Engineering the Cambrian explosion: the earliest bioturbators as ecosystem engineers

Abstract By applying modern biological criteria to trace fossil types and assessing burrow morphology, complexity, depth, potential burrow function and the likelihood of bioirrigation, we assign ecosystem engineering impact (EEI) values to the key ichnotaxa in the lowermost Cambrian (Fortunian). Surface traces such as Monomorphichnus have minimal impact on sediment properties and have very low EEI values; quasi-infaunal traces of organisms that were surficial modifiers or biodiffusors, such as Planolites, have moderate EEI values; and deeper infaunal, gallery biodiffusive or upward-conveying/downward-conveying traces, such as Teichichnus and Gyrolithes, have the highest EEI values. The key Cambrian ichnotaxon Treptichnus pedum has a moderate to high EEI value, depending on its functional interpretation. Most of the major functional groups of modern bioturbators are found to have evolved during the earliest Cambrian, including burrow types that are highly likely to have been bioirrigated. In fine-grained (or microbially bound) sedimentary environments, trace-makers of bioirrigated burrows would have had a particularly significant impact, generating advective fluid flow within the sediment for the first time, in marked contrast with the otherwise diffusive porewater systems of the Proterozoic. This innovation is likely to have created significant ecospace and engineered fundamentally new infaunal environments for macrobiotic and microbiotic organisms alike.

The development of infaunal bioturbation is one of the key Phanerozoic innovations and its first appearance is used to define the Ediacaran-Cambrian boundary (Brasier et al. 1994). The evolution of burrowing, and especially bioirrigation, by macrobiotic organisms is a key aspect of the ecological changes seen at the base of the Cambrian. It is likely to have been a fundamental evolutionary stimulus, since the seafloor and its nutrient flux underwent fundamental changes from the matground-dominated Proterozoic -which had only meiofaunal bioturbation/mixing -to the Phanerozoic style of infaunal bioturbation (Brasier & McIlroy 1998;McIlroy & Logan 1999;Mazurek 2013;Tarhan et al. 2015). In this study, we apply functional bioturbation analysis (cf. Solan & Wigham 2005;Bambach et al. 2007) to this turning point in Earth history, which will allow the assessment of the relative importance of different trace fossil types as ecosystem engineers.

Bioturbators as ecosystem engineers in modern ecosystems
Ecosystem engineering is the construction, modification and maintenance of environments by organisms, which may be classified as being either autogenic or allogenic ecosystem engineers (Jones et al. 1994). All ecosystem engineers modify the flow of resources within their environment. Autogenic ecosystem engineers, such as reefbuilding corals, provide a physical structure for other taxa to utilize (Jones et al. 1994;Wood 2016). Allogenic ecosystem engineers create new habitats and resource flows by redistributing materials from their environment, such as dam-building beavers (Pollock et al. 1995) and mound-forming termites (Dangerfield et al. 1998), but also those bioturbating organisms that significantly affect the distribution of both particles and solutes. The structures created by ecosystem engineers commonly persist long after the engineer has either died or relocated (Jones et al. 1997).
