Giving the Early Fossil Record of Sponges a Squeeze

Home , Ausia (animal), Calcareous sponge, Grypania, Otavia, Palaeophragmodictya, Thectardis

Biol. Rev. (2014), pp. 000–000. 1 doi: 10.1111/brv.12090 Giving the early fossil record of sponges a squeeze

Jonathan B. Antcliffe1,2,∗,RichardH.T.Callow3 and Martin D. Brasier4,5 1Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, U.K. 2Department of Earth Sciences, University of Bristol, Wills Memorial Building, Queen’s Road, Bristol BS81RJ, U.K. 3Department of Geology and Petroleum Geology, School of Geosciences, University of Aberdeen, Meston Building, Aberdeen AB24 3UE, U.K. 4Department of Earth Sciences, Oxford University, Parks Road, Oxford OX13PR, U.K. 5Department of Earth Sciences, Memorial University of Newfoundland, 300 Prince Philip Drive, St John’s A1B 3X5, Canada

ABSTRACT

Twenty candidate fossils with claim to be the oldest representative of the Phylum Porifera have been re-analysed. Three criteria are used to assess each candidate: (i) the diagnostic criteria needed to categorize sponges in the fossil record; (ii) the presence, or absence, of such diagnostic features in the putative poriferan fossils; and (iii) the age constraints for the candidate fossils. All three criteria are critical to the correct interpretation of any fossil and its placement within an evolutionary context. Our analysis shows that no Precambrian fossil candidate yet satisﬁes all three of these criteria to be a reliable sponge fossil. The oldest widely accepted candidate, Mongolian silica hexacts from c. 545 million years ago (Ma), are here shown to be cruciform arsenopyrite crystals. The oldest reliable sponge remains are siliceous spicules from the basal Cambrian (Protohertzina anabarica Zone) Soltanieh Formation, Iran, which are described and analysed here in detail for the ﬁrst time. Extensive archaeocyathan sponge reefs emerge and radiate as late as the middle of the Fortunian Stage of the Cambrian and demonstrate a gradual assembly of their skeletal structure through this time coincident with the evolution of other metazoan groups. Since the Porifera are basal in the Metazoa, their presence within the late Proterozoic has been widely anticipated. Molecular clock calibration for the earliest Porifera and Metazoa should now be based on the Iranian hexactinellid material dated to c. 535 Ma. The earliest convincing fossil sponge remains appeared at around the time of the Precambrian-Cambrian boundary, associated with the great radiation events of that interval.

Key words: Precambrian, Porifera, fossil calibration, sponges, Ediacaran, origin of animals, macroevolution, Cambrian explosion.

CONTENTS I. The Cambrian explosion ...... 2 II. The phylogeny of animals and the characteristics of Porifera ...... 2 (1) Potentially useful characteristics of hexactinellids and demosponges ...... 4 (2) Potentially useful characteristics of calcareous sponges ...... 4 III. Case studies from the Cambrian: Archaeocyatha and stem group sponges ...... 5 IV. The earliest poriferan candidates ...... 8 (1) An organic walled vesicle, Wynniatt Formation, Canada (between 723 and 1077 Ma) ...... 9 (2) Jacutianema and related forms, Lakhanda Sequence, Russia, c. 1020 Ma ...... 10 (3) Otavia antiqua, a ‘sponge-like’ fossil from Namibia c. 760 Ma ...... 11 (4) ‘Sponge spicules’ from Nevada, USA c. 750 Ma ...... 11 (5) Organic ‘demosponge biomarkers’ from the Cryogenian of Oman ...... 11 (6) ‘Pre-Varangerian’ monaxons from Alaska and northwest Canada ...... 13 (7) ‘Pre-Marinoan bioclasts’, South Australia ...... 13 (8) Monaxons and body fossils from Doushantuo Formation, China ...... 13 (9) Thectardis avalonensis from Newfoundland c. 565–555 Ma ...... 14

* Address for correspondence (E-mail: [email protected]).

(10) Fedomia mikhaili from Russia c. 558–555 Ma ...... 14 (11) Ausia from Namibia c. 555 Ma ...... 15 (12) Vaveliksia vana from Russia c. 555 Ma ...... 16 (13) Palaeophragmodictya from Australia c. 555 Ma ...... 16 (14) Cucullus fraudulentus from the Miaohe Biota, China c. 555 Ma ...... 17 (15) Spicules from the Doushantuo and Dengying Formations, China, c. 550 Ma ...... 17 (16) Namapoikia, Namacalathus and spicules from Namibia c. 549 Ma ...... 17 (17) Hexaxon spicules from Mongolia c. 545 Ma ...... 18 (18) Spicules from the Lower Tal Formation, Lesser Himalaya, India ...... 18 (19) Spicules from the Soltanieh Formation, Northern Iran c. 535 Ma ...... 21 (20) Fossils from various localities in Siberia, Lower Cambrian ...... 23 V. A discussion of our search criteria for the earliest sponge fossils ...... 24 VI. A discussion constraining the origin of sponges ...... 25 VII. Conclusions ...... 29 VIII. Acknowledgements ...... 29 IX. References ...... 29

I. THE CAMBRIAN EXPLOSION (Seilacher, 1977; Bottjer et al., 2000; Orr, Benton & Briggs, 2003). The Precambrian-Cambrian boundary is a fundamental These many lines of evidence suggest that animal groups dividing line in the history of life. This boundary separates emerged near the base of the Cambrian Period and an earlier world with no certain animal remains, lasting for quickly established ecological feedbacks resulting in the some 4000 million years, from a later world where undoubted construction of these first complex ecosystems (Butterfield, animal remains abound. The fossil record indicates that 2007). Precambrian fossils, however, including candidate all the major animal phyla appeared in an interval of animals, have been much more difficult to interpret. Finding approximately 30 million years following the Ediacaran- definitive Precambrian crown-group animal characteristics Cambrian boundary. However, Precambrian ancestors is the only way to refute a relatively late radiation of animals. to the Metazoa have long been sought (e.g. Glaessner, In practice this is complicated by the lack of generally agreed 1984; Gehling, 1991; Runnegar & Fedonkin, 1992; criteria for what such organisms might look like, or how Fedonkin & Waggoner, 1997; Sperling, Pisani, & Peterson, we might otherwise recognize them. This is particularly 2007). important for sponges, both as animals that might have a The Cambrian fossil record of animals contains a low preservation potential, and because they are generally coherent and well-documented set of evidence. This includes recovered as being basal or a sister group to all other animals. evidence for: the origin of bioturbation (Cowie & Brasier, Our aim is thus to erect criteria and establish a methodology 1989; McIlroy & Logan, 1999); the first direct evidence for assessing such (generally controversial) claims. In turn of zooplankton (e.g. the phosphatocopid Klausmuelleria this allows us to address a series of exciting questions, such salopiensis from the Lower Cambrian Comley Limestone; as: did deep ancestors have negligible fossilization potential see Hinz, 1987; Siveter, Williams, & Waloszek, 2001; or has the fossil record of early animal evolution been Siveter, Waloszek, & Williams, 2003; Harvey, Velez,´ & misinterpreted? Does our approach to classifying fossil taxa inevitably lead us to assign them to relatively derived, rather Butterfield, 2012); the rapid expansion of biomineralisation than more basal positions? Or have fossil problematica, (Porter, 2007, 2010; Brasier, Antcliffe & Callow, 2011); such as the Ediacaran biota, been incorrectly assigned to the construction of the first animal reef systems (Gandin plesiomorphic positions within metazoan phylogeny? & Debrenne, 2010); the assembly of solid exoskeletons from multiple skeletal elements (e.g. Bengtson, 1985, 1992; Bengtson et al., 1990); the diachronous geographical spread of new metazoan groups (Cowie & Brasier, 1989); and II. THE PHYLOGENY OF ANIMALS AND THE the continued radiation of these new groups (Marshall, CHARACTERISTICS OF PORIFERA 2006). Changes in ocean chemistry also appear to have followed the emergence of early animal activity, including Sponges are positioned near the base of the animal a reduction of oceanic dissolved organic carbon owing group (Halanych, 2004; Srivastava et al., 2010) and so to zooplankton ventilaton of the oceans (Logan et al., have been considered top candidates for metazoan fossil 1995; Bottjer, Hagadorn, & Dornbos, 2000; Butterfield, remains in the Precambrian, for example the ‘long expected 2007), a decrease in carbonate saturation of the oceans sponges’ of Gehling & Rigby (1996). This has resulted due to the evolution of calcium carbonate biominerals in numerous claims for Precambrian Porifera and has (Brasier et al., 2010), and the aeration of the sediment stimulated interested in Precambrian palaeontology in recent profile by the activities of bioturbation and bioirrigation decades. Our aim herein is to assess the reliability of these

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society The early fossil record of sponges 3 claims, constrain their potential sources of error, and discuss there is no reason to suppose that sponge apomorphies are alternative hypotheses for their interpretation. any more likely to be found in the Precambrian than the Porifera are traditionally divided into three main groups: characters of any other animal phylum. Therefore, the safest the Calcarea (calcareous sponges), Hexactiniellida (silica or approach when examining the Precambrian fossil record is glass sponges), and Demospongia (spongin and/or silicate to deal with each candidate fossil on its own merits and spiculed ‘horn’ sponges). There has been extensive recent not with the expectation that we should be finding sponge debate regarding the phylogenetic relationships of these fossils. We should also consider what characters to look for in sponge groups, to each other and to wider metazoan the Precambrian fossil record to convince ourselves that we phylogeny. The relationships of the modern sponge groups to have found a reliable sponge fossil. Doing so is not a trivial extinct major groups of sponges, such as the Archaeocyatha, problem at all. are also controversial. The main debate concerns whether Phylogenetics clearly has a role to play in shaping our sponges are monophyletic or paraphyletic. The hypothesis expectations of the appearance of basal animals but we that sponges are paraphyletic and represent a grade of should be careful not to confuse phylogenetically useful morphological organisation has been popular over the last characters with characters that are useful for understanding decade (Borchiellini et al., 2001), and the Hexactinellida the affinities of enigmatic fossils. Phylogenetics in this regard have also been regarded as paraphyletic (Dohrmann et al., has warped our thinking about the early fossil record. To 2008.) This is supported by molecular data (Sperling et al., identify a Precambrian animal we don’t need to demonstrate 2007, 2010) and has had considerable influence on views of convincingly that the fossil possess characters that are Precambrian evolution. This is because sponge paraphyly synapomorphic to all Metazoa, or even just synapomorphic would indicate that the last common ancestor of the to Porifera. All we need to do is demonstrate that the fossil Eumetazoa was a sponge, and as Hexactinellida is the possesses any apomorphic character that convinces us it first class to diverge then the most probable Precambrian was an animal. Alternatively we could recognise a suite of animal would be a sponge with hexactinellid characteristics. features that are collectively diagnostic of sponge affinity. Further, recent molecular clocks based on a small number The problem with many defining characteristics of the of nuclear housekeeping genes dated animal phylogenies Porifera is that they have limited fossilisation potential – for showing sponge paraphyly (Sperling et al., 2007, 2010; instance the structure of the mesohyl, or the association Erwin et al., 2011) and recovered very deep Precambrian of amoebocytes, choanocytes and porocytes. However, divergences for the origin of the Metazoa, in the range of c. Demospongiae, Hexactinellida, and Calcarea, all possess 800 million years ago (Ma). many apomorphies with good fossilisation potential. A growing body of evidence has now shown that the This approach does not imply that the only certain way apparent paraphyly of the sponge classes is actually an to recognise Precambrian sponges is to discover crown- artefact of very restricted data sets and limited taxon group characteristics in Precambrian fossils. It does however sampling. These studies find good support for the monophyly suggest useful search criteria for Precambrian sponges; we of sponges from genomic data (Srivastava et al., 2010) as well have to start somewhere and the modern suite of crown- as transcriptomic/Expressed Sequence Tags (EST) analyses group sponges is our main point of anatomical comparison. with much denser taxon sampling (Philippe et al., 2009; Stem-group search criteria are also receiving attention at Pick et al., 2010). The sponge paraphyly studies by Sperling the moment, particularly with regard to possible stem-group et al. (2007) and Sperling et al. (2010) may also have been sponges that have already been found in the Cambrian. biased by the choice of inappropriate outgroups (Philippe These will be discussed in the next section. et al., 2011). Worheide¨ et al. (2012) provide a comprehensive It has recently been theorised that basal animal phylogeny review of sponge phylogenetics, including an in-depth look may yet be resolved by new and better understandings of at the paraphyly versus monophyly debate, and conclude stem-group and crown-group concepts (Budd & Jensen, that ‘Sponge monophyly is (a) supported by currently the largest 2000) and this has helped a great deal in terms of amount of phylogenomic data (in terms of amino acid positions and constraining relationships of Cambrian taxa (e.g. Daley in-group taxon sampling) ... and (b) is congruent with cladistic et al., 2009) and understanding models of Cambrian diversity analyses of morphological characters’ (p.13). Nosenko et al. (2013) (Budd & Jensen, 2000). However, resolving basal animal demonstrated that the paraphyletic topology can also arise phylogeny and the sequence of character acquisition has as a consequence of data partitioning between ribosomal proved complex (Marshall, 2006). Rapid rates of evolution and non-ribosomal genes. The monophyly of the sponge may make it difficult to unravel the sequence of character classes has significance for understanding the Precambrian evolution in fossiliferous strata that are time averaged. This fossil record. It means that there is no need to argue that the is due to the interplay of rapid evolutionary rates and low most recent common ancestor of eumetazoans was a sponge. preservation potential of diagnostically useful characters. The As a result, crown-group sponges would be as far from the slow radiation of Mammalia, from the Permian through to last universal common ancestor of animals as are the crown- the Cenozoic [a duration of around 200 million years (Myr)], group characters of any other animal phylum (Jenner, 2006). is comparably well illuminated by fossils (Kemp, 1999, 2004), That is to say that the modern crown-group sponges would implying that this is not a general problem of the fossil record not be a good model for the ancestral animal and as a result per se in constraining major evolutionary hypotheses, but

