Arumberia is an enigmatic sedimentary surface texture that consists of parallel, sub-parallel or radiating ridges and grooves, most commonly reported from upper Neoproterozoic – lower Palaeozoic strata. It has variably been interpreted as the impression of a small metazoan, a ‘vendobiont’, a physical sedimentary structure formed on a substrate with or without a microbial mat covering, or a non-actualistic microbial community. In this paper we contribute new insights into the origin of Arumberia, resulting from the discovery of the largest contiguous bedding plane occurrence of the texture reported to date: a 300 m2 surface in the lower Cambrian Port Lazo Formation of Brittany, NW France. We compare the characteristic features of Arumberia at this locality with 38 other global records, revealing four defining characteristics: (1) the three-dimensional (3D) morphology of exposed Arumberia lines (either positive relief ‘ridges’ or negative relief ‘grooves’) records fully preserved cords within clay laminae; (2) lines may transition laterally into reticulated patterns; (3) characteristic parallel and sub-parallel Arumberia lines can become modified by desiccation on emergent substrates prior to interment; and (4) Arumberia are streamlined with palaeoflow in successions showing evidence of unidirectional currents, but are organized parallel to ripple crests where strata were sculpted by oscillatory flows. These characteristics indicate that Arumberia records a 3D entity, distinct in material properties from its host sediment, which occurred in very shallow water settings where it was prone to passive reorganization in moving water, and desiccation when water drained. A literature survey of all known Arumberia occurrences reveals that the most reliable examples of the form are stratigraphically restricted to a 40 Ma interval straddling the Ediacaran–Cambrian boundary (560–520 Ma). Together these characteristics suggest that Arumberia records the remains of extinct, sessile filamentous organisms (microbial or algal?) that occupied very shallow water and emergent environments across the globe at the dawn of the Phanerozoic Eon.

Arumberia is a sedimentary surface texture comprising a series of low-relief, sharply defined sub-parallel lines (Fig. 1), which is often recognized on Precambrian and lower Palaeozoic bedding planes. It was originally reported from the Ediacaran Arumbera Sandstone of Australia by Glaessner & Walter (1975), who interpreted it as an impression fossil of a small cup-shaped metazoan. Subsequent Arumberia discoveries have shown that the cup-shaped gross morphology of the original specimens is not indicative of its true form, which can extend over multiple square metres on planar or rippled bedding planes (Fig. 1i; Table 1). The characteristic features of Arumberia have been explained in several mutually exclusive ways. Dominant among these are hypotheses that the texture is either the ‘body-fossil’ impression of a mat-like organism or community attached to a sedimentary substrate (Bland, 1984; Kumar & Pandey, 2008, 2009; Kolesnikov et al. 2012, 2015; Arrouy et al. 2016), or a physical sedimentary structure formed with or without a microbial mat covering (Brasier, 1979; McIlroy & Walter, 1997; Jensen et al. 2005; McIlroy et al. 2005; Allen & Leather, 2006; Seilacher, 2007; Netto, 2012; Liu et al. 2013; Sharma & Mathur, 2014; Becker-Kerber et al. 2020). Notable among previous work is the 1984 review paper by Bland, which described and reviewed all prior known Arumberia occurrences. Bland concluded that only a fraction of reported Arumberia specimens exhibited the original diagnostic criteria of Glaessner & Walter (1975), and that they represented large impression fossils restricted to uppermost Precambrian and lower Cambrian strata deposited in very shallow-marine to non-marine environments.

In this study we critically assess the nature of Arumberia using new evidence from the largest reported instance so far discovered in the global rock record: a 300 m2 bedding plane in the lower Cambrian Port Lazo Formation of Brittany, France (Fig. 2). Isolated Arumberia have previously been reported from this locality (Bland, 1984), but the extensive low-lying coastal outcrop we describe here was fortuitously exhumed during an interval when the modern beach profile was lowered. Supporting this case study, we revisit Bland’s (1984) assessment of Arumberia in light of 22 subsequent reports, in addition to the 18 instances that he previously reviewed (Tables 1, 2). Our literature survey has been ground-truthed by field visits to six of these Arumberia-bearing formations (Table 1; the Arumbera Sandstone of Northern Territory, Australia; the Crown Hill, Ferryland Head, Gibbett Hill and Maturin Ponds formations of Newfoundland, Canada; and the Synalds Formation of Shropshire, England). These sites have provided further observations to supplement evidence from the Port Lazo bedding plane and support our interpretation of Arumberia.

