Opinion

Ediacaran Extinction and CambrianExplosion

Simon A.F. Darroch,1,* Emily F. Smith,2 Marc Laflamme,3 and Douglas H. Erwin4

HighlightsWe provide evidence for a two-phasedbiotic turnover event during theEdiacaran–Cambrian transition (about550–539 Ma), which both comprisesthe Earth’s first major biotic crisis ofmacroscopic eukaryotic life (the disap-pearance of the enigmatic ‘Ediacarabiota’) and immediately precedes theCambrian explosion.

Wesummarizetwocompetingmodelsforthe turnover pulses – an abiotically drivenmodel(catastrophe)analogoustothe ‘Big5’ Phanerozoic mass extinction events,and a biotically driven model (biotic repla-cement) suggesting that the evolution ofbilaterian metazoans and ecosystemengineering were responsible.

We summarize the evidence in supportof both models and identify several keyresearch questions which will help todistinguish between them, and thusshed light on the origins of the modern,animal-dominated biosphere.

We argue that the first turnover pulse(about 550 Ma) marks the greatestchange in organismal and ecologicalcomplexity, leading to a more recog-nizably metazoan global fauna that hasmuch more in common with the Cam-brian than the earlier Ediacaran. Thislatest Ediacaran interval (the ‘Nama’)may therefore represent the earliestphase of the Cambrian Explosion.

1Vanderbilt University, Nashville, TN37235-1805, USA2Johns Hopkins University, Baltimore,MD 21218-2683, USA3University of Toronto Mississauga,Mississauga, ON L5L 1C6, Canada4Smithsonian Institution, PO Box37012, MRC 121, Washington, DC20013-7012, USA

*Correspondence:simon.a.darroch@vanderbilt.edu(Simon A.F. Darroch).

The Ediacaran–Cambrian (E–C) transition marks the most important geobio-logical revolution of the past billion years, including the Earth’s first crisis ofmacroscopic eukaryotic life, and its most spectacular evolutionary diversifica-tion. Here, we describe competing models for late Ediacaran extinction,summarize evidence for these models, and outline key questions which willdrive research on this interval. We argue that the paleontological data suggesttwo pulses of extinction – one at the White Sea–Nama transition, which ushersin a recognizably metazoan fauna (the ‘Wormworld’), and a second pulse at theE–C boundary itself. We argue that this latest Ediacaran fauna has more incommon with the Cambrian than the earlier Ediacaran, and thus may representthe earliest phase of the Cambrian Explosion.

Evolutionary and Geobiological Revolution in the EdiacaranThe late Neoproterozoic Ediacara biota (about 570–539? Ma) are an enigmatic group of soft-bodied organisms that represent the first radiation of large, structurally complex multicellulareukaryotes. Although many of these organisms may represent animals (Metazoa), few shareany synapomorphies with extant metazoan clades and thus may represent extinct groups withno modern representatives [1] (Figure 1). Consequently, the position of the Ediacara biota in thecontext of the late Neoproterozoic–Cambrian rise of animals remains poorly understood. Afternearly 30 million years (my) as the dominant components of benthic marine ecosystems, theEdiacara biota disappeared in the first major biotic transition of complex life, succeeded by aspectacular biotic radiation – the ‘Cambrian Explosion’ – which marks the appearance of amajority of extant animal clades and the establishment of metazoan-dominated ecosystems[2,3]. Understanding the cause(s) behind the disappearance of the Ediacara biota and itsrelationship to the subsequent Cambrian radiation is thus key to understanding the origins ofthe modern biosphere, and is an important research goal in geology, paleontology, andevolutionary biology. However, the mechanisms behind this transition are poorly understood,with recent research providing support for different models. One model posits an abiotic driversimilar to the causes of the Phanerozoic ‘Big 5’ mass extinctions (hereafter termed the‘catastrophe model’), while a separate – but not necessarily mutually exclusive – scenarioinvokes a biotic driver linked to the radiation of metazoans with new behaviors and feedingmodes (the ‘biotic replacement’ model). An alternative model suggesting that the apparentextinction is an artifact of a closing preservational window has been rejected, on the basis thatEdiacaran-style matgrounds (and thus the conditions necessary for fossilization) persist into theCambrian [4]. The catastrophe model would place the Cambrian Explosion as a postextinctiondiversification, similar in form if greatly different in result, to the early Cenozoic diversifications ofbirds and placental mammals. By contrast, a biotic replacement model potentially creates amore gradual transition from the latest Ediacaran faunas into the earliest Cambrian. Here, wesynthesize recent work on this interval and identify a series of fundamental research goals.Answering these questions will shed light on the sequence of events leading to the origin of

Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9 https://doi.org/10.1016/j.tree.2018.06.003 653© 2018 Elsevier Ltd. All rights reserved.

(A) (B)

(E) (F)

(G)

(K)(J)

(L)

(M)

(I)

(H)

(C) (D)

Figure 1.

(Figure legend continued on the bottom of the next page.)

