The rise and early evolution of animals: where do we stand from a trace-fossil perspective?

Abstract

The trace-fossil record provides a wealth of information to track the rise and early evolution of animals. It comprises the activity of both hard- and soft-bodied organisms, is continuous through the Ediacaran (635–539 Ma)– Cambrian (539–485 Ma) transition, yields insights into animal behaviour and their role as ecosystem engineers, and allows for a more refined characterization of palaeoenvironmental context. In order to unravel macroevolutionary signals from the trace-fossil record, a variety of approaches is available, including not only estimation of degree of bioturbation, but also analysis of ichnodiversity and ichnodisparity trajectories, and evaluation of the occupation of infaunal ecospace and styles of ecosystem engineering. Analysis of the trace-fossil record demonstrates the presence of motile benthic bilaterians in the Ediacaran, mostly feeding from biofilms. Although Ediacaran trace fossils are simple and emplaced at or immediately below the sediment surface, an increase in ichnofossil complexity, predation pressure, sediment disturbance and penetration depth is apparent during the terminal Ediacaran. Regardless of this increase, a dramatic rise in trace fossil diversity and disparity took place during the earliest Cambrian, underscoring that the novelty of the Fortunian (539–529 Ma) cannot be underestimated. The Fortunian still shows the persistence of an Ediacaran-style matground ecology, but is fundamentally characterized by the appearance of new trace-fossil architectural plans reflecting novel ways of interacting with the substrate. The appearance of Phanerozoic-style benthic ecosystems attests to an increased length and connectivity of the food web and improved efficiency in organic carbon transfer and nutrient recycling. A profound reorganization of the infaunal ecospace is recorded in both high-energy sand-dominated nearshore areas and low-energy mud-dominated offshore environments, during the early Cambrian, starting approximately during Cambrian Age 2 (529–521 Ma), but continuing during the rest of the early Cambrian. A model comprising four evolutionary phases is proposed to synthetize information from the Ediacaran–Cambrian trace-fossil record. The use of a rich ichnological toolbox; critical, systematic and comprehensive evaluation of the Ediacaran–Cambrian trace-fossil record; and high-resolution integration of the ichnological dataset and sedimentological information show that the advent of biogenic mixing was an important factor in fully marine environments at the dawn of the Phanerozoic.

1. Introduction

The trace-fossil record, essentially consisting of all sorts of bioturbation, bioerosion and biodeposition structures, represents a valuable archive to decipher the history of life, and its importance to reconstruct macroevolutionary events has been known for over 60 years [1]. This has not been, however, the dominant trend in trace-fossil research, which has mostly been focused on the successful application of ichnological analysis to address sedimentological questions, in cases with the ultimate goal of solving problems in the petroleum industry. This mainstream line of research has proved to be extremely rewarding in terms of gaining profound understanding on how organisms react to environmental parameters, how trace-fossil distribution can reflect relative changes in sea level and how bioturbation modifies reservoir properties [2–4]. Although coexisting with other trends, a renewed interest in viewing the trace-fossil record as a valuable archive for understanding macroevolution has characterized the last two decades (e.g. [5–14]). An up-to-date perspective of this line of research has been recently summarized in a more comprehensive fashion [15,16]. Ichnology, the science of organism–substrate interactions, is at the interface of two sets of disciplines, one set encompassing biology and the other one in connection with sedimentary geology. This double nature of ichnology defines its unique character, potentials and limitations. The ichnological toolbox is quite rich [17] (table 1), but the available methods and conceptual framework have not been extensively applied from a macroevolutionary perspective. These tools should be used in an integrated fashion in order to offer a more accurate and holistic reading of the trace-fossil record. Central questions that this approach may help to elucidate are the present controversies on the timing of emergence and role of bioturbation during this critical span in Earth history, including insights into whether secular changes between Ediacaran and Cambrian biotas are best explained by successive, transitional radiation events or involve larger-scale evolutionary events [5,12,13]. Analysis of organism–substrate interactions offers a way to contrast Ediacaran and early Phanerozoic communities, and also to trace the continuity of animal activity through the Ediacaran–Cambrian transition. The aims of this paper are to: (i) review the potential and limitations of trace fossils to unravel the rise and early evolution of animals, using up-to-date concepts and methodological tools of ichnological analysis, (ii) highlight animal–substrate interactions in Ediacaran ecosystems, tracking trends and changes between the different Ediacaran assemblages, (iii) contrast Ediacaran and Fortunian ichnodiversity and ichnodisparity levels in order to outline their distinct natures, and (iv) unravel the palaeoecological and geobiological significance of bioturbation during the rest of the early Cambrian.

Table 1.

Summary of the ichnologic toolbox in palaeobiology and evolutionary palaeoecology.

concept	tool	characteristics and approach	advantages, potential and applications	limitations and pitfalls	key references
measuring degree of bioturbation	ichnofabric index (ii)	estimation of degree of bioturbation in cross-section through the use of ichnofabric flashcards (i.e. standardized visual representations of the biogenic reworking of sedimentary fabrics)	proxy to make inferences on degree of infaunalization allows tracking changes in intensity of bioturbation along the depositional profile through time (i.e. time–environment matrix) meaningful results in the case of simple ichnofabrics (i.e. those that result from the activity of a single infaunal community at a given moment)	all metrics proposed based on a measurement of the percentage area bioturbated, but bioturbation should be measured as a unit volume per unit time use of flashcards quite limited for composite ichnofabrics (i.e. those produced by the upward migration of a tiered community or by the replacement of successive communities through time) estimations may be problematic if colonization surface is not located commonly used without considering other tools of the ichnofabric approach (e.g. tiering analysis) ambiguous characterization of and differentiation between ii 5 and ii 6	[18]
	bioturbation index (BI)	descriptive grades assigned to the degree of bioturbation as seen in cross-section	proxy to make inferences on degree of infaunalization allows tracking changes in intensity of bioturbation along the depositional profile through time (i.e. time–environment matrix) meaningful in the case of simple ichnofabrics more accurate characterization of composite ichnofabrics because the different indexes consider not only burrow density, but the amount of burrow overlap and the sharpness of the original fabric commonly used in conjunction with analysis of tiering structure, evaluation of taphonomic filters, ichnoguild characterization, and ichnofabric constituent diagrams	all metrics proposed based on a measurement of the percentage area bioturbated, but bioturbation should be measured as a unit volume per unit time estimations may be problematic if colonization surface is not located	[19–21]
	bedding- plane bioturbation index (BP-BI)	visual estimation of density of trace fossils on these horizontal surfaces with the aid of flashcards	particularly suitable for the analysis of shallow-tier to superficial trace fossils preserved along lithological interfaces (e.g. most Ediacaran–Fortunian) important to assess synecological aspects of benthic faunas useful synergies if combined with methods developed to evaluate the degree of patchiness in horizontal distribution of trace fossils	in strict sense not a real metric of degree of bioturbation, which typically implies disruption of the sedimentary fabric	[22–24]
assessing trends in ichnodiversity and ichnodisparity	ichnodiversity	number of ichnotaxa present (commonly at ichnogeneric rank)	consideration of ichnotaxonomic composition construction of ichnodiversity curves useful for reconstructing diversity trajectories allows to unravel the macroevolutionary dynamics of evolutionary radiations allows to evaluate colonization trends in distinct environments distinction between α, β, γ and global ichnodiversity useful to understand diversification at different scales	more a measurement of richness rather than of diversity (but recent attempts to quantify abundance allow evaluating both the number of ichnotaxa and the degree of dominance) ichnodiversity is a different metric from biodiversity, because the former reflects the diverse ways in which animals interact with the substrate, not the diversity of organisms per se affected by taphonomic filters requires constant update due to continuous introduction of new ichnotaxa and revision of previously established ones	[25–27]
	ichnodisparity	variability of general morphologic plans in trace fossils assessed in terms of categories of architectural designs	captures higher-rank changes in animal–substrate interactions that are more substantial than those captured by ichnodiversity yields insights into large-scale innovations in body plan, locomotory system and/or behavioural strategies flexible matrix able to accommodate basic morphologic patterns that reflect our present understanding of trace-fossil architecture and its fabrication, rather than a rigid framework allows evaluation of contrasting evolutionary scenarios (e.g. coupling or decoupling between ichnodiversity and ichnodisparity) free of some of the vagaries involved in ichnotaxonomy	categories of architectural designs are heterogeneous in nature requires constant update due to continuous introduction of new ichnotaxa and revision of previously established ones	[25,28]
evaluation of ecospace utilization and ecosystem engineering	ichnoguild	evaluation of ecospace utilization by assessment of three parameters, bauplan, food source and use of space	allows assessing how organisms tend to group together within the same tier to exploit the same resources in similar ways helps understanding patterns of ecospace utilization in different environments through geologic time valuable tool to understand the adaptive strategies displayed by benthic organisms allows characterizing the ecological complexity of ichnofaunas	difficult to apply in some circumstances due to various factors, such as temporal instability of community structure, time averaging and limited cross-cutting relationships	[29,30]
	ecospace and ecosystem engineering cubes	characterization of tiering, motility, feeding mode, ways in which organisms interact with and modify the sediment, and mode of penetration in the substrate	allows integration with conceptual frameworks in marine benthic ecology adjustment of categories from marine benthic ecology to better suit the specificities of trace fossils allows detailed evaluation of changes in ecospace utilization and ecosystem engineering in specific ecosystems through geologic time suitable for applying to the analysis of large datasets and databases	attribution of categories of interaction with the sediment and of modification of the sediment is not always straightforward	[14,17]
	ecosystem engineering impact index (EEI)	sum of the contribution of tier, functional group of the bioturbator, and the likelihood of bioirrigation	allows integration with conceptual frameworks in marine benthic ecology simple way of characterizing ecosystem engineering	attribution of categories of functional group is not always straightforward categories taken from marine benthic ecology without adjustment overlap of functional group (e.g. epifaunal bioturbators, surficial modifiers) with tiering position, making the final estimation of ecosystem engineering impact potentially redundant lack of common language with schemes of widespread use in palaeobiology, making comparisons with the body-fossil record more difficult	[31]

2. The Ediacaran–Cambrian trace-fossil record: potential and caveats

2.1. Evidence of activity of soft-bodied and poorly skeletonized animals

In contrast with the body-fossil record, which is overwhelmingly biased towards organisms with hard parts, the trace-fossil record comprises the activity of both hard- and soft-bodied organisms, as well as poorly skeletonized animals. Accordingly, the trace-fossil record provides valuable information on organisms that tend to be underrepresented in the body-fossil record. Soft-bodied and small organisms typically make the vast majority of the biomass and their assessment has been referred to as ‘tackling the 99%’ [32]. Most importantly, even if metazoan skeletonized clades could have evolved previous to the acquisition of a hard body [33], ichnofossils could track this early history (granting ancestral metazoans were benthic [1,5]). In fact, the rich Fortunian ichnological record testifies that non-skeletonized and poorly skeletonized stem-group members of modern phyla were active players in fully marine settings [12]. For example, the regularly sinusoidal trace fossil Cochlichnus, typically attributed to nematodes in younger Paleozoic rocks, is apparently absent in Ediacaran strata [34], but quite abundant in both shallow- and deep-marine Fortunian deposits [34–39].