In marine environments, the key form of allogenic ecosystem engineering is bioturbation, in which burrowing organisms create and modify sedimentary environments. The impact of bioturbation on benthic ecosystems is broadly two-fold, leading to: (1) the creation of new habitats through the modification of the two-dimensional sediment -water interface into a three-dimensional complex (McIlroy & Logan 1999), which additionally varies through time (e.g. Solan & Kennedy 2002); and (2) the physical redistribution of nutrients and particulate matter (Aller 1982;McIlroy & Logan 1999;Michaud et al. 2005Michaud et al. , 2006. Bioturbation brings about several key physicochemical changes to the sediment, including: alteration of the porewater chemistry (Aller 1982); increasing the depth of the redox potential discontinuity below the sediment-water interface from its Proterozoic position close to the sediment -water interface (McIlroy & Logan 1999); increased water content (Rhoads 1970); and modifications to the topography and roughness of the sediment -water interface (Orvain 2005). Marine bioturbation is therefore one of the most important forms of allogenic ecosystem engineering (Levinton 1995). All bioturbating organisms are ecosystem engineers to some extent, particularly when taken in the context of the ichnological expression of the Cambrian explosion (McIlroy & Brasier 2016), when there was a fundamental change from small, essentially interface, burrows to deeper, more pervasive burrowing. It is clear that different types of bioturbation have different effects on microbial processes and nutrient fluxes within the sediment and therefore on benthic ecosystems (McIlroy & Logan 1999;McIlroy & Brasier 2016). The impact of a burrowing taxon is greatest where its bioturbation produces conditions that contrast most significantly with the pre-existing environment, such as particle and solute redistribution, particularly through bioirrigation (Mermillod-Blondin & Rosenberg 2006). The basal Cambrian expansion of infaunal bioturbation was one of the most profound changes of environment caused by organisms in the history of life on Earth and this ichnological diversification is used to help define the base of the Phanerozoic (McIlroy & Brasier 2016;Geyer & Landing 2016).
(1) Epifaunal bioturbators, i.e. mobile organisms that live on the substrate and do not penetrate the sediment -water interface to any significant degree (e.g. pagurid arthropods), producing structures such as locomotion trails (cf. the trace fossil Rusophycus). As such, their ecological impact on modern sedimentary environments is minimal.
(2) Surficial modifiers, i.e. taxa that live in the uppermost layers of sediment, typically at depths of no more than 2 cm (e.g. modern clypeasteroid echinoderms; cf. Cambrian traces such as Gordia isp.). They have a low impact on sedimentary ecosystems because their activities are restricted to short distance particle redistribution at, or at very shallow depths beneath, the sediment -water interface (Solan & Wigham 2005). Rather than engineering new ecosystems, the behaviour of some surficial modifiers may actually lead to the exclusion of other, potentially more significant, ecosystem engineers through predation on larval forms and sediment disturbance that precludes larval settling (DeWitt & Levinton 1985;Snelgrove 1999), although the disaggregation of surficial microbial mats may introduce mucus into the sediment, which significantly affects microbial productivity (McIlroy & Logan 1999 (Francois et al. 1997). As such, they are more significant ecosystem engineers than surficial modifiers. (4) Regenerators are taxa such as fiddler crabs that excavate semi-permanent burrows in the sediment, the removed sediment being transported back to the sediment -water interface, and can therefore be responsible for relatively large particle fluxes and the generation of significant amounts of new ecospace, which is essentially a vertical extension of the sediment -water interface (cf. the trace fossils Gyrolithes and Skolithos). (5) Upward conveyors are organisms that occupy vertical burrows in a head-down orientation, transferring sediment actively or passively toward the surface, often by deposit feeding at depth (e.g. arenicolid polychaetes). Larger organisms in deeper burrows can transport significant volumes of buried sediment (Treptichnus is a possible trace fossil analogue to modern upward conveyors).
(6) Downward conveyors (e.g. capitellid polychaetes) also live in vertical burrows, but with their head at or close to the surface, and move sediment down into the subsurface, either by ingestion and excretion, or by passive movement under gravity within the burrow. Some taxa may exhibit both upward-and downward-conveying behaviour at different times. Downward conveyors can transfer nutrients to depths of many centimetres, creating sedimentary habitats that would otherwise not exist deep within the sediment (e.g. Herringshaw et al. 2010;Kristensen et al. 2012; trace fossil examples include Halopoa, Asterosoma and similar mud-packed taxa). (7) Gallery biodiffusors, exemplified by nereidid polychaetes (Francois et al. 2002), build complex burrow systems within the sediment, with galleries connected to the sediment -water interface by one or more openings. Essentially, this behaviour combines short distance biodiffusive particle transport around the burrow walls with the rapid, larger scale flow of nutrients and sediment throughout the burrow network. Intense burrowing and bioirrigating activity within such galleries means that the trace-makers can potentially engineer significant volumes of new ecospace (Multina is a common Cambrian trace fossil exhibiting this type of behaviour).