Enigmatic Fossils from the Ediacaran. (A–D) Soft-bodied Ediacara biota belonging to the ‘White Sea’ assemblage, and which disappear prior to theonset of the late Ediacaran ‘Nama’ interval about 550 Ma (A, Andiva; B, Dickinsonia; C, Yorgia; D, Tribrachidium). (E–H) Soft-bodied erniettomorph and rangeomorphEdiacara biota belonging to the relatively species-poor ‘Nama’ assemblage (E, Swartpuntia; F, Rangea; G, Ernietta; H, Pteridinium). (I–M) Characteristic metazoan

654 Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9

modern-looking ecosystems, as well as add valuable new information on the processes thathave driven both mass extinctions and evolutionary radiations over the last about 550 my.

Ediacaran–Cambrian BioeventsThe Ediacaran [635–539? million years ago (Mya)] marks the transition from Proterozoic toPhanerozoic, and is the most recently named geological period. The base of the Ediacaran isdefined by a Global Boundary Stratotype Section and Point (GSSP) in South Australia marking theend of the Marinoan Snowball Earth glaciation [5], while its top (and base of the Cambrian) isintended to coincide with the first appearance of the trace fossil Treptichnus pedum in southeastNewfoundland [6,7]. Fossils of soft-bodied Ediacara biota first appear in the Avalon Peninsula(Newfoundland) in sequences that overly diamictite recording the Gaskiers glaciation. Recentdates from the Avalon and Bonavista peninsulas give a maximum age of 570.94 � 0.38 Ma for thefirst appearance of Ediacaran fossils [8]. New ashes that have been dated at the onset and at thenadir of the basal Cambrian d13C excursion in South China constrain this secondary chronostrati-graphic marker of the Ediacaran–Cambrian (E–C) boundary to about 539 Ma, consistent with re-dated ashes from southern Namibia [9]. Within this interval, three temporally and faunally distinc-tive assemblages have been recognized: the ‘Avalon’, ‘White Sea’, and ‘Nama’ [10,11] (Figure 2).The White Sea assemblage (about 558–550 Ma) marks the apex of Ediacaran diversity, with manyiconic groups (sometimes clades, see [12]), such as the Dickinsonimorpha, Bilateromorpha, andTriradialomorpha, well represented. By contrast, the youngest Ediacaran assemblage – the Nama(550–539? Ma) – contains a relatively species-poor community overwhelmingly dominated by theRangeomorpha and Erniettomorpha, with many other Ediacaran groups apparently alreadyextinct [13,14]. The Nama assemblage also records a proliferation of organic-walled tubularorganisms, a diversification in bilaterian trace fossils [15], and the appearance of the calcifyinginvertebrates Cloudina, Namacalathus, and Namapoikia, marking a fundamental shift to commu-nities with a much higher proportion of metazoans [16–18]. This proliferation of vermiformmetazoans has led to the latest Ediacaran Nama assemblage being referred to as ‘Wormworld’and may indicate an escalation in metazoan ecosystem engineering and the evolution of theEltonian pyramid (i.e., the appearance of primary and secondary metazoan consumers) [19]. Witha few rare (and contested) exceptions, the Ediacara biota disappear entirely at the Cambrianboundary, along with the majority of the ‘Wormworld’ fauna [20–23].

We argue that the biostratigraphic evidence points to two episodes of biotic turnover andextinction: one at about 550 Ma at the transition between the White Sea and Nama assemb-lages, and a second at the E–C boundary itself. This second event was closely followed by theCambrian Explosion, marked by the appearance of a diverse array of metazoan groups, and themost dramatic interval of metazoan morphologic innovation in the history of life [24].

Ediacaran–Cambrian Record of Environmental ChangeOne of the great challenges of the E–C transition is understanding the possible causalconnections between the dramatic biotic shifts and coeval environmental changes. One featureof Phanerozoic mass extinctions is that they are frequently accompanied by negative carbonisotope excursions, which record perturbations to global marine oceans. Rather than exploringthe list of possible environmental triggers for the Cambrian radiation, we focus our discussionhere on the two large, possibly global perturbations to the carbon cycle at the close of theProterozoic: the Shuram and the Basal Cambrian Carbon Isotope Excursion (the ‘BACE’).

‘Wormworld’ body fossils and trace fossils that existed alongside Ediacara biota in the Nama assemblage, representing the onset of biomineralization, diversification offeeding behaviors, and intensification of animal ecosystem engineering (I, Gaojiashania; J, Namacalathus; L, unnamed trace fossil; K, Shaanxilithes; M, large,treptichnid-like traces from the late Ediacaran of Namibia). Filled scale bars 1 cm, open 5 mm.

Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9 655

Global diversity ‘Ediacara biota’Carbon isotope

(δ13C) recordMetazoans and

‘ecosystem engineering’

Calcifyingmetazoans

Organic-walledmetazoans

(’Wormworld’fauna)

Exnc

ons

‘Wormworld’

Arbo

reom

orph

a

Rang

eom

orph

a

Erni

eom

orph

a

Pent

arad

ialo

mor

pha

Tetr

arad

ialo

mor

pha

Dick

inso

nim

orph

a

Bila

tero

mor

pha

Trira

dial

omor

pha

0 0

0

20 20

20

40 40

40

60 80 100 120529

533

538

543

548

553

558

563

568

529

533

538

543

548

553

558

563

568

5710 0 0 0 0 0 0 0 0 02220 40 60 80 100 120

Cam

bria

nEd

iaca

ran

Cam

bria

nEd

iaca

ran

Fort

unia

nN

ama

Whi

te S

eaAv

alon

Fort

unia

nN

ama

Whi

te S

eaAv

alon

# Genera # Genera4 4 4 4 4 4 4 4 4 4 4 -12 -6 +648 8

δ13C (‰VPDB)

?

