Fossils of the Ediacaran macrobiota ( from about 571 to 539 million years ago) record phylogenetically diverse marine palaeocommunities, including early Animals, which pre-date the Cambrian Explosion. Benthic forms with a frondose gross morphology, assigned to the morphogroups Rangeomorpha and Frondomorpha, are among the most temporally wide-ranging and environmentally tolerant members of the Ediacaran macrobiota and dominated deep-marine ecosystems between about 571 and 560 million years ago. Investigations into the morphology, palaeoecology, reproductive strategies, feeding methods, and morphogenesis of frondose taxa together constrain their phylogenetic position to the Metazoan (for Rangeomorpha) or Eumetazoan (e.g., Arborea) total groups, but tighter constraint is currently lacking..
In a paper published in the journal Current Biology on 6 April 2020, Alexander Liu of the Department of Earth Sciences at the University of Cambridge, and Frances Dunn of the School of Earth Sciences at the University of Bristol, the British Geological Survey, and the Oxford University Museum of Natural History, describe fossils of abundant filamentous organic structures preserved among frond-dominated fossil assemblages in Newfoundland (Canada). The filaments constitute a prominent component of the ecosystems, and exhibit clear physical associations with at least seven frondose taxa. Individual specimens of one uniterminal Rangeomorph taxon appear to be directly connected by filaments across distances of centimeters to meters. Such physical linkages are interpreted to reflect evidence for stolonic connections: a conclusion with potential implications for the phylogenetic placement and palaeoecology of frondose organisms. Consideration of extant stoloniferous organisms suggests that Ediacaran frondose taxa were likely clonal and resurrects the possibility that they may have been colonial.
Fossilized macroscopic filamentous structures are here reported from 38 unique bedding plane horizons (out of 183 studied fossil-bearing horizons) on the Avalon and Bonavista peninsulas of Newfoundland. Filamentous structures manifest as low (less than 1 mm) positive epirelief impressions, with no visible cell walls, membranes, external ornamentation, or disarticulation. Filaments are typically 100–1000 μm in width and 2–40 cm in length, although the longest and thickest examples Liu and Dunn have observed (on the LC6 surface of the Catalina Member of the Trepassey Formation) measure over 4 m in total length. Filament densities vary between different bedding planes, ranging from occasional individual strands to hundreds per square meter (extrapolated estimates suggest over 580 filaments per m² from sections of the MUN Surface of the Port Union Member of the Trepassey Formation), but densities are largely uniform within individual bedding plane assemblages. Individual specimens possess broadly constant widths and traverse bedding planes in multiple directions, Where filaments meet, they are typically superimposed rather than cross-cutting, strongly suggesting that they are not trace fossils. Rarely, observed filamentous structures dichotomously bifurcate, while some examples are seemingly arranged into bundles from which individual filaments can radiate. Small bulges along the length of some filaments are also observed, often at triple junction branching points.
Filamentous Macrofossils from the Bonavista Peninsula, Newfoundland. Filaments are preserved as positive epirelief impressions beneath fine-grained tuffs. (A) Negative epirelief holdfast, with positive epirelief filaments running across (inferred to be beneath) and around it, PU13 Surface. (B) Dense superimposed filament assemblages, showing superposition and directional changes, PU13 Surface. (C) Abundant filaments from the MUN Surface. Note filament superposition (arrowed at left) and bundling (arrowed at right). Inset: orientations of all filaments present on cast CAMSM X 50340.1 CST1, from the MUN Surface. Orange arrows denote the range of orientations of frondose taxa (indicating perceived current direction). Blue bars indicate primary fracture directions. (D) Further filaments, including one specimen that overlies another (arrowed, PU13 Surface). (E) Filamentous structure (white arrows) seemingly wrapped around a concentric holdfast disc (black arrow). Scale bars in (A)–(D) 10 mm and in (E) 10 cm. Liu & Dunn (2020).
Filaments follow relatively straight paths, but slight to significant curvature in most specimens (even doubling back on themselves in places), and bending around the holdfast structures of frondose macrofossils indicates that they were originally flexible structures. Across studied filament populations, filaments show no consistent preferential alignment with fracture or cleavage planes, or frond orientations. Thin sections reveal no three-dimensional sub-surface expression or preserved organic material and confirm that filaments are not associated with sub-surface fracture planes. Together, these observations imply that the filamentous structures were benthic, and Liu and Dunn interpret observed specimens to have lain above/on seafloor-covering microbial mats at the point of burial. However, the gradual fading of many specimens into bedding surfaces suggests that filaments may also have lain partially beneath the sediment, or within microbial mats, outside the plane of preservation. Liu and Dunn cannot refute the possibility that smaller filaments may reflect torn, fragmented, or partially degraded specimens.