The most effective ecosystem engineers are those bioturbators that move substantial quantities of particulate material through large and/or deep burrow systems (i.e. upward and downward conveyors and gallery biodiffusors; Fig. 1). Taxa that utilize one or more of these bioturbation modes over considerable periods of time, such as polychaete annelids and thalassinid crustaceans ( Fig. 1 In addition to bioturbation type, the impact of a burrowing organism on the sedimentary environment is greatly increased if it bioirrigates its burrow. Bioirrigation is used by most macroscopic burrowing organisms to oxygenate the burrow system and create currents that eject metabolites and faecal material from the burrow (Kristensen & Kostka 2005). The net effect of bioirrigation is to elevate nutrient levels and microbial productivity in and around the burrow; it can also change the permeability and grain size profiles of the substrate (Herringshaw & McIlroy 2013). It has the greatest ecological impact in low permeability substrates where the bioirrigated burrows create advective flow pathways in otherwise diffusive porewater systems (Mermillod-Blondin & Rosenberg 2006).

Ichnology of the Ediacaran -Cambrian transition
Surface traces from the Conception Group (565 Ma) of Newfoundland are the first reliable evidence for macrofaunal animal locomotion (Liu et al. 2010a(Liu et al. , b, 2014. Simple surficial grazing and feeding trails (e.g. Helminthorhaphe and Archaeonassa) and slightly more penetrative horizontal burrows (e.g. Planolites) become abundant in the late Ediacaran (e.g. Jensen 2003;McIlroy & Brasier 2016). The most common Ediacaran trace fossils are tiny horizontal traces, often attributed to burrowing within microbial mats, and commonly termed 'undermat miners' (Seilacher 1999), although this is something of a misnomer as many such trace fossils show no evidence of systematic mining behaviour and their mode of feeding is not based on any objective evidence.
Eustatic sea-level rise in the late Ediacaran and early Cambrian led to an increase in the surface area occupied by shallow marine environments (e.g. Brasier 1992;Peters & Gaines 2012). This transgression also led to decreased sediment flux to the continental shelf, resulting in an increase in the ratio of organic to inorganic material in shelf sediments (McIlroy & Logan 1999). The build-up of organic nutrient reserves beneath the sedimentwater interface has been proposed as a major trigger for the development of bioturbation across this interval (McIlroy & Logan 1999). It has been noted that trace fossils appear in a broadly consistent order in Ediacaran-Cambrian successions (Gehling et al. 2001;McIlroy & Brasier 2016). This rapid increase in both bioturbation intensity and depth of bioturbation across the Ediacaran -Cambrian transition is sometimes known as the 'agronomic' or 'substrate revolution' (Seilacher & Pflüger 1994;Bottjer et al. 2000). The increase in depth of bioturbation coupled with the widespread expansion of the palaeobathymetric range of exploited ecospace (from shallow-to deeper-marine environments), represents the first example of large-scale manipulation of the sedimentary environment by macroscopic animals (e.g. Marenco & Bottjer 2008;Dufour & McIlroy 2016).