?

?

Tim

e (m

ya)

Plausible metazoans

Ediacara biota

? ‘BACE’

‘Shu

ram

’

Clou

dina

Nam

acal

athu

s

Shaa

nxili

thes

Wut

ubus

Sekw

itubu

lus

Gaoj

iash

ania

Cono

tubu

s

‘Com

plex

’ bur

row

ing

Mac

rosc

opic

pre

datio

n?

?

# Bilaterian ichnogenera

Figure 2. Ediacaran–Cambrian Timescale. This illustrates the stratigraphic distribution and gamma diversity of enigmatic Ediacaran groups (Ediacara biota), themore recognizably metazoan ‘Wormworld’ fauna (both ‘calcifying’ and ‘organic-walled’ metazoans), diversity of bilaterian ichnogenera, and the onset of putative‘ecosystem engineering’ effects, alongside the record of environmental perturbation (as compilation d13C curve). Diversity of Ediacara biota and bilaterian ichnogenerareplotted from [11] and [15], respectively. d13C data compiled by [26] and [58]. BACE, Basal Cambrian Carbon Isotope Excursion; mya, million years ago; VPDB, ViennaPee Dee Belemnite standard.

The Shuram event is a large negative carbon isotope excursion that reaches values as low as�12m, much lower than the modern mantle carbon value of �5m [25]. This excursion iscommonly thought to end at about 551 Ma [26], however, there are poor radiometric ageconstraints for its onset, duration, and return to baseline values; better age constraints arecritical for demonstrating its global synchroneity [27]. Diagenesis has been proposed as apossible explanation for the extreme negative d13C values (e.g., [28]). More recently, however,Ca and Mg isotopic data have been used to argue for a primary nature of the Shuram [29]. Whatremains to be determined is whether or not the Shuram d13C values recorded isotopic values ofglobal dissolved inorganic carbon (DIC) at the time of deposition or, as has been suggested forother large Proterozoic excursions, a local phenomenon [30].

The BACE, a negative d13C excursion that accompanies the biotic shift that occurs across theE–C boundary [31], has better radiometric constraints than the Shuram, but is similarly lacking amechanistic explanation. As with the Shuram, it is not clear whether this signal is primary ordiagenetic, local or global in nature.

Competing Models for End-Ediacaran TransitionSeveral alternative scenarios have been posited for the shift between the Ediacara and earliestCambrian biotas. An abiotic driver of extinction is invoked by the ‘catastrophe’ model,potentially represented by the Shuram and/or the BACE. If globally synchronous, theseexcursions represent major perturbations to the global carbon cycle, and could reflect a varietyof possible kill mechanisms. The ‘biotic replacement’ model, by contrast, suggests that bioticmodification of substrate and other environments by emerging metazoans may have beenresponsible. Ecosystem engineering encompasses physical, ecological, and chemical modifi-cation of their environment by organisms, which may create, modify, or destroy habitats for

656 Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9

other organisms [33,34]. Unlike much of the Ediacara biota, which were dominantly sessile andused only a limited variety of feeding modes [21], many extant metazoan groups have importantecosystem engineering and trophic impacts. Examples include the ability to vertically mix thesediment column (thus aerating the subsurface and creating new habitat), actively ‘pumping’water to aid in filter feeding (removing suspended organic matter and ventilating the watercolumn), and predatory feeding modes, which can precipitate predator–prey arms races, oralternatively eliminate prey species that are unable to adapt [19]. Although there is no directevidence for the early bilaterians recorded in the Nama ‘Wormworld’ outcompeting Ediacarabiota in the Ediacaran [35], this model suggests that the activities of evolving metazoans weredeleterious to soft-bodied Ediacara biota and led to their disappearance prior to the Cambrian.

These differences in extinction drivers between competing models – ‘catastrophe’ and ‘bioticreplacement’ – can be illustrated using the predicted trajectories of both soft-bodied Ediacarabiota and the burgeoning Cambrian-style metazoan fauna (Figure 3). ‘Biotic replacement’predicts that depauperate latest Ediacaran ecosystems reflect a fauna that was beingprogressively stressed, marginalized, and replaced, while also showing increased evidencefor diversification and ecosystem engineering by new metazoan lineages. By contrast,‘catastrophe’ would predict that Ediacaran ecosystems should remain relatively high diversityand ‘unstressed’ until the E–C boundary, at which point environmental disturbance triggersextinction in both Ediacara biota and metazoans alike. These scenarios are end members of acontinuum, with additional permutations possible: the simple ‘catastrophe’ model does notaccount for the two apparent episodes of extinction, and sequential environmental events(pulsed catastrophe) may have driven biotic turnover at the start and end of the Nama interval(represented by the Shuram and BACE events, respectively).