In the case of the Ediacaran–Cambrian transition, a thorough scrutiny of the trace-fossil record allows challenging some hypotheses on the origin of bilaterians and their ecologic role in Ediacaran and early Phanerozoic communities. This makes possible tracking the appearance of benthic behavioural innovations and how they may be calibrated with metazoan diversification or fuelled by innovations taking place in the planktonic system [5,40–44]. The fact that a dramatic increase in trace-fossil diversity occurred in the early Cambrian [1,12,45–47] argues against the view of the Cambrian explosion as an artefact resulting from the acquisition of mineralized skeletons in the early Cambrian, but also argues against major phylum-level cladogenesis taking place during the Ediacaran. A recent compilation of changes in ichnodiversity through geologic time shows that the rise in ichnodiversity that took place in the earliest Cambrian (433% of ichnogeneric increase in comparison with the Ediacaran) is unparalleled in the rest of the Phanerozoic, emphasizing the uniqueness of this event [48].

2.2. The continuity of the trace-fossil record across the Ediacaran–Cambrian boundary

There is an almost 20 Ma gap between the youngest exceptional preservation of Ediacaran body-fossil assemblages, the Nama assemblage (549–539 Ma), and those of Cambrian Age 3 (521–514 Ma), represented by Burgess Shale-type fossil sites, such as Chengjiang, Sirius Passet and Qingjiang (521–514 Ma) [49–52]. This gap encompasses a critical time to establish a chronology of early metazoan evolution, which is represented in the stratigraphic record essentially by small shelly fossils [53,54], trace fossils [12] and, towards Cambrian Age 2, by small carbonaceous fossils [55]. The trace-fossil record is continuous through this transition, with 72 ichnofossiliferous stratigraphic successions documented worldwide, allowing evaluation of the timing of diversification in strategies of animal–substrate interactions and patterns of ecosystem construction [12]. This record through the Ediacaran–Cambrian transition indicates that only a few ichnotaxa are restricted to the Ediacaran (i.e. those produced by iconic elements of the Ediacara biota, such as Epibaion), whereas a few poorly specialized forms cross the boundary, and a large number of distinctive trace fossils appeared for the first time in the early Cambrian [12,48]. This pattern provides key information on both the novelties that signalled the early Cambrian and the Ediacaran roots of the diversification.

2.3. The link between trace fossils, behaviour and ecosystem engineering

Trace fossils are evidence of behaviour [56]. Accordingly, autecological analysis of ichnotaxa forms the basis to read the trace-fossil record from an ethological and functional perspective. In addition, synecological analysis of trace-fossil assemblages yields insights into ecological aspects of benthic communities; this scale of analysis is critical to track changes at ecosystem level. Also, there has been a shift in paradigm in ichnology from the classic model of benthos controlled by the physical parameters of the environment to a more interactive model where animal behaviours (e.g. burrowing, feeding, excretion) play a significant role as architects of their own environments [57]. Within this framework, animal activity represented by trace fossils can reveal the work of ecosystem engineers that actively modified their hosting substrates, altering physico-chemical conditions and creating habitats through niche construction [58]. In fact, the process of bioturbation itself can be viewed as an archetypal example of ecosystem engineering [59]. Although the complexity of geobiological feedback loops is not fully understood, there is an increased recognition that bioturbation played a major role as a force of the evolutionary changes that took place during the Ediacaran–Cambrian transition [12,40,42,58–64].

2.4. The link between trace fossils, substrates and environments

Trace fossils are produced through the interaction of organisms and the substrate. Evaluation of substrate, central in our understanding of bioturbation and bioerosion, provides added value to trace fossils over body fossils, but substrate can also pose significant challenges, particularly in the evaluation of taphonomic overprint and an adequate interpretation of morphology. The final morphology of a trace fossil is the result of variable combinations of three factors: the anatomy of the animal, its behaviour (regulated by its biomechanical capabilities) and conditions of the substrate [65]. Trace fossils record the activities of animals, which in turn are strongly influenced by environmental parameters. Accordingly, the link between trace fossils and their associated substrates and environments is typically much stronger than that recorded by body fossils. This is the underlying reason why ichnology has been particularly successful in sedimentary geology through its applications in facies analysis and sequence stratigraphy. Ichnological studies typically involve a more refined characterization of the associated sedimentary facies than studies based on body fossils (e.g. [66,67]). Palaeobiological studies using trace fossils should benefit from the potential of ichnology to provide high-resolution palaeoenvironmental and stratigraphic frameworks. For example, in the debates with respect to early metazoan evolution, it is essential to discriminate between environmental and evolutionary controls on specific ichnofaunas (e.g. [36,68]).

Extreme care should be exercised before attributing macroevolutionary significance to sparsely bioturbated intervals because local parameters (e.g. freshwater discharge, oxygenation) exert a primary control on the benthic community, affecting biogenic disruption of the sedimentary fabric. For example, a fully bioturbated offshore deposit may grade laterally or vertically into a sparsely or unbioturbated prodelta deposit due to the proximity of a deltaic distributary mouth, resulting in increased sediment and freshwater discharge (e.g. [67,69]). Accurate evaluation of the role of local parameters may require the analysis of multiple, coeval stratigraphic sections, rather than a single vertical column. Averaging degree of bioturbation to infer macroevolutionary significance of sediment mixing in this case would be misleading, because it is the fully bioturbated facies the one that provides key information about the ability of organisms to mix sediment at a given time in geologic history [62].

There has been a notable shift in our understanding of the role that oxygen may have played in the origin and evolution of early animals [70–76]. Recent theoretical and experimental research suggests that early metazoans may have lived and thrilled under strong dysoxic conditions, essentially challenging the idea of oxygen as the fundamental trigger in metazoan evolution [71–74]. In fact, it has been proposed that the oxygen equation could be reversed, animals being the cause of oxygen availability, decreasing water turbidity and having a positive effect on ocean ventilation [42,72,77]. In any case, oxygen is an important environmental variable that can influence diversity, size, activity and feeding type of metazoans [73,78–80]. Accordingly, discrimination of ecological (local) and evolutionary controls in trace-fossil assemblages is an extremely sensitive matter when analysing the Ediacaran–Cambrian transition, a time characterized by unstable environmental conditions in terms of oxygen, nutrients and other factors.

2.5. The uncertainties in attributing ichnotaxa to their producers

Undoubtedly, the main disadvantage of the trace-fossil record is the uncertainty to establish a solid link between specific ichnotaxa and their producers, strongly limiting the use of trace fossils in phylogenetic analysis. This general principle, however, finds remarkable exceptions, relying on specific taphonomic windows or on complex architectures and mode of construction of some trace fossils that offer a strong link to decipher the phylogenetic affinities of the producer (e.g. [81]). Of particular interest are those structures where distinct morphological features can be related to diagnostic or significant anatomical traits (e.g. insect tagmosis, long abdomen with leglets, head and palps [82]; minimum number of locomotory appendages [83]), providing crucial information to pinpoint the most likely candidate. The early metazoan trace-fossil record, however, is composed vastly of unbranched, horizontal, shallow trails and burrows [1,84–87], which do not display enough morphological complexity to provide a link to a particular group, and are commonly attributed to vermiform producers of unknown affinities.

In spite of these uncertainties, detailed studies of Ediacaran–Cambrian trace fossils have proved instrumental to assess evolutionary innovations. First, there are extraordinary cases of producers in close connection to their traces, such as Yorgia and Dickinsonia with Epibaion [87–92]. Second, evidence of locomotory appendage imprints in bilobate trails and resting structures and trackways are uncontroversial proof of production by euarthropods [5,38,43,93]. Third, even in the case of worm-like organisms, neoichnological studies and the use of modern analogues may help to narrow down the spectrum of producers and ecologies involved, at least strongly arguing in favour of particular tracemakers, such as priapulids as the producers of Treptichnus pedum [94,95] or nematodes as producers of Cochlichnus [34].

3. Ichnological evidence of early metazoan life and ecology: Avalon versus White Sea assemblages

The Precambrian trace-fossil record has been critically reviewed in many studies [6,7,12,34,87]. As a result of these reviews, the number of Ediacaran ichnotaxa has decreased dramatically in comparison with previous estimations [96,97]. Also, the age of the oldest animal trace fossils has been a contentious issue [87]. Bilaterian trace fossils have been described from postglacial strata in Uruguay that were regarded as at least 585 My old [98]. However, there is growing evidence that the trace fossil-bearing strata are Permian postglacial deposits that host an ichnofauna that is characteristic of the Parana and Karoo basins [87,99,100]. In the original report, only horizontal grazing trails (Helminthoidichnites-like) were documented [98]. However, ongoing collections from the same localities have provided arthropod ichnotaxa, including the distinctive Permian arthropod resting trace Gluckstadella (R. Netto, personal communication, 2020). Although behavioural convergence is a common motif in the ichnologic record, the fact that classic Permian arthropod ichnotaxa from the Parana and Karoo basins were found makes the Ediacaran age claim highly implausible.

Possible animal trace fossils have been documented from the 565 My-old Mistaken Point Formation, the best-known unit recording the Avalon assemblage [101]. However, alternative interpretations, including the dragging or uprooting of fronds through a microbial surface, have been suggested [87]. Based on these uncertainties, the Fermeuse Formation of Newfoundland (approx. 560 Ma) has become a key unit to detect ichnological evidence in the Avalon assemblage [102,103]. The close association and apparent intergradations of structures described as trace fossils from this unit [103] with the body fossil Palaeopascichnus suggest that these structures may represent a continuum of taphonomic variants of this body fossil [87]. Interestingly, other structures in the same outcrop, specifically crescentic horizontal structures connected with Aspidella-discs and vertical equilibrium structures [102], display morphologic features that can be explained in terms of animal–substrate interactions. In particular, these structures seem to record vertical movement (fig. 3A–E in [102]) and lateral displacement, and have been interpreted as formed by an organism probably of cnidarian grade (diploblastic) [102]. Based on morphologic evidence, these trace fossils are included in Bergaueria sucta [87] (table 2) and represent one of the earliest records of animal behaviour.

Table 2.