Despite the increase in ichnodiversity in the Cambrian portion of the Chapel Island Formation, both the intensity of bioturbation and trace fossil size are generally low throughout the succession (McIlroy & Logan 1999). Most trace fossils are of small diameter and there is a lack of burrows deeper than a few centimetres, even in sedimentary facies where deeper bioturbation might be expected (Crimes & Anderson 1985). The H. podolica Zone has an ichnofabric index (ii) of 0-2 at the lamina scale and burrow depths and diameters ,10 mm (Fig. 2). The T. pedum ichnozone has a maximum burrow diameter of 30 mm (typically c. 10 mm), a maximum tiering depth of 20 mm and a typical ii of 2 ( Fig. 2; McIlroy & Logan 1999). In the R. avalonensis ichnozone, the maximum burrow depth recorded is 70 mm, but is typically 20 -30 mm; the maximum burrow diameter is 30 mm (typically c. 10 mm) and the ii value is generally 2, only occasionally rising to 3 or 4 at the bed scale ( Fig. 2; McIlroy & Logan 1999). Recent work by Tarhan et al. (2015) argues that low levels and depths of bioturbation persisted in siliciclastic marine environments from the Early Cambrian to the Silurian, although the intensity of bioturbation is clearly a function of the relative rates of bioturbation v. sedimentation (McIlroy 2004(McIlroy , 2008. The same researchers recognized the evolution of many infaunal behavioural types in the Early Cambrian, but suggest that major sediment mixing by 'bulldozing' infauna did not evolve until the mid-late Palaeozoic (Tarhan et al. 2015).

Methodology
This work documents the distribution of trace fossils through the late Neoproterozoic to pre-trilobitic Lower Cambrian (Terreneuvian) Chapel Island Formation of Fortune Head, SE Newfoundland (Crimes & Anderson 1985;Narbonne et al. 1987). This stratigraphically continuous, ichnologically rich section (Fig. 2) is the Global Stratotype Section and Point (GSSP) for the Precambrian (Ediacaran) -Cambrian boundary, with the boundary itself being defined on the appearance of the trace fossil T. (Phycodes) pedum at the boundary between the H. podolica Zone and the T. pedum ichnozone (Brasier et al. 1994;Landing 1994;Geyer & Landing 2016;McIlroy & Brasier 2016). As the stratotype succession was deposited exclusively in a shelf setting, it is possible to examine evolutionary changes in trace fossil assemblages without the biases imparted by facies and palaeoenvironmental shifts (Landing et al. 1988). For these reasons, the Chapel Island Formation also represents an ideal case study to examine the role of ecosystem engineers during the Ediacaran -Cambrian transition.
We examined 11 of the key Ediacaran -Cambrian ichnotaxa from the Chapel Island Formation: Planolites, Gordia, Treptichnus, Monomorphichnus, Skolithos, Palaeophycus, Psammichnites, Helminthoidichnites, Teichichnus, cf. Thalassinoides and Gyrolithes (Table 1; Figs 3 & 4). To assess the potential impact of the trace-makers as ecosystem engineers, we assessed each ichnotaxon based on three criteria: the infaunal ecospace created by the ichnotaxon; burrow function(s); and the likelihood of bioirrigation behaviour. To code for the ecospace created, we assigned a numerical value to the depth of the trace or burrow (methodology modified from Bambach et al. 2007): (1) surface traces; (2) semi-infaunal traces (shallowly penetrative structures in which the trace-maker was partly infaunal and partly exposed to the water column); (3) shallow infaunal (,5 cm deep) burrows; and (4) deep infaunal (.5 cm deep) burrows. The likely behaviour of the trace-maker and the functional/ethological group or groups to which the trace-maker could have belonged were then assigned (method of Solan et al. 2004;Solan & Wigham 2005;after Francois et al. 1997after Francois et al. , 2002. Using this functional group classification, the inferred relative sedimentary particle and nutrient fluxes associated with each ichnotaxon could be assigned, with increasing values indicating increasing impact on particle transport and nutrient flux. The functional groups were assigned as follows: (1) epifaunal locomotion; (2) surficial modification; (3) biodiffusion; (4) regeneration; (5) downwardand/or upward-conveying; and (6) gallery biodiffusion. Where a number of functions were considered to be likely for an ichnotaxon, more than one value was assigned to the same taxon.