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society 4 Jonathan B. Antcliffe and others may actually reflect the scale and rapidity of the events instance, how would one visualise a stem sponge that lacked taking place during the Cambrian. The systematic study of porocytes? Is this even mechanically viable? Would it even the taphonomy of phylogenetically informative characters is constitute a sponge or should it be considered a non-animal beginning to help shed some light on the early evolution of colony of single-celled eukaryotes. animals, although most recent studies (e.g. Sansom, Gabbott The only safe course of action when investigating the & Purnell, 2010) focus too far up the tree to help specifically early fossil record of the Porifera, therefore, is to explore with basal animal radiations. Projects are ongoing to try to evidence for the oldest certain fossil representatives of sponges, produce comparable taphonomic data for the Cnidaria and rather than the oldest possible representatives of sponges. the protostomes. The burden of proof must be to demonstrate the presence The following hypotheses could arise when dealing with of definitive characteristics, rather than to assume that fossils lacking any definitive crown-group characteristics. candidates represent poriferans because of their great age Firstly, one could assert that the fossil is a member of a crown and their coincidence with prediction of divergence times sponge group but that the definitive characteristics are absent from some molecular clocks. We urge that more emphasis for taphonomic reasons. Conclusions made along these lines be placed upon the meeting of agreed criteria (and that usually result in authors referring to the fossil as a ‘probable there should be a debate about what these are) and less be sponge’, or ‘sponge-like’. Secondly, one could assert that placed upon the meeting of conjectures about evolutionary the fossil is a member of the stem lineage of a sponge trajectories. We also argue that the explosion of animal group, meaning that definitive characteristics have not yet forms near to the base of the Cambrian – as demonstrated evolved. Unfortunately, this evades the critical question of by an intensely studied fossil record – provides the only character recognition, whilst simultaneously conforming with reasonable null hypothesis. Our intention is to present a a hypothetical pattern of a deep Precambrian history for guiding methodology for dealing with early enigmatic fossils, animal groups. Thirdly, an author may hypothesise that looking at both the nature of the evidence required and at there is no demonstrable affinity to modern sponges. This the need to test non-animal hypotheses. latter course of action is, in our opinion, by far the safest. A stem group by definition consists of extinct organisms in possession of some but not all of the crown-group characters. (1) Potentially useful characteristics of As such, it is more closely related to the crown group of hexactinellids and demosponges the phylum than it is to any other group (Budd & Jensen, The search for siliceous spicules in the Precambrian has a 2000). However, Precambrian macrofossils arguably lack long history. Many of the candidates that we consider in the characters of any modern group and yet they must Section IV are based on the preservation of possible siliceous have a closest living relative. A further problem concerns spicules. This is a reasonable search criterion as spicules are the lack of constraints on the expected morphology of highly distinctive of Porifera, usually occur in large numbers, basal groups when projected far back beyond the last and have a high fossilisation potential. However, there are certain member of the group. When dealing with basal numerous abiological mechanisms, such as silica precipita- groups, such as sponges, the matter is rendered yet more tion, which can generate putative spicular structures and this difficult because the morphological starting point for this means that claims must be examined very carefully based on group is not well understood. Compare the origin of their taphonomic and diagenetic history. There are numer- sponges to the well-constrained evolutionary sequence from ous distinctive forms that spicules can take that cannot easily reptiles to mammals, or dinosaurs to birds, where the start be formed by abiogenic processes (Fig. 1). Beyond these many and end points are comparatively well understood. The transition from single-celled eukaryotes to complex Metazoa apomorphic characters, it is possible to identify genuine hex- is substantially more difficult to constrain. Furthermore, act spicules if layered secretion structures and axial canals are stem-group concepts are much easier to apply, for instance, present. Such spicules should be extracted from the surround- to the study of stem arthropods in the Cambrian where ing rock matrix and be expected to occur in large numbers, fossils of the crown group are temporally close and direct with regular morphology in a given rock sample. It is difficult ancestors and descendants are well understood from fossil to form many regular spicules with axial canals and layered examples (Daley et al., 2009) yet even in this case, it is difficult secretion structures by abiological silica crystallisation. to understand the sequence of character acquisition and debate is ongoing. This is more difficult when considering (2) Potentially useful characteristics of calcareous the study of stem sponges in the Precambrian, because there sponges is no good understanding of character development (see Halanych, 2004), and their single-celled sister group (the The Calcarea are massive carbonate-secreting sponges and choanoflagellates) has no fossil record. Sponges are, at least are substantially less diverse than the other two main groups in some sense, a cellular association related to differentiated in modern oceans. They lack siliceous spicules so the search colonial choanoflagellates. But it is not clear which cell type criteria differ from those already discussed. Calcareous differentiated first or was acquired first, nor indeed what sponges in general are confined to waters above the carbonate effect this would have had upon overall morphology. For compensation depth, and usually far above it, so as to

secreted, and the complex and varied ontogenetic modes. All of these characteristics are connected in many ways to the debate surrounding why the enigmatic Cambrian fossil group the Archaeocyatha are considered to be sponges. This is discussed in the following section.

III. CASE STUDIES FROM THE CAMBRIAN: ARCHAEOCYATHA AND STEM GROUP SPONGES

Numerous fossil groups previously regarded as enigmatic have over recent decades been conclusively shown to be sponge fossils. The enigmatic early Cambrian calcareous- bodied Archaeocyatha (see Debrenne, Zhuravlev & Kruse, 2012, p. 1) ‘were the first Paleozoic metazoans to engage in extensive bioconstruction, in some regions building reef complexes rivalling those of the present’. The history of their phylogenetic interpretation is complex (Debrenne & Vacelet, 1984; Debrenne, Zhuravlev & Kruse, 2012b; Debrenne et al., 2012) as they have been variously interpreted as algae (Opik,¨ 1975), foraminifera (Meek, 1868), cnidarians or coelenterate grade (Walcott, Fig. 1. Phylogeny of Hexactinelida based on molecular 1894), as a unique phylum (Hill, 1972), or even a new data, mapping on apomorphic spicule characters modified kingdom (Zhuravleva, 1970). It was during the late 1970s that after Dohrmann et al. (2008). Illustration and description of convincing evidence began to emerge that archaeocyathans characters after Boury-Esnault & Rutzler (1997). The characters are an extinct group of sponges. We reiterate that evidence shown help to illustrate the great diversity in the diagnostic here, as it is useful when analysing claims for the earliest morphologies of siliceous sponge spicules. Characters are as candidate fossil sponges. Brasier (1976) documented a follows: (1) triaxonic spicules – spicules with three axes (these series of competitive interactions, phobotactic behaviour may possess five or six rays). (2) Hexasters – microscleres with and growth structures that are strongly indicative of a six branching rays. (3) Hypodermal pentactins – pentactines metazoan grade of organisation. Such macroscopic structures are triaxonic spicules with five rays; hypodermal refers to the position of the spicule on the surface of the sponge with one of demonstrating competitive interactions are not known from the rays protruding inside the body and the others tangential outside the Metazoa. to the external surface of the body. (4) Amphidiscs – spicules In general sponges contain two forms of skeleton, the with a single shaft and umbrella-shaped endings. (5) Uncinates primary spicular skeleton, which may be constructed of perpendicular to dermal surface – uncinates are diactin (two- carbonate or silica spicules depending on the group, and rayed) spicules with small barbs. (6) Dictyonal skeleton – a the secondary massive/aspicular skeleton, which if present three-dimensional network comprised of fused rigid spicules; is typically built of carbonate. It was thought that the different types of dictyonal skeleton are distinguished as primary spicular skeleton was always present in sponges, characters 6a and 6b. (7) Sceptrules – monactine (one-rayed) therefore, the absence of spicules in the massive calcareous spicules which protrude from the surface of the sponge and bodies of archaeocyathans was problematic for confident terminate at one end with a complex structure; several examples placement of this group within the Porifera. When Vacelet are shown. (8) Choanosomal pentactines – pentactines that are used to support the canal system often referred to (1977) described a living sponge, Vacelitia crypta,withan as constructing the ‘primary/principal/main skeleton’. (9) aspicular calcareous, skeleton this perfectly reasonable Scepters – uncinate spicules which terminate at one end in objection to the poriferan hypothesis was overcome. Addi- four tiny actins. (10) Floricomes – hexasters with S-shaped rays tionally, Vacelitia shares numerous morphological features that end in ornamented umbrella shapes. (11) Diarhyses – a with both archaeocyathans (specifically Archaeosyconoida, radial shape in the canals of the skeleton. (12) Plumose Warriootacyathus, Ardrossacyathus:seeWorheide,¨ 2008) and orientation of dermal diactins (two-rayed spicules on the external another enigmatic (probably polyphyletic) sponge grouping, body surface) – plumose morphology is when spicules radiate the ‘sphinctozoans’. This is interesting because although obliquely from each other. growth-mode features may not help to place an enigmatic fossil in a particular sponge group, it seems that they are have readily available carbonate. They are universally pore constraining enough to assure a poriferan affinity. bearing, using the pores in their carbonate skeleton to draw Growth and ontogeny models are critical to the sea water which they filter for the bacteria they consume for phylogenetic placement of the Archaeocyatha (see Fig. 2A–F food. The useful fossil characteristics of calcareous sponges for an illustration of the different forms). The mode of are related to their porous massive skeleton, how it is growth in Vacelitia crypta is very similar to that seen in

(B) (C)

(A) (D)

(E) (F)

(G) (H)

(I) (J)

Fig. 2. Different growth and architecture schematics of the Archaeocyatha, parts A–F after Debrenne et al. (2012). (A) Single-walled conical archaeocyathid type. (B) Double-walled conical archaeocyathid type. (C) Multi-chambered thalamid type. (D) Syringoid type. (E) Single-chambered spherical type. (F) Chaetitid type. (G) Thin section of sample from the basal Tommotian Nochoroicyathus sunnaginicus Zone showing a single-walled conical archaeocyathid-type sponge called Archaeolynthus polaris (preserved in pale calcite), alongside phosphatised small shelly fossils including Tiksitheca licis (brown circle at top left), from sample BS1b, Dvortsy, Aldan River. (H) Fragments of the single wall of Ar. polaris, here shown alongside hyolith Allatheca sp., from the same sample. (I) A two-walled archaeocyathan from the N. kokoulini Zone, Archchagyi Kyyry-Taas, Oxford University Museum of Natural History OUMNH AY.199 with field samples numbers AKT 15–58. (J) An archaeocyathan from the upper Atdabanian Wilkiwillina Limestone of South Australia, showing the granular, layered, aspicular skeleton (dark grey calcite, with later white calcite infilling internal pore spaces), OUMNH AY.200 from field sample number AC35.1 dL.

‘sphinctozoan’ fossils and some Archaeocyatha (the ‘thalamid numerous tubular pores, of which the most primitive forms type’, see Fig. 2C), where pored calcareous chambers of had just a single wall (Fig. 2G, H). More advanced forms had the secondary skeleton are added at the margin distal to a double wall connected by porous septa. The ancestral forms the attachment site. However, many of the Archaeocyatha underwent rapid morphological diversiﬁcation through the do not show this sphinctozoan/thalamid type of growth. Tommotian to Botoman stages, involving a series of The simplest archaeocyathans had aspicular walls bearing innovations to the nature of their inner and outer wall

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society The early fossil record of sponges 7 pores (see Fig. 2I, J). These innovations resulted in the topology. Quite rightly Sperling et al. (2010) discussed secretion of complex spines, ridges or microporous sheaths. whether siliceous spicules should be basal to a crown The porous double wall raised the efficiency of filtration demosponge/hexactinellid sister group with some secondary following simple Bernoulli fluid dynamics. This is critical losses, or basal to a hexactinellid/spiculate demosponge because it validates a model of water filtration for very group. Such contrasts are germane to our understanding small food particles like bacteria. All modern sponges eat of the sequence of character acquisition in early animal bacteria, and almost always as the primary source of food phylogeny, and this to some degree informs our expectations (for instance, some sponges are secondarily photosynthetic of the characters most likely to be found in the early fossil and cultivate a range of symbionts). The conical porous record. The Cambrian fossil record shows that there are double-wall structure of the Archaeocyatha is unique to that unusual and unexpected character combinations in some group. But there are also good reasons for uniting double early sponge fossils that would not be predicted by a study walled archaeocyathans with those with other growth modes. of molecular biology alone. The case of the aspiculate The chaetetid-type growth mode proceeds from a juvenile Archaeocyatha as described above is a good example of with double-walled archaeocyathid type, while the syringoid the fossil record informing our understanding of character type evolved from the chaetetid type (Debrenne et al., 2012). diversity beyond what was known from studies of living These growth modes also occurred amongst the Sphinctozoa sponges. Understanding the body plan evolution of early during the Mesozoic. The presence of spines on the septa sponges is made particularly difficult as ‘most early sponges (intracalicular spines) along with astrorhizae provides a did not have their skeletal spicules fused’ (Botting et al., feature that confidently links chaetitid-type archaeocyathans 2012). This means that the organisms typically disarticulate with chaetitid demosponges (Debrenne & Zhuravlev, 1994). on death and that complete (articulated) bodies are very Thus to place massive calcareous aspiculate fossils with their rare, and as a result appreciating character combinations closest living relatives we must carefully examine the growth during early sponge evolution is difficult. There have been dynamics of the fossils. numerous articulated stem-group sponges described from In terms of establishing a definitive poriferan affinity for all the Cambrian, the most well known of which is probably archaeocyathans, only the single-walled type remain poten- Eiffelia from the middle Cambrian Burgess Shale, Canada tially problematic. These forms, such as Archaeolynthus polaris (Botting & Butterfield, 2005), although there are several (see Fig. 2G), make their first appearance in the fossil record other interesting cases, such as Lenica from the lower alongside double-walled archaeocyathid-type forms that Cambrian (Tommotian/Atdabanian, i.e. Stage 2 to early define the Nochoroicyathus sunnaginicus Zone at the base of the Stage 3) of China (Botting et al., 2012). Both Eiffelia and Tommotian. Thus the supposed ancestral single-walled form Lenica have been argued to provide evidence for stem does not extend the evolutionary history beyond that confi- sponges which show character combinations of calcareous dently described by the double-walled type. The reason for and silicate sponges. Eiffelia, a heteractinid calcarean, shows regarding single- and double-walled types as part of the same possible compound spicules of silica and calcium carbonate group is again because of shared ontogenetic features. Both (Botting & Butterfield, 2005) and a mix of hexactine develop from an aporous cup in the early ontogenetic stage spicules (like Silicea) and hexaradiate spicules (like Calcarea). which as the organism grows becomes the cemented attach- Eiffelia also seems to possess an organic spicule sheath ment structure (Debrenne et al., 2012) then pores develop found only in modern calcareans (Botting & Butterfield, on ontogenetically later parts of the wall (this is shared with 2005). Harvey (2010) described a suite of fossils with chaetitid and thalamid archaeocyathans; see Fig. 2C, F). ‘carbonaceous composition of calcarean-type spicule sheaths, Several important characters can help us identify ancestral but hexactinellid-type morphology’ (Harvey, 2010, p. 834) calcareous sponges in regard to the secondary calcareous from the lower Cambrian (stage 4) Forteau Formation, skeleton, regardless of whether they are ultimately of Newfoundland, Canada. Botting et al. (2012) suggests that Demospongea or Calcarea affinity. Firstly, understanding the these carbonaceous sheaths further support the idea of mode of growth of the specimens is critical to their attribution Calcarea and Silicea spicular homology. Harvey (2010) is to the Porifera; the ontogenetic dynamics are well mapped not convinced by such homology, however, and suggests that out in both extinct and extant forms (see Fig. 2). Secondly, apomorphic convergence of the carbonaceous sheath is more understanding the filtration system and its fluid dynamics parsimonious, although perhaps more surprising given the is critical to understanding whether the fossil could have exclusivity of spicular sheaths to modern calcareans. These functioned as a sponge. Thirdly, the presence of pores on the discoveries taken together provide compelling evidence for external body wall which penetrate through to the inner part early complexity on the sponge stem lineage(s) and mean that of the organism and are arranged in ordered rows. These there is good reason to suppose that early sponge fossils can pores are not simple holes, they are secreted by a process of be rich with apomorphies that allow fossils to be confidently biologically controlled mineralisation as pores with regular placed within the Porifera. So it seems that the disparity of position, form, and size. The position of the pores on the characters in early to middle Cambrian sponges may actually body is an evolutionary adaptation to maximising fluid flow. have been higher than in later sponge groups (Harvey, 2010). The debate concerning stem-group characteristics of This is encouraging because stem-group fossils depauperate sponges is closely linked to views of their phylogenetic in fossilisable characters do not make tracing the early