Universal diagnostic criteria for Arumberia have proved challenging to isolate (Glaessner & Walter, 1975; Bland, 1984; McIlroy et al. 2005; Kumar & Pandey, 2008, 2009; Kolesnikov et al. 2012; Retallack & Broz, 2020), resulting in continuing uncertainty surrounding its affinities and origin (Table 3). Three factors have hampered understanding, described in the following.

  • (1) Preservation and ease of observation: it is challenging to describe individual specimens of Arumberia because the lateral margins of the texture are commonly either indistinct, or incomplete at the scale of observation. This makes it difficult to accurately define its areal extent. Additionally, while sub-parallel to parallel lines constitute all reported Arumberia (Fig. 1), the morphology and relief of these can be highly variable even over short distances, with spatially patchy expressions of branching, spacing, curvature and reticulation. Compounding these practical issues of description, Arumberia frequently has a topographic relief of < 1 mm, meaning that its visibility in natural outcrops can be hampered by poor light conditions, a lack of magnification and degradation through weathering.

  • (2) Changing diagnostic criteria and loaded terminology: the original type material of Arumberia (Glaessner & Walter, 1975, redescribed by Retallack & Broz, 2020) lacks the full range of characteristics and spatial extents that have been revealed and incorporated into taxonomic diagnoses by later discoveries. Furthermore, some characteristics originally thought to be universal (e.g. conical forms with blunt apices, regular radial patterns of lines) are now known to be ancillary features, present in only a subset of instances. Arumberia has been reinterpreted several times as a body fossil (e.g. Retallack & Broz, 2020) or a pseudofossil (e.g. McIlroy et al. 2005), with resultant emended diagnoses changing to reflect those interpretations through the addition or removal of diagnostic features. As a result, some of these criteria employ terminology that is loaded with an implication of biogenicity, rather than passive descriptive terms (e.g. description of the constituent lines as ‘rugae’; Kolesnikov et al. 2015, 2017; McMahon et al. 2017), further hindering objective interpretation.

  • (3) ‘Arumberia’ as a bucket-term: the most frequently observed characteristic of Arumberia is a set of closely spaced sub-parallel to parallel lines on a bedding plane (Fig. 1). Interpreting such a simple form in isolation encounters the challenge of equifinality: such phenomena may have plural potential explanations, both biotic and abiotic, and the cause behind a specific feature may be underdetermined by available evidence (e.g. Davies et al. 2016, 2020). There appear to be instances in existing literature where the term ‘Arumberia’ has been liberally used to describe any observations of (sub)parallel lines: in some of these instances, the forms share no further characteristics in common with other described Arumberia (e.g. in terms of physical dimensions, accessory characteristics or host facies) (Table 1).

Below, we attempt to redress these outstanding issues by presenting observations of exceptionally extensive Arumberia from a newly discovered bedding plane in the lower Cambrian Port Lazo Formation in France, augmented with observations of well-preserved examples from elsewhere in the rock record. We use these observations to amend the diagnostic criteria of Glaessner & Walter (1975) in light of those characteristics that are universal, common or infrequent attributes of Arumberia, and highlight previously reported instances of ‘Arumberia’ that can be considered doubtful. Our reassessment shows that the least equivocal instances of Arumberia exhibit a narrow stratigraphic and palaeoenvironmental range (cf. Bland, 1984).