Rangeomorph Fronds and Associated Filaments on the LC6 Surface, Little Catalina, Newfoundland (A) Two large Rangeomorph fronds, seemingly connected by a filamentous structure (white arrows) that transits between their holdfasts (black arrows) and then continues across the surface, tracing an inverted ‘z’ shape on the surface (photograph from cast CAMSM X 50341.4 CST1). Inset: orientations of all filaments present on this cast. Orange arrows denote the range of orientations of frondose taxa (indicating perceived current direction). Blue bars indicate primary (thick) and secondary (thin) cleavage directions. (B)–(F) Further examples of multiple filaments (arrowed) converging on the holdfasts (circled) of Rangeomorph specimens. The specimens in (D),CAMSM X 50341.2 CST1) and (E) CAMSM X 50341.3 CST1 lie along the same filament, which continues beyond both of them. All examples are of the same, as yet unnamed, Rangeomorph taxon. Scale bars in (A) 10 cm, and in (B)–(F) 10 mm. Liu & Dunn (2020).
Ediacaran frondose taxa are typically constructed of one or multiple fronds and often possess a basal holdfast structure interpreted to have anchored them to the seafloor, as well as a stem to elevate the frond into the water column. Filaments occur alongside all frondose Ediacaran macrofossil taxa described from Newfoundland to date and could both overlie, and lie beneath, the fronds and stems of such organisms. Of the 38 surfaces on which we have documented filamentous impressions, they occur alongside frondose taxa on 27 surfaces, alongside only discoidal specimens on nine surfaces, and as the only fossil impressions on two surfaces. Several specimens of an undescribed uniterminal rangeomorph taxon on the LC6 surface exhibit filaments terminating at/converging upon the outer margin of their holdfast discs. In one specimen, a large, unbroken filament traverses the bedding plane for 4.1 m and terminates at the holdfast of a frond. It then doubles back for 46 cm and terminates at the holdfast of another similarly sized specimen of the same taxon, before continuing on a curving trajectory for 90 cm to terminate at a small circular bulge, from which two additional filamentous impressions radiate. These specific filaments can exhibit branching along their length, and in places comprise multiple discrete strands. A second pair of fronds of the same taxon lie along another single filament of over 2.23 m in length, and at least three other specimens of the same taxon on that surface possess holdfasts that exhibit direct contact with filamentous structures, many of which clearly change their course to converge on the holdfasts.
Close-up Images of Seemingly Connected Rangeomorphs on the LC6 Surface (cast CAMSM X 50341.4 CST1). (A)–(B) Rangeomorph frond with multiple filamentous structures converging on its holdfast disc (B). Close-up of the holdfast region in (A). (C)–(D) Large Rangeomorph showing the spatial relationship between its holdfast and prominent filaments, which terminate at the holdfast margin. (D). Close-up of the holdfast region in (C). (E) Zoomed-out view of the frond in (C) showing how the filament leading to the second frond branches about 20 cm before reaching that specimen, with the branching filament (arrowed) possessing a trajectory that directly intersects the holdfast of the frond in (C). (F) Close-up of the bulbous branching junction (arrowed) between the filaments in (E). n.b. additional thinner filaments traverse the surface in multiple directions nearby. Scale bar gradations are in centimeters and millimeters. Liu & Dunn (2020).
Seven specimens of small frondose organisms termed ‘Ostrich Feathers’ on the LC6 surface are observed to possess filamentous structures of variable length that radiate from their holdfast margins. This variation in length in individual specimens is distinct from the radial ‘rays’ possessed by contemporary Hiemalora discs, which are typically of equal length in individual specimens.
Filamentous Macrofossils (arrowed) Terminating at Ediacaran Frondose Taxa. (A) Fractofusus andersoni from the Brasier Surface, Mistaken Point Ecological Reserve, with a filament seemingly extending from one end of the specimen midline. (B) Filaments on the MUN Surface, including one specimen that bisects the holdfast disc (circled) of a small Charniodiscus specimen. (C) Primocandelabrum sp. (MUN Surface), with associated filaments (arrowed) that appear to terminate at its holdfast. (D) Charnia masoni (cast CAMSM X 50341.5 CST1) from the LC6 surface, associated with two prominent curving filaments (arrowed) that converge on its holdfast. (E) ‘Ostrich feather’ specimen from the LC6 surface. Note the ray-like projections of variable length emanating from the holdfast disc (black arrows), with one filament (white arrow) extending from the holdfast over a greater distance of several centimeters. Scale bars 10 mm. Liu & Dunn (2020).