Ichnologic composition of Ediacaran strata. Stratigraphic ranges based on [12,48] and categories of architectural designs after [28]. Cyanobacterial microborings are not included.

assemblage	category of architectural designs	ichnogenera
Avalon	34-vertical plug-shaped burrows	Bergaueria
White Sea	1-simple horizontal trails	Archaeonassa
	34-vertical plug-shaped burrows	Bergaueria
	32-oval-shaped imprints	Epibaion
	1-simple horizontal trails	Gordia
	1-simple horizontal trails	Helminthoidichnites
	1-simple horizontal trails	Helminthopsis
	9-fan-shaped to radiating scratched imprints	Kimberichnus
	11-simple actively filled (massive) horizontal to oblique structures	Nenoxites
	10-passively filled horizontal burrows	Palaeophycus
	11-simple actively filled (massive) horizontal to oblique structures	Torrowangea
Nama	1-simple horizontal trails	Archaeonassa
	1-simple horizontal trails	Circulichnis
	34-vertical plug-shaped burrows	Conichnus
	34-vertical plug-shaped burrows	Bergaueria
	1-simple horizontal trails	Gordia
	1-simple horizontal trails	Helminthoidichnites
	1-simple horizontal trails	Helminthopsis
	63-circular holes and pit-shaped borings	Oichnus
	10-passively filled horizontal burrows	Palaeophycus
	14-complex actively filled horizontal structures	Parapsammichnites
	11-simple actively filled (massive) horizontal to oblique structures	Torrowangea
	17-horizontal burrows with horizontal to vertical branches	treptichnids

Bilaterian (triploblastic) trace fossils have been documented from the White Sea assemblage [6,7,87,104,105]. A total of 10 ichnogenera comprising six categories of architectural designs have been identified (table 2). By far, the most common trace fossils are simple horizontal trails, namely Helminthoidichnites, Helminthopsis (figure 1a), Archaeonassa and Gordia (figure 1b) [6,12,87,106]. Alternative interpretations have been proposed in many instances [107–110], and a case-by-case evaluation is invariably needed before affirming the bilaterian origin of any alleged trace fossil [111]. The nature of Nenoxites has been controversial and both body fossil and trace-fossil interpretations have been suggested [34]. Most structures attributed to Nenoxites are best regarded as body fossils [112–116], but a critical revision of the type specimen is awaiting in order to evaluate its true nature. Regardless, the presence of Torrowangea, an ichnogenus showing evidence of active infill, indicative of a gut, suggests the presence of an internal cavity (e.g. coelom) and possibly peristalsis to penetrate into the substrate [117]. Structures attributed to the vertical plug-shaped burrow Bergaueria and the horizontal passively filled burrow Palaeophycus, although most likely present, are hard to differentiate from body fossils in many instances [7,34,118,119]. The trace-fossil nature of simple plug-shaped structures can only be confirmed with certainty when detailed morphologic analysis allows inferring escape behaviour or adjustment to sedimentation rate [102]. Particularly noteworthy are Epibaion (figure 1c) and Kimberichnus (figure 1d), ichnotaxa portrayed as directly associated with their producers, dickinsonids and Kimberella, respectively [88–92,120–122]. The suggestion that dickinsonids were passively transported by currents [123] has been rejected on the basis that the successive Epibaion and the directly associated body fossil show a curved trajectory and a preferred direction of forward movement [92,124]. Kimberichnus has been interpreted as a trace fossil produced by scratching of the mat by Kimberella [120–122]. However, ongoing research is casting doubts on the tracemaker–trace fossil connection between Kimberella and Kimberichnus, the latter having been regarded as a microbial texture rather than a trace fossil [125].

Figure 1.

Representative trace fossils of the White Sea assemblage. (a) Helminthopsis tenuis, (b) Gordia marina, (c) Epibaion costatus, represented by four overlapping resting trace fossils, associated with its producer, Dickinsonia costata. Photograph by Jakob Vinther. (d) Kimberichnus teruzzi associated with its producer, Kimberella quadrata. Photo from a cast at the exhibition Fossil Art. All photos are from specimens preserved as positive hyporelief, Ediacara Member, Rawnsley Quartzite, Flinders Ranges, South Australia. Images from [87] and reproduced with permission of Springer.

Various styles of organism–microbial mat interactions can be recognized from the ichnoguild perspective [87] in localities hosting the White Sea assemblage. In contrast with Phanerozoic examples, the vast majority of Ediacaran ichnoguilds are monospecific, revealing a simple ecological structure of epifaunal and shallow infaunal communities. Overall, these are very shallow-tier to superficial trace fossils that produced almost no disturbance of the primary fabric, with bioturbation index (sensu [21]) being typically 0 [12,126]. The Helminthopsis ichnoguild consists of transitory, near-surface to very shallow-tier, mat-grazer structures produced by vagile animals. Although the precise phylogenetic position of these early tracemakers is uncertain, their mode of production, particularly in the case of shallow burrows, indicates a triploblastic animal [84]. These organisms feed from organic matter concentrated within epibenthic or endobenthic microbial mats [127]. This ichnoguild is the most abundant in the White Sea assemblage and includes the ichnogenera Archaeonassa, Gordia, Helminthoidichnites and Helminthopsis.

In short, based on all available evidence, there is little doubt that bilaterian trace fossils were part of the White Sea communities and their producers somehow ecologically related to other members of the community [6,13,84]. In this framework, it has been suggested that the Ediacara biota played an enabling role in bilaterian evolution, providing enriched organic carbon sites (the so-called Savannah hypothesis [13]). Recently, the presence of grazing trails cutting through the external moulds and casts of several Ediacara body fossils prompted the interpretation that scavenging may have been involved [128]. The association of non-biomineralized skeletons and trace fossils is well known from Burgess Shale-type deposits [129–133], and has been interpreted as a record of the increasing heterogeneity of food resources on the Cambrian seabed [132]. Interestingly, Helminthoidichnites (an Ediacaran survivor) displays high-density occurrences in Tuzoia and Arthroaspis carcasses. In spite of this association, morphological features in Helminthoidichnites, such as the absence of active infill and curved to straight course, are suggestive of grazing, linked to targeting of bacteria hotspots, rather than scavenging in strict sense [132,133]. In this scenario, distribution in close proximity to body fossils may simply reflect bacterial enrichment (not necessarily implying feeding on decaying soft tissue). Evidence of scavenging in early Palaeozoic carcasses is represented by abundant pellet-infilled and backfilled annulated burrow structures [132,134], which have not been recorded in trace fossils associated with Ediacara body fossils so far.

The Epibaion ichnoguild comprises transitory, surface mat-digester structures produced by vagile animals of debatable affinities [88–92]. Direct absorption of nutrients from the underlying mat across their lower surfaces is envisaged, although the precise feeding mechanism remains unresolved [88,135]. The Kimberichnus ichnoguild consists of transitory, surface mat-scratcher structures produced by vagile animals, probably of mollusc affinities [136,137] (but this is still controversial [13]). The classic interpretation is that these are rasping traces produced by a proboscis-like structure with paired teeth [34,88,89,120,121,138,139]. However, the link between Kimberella and Kimberichnus has been questioned recently [125], and a thorough re-evaluation of this ichnotaxon is due. Other ichnoguilds do not seem to show any relationship with microbial mats, underscoring the fact that Phanerozoic-style feeding strategies were subordinately present. These include the Torrowangea and Bergaueria–Palaeophycus ichnoguilds. The Torrowangea ichnoguild consists of transitory, very shallow-tier structures of vagile deposit feeders. Torrowangea and, less likely, Nenoxites are included in this ichnoguild. The Bergaueria–Palaeophycus ichnoguild comprises semi-permanent, very shallow-tier structures produced by suspension feeders or passive predators.

4. Evolutionary innovations at the terminal Ediacaran: the Nama assemblage

A number of studies have contrasted the elements of the Nama assemblage (sensu [7]) with those from the Avalon and White Sea assemblages [61,64,140–145]. From an ichnological perspective, the Nama assemblage illustrates the persistence of previous styles of animal–substrate interactions, but most importantly, the disappearance and appearance of others. A total of 12 ichnogenera comprising seven categories of architectural designs have been identified (table 2). Accordingly, no significant increase in ichnodiversity and ichnodisparity seems to have occurred. Simple horizontal trails, such as those present in the White Sea assemblage, continued being the most abundant ichnotaxa [146–148]. In a similar fashion, actively filled burrows are represented by Torrowangea [149–151], passively filled burrows by Palaeophycus [150,151] and plug-shaped burrows by Conichnus and Bergaueria [148,152]. Kimberichnus and Epibaion have not been recorded, which would be consistent with the potential extinction of their producers [119,153]. However, Kimberella has been recently documented in the terminal Ediacaran of Iran [154]. If a mollusc affinity is assumed, then this new record would fill in the stratigraphic gap between stem-molluscs from the White Sea assemblage and the crown-group molluscs documented from the Cambrian, therefore suggesting a continuity of this lineage through the Ediacaran–Cambrian transition [154].

Emphasis in recent years has been placed on the appearance of complex styles of animal–substrate interactions in terminal Ediacaran strata [148,155–164]. An increase in the density of trace fossils and the ability to penetrate into the sediment is also apparent [158–161,164]. Ichnofossil occurrences in terminal Ediacaran strata have been critically evaluated elsewhere [87], and we will focus in this section only on more recent reports. The vast majority of these new findings come from the Shibantan Member, Dengying Formation [155,160,165]. Most significantly, these new findings include some proposed bilaterian trackways (TW1 and TW2) [160], which are interpreted as recording paired appendages of a producer of unknown phylogenetic affinity (hypothesized candidates are arthropods, annelids and the last common ancestor of bilaterians). These putative trackways display a diversity of track shapes, wide track size range and poor organization, which the authors attributed to poor coordination in locomotion (pre-highly coordinated metachronal rhythm) and/or diagenesis of the carbonate surface. Indeed, trackways are real oddities in carbonate rocks due to the common absence of good lithological contrast and tendency to dissolution, particularly at interfacial surfaces. The trackways are associated with undermat burrows (BW1–3), and a possible scenario of undermat mining and walking on the mat was suggested. The material presents interesting challenges, such as possible undertracks (sensu [166]), produced by imprinting of appendages on an overlying surface seamlessly connected to an undermat burrow exit (see fig. 2, TW2 in continuity with UB3 in [160]). This palimpsest of two structures formed at different stratigraphic levels poses difficulties to the envisioned scenario, which are not easily resolvable based on documented evidence. The shorter trackway (TW1) is described as consisting of up to four sets of tracks [160, fig. 3a], organized in oblique series of four or more tracks, whereas no sets or distinct series can be identified in TW2. Interestingly, TW1 seems to display opposite symmetry (but see [160]), which in arthropods has been related to jumping rather than walking patterns [166, Pl. 7]. Another noteworthy point that arises from detailed analysis of TW2 and its contiguous burrow (UB3) is the inferred position of walking appendages in the producer. This lineal continuity (note same external width of TW2 and UB3) would imply extremely short, lateroventral articulated appendages. In short, TW1 seems to display some sort of rough imprint organization; however, the extremely irregular morphology and wide size range of tracks, as well as the trackway–burrow continuity, deserve further investigation and preclude from a definite assessment.