Previous reports of Arumberia from the Port Lazo Formation in Brittany, France (Bland, 1984; Davies et al. 2016; McMahon et al. 2017; Went, 2017, 2021) have not described the feature in detail. The Port Lazo Formation is a constituent of the ‘Series Rouge’, a predominantly red-bed succession that crops out in coastal exposures across NW France and the neighbouring Channel Islands (Fig. 2a) (McMahon et al. 2017). Multiple potential ages have been suggested for the Port Lazo Formation. The French Geological Survey maps it as Early Ordovician in age, based on correlation to stratigraphy at Crozon that lies 125 km to the south across the North Armorican Shear Zone in the central Armorican domain (Egal et al. 1995). This correlation is consistent with the Port Lazo Formation being bounded by underlying Brioverian metasediments, which have recently yielded zircon U–Pb radiometric dates of 551 ± 7 Ma and 540 ± 5 Ma (Gougeon et al. 2021), and a 472 ± 5 Ma age from overlying andesites (Fig. 2a) (Auvray et al. 1980). However, regional variability in tectonic setting and sedimentary environments favours an early Cambrian age for the Port Lazo Formation, with deposition having commenced in extensional basins shortly after the Cadomian orogenic climax c. 540 Ma (Peucat, 1986). Three lines of evidence support this contention. (1) The strata had a source area to the west (McMahon et al. 2017), which is consistent with regional palaeogeography during the Cambrian Period (Went & McMahon, 2018; Went, 2021), but unlikely during the Early Ordovician Epoch when surrounding areas witnessed the deposition of the well dated marine ‘Grès Armoricain’ (Paris et al. 1999; Dabard et al. 2007). (2) Red-bed sequences in Normandy, which may be correlated with the Port Lazo Formation (Went, 2021), underlie Cambrian Stage 3 (521–514 Ma) limestones (dated by the presence of the trilobite Bigotina Pillola, 1993) and Cambrian Stage 2 (529–521 Ma) siliciclastic strata. (3) Radiometric ages from overlying andesites (Auvray et al. 1980), which provide crucial support for a posited Ordovician age, are reliant on the igneous bodies being extrusive; however, the andesites have more recently been demonstrated to be a mixture of disconformable lavas and intrusive bodies (Went, 2017). Furthermore, the presence of Arumberia itself may provide circumstantial evidence for a Cambrian age once the global stratigraphic record of this feature is considered (see Section 5).

The majority of Series Rouge outcrops are exposed within the limbs of basin-wide synclines (Fig. 2a), but the studied section of the Port Lazo Formation occurs as an outlier to these, cropping out at Bréhec Bay (Fig. 2b). Sedimentological observations made over the course of this study, and previously, have shown that Port Lazo Formation specimens of Arumberia are restricted to very shallow-water facies that were intermittently subject to sub-aerial exposure (Davies et al. 2016; McMahon et al. 2017; Went, 2017, 2021).

The Port Lazo Formation is separated into two distinct members (Went, 2017), with the bedding plane central to this study occurring in the Port Lazo Lower Member (Fig. 2c, d). The surface has a pervasive covering of polygonal mud cracks (Fig. 2f), in addition to raindrop impressions and wave-ripple marks (Fig. 2g). These characteristics indicate that Arumberia occurs on a true substrate, that is, a bedding plane that faithfully records a synoptic topography that formed at the interface of sediment and water or air (Davies & Shillito, 2018, 2021). Its c. 300 m2 surface area provides an opportunity to assess Arumberia morphology across a far larger area than has previously proved possible (Fig. 2e). Approximately 50% of the bedding plane is covered by Arumberia (although contiguous individual Arumberia are likely less extensive than this, due to their occurrence within multiple successive < 1 mm thick clay laminae that are impractical to correlate across the entire surface). Additional surfaces bearing Arumberia are also present at higher stratigraphic levels in the Port Lazo Formation Upper Member (Fig. 2c). These examples, which are far more limited in spatial extent, were preserved in shallow subaqueous, nearshore marine settings (Went, 2017; Table 2).

3.a. Arumberia lines

The Port Lazo Arumberia consist of a series of parallel or sub-parallel lines, usually < 2 mm in relief, 0.5–1.5 mm wide and spaced 1–3 mm apart (Fig. 3). Line relief and spacing can be variable even across individual specimens. The lines are most frequently low sinuosity and can exhibit intermittent bifurcations (Fig. 3a–d, g, h). A subordinate expression is present where Arumberia specimens are radially arranged, curving gently away from a single apex through an angle of up to 40° (Fig. 3e, f). Arumberia lines can sometimes be seen to transition along their length into a reticulated pattern (Fig. 3i, j). Reticulated patterns cover areas of up to 30 cm2 and consist of intersecting lines that form irregularly shaped, sharp-edged polygons < 1 cm in diameter.