The likelihood that a trace was bioirrigated was given one of three values: (1) highly improbable (surficial locomotion or grazing traces for which irrigation is unlikely to have been necessary); (2) possible (shallow, simple burrows, possible dwelling structures, which might have required at least short-term irrigation); or (3) probable (deeper and/ or more complex burrow systems, probably dwelling structures, in which irrigation is likely to have been required for longer term occupation).
By then combining the ecospace, functional group and bioirrigation values, the relative ecosystem engineering impact (EEI) of each ichnotaxon can be estimated. A non-bioirrigated surface locomotion trail therefore has an EEI of 3 (1 + 1 + 1), whereas a deep infaunal network with gallery biodiffusion and bioirrigation has an EEI of 12 (3 + 6 + 3). For most ichnotaxa, where more than one function is possible, a range of EEI values was obtained. It should be noted that the value of the EEI obtained for a taxon does not imply position on a linear scale of impact, but is a means of classifying the impact of an organism on its environment.

Results
The results of the EEI analysis of key Ediacaran -Cambrian ichnotaxa are summarized in Table 1. In the H. podolica ichnozone, all ichnogenera presente.g. Planolites, Harlaniella, Gordia and Torrowangea -are bedding-parallel, shallow tier trace fossils with little vertical penetration of the sediment (cf. Marenco & Bottjer 2008). The biological affinity of Palaeopascichnus is currently unclear (e.g. Dong et al. 2008;Antcliffe et al. 2011), but it is unlikely to be a trace fossil and is therefore not considered here. Planolites and Gordia (along with Harlaniella and Torrowangea) are interpreted to represent biodiffusive bioturbation. Both Planolites and Gordia are essentially surficial traces, although Droser et al. (2002) interpret Planolites as having been shallowly penetrative, in which case Planolites trace-makers could also have been biodiffusors. Gordia most probably represents only surficial modification and it is therefore improbable that Gordia burrows were bioirrigated, but possible that Planolites traces were. As such, Gordia is given a maximum EEI value of 5 and Planolites a maximum EEI value of 8 (Table 1). The first occurrence of burrows such as Gordia and Planolites increased the surface area of the sediment -water interface (McIlroy & Logan 1999) and initiated the development of a sediment mixed layer (Savrda & Bottjer 1989;Bromley 1996;Droser et al. 2002Droser et al. , 2004Tarhan  In the succeeding T. pedum ichnozone (Fig. 2), a greater diversity of ichnotaxa and behaviours is encountered. Not only do horizontal trace fossils become more morphologically diverse, with the appearance of ichnotaxa such as Helminthoidichnites and Psammichnites (used in preference to Taphrhelminthopsis and Archeonassa, cf. Jensen 2003;Jensen & Palacios 2016), and the arthropod scratch marks of Monomorphichnus ( Fig. 3f; Table  1), but vertical burrows, such as Skolithos, also appeared for the first time. The dwelling trace Skolithos did not create significant volumes of ecospace and any bioturbation associated with these burrows is likely to have been essentially biodiffusive. Nonetheless, the depth and abundance of Skolithos burrows in some Early Cambrian successions -often forming 'pipe rock' ichnofabrics (McIlroy & Garton 2004) -indicates a fundamental change in infaunal habitats. A greater diversity of functional groups and the increased generation of ecospace is implied for the T. pedum ichnozone owing to the occurrence of more three-dimensional ichnogenera, most notably the branching, penetrative trace fossils Treptichnus and Gyrolithes, but also Arenicolites, which indicate the increasing behavioural complexity of infaunal organisms (Narbonne et al. 1987 and herein). The diagnostic trace fossil Treptichnus is widely taken as the first burrow that shows probable shallow tier gallery diffusion, although it has also been interpreted as the probing burrow of a carnivorous priapulid (Dzik 2005;Vannier et al. 2010). As such, its EEI value has a range of 7-12.
