface of the carbonate body. The micrite is homogeneous and has the same mottled to speckly moderate yellowish orange luminescence as the clotted micrite.

Both types of carbonate bodies sometimes have a laminated fabric that appears to form on the outer surface of the body (e.g. Figs 8e and 10a, b). Internal pores within both clotted and reticulate micrite are lined by scattered crystals of brightly luminescent calcite followed by a phase of pore-filling non-luminescent blocky calcite spar (Figs 9 and 10). A set of sub-parallel microfractures infilled by brightly luminescent zoned calcite often exploit pre-existing crystal boundaries and cross-cut pressure dissolution seams between carbonate bodies (Figs 9a, b and 10a, b). Any remaining porosity is infilled by brightly luminescent calcite cement (e.g. Fig. 10c and d).

Stewart (2002) interpreted the thickly bedded structureless sandstone that under- and overlies the two logged sections (localities 1 and 2) as a lake basin facies. The two logged sections are interpreted as emergent to very shallow water lake margin facies (Fig. 11a).

The predominance of fine-grained laminated sediment indicates deposition in overall low-energy conditions. The symmetrical bimodal ripples are interpreted as wave ripples, resulting from wind action in shallow water rather than ripples formed in a unidirectional current. Cut and fill structures are interpreted as minor scours, possibly associated with the distal parts of small channels. The convolute and chaotic bedding and the foundering of ripples and cut and fill structures indicate dewatering as a result of deposition on unconsolidated sediment. These were deposited in shallow water, possibly not more than a few centimetres deep at a lake margin. Fine and medium sandstone layers with tabular mudstone clasts may indicate redeposition of desiccated muddy overbank or lake margin sediment during occasional floods. Polygonal fractures infilled by sand indicate desiccation and emergence, and imply deposition in higher parts of the lake margin prone to episodic exposure. The sinuous nature of the fractures suggests that they formed in soft sediment and were compacted.

The stringers of silt- and fine sandstone between low-relief laterally linked domes on mudstone and siltstone bedding planes are interpreted as infills of microchannels possibly not more than a few millimetres deep. This suggests that the domes formed micro-topographic relief and were resistant to erosion, implying that the sediment was bounded by either a microbial mat or a biofilm. The associated polygonal sinuous fractures are interpreted as desiccation fractures and indicate fluctuating water levels in very shallow water. The polygonal network of ridges seen in bedding plane exposures of carbonate sheets are also interpreted as desiccation structures, but the lack of associated fractures suggests that they represent desiccation curls that form when a sediment bounded by a cohesive mat undergoes shrinkage as it dries out (e.g. Shinn 1983).

The steep, sometimes vertical margins of both smooth and cuspate domes suggest that the sediment was bounded and was able to support steep depositional slopes. The onlap of the domes by layered sediment suggests that they were depositional features and the similarity of the form of internal carbonate laminae with their external morphology and the thinning of carbonate layers towards the margins of the domes suggests that that they are constructional features. The general symmetry of the smooth and crinkly domes suggests that they formed in standing or gently agitated water.

Brasier et al. (2017, 2019) interpreted the carbonate sheets and domes to be dewatering structures, in some cases modified by desiccation. Although dewatering structures are present in associated sediments, the majority of domical structures are onlapped by undisturbed planar laminated sediment. The smooth domes also have a consistent sub-symmetrical convex-up attitude with sub-parallel tops, bases and internal lamination. Their alternating back-stepping and progradational morphology (e.g. Fig. 4b) cannot be explained by dewatering. Cuspate domes (e.g. Fig. 4c) have flat or gently undulatory bases above which the internal lamination shows a progressive upward tightening of any apparent ‘folds’. The carbonate laminae are continuous across these structures with no penetration in the cores of the upright ‘folds’ that might have been the result of upward injection of fluidized sediment. Further, foundered ripples and convolute lamination in dewatered layers sometimes contain overturned or recumbent folds (e.g. Fig. 3e), geometries that have not been observed in either smooth or cuspate domes. The general convex-up geometry and the outward dip of laminae at the margins of the domical structures are also inconsistent with foundering of unconsolidated sediment and the upward injection of fluidized sediment.

The onlap of smooth and cuspate domes and the similarity of the form of the internal laminae and their external morphology surrounding sediments suggests that they were constructional structures that formed areas of positive depositional relief. The vertical and steep slopes of margins of some domes suggest that the sediment was bounded. The alternating back-stepping and progradational morphology of domes and the alternations between planar and crenulate growth forms (e.g. Fig. 4a) that occur at the same level in adjacent domes indicate a biological origin.