Arumberia lines frequently exhibit three-dimensionality to their preservation, implying that they originally had a cord-like form (in agreement with Bland, 1984; Becker-Kerber et al. 2021; see also Fig. 1h). On bedding surfaces, lines can occur in both positive epirelief (ridges) and negative epirelief (grooves), and both expressions of this relief can be seen along an individual line (a ‘ridge–groove’ transition) (Fig. 4). With magnification, they can be seen to be an artefact of partial weathering: the ridges record positive epirelief on the upper surface of the line-hosting lamina, and the grooves reflect casts of positive hyporelief in an underlying lamina (Fig. 4b). Ridge–groove transitions may reflect instances where Arumberia lines have been preserved in full relief between successive, very thin (c. 0.3 mm) clay laminae, or they may reflect moulding by sub-millimetre-thick laminae. In order to preserve the original three-dimensional (3D) morphology, Arumberia colonized the substrate during a net interval of sedimentary stasis, with subsequent deposition (in this instance of clay) covering Arumberia’s external surface, preceded by little or no scour (e.g. Davies & Shillito, 2021).

The three-dimensionality of the lines, and transitions from directed linearity to reticulated patterns, contradict previous assertions that Arumberia could be a physical sedimentary structure arising from unidirectional flow (i.e. a surficial texture registered on the substrate, such as current rills or flute marks: Brasier, 1979; Jenkins et al. 1981; McIlroy & Walter, 1997; Seilacher, 2007). Instead, these combined characteristics indicate the prior existence of a feature that was external to the sediment in which it is now hosted, and which responded variably along its length. The reticulation bears similarities to the patterning that emerges when buoyant filaments of bacteria or algae become entangled in standing or draining water (Shepard & Sumner, 2010; Davies et al. 2016; Cuadrado & Pan, 2018).

On the large bedding plane exhibiting Arumberia (Fig. 2d), the drainage of water and associated drying and emergence is indicated by the presence of polygonal mud cracks and raindrop impressions (Fig. 2f, g). The occurrence of raindrop impressions overprinted on Arumberia implies that Arumberia was an ineffective buffer against droplet impact (Davies et al. 2016) (see also Fig. 1j). When in proximity to a desiccated plate margin, Arumberia can be seen to have deformed alongside the shrinking cohesive muds, and individual lines are seen to curl to form irregular, broadly concentric circles (Fig. 5). These instances demonstrate that Arumberia was already present on the muddy surface before desiccation. Small positive epirelief ‘blisters’ were observed at the margins of the deformed Arumberia (Fig. 5f), which potentially also formed during emergence as water escaped from the surficial clay veneer or a microbial mat, producing transient bubbles that subsequently ruptured at the surface. That Arumberia ridges and grooves document discrete 3D forms is further supported by late Ediacaran specimens from Newfoundland, Canada, in which individual Arumberia lines are seen to consist of multiple, finer lines that can appear to be superimposed upon one another (Fig. 1d, h).

3.b. Arumberia orientation

Directional measurements of Arumberia lines have a tight spread of WNW–ESE orientations over the total area of 300 m2 and through < 0.1 m of vertical succession (Fig. 6). Palaeocurrent measurements taken throughout the succession indicate that dominant current flow was near perpendicular to this axial orientation (n = 61), and symmetrical ripple mark crest lines on overlying beds have near-identical strikes (n = 14). The evident spatial arrangement of Arumberia lines with respect to measured palaeoflow implies a hydrodynamic role in morphogenesis.

3.c. Association with dimple-pimple marks

Arumberia lines may be associated with 0.5–1.5 mm diameter circular features that lack internal structure (Figs 3d, h, 7). These features have reliefs of < 1 mm and can occur in dense populations consisting of hundreds of examples. They are generally found in negative epirelief (dimples) and positive epirelief (pimples) within < 1 mm thick clay veneers (reflecting full relief orbs, in the same way that Arumberia lines reflect full relief cords). One surface shows surficial dimples and basal pimples, effectively recording impression of clay laminae from above and below by two opposing hemispheres (Fig. 7d). Elsewhere, at least 10 fully illustrated and described global occurrences of Arumberia exhibit similar features (Table 1), which have variably been suggested to be resting cysts of Arumberia (Bland, 1984), structures which mark a distinct stage in the development of Arumberia (Becker-Kerber et al. 2020), independent body fossils assigned to the taxon ‘Beltanelliformis minutae’ (McIlroy et al. 2005), pseudofossils associated with the perforation of microbial matgrounds (Menon et al. 2016) or volcanic lapilli (McIlroy & Walter, 1997). While the features remain enigmatic, the local palaeogeographic context of the Port Lazo specimens eliminates the possibility of volcanic lapilli.