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society 8 Jonathan B. Antcliffe and others evolution of a major group any easier. The main significance the claim could present a case for good characters that are of these Cambrian (stem) sponges is to help understand the unique to sponges (thereby passing the characters criterion) sequence of character acquisition of the major sponge groups but fail to convincingly show their presence in the fossil can- and help to inform phylogenetic topology. However, here didate (thereby failing the diagnosis criterion). We term this we are concerned more with timing than topology. What are ‘error type 2’. The reverse is also possible. The claim could the oldest reliable sponge fossils? demonstrate characters that are definitely present in the fossil candidate (passing the diagnosis criterion), but these characters may not be unique or limited to sponges (so failing the IV. THE EARLIEST PORIFERAN CANDIDATES characters criterion). We term this ‘error type 3’. As we review the candidates below, the different types of error advance the debate in different ways. An error of type 2 still helps to Agreement about the status of early animal fossil candidates can help to constrain some important inferences. On one develop understanding of taphonomy and diagenesis and to hand, they can help constrain molecular rates of evolution what extent abiogenic or microbial processes may develop and estimates of divergence times (Peterson et al., 2004; Rota- structures that are putatively similar to characters that would Stabelli et al., 2011). On the other hand, they can help to otherwise be considered unique to sponges. An error of type constrain understanding of the evolutionary processes acting 3 still helps to develop consideration of what characters are during major radiations and their causes (Budd, 2008). Here definitive and useful for detecting sponges in deep time. we shall review all the main candidates claimed to be the Using geochronological minimum constraints is best earliest Porifera. We will not cover claims for the origin of practice when constructing molecular clock analyses and multicellularity (e.g. El Albani et al., 2010), because many shaping views of evolutionary time (Benton, Donoghue & other groups (see Butterfield, 2009; Knoll, 2011) have in Asher, 2009). Following the methodology established by some way transitioned from the unicellular to the multi- Benton et al. (2009), if a fossil is reported as being between cellular (see Butterfield, 2009; Donoghue & Antcliffe, 2010). 600 and 800 Myr old with an error bar of 10 Myr on both Distinct uncertainties that need to be addressed for each can- the upper and lower boundary, then the appropriate age didate are as follows: (i) the morphological characters needed to consider the specimens would be 590 Ma. Therefore our to categorize a particular animal group in the fossil record focus for satisfying the age criterion is the accuracy and (characters criterion) i.e. whether the claimed for characters precision of the minimum age constraint, which should are useful to help to narrow the possible affinities of the be based on absolute dating, or well-established regional fossil; (ii) the uncertainty regarding whether the characters correlations to dated sections. We regard it as a matter are actually present in the fossil (diagnosis criterion); and (iii) of bad practice for researchers to report a fossil candidate the uncertainty about the age constraints for the candidate occurring at a headline age which is at or much closer to the fossils (age criterion). Thus a candidate could fail these tests maximum end of the possible age range with the details of by failing either or both of our first two criteria (see Table 1). the dating uncertainty hidden away in technical details. This We term failing both characters and diagnosis criteria ‘error creates an artificial sense of the antiquity of specimens which type 1’ which means that not only do the characters described is often then overlooked by future researchers using these not categorize a particular fossil group but they are not even dates as calibrations, or in narratives of the history of life. We adequately shown to be present in the fossil. Alternatively urge future authors making claims for early fossil candidates

Table 1. Explanation of criteria and possible interpretations

Characters Diagnosis Result for interpretation Error criterion criterion Meaning as early sponge number Fail Fail The characters claimed for are not useful for identifying Fail 1 fossil sponges, e.g. non-sponge groups also possess these characters. These characters have not been reliably shown to be present in the candidate fossil(s), e.g. there is little account of how taphonomy or abiogenic mechanisms could have generated the putative character. Pass Fail The characters are useful for detecting sponges in the fossil Fail 2 record but have not been reliably shown to be present in the particular candidate fossil. Fail Pass The characters claimed for are reliably present in the Fail 3 candidate fossils but these characters are not useful for identifying fossil sponges. Pass Pass The characters claimed for are useful for detecting sponges Pass — in the fossil record and have been reliably shown to be present in the particular candidate fossil.

Fig. 3. Generalised geological column spanning the interval from ∼820 to 490 Ma, showing the horizons of candidate early sponge records, calibrated against major carbon isotope oscillations within the Neoproterozoic (after Halverson et al., 2005) and the Cambrian (after Brasier & Sukhov, 1998). Error bars on the right show uncertainty on dating of various geological horizons discussed in the text, highlighted because it is critical to understand the uncertainty of dates when examining claims for the earliest fossil representatives of a group. The four shaded sections show detail of late Ediacaran to early Cambrian horizons used for correlation throughout the text, used consistently throughout the figures: orange = ‘Cloudina zone/pre-Anabarites shelly fossil zone’ characterised by a large positive carbon isotope excursion; yellow = Anabarites trisulcatus zone; green = Purella antiqua zone; blue = old ‘Tommotian stage’ now the lower part of Series 1, Stage 2. of the Metazoa to report ages at the minimum boundary or & Xiao, 1998) that was ‘tentatively identified as an early sponge to state the possible range prominently. or choanoflagellate’ (Butterfield, 2009, p. 207). This solitary Below, we review some 17 candidates of Proterozoic age specimen, approximately 100 μm across, was identified in a of possible early sponge fossils, taking them in chronological palynological preparation of mudstones from the Wynniatt order, from the oldest onwards (see Fig. 3). These are Formation, Victoria Island, Arctic Canada (Fig. 4A). followed by a review of three candidate poriferan-bearing The minmum age of this formation is constrained by cross Cambrian localities and their suites of fossils, including the cutting Diabase intrusions (723 Ma) and the maximum age by first full and illustrated report of sponge spicules from the detrital zircons from the underlying Nelson Head Formation Cambrian of Iran. (1077 Ma; Butterfield, 2005). While a range of 350 Ma provides weak constraints, it is noted that the fossils come from strata near the top of the Wynniatt Formation and (1) An organic walled vesicle, Wynniatt Formation, demonstrably are Proterozoic in age. Since the illustrated Canada (between 723 and 1077 Ma) acritarch shows none of the distinctive morphological A review of Precambrian multicellular organisms described characteristics of sponges, we feel that the author was an organic walled vesicle (i.e. an acritarch sensu Zhang, Yin rightly cautious. Here we question whether these structures

(B) (C) (A)

(E)(D) (F)

Fig. 4. Questionable early sponge microfossils revealed by petrographic studies. (A) Acritarch with ‘cruciform structure’ from the Wynniatt Formation, Canada, constrained between 1077 and 723 Ma (after Butterfield, 2007). (B, C) Otavia antiqua (after Brain et al., 2012) abiogenic calciphosphate sedimentary grains from the Otavi Group, Namibia c. 760 Ma. (D) Acicular silica crystals from the Upper Tindir Group, Alaska with an upper bounding date of c. 570 Ma (after Allison & Awramik, 1989). (E, F) Sedimentary structures, mud chips, and microbialites from Trezona Formation constrained to between 664.7 and 570 Ma (after Maloof et al., 2010). are ‘cruciform’ (Butterfield, 2009, p. 207, fig. 7). Firstly, with structures they interpret as mesoglea, sponge larvae, ‘cruciform’ structures are not diagnostic for the Porifera. and oscula. Further, German & Podkovyrov (2012, p. 219) Instead we consider this structure better explained by the describe the fossils as ‘in association with nematode-like organisms’. hypothesis of a series of randomly orientated wrinkles and These nematode-like organisms contain none of the features ridges in a cyst wall. The specimen seems little different from that one would expect to see in a nematode fossil, in fact they other Ediacaran and pre-Ediacaran cellular clusters and seem to be featureless and compressed organic tubes (even ornamented algal cysts, which are very common through showing full overfolds of the tube), akin to the many microbial this interval (e.g. Butterfield, Knoll & Swett, 1994). An structures well known globally at around this age [see for ornamented cyst or cellular stage is so common across example the well-known form Grypania (Walter, Oehler & the tree of life that the only phylogenetic hypothesis that Oehler, 1976), particularly their Plate 2, figs 2 and 3]. can probably be rejected on this basis would be that The interpretation as sponge embryos is similarly suspect. of the Bacteria and Archaea kingdoms (but see Cavalier- Firstly, there seems to be no consideration of taphonomic Smith, 2002, 2006, who disagrees even on this point). This process, as the material is preserved as impressions in claim does not meet the three criteria due to the lack of siliciclastics, probably an impossible medium for the diagnostic characters established in the fossils and weak preservation of an embryo. Secondly, there are no diagnostic geochronological constraints. criteria present in these specimens that lead logically to the conclusion of a sponge affinity. Jacutianema itself seems (2) Jacutianema and related forms, Lakhanda comparable to the very many microbial sheaths and filaments Sequence, Russia, c. 1020 Ma found in similar Mesoproterozoic sediments around the German & Podkovyrov (2012, p. 219) recently reported world. Finally, surfaces on which these small specimens are a variety of organic structures from the Mesoproterozoic found are covered in fossilised microbial mats of the type (‘Riphaean’) sections of Siberia, Russia to be ‘presumably commonly referred to as classic ‘elephant skin’, and now of sponge organization grade’. Dated to between 1025 and officially categorised as microbially induced sedimentary 1015 Ma, the material was described as sponge ‘blastula’, structures (MISS) sensu Noffke et al. (2001). The ‘fossil’

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society The early fossil record of sponges 11 specimens grade into the MISS and evidently are part of and microbial colonies are also macroscopic. No features the same microbial structures. This claim suggests characters are presented that are diagnostic of sponges. We interpret that would be distinctive for the Porifera such as mesoglea and these objects as calciphosphate grains that have been pitted distinctive sponge larvae. However, these characters have not by sediment reworking. This claim does not meet any of our been demonstrated to be present in these fossils, and there three criteria. is no consideration of the taphonomic process. Therefore the claim passes the characters criterion (for suggesting (4) ‘Sponge spicules’ from Nevada, USA c. 750 Ma potentially good characters) but fails the diagnosis criterion (as the characters are not convincingly present in the fossil). The oldest claimed sponge spicules have been reported from The geochronology seems to be well constrained. the Noon Day Dolomite in Nevada (Reitner & Worheide,¨ 2002) and were described as follows (Reitner & Worheide,¨ (3) Otavia antiqua, a ‘sponge-like’ fossil from 2002, p. 53): ‘The oldest spicules with demosponge affinities were found Namibia c. 760 Ma by the senior author (Reitner) in ca. 750 my old Noon Day Dolomite in Nevada’. The Noon Day Dolomite is a cap carbonate These small calc-phosphatic objects, from the Otavi group (and therefore a chemical precipitate) of Neoproterozoic of northern Namibia and also from the probably younger age (Corsetti & Kaufman, 2005). Unfortunately, the putative Nama Group of southern Namibia, have been claimed to be sponge material has not been formally described; images have the oldest fossilised evidence for animals (Brain et al., 2012). not yet been published; stratigraphic and locality information These irregularly pitted grains range from 0.3 to 5 mm and seem loosely constrained; and published catalogue numbers were isolated from ‘black limestones’ by acid maceration for a depository in a museum have not been provided. Since (Fig. 4B, C). However, alternative hypotheses have not been no evidence has yet been presented for these structures or sufficiently explored and/or have been rejected without their interpretation as sponge spicules, we are unable to sufficient reason. For instance, Brain et al. (2012, p. 2) deem them as meeting our characters and diagnosis criteria, examine whether there is ‘a priori quantitative evidence against although we, of course, would welcome the presentation biogenicity’ and use a rather unusual statistical treatment to and detailed documentation of this material in the scientific conclude that there is not, adding ‘this conclusion provides no literature. We include the statements of the Noon Day evidence of biogenicity’. However, the burden of proof should material here as a matter of completeness. Despite the lack be on those authors to conclusively demonstrate biogenicity. of any presented data this claim has occasionally been cited Brain et al. (2012) dismiss the idea that these structures could as evidence for the earliest sponges, including recently by be bacterially induced due to a lack of micro-scale coccoids German & Podkovyrov (2012). or botryoidal surfaces, but this could also be taphonomic. It is always very difficult to argue an affinity based on negative (5) Organic ‘demosponge biomarkers’ from the evidence. If one wants to reject a bacterial origin for a Cryogenian of Oman particular fossil it is much more reliable to find a character present in the fossil that logically demonstrates a eukaryotic High concentrations of the compound 24-isopropyl grade of organisation, and could not have been produced cholestane in Cryogenian and Late Ediacaran rocks of the by any bacteria (see Antcliffe & McLoughlin, 2008, for a Huqf Supergroup, Oman (McCaffrey et al., 1994; Hod et al., discussion). To say the structure lacks a bacterial flagellum 1999; Love et al., 2009) have been widely cited as providing is not diagnostic for animals, because it is also true of many the oldest evidence for sponges in the Proterozoic (e.g. Maloof bacteria. Brain et al. (2012, p. 4) reject an amoebozoan affinity et al., 2010; Sperling et al., 2010). These organic compounds because, firstly, Otavia ‘has a phosphatic shell’ which they state are referred to as exclusive indicators of certain Demospongia is uncommon amongst Amoebozoa, and secondly because (Love et al., 2009) and are found from sediment beneath the ‘foraminifera are typically smaller (generally < 200 μm) than Otavia’. Marinoan cap carbonate (c. 635 Ma; Hoffmann et al., 2004; However, Brain et al. (2012, p. 4) also state ‘it is assumed that Condon et al., 2005) and above Sturtian glacial sediments Otavia was originally composed of calcium carbonate’. It would seem (∼713 Ma; Brasier et al., 2000). There are numerous serious strange to make a phylogenetic case based on the diagenetic problems with this evidence as recently argued by Antcliffe replacement of calcium carbonate with phosphates. Even (2013). The principal objections are outlined below. so, phosphate biomineralisation is well known amongst Problems arise when using biomarkers as evidence for protists and has a respectable record well back into the the earliest sponges since it assumes the modern biota Precambrian (see Cohen et al., 2011) so the possible presence represents all the possible producers and cannot account of phosphate is a very unreliable basis for rejecting all for the presence of this biomarker in stem groups (see unicellular eukaryote affinities. Further, Amoebozoa, and Brocks & Butterfield, 2009). The argument requires the in particular Foraminifera, regularly attain sizes greater demonstration of an exclusive relationship between a certain than 200 μm, making size a very unreliable criterion for biomolecular group and a given biological group, which dealing with phylogenetic hypotheses for Precambrian fossils is a challenge given the non-availability of extinct forms. (see Antcliffe, Gooday & Brasier, 2011). Being macroscopic For the case in hand, we find that published claims for does not logically lead to the conclusion of an animal biomarkers typically state that these compounds ‘are primarily affinity because plants, fungi, macro-algae, macro-ciliates, found in certain genera of Demospongiae’(Loveet al., 2009, p.719).