A depositional model is shown by Figure 11b. Interval 1 is interpreted as a lake margin environment with planar and ripple-bedded sandstone, siltstone and mudstone deposited in an overall low-energy setting in shallow water, probably not more than a few centimetres in depth. Horizons of polygonal fractures represent episodes of desiccation indicating fluctuation of the water level. The acicular and chevron-shaped pseudomorphs are interpreted as evaporite crystals that grew in the sediment, implying concentration of water by evaporation. The tufted carbonate sheets with their consistently flat bases and upward conical projections onlapped by sediment are interpreted as calcified mats in agreement with Upfold (1984) and not as replaced evaporites (Stewart 2002; Brasier 2011) that would have produced more irregular nodular forms. Laterally linked symmetrical domical structures seen on siltstone and mudstone bedding surfaces are associated with desiccation surfaces, suggesting that they formed in very shallow water to emergent conditions in shallow lake margin settings. Their symmetrical, subrounded dome-like structures suggest that these formed by constructional growth that was resistant to erosion. The beds of structureless fine sandstone are interpreted as unconfined mud flows or high-density flows deposited on a lake margin.

The thicker laminated carbonate structures in intervals 2 and 3 are interpreted to have developed by the progressive amalgamation and stacking of individual carbonate mats. The mimicking of external morphology and the offlapping and amalgamation of internal domical structures are interpreted as constructional features. The carbonate unit at the top of interval 2 is interpreted as a biohermal structure that developed in an abandoned channel. The carbonate unit that caps interval 3 is interpreted as a biohermal structure several metres thick and several tens of metres in lateral extent that formed in water several metres deep.

Stewart and Parker (1979) and Parnell et al. (2015) interpreted the presence of evaporite pseudomorphs, elevated boron in illite and molybdenum in this section to indicate an evaporitic lacustrine environment. Stüeken et al. (2017) proposed a marginal marine setting based on flaser bedding and herringbone cross-stratification with mud drapes; however, bimodal cross-bedding has been found at other localities where it has been interpreted as the result of wave reworking in a lake margin setting. Stüeken et al. (2017) also presented geochemical evidence for marine influence in this section based on isotopically enriched molybdenum (+1.19‰ $δ$⁹⁸Mo) and the ⁸⁷Sr/⁸⁶Sr ratio of 0.707 in carbonate indicating precipitation from mixed freshwater and marine water. The depositional setting is interpreted as a marine-influenced lake margin or part of an estuary.

The very coarse clast size, angularity and poorly sorted nature of the breccio-conglomerates and the offlapping geometry with clinoforms suggest that they represent offlapping proximal fan-delta lobes (Fig. 11c). Clinoforms within and at the termination of these bodies are interpreted as the intermediate or final positions of the fan-delta lobe during episodic progradation.

The coated conglomerate is interpreted as a fluvial facies deposited on the topographically flat inner fan area, possibly feeding laterally into a fan delta. The lack of desiccation features within the calcareous red mudstone suggests that the coating formed in a submergent setting. The occurrence of finer-grained sandy layers interbedded with the red mudstone coating is interpreted as periods of lower energy discharge. The calcareous red mudstone is not a concentric oncolitic coating of individual clasts (e.g. Flügel 2004) but is interpreted as a superficial binding of the alluvial surface that formed during periods of non-deposition in an otherwise high-energy setting. Clinoform surfaces were also coated by the red mudstone during periods of fan-lobe inactivity or following abandonment (e.g. Figs 7b and 11c).

The calcareous laminated red mudstone forms a thick coating of abandoned fan-delta lobes forming an offlapping biohermal structure composed of broad overlapping domes. A veneer of rounded cobbles was deposited on the lithified surface of the bioherm with some cobble-sized clasts that settled between domes (Fig. 7g). These clasts were then coated by the laminated calcareous mudstone, some of which was near-vertical to overhanging, to form domical structures centred on the clasts that further expanded to produce a larger structure of numerous broad overlapping domes.

The similarity between the form of carbonate laminae and the external mounded and undulatory geometry of the structure indicates that these domical structures represent a series of constructional structures. The symmetry of the domes suggests that there was no preferential current. These laminated domical structures at Enard Bay are of similar dimensions to domical structures that occur as drapes over cobbles and flat and contorted mats in the Mesoproterozoic Copper Harbor Group described by Elmore (1983), Wilmeth et al. (2014) and Fedorchuk et al. (2016).