The differences in extinction models can also (potentially) be distinguished on the basis of rateand timing of biotic turnover. Of the ‘big 5’ Phanerozoic events known to have been caused by‘catastrophe’-type events – for example, the Permian–Triassic (P–T) and Cretaceous–Paleo-gene (K–Pg) – the rate of ecosystems collapse is fast on geological timescales (virtuallyinstantaneous in the K–Pg, and about 103–104 years in the case of the P–T, see [36]).Consequently, a ‘catastrophe’-type event in the Late Ediacaran might be expected to resultin geologically rapid turnover. ‘Biotic replacement’, by contrast, relies on the appearance ofnew behaviors and traits in multiple metazoan clades, and therefore might be expected to resultin more protracted biotic turnover on evolutionary timescales (105–106 years). However, thislatter point is more speculative, because unlike ‘catastrophe’, large-scale ‘biotic replacement’has no direct analog in the Phanerozoic. Although some authors [14] have suggested that theongoing diversity crisis (the ‘6th mass extinction’) might represent a similar process, the extentto which current extinction rates are comparable with mass extinctions in the geological past isuncertain [37], and thus we argue that ‘fast’ and ‘slow’ are useful predictions for each model.

In the following sections, we outline several key questions that will help discriminate betweenthese models, resolve the cause(s) of the E–C transition, and shed new light on the drivers of(and links between) mass extinctions and subsequent diversification.

Key Questions and Research Priorities for the Ediacaran–CambrianWhat, and When, Was the Shuram?Can the large environmental perturbations represented by the Shuram be stratigraphically andradiometrically linked to the biotic shift, and possible extinction event, that happens across theWhite Sea–Nama transition? This interval is thought to coincide with the Shuram excursion [38],but radiometric constraints on the Shuram are poor, with some estimates placing the start of

Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9 657

1.) ‘Catastrophe’

2.) ‘Pulsed catastrophe’

3.) ‘Bio c replacement’

Ediacara biota

Ediacara biota

Ediacara biota

Metazoan‘ecosystem engineers’

Metazoan‘ecosystem engineers’

Metazoan‘ecosystem engineers’

Comm

unity metric

(abundance/evenness/diversity)

White Sea NamaEdiacaran Cambrian

550 530Age (Ma)

Figure 3. Modeled Trajectories of Different Biological Groups (Ediacara Biota and Cambrian-StyleMetazoan ‘Ecosystem Engineers’) and Predicted Fossil Community Responses (Community Metric), underThree Proposed Extinction Scenarios. Figure modified from [3].

the excursion at about 560–550 Ma [26,39], while others suggesting that Shuram-like excur-sions may have begun as early as about 609 Ma [40], predating the Gaskiers glaciation. Insections where the Shuram and Ediacaran fossils are both present, the excursion is typicallyfinished before White Sea assemblage fossils appear [41]. In addition, some studies suggestthat the Shuram may have been a protracted event lasting up to about 50 my [40]. With itsextreme nadir and unusually long duration, the Shuram requires either an Ediacaran carboncycle fundamentally different from the modern one (e.g., [25]), a global diagenetic event [42], orthe product of authigenic carbonate formation [43,44]. Adequately testing this facet of the‘catastrophe’ hypothesis requires (i) establishing radiometric ages on the onset and terminationof the Shuram; (ii) demonstrating that the Shuram is synchronous globally; and (iii) linking theseenvironmental changes with the fossil record (see Outstanding Questions).

What, and When, Was the BACE?Until it has been demonstrated that it is a synchronous global chemostratigraphic marker, it isnot clear whether the d13C chemostratigraphic profiles that have been documented acrossseveral E–C boundary sections record coeval changes in DIC (e.g., [30]). Although the BACEhas better radiometric age constraints than the Shuram excursion [9,20,45], it has not beenprecisely dated radiometrically in more than one region. If it is established that BACE records aglobal signature of DIC, follow-up questions about its driving mechanism and causal relation-ship to biotic turnover can be addressed, including: (i) What is the extent and pace of bioticturnover? Does abrupt biotic turnover coincide with an environmental perturbation[20,22,46,47], or is the biotic transition gradual and uncorrelated to environmental changes?

658 Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9

[31]; (ii) What is the mechanism driving the BACE? Only with absolute ages can the carbonfluxes, reservoir sizes, and oxidant demands needed to drive the BACE and other large E–Cexcursions be rigorously constrained (e.g., [48]).

Can We Disentangle Correlation versus Causation in the Late Ediacaran Extinction Event?Establishing proximal extinction drivers is a major challenge in deciphering the causes of bioticturnover in the late Ediacaran. In terms of ‘catastrophe’, establishing a temporal correlationbetween isotope excursions and biotic impact is crucial, however, this says little aboutcausation. Taking the P–T mass extinction (251 Ma) as an example, a negative d13C changeat the end-Permian has been convincingly linked to volcanism associated with the SiberianTraps; this in turn drove global warming, ocean acidification, elevated carbonate saturationstates, and primary productivity collapse, all of which would have devastated the biosphere andcomprised proximal extinction drivers [36,49]. However, establishing similar causal connec-tions in the late Ediacaran is far more challenging, not least from the difficulty of finding sectionswhere it is possible to link geochemical proxies with the paleontological record. Considerabledebate persists over the mechanism causing the Shuram excursion, and similar problems areassociated with the source of the BACE [22]. This uncertainty in the ultimate source ofperturbation to the late Ediacaran carbon cycle renders any discussion of proximal extinctiondrivers associated with them speculative.