The secondary producers (see supplementary information Organic compounds should be treated like all other published online by Love et al., 2009) are various marine fossils, with original form that decays and then reacts to the algae. Love et al. (2009) use the absolute abundance of these taphonomic process. Like all characters used unequivocally compounds produced by modern demosponges and algae to diagnose the presence of a biological group in the fossil to argue that because demosponges produce the compounds record, they must be unique and distinctive. We must avoid in greater quantity they are more likely to be the producers basing our analyses on convergent characters. There is no responsible for the related compounds in Proterozoic strata. reason why the case should be different for the character Love et al. (2009) note that the choanoflagellate Monosiga of molecular fossils. These data provide no diagnostic brevicollis does not produce precursor sterols, acknowledging sponge-like character, and other non-metazoan groups are the need to constrain the exclusivity of these compounds to documented as capable of biosynthesising these compounds. sponges. We also know that other modern sponge groups, There are considerable uncertainties with the dating such as the Hexactinellida do not produce these organic of these sequences, which is based on uncertain regional compounds (Blumenberg et al., 2002). As we know that the correlations and geochronological dating from detrital compounds are not unique to sponges (because they are zircons. The application of precise dates requires regional also produced by algae) then in addition to the models stratigraphic comparisons between the glacial deposits presented by Sperling et al. (2010) these compounds could in Oman and other dated sequences. Such correlations be a convergent character within at least one algal group are difficult because numerous glacial deposits in the and also the demosponges. This possibility means that these Neoproterozoic show geological sequences that may be compounds are not phylogenetically distinctive in any way compared with those in the Huqf Supergroup (Brasier et al., + for total group sponges, or choanoflagellates sponges, or 2000). The geochronological arguments are also reviewed just demosponges. This convergent evolution of an organic by Antcliffe (2013) to which the reader is referred for further compound is clearly a rational possibility and makes a detail. The widely reported dates of 751 Ma (Love et al., false dichotomy of the two models in Sperling et al. (2010) 2009) are regarded as incorrect for several reasons. Firstly, where the known algal source for such sterols is never when working with detrital zircons, the youngest date is discussed. always chosen, for the obvious reason that it is not possible Interestingly, the Huqf sediments are also found to to incorporate a younger grain into an older rock. Therefore contain 24-n-propylcholestane, a compound found within the current youngest date of 713 Ma (Bowring et al., 2007) modern marine algae (Moldowan et al., 1990; Love et al., is the one we must consider to be the most reliable. The 2009). The only difference between 24-n-propylcholestane alternative and arguably more likely interpretation is that and 24-isopropylcholestane is the attachment site of the the oldest chosen date of 751 Ma (from Love et al., 2009) has propyl∼ substituent to the cholestane. There are only two been inherited from the proximal basement. Secondly, in options in the three carbon-chain propyl: either as an end carbon (in which case it is n-propyl∼); or as the central carbon the Miqrat 1 core which yields the oldest of the biomarkers, (inwhichcaseitisisopropyl∼). In modern terminology, they the correlated glacial deposits must be from the Ghrubah are referred to as 24-propan-1-ylcholestane and 24-propan- Member of the Ghadir Manqil Formation if the biomarkers 2-ylcholestane, respectively. These molecules have exactly are to date from the 713 Ma date given. This formation the same chemical composition and there is only one minor contains two glacial deposits; the older from the Gubrah structural difference. Some algae produce the isopropyl form, Member (dated to 713 Ma) and the younger from the and it is very similar to the n-propyl form that they are well Fiq Member (dated to younger than 645 Ma; see Fig. 3). known to produce in abundance (Moldowan et al., 1990). Curiously, it seems that the oldest possible date of the Recent work has examined the ecology of bacteria formation has been given for these biomarkers, even though associated symbiotically with marine sponges (Siegl et al., they seemingly come from its youngest beds, the Fiq Member. 2011). This suggestive work found that a group of bacteria, As a result the appropriate maximum date becomes 645 Ma the Poribacteria, seem to be exclusively symbiotic with and not 751 Ma as given in Love et al. (2009). Thirdly, when sponges and contain the same genes thought to be responsible proposing calibration points for major evolutionary events, for the production of the biomarker compounds in sponges geochronological minimum dates should be reported as best (Siegl et al., 2011). It is at best unclear what the implications practice (Benton et al., 2009). Currently, the date range of this are, and how valid these findings are. It is possible that for the oldest of the biomarkers is constrained between the Poribacteria are producing the biomarker compound 645 and c. 580 Ma if we use the Shuram Formation in addition to, or instead of, the demosponge host. This excursion correlation as a minimum constraint or 645 and needs to be validated by future work to demonstrate that the c. 635/580 Ma if we use the Fiq Glaciation as the Marinoan bacteria are actually capable of producing the compounds. Glaciation correlation. If indeed the Poribacteria can produce the biomarker then These claims fail all three criteria because there is no reliability can be placed in the Oman sponge biomarkers. substantial doubt over whether this molecular character Certainly this case raises the idea that we must look much is exclusive to sponges, and whether the ratios that it is based more carefully at the microbial ecosystem associated with all on are a true representation of the original biology. The animal biomarkers. geochronology is also poorly constrained (oldest biomarkers

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society The early fossil record of sponges 13 between 645 and 580 Ma) and the published headline dates that are distinctive for sponges are presented by the authors. (751 Ma) are substantially misleading. Instead, support for their poriferan nature was inferred from the presence of sponge biomarkers in rocks elsewhere of (6) ‘Pre-Varangerian’ monaxons from Alaska and approximately this age (but see Section IV.5), and from northwest Canada claims for sponge fossils in somewhat later Ediacaran formations. Nor has the possibility of a microbial origin Thin-walled and possibly hollow spicule-like structures been rejected. Given that extensive stromatolitic material is (Fig. 4D) have been claimed to be hexactinellids (Allison present at this locality, it appears to us that an origin of these & Awramik, 1989), and have been reported from Alaska ‘bioclasts’ from the break-up of calcified microbial mats has (Allison & Awramik, 1989) and the Upper Tindir Group not yet been firmly rejected. The assertion that stromatolitic (Unit 5) of northwestern Canada (Debrenne & Reitner, structures are ‘commoninthefossilrecord, but have laminar 2001). These strata have not been dated directly and their or columnar structures’ (Maloof et al., 2010, p. 656) does not Pre-Varangerian age is based upon global correlation with take into account the complex morphologies found within other stratigraphic sections (see Macdonald et al., 2010). calcimicrobes, stromatolites and thrombolites (Antcliffe & Chemostratigraphy and acritarch biostratigraphy have been McLoughlin, 2008; McLoughlin, Wilson & Brasier, 2008; used to argue that the ‘Tindir unit 5 is Neoproterozoic in Reitner, Queric & Arp, 2010). Many of the claimed sponge age and most likely predates the Varangian ice age’ (Kaufman, Knoll & Awramik, 1992, p. 184). The Varangian Glaciation body fossils appear to simply be mud chips, ripped up in northern Norway is now thought to be equivalent to by the same storms that caused the disaggregation of the the Marinoan Glaciation of Australia (Gradstein, Ogg & stromatolitic bioherm and are clearly not transported in Smith, 2004, fig. 9.4, p, 138). The Marinoan glacial rocks situ sponge reefs. These fossils have no morphology that is form part of the global stratotype for the base of the diagnostic of sponges and should in our view be more readily Ediacaran Period (Knoll et al., 2004) and various ages for ascribed to the calcimicrobes that abound at these localities. the Marinoan Glaciation have been given: 630 Ma (Ogg, These fossils do not satisfy our characters and diagnosis 2004), 580 ± 10 Ma (Gradstein et al., 2004), 635–599 Ma criteria. The dating is well constrained. (Knoll et al., 2004). Biologically, however, these specimens are not convincing as early sponge remains. Allison & Awramik (1989, p. 253) were mainly concerned with (8) Monaxons and body fossils from Doushantuo reporting ‘mineralized and organic walled microfossils ... in chert Formation, China nodules and chert beds’ within Tindir Group limestones. Given Monaxon spicules in siliceous phosphate within the the association of the putative monaxons with abiogenic Doushantuo Formation (Fig. 5), Weng’an, Guizhou silica precipitates, we conclude that it is highly likely that Province, China, 630–555 Ma (Tang, Zhang & Jiang, these are acicular crystals associated with silicate nodules 1978; Li, Chen & Hua, 1998) have been interpreted as and cement precipitation, as was common during the late siliceous demosponge spicules. Co-occurring ‘soft tissue’was Precambrian (see Maliva, Knoll & Siever, 1989; Maliva, interpreted as evidence for ‘epidermis, porocytes, amoebocytes, Knoll & Simonson, 2005). The irregular morphology of sclerocytes, and spongocoel’(Liet al., 1998, p. 880). However, these specimens and their lack of discernible canal structures the published images suggest that such detail is not present (see fig. 5 in Allison & Awramik, 1989; Fig. 4D) and more in this material. While porocytes are tubular, so are many complex triaxon or hexaxon forms, make it impossible to structures in nature. Amoebocytes have highly variable form separate these specimens from abiogenic acicular crystals. and possess pseudopodia. The pseudopodia are common to Thin sections showing the internal structure and a complete many eukaryotes (Cavalier-Smith, 2006). Sclerocytes form diagenetic history detailing the diagenetic evolution of the cellular triads but so do many other cellular associations and acicular crystals in contrast to the surrounding matrix would again no evidence is shown to demonstrate their presence be required to demonstrate a poriferan origin for these in these rocks. Yin, Xiao & Yuan (2001, p. 1828) were also structures. Currently this claim does not meet our characters critical of the poriferan interpretation of the spicular material and diagnostic criteria and further work would need to ...... substantiate how these could be met for such simple acicular stating that ‘ light microscope and SEM observations do not crystals. substantiate a sponge spicule interpretation of spicular structures. No convincing axial canals have been seen in the observed spicule population ... the coexistence of some monaxonal spicules with clearly diagenetic (7) ‘Pre-Marinoan bioclasts’, South Australia crystal fascicles ... suggest that these monaxons may also be diagenetic A suite of putative sponge fossils from the Trezona Formation in origin. Our preliminary EDS analyses detect no significant silicon of South Australia dated to a maximum age 659.7 ± 5Ma in spicule-containing intraclasts including abiotic clasts and fragments and constrained by the Marinoan glacial deposits above of sphaeromorphic acritarchs and algal thalli, that contain abundant at c. 635 Ma (but see Section IV.5) were found in a monaxonal spicular structures.’ The characters suggested for these stromatolitic bioherm (Maloof et al., 2010) These fossils ‘body fossils’ are useful for distinguishing sponges, however have been interpreted as fragmented bioclasts amongst the they have not been convincingly shown to be present in these stromatolitic reef (see Fig. 4E, F). However, no characteristics fossils. The dates are also not well constrained.

(A) instance, if a specimen had a length to width ratio of 0.9, one could reject the sponge hypothesis, assuming the validity of the test. However, a ratio of 1.8 does not confirm the sponge hypothesis because this ratio is not necessarily exclusive in nature to conical sponges. Many conical/triangular forms from across the tree of life also satisfy this test, for example, seaweeds, orthocone nautiloids, many solitary corals, and conical stromatolites. Further, we doubt the validity of this test for the following reasons. Firstly, it is based on the idea that sponges are perfectly conical, and that they require these overall body ratios in order successfully to maintain water exchange over their surface owing to the continuity equations of fluid dynamics. But sponges are not perfectly conical, and commonly deviate far from this. For instance Tethya represents a large family of demosponges (the Tethyidae) which are superficially spherical (Branch, Griffiths, Branch (B) et al., 2010) and constructed of complex internal bifurcations that are non-uniform and ramifying (Hammel et al., 2011). Further, the different internal organisation of sponges (i.e. asconoid, syconoid, and leuconoid) greatly affects their water-processing ability by creating complex folded surfaces which serve to increase their overall surface area and alter the internal rates of fluid flow (Ruppert, Fox & Barnes, 2003). Secondly, the accuracy of ratio measurements from fossils has to be critically considered. The measurements are taken from retro-deformed images based on assumed amounts of regional tectonism. However, there are no independent measures of tectonic strain for the Ediacaran of Newfoundland (see Liu et al., 2011). Therefore it is not possible to derive such accurate biometric measurements from retro-deformed images at these sites. Thirdly, it is also based on the interpretation that Thectardis avalonensis was Fig. 5. (A, B) Petrographic thin sections through phosphatic conical in life. However there are no structures such as rocks of the early Ediacaran Doushantuo Formation, collapse marks, fractures, or wrinkles on these specimens constrained between 635 and 555 Ma, in south China, showing of the kind expected when a conical organism is flattened. acicular silica crystals previously interpreted as sponge spicules We therefore suggest an alternative hypothesis for testing (e.g. Li et al., 1998). Images provided by Chuanming Zhou. with Thectardis: that it is a highly effaced remnant of some other Edicaran organism representing a stage of extreme (9) Thectardis avalonensis from Newfoundland decay, where nearly all body morphology is gone and all c. 565–555 Ma that remains are microbial mats veiling the smoothed carcass (following Liu et al., 2011). There are comparable examples of Soft-bodied triangular-impression fossils have been reported Charniodiscus from Mistaken Point E surface that are so effaced from tuff-turbidite interfaces in the Pigeon Cove locality in in their preservation that they appear very like Thectardis. the Drook Formation, around 1500 m stratigraphically below The interpetation of Thectardis avalonensis as a fossil sponge the famous Mistaken Point Surfaces (Clapham et al., 2004), is therefore highly problematic. The age of the succession Newfoundland, Canada, c. 565–555 Ma. These fossils have is reasonably well constrained. The characters suggested by recently been interpreted as early evidence for sponges, with Sperling et al. (2011) are definitely present in Thectardis to at the triangular impression reconstructed as a conical body in least some degree (although see argument concerning retro- life (Sperling, Peterson & Laflamme, 2011; see Fig. 6A). They deformation) but they are not characteristic for sponges in use a biomechanical constraint as key evidence for a sponge any way. This claim therefore fails the characters criterion. affinity. Sperling et al. (2011), p. 25) state that ‘a perfectly conical sponge will require a length–width ratio greater than ∼1.6 to ensure that (10) Fedomia mikhaili from Russia c. 558–555 Ma the inhalant surface area is equal to or greater than that of the osculum’ and then further that ‘the full flattening of a cone to a 2D plane Fedomia mikhaili is a soft-bodied, bag-like fossil (Fig. 6B) from would reduce the necessary length–width ratio to ∼1.1’. However, the fossiliferous deposits of the Upper Vendian Verkhovka the application of such as test assumes a knowledge of the Formation (558 ± 1 to 555.3 ± 0.3 Ma: dates are based on original biology of the fossil. Such a test could successfully regional correlation, not direct dating); interpreted as a refute a hypothesis of affinity, but it cannot confirm it. For sponge by Serezhnikova & Ivantsov (2007). No arguments

(A) (B) (C)

(D)

(E) (F)