The presence of calcite cement infilling internal pores within the clotted and reticulate micrite indicates that the carbonate bodies were originally porous and the original texture is interpreted as a framestone. The preferential alignment of the reticulate fabric and the presence of laminated micrite at the margins of some carbonate bodies suggest that they may represent biologically mediated production that formed calcified structures. The framestone fabric also suggests that the carbonate bodies were brittle during deposition and were unlikely to survive reworking and transportation. They are therefore unlikely to have been detrital clasts that accumulated on the laminated structures. The laminated calcareous red mudstone is interpreted as a composite biohermal structure that developed by local biologically mediated production of carbonate and binding and stabilization of sediment represented by the red mudstone. The bioherms nucleated over the terminations of fan lobes and adjacent parts of the submerged Lewisian surface at a lake margin. The thickness of the structure suggests that the lake was several metres deep at this point (Fig. 11c).

Facies analysis of associated sediments indicates that the laminated carbonate structures formed in water depths ranging from a few millimetres to several metres deep in settings that included very shallow water freshwater and marine-influenced lake margins that were episodically emergent, abandoned channels and draping the terminations of fan-delta lobes that prograded into a lake. Stewart (2002) and Brasier et al. (2017, p. 133; 2019, p. 314) considered the laminated carbonate structures to be of clastic origin incorporating detrital clasts of metamorphic carbonates but did not expand on possible mechanisms; however, the following observations point to biological accretion of these structures:

• onlap of sediment indicates that the domes formed depositional structures with topographic relief (e.g. Fig. 4a and b);

• mimicking of internal geometry of the laminae with external form indicates that they represent constructional features (e.g. Figs 4b, 6g, h and 7h, i);

• progressive offlapping morphology and progressive expansion and amalgamation of domes forms larger structures (e.g. Figs 6h and 7g–i);

• back-stepping alternates with progradational domes (e.g. Fig. 4b);

• broad smooth domes alternate with smaller isolated domes (e.g. Fig. 4b);

• carbonate laminae within domes often show morphological variations throughout the development of an individual dome that can sometimes be matched between adjacent domes, suggesting that they may have been subject to external environmental changes (e.g. Fig. 4a).

Fedorchuk et al. (2016) outlined further criteria for the possible biogenic origin of fluvio-lacustrine stromatolites from the Mesoproterozoic Copper Harbor Conglomerate in Michigan. They included evidence for sediment binding such as crinkly carbonate fabric that traps and preserves detrital grains beyond their natural angle of repose. The crinkly carbonate fabric forming cuspate domes at Clachtoll Bay shows upward projections with steep, sometimes subvertical sides that include trapped detrital grains (e.g. Fig. 4c). Smooth domes and biohermal structures also have some steep-sided to subvertical and overhanging margins (e.g. Figs 4a and 6g) that require a binding mechanism to support such slopes. Smooth domes are made up of carbonate microspar laminae alternating with mudstone and siltstone layers suggest that binding of detrital sediment as well as carbonate precipitation is required for the formation of these domes (e.g. Fig. 4a). Fedorchuk et al. (2016) also mentioned asymmetrical growth in response to light and the presence of fenestrae resulting from the photosynthetic production of O₂ as indicators of biogenesis, but these features were not observed.

Carbonate fabrics described here include microspar and clotted and reticulate textures; the latter two forming a framestone texture were regarded as biogenic indicators by Fedorchuk et al. (2016). In contrast, microfabrics such as radial fibrous calcite fans, radiaxial fibrous and fascicular optic calcite or microfabrics characteristic of carbonate cement resulting from abiogenic inorganic or physicochemical precipitation were not observed (e.g. Flügel 2004; Bosence et al. 2015; Fedorchuk et al. 2016). Kah and Bartley (2021) recognized a range of microfabrics in Mesoproterozoic carbonates and distinguished tufted, pustular and irregular micritic laminae, equivalent to the tufted, clotted and reticulate fabrics described here, as associated with microbial fabrics often associated with stromatolites. Mattes and Conway Morris (1990), Le Ber et al. (2015) and Mettraux et al. (2015) also described microspar laminae and clotted and reticulate micrite from Neoproterozoic microbial stromatolites and thrombolites from Oman that resemble the textures described from the Stoer Group. A further point of similarity is that standard and CL petrographic studies by the above-cited researchers showed that the clotted and reticulate micrite masses form a framework cemented by calcite spar or with preserved porosity. Other examples of clotted and reticulate micrite have been described from microbial stromatolites and thrombolites from the Eocene Green River Formation, the recent Great Salt Lake in Utah and Lake Thetis in western Australia (Grey et al. 1990; Chidsey et al. 2015; Baskin et al. 2022).