The recent discovery of body fossils of an elongate, segmented bilaterian (Yilingia spiciformis) in direct association with horizontal structures interpreted as locomotion and resting trace fossils [155] is fascinating and equally intriguing. One structure directly connected to Yilingia spiciformis is interpreted as a mortichnion (i.e. final walk to death sensu [166]). Although phylogenetic affinities remain uncertain, panarthropod or annelid affinities were suggested [155]. The studied assemblages seem to involve both trace fossils and body fossils associated with a microbial mat. However, some of the morphologic details of the illustrated structures regarded as trace fossils may provide room for an alternative interpretation. In particular, the possibility of Yilingia displaying a wide range of preservational variants needs to be assessed. Regardless of the phylogenetic affinity of the Yilingia producer, its functional morphology (i.e. vermiform body plan and segment-like body in the presumed absence of appendages) is compatible with a benthic mode of life and potentially an infaunal burrower. Some ‘Helminthoidichnites-like’ trace fossils displaying width ranges compatible with Yilingia size range could prompt the interpretation of a trace fossil–producer connection based on direct association (cf. [155]). However, the interpreted resting structures displaying chevron-like morphology (e.g. fig. 2a–d; Extended Data fig. 8c in [155]) seems more suggestive of poorly preserved body fossils than of trace fossils. Truncation of a previously emplaced Helminthoidichnites-like trace fossil is mentioned as evidence of the ichnofossil nature of the structures connected to Yilingia spiciformis, but the same relationship could be attained by a body fossil overprinted on a bioturbated surface disturbed by very shallow-tier burrows. Levees are unconvincing (and locally associated with body fossils, e.g. Extended Data fig. 4b in [155]), as it is the narrow area of adjacent sediment disturbance (in many photos hard to differentiate from the pustule texture of the microbial mat).

A supposed new ichnotaxon, Yichnus levis, consisting of short, disconnected, aligned spindle-shaped, horizontal segments, was recently defined in the same unit [165]. These structures are interpreted as transient burrows, resulting from the tracemaker moving repeatedly in and out of the microbial mat, with mat burrowing alternating with epibenthic intermissions, as a response to redox conditions in the water column and the mat [165]. Regardless of the controversial features of these structures, which display some morphological characteristics that are not classic of trace fossils (e.g. drastic changes in width along a segment), there are some significant problems with this interpretation and with the erection of a new ichnotaxon. This is a simple structure, passively infilled, displaying a vertical, undulatory component (the so-called in-and-out behaviour [165]), involving exiting and re-entering the substrate. The behavioural complexity inferred (i.e. as up-and-down movement interpreted as regulated by diurnal redox cycles of the cyanobacterial mat [165]) would imply daily relocation of the previous burrow to produce successive segments along the same undulatory structure (fig. 2B in [165]). In short, complexity derives from this interpretation, but the simple architecture of this structure militates against the erection of a new ichnotaxon.

A critical review of the Nama assemblage trace-fossil record suggests that three main innovations are apparent in terminal Ediacaran strata. First, three-dimensional burrow systems displaying an apparent branching pattern, referred to as treptichnids, reveal new modes of penetrating the substrate [156,157,163] (figure 2a). The ichnogenus Streptichnus shows even more sophisticated branching patterns [148], but this is now thought to occur right above the Ediacaran–Cambrian boundary [168]. Second, horizontal, increasingly more complex grazing and deposit feeding patterns (e.g. fig. 4.6 in [106]) indicate a more efficient use of food resources at the sediment–water interface, but also within the infaunal ecospace. In this context, bilobate structures displaying transverse, ring-like constrictions, referred to as Parapsammichnites, provide evidence of incipient sediment bulldozing [164] (figure 2b). Third, circular holes in the tubular fossil Cloudina, included in the ichnogenus Oichnus, represent evidence of metazoan predation of unknown affinity in the terminal Ediacaran [167,169] (figure 2c). To summarize, ichnological evidence shows the advent of an incipient Phanerozoic-style ecology, namely increased sediment mixing, substrate penetration, sediment bulldozing and drilling predation during the terminal Ediacaran. The ability to disturb the primary fabric is reflected by the increase in intensity of bioturbation (bioturbation index up to 2) in terminal Ediacaran strata [170]. Overall, this has resulted in an increased capacity to engineer shallow-marine environments [170]. Interestingly, high-resolution integration of palaeontological, stratigraphical and geochronological data from Namibia indicates that these evolutionary inovations took place essentially by the very end of the terminal Ediacaran [168].

Figure 2.

Representative trace fossils of the Nama assemblage. (a) Treptichnids and possible minute grazing trails, preserved as positive hyporelief. Huns Member, Urusis Formation, Schwarzrand Subgroup, Nama Group, Arimas Farm, Namibia. Photograph by Sören Jensen of a specimen originally published in [156]. (b) Parapsammichnites pretzeliformis, preserved as positive hyporelief. Spitskop Member of the Urusis Formation, Schwarzrand Subgroup, Nama Group, Koelkramps, Namibia. (c) Oichnus isp. in the tubular fossil Cloudina hartmannae. Dengying Formation, Lijiagou, Shaanxi Province, China. Photograph by Stefan Bengtson of specimens previously illustrated in [167].

From an ichnoguild perspective, the Helminthopsis, Torrowangea and Bergaueria–Palaeophycus ichnoguilds continued into the Phanerozoic. An incipient Treptichnus ichnoguild (or, more accurate, treptichnid ichnoguild), which consists of semi-permanent, shallow-tier, predation or, less likely, deposit-feeding structures of vagile to semi-vagile organisms, is present in terminal Ediacaran strata, and shows clear link between latest Ediacaran and early Cambrian ichnofaunas [37,94]. The Parapsammichnites ichnoguild comprises transitory, very shallow-tier, detritus-feeder structures produced by vagile animals [164]. The Oichnus ichnoguild consists of transitory, surface, predation structures produced by vagile animals [167,169]. Most ichnoguilds continued being monospecific. Noteworthy, none of the three newly added ichnoguilds show any dependency with microbial mats. Seilacher [60] portrayed undermat miners as significant components, recording the work of bilaterians, in his Ediacaran-aged, pre-Agronomic Revolution world. Paradoxically, only limited increase in ichnodisparity and ichnodiversity of this trophic group is recorded during the Ediacaran, with examples mostly concentrated in the terminal Ediacaran as shown by occurrences from Namibia [171] and China [158,161]. In other words, the real explosion in undermat miners seems to have taken place by the earliest Cambrian [38,127].

5. The dual nature of the Fortunian: relicts and novelties

The body-fossil record of the Fortunian is notoriously poorly diverse, and mostly dominated by small shelly fossils of uncertain affinity [54,172–175]. Cnidarians and sponges are relatively common, but bilaterians, although present (e.g. scalidophorans, molluscs), are not so well represented. The bilaterian small shelly fossil record, dominated by Spiralia, is in sharp contrast with the diverse Fortunian trace-fossil record [12], which displays, among many other burrow architectures, the novelty of the first ichnofossils produced by euarthropods [38,43,93,149,176]. A total of 46 ichnogenera comprising 24 categories of architectural designs have been identified (table 3). This represents the largest increase in ichnodiversity and ichnodisparity in the history of the marine biosphere [48]. Notably, this explosion in styles of animal–substrate interactions predates the increase shown by the body-fossil record (i.e. the explosion in hard parts) by approximately 17 My [12]. With the exception of the recently described ichnogenus Parapsammichnites, all the other terminal Ediacaran ichnogenera persisted into the Phanerozoic. The most outstanding aspect of the diversification event is the establishment of a wide variety of novel styles of animal–sediment interactions, reflecting both the appearance of new body plans and strategies to colonize the substrate [7,12,47] (figure 3a–g). This increase is restricted to bioturbation structures, but the Cambrian explosion did not represent a similar breakthrough for bioerosion structures, revealing that hardground communities remained poorly diverse [177]. Recent detailed analysis of ecospace utilization in the Treptichnus pedum zone also shows the protracted nature of the Fortunian diversification event [93].

Table 3.

Ichnologic composition of Terreneuvian strata. Stratigraphic ranges based on [12,48] and categories of architectural designs after [28]. Cyanobacterial microborings are not included.