Fig. 6. Soft-bodied Ediacaran biota fossils previously interpreted as sponges. (A) Thectardis avalonensis from Mistaken Point E surface, Newfoundland, Canada (565–555 Ma), image of cast of Mistaken Point surface. (B) Fedomia mikhaili from Verkhovka Formation (559–555 Ma) of Russia, after Serezhnikova & Ivantsov (2007). (C) Ausia fenestrata from Namibia (555 Ma), also described from White Sea region of Russia (556 Ma). (D) Vaveliksia vana from Mezen Formation (c. 555 Ma), White Sea region of Russia, after Ivantsov et al. (2004). (E) Palaeophragmodictya from Flinders Ranges, South Australia, (c. 555 Ma) after Gehling & Rigby (1996). (F) Primocandelabrum heimaloranum (= Heimalora) from South East Newfoundland (c. 560 Ma) field image with one British penny for scale, comparative form to Palaeophragmodictya but demonstrated as an attachment organ for Ediacaran fronds. were presented by those authors for this interpretation, nor not satisfy our diagnosis or characters criteria as no definitive is it easy to discern any distinctly sponge-like characteristics characters were presented. in the published material. We would greatly welcome an expansion of their hypothesis focussing on which characters (11) Ausia from Namibia c. 555 Ma led them to the conclusions they reached. As with all soft-bodied Ediacaran organisms, we suggest that they are This triangular soft bodied impression shows irregular ovate best understood in terms of their co-occurring biota (i.e. surface structures (Fig. 6C) and was first described by Hahn & Pflug (1985) from the Nama Group of Namibia the Vendobionta sensu Seilacher, 1992). See for instance (c. 545 Ma) and regarded as a pennatulacean cnidarian. It Grazhdankin & Gerdes (2007) for a critical discussion of was reinterpreted by McMenamin (1998, p. 38–39, fig. 2.24) bag-shaped and blob-shaped Ediacaran enigmata and Liu as having potential sponge affinity. The presence of pore-like et al. (2011) for an examination of taphomorphic variety seen depressions was used to support this interpretation. It has in similar fossils. Following those works we suggest that the also been cited as an ancestor of Cambrian archaeocyathan most likely hypotheses for this fossil material is that of a sponges (Fedonkin, 1996) (see Section III). Specimens have microbial colony or macro-algal sheath. These specimens do since been described from the Ust-Pinega Formation, White

Sea Region, Russia (c. 556 Ma; see Grazhdankin, 2004). agglutination is known from many non-metazoan groups. In the Russian sections, Ausia is found within sandstone For instance, the Rhizaria and Amoebozoa are capable of beds in close association with the Ediacaran organisms forming complex macroscopic agglutinated structures e.g. Onegia, Bomakellia and Palaeoplatoda (Grazhdankin, 2004, p. the foraminifera Playsolenites from the lower Cambrian of 209). The sandstones have been interpreted as fluvially Avalonia (McIlroy, Green & Brasier, 2001). Recent reports influenced, shallow marine deposits (Grazhdankin, 2004). have suggested the presence of agglutinated amoebozoans Two specimens of Ausia have been described and illustrated (vase-shaped microfossils) from strata older than 700 Ma from the Namibian sections (see Vickers-Rich, 2007, fig. 131, (Porter, Meisterfeld & Knoll, 2003). Vaveliksia clearly warrants p. 84) which are found in association with specimens of further analysis to see if there are any features that can Ernietta. It has even been suggested that Ausia ‘may indeed demonstrate a macro-single-celled eukaryotic interpretation represent a group related to the chordates’ (Vickers-Rich, 2007, (see Antcliffe, Gooday & Brasier, 2011 for criteria). Vaveliksia p. 79; Fedonkin et al., 2012) although no evidence for this vana lacks any definitive poriferan characteristics and fails interpretation has been presented. Hahn & Pflug (1985) the characters and diagnosis criteria. If agglutination can be considered Ausia to be closely related to Kuibisia which demonstrated and further analysis is done, it may well form Runnegar (1992) regarded as a preservational variant of part of an interesting narrative concerning Precambrian Ernietta. One possibility, therefore, is that Ausia canbebest Rhizaria, although this is currently far from established. understood as a variant of Ernietta or its close ally Onegia. Validation of this suggestion would require detailed study of (13) Palaeophragmodictya from Australia c. 555 Ma the taphonomy of these fossils. The phylogenetic affinities of Ausia, Kuibisia,andErnietta must be dealt with together Ediacaran soft-bodied impressions from the Ediacara as a group. Because Kuibisia and Ernietta are non-porous Member of the Rawnsley Quartzite, Flinders Ranges, South and because Ernietta is certainly infaunal (see the closely Australia, 555Ma (Fig. 6E), have been interpreted as sponge related form Namalia and the very common Pteridinium from (Gehling & Rigby, 1996). This was based on ‘comparison Grazhdankin & Seilacher, 2002) it would have been difficult of preserved characters to Paleozoic hexactinellid dictyosponges’ for these organisms to function as sponges. The suggestion (Gehling & Rigby, 1996, p.185). Palaeophragmodictya has of pores that show some degree of regularity is interesting, subsequently been interpreted as an attachment disc of but there is insufficient consideration of the taphonomy of another Ediacaran organism by Serezhnikova (2007a), who the material and the affinities of the co-occurring forms then further suggested a possible relationship to sponges which are clearly taphonomically related in some way. This or cnidarians. While we agree with Serezhnikova (2007a) claim therefore fails our diagnosis criteria. The dates are that some of the material is likely to represent the holdfast reasonably well constrained. of other Ediacaran-age organisms, we question whether these specimens represent sponges of any kind. We also draw attention to the difficulty of demonstrating cnidarian (12) Vaveliksia vana from Russia c. 555 Ma affinities for such material (see Antcliffe & Brasier, 2007, Vaveliksia vana is a soft-bodied, sac-like fossil preserved as a 2008; Brasier & Antcliffe, 2009). No compelling arguments ‘natural mold’ (Ivantsov, Malahovskaya & Serezhnikova, 2004) are presented that Palaeophragmodictya should be considered from the Zimnii Bereg (Winter Coast) of the White Sea near separately from other disc-like structures of Ediacaran age. A the Zimnegorsk lighthouse, in the Mezen Formation, dated recent review of discoid genera concluded that the majority to around 555 Ma (Grazhdankin, Podkovyrov & Maslov, of forms could either be ascribed to attachment discs of 2005). No definitive poriferan characteristics are present Ediacaran-age organisms, or were completely microbial in these specimens, which are simply sac-shaped and may in origin (Grazhdankin & Gerdes, 2007). Specifically, have been anchored to the sediment surface at one end Grazhdankin & Gerdes (2007, p. 205) state that ‘the peculiar by a small attachment disc (Fig. 6D) comparable to that taphonomy and morphological features of the positive hyporelief moulds seen in other Ediacaran-age organisms. Ivantsov et al. (2004, suggest that Cyclomedusa, Paliella, and Ediacaria were discoidal p. 6) interpret several shallow impressions on the surface structures consisting of microscopic filaments and growing by accretion body as evidence that ‘the walls were supported by spicula-like within microbial mats’. These three genera have previously structures’. An alternative hypothesis would be that these been interpreted as Precambrian jellyfish (Glaessner, 1984; marks are left by the contraction and crushing of a three- Gehling, 1991). The interpretation of microscopic filaments dimensional, soft-bodied object. There is nothing to suggest is critical when considering the arguments regarding the biomineralization or the presence of spicules. As discussed morphology of Palaeophragmodictya.The‘quadrate spicular mesh’ in Sections II and III, making a convincing case for spicules found in the holotype specimen (South Australian Museum, either requires that apomorphic spicule characters should be Adelaide, Australia SAM P32324a) and interpreted as present or there should be evidence of biological secretion the spicular framework of a poriferan (Gehling & Rigby, structures and an axial canal. Ivantsov et al. (2004, p. 6) 1996, p. 192), could therefore have been produced by also argued that the body was ‘soft or composed of sediment microbial organisms growing on the Ediacaran sea floor. particles imbedded into the organic matrix (in the manner of some Indeed, microbial mats are extensively preserved as the sponges)’. It is true that some sponges do agglutinate sediment famous ‘elephant skin’ texture of Ediacaran bedding planes (e.g. Cliona and Chondrosia, see Cerrano et al., 2007) but from Australia (Glaessner, 1984; Gehling, 1999; Brasier

& Antcliffe, 2008; Brasier, Antcliffe & Callow, 2011), and (15) Spicules from the Doushantuo and Dengying mats of this age with sharply defined, spicule-like microbial Formations, China, c. 550 Ma filaments are now being reported (Callow & Brasier, 2009a,b). Calcareous triaxon spicules have been described from the That raises the question as to whether the ‘quadrate spicular Shibantan Member, Dengying Formation, Yangtze Gorges mesh’ (Gehling & Rigby, 1996 p. 192) was truly either region, China, 555–551 Ma (Steiner et al., 1993). Although quadrate or spicular; and whether the ‘peripheral frill with its Neoproterozoic age is assured, constraining the exact age closely spaced, straight, radial spicular ridges’ (Gehling & Rigby, of the Doushantuo Formation has proved to be problematic. 1996 p. 190) can safely be used in support of a poriferan The base of the formation has been estimated between 635 interpretation. Comparable material has recently been and 621 Ma, and the top between 555 and 551 Ma (see described from the Ediacaran of Newfoundland, Canada Condon et al., 2005; Zhang et al., 2005). Spicule-like objects (Hofmann, O’Brien & King, 2008), which unambiguously that somewhat resemble sponge hexacts in thin section have forms part of the basal attachment disc of the frondose also been reported from the Ediacaran Doushantuo (see organism Primocandelabrum sp.(Fig.6F;Hofmannet al., 2008, Section IV.8 this manuscript) and Dengying formations of figs 9 and 12). Similar discoid fossils from Siberia have also Hubei Province (Tang et al., 1978; Zhao et al., 1988; Steiner recently been reinterpreted as attachment discs of Ediacaran et al., 1993). This material has already been well discussed in fronds (Serezhnikova, 2007b). It has been suggested that the literature, from which it has been concluded that these these filaments extended like a root system to help anchor the structures are pseudofossils composed of inorganic crystals organism to the sediment. Unfortunately, Palaeophragmodictya (Zhou, Yuan & Xue, 1998). Zhang, Yin & Xiao (1998) found is only known from a few specimens that have a simple some evidence that several specimens may represent the morphology which is easily misinterpreted. Like all soft- remains of spiny organic-walled vesicles of likely algal origin bodied fossils of Ediacaran age, it is better understood by (acanthomorph acritarchs). Hence, there is already great examining the suite of co-occurring fossils and also the doubt regarding the sponge affinities of this material, and we taphonomic role of microbial mats. This candidate fails the refer readers to those works. We propose here that the null characters criterion. hypothesis must be that the spicules grew abiotically during carbonate precipitation, dolomitisation, or metasomatism. These specimens fail the diagnosis criterion. (14) Cucullus fraudulentus from the Miaohe Biota, China c. 555 Ma (16) Namapoikia, Namacalathus and spicules from Macroscopic carbonaceous compression fossils known as Namibia c. 549 Ma Cucullus occur in the Neoproterozoic Doushantuo Formation, Putative biomineralized calcium carbonate skeletons have South China. These fossils were first described by Steiner been described from the Omkyk Member, Nama Group, (1994), and then interpreted by Steiner & Reitner (2001) 549 Ma (Grotzinger, Watters & Knoll, 2000; Wood, as prokaryotic colonies. Wang & Wang (2011) reiterated Grotzinger & Dickson, 2002) and acicular crystals claimed interpretation of this material as a possible demosponge to be spicules of demosponges have also been found representative despite earlier dismissals of this idea by in these Namibian sections (Reitner & Worheide,¨ 2002). Xiao et al. (2002); Yuan et al. (2002) and Finks & Rigby Among these, Namapoikia rietoogensis was claimed by Wood (2004). The putative sponges of Cucullus are preserved at et al. (2002) to show a robust biomineralised but aspicular multiple horizons within black shales towards the top of skeleton. Though the dating of these sequences seems to the Doushantuo Formation, as a component of the ‘Miaohe be fairly reliable, based on volcanic ash found in the Biota’ (Chen, Xiao & Yuan, 1994). The fossils consist of stratigraphic section and good global correlation (Grotzinger, tubular, sac-like, carbonaceous compressions up to 30 mm in Adams&Schroder,¨ 2005), these specimens do not yet width and 200 mm in length. Wang & Wang (2011) describe meet our other criteria. Diagnostic poriferan morphological a two-layered wall composed of interweaved ‘spongin characteristics are lacking. Namapoikia has an irregular filaments’ (with mesh-like relief) that they say delineate morphology consisting of carbonate walls arranged in a polygonal openings comprising a canal system. A central spongiform pattern. The concern here is that such a structure upper opening surrounded by concentric openings was also could arrive from the calcification of microbial colonies. This described. There is however, no unequivocal morphological fossil is interesting, but cannot be deemed have to met our evidence to support assignment to the Porifera. No clear criteria without high-resolution analysis of both composition canal system or central opening is visible, and the ‘spongin and morphology. The rarity of this material naturally limits fibre’ networks may alternatively be explained as wrinkles such destructive analysis but, in the future, further finds could and folds produced during burial and compaction of the allow more detailed analyses to be performed. Namacalathus soft organism. Furthermore, no spicules have been described consists of a globular body attached to a stalk-like tube. in association with Cucullus. We conclude that assignment The globular body is provided with rather regular lateral to the Porifera is highly speculative and that Cucullus fails openings, typically between five and seven, see Grotzinger the diagnosis test as the possibility that the specimens are et al. (2000). There is no comparable form amongst any microbialites seems to us much more likely. living sponge group, although there is a crude similarity