Although many new horizontal trace fossils appear in the Rusophycus avalonensis Zone, such as the geometrical graphoglyptid trace fossil Squamodictyon -which is considered to represent shallow gallery diffusion -along with additional vertical burrows, the functional groups of tracemakers in this interval appear to have been essentially the same as those of the T. pedum Zone. The most significant first appearance datums are those of Teichichnus, and possibly Diplocraterion, which allow the inference of the evolution of vertical and horizontal biodiffusion, upward-and downwardconveying and regeneration (Table 1; McIlroy & Logan 1999). Teichichnus has its first reliable occurrence at around the base of Cambrian Stage 2 (Landing et al. 1988;McIlroy & Brasier 2016), but purported Diplocraterion (Narbonne et al. 1987), which can be difficult to distinguish from cross-sections of Teichichnus, could not be confirmed in this study.

Discussion
Although the diagnostically Lower Cambrian ichnotaxon T. pedum has a relatively wide range of potential EEI values, depending on the interpretation of the behaviour of the trace-maker (Table 1), it might not have been the trace fossil with the highest ecosystem impact. Ichnotaxa such as Teichichnus have a narrower range of EEI values and their tracemakers can be interpreted more confidently as higher impact ecosystem engineers. Nonetheless, the effects of Treptichnus and similar trace fossils on the Cambrian seafloor should be seen in the context of change relative to the minimal ecosystem engineering potential of the latest Ediacaran trace fossil assemblages. When viewed in this manner, it is clear that T. pedum and similar trace fossils are likely to have had a profound effect on the marine ecosystem. Our assessment of the ichnology of the Chapel Island Formation suggests that most, if not all, of the major functional groups of modern bioturbators appear to have evolved by the base of Cambrian Stage 2 in the Global Stratotype Section -that is, within the first 12 million years of the Phanerozoic (corroborated by findings from northern Norway by McIlroy & Brasier 2016).
It is well known that matground facies persisted locally in marine environments into the Silurian (e.g. Tarhan et al. 2015), but they did so in increasingly restricted, commonly nutrient-starved sedimentary environments (Harazim et al. 2013;McMahon et al. 2016). This marginalization of the matground ecosystem to environments with ecologically harsh conditions, such as brackish and freshwater facies, reflects the effects of the Cambrian-style ecosystem engineering bioturbators that gradually shaped the marine realm toward its modern state.
As with any ichnological study that requires the functional interpretation of trace fossils, there is always some ambiguity as to the life activity of the trace-maker, indeed many modern burrows can reflect a diverse range of behaviours that can change due to ontogenetic or palaeoenvironmental controls as well as ecosystem engineering (Herringshaw et al. 2010). The accuracy/importance of the functional groups assigned to the ichnotaxa has significant potential bearing on their EEI value. We addressed this issue by assigning a potential range of ecosystem impacts to some trace fossils, which affects the EEI value (i.e. their inferred significance as ecosystem engineers). The magnitude of the impact that a burrowing organism has on its environment should be considered in the context of its stratigraphic level and whether the sedimentary environment it occurred in was diffusion-or advection-dominated.

Ecosystem engineering of fine-grained Cambrian sediments
Bioirrigating organisms that construct networks or galleries in fine-grained sediments (mud and silt) create a greater contrast to natural sedimentary flow conditions than those that burrow in sands (cf. Mermillod-Blondin & Rosenberg 2006;Meysman et al. 2006). The shallow, narrow, vertical tube Skolithos (Fig. 3a) and the similar, but spiralling, Gyrolithes (Fig. 3b), first encountered in the T. pedum ichnozone of the Early Cambrian (Fig. 2), are blindended subvertical burrows, whose trace-makers may well have bioirrigated their burrows. The evolution of such behaviour would have had a particularly profound impact on infaunal ecosystems, particularly in terms of microbial and meiofaunal activity in the near-burrow zone (cf. Reise 1981;McIlroy & Logan 1999). As such, morphologically comparable ichnotaxa found in early Phanerozoic mudstones and siltstones, which would have been diffusive sedimentary environments, are of greater significance than those found in sandstones, which are likely to have been advective.