Inferring extinction drivers associated with ‘biotic replacement’ is no easier. Although thediversification of metazoan trace and body fossils in the Nama indicates an escalation inmetazoan ecosystem engineering, how this may have affected the soft-bodied Ediacara biotahas not been established. Suggestions that expansion of grazing and burrowing in metazoansdisrupted microbial matgrounds that served as food for some Ediacaran organisms have beenlargely rejected, since similar substrates persisted into the Cambrian [4]. Given the discovery ofdrill holes in biomineralizing Cloudina [50] it is tempting to suggest the evolution of macroscopicmetazoan predation as a plausible extinction driver, however, no fossil evidence has beenfound for metazoans preying directly on soft-bodied Ediacarans (in fact, slabs preserving bothEdiacara biota and bilaterian trace fossils in the Nama are rare, suggesting a degree of nichepartitioning [35]). Although there is some limited evidence for large populations of metazoanpassive predators living in the same communities as Ediacara biota (and thus possiblyrepresenting a significant ecological pressure upon Ediacaran larvae in the water column –

see [35]), these sorts of biotic interactions are hard to demonstrate convincingly.

In sum, although demonstrating a close temporal correlation between isotope excursions andglobal faunal turnover may be enough to support the ‘catastrophe’ model, support for ‘bioticreplacement’ will hinge on demonstrating a plausible extinction driver stemming from earlyanimal activity. Interrogating the ‘biotic replacement’ model thus necessitates a deeper under-standing of Ediacaran paleoecology, in particular establishing how the Ediacara biota inter-acted with the emerging metazoan fauna. Interestingly, examining the survivors of the firstextinction pulse may be a key to understanding the proximal extinction drivers associated witheither the ‘catastrophe’ or ‘biotic replacement’ models. Although many Ediacaran groupsapparently perished during the White Sea–Nama transition, the Rangeomorpha and Ernietto-morpha survived; representatives of these groups have been found in Nama-aged assemb-lages on three paleocontinents, suggesting that they are some of the few Ediacaran organismsto have persisted into the metazoan-dominated ‘Wormworld’. These two clades may havebeen ecological generalists (in terms of habitat preferences and/or modes of nutrient acquisi-tion) [14], or may have possessed characteristics that made them resistant to the stressesassociated with the extinction pulse (e.g., tolerance to low oxygen). Either way, establishing

Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9 659

how and why these groups survived will offer valuable clues as to why many others did not. Newtechniques involving computational fluid dynamics [51,52] provide a powerful new way ofexamining Ediacaran paleobiology (including determination of feeding mode and mobility), andso offer one potential way forward to testing these hypotheses.

What Role Did End-Ediacaran Extinction Play in the Cambrian Explosion?Several of the Phanerozoic ‘Big 5’ mass extinctions were followed by the origin of new clades,expansions of morphologic disparity, and rapid diversification of new taxa. For example,bivalves show transient increases in origination rate after the three most recent mass extinctionevents [53]. Given that the Cambrian Explosion remains one of the most rapid and dramaticbiotic radiations in Earth history, there has been speculation that this event may have beentriggered by environmental perturbation across the E–C boundary (e.g., [46]). However,demonstrating this has remained elusive; small shelly faunas show a relatively steady patternof diversification across the E–C transition [32], and there is little evidence for a postmassextinction ecological vacuum or ‘disaster fauna’ in the earliest Cambrian. In addition, the overallnotion of ecological opportunity as a driver of biotic radiations is being re-evaluated, with therecognition that most are associated with macroevolutionary ‘lags’ between the first appear-ance of new organismal traits and their eventual ecological dominance. This suggests thatmorphological novelty often arises long before the environmental or ecological conditions thatmight guarantee their success [54]. The discovery that many complex bilaterian behaviors thatwere initially thought to be restricted to the Cambrian actually arose in the late Ediacaran about548 Ma (albeit in a primitive/preliminary form; see [55]) supports the notion that forerunners tothe Cambrian evolutionary fauna were present millions of years before the Cambrian boundary.

That Cambrian-type metazoan organisms are present and apparently diversifying in the latestEdiacaran poses an alternative question: does the Nama ‘Wormworld’ interval represent theearliest phase of the Cambrian Explosion? Although opinion is still divided on whether many ofthe White Sea fauna represent metazoans (Box 1), the appearance of biomineralization,macroscopic predation, and complex, Treptichnus-like burrowing behaviors are arguablymetazoan traits that have far more in common with the Cambrian evolutionary fauna thanthe Ediacaran White Sea fauna preserved in South Australia and Russia. If all the ecological andevolutionary changes that take place over the Ediacaran and Cambrian are considered

Box 1. Which of the Ediacara Biota Represent Metazoans?