stage	category of architectural designs	ichnogenera
Fortunian	6-trackways and scratch imprints	Allocotichnus
	1-simple horizontal trails	Archaeonassa
	6-trackways and scratch imprints	Asaphoidichnus
	31-pentameral-shaped imprints and burrows	Asteriacites
	3-vertical plug-shaped burrows	Bergaueria
	41-burrows with shaft or bunch with downwards radiating probes	Chondrites
	1-simple horizontal trails	Circulichnis
	1-simple horizontal trails	Cochlichnus
	34-vertical plug-shaped burrows	Conichnus
	5-bilobate trails and paired grooves	Cruziana
	2-trilobate flattened trails	Curvolithus
	19-radial to rosetted structures	Dactyloidites
	47-radial graphoglyptids	Dendrorhaphe
	5-bilobate trails and paired grooves	Didymaulichnus
	6-trackways and scratch imprints	Dimorphichnus
	6-trackways and scratch imprints	Diplichnites
	5-bilobate trails and paired grooves	Diplopodichnus
	1-simple horizontal trails	Gordia
	40-vertical helicoidal burrows	Gyrolithes
	19-radial to rosetted structures	Heliochone
	1-simple horizontal trails	Helminthoidichnites
	1-simple horizontal trails	Helminthopsis
	9-fan-shaped to radiating scratched imprints	Kimberichnus
	6-trackways and scratch imprints	Monomorphichnus
	16-horizontal branching burrow systems	Multina
	14-complex actively filled horizontal structures	Nereites
	18-surface-coverage branching burrows	Oldhamia
	10-passively filled horizontal burrows	Palaeophycus
	17-horizontal burrows with horizontal to vertical branches	Phycodes
	16-horizontal branching burrow systems	Pilichnus
	11-simple actively filled (massive) horizontal to oblique structures	Planolites
	49-regular to irregular network graphoglyptids	Protopaleodictyon
	14-complex actively filled horizontal structures	Psammichnites
	26-burrows with horizontal spreiten	Rhizocorallium
	8-bilaterally symmetrical short, scratched impressions and burrows	Rusophycus
	17-horizontal burrows with horizontal to vertical branches	Saerichnites
	17-horizontal burrows with horizontal to vertical branches	Streptichnus
	6-trackways and scratch imprints	Tasmanadia
	22-horizontal burrows with simple vertically oriented spreiten	Teichichnus
	50-maze and boxwork burrows	Thalassinoides
	11-simple actively filled (massive) horizontal to oblique structures	Torrowangea
	17-horizontal burrows with horizontal to vertical branches	Treptichnus
	41-burrows with shaft or bunch with downwards radiating probes	Trichichnus
	22-horizontal burrows with simple vertically oriented spreiten	Trichophycus
	19-radial to rosetted structures	Volkichnium
	27-burrows with helicoidal spreiten	Zoophycos
stage 2	13-simple, actively filled (pelletoidal) horizontal burrows	Alcyonidiopsis
	35-vertical unbranched burrows	Altichnus
	1-simple horizontal trails	Archaeonassa
	36-vertical single U- and Y-shaped burrows	Arenicolites
	31-pentameral-shaped imprints and burrows	Asteriacites
	34-vertical plug-shaped burrows	Astropolichnus
	34-vertical plug-shaped burrows	Bergaueria
	8-bilaterally symmetrical short, scratched impressions and burrows	Cheiichnus
	1-simple horizontal trails	Cochlichnus
	34-vertical plug-shaped burrows	Conichnus
	5-bilobate trails and paired grooves	Cruziana
	2-trilobate flattened trails	Curvolithus
	42-vertical concentrically filled burrows	Cylindrichnus
	47-radial graphoglyptids	Dendrorhaphe
	5-bilobate trails and paired grooves	Didymaulichnus
	6-trackways and scratch imprints	Dimorphichnus
	6-trackways and scratch imprints	Diplichnites
	36-vertical single U- and Y-shaped burrows	Diplocraterion
	5-bilobate trails and paired grooves	Diplopodichnus
	1-simple horizontal trails	Gordia
	19-radial to rosetted structures	Guanshanichnus
	40-vertical helicoidal burrows	Gyrolithes
	1-simple horizontal trails	Helminthoidichnites
	1-simple horizontal trails	Helminthopsis
	9-fan-shaped to radiating scratched imprints	Kimberichnus
	35-vertical unbranched burrows	Laevicyclus
	35-vertical unbranched burrows	Lingulichnus
	6-trackways and scratch imprints	Monomorphichnus
	16-horizontal branching burrow systems	Multina
	63-circular holes and pit-shaped borings	Oichnus
	18-surface-coverage branching burrows	Oldhamia
	10-passively filled horizontal burrows	Palaeophycus
	49-regular to irregular network graphoglyptids	Paleodictyon
	6-trackways and scratch imprints	Petalichnus
	17-horizontal burrows with horizontal to vertical branches	Phycodes
	11-simple actively filled (massive) horizontal to oblique structures	Planolites
	49-regular to irregular network graphoglyptids	Protopaleodictyon
	3-chevronate trails	Protovirgularia
	14-complex actively filled horizontal structures	Psammichnites
	26-burrows with horizontal spreiten	Rhizocorallium
	42-vertical concentrically filled burrows	Rosselia
	8-bilaterally symmetrical short, scratched impressions and burrows	Rusophycus
	17-horizontal burrows with horizontal to vertical branches	Saerichnites
	35-vertical unbranched burrows	Skolithos
	25-burrows with complex vertically oriented spreiten	Syringomorpha
	12-simple actively filled (meniscate) horizontal to oblique structures	Taenidium
	22-horizontal burrows with simple vertically oriented spreiten	Teichichnus
	17-horizontal burrows with horizontal to vertical branches	Treptichnus
	41-burrows with shaft or bunch with downwards radiating probes	Trichichnus
	22-horizontal burrows with simple vertically oriented spreiten	Trichophycus

Figure 3.

Representative trace fossils of the Fortunian. (a) Oldhamia alata, preserved as positive hyporelief. Puncoviscana Formation, El Mollar, Quebrada del Toro, northwest Argentina. (b) Gyrolithes scintillus, vertical cross-section view. Member 2, Chapel Island Formation, Fortune Head, Burin Peninsula, Newfoundland, Canada. Image from [47] and reproduced with permission of Springer. (c) Psammichnites isp., preserved as negative epirelief. Member 3, Chapel Island Formation, Grand Bank Head, Burin Peninsula, Newfoundland, Canada. (d) Cochlichnus anguineus, preserved as positive hyporelief. Member 2, Chapel Island Formation, Grand Bank Head, Burin Peninsula, Newfoundland, Canada. (e) Monomorphichnus isp., preserved as positive hyporelief. Chapel Island Formation, Lewin's Cove, Burin Peninsula, Newfoundland, Canada. (f) Tasmanadia cachii, preserved as positive hyporelief. Puncoviscana Formation, Cachi, northwest Argentina. (g) Treptichnus pedum, preserved as positive hyporelief. Lower Member, Wood Canyon Formation, Death Valley, western USA.

The marked increase in ichnodiversity and ichnodisparity coexisted with the persistence of Ediacaran-style feeding strategies, giving the Fortunian a so-called dual nature [47,62]. This is revealed by the common presence of microbially induced sedimentary structures in direct association with trace fossils [38,47,62,178]. From an ichnoguild perspective, the common presence of the Helminthopsis ichnoguild in microbial mats gives Fortunian deposits an Ediacaran flavour. Scratch impressions comparable with radular imprints produced by modern molluscs, originally assigned to Radulichnus, have been recorded from Fortunian matgrounds [178]. The most common mat scratchers in the Fortunian, however, seem to have been arthropods, as indicated by superbly preserved trace fossils recording the action of legs against a microbially stabilized surface and showing that even some of the new tracemakers may have adopted Ediacaran-style ecologies [38]. However, inferring trophic types from arthropod trackways and scratch imprints is not straightforward, and the possibility of these trace fossils representing predation on mat grazers cannot be disregarded [83]. In addition, the dual character of the Fortunian is also evidenced by the fact that the exploitation of microbial mats was accomplished using innovative, energy-efficient feeding strategies unknown in the Ediacaran. One of the best examples of these complex strategies is recorded by Oldhamia, a specialized undermat miner typical of the early Cambrian [34,60,179] (figure 3a).

Overall intensities of bioturbation remained quite low during the Fortunian, reaching up to bioturbation index 3 [12]. In fact, the vast majority of trace fossils are preserved along lithological interfaces, revealing the work of semi-infaunal to shallow-tier structures in low- to moderate-energy settings [12,37,38,180–183]. High-energy, nearshore areas were essentially uncolonized by the infauna [62]. However, uncontroversial vertical burrows, such as Gyrolithes (figure 3b), have been recorded for the first time from the early Fortunian, revealing a more nuanced scenario [157,184–187]. These Fortunian Gyrolithes have been recently interpreted as permanent domiciles with the main purpose of farming bacteria, with the helical morphology both serving as sediment holdfast and providing a maximum surface area for bacterial gardening [187]. Construction of spiral burrows in strongly oxygen-stratified sediment sealed by microbial mats may have resulted in a sharp redox gradient across the burrow boundary, attracting bacteria that provided the source of food for the Gyrolithes producer [187]. Notably, farming strategies tend to be common in food-depleted settings, which may have been the case of Cambrian seas, a scenario consistent with the onshore–offshore trends reflected by the farming burrows referred to as graphoglyptids, which have been recorded in shallow water in the early Cambrian, becoming later essentially restricted to the deep sea [188]. This trend is consistent with the fact that Gyrolithes has been recorded in association with the graphoglyptid Dendroraphe in the Fortunian [184].

In addition, incipient sediment bulldozing became more important particularly later in the Fortunian as shown by the appearance of the ichnogenera Psammichnites (figure 3c), Didymaulichnus and Curvolithus [35,149,186,189]. Some degree of sediment mixing may have been attained by benthic trilobites and other arthropods, as well as by various worm-like animals. Albeit limited in intensity and depth, bioturbation styles show a marked departure from those in the Ediacaran, most likely having played a role in triggering changes in the sediment and the water column. Biogenic activity, although relatively close to the sediment–water interface, produced some disruption of the primary fabric, deepening of the redox discontinuity surface and generating concomitant improved substrate ventilation, and flux exchange with the water column [12,190].

6. The rise of complex ecosystems: Cambrian Ages 2–4

Regardless of the seemingly minor increases in ichnodiversity and ichnodisparity recorded during Cambrian Age 2 (50 ichnogenera and 28 categories of architectural designs) (table 3), significant change is revealed by other aspects of the trace-fossil record (e.g. tiering structure, degree of bioturbation), as well as by other lines of palaeobiological evidence. Examination of the trace-fossil record suggests that Cambrian Age 2 signalled the rise of complex shallow-marine ecosystems of modern aspect. Accordingly, it has been suggested that the Seilacherian Agronomic Revolution (i.e. the replacement of matgrounds by mixgrounds) did not coincide with the Ediacaran–Cambrian transition, but actually took place during Cambrian Age 2 [12,62]. Major breakthroughs were attained in two distinct shallow-water settings. First, a deep-tier infauna dominated by suspension feeders became established in nearshore, relatively high-energy, mobile sandy substrates considerably expanding the range of environmental settings colonized during the earliest Cambrian [12,62]. Although ichnofabrics dominated by vertical burrows of suspension feeders occurred for the first time in Cambrian Stage 2 sandstone (e.g. [191,192]), these relatively dense vertical burrow fabrics became even more common later in the early Cambrian (e.g. [193–196]). Typical ichnofabrics are represented by Skolithos (figure 4a), but other elements, such as Arenicolites and Diplocraterion (figure 4c), are important elements as well (e.g. [197,198]). In some examples, vertical burrows of detritus feeders, such as Rosselia (figure 4b), are locally present [196]. Second, a significant increase in the intensity of biogenic mixing, mostly by deposit and detritus feeders, took place in lower-energy offshore settings, leading to the establishment of intensely bioturbated sediments under fully marine conditions [12,47,62,186,198–203]. In addition to sediment bulldozers (figure 4d), dense occurrences of the spreiten burrow Teichichnus and mottled ichnofabrics became common [199,201,204] (figure 4e). For the first time, a maximum bioturbation index of 6 was attained [12]. These trends were accentuated during the rest of the early Cambrian. During Cambrian Ages 3–4, a new phase of ichnodiversification took place (67 ichnogenera in Cambrian Age 3), although not followed by an equally significant increase in ichnodisparity (table 4).

Figure 4.

Representative trace fossils and ichnofabrics of Cambrian Stage 2 and Series 2. (a) Skolithos linearis forming a typical piperock, vertical cross-section view. Lake O'Hara Member, St Piran Formation, Gog Group, Lake O'Hara, Canadian Rockies. Photograph by Patricio Desjardins. (b) Rosselia isp. vertical cross-section view. Lake O'Hara Member, St Piran Formation, Gog Group, Lake Magog, Canadian Rockies. Photograph by Patricio Desjardins. Image from [47] and reproduced with permission of Springer. (c) Diplocraterion parallelum, vertical cross-section view. Dividalen Group, Imobekken, northern Norway. Image from [47] and reproduced with permission of Springer. (d) Psammichnites gigas, as seen on a sandstone base. Shiyantou Formation, Meishucun, Yunnan Province, China. (e) Teichichnus rectus (arrowed) superimposed on typical mixed-layer burrow mottling. Member 5, Chapel Island Formation, Little Dantzic Cove, Burin Peninsula, Newfoundland, Canada. Photograph by Romain Gougeon. (f) Dense occurrence of Thalassinoides isp. ichnofabrics, vertical cross-section view. Zhushadong Formation, Guankou, Henan Province.