The carbonate component is interpreted to form a template around biofilms formed by filamentous and coccoid cyanobacteria or nucleated on extracellular polymeric substances by the interaction of microbial metabolism with supersaturated carbonate pore fluid (e.g. Riding 2002, 2008; Bosence et al. 2015). The main mechanisms of precipitation are interpreted as biologically induced calcification and biologically influenced calcification (Dupraz et al. 2009). Reid et al. (2000) and Lan et al. (2020) suggested that modern and ancient marine stromatolites represent a dynamic balance between sedimentation and the growth and intermittent lithification of biofilms formed by filamentous and coccoid cyanobacteria. It is possible that some of the structures in the Stoer Group may have formed by a combination of calcifying and non-calcifying microbial mats. The variations of microfabric may also reflect varying pore fluid chemistry or differing microbial assemblages. Geochemical evidence for possible microbial activity during deposition of the Stoer Group was also provided by Parnell et al. (2010), who attributed S isotope fractionation to microbial sulfate reduction, and Spinks et al. (2010), who identified vanadium-rich micas at the core of reduction spots in shales as a proxy for microbial activity.

Brasier et al. (2017, p. 133; 2019, p. 314) also proposed that the carbonate bodies in the Enard Bay bioherm represent detrital clasts reworked from metamorphic carbonates. However, the carbonate microfabric is characteristic of microbially mediated precipitates and does not display annealed textures, which would be expected in reworked metamorphic carbonates. The clumped isotope temperature of 160°C obtained by Brasier et al. (2019) indicates that these carbonates have not reached the onset of low-grade metamorphism defined as 200–250°C (Immenhauser 2022) but is more likely to reflect later Mesoproterozoic burial (Stewart 2002) and not an inherited metamorphic temperature.

The laterally linked domes on mudstone and siltstone bedding planes (e.g. Figs 3h and 6b) may provide evidence of sediment stabilization by microbial mats or biofilms that support the domical topography and provide resistance to erosion with the preservation of sediment in microchannels between domes. Desiccated carbonate sheets showing polygonal ridges (Fig. 3g) are also interpreted as indicators of sediment binding. The coated conglomerate also shows evidence of sediment binding of sand layers between pebbles, fluvial surfaces and inclined surfaces of the intermediate or abandonment surfaces of fan-delta lobes. The origin of these features is uncertain but has been attributed to microbially induced sedimentary structures (e.g. Prave 2002; Brasier et al. 2017, 2019; McMahon and Davies 2018), implying that they result from the biostabilization of sediment by the presence of biofilms and microbial mats as defined by Noffke (2009). However, the biological significance of these structures has not been conclusively proven and the term sediment surface texture is used instead (e.g. Davies et al. 2016).

The difference between the development of sediment surface texture as opposed to stromatolites may have been the level of carbonate saturation of the groundwater and surface water. Microbial communities may have been able to fix carbonate only in porewater that was sufficiently supersaturated with calcium carbonate. An alternative explanation is that sediment surface textures were formed by non-calcifying microbial communities whereas stromatolites represent a differing assemblage of microbial species that were able to calcify. Grey et al. (1990), Chidsey et al. (2015) and Baskin et al. (2022) described modern saline lakes including Lake Thetis in Western Australia and the Great Salt Lake in Utah that are host to a wide range of microbial communities restricted to differing sub-environments, each giving rise to a range of sediment surface textures and stromatolites.

The supply of dissolved sulfate and carbonate is a prerequisite to support any possible microbial metabolism and biologically mediated calcification. This section addresses the nature and possible origins of depositional and early diagenetic pore fluids during deposition of the Stoer Group.

Wacey et al. (2017) showed that walls of microfossils extracted from the Stoer Group have been pyritized, suggesting the presence of a sulfate-rich, phosphate-poor pore fluid. Parnell et al. (2014) described early diagenetic greenockite in the Poll a'Mhuilt Member at the Bay of Stoer. Formation of greenockite requires an anoxic pore fluid enriched in cadmium, suggesting that the low partial pressure of oxygen in the Precambrian atmosphere was insufficient to have produced near-surface oxic diagenesis (Dyrssen 1988; Parnell et al. 2014). Wacey et al. (2014, 2017) suggested that there was sufficient partial pressure of oxygen in the atmosphere to cause oxidative weathering of sulfides; however, local conditions, such as high rates of erosion or deposition, may have allowed transport and preservation of detrital sulfides and magnetite (e.g. Stewart 1991, 2002; Reinhard et al. 2009). Potential sources of sulfate in wholly freshwater environments may have been by oxidative weathering of detrital and in situ sulfides in the Lewisian (e.g. Jones et al. 1987; Strachan et al. 2012; Park 2022). Dissolved sulfate in the basal Poll a'Mhuilt Member at the Bay of Stoer may have been sourced additionally from marine water.