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society 18 Jonathan B. Antcliffe and others with Cambrian archaeocyathan sponges (Grotzinger et al., sponges (Hooper & Van Soest, 2002). Third, there is no 2000). Archaeocyathans typically have numerous vertically evidence for concentric layered silica secretion, a hollow axial arranged pores for which there is no evidence in Namacalathus. canal, or for an organic axial filament which would support The growth program of the Archaeocyatha is also well a poriferan affinity (see Finks et al., 2003). Fourth, many constrained and occurs by secretion of the cone along structures which appear superficially similar to multirayed the distal margin. There is no comparable evidence for sponge spicules (e.g. stauractines, Fig. 7A–D) are shown growth currently available for Namacalathus. Furthermore, to be artefacts produced by randomly overlapping and carbonate mineralisation alone cannot be taken as indicative closely adjacent acicular crystals (Fig. 7D). Lastly, scanning of metazoan affinities because it is widespread in many electronmicroscopy (SEM)–EDX analysis indicates that the different microbial and algal groups (e.g. cyanobacteria, spicules contain high concentrations of arsenic, iron and rhodophytes and chlorophytes). The spicules reported by sulphur (Fig. 7E). Reitner & Worheide¨ (2002) are small monaxons (< 200 μm These results suggest that these structures are crystals in length). The null hypothesis here is for an origin from of arsenopyrite (see King, 2002), that have experienced abiogenic acicular crystalisation processes, of which there later alteration to products such as iron oxide and, locally, are several candidates in the massive carbonates in which to barium sulphate. Tellingly, they are comparable with they are found. abiogenic arsenopyrite needles from the c. 3.4 Ga Strelley Pool Sandstone of Western Australia, (e.g. Wacey et al., (17) Hexaxon spicules from Mongolia c. 545 Ma 2011). This demonstrates clearly that complex spicule-like structures can be found in rocks that are assuredly too old to An assemblage of sponge-like spicules was described from have contained sponges. Taken together, we conclude that the lower part of the Tsagaan Oloom Formation of the the Mongolian ‘sponge spicule’ clusters should no longer be Gobi-Altay region in Mongolia (Fig. 7; see Brasier, Green & accepted as providing evidence for the earliest hexactinellids Shields, 1997; sample TS29, from bed 25). This assemblage in the rock record. Instead, they appear to be abiogenic occurs in small chalcedonic silica nodules within shallow arsenopyrite crystals which occasionally show complex marine carbonates that yield a carbon and strontium crystal twins. We therefore put forward the following null isotope signal consistent with an age close to the Ediacaran- hypothesis for these specimens: that they are pseudofossils Cambrian boundary, lying beneath the negative excursion that grew within a hydrothermally influenced setting, perhaps known as Anomaly W (Brasier et al., 1996, 1997). This similar to the model invoked for the deposition of Ba-rich indicates a latest Ediacaran age, because it also lies beneath and As-rich rich shales of comparable age in the Lower the first clear assemblages containing either Protohertzina (Chopoghlu) Shale Member of Iran (see Section IV.19). anabarica, Anabarites trisulcatus or Cambrian-type trace fossils Therefore the candidate fails the diagnosis criterion. (Brasier et al., 1997). The specimens are translucent to opaque rods with (18) Spicules from the Lower Tal Formation, Lesser tapering terminations, arranged in patterns that have Himalaya, India been compared to sponge hexacts, pentacts, stauracts and monaxons (Fig. 7; Brasier et al., 1997; Finks, Reid & Deposits in the Lesser Himalaya of India preserve what Rigby, 2003). Others show complex patterns, with seven are arguably the oldest examples of sponge spicule or more rays joined together (see Brasier et al., 1997). cherts – spiculites – in the rock record (Figs 8 and 9). These Until now, the position of these structures as the earliest occur within the chert-phosphorite member of the Lower remains of hexactinellid sponge spicules (Brasier et al., 1997) Tal Formation (Brasier, 1989a, p. 52), for which both has remained unchallenged. New analysis of the material biostratigraphy and carbon isotope stratigraphy suggest a described by Brasier et al. (1997); housed in the Oxford basal Cambrian age (Anabarites trisulcatus Zone; Fig. 8A; University Museum of Natural History - OUMNH AY35-41) Brasier & Singh, 1987; Brasier, 1989a; Azmi & Paul, 2004; now allows us to re-examine this interpretation. Kaufman et al., 2007; Tiwari, 1999). Cruciform structures Thin and polished sections were examined in transmitted of putative hexactinellid origin have been illustrated from light using oil-immersion microscopy and automontage these sites (Mazumdar & Banerjee, 1998; Tiwari, Pant & reconstruction techniques. A scanning electron microscope Tewari, 2000). Further work by us shows that many such equipped with an Energy Dispersive X-ray Spectrometer cruciform structures (Fig. 8B) in this unit may be better (EDX) was used for imaging and compositional analysis of explained as abiogenic ambient inclusion trails (AITs, see structures within polished thin sections. Our re-examination Fig. 8C; see also Wacey et al., 2008a,b). These are formed of the morphology and geochemistry of this assemblage when small pyrite grains are propelled through a medium allows us to shed new light on the nature of these spicule- of chert or phosphate by gaseous pressure during burial, to like structures. First, we find that while the spicules can be leave behind fossil-like trails. Such AITs are often curved or arranged in hexact-like arrays, the spicules themselves are spiral, and have margins that are striated parallel to the axis, not fused together, but show separate, angular terminations cross sections that are polygonal, abrupt terminations and (in (Fig. 7D, E). Second, in transverse section the spicular some cases) a pyrite crystal at these terminations (Fig. 8C). structures are rhombic or diamond shaped, rather than We find that cruciform structures resembling hexactinellid rounded, or triangular as is common for hexactinellid spicules in the lower Tal Formation also show these features,

(A) (B)

(E)

Fig. 7. Clusters of spicule-like structures obtained from latest Ediacaran cherts of the Tsagaan Oloom Formation in Mongolia. (A–D) Thin-section images showing stauract- and hexact-like rods of opaque sulphide mineral floating within chalcedonic silica catalogued in the Oxford University Museum of Natural History (OUMNH) as (A) OUMNH AY.38; (B) OUMNH AY.37; (C) OUMNH AY.41; (D) OUMNH AY.36. (E) Suite of images obtained during Energy Dispersive X-ray of cruciform material from OUMNH AY.38, revealing concentrations of arsenic, silicon and sulphur (top row) and calcium, iron and oxygen (bottom row). Note also the rhombic cross sections of needles, and euhedral crystal terminations at the centre of the cruciform structure. These features are here regarded as consistent with abiogenic arsenopyrite crystals. making their acceptance as the oldest stauracts or pentacts in Fig. 8E). The question that remains here is whether doubtful. these rods may be the remains of oxea-like megascleres. Of Also of interest are siliceous rods (Fig. 9) which are found greatest interest within this assemblage are grains of clear and in large numbers within these chert beds. These rods are finely crystalline chalcedony, which have pinched margins typically about 200 μm long and up to 50 μm wide, with consistent with compression during burial (Fig. 8E). High- circular cross sections and pointed terminations at each magnification optical analysis reveals hundreds of closely end. In many layers, they have been broken into angular fragments prior to deposition (Fig. 8D). While a few appear packed silica objects that range in shape from rounded filled with secondary silica, pyrite or phosphate, the majority spheres, through to spindle-shaped rods (Fig. 9). Most of consist of microcrystalline chalcedonic silica, consistent with these rods are ∼5 μm in diameter and less than 20 μm in recrystallization from an opaline precursor (see large clast length, none of which appear fragmented. In shape, they

(A)

(B) (C)

(D) (E)

Fig. 8. Context for early Cambrian sponge evolution in India. (A) Generalised stratigraphic column showing the Ediacaran Krol Formation and the earliest Cambrian Lower Tal Formation, with chert-phosphate member at the base (dashed), here calibrated against the carbon isotope record from the Ediacaran (after Kaufman et al., 2007) and Cambrian (after Banerjee et al., 1997). The two shaded sections show detail of late Ediacaran to early Cambrian horizons used for correlation throughout the text, and is used consistently throughout the ﬁgures: orange = ‘Cloudina zone/pre-Anabarites shelly fossil zone’ characterised by a large positive carbon isotope excursion; yellow = Anabarites trisulcatus zone. (B, C) Thin sections from the chert-phosphorite member showing cruciform and spiral patterns, here attributed to abiogenic processes called ambient inclusion trails (see text for details). (D, E) Thin-section images showing brown phosphate layers containing abundant rods of chalcedonic quartz (ﬁeld sample PH2 from Martin Brasier’s collection). resemble miniature examples of the larger silica rods that silica needles while coeval deposits in the south of Oman make up much of the chert in this unit. are said to consist entirely of abiogenic silica (Schroder¨ & Of comparable age and appearance are some spindle- Grotzinger, 2007). Such occurrences need much further shaped and arcuate oxeas preserved in silica, reported investigation. Stratigraphic sections from Oman through from basal Cambrian cherts of Shaanxi in south China (e.g. Braun et al., 2007). Cherts from the base of the Fara India to China provide evidence for widespread silica- and Formation in the north of Oman, of basal Cambrian age nutrient-enriched and oxygen-depleted water masses at the (Brasier et al., 2000) also contain abundant monaxon-like base of the Cambrian, in which early sponges may have

(A) (B) the Elburz Mountains (Fig. 10; see Hamdi et al., 1989; Zhiwen, Brasier & Hamdi, 1989; Brasier, 1989a, 1990). The upper assemblage consists of tiny pyritized monaxons, stauracts, pentacts and hexacts preserved within thin sections of phosphatic dolomite (see Figs 11 and 12). At this level, they occur alongside some of the first known siphogonuchitid sclerites and problematic tubes of Tiksitheca licis (Hamdi et al., 1989; Zhiwen et al., 1989). The lower assemblage consists of exquisitely preserved spicular structures isolated using acid maceration from the lower part of the Middle Dolomite (Fig. 12). These consist of small, delicate stauracts (Fig. 12A) and pentacts (Fig. 12B–G) preserved in white translucent silica. The structures have rounded (rather than rhombic) cross sections and, unlike Fig. 9. (A, B) Earliest Cambrian siliceous monaxons from strata the Mongolian structures, show rounded terminations rather in India correlative with the Anabarites trisulcatus zone. Detail of than sharp, pointed terminations. Axial canals and layered the large clast shown in Fig. 8E, revealing slightly curved oxea- secretion structures are well represented in the specimens. like monaxons with tapered ends, in a matrix of fine grained Their morphology compares well with pentacts isolated by chalcedony containing decayed organic matter. us from the younger (Atdabanian) Wilkawillina Limestone of Australia (Fig. 12H, I). These spicules occur within a flourished (Brasier, 1990a,b, 1991, 1992, 1997; Schroder¨ & small shelly fossil assemblage zone dominated by Protohertzina Grotzinger, 2007). anabarica, and alongside Olivooides multisulcatus. This correlates with the Protohertzina anabarica zone in China, provisionally suggested to date to 537 to 532 Ma (Gradstein et al., (19) Spicules from the Soltanieh Formation, 2012). However, our integrated chemostratigraphy and Northern Iran c. 535 Ma biostratigraphy (Fig. 11B) here shows that they occur as low Hexactinellid sponge spicules have been reported from Iran as the negative δ13C Anomaly W that marks the base of the (Hamdi, Brasier & Jiang, 1989; Brasier, 1989a) although little Cambrian Period in Siberia (Fig. 11B), which is now called biological description was provided. These microfossils occur the Basal Cambrian Carbon Isotope Excursion (‘BACE’) that at two levels within the lowermost Cambrian Protohertzina characterizes the base of the Cambrian worldwide (Gradstein anabarica Zone (537–532 Ma see Gradstein et al., 2012) of et al., 2012). In Iran, these spicules occur alongside these the Middle Dolomite Member, Soltanieh Formation, in early skeletal fossils of early Nemakit-Dakldynian (Brasier

Fig. 10. Map showing location of the Iranian sections around the Elburz Mountains that run along the coast of the south end of the Caspian Sea just north of Tehran. Critical sections are found near Soltanieh, Dalir, and Valiabad.

(A) (B)

(E) (F)

Fig. 11. Context for early Cambrian sponge evolution in Siberia and Iran compared. (A) Stratigraphic column for Siberia, here calibrated against the carbon isotope record (after Brasier et al., 1994a,b). Abbreviations for biozones are as follows: Ns = Nochoroicyathus sunnaginicus; Dr = Dokidocyathus regularis; Dl = Dokidocyathus lenaicus; Pj = Profallotaspis jakutensis; Rz = Retecoscinus zegebarti; D. anabara = Delgadella anabara; C. pinus = Carinacyathus pinus (D. anabara and C. pinus are equivalent to Pagetiellus anabarus of Gradstein et al., 2012); F = Fallotaspis; Nk = Nochoroicyathus kokoulini; Fl = Fansycyathus lermontovae; Bm = Bergeroniellus micmacciformis; Bg = Bergeroniellus guararii; Ba = Bergeroniellus asiaticus. (B) Stratigraphic column for the Elburz Mountains of Iran, here calibrated against the carbon isotopic curves of Brasier et al. (1990; black line) plus Kimura et al. (1997; grey line). LDM, Lower Dolomite Member. The precise correlation of the LDM awaits confirmation of its palaeontological character as carbon isotopic markers are not sufficiently distinctive. The three shaded sections in (A and B) show detail of late Ediacaran to early Cambrian horizons used for correlation throughout the text: yellow = Anabarites trisulcatus zone; green = Purella antiqua zone; blue = old ‘Tommotian stage’ now the lower part of Series 1, Stage 2. W shows the position of the basal Cambrian excursion. 13v and 7v mark the position of the sample collection points for AY.201 (D–F) and AY.202 (Fig. 12) respectively and represent the thirteenth and seventh beds in Valiabad section. (C–F) Images of cruciform sponge spicules, preserved as iron sulfide replacements after primary silica, set within a matrix of fine-grained calciphosphate (from the Anabarites trisulcatus zone of the earliest Cambrian, Middle Dolomite Member, sample OUMNH AY.201 (field number 87H.13v).

antiqua assemblage appears close to the first positive peak (B) of the succeeding Zhujiaqing Carbon isotope Excursion (A) [known as ‘the ZHUCE’; e.g. Dahai of Brasier et al. (1990) in China; L4 of Li et al. in China (2009); C of Brasier et al. (1996) in Mongolia; Z, I and I’ of Brasier et al. (1994) in Siberia; ZHUCE of Gradstein et al. (2012). See Discussion in (C) Landing et al. (2013)]. In Iran, Anabarites trisculcatus followed by Purella antiqua assemblages can be calibrated against a carbon isotopic profile that is much like that in Siberia but there are differences too. In Iran, a positive carbon isotope excursion is inferred to lie below the local first appearance (D) of the A. trisulcatus assemblage (Fig. 11B), but it is important (E) to note that this small excursion is an extrapolation derived 13C Corg 13 δ13 = 12C − × from δ Corg (where Corg 13C 1 1000 12C PDB to return a result in parts per thousand. Corg is the value measured from organic carbon in the sample. PDB is the (F) (G) standard from the Pee Dee Belemnite) signals in carbonate- free shales, and therefore comes without corroboration from carbonate values (Kimura et al., 1997; Kimura & Watanabe, 2001). That this is not the ZHUCE peak is shown by the presence of a similar carbon isotope excursion well (I) within the A. trisculatus zone in Mongolia (e.g. excursion (H) AofBrasieret al., 1996). Together, these data can be taken to suggest that the assemblage of hexactinellid spicules likely belongs in the middle part of the Anabarites trisculatus zone (c. 535 Ma), above the BACE and below the ZHUCE excursions. Mineralic needles of non-poriferan origin are known from Fig. 12. Isolated siliceous stauracts and pentacts extracted from lower in the stratigraphy, within the Ediacaran Lower Shale the Protohertzina anabarica zone (containing correlating Anabarites Member, alongside large organic-walled vesicles of Chuaria trisulcatus) of the earliest Cambrian in Iran (A–G; Middle circularis (Hamdi et al., 1989). These brown weathering Dolomite Member, OUMNH AY.202 (field sample number sulphide crystals are up to 10 mm long, tapering towards 87H.7v). Note distinct axial canals in (D and E), a diagnostic each end and seemingly round in cross section. They marker of sponge spicules, compared with similar pentacts from occur within planar laminated black argillites that bear the Atdabanian of Australia (H, I; Wilkawillina Limestone). μ markedly negative carbon isotope values (Brasier et al., 1990; Scale bar for all images, 100 m. anomaly W of Brasier et al., 1996) and barium anomalies (Kimura et al., 1997; Kimura & Watanabe, 2001) that et al., 1996) or Fortunian age (Landing et al., 2007) age. The are characteristic of the Ediacaran-Cambrian transition. Iranian spicules seemingly lie beneath the level of carbon These are of interest because they provide yet another isotope excursion Z within the earliest Cambrian in Siberia, example of abiogenic spicule-like crystals, comparable with which here and elsewhere typically contain examples of the Mongolian, Australian and Chinese structures discussed mollusc-like shells of Purella antiqua and Maikhanella multa, above. of late Nemakit-Dakldynian (Brasier et al., 1996) or middle Fortunian (Landing et al., 2007) age. However, there needs (20) Fossils from various localities in Siberia, to be some consideration of the chemostratigraphy here Lower Cambrian (Gradstein et al., 2012) since there are no defined biozones below the Protohertzina anabarica zone, which is usually defined Records of Cambrian fossil sponges have been extensively in China. Therefore, if Protohertzina extends deeper in time reviewed (Rigby, 1986; Rigby & Hou, 1995; Kouchinsky by a few million years to the W and Z anomaly range in et al., 2005, 2007). Important among these are rich Iran, then this could account for this small discrepancy. assemblages from the early Cambrian of the Siberian In southern China and in Siberia, the first appearance of platform (Fig. 13), the stratigraphy of which is well the Anabarites trisulcatus assemblage occurs somewhat above constrained (see Kochnev & Karlova, 2010). Hexactinellid the negative peak of ‘the BACE’ (e.g. the L1 of Li et al., spicules seen in thin sections from this region are commonly 2009 in China), whereas in Mongolia the assemblage locally placed within the Protospongidae gen. et sp. indet. Such appears at or below the peak (the A peak of Brasier et al., remains have been reported but not illustrated from the 1996). In these same areas, the succeeding Watsonella-Purella level of the Purella antiqua Zone (Terreneuvian, Upper

(A) (B)

Fig. 13. Context for early Cambrian sponge evolution as seen on the Siberian platforms. (A) Chronostratigraphy and chemostratigraphy of the early Cambrian of Siberia (after Brasier et al., 1994a,b). Abbreviations for biozones are as in Fig. 11A. (B) An abrupt change in strontium isotopes at the base of the Tommotian [after Derry et al. (1994) and Brasier et al. (1996)], was followed by a rapid increase in archaeocyathan sponge generic diversity around the globe (after Brasier et al., 1994b), which forms the basis for Early Cambrian biostratigraphy and correlation from the Tommotian to Botoman stages (see Rozanov et al., 1992). This involved innovations in inner and outer wall porosity, here summarised as follows: 1, basal Tommotian, simple outer walls, some with branched pores; simple inner walls; maybe spines and small bracts on inner wall; septa with normal porosity. 2, Upper Tommotian, spines on outer wall. 3, Diaphragm pores, bracts, canals with bracts, or microporous sheath on outer wall; bracts, spines, or annulae on inner wall; simple canals in inner wall; double inner wall. 4, Lower Atdabanian, multiple tumuli on outer wall. 5, Upper Atdabanian, stirrup canals, communicating canals or microporous sheaths on inner wall. 6, Ethmophylloid inner wall. 7, Basal Botoman, clathrate, annular, erbocyathoid, Tylocyathus or Clathricoscinus types of outer wall. 8, Lower Botoman, redimiculi or spongy mass on inner wall.