Table 4.

Ichnologic composition of Cambrian Series 2 strata. Stratigraphic ranges based on [12,48] and categories of architectural designs after [28]. Cyanobacterial microborings are not included.

stage	category of architectural designs	ichnogenera
stage 3	35-vertical unbranched burrows	Altichnus
	1-simple horizontal trails	Archaeonassa
	36-vertical single U- and Y-shaped burrows	Arenicolites
	33-dumbbell- and arrow-shaped burrows	Arthraria
	31-pentameral-shaped imprints and burrows	Asteriacites
	43-horizontal, branched concentrically filled burrows	Asterosoma
	34-vertical plug-shaped burrows	Astropolichnus
	34-vertical plug-shaped burrows	Bergaueria
	33-dumbbell- and arrow-shaped burrows	Bifungites
	8-bilaterally symmetrical short, scratched impressions and burrows	Cheiichnus
	41-burrows with shaft or bunch with downwards radiating probes	Chondrites
	1-simple horizontal trails	Cochlichnus
	34-vertical plug-shaped burrows	Conichnus
	34-vertical plug-shaped burrows	Conostichus
	45-guided meandering graphoglyptids	Cosmorhaphe
	5-bilobate trails and paired grooves	Cruziana
	2-trilobate flattened trails	Curvolithus
	42-vertical concentrically filled burrows	Cylindrichnus
	19-radial to rosetted structures	Dactyloidites
	5-bilobate trails and paired grooves	Didymaulichnus
	6-trackways and scratch imprints	Dimorphichnus
	6-trackways and scratch imprints	Diplichnites
	36-vertical single U- and Y-shaped burrows	Diplocraterion
	5-bilobate trails and paired grooves	Diplopodichnus
	1-simple horizontal trails	Gordia
	40-vertical helicoidal burrows	Gyrolithes
	21-horizontal burrows with serial chambers	Halimedides
	22-horizontal burrows with simple vertically oriented spreiten	Halopoa
	1-simple horizontal trails	Helminthoidichnites
	1-simple horizontal trails	Helminthopsis
	35-vertical unbranched burrows	Laevicyclus
	35-vertical unbranched burrows	Lingulichnus
	47-radial graphoglyptids	Lorenzinia
	34-vertical plug-shaped burrows	Mammillichnus
	49-regular to irregular network graphoglyptids	Megagrapton
	19-radial to rosetted structures	Monocraterion
	6-trackways and scratch imprints	Monomorphichnus
	26-burrows with horizontal spreiten	Multilamella
	23-horizontal spiral burrows	Multilaqueichnus
	16-horizontal branching burrow systems	Multina
	14-complex actively filled horizontal structures	Nereites
	63-circular holes and pit-shaped borings	Oichnus
	18-surface-coverage branching burrows	Oldhamia
	10-passively filled horizontal burrows	Palaeophycus
	49-regular to irregular network graphoglyptids	Paleodictyon
	6-trackways and scratch imprints	Petalichnus
	17-horizontal burrows with horizontal to vertical branches	Phycodes
	11-simple actively filled (massive) horizontal to oblique structures	Planolites
	37-vertical multiple U- and Y-shaped burrows	Polykladichnus
	3-chevronate trails	Protovirgularia
	14-complex actively filled horizontal structures	Psammichnites
	30-isolated and serial oval to almond-shaped burrows	Ptychoplasma
	26-burrows with horizontal spreiten	Rhizocorallium
	42-vertical concentrically filled burrows	Rosselia
	8-bilaterally symmetrical short, scratched impressions and burrows	Rusophycus
	17-horizontal burrows with horizontal to vertical branches	Saerichnites
	19-radial to rosetted structures	Scotolithus
	35-vertical unbranched burrows	Skolithos
	25-burrows with complex vertically oriented spreiten	Syringomorpha
	12-simple actively filled (meniscate) horizontal to oblique structures	Taenidium
	22-horizontal burrows with simple vertically oriented spreiten	Teichichnus
	50-maze and boxwork burrows	Thalassinoides
	17-horizontal burrows with horizontal to vertical branches	Torrowangea
	17-horizontal burrows with horizontal to vertical branches	Treptichnus
	22-horizontal burrows with simple vertically oriented spreiten	Trichophycus
	59-cylindrical vertical to oblique borings	Trypanites
	27-burrows with helicoidal spreiten	Zoophycos
stage 4	13-simple, actively filled (pelletoidal) horizontal burrows	Alcyonidiopsis
	1-simple horizontal trails	Archaeonassa
	36-vertical single U- and Y-shaped burrows	Arenicolites
	33-dumbbell- and arrow-shaped burrows	Arthraria
	31-pentameral-shaped imprints and burrows	Asteriacites
	43-horizontal, branched concentrically filled burrows	Asterosoma
	34-vertical plug-shaped burrows	Astropolichnus
	34-vertical plug-shaped burrows	Bergaueria
	33-dumbbell- and arrow-shaped burrows	Bifungites
	8-bilaterally symmetrical short, scratched impressions and burrows	Cheiichnus
	41-burrows with shaft or bunch with downwards radiating probes	Chondrites
	1-simple horizontal trails	Cochlichnus
	34-vertical plug-shaped burrows	Conostichus
	45-guided meandering graphoglyptids	Cosmorhaphe
	5-bilobate trails and paired grooves	Cruziana
	2-trilobate flattened trails	Curvolithus
	42-vertical concentrically filled burrows	Cylindrichnus
	19-radial to rosetted structures	Dactyloidites
	5-bilobate trails and paired grooves	Didymaulichnus
	6-trackways and scratch imprints	Dimorphichnus
	6-trackways and scratch imprints	Diplichnites
	36-vertical single U- and Y-shaped burrows	Diplocraterion
	1-simple horizontal trails	Gordia
	19-radial to rosetted structures	Guanshanichnus
	25-burrows with complex vertically oriented spreiten	Gyrochorte
	40-vertical helicoidal burrows	Gyrolithes
	21-horizontal burrows with serial chambers	Halimedides
	22-horizontal burrows with simple vertically oriented spreiten	Halopoa
	1-simple horizontal trails	Helminthoidichnites
	1-simple horizontal trails	Helminthopsis
	35-vertical unbranched burrows	Laevicyclus
	35-vertical unbranched burrows	Lingulichnus
	47-radial graphoglyptids	Lorenzinia
	34-vertical plug-shaped burrows	Mammillichnus
	49-regular to irregular network graphoglyptids	Megagrapton
	19-radial to rosetted structures	Monocraterion
	6-trackways and scratch imprints	Monomorphichnus
	26-burrows with horizontal spreiten	Multilamella
	16-horizontal branching burrow systems	Multina
	14-complex actively filled horizontal structures	Nereites
	63-circular holes and pit-shaped borings	Oichnus
	18-surface-coverage branching burrows	Oldhamia
	10-passively filled horizontal burrows	Palaeophycus
	49-regular to irregular network graphoglyptids	Paleodictyon
	6-trackways and scratch imprints	Petalichnus
	17-horizontal burrows with horizontal to vertical branches	Phycodes
	11-simple actively filled (massive) horizontal to oblique structures	Planolites
	37-vertical multiple U- and Y-shaped burrows	Polykladichnus
	6-trackways and scratch imprints	Protichnites
	3-chevronate trails	Protovirgularia
	14-complex actively filled horizontal structures	Psammichnites
	30-isolated and serial oval to almond-shaped burrows	Ptychoplasma
	26-burrows with horizontal spreiten	Rhizocorallium
	42-vertical concentrically filled burrows	Rosselia
	8-bilaterally symmetrical short, scratched impressions and burrows	Rusophycus
	17-horizontal burrows with horizontal to vertical branches	Saerichnites
	19-radial to rosetted structures	Scotolithus
	11-simple actively filled (massive) horizontal to oblique structures	Sericichnus
	35-vertical unbranched burrows	Skolithos
	25-burrows with complex vertically oriented spreiten	Syringomorpha
	22-horizontal burrows with simple vertically oriented spreiten	Teichichnus
	50-maze and boxwork burrows	Thalassinoides
	17-horizontal burrows with horizontal to vertical branches	Treptichnus
	22-horizontal burrows with simple vertically oriented spreiten	Trichophycus
	59-cylindrical vertical to oblique borings	Trypanites
	27-burrows with helicoidal spreiten	Zoophycos

These evolutionary innovations were conducive to profound restructuring of shallow-marine ecosystems, promoting a more efficient vertical and horizontal transfer of organic carbon from the water column to the substrate and within the sediment, respectively [12,44,205–207]. The evolution of filter-feeding mesozooplankton and filter-feeding nekton resulted in repacking of small, unicellular phytoplankton into larger, more rapidly sinking biomass particles, which fertilized the benthos [55,205,206,208]. This addition of the tier of primary consumers and the improved delivery of organics to the sediment resulted in a dramatic increase in quality (eukaryote- and pellet-dominated) and amount of food available for the benthic system. These changes facilitated a more efficient use of the infaunal ecospace [12,209] and an increase in the complexity (i.e. length and connectivity) of food webs [210–216]. The basic principles regulating modern ecosystem functioning were already in place by Cambrian Epoch 2, and trophic webs consisted of primary production by phytoplankton, primary zooplanktonic consumers, secondary planktonic and nektonic consumers, suspension feeders, detritivores and scavengers, and benthic–nektobenthic primary consumers [212,213]. This picture is complemented by available ichnological evidence of predation on the basis of bioturbation [217] and bioerosion [218] structures.