Parnell et al. (2014) indicated that groundwater during deposition of the Stoer Group was carbonate-rich and suggested that early oxidation of organic matter was a possible source of carbonate. Brasier et al. (2019), using Raman spectroscopy, showed that the Stoer Group contains kerogenous organic carbon and externally sourced detrital graphite. Shales, such as those of the Poll a'Mhuilt Member with an average total organic carbon content of 0.24–0.37 wt%, may have provided the required organic matter during deposition (Stewart 2002; Stüeken et al. 2017). Cerling (1994) emphasized the importance of arid weathering of basic and ultrabasic volcanic rocks to create high carbonate alkalinity in present-day African Rift Valley lakes. Although there are no volcanic rocks in the hinterland of the Stoer Group, semi-arid weathering of basic and ultrabasic Scourian dykes may have generated carbonate-rich groundwater. Marine water may also have provided a source of dissolved carbonate either directly in marginal marine settings or in marine–meteoric mixing zones. Bartley and Kah (2004) estimated that Proterozoic marine water contained between two and 10 times more dissolved inorganic carbon compared with present-day ocean water, which would have sustained elevated carbonate saturation.

Smooth and cuspate domes and biohermal structures composed of laminated carbonate and fine siliciclastic laminae occur in the Mesoproterozoic Stoer Group in submerged and desiccated lake margins and at the termination of fan-delta lobes where they prograded into lakes. Evidence for the biological origin of these structures includes the following:

• steep, subvertical and locally overhanging margins of domes and bioherms indicating sediment binding;

• mimicking of internal geometry with external form indicating that these structures represent constructional features;

• alternation of back-stepping with progradational morphological forms;

• progressive offlapping morphology and progressive expansion and amalgamation of domes to form larger structures;

• microspar, clotted and reticulate carbonate fabrics forming a framestone texture that resembles documented microbial carbonate fabrics.

The smooth and cuspate domes and biohermal structures fulfil the definition by Riding (1999, 2011a, b) of microbial stromatolites as ‘macroscopically layered authigenic microbial sediments with or without abiogenic precipitates’ and thus represent the oldest microbial stromatolites in the British Isles.

The laminated carbonate structures are associated with sediment surface textures including laterally linked domes on mudstone bedding planes, desiccated carbonate sheets and coated conglomerates. These also imply sediment binding; however, it is not clear to what extent any biological influence was involved. Sediment surface textures may have developed instead of stromatolites because of the level of carbonate saturation of the groundwater and surface water. Microbial communities may have been able to fix carbonate only in porewater that was sufficiently supersaturated with calcium carbonate. Alternatively, differing assemblages of microbial species may have been responsible for sediment surface textures and stromatolites.

Figure 12 shows the laminated carbonate structures and sediment surface textures in terms of their depositional fabric on the triangular diagram showing the inferred processes of microbialite formation (after Riding 1977; Burne and Moore 1987; Bosence et al. 2015). The smooth and cuspate domes and the Bay of Stoer biohermal stromatolites contain layers of detrital sediment alternating with microbially mediated carbonate layers and represent structures with biologically influenced carbonate precipitation combined with trapping and binding of detrital sediment.

Dissolved sulfate was supplied by weathering of detrital and in situ sulfides derived from the Lewisian hinterland, and carbonate may have been supplied by semi-arid weathering of basic and ultrabasic Scourian dykes. Marine–meteoric mixing zones are an additional possible source of dissolved sulfate and carbonate.

I thank editor A. Dunhill and reviewers W. McMahon and P. Wright for their time and comments; also J. Garland and A. Horbury of Cambridge Carbonates Ltd, who allowed use of their cathodoluminoscope. This paper is dedicated to my Leeds University mentors H. Clemmey, M. Leeder and the late M. Coward.

PG: formal analysis (lead), writing – original draft (lead)

The author declares that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The only permanent data generated by this study are a set of thin sections. I am happy to make these available to interested parties.