The original diagnosis of Arumberia provided by Glaessner & Walter (1975) was based on specimens from the Arumbera Sandstone, but 39 additional Arumberia reports from around the globe (Tables 1, 2), in addition to key characteristics recognized in the Port Lazo examples, necessitate modification of its diagnosis.

Arumberia: Glaessner & Walter 1975 (figs 6a–c, 7a–c, 8).

Emended diagnosis: Surface texture comprising parallel, sub-parallel or radiating lines 0.4–4.0 mm wide and spaced 0.25–30.0 mm apart. Lines can present in both positive and negative epirelief and hyporelief, with total line relief rarely exceeding 2.0 mm. Lines are commonly several centimetres long but may continue for tens of centimetres. Lines may bifurcate, but typically maintain constant width along their lateral extent. Edges are indistinct. Lines rarely overlap, but can transition laterally into reticulated networks.

Distribution: Arumberia has been described from every continent except Antarctica (see full list of localities in Table 2). Its temporal range is discussed in Section 5.

Discussion: Emended diagnoses for Arumberia and Arumberia banksi were recently presented by Retallack & Broz (2020), following re-examination of the type material of Arumberia banksi and other specimens from Central Australia. They interpreted features including septae, chambers and radiating ribs within their studied material, and included these within their emended diagnosis. We find no evidence of septae or chambers in any of the Arumberia specimens we have studied, and also find no reason to consider the radiating pattern of ribs to be a diagnostic feature. As such, we do not include these features in our emended diagnosis.

There has been disagreement regarding whether Arumberia should be considered a genus (Glaessner & Walter, 1975), arising from the uncertainty regarding its origin (see Section 6) and the absence of sufficient fossil material to permit more than a subjective hierarchical context. Recent formalization within the class Vendobionta (Seilacher, 1992; Retallack & Broz, 2020) is not followed here, since: (1) a discrete macroorganism interpretation for Arumberia is seemingly incompatible with its ability to express quite different morphological forms (e.g. linear, Fig. 3a–h; reticulated, Fig. 3i, j; curled, Fig. 5) dependent on the prevailing environmental conditions; and (2) we do not consider Vendobionta to be a valid clade (Dunn & Liu, 2019). McIlroy et al. (2005) opted to classify it as a pseudofossil under ‘microbially modified sedimentary structures’. While we agree that there is the potential for a microbial role in the formation of this surface structure, we argue that the 3D morphology of individual lines indicates that they preserve casts of fossil organisms, rather than modifications of sedimentary surfaces. Our interpretation of Arumberia as the remains of a distinct organism or population of organisms requires that it retains its Linnaean terminology.

Glaessner & Walter’s (1975) type material was given the species name Arumberia banksi. Five additional Arumberia species have since been proposed (A. vindhyanensis, ‘A. usvaensis’, ‘A. multykensis’, ‘A. beckeri’ and ‘A. ollii’: Kumar & Pandey, 2008; Kolesnikov et al. 2012). Presently, only A. banksi and A. vindhyanensis are considered valid (Kolesnikov et al. 2017), and these are differentiated by the frequency of line bifurcation. Arumberia banksi was most recently re-described by Retallack & Broz (2020), while the systematic description of Arumberia vindhyanensis can be found in Kumar & Pandey (2008). In the Port Lazo Formation, specimens that could be attributed to both of these species can be seen to grade into one another (compare Fig. 3a and Fig. 3g), suggesting that subdivision to species level may have limited biological meaning (see Section 6.d). Morphological differences between forms may instead stem from external hydrodynamic or environmental factors.

Arumberia has a narrow stratigraphic range, spanning the Neoproterozoic to lowermost Cambrian strata (Table 2), a trend that becomes particularly pronounced when equivocal reports of the surface texture are omitted (Fig. 8). These equivocal reports include 14 records where Arumberia is reported but not figured or described in a paper (i.e. with no evidence presented to verify or falsify the claim), and the following six instances that are here rejected as being Arumberia because they are either markedly dissimilar to both the original (Glaessner & Walter, 1975) and emended diagnoses (Section 4), or are insufficiently well preserved to confirm their identification as Arumberia.

  • (1) Bland (1984) proposed that a sedimentary surface texture illustrated by Miller (1975, fig. 2) from the Mesoproterozoic Auborus Formation of Namibia may be Arumberia. However, the original figure lacks the definition to confirm this, and Miller (1975) interpreted the texture as forming abiotically due to aeolian adhesion on a mudflat.