Several other frondose taxa exhibit one or multiple filaments terminating at or bisecting their holdfast margin (e.g., the Frondomorph/Arboreomorph Charniodiscus, and the Rangeomorphs Charnia and Primocandelabrum) Liu and Dunn also observed rare examples of single filaments terminating at one end of small Fractofusus andersoni specimens on the Brasier and H14 surfaces aligning with the trajectory of the organism’s midline and not emerging on the other side of the specimen.
Comparable filamentous structures to those seen in Newfoundland are recognized from the Memorial Crags and ‘Bed B’ surfaces of Charnwood Forest (UK), occurring in relatively low densities directly adjacent to frondose macrofossils. Negative hyporelief linear structures in the frond-bearing Ediacara Member of South Australia, and the Lyamtsa and Verkhovka formations of the White Sea region, Russia, share morphological (e.g., their size and shape) and taphonomic (negative hyporelief/ positive epirelief surface impressions of low topography) similarities with the Newfoundland structures but require further investigation to confirm a common origin.
Looped filamentous macrofossil from the Avalonian Ediacaran deposits in the UK. Arrow indicates a point where the filament crosses over itself. Sale bar is 10 mm. Helen Boynton in Liu & McIlroy (2015), in McIlroy (ed) (2015).
The thousands of filamentous fossils in Newfoundland do not exhibit cellular preservation, annulations, striations, or ornamentation, and maintain constant width along their length. Specimens could reach large size, appear to have been flexible, could dichotomously branch, are inferred to have been benthic, and could terminate at (or radiate from) holdfast structures or assumed growth axes of frondose taxa. There is no link between the filaments and cleavage or fracture planes either at the surface or in thin sections, ruling out a tectonic origin. The non-uniform orientations of filaments on bedding planes indicate that they have not undergone significant current alignment and were therefore unlikely to have been tethered to the substrate at just one point.
Previously described Ediacaran filamentous macrofossil impressions are not directly comparable to those described by Liu and Dunn. Filamentous structures from Spain and Namibia interpreted as Vendotaenids, as well as structures from the Drook Formation of Newfoundland, can be of comparable width but are typically just a few centimeters in length, are preserved in far lower densities, and possess more sinuous morphologies than these Newfoundland specimens. Possible Algal fossils described from shallow marine assemblages of the White Sea only reach a few millimeters in length and are found in small, dispersed clusters on the bedding surfaces. A single figured specimen from the Khatyspyt Formation of Siberia documents physical filamentous connections between macroscopic circular carbonaceous compression fossils within successions that contain frondose taxa but includes no further description.
Other modern and extinct organisms with a macroscopic filamentous appearance include several Neoproterozoic forms of a few centimeters in length, which have been compared with Macroalgae, Metazoans, or the sheathes of Sulphur Bacteria, and Cyanobacteria.The filaments Liu and Dunn describe are too large to be attributed to most extant Bacterial groups, including Giant Bacteria and those capable of undergoing filamentation. Algal fossils can show some similarities to this material, but the deep-marine depositional setting inferred for the Conception Group in Newfoundland would preclude benthic photosynthetic lifestyles. Algae could have been washed into these depositional settings, but the abundance and extensive lateral distributions of filaments on bedding planes, and their apparent connections to holdfasts of frondose taxa, are difficult to explain in that scenario. The taphonomic style and branching of the filaments bears passing resemblance to certain late Ediacaran biotic sedimentary surface textures (e.g., ‘Arumberia'), but such impressions usually show a preferential alignment and regular spacing on a given surface and overwhelmingly occur in shallower sedimentary facies that do not contain Ediacara-type macrofossils.
The filaments described by Liu and Dunn exhibit widths at least an order of magnitude larger than those of the largest modern Fungal hyphae. Meanwhile, clear superposition rather than truncation renders an ichnological explanation unlikely.Filamentous components of contemporary Ediacaran macrofossils such as the long filamentous ‘string’ of Hadrynichorde or the radial ‘rays’ of Hiemalora are distinctive structures, with consistent spatial associations relative to their respective body impressions. Hiemalora typically possesses roughly 10–80 individual rays, which radiate in all directions from an attachment point at the margin of the disc, and which usually all terminate at similar distances of a few centimeters. This is in contrast to the small number of filaments (less than 8) associated with individual holdfasts seen among Liu and Dunn's material, which can extend over distances of many centimeters. To the best of Liu and Dunn's knowledge, there are no described extant or fossilized discrete, filamentous organisms that exhibit all aforementioned characters.