Despite >80 years of study the biological affinities of the Ediacara biota remain enigmatic, principally because few shareany synapomorphies with Phanerozoic animal groups. Some, however, do preserve characters that align them withmetazoan phyla that subsequently diversify in the Cambrian. Among these, Kimberella (IV) is the best candidate for abilaterian metazoan, possessing a number of mollusk-like characteristics: Kimberella was likely mobile, preservesevidence for a muscular ‘foot’, and is commonly associated with scratch marks (Kimberichnus isp.), illustratingpossession of paired radula-like structures. Other Ediacaran groups, however, possess body plans entirely unknownamong extant metazoan phyla. For example, the Rangeomorpha (I) are a clade that possesses ‘fractal’ morphologycomposed of self-similar frondlets, while the Erniettomorpha are constructed from unbranched cylindrical modulesarranged in leaflike sheets. Although the Dickinsonimorpha (V), Triradialomorpha (VI), and Bilateromorpha (VII) have all atone time been placed somewhere on the metazoan tree (e.g., the Dickinsonimorpha with Placozoa, see [59]), there hasbeen no scientific consensus on these placements (Figure I). As a result, there is currently no basis for assigning thesegroups to the Metazoa. We therefore split soft-bodied Ediacaran fossils in Figure 2 into either ‘Ediacara biota’ (generawith no incontrovertible metazoan synapomorphies) or ‘plausible metazoans’. The former group includes many iconicEdiacaran fossils, such as Dickinsonia, Tribrachidium, and Rangea (Figure 1), while the latter includes Kimberella, as wellas possible sponges such as Thectardis, Rugoconites, and Paleophragmodictya. The distinctive tubular and calcifyingforms of the ‘Wormworld’ fauna during Nama time are likely metazoans (for example, [18]). Splitting the Ediacaran fossilassemblage in this way oversimplifies much of the phylogenetic uncertainty surrounding some of these organisms,however, it illustrates the rise of organisms in the late Ediacaran, which bear a stronger resemblance to the Cambrianevolutionary fauna.

660 Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9

Outstanding QuestionsWhat, and when, was the Shuram?

What, and when, was the BACE?

Can we disentangle correlation andcausation in Ediacaran extinction?

What role did end-Ediacaran extinctionplay in the Cambrian Explosion?

Paleophragmodictya Inaria Kimberella

Avalofractus

DeuterostomiaPorifera Placozoa Cnidaria Lophotrochozoa Ecdysozoa

Rangeomorpha?Examples of unknown

groups

Bilateria

Dickinsonia Tribrachidium Spriggina

II III IV

I

V VI VII

Figure I.

together, it could be argued that the biggest change in organismal and ecological complexityarrives at the White Sea–Nama boundary, rather than at the E–C boundary itself. In this context,establishing the timing and driver of the Shuram excursion takes on added importance, as inthis scenario it would coincide not only with the earliest major turnover in complex macroscopiclife, but also with the onset of the most dramatic evolutionary radiation in Earth History.

Concluding RemarksUnderstanding the drivers and pacing of late Ediacaran extinction, as well as how it isconnected to the evolutionary radiation that followed, will reveal the origins of the modernmetazoan-dominated biosphere. On the one hand, support for the ‘catastrophe’ model wouldillustrate a crucial role played by environmental fluctuations in driving early metazoan evolution.On the other, support for ‘biotic replacement’ would instead represent the first evidence for abiological driver of mass extinction, and suggest that evolutionary innovation, ecosystemengineering, and biological interactions ultimately caused the first major biotic turnover ofcomplex life. However, we also note that these two models are not mutually exclusive; either orboth could be operating in the late Ediacaran and be responsible for either of the two turnoverevents. Even more intriguingly, the two could be mechanistically linked – the effects ofbioturbation on sediment and ocean chemistry are well known (e.g., [56]), and so a model

Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9 661

whereby the evolution of new metazoan behaviors were responsible for perturbations to theglobal carbon cycle (the Shuram and BACE) is one that deserves testing. In summary, the E–Ctransition is undoubtedly a complex interval and determining between abiotic and biotic driverswill be challenging. However, the questions we pose above are all readily answerable with newdata, and the recent discovery of continuous sections through the E–C boundary preservingboth a rich fossil assemblage and a detailed chemostratigraphic record [57] offers muchpromise for understanding this critical interval in Earth History.

AcknowledgmentsS.A.F.D. acknowledges generous funding from National Geographic (Grant No. 9968-16), and a Paleontological Society

Arthur Boucot Award. E.F.S. was supported by the Smithsonian Institution Peter Buck Fellowship Program and The

Palaeontological Association (Award ID PA-RG201703). M.L. acknowledges funding from NSERC Discovery Grant

(RGPIN 435402). D.H.E. was supported by NASA through National Astrobiology Institute (Grant No. NNA13AA90A)

to the MIT node. This manuscript was considerably improved after constructive reviews by Susannah Porter and two

anonymous reviewers.