Ecosystem engineering in coarse-grained Cambrian sediments
In coarser grained Lower Cambrian successions, from about the base of Cambrian Stage 3, the vertical tubular trace fossil Skolithos is commonly found in composite ichnofabrics, sometimes in very high densities (up to 7500 burrows/m 2 , Davies et al. 2009) that may have been due to very dense assemblages of organisms in the typical 'pipe rock' biotope (McIlroy & Garton 2004. High densities of burrowing organisms may have increased the sedimentary impacts of such simple burrows (cf. Miron et al. 1992;Volkenborn & Reise 2006;Herringshaw et al. 2010), but their potential for ecosystem engineering is likely to have been low because the sandy sediments in which they formed would already have had advective flow through them. The impact of Skolithos on sediments and nutrients is likely to have been restricted to upward-conveying, arguably during burrow excavation only, with associated biodiffusion and possibly downward-conveying during the period that the burrow was inhabited. The amensalism that might have resulted from dense assemblages of vertical burrow dwellers that is evident in the low diversity of pipe rock facies is difficult to prove owing to the palimpsesting of assemblages (McIlroy & Garton 2010).

Comparison of ichnological trends with other Lower Cambrian successions
The trends indicated by the ichnological data from the Chapel Island Formation are augmented by contemporaneous successions bearing trace fossils elsewhere. In western Mongolia, the Ediacaran -Cambrian transition is preserved in a mixed carbonate -siliciclastic succession with a trace fossil fauna showing similar patterns to those observed in Newfoundland (Goldring & Jensen 1996). Questionable short vertical and horizontal burrows are described from the latest Ediacaran Tsagaan Oloom Formation (Goldring & Jensen 1996;Khomentovsky & Gibsher 1996;Brasier et al. 1997), while the ichnogenera Didymaulichnus, Helminthoidichnites and Planolites were described from the earliest Cambrian Bayan Gol Formation. These simple, bedding-parallel ichnotaxa are morphologically and behaviourally comparable with those of the H. podolica assemblage Zone of Newfoundland. The overlying unit of the Bayan Gol Formation contains these taxa, as well as Cochlichnus, Homosiroidea (probably a taphomorph of T. pedum), Palaeophycus, Phycodes (?Trichophycus) and Treptichnus. As with the trace fossil assemblage of the T. pedum assemblage Zone of Newfoundland ( Fig. 2; Geyer & Landing 2016), the upper unit includes increasingly complex burrow morphologies. Trace fossils in the Mongolian successions therefore show progressively greater vertical depths of burrow penetration, and an increasing abundance of burrows that are inferred to have been created for shallow gallery diffusion, through the Lower Cambrian succession.
The Ediacaran -Cambrian succession of lower shoreface-offshore transition zone sediments in western Canada also show comparable ichnological patterns to those of the Chapel Island Formation (MacNaughton & Narbonne 1999;Carbone & Narbonne 2014), which might be elucidated further from this type of ecological assessment. In a review of Proterozoic -Cambrian ichnology, Jensen (2003) described the increase in morphological and behavioural complexity across the Ediacaran-Cambrian boundary, noting trends that are very similar to the specific patterns described herein from the GSSP. This stratigraphic trend toward both increasing diversity and behavioural complexity is consistently found in Ediacaran to Lower Cambrian siliciclastic sediments around the world (Gehling et al. 2001;Geyer & Landing 2016;McIlroy & Brasier 2016). Some Ediacaran to Cambrian sections have been considered to have Cambrian-type trace fossils slightly below the Cambrian boundary and overlapping with the Ediacaran biota (e.g. Jensen & Runnegar 2005; Darroch et al. 2015). The precise level at which trace fossils of Cambrian aspect appear is significant for global correlations, but it is perhaps to be expected that the event was somewhat diachronous. This does not detract from the fact that this emerging biota fundamentally changed the marine realm and that its development occurred globally within a geologically short interval of time.