Fortunian Stage c. 532–529 Ma; Gradstein et al., 2012) in also Rozanov et al., 1969, 1992, 2008; Shabanov et al., 2008; the upper Ust’-Yudoma Formation (e.g. Khomentovsky, Riding & Zhuravlev, 1995; Kruse et al., 1995; Khomentovsky Val’Kov & Karlova, 1990; Khomentovsky & Karlova, 1993) et al., 1990) as well as at coeval levels elsewhere on the and can be observed in thin sections and acid residues from Siberian Platform (Kouchinsky et al., 2012). the lowermost Tommotian strata Nochoroicyathus sunnaginicus Zone (Sokolov & Zhuravleva, 1983; Pel’man et al., 1990; Rozanov & Zhuravlev, 1992; Khomentovsky & Karlova, V. A DISCUSSION OF OUR SEARCH CRITERIA 2002). Demosponge-type spicules have also been reported FOR THE EARLIEST SPONGE FOSSILS from the much younger Botoman Stage (Ivantsov et al., 2005; Ponomarenko et al., 2005). Accurately constraining the origin of sponges is closely linked Higher levels within the Tommotian Stage are suggested to models concerning the origin of animals because of the to contain the earliest known calcareous sponge spicules basal position of sponges in the Metazoa (Peterson et al., 2004; (Kruse, Zhuravlev & James, 1995). The first aspicular Philippe et al., 2009; Pick et al., 2010; Srivastava et al., 2010). calcareous sponges are represented by the porous calcite Unlike many other organisms which sit in a comparably cups of archaeocyathans (Brasier, 1976; Wood, 1998). basal position (e.g. the Placozoa or the Ctenophora), the Archaeocyatha first appear within the Nochoroicyathus sponges have very good preservation potential. Furthermore, sunnaginicus Zone, at the base of the Tommotian Stage their diagnostic characteristics are readily identifiable in (Terreneuvian, Stage 2, c. 525.5–523.5 Ma; Gradstein et al., fossilised forms throughout the whole Phanerozoic, unlike 2012) along the Aldan and Lena Rivers (Fig. 2G, H; see many other organisms that leave an extensive although

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society The early fossil record of sponges 25 potentially phylogenetically ambiguous fossil record, as with fossil may easily also be misinterpreted as an individual as many cnidarian groups. As such, the sponges offer great opposed to a part of a larger organism or microbial structure potential for being the first recognisable animals in the fossil (as in the case of Section IV.7). It is also easy to make strong record. This means that claims for early sponge fossils in deep claims from a small number of samples, whereas it is harder time have abounded over the last decade. Paradoxically, this and more complex to understand ontogenetic dynamics burgeoning of early sponge claims has taken place against and population analyses. Finally, there arguably needs to a backdrop of ever more sceptical treatment of the classic be more awareness of the potential for analysing growth Ediacaran biota and other Ediacaran-age enigmata (see dynamics of a given fossil organism for ascribing sponge for instance Budd & Jensen, 2000; Antcliffe & Brasier, 2007, or other affinities in enigmatic fossil groups (Brasier, 1976; 2008; Brasier & Antcliffe, 2008). Hence, those who have used Antcliffe & Brasier, 2008; Antcliffe, Gooday & Brasier, 2011; molecular clocks to argue for a deep Precambrian ancestry Debrenne et al., 2012). to modern animal groups (e.g. Peterson et al., 2004; Sperling The sponge fossil record of the Lower Cambrian is et al., 2007; Erwin et al., 2011), have become increasingly much clearer and more coherent. Extensive suites of reliant upon sponges as a search criterion (e.g. Sperling et al., spicular material appear across a large Palaeotethyan belt 2010). Our approach compares with the critical analysis stretching from Iran, through India, and into China. of Proterozoic claims of Bilateria (Budd & Jensen, 2000) Comparable material is also found across Siberia. This or the earliest Archaean fossil record (Brasier et al., 2006). stands in stark contrast to the rare and isolated specimens A clear way forward here has been the development of of just a few million years earlier. The difference between a null hypothesis (e.g. Brasier et al., 2002). Following this candidate fossils from 555 to 545 Ma (Section IV.10–18) approach Antcliffe & McLoughlin (2008) argued that when and those from c. 535 Ma (Section IV.19) is profound. analysing enigmatic Proterozoic fossils one should start with By c. 525 Ma, the archaeocyathan sponges were globally the null hypothesis that the fossil is abiogenic until otherwise developed and reef building. Table 3 presents a compilation disproved. It should then be assumed that the fossil is of how we scored each criterion for each possible poriferan microbial (or a microbial colony) until this is disproved. candidate. In almost half of all Precambrian cases (error = An hypothesis of animal affinity should not be the starting type 1 9 candidates of 19; 47%) there was no attempt point when examining any Precambrian fossil. made to demonstrate that the fossils showed any diagnostic A case in point is the evidence for early sponges adduced characters, and there was little consideration of possible from biomarkers (see Section IV.5), in which predictions from alternative hypotheses or the taphonomic processes. In over = molecular divergence date estimates are also involved. If one a third of Precambrian cases (error type 2 7 candidates assumes that both biomarkers and molecular divergence of 19; 37%) useful characters were suggested but obvious dates are correct, it can appear reasonable to interpret an alternative hypotheses (e.g. abiogenic, microbial etc.) and Ediacaran fossil as a sponge. But such an interpretation taphonomy were not considered sufficiently. In several cases = is reliant upon the notion that molecular clocks provide (error type 3 3 candidates of 19; 16%) characters were evidence that is independent of the fossil record and its discussed that were demonstrated to be present in the fossil interpretation. Unfortunately, this assumption is problematic usually through appropriate taphonomic analyses. However, because molecular clocks are probabilistic models used to these characters are not distinctive for sponges and in general join up the known calibration points. These calibration alternative hypotheses were not sufficiently explored. Too points rest upon the veracity of fossil evidence. If critical often claims for Precambrian sponges are arguments from calibration fossils are removed from molecular clock studies authority, with weak circumstantial evidence supporting then the analysis becomes model dependent. This means that confirmation bias, and insufficient hypothesis testing. It is molecular clocks, particularly for basal animal divergences, clear from these results that much more careful testing of can verge on concluding their own assumptions. alternative hypotheses and of the taphonomic process is essential as well as meaningful consideration of what the It is clear from the review of the candidates above that diagnostic characters of sponges actually are. We have tried many examples of putative Precambrian sponge fossils are to suggest in Sections II and III which sponge characters more easily explained by other hypotheses (see Tables 2 may be useful for investigating the early fossil record. and 3). For the most part, previous studies have not tested a series of alternative hypotheses, notably of an origin from abiological causes, and then an origin from microbial/sedimentary processes and interactions. Next, VI. A DISCUSSION CONSTRAINING THE ORIGIN there has been a tendency to select end-members that look OF SPONGES like sponges, rather than analyse the whole population of structures in which they occur. This lack of population The fossil record currently provides a firm constraint of analysis is a major concern, for example, in cases where a the earliest Cambrian c. 535 Ma for the origin of the morphological continuum may exist between structures that Hexactinellida, with specimens from the Middle Dolomite appear sponge-like (e.g. Palaeophragmodictya) and others that Member, Elburz Mountains, Iran. These hexactine spicules are not at all sponge-like (Hiemelora with Primocandelabrum; have axial canals and layered secretion structures, a character e.g. Hofmann et al., 2008). This case also shows how a combination convincingly not produced by abiogenic or

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society 26 Jonathan B. Antcliffe and others or Ernietta not rejected possibly another member of Ediacaran biota. macro-algae. crystals and microbialites. member of the Ediacaran biota. crystals associated with chert nodules. such as mud chips and microbial mat fragments. published images and no repository numbers. compounds produced by marine algae or possibly symbiotic bacteria. calc-phosphate grains. microbialites. Alternative hypotheses Taphomorph of Microbial colony or Abiogenic acicular silicate Taphomorph of a Sedimentary structures Long chain carbon Abiogenic sedimentary Microbial sheaths and hypothesis Insufﬁcient consideration of taphonomy and co-occurring fossils. Invalid consideration of alternative hypotheses. Claimed cellular structures are not demonstrated. Arguments based on invalid biomechanical constraints. and based on temporal correlation with other claims. Compounds known to also be produced by modern marine algae. Invalid rejection of alternative hypotheses. Invalid rejection of alternative hypotheses. Problems with sponge No diagnostic characters. No arguments presented. No data available as no No diagnostic characters. Age (Ma) . 570 No diagnostic characters. Abiogenic acicular silica c (Russia) 555 (Namibia) 556 c. 750 Before c. 760 Between 1025 and 1015 No diagnostic characters. Location and Ust Pinega Fm., White Sea, Russia Nevada Alaska Namibia Siberia, Russia Verkhovka Fm., Russia Between 559 and 555 No diagnostic characters. Noon Day Dolomite, Lakhanda Sequence, (2011) Newfoundland Canada Range 565–555 No diagnostic characters. ¨ orheide (2010) Trezona Fm., Australia Between 664.7 and 570 No diagnostic characters (2012) Otavi Group, Northern (2009) Namibia Between 713 and 570 No diagnostic characters. et al. et al. (1998) Doushantuo Fm., China Between 635and 555 No diagnostic characters. et al. et al. et al. (2007) (2002) (2012) 5Li — German & Podkovyrov 6B Serezhnikova & Ivantsov 6A Sperling 4A Butterﬁeld (2007) Wynniatt Fm., Canada Between 1077and 723 No diagnostic characters. Algal cyst. 6C McMenamin (1998) Nama Group, Namibia 4B, C Brain Sponge candidate Figure Reference fossils ‘cruciform structure’ Ausia Thectardis Fedomia Otavia Jacutianema 11 9 10 7 ‘Bioclasts’ 4E, F Maloof 8 ‘Spicules’ and body 5 ‘Biomarkers’ — Love 6 ‘Spicules’ 4D Allison & Awramik (1989) Upper Tindir Group 4 ‘Spicules’ — Reitner & W 3 Table 2. Summary of rejected Precambrian candidate fossils 1 Acritarch with 2

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society The early fossil record of sponges 27 .Itisan : Coralline algae not rejected Alternative hypotheses crystals associated with acanthomorphic acritarchs. crystals, acicular crystal twins composed of arsenopyrite. produced by pyrite crystal moving through chert or phosphate gel. microbialites. or encrusted algae Spicules: Abiogenic acicular crystals. Primocandelabrum organ not a separate organism. possibly a rhizarian. Namacalathus Hydrothermal abiogenic Prokaryotic colony, Attachment disk of If agglutination is confirmed hypothesis Cryptic ontogeny unmatched in modern sponges. Invalid use of biomineralisation to constrain phylogeny. spicule but crystal twins of acicular crystals. No diagnostic characters (e.g. layered silica, axial canals). ‘spongin filament’ are not evidenced. ‘spicular mesh’ is not mineralised. Insufficient consideration of taphonomy. ‘spicules’ are not mineralised. Problems with sponge No diagnostic characters. Ambient inclusion trails No diagnostic characters. Hexacts are not fused No diagnostic characters. Abiogenic acicular silicate No diagnostic characters. No diagnostic characters. No diagnostic characters. c. 543 c. 549 c. 550 c. 545 c. 555 c. 555 c. 555 Age (Ma) Location Himalaya,India Mongolia Doushantuo Fm., China South Australia Russia Lower Tal Fm., Lesser Nama Group, Namibia Oloom Fm, Gobi-Altay, Dengying Fm., China ¨ orheide (2004) Mezen Fm, Zimnii Bereg, (1993) (1997) (2002) and et al. Reference et al. et al. et al. (1998) (2002) Reitner & W 8 Mazumdar & Banerjee — Wood — Wang & Wang (2011) Miaohe Biota, 6E Gehling & Rigby (1996) Rawnsley Quartzite, 6D Ivantsov ,and structures’ ‘spicules’ Namacalathus Cucullus Sponge candidate Figure Vaveliksia Palaeophragmodictya 16 18 ‘Cruciform 15 ‘Spicules’ —17 Steiner Hexact ‘spicules’ 7 Brasier 14 Table 2. Continued 12 13