References

1. Xiao, S. and Laflamme, M. (2009) On the eve of animal radiation:

phylogeny, ecology, and evolution of the Ediacara biota. TrendsEcol. Evol. 24, 31–40

2. Droser, M.L. et al. (2017) The rise of animals in a changingenvironment: global ecological innovation in the Late Ediacaran.Annu. Rev. Earth Planet. Sci. 45, 593–617

3. Muscente, A.D. et al. (2018) Environmental disturbance, resourceavailability, and biologic turnover at the dawn of animal life. EarthSci. Rev. 177, 248–264

4. Buatois, L.A. et al. (2014) Ediacaran matground ecologypersisted into the earliest Cambrian. Nat. Commun. 5, 3544

5. Knoll, A.H. et al. (2006) The Ediacaran Period: a new addition tothe geologic time scale. Lethaia 39, 13–30

6. Landing, E. (1994) Precambrian-Cambrian boundary global stra-totype ratified and a new perspective of Cambrian time. Geology22, 179–182

7. Gehling, J.G. et al. (2001) Burrowing below the basal CambrianGSSP, Fortune Head, Newfoundland. Geol. Mag. 138, 213–218

8. Pu et al. (2017) Dodging snowballs: geochronology of the Gask-iers glaciation and the first appearance of the Ediacaran biota.Geology 44, 955–958

9. Tsukui, K. et al. (2016) High-precision temporal calibration of theearly Cambrian biotic and paleoenvironmental records: new U-Pbgeochronology from eastern Yunnan, China. In AmericanGeophysical Union Annual Meeting. San Francisco, CA,American Geophysical Union Abstract ID: PP31A-2260

10. Waggoner, B.M. (2003) The Ediacaran biotas in space and time.Integr. Comp. Biol. 43, 104–113

11. Boag, T.H. et al. (2016) Ediacaran distributions in space and time:testing assemblage concepts of earliest macroscopic bodyfossils. Paleobiology 39, 591–608

12. Dececchi, T.A. et al. (2017) Relating Ediacaran fronds. Paleobiol-ogy 43, 171–180

13. Gehling, J.G. and Droser, M.L. (2013) How well do fossil assemb-lages of the Ediacara biota tell time? Geology 41, 447–450

14. Darroch, S.A.F. et al. (2015) Biotic replacement and mass extinc-tion of the Ediacara biota. Proc. Biol. Sci. 282, 20151003

15. Mangano, M.G. and Buatois, L.A. (2014) Decoupling of body-plan diversification and ecological structuring during theEdiacaran–Cambrian transition: evolutionary and geobiologicalfeedbacks. Proc. Biol. Sci. 281, 20140038

16. Grotzinger, J.P. et al. (2000) Calcified metazoans in thrombolite-stromatolite reefs of the terminal Proterozoic Nama Group,Namibia. Paleobiology 26, 334–359

17. Wood, R. et al. (2002) Proterozoic modular biomineralized meta-zoan from the Nama Group, Namibia. Science 296, 2383–2386

662 Trends in Ecology & Evolution, September 2018, Vol. 33, No

18. Zhuravlev et al. (2015) Ediacaran skeletal metazoan interpreted asa lophophorate. Proc. Biol. Sci. 282, 20151860

19. Schiffbauer, J.D. et al. (2016) The latest Ediacaran Wormworldfauna: setting the ecological stage for the Cambrian Explosion.GSA Today 26, 4–11

20. Amthor, J.E. et al. (2003) Extinction of Cloudina and Namacala-thus at the Precambrian–Cambrian boundary in Oman. Geology31, 431–434

21. Laflamme, M. et al. (2013) The end of the Ediacara biota: extinc-tion, biotic replacement, or Cheshire cat? Gondwana Res. 23,558–573

22. Smith, E.F. et al. (2016) Integrated stratigraphic, geochemical,and paleontological late Ediacaran to early Cambrian recordsfrom southwestern Mongolia. Geol. Soc. Am. Bull. 128, 442–468

23. Smith, E.F. et al. (2017) A cosmopolitan late Ediacaran bioticassemblage: new fossils from Nevada and Namibia support aglobal biostratigraphic link. Proc. Biol. Sci. 284, 20170934

24. Erwin, D.H. and Valentine, J.W. (2013) The Cambrian Explosion:The Construction of Animal Biodiversity, Roberts and Company

25. Kump, L.R. and Arthur, M.A. (1999) Interpreting carbon-isotopeexcursions: carbonates and organic matter. Chem. Geol. 161,181–198

26. Condon, D. et al. (2005) U-Pb ages from the NeoproterozoicDoushantuo Formation, China. Science 308, 95–98

27. Grotzinger, J.P. et al. (2011) Enigmatic origin of the largest-knowncarbon isotope excursion in Earth’s history. Nat. Geosci. 4,285–292

28. Knauth, L.P. and Kennedy, M.J. (2009) The late Precambriangreening of the Earth. Nature 460, 728–732

29. Husson, J.M. et al. (2015) Ca and Mg isotope constraints on theorigin of Earth’s deepest d13C excursion. Geochim. Cosmochim.Acta 160, 243–266

30. Higgins, J.A. et al. (2018) Mineralogy, early marine diagenesis,and the chemistry of shallow-water carbonate sediments.Geochim. Cosmochim. Acta 220, 512–534