  • (2) Bland (1984) suggested that a specimen illustrated by Kaudern (1932, fig. 5) from the Neoproterozoic Visingsö Formation of Sweden may represent poorly preserved casts of Arumberia. Kaudern’s (1932) original figure shows only a small hand specimen with no sedimentological context, and the sample is therefore an unconvincing example of Arumberia given its poor preservation.

  • (3) Callow et al. (2011, fig. 10) reported and figured surface textures described as Arumberia from the early Neoproterozoic Diabaig Formation of Scotland. However, these surface textures differ from both the original and revised diagnoses of Arumberia, with lines being sporadically spaced, changing width along their lateral extent and showing a wide disparity of orientations even across small bedding planes. The interpretation of these specific structures by Callow et al. (2011) as forming due to the moulding of entwined microbial filaments appears robust, but they are dissimilar to classic Arumberia.

  • (4) Reported Arumberia from the Ediacaran Masirah Bay Formation of Oman were figured by Allen & Leather (2006, figs 14c, d) and Brasier et al. (2011, fig. 11b). Brasier et al. (2011) noted that the features extended for up to “hundreds of square kilometres” and suggested that they reflected the flow of bottom currents interacting with possible filamentous microbes or algae. In both publications the small dimensions of the figured specimens make an assured diagnosis difficult, but their apparent and regular intertwining of sinuous ridges is not present in any other known Arumberia specimens (Table 1).

  • (5) Arumberia reported from the Upper Cambrian Booley Bay Formation of Ireland (MacGabhann et al. 2007, fig. 15) do not have a uniform line thickness across each bedding plane, unlike Arumberia examples described elsewhere (Table 1). Additionally, many individual lines exceed 0.5 cm in diameter, greater than any other previously described Arumberia.

  • (6) Liu (1981) described Arumberia from a limestone bed of the Shichang Member in China, the only report from a carbonate lithology. The small figured specimen makes confirmation difficult, but the positive epirelief lines appear to be an order of magnitude greater in size than any other Arumberia reports (Table 1).

When these examples are excluded, all remaining reports of Arumberia-hosting strata date to 520–680 Ma in age (Fig. 8). When considering the overlapping ranges of posited ages, the stratigraphic range can be further refined to 520–560 Ma in age (this is also true for 12 of the 14 instances of reported Arumberia that were not figured in their original reports).

The limited range of consistently expressed morphological characteristics of ‘true’ Arumberia, and the inclusion of a wide variety of instances of clustered lines on bedding planes within the genus in the past, have resulted in ongoing debate regarding the biogenicity of this feature. Three lines of evidence are discussed here to interrogate these competing claims: similarity to abiotic forms, the tight global stratigraphic range and the preferred orientation of specimens.

6.a. Similarity to abiotic forms

Features with a passing resemblance to some Arumberia specimens were created abiotically by Dżułyński & Walton (1965) in flume tank experiments devised to study sedimentary features associated with deep marine flysches. Brasier (1979) was the first to note this similarity, and it has subsequently been invoked to contend that Arumberia was, at least in part, hydrodynamic in origin (e.g. Kumar & Pandey, 2008; Callow et al. 2011; Menon et al. 2016). We note that the structures developed by Dżułyński & Walton (1965) occurred within turbidity-current-generated sand, while the majority of Arumberia worldwide come from mudstones and siltstones deposited in emergent and nearshore marine environments (Table 1; a notable exception is the Masirah Bay Formation ‘Arumberia’ from lower shoreface/offshore facies (Allen & Leather, 2006) discussed in Section 5). Tangible characteristics of exposed Arumberia in the Port Lazo Formation (3D cord-like form, reticulated transitions, potential for cords to curl when subject to emergence) cannot, in concert, be explained by purely hydrodynamic mechanisms. Cohesive mud and granular sand have very different responses to hydrodynamic forces, and so the similarity to Dżułyński & Walton’s (1965) sedimentary structures is most likely an artefact of equifinality. Furthermore, the observation that exceptionally well preserved specimens comprise finer linear structures (Fig. 1d, h) rules out abiotic formation mechanisms that produce only surface impressions. The bundling of these cords is consistent with them preserving discrete filamentous organisms such as bacteria or algae, and is inexplicable by fluid dynamics alone.