Multiple lines of evidence suggest that these changes have had significant geobiological impact with complex feedback loops acting at different scales [12,47,62,219,220]. The establishment of densely packed, deep-tier, suspension-feeder communities in water-agitated nearshore areas may have helped to regenerate nitrogen and phosphorus to the water column [221,222]. Suspension feeders typically consume large amounts of seston and filter vast volumes of water, engineering their environment by preventing eutrophication, decreasing water turbidity and making light available for microphytobenthos [221–223]. This style of bioturbation promotes ventilation and transport of solutes within the burrow and into the host sediment due to the expansion of the area of sediment–water interface available for exchange [59,224–227]. High densities of burrow mazes (figure 4f), which are typically present in lower-energy settings, also tend to favour expansion of aerobic bacteria, fluxes of nitrogen and ammonium across the sediment–water interface, and increase in the rate of organic matter decomposition and the regeneration of nutrients [228]. The overall high density of burrows allows their oxygenated zones to meet, further expanding the effect of bioturbation [228,229]. Deeply emplaced burrows also contribute to the advective transport of sediment to the surface from below. Intense bioturbation by the activity of mobile detritus and deposit feeders in low-energy offshore areas produced significant churning of the primary fabric [12,47,62,198,199,201–203]. This style of biogenic mixing also affected geochemical cycles in many ways, including enhancing fluxes of organic carbon and dissolved inorganic nitrogen into the sediment [199]. Overall, the increase in depth of the bioturbated zone was accompanied by a concomitant downward shift of the redox discontinuity surface, opening new ecological niches [12,199].

7. Discussion

Understanding the underlying mechanisms, pattern of evolution (i.e. phylogeny) and timing of the rise of animals has been the focus of intense research during the last few decades. There is general consensus that significant advance can only be obtained through the integration of multiple datasets. In this context, our up-to-date systematic and critical review of the trace-fossil record, highlighting a full artillery of ichnological concepts and recently developed analytical tools, attempts to contribute to the debate. Continuous documentation of new ichnological findings and re-interpretation of Ediacaran trace fossils have been the bulk of most research. Available evidence previously discussed strongly supports that the earliest uncontroversial records of bilaterian activity are approximately 560–555 My old. Remarkably, after more than 15 years of close scrutiny of Ediacaran sections worldwide, this estimation is still in line with that based on a previous critical and comprehensive evaluation of the trace-fossil record [6].

The discussion about small size in early metazoans is a topic that is recurrently revisited [e.g. 230–234]. The basic idea is that small size might have played a crucial role in their virtual absence in the Precambrian fossil record, favouring a strong taphonomic bias over actual non-existence. Small size could be an unsurmountable barrier for the fossil record as a whole (including both body and trace fossils). The underlying reasoning is that the trace-fossil record consists only of structures produced by the displacement of relatively large organisms ploughing the sea floor, and that tiny trace fossils (less than 1 mm), if produced, are not expected to survive the taphonomic barrier. However, the entrenched idea of non-preservability of small traces has been repeatedly falsified by the increasing number of reports of meiofaunal trace fossils in Phanerozoic siliciclastics and limestones [132,133,235,236], and even one putative record in the terminal Ediacaran [237]. In the absence of profuse and penetrative vertical bioturbation, and mediated by a microbially stabilized substrate, the anactualistic conditions for the preservation of tiny, including meiofauna-size, trace fossils are met (e.g. [127,132,235]). The widespread presence of microbially stabilized substrates opened a high-fidelity taphonomic window that persisted through the Ediacaran–Cambrian transition, strongly arguing against preservational bias as detrimental of small size. In fact, the level of fidelity shown by the Ediacaran–Cambrian surficial and shallow-tier trace-fossil record is mostly unparalleled in younger strata [38,43,87,181–183].

The trace-fossil record of early bilaterians may be understood as consisting of four main evolutionary phases. These phases track the history of bilaterian evolution through the lens of ichnological evidence from the first records, through the major radiation phase of the Cambrian explosion, to the advent of a Phanerozoic-style benthic ecology, encompassing roughly 50 My of evolutionary change (figure 5). The first phase (approx. 560–540 Ma) is herein viewed as comprising the early phase of bilaterian evolution. Regardless of intense scrutiny, the vast majority of the ichnologic record is essentially represented by very simple (unbranched) trails and shallow burrows that can be confidently assigned to the work of bilaterians [6,84]. Morphologic features of some of the ichnotaxa present suggest animals with some kind of internal cavity and a hydrostatic skeleton, particularly when dealing with shallow-tier structures that require some penetration in the sediment. Inferred bilaterian locomotion mechanisms, such as slide and push-and-pull (e.g. Archaeonassa) and possibly peristalsis (e.g. Torrowangea), have been proposed [117]. In addition, the presence of biogenic structures, such as a series of Epibaion occasionally preserved together with its producer, demonstrates active motility by some elements of the Ediacara biota exploiting mat resources [89–92,238]. Additional evidence of motility in members of the Ediacara biota is based on body fossils, such as Parvancorina [239–241]. Food availability was dominantly governed by the presence of microbial films and mats, and exploited through grazing (both superficial and within the mat), undermat mining, direct absorption through ventral surfaces and potentially scratching. These feeding types, although some continued to be present in Phanerozoic ecosystems, define completely different ecological niches, and depict an anactualistic synecological scenario. This phase is recorded by the White Sea assemblage and a significant part of the Nama assemblage.

Figure 5.

Summary of ichnodiversity and ichnodisparity changes during the Ediacaran–Cambrian Series 2, and iconic representation of evolutionary phases (modified from [12]). Phase 1 (Ediacaran) shows the earliest ichnological evidence of locomotion by bilaterians. Simple grazing trails are dominant (e.g. Gordia, Helminthoidichnites, Helminthopsis), but other classic members of the Ediacara biota also record motility (e.g. resting traces of dickinsonids: Epibaion). This phase is typically illustrated by the White Sea assemblage, but the oldest deposits representative of the Nama assemblage are part of this phase as well. Phase 2 (Ediacaran) records a prelude to the Cambrian explosion, as evidenced by the appearance of more complex and penetrative burrows (treptichnids), trace fossils of bulldozers (Parapsammichnites) and drilling predation in Cloudina (Oichnus). Patches of more bioturbated sediment are apparent in shallow, fully marine settings. This phase is recorded in strata corresponding to the youngest representatives of the Nama assemblage. Both phases 1 and 2 are clearly manifested in environments between the fair-weather and storm wave bases (i.e. offshore). Phase 3 (Fortunian) is characterized by the appearance of novel architectures reflecting the evolution of novel body plans, significantly euarthropods (e.g. Diplichnites isp., Rusophycus avalonensis), more penetrative burrows (e.g. Treptichnus pedum, various ichnospecies of Gyrolithes) and patterned feeding strategies (e.g. various ichnospecies of Oldhamia) that allowed more efficient exploitation of fine-grained sediment and microbial mats. Although most of the ichnologic information comes from offshore settings, evidence of colonization of very shallow intertidal settings and deep-marine environments is recorded. Phase 4A (Cambrian Age 2) is signalled by an increase in depth and extent of bioturbation. Nearshore areas were colonized by a deep-tier, suspension-feeding (e.g. Skolithos, Diplocraterion) and detritus-feeding (e.g. Rosselia) infauna, and offshore environments by a stationary and vagile deposit- and detritus-feeder benthos. Phase 4B (Cambrian Ages 3–4) is similar to the previous one, but recording a renewed diversification at ichnogenus level and more extensive colonization of agitated nearshore settings as shown by the widespread occurrence of Skolithos piperock.

The second phase (approx. 540–538 Ma) records the proximate roots of the Cambrian explosion, and is represented by the appearance of more specialized trace fossils, reflecting more efficient ways of interacting with the substrate and modifying the primary sedimentary fabric. Exploitation of microbial mats persisted, but the sea floor may have become increasingly heterogeneous [106] with patches resulting from the activity of sediment bulldozers reaching moderate intensity of bioturbation [164]. It has been hypothesized that increased deposit-feeding activity may have negatively impacted on microbially stabilized surfaces and sessile suspension-feeding and/or osmotrophic communities, representing an early example of trophic-group amensalism [164]. In comparison with the previous phase, the role of bioturbators as ecosystem engineers able to modify their environment and to impact on structure and functioning of shallow-marine ecosystems may have been considerably higher [170]. Incidentally, the increase in density of trace fossils as seen on bedding planes commonly results in burrow overlap, which is rare in older strata. The downside of this is that burrow overlap can be confused with branching, leading to the mistaken assumption of complex burrow architectures and behaviours. This phase encompasses the younger part of the Nama assemblage. High-resolution dating in the Spitskop Member and the Nomtsas Formation (Nama Group) at the classic locality of Swarpunt Farm may indicate that this phase may have been quite short [168]. However, treptichnids are present in the older Huns Member [156], suggesting that some of these innovations can be tracked farther down into the Nama Group. Therefore, the 540 My lower boundary of this phase should be taken as a minimum age; the scarcity of high-resolution radiometric dating through this part of the Nama succession prevents further precisions at this moment. In the same vein, a refined chronostratigraphic framework of the terminal Ediacaran succession in the Yangtze Gorges area (South China) would allow placing some of the more recent findings in the Dengying Formation [155,158,160,161,165] within a more refined chronology. It is even possible that the lower limit of this phase may have been coincident with the appearance of organisms capable of developing hard parts, such as tubular fossils [242–244] and reef-forming organisms [245–248].

The third phase (approx. 538–529 Ma) is characterized by an unparalled increase in the number of trace-fossil morphologies, essentially providing overwhelming evidence of a rapid diversification of body plans, earlier than suggested by the body-fossil record [12]. This phase was characterized by the appearance of sophisticated detritus and deposit feeding and grazing strategies able to efficiently exploit fine-grained clastic sediment and microbial mats. Patchy distribution pattern of trace fossils suggests concentration of resources in an increasingly more heterogeneous sea bottom [12,47,62,249,250]. Detritus- and shallow deposit-feeding strategies are illustrated by a variety of highly sinuous grazing trails, showing evidence of strophotactic, phobotactic and thigmotactic behaviours. These complex structures targeting organic-rich concentrations indicate a diversity of animals equipped with finely tuned navigational devices [12,249,250]. This level of complexity in sensorial systems is also revealed by undermat miners. Although microbial biofilms were the main food resource available to early bilaterians during the Ediacaran [106], they were exploited in the Fortunian in a completely innovative way, unquestionably recording the work of new producers. The iconic ichnotaxon Oldhamia best illustrates the wide variety of highly patterned morphologies, most likely reflecting the radiation of a clade using the resource of microbially stabilized surfaces [34,179]. Fortunian interactions were mostly restricted to diffusion-dominated benthic systems [12]. Precise chronostratigraphic dating is lacking in many sections, preventing establishing a robust chronology of these innovations through the Fortunian. However, detailed work in the Chapel Island Formation of Newfoundland seems to show an increase in infaunalization through the Fortunian [199]. Noteworthy, however, is the occurrence of Gyrolithes scintillus, the earliest vertical burrow recorded at the base of member 2 of the Chapel Island Formation, immediately above the Precambrian–Cambrian GSSP [187].