31. Zhu, M. et al. (2006) Advances in Cambrian stratigraphy andpaleontology: integrating correlation techniques, paleobiology,taphonomy and paleoenvironmental reconstruction. Palaeoworld15, 217–222

32. Zhu, M. et al. (2017) A deep root for the Cambrian explosion:implications of new bio- and chemostratigraphy from the SiberianPlatform. Geology 45, 459–462

33. Jones, C.G. et al. (1994) Organisms as ecosystem engineers.Oikos 69, 373–386

34. Erwin, D.H. (2008) Macroevolution of ecosystem engineering,niche construction and diversity. Trends Ecol. Evol. 23, 304–310

. 9

35. Darroch, S.A.F. et al. (2016) A mixed Ediacaran-metazoanassemblage from the Zaris Sub-basin, Namibia. Palaeogeogr.Palaeoclimatol. Palaeoecol. 459, 198–208

36. Burgess, S.D. et al. (2014) High-precision timeline for Earth’smost severe extinction. Proc. Natl. Acad. Sci. U. S. A. 111,3316–3321

37. Hull, P. et al. (2015) Rarity in mass extinctions and the future ofecosystems. Nature 528, 345–351

38. Macdonald, F.A. et al. (2013) The stratigraphic relationshipbetween the Shuram carbon isotope excursion, the oxygenationof Neoproterozoic oceans, and the first appearance of theEdiacara biota and bilaterian trace fossils in northwesternCanada. Chem. Geol. 362, 250–272

39. Xiao, S. et al. (2016) Towards an Ediacaran time scale: problems,protocols, and prospects. Episodes 39, 540–555

40. Le Guerroue, E. et al. (2006) 50 Myr recovery from the largestnegative delta C-13 excursion in the Ediacaran ocean. Terra Nova18, 147–153

41. Williams, G.E. and Schmidt, P.W. (2018) Shuram–Wonokacarbon isotopeexcursion: Ediacaran revolution in the worldocean’s meridional overturningcirculation. Geosci. Front. 9,391–402

42. Swart, P. and Kennedy, M.J. (2012) Does the global stratigraphicreproducibility of d13C in Neoproterozoic carbonates require amarine origin? A Pliocene–Pleistocene comparison. Geology 40,87–90

43. Schrag, D. et al. (2013) Authigenic carbonate and the history ofthe global carbon cycle. Science 339, 540–543

44. Cui, H. et al. (2017) Was the Ediacaran Shuram Excursion aglobally synchronized early diagenetic event? Insights from meth-ane-derived authigenic carbonates in the uppermost DoushantuoFormation, South China. Chem. Geol. 450, 59–80

45. Grotzinger, J.P. et al. (1995) Biostratigraphic and geochronologicconstraints on early animal evolution. Science 270, 598–604

46. Knoll, A.H. and Carroll, S.B. (1999) Early animal evolution: emerg-ing views from comparative biology and geology. Science 284,2129–2137

47. Corsetti, F.A. and Hagadorn, J.W. (2000) Precambrian-Cambriantransition: Death Valley, United States. Geology 28, 299–302

48. Bristow, T.F. and Kennedy, M.J. (2008) Carbon isotopeexcursions and the oxidant budget of the Ediacaran atmosphereand ocean. Geology 36, 863–866

49. Shen, S.Z. et al. (2011) Calibrating the end-Permian massextinction. Science 334, 1367–1372

50. Bengston, S. and Zhao, Y. (1992) Predatorial borings in lateprecambrian mineralized exoskeletons. Science 257, 367–369

51. Rahman, I. et al. (2015) Suspension feeding in the enigmaticEdiacaran organism Tribrachidium demonstrates complexity ofNeoproterozoic ecosystems. Sci. Adv. 1, e1500800

52. Darroch, S.A.F. et al. (2017) Inference of facultative mobility in theenigmatic Ediacaran organism Parvancorina. Biol. Lett. 13,20170033

53. Krug, A.Z. and Jablonski, D. (2012) Long-term origination ratesare reset only at mass extinctions. Geology 40, 731–734

54. Erwin, D.H. (2015) Early metazoan life: divergence, environmentand ecology. Philos. Trans. R. Soc. Lond. B Biol. Sci. 370,20150036

55. Jensen, S. et al. (2000) Complex trace fossils from the terminalProterozoic of Namibia. Geology 28, 143–146

56. Canfield, D. and Farquhar, J. (2009) Animal evolution, bioturba-tion, and the sulfate concentration of the oceans. Proc. Natl.Acad. Sci. U. S. A. 106, 8123–8127

57. Smith, E.F. et al. (2016) The end of the Ediacaran: two newexceptionally preserved body fossil assemblages from MountDunfee, Nevada, USA. Geology 44, 911–914

58. Maloof, A.C. et al. (2010) The earliest Cambrian record of animalsand ocean geochemical exchange. Geo. Soc. Am. Bull. 122,1731–1774

59. Sperling, E.A. and Vinther, J. (2010) A placozoan affinity forDickinsonia and the evolution of late Proterozoic metazoanfeeding modes. Evol. Dev. 12, 201–209

Trends in Ecology & Evolution, September 2018, Vol. 33, No. 9 663