6.b. Stratigraphic range

McIlroy & Walter (1997) used the tight stratigraphic range of Arumberia to suggest that it is an anactualistic fabric on sedimentary surfaces resulting from the interplay of microbial mats with physical hydrodynamic processes. However, this interpretation cannot explain the absence of Arumberia from Phanerozoic or lower Precambrian strata. Both of these intervals host abundant sedimentary surface textures attributable to the interplay of fluids and microbial mats (Davies et al. 2016). Arumberia is absent from comparable Phanerozoic sedimentary facies where bedding planes have received intensive study: for example, there are no reports of Arumberia in voluminous desiccated red mudrocks of the Siluro-Devonian Old Red Sandstone and equivalent strata, despite detailed investigations of surficial trace fossils (e.g. Gordon, 1988; Smith et al. 2003; Shillito & Davies, 2017) and microbial sedimentary surface textures (Davies et al. 2006, 2016; Marriott et al. 2013) in littoral facies. Similarly, Arumberia is also unknown from pre-Ediacaran successions that have been interrogated for microbial sedimentary structures (e.g. Eriksson et al. 2012; Lan et al. 2013; Beraldi-Campesi et al. 2014).

Given that no physical hydrodynamic fluid processes are likely to have been unique to the Ediacaran–Cambrian transition, the most parsimonious explanation for the stratigraphic restriction of Arumberia (Fig. 8) is that there was a dominant biological component to its formation (Bland, 1984; Kumar & Pandey, 2008, 2009; Kolesnikov et al. 2012, 2015, 2017; Arrouy et al. 2016). This is also implied by the variability in individual Arumberia linearity (Fig. 5), reticulation (Fig. 3i, j) and spacing (Fig. 3a–h) observed within the large bedding plane studied at Port Lazo. Such morphological diversity is challenging to account for solely through hydrodynamic processes, but can be explained as different manifestations of a living system.

Forms resembling Arumberia have been observed on the surface of biofilms in modern salterns, with radiating and curved macroscopic ridges interpreted to form as a microbial community responds to changing environmental pressures (Kolesnikov et al. 2017). While these extant ‘arumberiamorph structures’ demonstrate a further mechanism by which equifinal Arumberia-like morphologies can form, they are unlikely to be direct analogues for the ancient Arumberia described here, given the global abundance of such forms in Ediacaran–Cambrian strata and the total absence of verified instances in the post-Cambrian rock record. The absence of Arumberia in any Phanerozoic strata, despite directed searching of equivalent facies, is most prosaically explained as a faithful historic testimony of its disappearance. This is because a Cambrian to Recent gap in the record of any sedimentary or fossil feature is significant: the notion that it is explainable by a > 500 Ma duration failure of the sedimentary-stratigraphic record to archive a relatively mundane sedimentary environment is unsubstantiated.

6.c. Orientation of Arumberia lines

One consensus viewpoint regarding Arumberia is that there is a dominant directionality to its orientation on individual surfaces, which has been considered to result from either physical (e.g. Brasier, 1979; Jenkins et al. 1981; McIlroy & Walter, 1997; Seilacher, 2007) or biological (Glaessner & Walter, 1975; Bland, 1984) mechanisms of alignment. Of the 40 reports in Table 1, 17 consider Arumberia line orientations with respect to the local current (as measured by ripple mark strike-lines on bedding planes hosting Arumberia, or indirectly inferred from palaeoflow indicators measured throughout the host succession). These case studies reveal that Arumberia may in fact strike both parallel and perpendicular to the original flow direction (Fig. 9a–f; Table 1). Within the Port Lazo Formation studied here, a strong perpendicular arrangement is apparent (Section 3.b; Fig. 6).

Kolesnikov et al. (2017) suggested that the organization of Arumberia lines may not necessarily be current-induced, but instead archive a trophic response of a microbial community in competition for sunlight. However, if this were the case, line orientation should be expected to be random relative to flow, rather than preferentially oriented with or normal to the prevailing current direction. We note that in several cases, Arumberia is aligned with flow in those environments where there was a unidirectional component (as shown by preserved linguoid ripple marks with superimposed, parallel Arumberia lines; Fig. 9b, c), but lies perpendicular to flow in settings with dominantly oscillatory flow conditions (as shown by Arumberia lines striking alongside preserved symmetrical wave-ripple marks; Fig. 9a, f). This would imply that Arumberia orientation could be in flux when ‘active’, and that the orientation of preserved Arumberia lines was dependent on external influences at the chance instant of burial. Lines aligned with ripple crests under gently oscillating flow, but were streamlined in the direction of a steady flow (e.g. see Davies et al. 2017). Reorganization in this way suggests that the Arumberia lines may record flexible cords that were pliant with hydrodynamic forcing. Such organization has analogue with extant marine algae, the flexible components of which reorganize and re-orientate under increasing unidirectionality and discharge of water flow (e.g. Denny & Gaylord, 2002).