Although the precise phylogenetic affinity of the producers cannot be determined with certainty, the fact that the Fortunian witnessed the first occurrence of a large number of new architectural designs reveals a diversity of body plans (i.e. new phyla). The presence of trackways and various types of scratch imprints (e.g. Diplichnites, Allocotichnus, Dimorphichnus, Monomorphichnus) and scratched bilobate trails and impressions (e.g. Cruziana, Rusophycus) provides solid evidence of euarthropods [5,36,38,43,93,149]. The systematic branching pattern displayed by Treptichnus pedum argues in favour of production by priapulids based on observations of modern structures [94,95]. The highly regular sinusoidal course that is diagnostic of the ichnogenus Cochlichnus is identical to locomotion traces produced by free-living nematodes [34,251,252], strongly suggesting the presence of this clade approximately 60 My earlier than indicated in previous studies [253] (but see [237]). Polychaetes and crustaceans have been proposed as producers of the vertical spiral burrow Gyrolithes in post-Palaeozoic strata, the latter based on intergradations with typical galleries confidently assigned to decapods [187]. None of these intergradations have been observed in Palaeozoic ichnospecies. In particular, polychaetes and enteropneusts have been proposed as potential producers of the early Fortunian Gyrolithes scintillus; notably, the size and architecture of the modern enteropneust Saccogossus inhacensis [254] being strikingly similar to this early Cambrian ichnotaxon [187]. In other cases, the link between a trace fossil and its producer is more tenuous, as illustrated by Oldhamia. However, even in this case, morphologic details of some of the Oldhamia ichnospecies (e.g. O. geniculata) indicate efficient burrowing suggestive of a coelomate producer, enteropneusts and annelids being the best potential candidates.

The fourth phase (approx. 529–509 Ma) is signalled by a profound ecosystem restructuring, involving the establishment of a suspension-feeder infauna in nearshore environments (i.e. shoreface, shallow subtidal to lower intertidal) and a highly mobile deposit- and detritus-feeder benthos in offshore settings. This restructuring involves substantial increase in the length and degree of connectivity of the trophic web and coupling of benthos and plankton systems [12]. In particular, the profuse deep-tier suspension-feeding infauna records the work of ecological engineers in advection-dominated benthic systems. In this scenario, ecosystem restructuring at macroevolutionary level is a necessary, but belated consequence of the ongoing radiation. The profound results of the so-called Agronomic Revolution are revealed by the advent of a modern benthic structure [12]. The internal dynamics of this phase may have been quite complex, with the initial ecological changes of Cambrian Age 2 (Phase 4A in figure 5) followed by further diversification in styles of animal–substrate interactions in Cambrian Series 2, most likely reflecting pulses of progressive environmental expansion (Phase 4B in figure 5). It has been hypothesized that the increase in depth and intensity of bioturbation during Cambrian Age 2 may have been the trigger of the subsequent diversification, representing ecological spillover [12]. High-resolution ichnological analysis in the Chapel Island Formation of Newfoundland showed that some innovations (e.g. increase in the degree of bioturbation by deposit and detritus feeders) may have been incipient by the very end of the Fortunian [199]. Ichnological evidence of profound ecosystem restructuring is consistent with geochemical proxies [219,255–257]. Based on sedimentary Mo and U concentration proxies, it has been proposed that unprecedented widespread oxygenation took place in the early Cambrian with modern-like levels of oxygen content reached around 521 Ma when it is inferred that strongly to moderately oxic water expanded over vast areas of the seafloor [257]. This oxygenation phase occurred during Cambrian Ages 2–3, and is mostly coincident with the Cambrian explosion based on body fossils [257].

Vertical burrows in near shore sandstone provide evidence of various clades of worm-like organisms able to penetrate deeply into the sediment. Of these, Skolithos has been attributed to a variety of infaunal producers able to feed on suspended particles of organic matter, typically lophophorate phoronids and tentacular-crowned polychaetes [258]. Based on observations in modern environments, the concentrically infilled and funnel- to spindle-shaped Rosselia has been attributed to terebellid, cirritulid and spionid polychaetes [259–262]. Interestingly, discrepancies persist between morphological cladistic and molecular phylogenetic datasets with respect to annelid phylogeny (and particularly the relationships of polychaetes and sipunculans), but it has been suggested that the concordance of the miRNA phylogeny and the fossil record may indicate that the earliest annelids were epibenthic, rather than infaunal [263]. This view on life habits is supported by the majority of Cambrian body fossils, which based on functional morphology, were dominantly epibenthic detritus feeders [264–266]. However, Peronochaeta dubia from the Burgess Shale may record an infaunal mode of life [267]. Other novel burrow morphologies are seen in lower-energy settings, as shown by some of the forms included in Psammichnites (e.g. P. gigas) having a complex internal backfill structure and a median groove, which is regarded as an expression of a sinuous siphon by a slug-like mollusc, maybe related to halkieriids [268,269].

These evolutionary changes did not take place in all environments along the marine depositional profile at the same time [47]. This diachronism has been detected in more detail in ichnological studies framed from a sedimentological and sequence-stratigraphical perspective [36,68]. The bulk of the trace-fossil record comes from siliciclastic settings located between the fair-weather and storm-wave bases, collectively referred to as the offshore (including the shallower-water offshore transition) [36,47]. Overall, innovations took place first in these settings, and further expanded into shallower- and deeper-water settings [47]. For example, two of the main evolutionary breakthroughs of the terminal Ediacaran, the appearance of three-dimensional branching burrows produced by infaunal organisms capable of more penetrative bioturbation [148,156,170] and of more sophisticated burrows produced by sediment bulldozers [164], occurred in offshore environments, with a third one, drilling predation [167,169], taking place in carbonate settings. The pattern of subsequent landward and seaward environmental expansion of evolutionary innovations has been repeatedly documented from younger strata as well [270,271]. As a result of diachronism, deposits from environments that have experienced delayed colonization may show anachronistic features. For example, some lower Fortunian deposits formed between the storm wave base and the shelf break contain a trace-fossil assemblage that is undistinguishable from an Ediacaran one [68]. However, environmental factors are never as pervasive as to completely bias detection of temporal changes through the Ediacaran–Cambrian transition if ichnological data are properly framed. The integration of sedimentological, ichnological and sequence-stratigraphical datasets shows that environmental controls are overprinted to first-order evolutionary signals [36]. In addition, some key ichnotaxa, such as Treptichnus pedum, conveniently display a wide range of environmental tolerance across the depositional profile (typically above storm-wave base), helping in establishing global correlations [36,272,273].

Overall, this review of the Ediacaran–Cambrian trace-fossil record reinforces the overarching notion that animal–substrate interactions have effectively engineered the biosphere (e.g. [12,41,58,62,64,71,255,256,274]). Subsequent to their origin most likely under low-oxygen conditions in the Neoproterozoic [72,73], animals may have altered oxygen content themselves triggering feedback loops [219,220,255–257] and in complex ways that we are just starting to decipher. Alternatively, it has been suggested that the role of bioturbation has been limited until the late Silurian [180,275]. There is overall agreement that bioturbation intensity and depth increased through geologic time parallel to an increase in ichnodiversity [48,177]. However, a scenario supporting little geobiological impact of bioturbation in the early Paleozoic is hard to reconcile not only with available ichnological data, but also with other lines of evidence. In addition to the problems outlined elsewhere [47,62], recent work in units previously considered as sparsely bioturbated [180] has revealed significant disruption of the primary sedimentary fabric and the overprinting of transition-layer burrows upon indistinct bioturbation mottlings, indicative of the presence of the mixed layer, up-section [199]. This is consistent with work elsewhere showing that bioturbation was intense under fully marine conditions during the early Paleozoic (e.g. [198,202,203,276]). Disparity in results most likely reflects biases in the methodological approach, most notably the targeting of heterolithic facies [180,275] which, by virtue of consisting of intercalations of discrete (i.e. not being biogenically mixed) sandstone and mudstone layers, invariably displays low to moderate bioturbation intensities regardless of their ages [62].

Finally, the reconstruction of global oceanic redox conditions and understanding of complex biotic feedbacks with large-scale oxygen fluctuations across the Ediacaran–Cambrian transition has been focus of intense research, and not without controversy (e.g. [74,219,220,255,256,274]). Geochemical proxies, however, typically allow estimating extensive global sea floor anoxia, without distinguishing between deep- and shallow-marine settings. Articulation of high-resolution local (ecological) data within the framework of shelf- to deep-marine transects at basin scale are only available for few basins [257,277–279]. Trace-fossil data can also contribute substantially to improve resolution at ecological scale, yielding insights into the ongoing debate between a laterally extensive redox stratification model of the oceans and a dynamic oxygen minimum zone model overlying oxic basin waters [277]. For example, integration of high-resolution ichnological and sedimentological data provides evidence of in situ communities in Fortunian deep-sea floors (e.g. Oldhamia assemblage in the Puncoviscana Basin [280]), recording oxygenated deep-sea bottoms and favouring the idea of spatially heterogeneous oxygenation of the deep-marine realm during the earliest Cambrian. Deep-time global inferences need to be supported by environmentally constrained, basin-normalized, robust geochemical data intimately integrated to body-fossil and trace-fossil data. Only through this laborious puzzle construction will we be able to articulate large datasets, unravel a patchy and complex scenario, and understand potential basin-scale controls affecting global models.

8. Conclusion

The significance of the trace-fossil record to provide insights into the rise and early evolution of animal life is highlighted. Critical assessment of the trace-fossil record indicates that the earliest uncontroversial records of bilaterian activity are approximately 560–555 My old, and that the trace fossils present in strata hosting the so-called White Sea assemblage mostly reflect feeding styles sustained by microbial mats. By the terminal Ediacaran (Nama assemblage), an increase in complexity, predation pressures, sediment disturbance and penetration depth is evident. However, body-plan diversity as recorded by trace-fossil architectures was still limited. The persistence of an Ediacaran-style matground ecology together with the appearance of disparate, new architectural plans reflecting novel ways of interacting with the substrate is the hallmark of the Fortunian. In turn, Cambrian Ages 2–4 are marked by the establishment of Phanerozoic-style ecosystems, typified by an increased length and degree of connectivity of the food web, and a profound reorganization of the infaunal ecospace. These changes took place in both high-energy, sand-dominated nearshore areas and low-energy, mud-dominated offshore environments. A model comprising four evolutionary phases is proposed to account for these evolutionary changes. The integration of high-resolution ichnological and sedimentological information can offer ecological data to reconstruct local redox conditions, which in turn can be used in basin transects and provide solid support for models at global scale. The present review highlights the significant role played by bioturbation in shaping large-scale macroevolutionary changes in fully marine environments at the dawn of the Phanerozoic.

Data accessibility

Data are included in the paper and tables.

Authors' contributions

M.G.M. took the leading role in gathering data and manuscript writing, whereas L.A.B. assisted in both activities.

Competing interests

We declare we have no competing interests.

Funding

We gratefully acknowledge funds from the Natural Sciences and Engineering Research Council of Canada (NSERC Discovery Grants 311727-15 to M.G.M. and 311726-13 to L.A.B.).