Strong circumstantial evidence for termination of filaments at frond holdfasts suggests a physical association with Ediacaran frondose taxa. Such an association could be direct (i.e., the filaments are part of the macro-organisms) or indirect (with the filaments being independent organisms engaging with the fronds passively, symbiotically or parasitically, as seen for example in the interactions between extant Plants and Mycorrhizal Fungal networks). An indirect relationship for the filaments with the frondose taxa cannot be ruled out but is considered less likely by Liu & Dunn since all observed filament-mediated connections between frondose specimens on individual surfaces are intraspecific. On the over 200 m² bedding plane LC6, which exhibits thousands of thin filamentous impressions, the majority of the few thick (at least 1 mm width) filamentous structures converge on holdfasts of a single, unnamed, Rangeomorph taxon seemingly passing adjacent holdfasts of other taxa without exhibiting any obvious relationship with them, despite high frond densities.
There is no indication that the filamentous structures were rigid (given their propensity to bend/change direction in many examples), implying that they were not biomineralized. Among extant marine taxa, non-mineralized filamentous outgrowths of comparable gross morphology occur in Algae (where they link individual fronds), certain Metazoans (where they link polyps/individuals), and Fungal mycelia. The outgrowths typically fulfill stabiliation, defense, nutrient transport, or (asexual) reproductive roles involving budding or stoloniferous growth, for example, in extant Algae (e.g., the Green Alga Caulerpa), terrestrial Plants, and Metazoans including Sponges, colonial Cnidarians, Entoprocts, and Bryozoans. These different functions of filamentous outgrowths are not mutually exclusive, and all remain potential candidates for the function of the Ediacaran filaments Liu and Dunn describe, given available evidence and sedimentary context.
Independent assessment of the spatial distribution of the Rangeomorph taxon Fractofusus on Ediacaran bedding planes predicted a stolon-like asexual reproductive strategy in the life cycle of that organism. Fractofusus specimens actually connected to each other by filaments have yet to be observed, but filaments are observed in abundance on several surfaces containing Fractofusus (e.g., bed H14), where they rarely terminate at the ends of small Fractofusus specimens. Fractofusus specimens possessing such filaments are never the very smallest but typically measure 1.5–3 cm in length. Further support for a stolon interpretation is provided by the presence of bulbous thickenings at filament branchpoints, which are morphologically comparable to the branch nodes seen in some stoloniferous Metazoans. If the filamentous structures do reflect stolon-like projections with a solely reproductive role, large specimens might be expected to connect to smaller ones. However, in examples of connected uniterminal Rangeomorph specimens on bed LC6, both specimens in any given pair are of a similar large size and are thus interpreted as ‘mature’ individuals of a similar developmental stage. This may indicate that, even if reproduction was the primary reason for stolon formation, the connections between specimens may have remained active for a considerable period following establishment of the individuals on the substrate, perhaps to facilitate nutrient transfer between individuals to counter the inferred nutrient-poor deep-water settings of the Conception Group. A stoloniferous habit is also consistent with observations that, in cases where fronds are seemingly connected, the filament often continues beyond the frond after meeting it, and that multiple filaments may converge upon a single holdfast.
The filamentous structures may ultimately provide novel morphological characters with which to assess Ediacaran fronds, but the prevalence of stolon-like structures among extant Eukaryotes means that, in isolation, stolonic growth cannot constrain the phylogenetic position of Ediacaran frondose taxa. However, multiple modern stoloniferous Eukaryotes, independent of phylogeny, are modular, clonal, and in some cases, colonial organisms. Ediacaran frondose taxa have previously been proposed to be clonal or colonial, albeit by viewing individual specimens as colonies on the basis of their highly compartmentalized morphology. In recent years, such interpretations have lost support as comparisons between frondose taxa and extant colonial Cnidarians have been questioned. A clonal facet to frond biology would raise the prospect that individual fronds were ‘unitary’ entities (ramets) within a larger benthic, interconnected clonal colony. This intriguing possibility could explain several aspects of frond palaeoecology (e.g., the dominance of particular taxa on individual surfaces) and has implications for our views of senescence, reproduction, and damage response within these early metazoan communities. Clonal reproduction in Ediacaran fronds could also have allowed for rapid colonisation of the seafloor or re-establishment of communities following sediment influx events. The observed filaments may therefore have favored rapid community succession by frondose taxa over non-frondose competitors in environments prone to episodic sedimentation, potentially in addition to engineering increased ecosystem habitability for those taxa by binding/stabilizing soft substrates.
Recognition of direct associations between organic filamentous structures and benthic frondose organisms offers new insight into late Ediacaran palaeocommunities. The profusion of filaments on Newfoundland bedding planes indicates that they were an important, and perhaps even integral, ecological component of frondose Ediacaran taxa and ecosystems. A stoloniferous interpretation of apparent filamentous connections between frondose taxa implies clonal reproduction in these organisms and may offer support to the view that these early macroscopic Metazoans were non-unitary.
Follow Sciency Thoughts on Facebook.