Perpendicular arrangements with the prevailing current flow, observed in the Port Lazo Formation (Fig. 6) and elsewhere (Table 1), require that there was a secondary, weaker current that was strong enough to re-orientate Arumberia but insufficient to reorganize the sediment grains on the colonized surface. One potential scenario arises when a subordinate current drains water away following an interval of elevated water level (e.g. in the receding waters of a tidal prism). Such conditions were likely frequently met in the deposition of Arumberia-bearing strata, which are composed of periodically emergent facies in the majority of unequivocal reports of the form (Table 1). In the Port Lazo Formation, evidence of water drainage and exposure is indicated by the occurrence of raindrop impressions (Fig. 2g) and desiccation cracks (Fig. 2f) alongside Arumberia, and possibly also through the reticulated Arumberia forms (Fig. 3i, j).

6.d. Arumberia is most likely a fossil organism

The arguments above strongly suggest that Arumberia records an extinct, environmentally restricted and sessile biological entity that was composed of non-biomineralized, 3D, flexible cords (see also Bland, 1984, p. 630). The origins of some of its associated features (e.g. dimple-pimple marks) remain enigmatic, and its precise biological affinity remains uncertain, even though it is parsimonious to assume that it was algal or microbial. While this conclusion is similar to that reached by Kolesnikov et al. (2012), who suggested that Arumberia records extinct, highly organized microbial colonies capable of producing 3D ‘rugae’, we caution against splitting the texture into biological ‘species’ given the propensity of these to grade into one another (Section 4). Differences between these forms (e.g. A. banksi, A. vindhyanensis and also non-Arumberia linear morphologies) may stem from external hydrodynamic or environmental factors, reflecting the form that happened to be adopted at the instant of burial. Clear transitions between linear (Fig. 3a–h), reticulated (Fig. 3i, j) and curled (Fig. 5) Arumberia rule out a recent ‘vendobiont’ fossil hypothesis (Retallack & Broz, 2020), since discrete macroorganisms would not have been expected to reorganize their form during emergence.

Arumberia consists of parallel, sub-parallel or radiating lines that present in both positive and negative epirelief. A detailed examination of the most contiguous Arumberia-bearing bedding plane so far discovered, from the Port Lazo Formation (NW France), combined with a literature survey of all known Arumberia reports, indicates:

  • the 3D morphology of Arumberia resemble cords or tubes distinct from the preserving sediment;

  • linear Arumberia cords transition laterally into reticulated networks, morphologies that bear strong resemblance to modern instances of reticulate marks formed as buoyant filaments of bacteria or algae become tangled in standing or draining water;

  • some Arumberia cords are demonstrably composed of finer, superimposed and overlapping threads;

  • Arumberia cords curl in proximity to desiccated plate margins, an apparent response to shrinking cohesive sediment undergoing emergence;

  • the orientation of Arumberia lines differs in different depositional settings, with lines apparently being streamlined in the direction of a unidirectional current, but oriented parallel to ripple crests when subject to oscillating flow; and

  • the significant majority of Arumberia occurrences worldwide are recorded from rocks dated to 520–560 Ma in age.

Together, these observations discredit the possibility that Arumberia is a purely physical sedimentary structure, and strongly favour a biological origin as the impression of an extinct, sessile Ediacaran–Cambrian organism or population of organisms composed of 3D, flexible cords.

Fieldwork in France was supported by Shell International Exploration and Production B.V under Research Framework agreement PT38181. Fieldwork in Canada was partially supported by Natural Environment Research Council grant NE/P002412/1, awarded to Dr Emily Mitchell, University of Cambridge. We thank Professor Sören Jensen and one anonymous reviewer for their constructive reviews of this paper, and editor Dr Bas Van de Schootbrugge for handling this paper.

This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.