In comparison with later Phanerozoic successions, one of the distinctive aspects of the Lower Cambrian is that most burrows were small and did not penetrate deeply into the sediment. Even though vertical bioturbation evolved in the late Ediacaran, it was not until the Cambro-Ordovician that significantly larger, deeper gallery networks such as Thalassinoides became prevalent (Sheehan & Schiefelbein 1984;Myrow 1995 (Myrow 1995).

Conclusion
In terms of the ecospace they created, the likely nutrient flux they generated and the probability that they bioirrigated the burrow systems, the tracemakers of Treptichnus would have been highimpact Early Cambrian ecosystem engineers. This is especially true when taken in the context of the probable ecological impact of Ediacaran trace fossils, which is likely to have been low. T. pedum is therefore unquestionably an important ichnotaxon for stratigraphic purposes because its first occurrence represents a major change in the seafloor ecosystems relative to those of the Ediacaran, but it is not the Early Cambrian trace fossil that is likely to have had the greatest EEI value.
The trace-makers of the diagnostically Cambrian deeper penetrative trace fossils Gyrolithes and the deposit-feeding burrow Teichichnus were potentially more significant ecosystem engineers than T. pedum. By comparison with modern burrows similar to these ichnotaxa (e.g. Dworschak & Rodrigues 1997;Gingras et al. 1999), we consider that both Gyrolithes and Teichichnus are likely to have been multifunctional, with the trace-maker occupying the same position for significant periods of time while bioirrigating the sediment. The spreiten burrow Teichichnus may have been particularly high impact because it is associated with the highest ichnofabric indices and deepest tiering in Early Cambrian successions (McIlroy & Logan 1999). The presence of spreiten in Teichichnus indicates bulk sediment processing behaviour during a relatively protracted period of burrow occupation. The trace-maker is likely to have bioirrigated the sediment -suggesting that the Teichichnus trace-maker was an engineer of new ecospace that caused significant biogenic particle flux. Its occurrence at about the base of Cambrian Stage 2, and its abundance in siliciclastic deposits of Cambrian age (Landing et al. 1988;McIlroy & Brasier 2016), make it a potentially useful marker for Cambrian Stage 2 in facies without small shelly fossils.
The trace fossil assemblage that marks the base of the Cambrian in the GSSP is herein demonstrated to result from the actions of animals that interacted with the seafloor much more significantly than their Ediacaran counterparts. Nonetheless, there are aspects of the sedimentary facies that need to be isolated from those of stratigraphy and evolution. It appears that, close to the boundary level, the first pervasive bioturbators evolved, with composite ichnofabrics of cf. Gyrolithes being common in the offshore facies and shallow tier Treptichnus-dominated assemblages occurring in association with fair weather mudstones in highenergy, lower shoreface facies, both in the GSSP (herein) and in Norway (McIlroy & Brasier 2016). The changes in microbial nutrient cycling that are likely to have accompanied the first bioturbators (McIlroy & Logan 1999) provide evidence to suggest that complex animals did indeed engineer the Cambrian explosion, irrevocably changing the biogeochemistry of sediments, porewaters and oceans alike (Brasier & McIlroy 1998;McIlroy & Logan 1999;Canfield & Farquhar 2009). Our work further refines that conclusion and identifies the basal Cambrian ichnogenera/ecosystem engineers that are likely to have had the most impact at this time. Martin Brasier, whose help and guidance stimulated and encouraged our interest in the ichnology of the Ediacaran-Cambrian of Newfoundland, is remembered with fondness and appreciation by us all. Support for this research came from NERC and an NSERC discovery grant to DMcI. This manuscript benefited greatly from the thoughtful and constructive reviews of Martin Solan and a second, anonymous, reviewer.