Table 3. Summary of candidates against assessment criteria and error types

Sponge candidate Age (Ma) Character Diagnosis Geochron Pass Error typea

1 Acritarch with ‘cruciform structure’ 1077 to 723√ X X √X X 1 2 Jacutianema 1025 to 1015 X X2 3 Otavia c. 760 XXXX1 4 ‘Spicules’ c. 750 XXXX1 5 ‘Biomarkers’ 713 to 570 X X X X 1 6 ‘Spicules’ Before c. 570 X X √X X 1 7 ‘Bioclasts’ 664.7 to 570√ X X X1 8 ‘Spicules’ and body fossils 635 to 555 √√XXX2 9 Thectardis Range 565 to 555 X √ X3 10 Fedomia 559 to 555 √X X √ X1 11 Ausia Range 556 to 555 X √ X2 12 Vaveliksia c. 555 XX√√X1 13 Palaeophragmodictya c. 555 √X √ X3 14 Cucullus c. 555 √ X √ X2 15 ‘Spicules’ c. 550 X √ X2 16 Namacalathus, and ‘spicules’ c. 549 √XX√ X1 17 Hexact ‘spicules’ c. 545 √ X √ X2 18A ‘Cruciform structures’ c. 543 √√X X2 18B Monaxons c.543 √√√√X X3 19 Iranian spicules c. 535 √√√√PASS 20 Siberian spicules c. 532.5 PASS aError type counts: Type 1 = 9 candidates (of 19), 47%; Type 2 = 7 candidates (of 19), 37%; Type 3 = 3 candidates (of 19), 16%. microbial processes. The hexactine morphology of the this has for the origin of animals. It is often assumed that spicules is diagnostic for Hexactinellida and allows confident burrowing is a characteristic of the Bilateria yet there is also assignment to that group. The origin of demosponges is best good evidence that diploblasts, particularly the Cnidaria, marked by the appearance of the archaeocyathans during are capable of extensive vertical burrowing. In fact many the Nochoroicyathus sunnaginicus zone (c. 526 Ma). We suggest cnidarians must burrow deeply to survive (Levin et al., 1994; that the ancestral archaeocyathan sponges must occur in the Levin & Dibacco, 1995), burrowing by peristalsis of the Purella antiqua Zone and would have consisted of small, simple central polyp foot, and this activity has even been found to rounded cups, each provided with a single, weakly calcified produce the basal Cambrian trace fossil marker Treptichnus wall, perforated by simple pores. The evolutionary trends (Phycodes) pedum in Miocene sediments of New Zealand seen in archaeocyathans (Fig. 13B) also lend support to the (Bradley, 1981). As such the very latest Precambrian (c. concept that the fossil record is a reliable witness for the early 542 Ma) to the earliest Cambrian (540 Ma) suite of burrows evolution of sponges. The fossil record shows how these stocks need only represent the origin of total group Eumetazoa. gradually assembled their skeletal framework. This radiation It specifically should not be taken as direct evidence for seems to have been delayed until the explosive radiation of the presence of crown group Bilateria at this time. The other phyla, notably brachiopods and molluscs near the base occurrence of trace fossils such as Treptichnus is considered of the Tommotian (Brasier, 1989b; Rozanov & Varlamov, to be highly reliable evidence of animal behaviour, and 2008). represents a first-appearance datum upon which many could There is currently a time lag between the phylogenetic agree. Further, use of this ichnological datum implies a ghost diversification of the basal Metazoa and the reliable range of only c. 5 Ma for the Porifera based on the earliest appearance of sponges in the fossil record. The base of reliable sponge fossils from Iran. This shows the utility of the Cambrian is defined by the first appearance of traces of searching for the oldest reliable fossil evidence for a particular animal activity [Treptichnus (Phycodes) pedum; Brasier, Cowie group rather than oldest possible evidence. This very small & Taylor (1994); Brasier et al. (1994a,c), at a section proposed ghost range is only 2.1% of the length of the 240 Ma ghost by Narbonne et al. (1987)]. Treptichnus isp. has been reported range implied by Sperling et al. (2010), and less than 1% of from the Ediacaran of Namibia (albeit in much simpler forms) the ghost range implied by Erwin et al. (2011). It also more some 450 m beneath the uppermost fossil attributed to the satisfactorily explains the absence of sponge fossils from a late Ediacaran biota (Grotzinger et al., 1995; Narbonne, Saylor & Precambrian record which abounds in soft-bodied remains in Grotzinger, 1997). This represents animal behaviour during siliciclastic deposits, and beautiful microscopic cellular fossils the late Precambrian before the diversification of behavioural preserved in chemically precipitated silica and phosphate. complexity at the base of the Cambrian (Gehling et al., It may be that future research will yield convincing sponge 2001). So it is pertinent to ask what creature or creatures fossils from Precambrian strata, but the methodology that we made these earliest vertical burrows and what implications have outlined herein will help those describing such fossils to

Biological Reviews (2014) 000–000 © 2014 The Authors. Biological Reviews © 2014 Cambridge Philosophical Society The early fossil record of sponges 29 convince others of their finds. Such a discovery would be very access to fossil material at the OUMNH. We would also like interesting and significant. Yet a latest Precambrian sponge to thank Françoise Debrenne for her hospitality over many would not greatly alter the narrative of major diversifications years and for her tutelage concerning the Archaeocyatha. of Metazoa at the base of the Cambrian, and the gradual Philip Donoghue and Nick Butterfield are thanked for construction of the animal stem lineages throughout the lower constructive discussion concerning this work. Allison Daley Cambrian. An early Neoproterozoic fossil with demonstrable read a late draft of the manuscript and provided critical crown-group sponge characteristics on the other hand, would feedback. Finally, we would like to thank Greg Edgecombe really shake things up and it is a possibility to which we must and Graham Budd for their very constructive reviews which remain open. helped us to greatly improve the manuscript.

VII. CONCLUSIONS IX. REFERENCES

Allison,C.W.&Awramik, S. M. (1989). Organic–walled microfossils from earliest (1) The evidence for sponges from Precambrian strata is Cambrian or latest Proterozoic Tindir Group rocks, northwest Canada. Precambrian circumstantial, often based on confirmation bias, and lacks Research 43, 253–294. testing of valid alternative non-animal hypotheses. Antcliffe, J. B. (2013). Questioning the evidence of organic compounds called sponge biomarkers. Palaeontology 56, 917–925. (2) The Cambrian explosion was an evolutionary event Antcliffe,J.B.&Brasier, M. D. (2007). Charnia and Sea Pens are poles apart. of great magnitude and closely connected to the origin of Journal of the Geological Society of London 164, 49–51. animals. Antcliffe,J.B.&Brasier, M. D. (2008). Charnia at 50: developmental analysis of Ediacaran fronds. Palaeontology 51, 11–26. (3) Hexact spicules from Mongolia at 545 Ma are Antcliffe,J.B.&McLoughlin, N. (2008). Deciphering fossil evidence for abiogenic artefacts. the origin of life and the origin of animals: common challenges in different (4) The earliest reliable sponge fossils are hexact spicules worlds. In Cellular Origins, Life in Extreme Habitats and Astrobiology: From Fossils to Astrobiology (Records of Life on Earth and the Search for Extraterrestrial Biosignatures from Iran dated to c. 535 Ma. (eds J. Seckbach and M. Walsh), pp. 211–229. Springer, Dordrecht. (5) The earliest reliable demosponges are archaeocy- Antcliffe,J.B.,Gooday,A.&Brasier, M. D. (2011). Testing the protozoan athans from Siberia c. 523.5–525.5 Ma. hypothesis for Ediacaran fossils: a developmental analysis of Palaeopascichnus. Palaeontology 54, 1157–1175. (6) By c. 525 Ma a variety of sponge fossils are widespread Azmi,R.J.&Paul, S. K. (2004). Discovery of Precambrian–Cambrian boundary across the Siberian platform. protoconodonts from the Gangolihat Dolomite of Inner Kumaun Lesser Himalaya: (7) Future work should report headline dates for fossils implication on age and correlation. Current Science 86, 1653–1660. Banerjee,D.M.,Schidlowski,M.,Siebert,F.&Brasier, M. D. (1997). at geochronological minimum constraints so as to not to Geochemical changes across the Proterozoic–Cambrian transition in the Durmala overstate the case and to lessen error in other analyses which phosphorite mine section, Mussoorie Hills, Garwhal Himalaya, India. Palaeogeography, use these dates. Palaeoclimatology, Palaeoecology 132, 183–194. Bengtson, S. (1985). Taxonomy of disarticulated fossils. Journal of Paleontology 59, (8) It is important to consider which characters are 1350–1358. diagnostic for sponges in deep time. These may be Bengtson, S. (1992). The cap-shaped Cambrian fossil Maikhanella and the relationship apomophies of any sponge clade rather than synapomorphies between coeloscleritophorans and molluscs. Lethaia 25, 401–420. Bengtson,S.,Conway Morris,S.,Cooper,B.J.,Jell,P.A.&Runnegar,B. of a major sponge group. N. (1990). Early Cambrian fossils from South Australia. Memoir of the Association of (9) A debate should begin about which diagnostic Australasian Palaeontologists 9, 1–364. characters have the best potential to be fossilised and what Benton,M.,Donoghue,P.C.J.&Asher, R. J. (2009). Calibrating and constraining molecular clocks. In Hedges,S.B.andKumar,S.(eds).The Timetree of Life,Oxford tests need to be done on candidate fossils to distinguish them University Press, Oxford, 35–86, 576 pp. from abiological processes that could produce putatively Blumenberg,M.,Thiel,V.,Pape,T.&Michaelis, W. (2002). The steroids of similar structures. hexactinellid sponges. Naturwissenschaften 89, 415–419. Borchiellini,C.,Manuel,M.,Alivon,E.,Boury-Esnault,N.,Vacelet,J. & Le Parco, Y. (2001). Sponge paraphyly and the origin of Metazoa. Journal of Evolutionary Biology 14, 171–179. Botting,J.P.&Butterfield, N. J. (2005). Reconstructing early sponge relationships VIII. ACKNOWLEDGEMENTS by using the Burgess Shale fossil Eiffelia globosa, Walcott. Proceedings of the National Academy of Sciences 102, 1554–1559. Botting,J.P.,Muir,L.A.,Xiao,S.,Li,S.&Lin, J.-P. (2012). Evidence for spicule We gratefully acknowledge the following funding bodies for homology in calcareous and siliceous sponges: biminerallic spicules in Lenica sp. from the Early Cambrian of South China. Lethaia 45, 463–475. their support: The Royal Commission for the Exhibition of Bottjer,D.J.,Hagadorn,J.W.&Dornbos, S. Q. (2000). The Cambrian substrate 1851 Research Fellowship and a Palaeontology Association revolution. GSA Today 10, 1–7. research grant (both to J.B.A.) as well as the support of Boury-Esnault,N.&Rutzler, K. (1997). Thesaurus of sponge morphology. Smithsonian Contributions to Zoology 596, 55 pp., 305 figures. the Oxford University Department of Zoology and the Bowring,S.,Grotzinger,J.P.,Condon,D.J.,Ramezani,J.,Newall,M. Department of Earth Sciences; NERC (grants to MDB J. & Allen, P. A. (2007). Geochronologic constrains on the chronostratigraphic and RHTC). Norman Charnley is acknowledged for his framework of the Neoproterozoic Huqf Supergroup, Sultanate of Oman. American Journal of Science 307, 1097–1145. expertise and patience with SEM-EDX analyses. Owen Bradley, J. (1981). Radionereites, Chondrites and Phycodes: trace fossils of anthoptiloid sea Green is acknowledged for his extensive support in the lab pens. Pacific Geology 15, 1–16. over many years. We would also like to thank Paul Taylor Brain,C.K.,Prave,A.R.,Hoffmann,K.-H.,Fallick,A.E.,Botha,A.,Herd, D. A., Sturrock,C.,Young,I.,Condon,D.J.&Allison, S. G. (2012). The first and Allison Daley at the NHM-London for help with access animals: ca. 760–million–year–old sponge-like fossils from Namibia. South African to fossils. We would like to thank Derek Siveter for help with Journal of Science 108, Article no. # 658, 8 pp. (doi: 10.4102/sajs.v108i1/2.658).

Branch,G.M.,Griffiths,C.L.,Branch,M.L.&Beckley, L. E. (2010). Two Butterfield, N. J. (2007). Macroevolution and macroecology through deep time. Oceans: A Guide to the Marine Life of Southern Africa. Random House Struik Publishers, Palaeontology 50, 41–55. Johannesburg. Butterfield, N. J. (2009). Modes of pre-Ediacaran multicellularity. Precambrian Brasier, M. D. (1976). Early Cambrian intergrowths of archaeocyathids, Renalcis,and Research 173, 201–211. pseudostromatolites from South Australia. Palaeontology 19, 223–245. Butterfield,N.J.,Knoll,A.H.&Swett, K. (1994). Paleobiology of the Brasier,M.D.(1989a). China and the palaeotethyan belt (India, Pakistan, Iran, Neoproterozoic Svanberg-fjellet Formation, Spitsbergen. Fossils & Strata 34, 1–84. Kazakhstan, and Mongolia). In The Precambrian-Cambrian Boundary, Oxford Monographs Callow,R.H.T.&Brasier, M. D. (2009a). A solution to Darwin’s dilemma of 1859: in Geology and Geophysics Number 12 (eds J. W. Cowie andM.D.Brasier), pp. 41–74. exceptional preservation in Salter’s material from the late Ediacaran Longmyndian Clarendon Press, Oxford. Supergroup, England. Journal of the Geological Society (London) 166, 1–4. Brasier, M. D. (1989b). Towards a biostratigraphy of the earliest skeletal biotas. In Callow,R.H.T.&Brasier,M.D.(2009b). The remarkable preservation of The Precambrian-Cambrian Boundary, Oxford Monographs in Geology and Geophysics Number microbial mats in Neoproterozoic siliciclastic settings: implications for Ediacaran 12 (eds J. W. Cowie and M. D. Brasier), pp. 117–165. Clarendon Press, Oxford. taphonomic models. Earth Science Reviews 96, 207–219. Brasier, M. D. (1990a). Nutrients in the early Cambrian. Nature 347, 521–522. Cavalier-Smith, T. (2002). The phagotrophic origin eukaryotes and phylogenetic Brasier, M. D. (1990b). Phosphogenic events and skeletal preservation across the classification of Protozoa. International Journal of Systematic and Evolutionary Microbiology Precambrian-Cambrian boundary interval. In Phosphorite Research and Development, 52, 297–354. Special Publication of the Geological Society of London 52 (eds A. G. Norholt and I. Cavalier-Smith, T. (2006). Cell evolution and Earth history: stasis and revolution. Jarvis), pp. 289–303. Geological Society of London, London. Philosophical Transactions of the Royal Society B 361, 969–1006. Brasier, M. D. (1991). Nutrient flux and the evolutionary explosion across the Cerrano,C.,Calcinai,B.,di Camillo,C.G.,Valisano,L.&Bavestrello, Precambrian-Cambrian boundary interval. Historical Biology 5, 85–93. G. (2007). How and why do sponges incorporate foreign material? Strategies in Brasier, M. D. (1992). Nutrient-enriched waters (NEW) and the early skeletal fossil Porifera. In Porifera Research: Biodiversity, Innovation, Sustainability (eds M. R. Custodio´ , record. Journal of the Geological Society (London) 149, 621–629. G. Lobo-Hajduˆ ,E.Hajdu and G. Muricy), pp. 239–246. Serie´ Livros 28 Museu Brasier,M.D.&Antcliffe, J. B. (2008). Dickinsonia from Ediacara: a new look Nacional, Rio de Janeiro. at morphology and body construction. Palaeogeography, Palaeoclimatology, Palaeoecology Chen,M.E.,Xiao,Z.Z.&Yuan, X. L. (1994). A new assemblage of megafossils: 270,311–323. mioahe biota from Upper Sinian Doushantuo Formation, Yangtze Gorges. Acta Brasier,M.D.&Antcliffe, J. B. (2009). Evolutionary relationships within the Palaeontologica Sinica 33, 391–403. Avalonian Ediacara biota: new insights from laser analysis. Journal of the Geological Clapham,M.E.,Narbonne,G.M.,Gehling,J.G.,Greentree,C.&Anderson, Society of London 166, 363–384. M. M. (2004). Thectardis avalonensis: a new Ediacaran fossil from the Mistaken Point Brasier,M.D.& Singh, P. (1987). Microfossils and Precambrian–Cambrian biota, Newfoundland. Journal of Paleontology 78, 1031–1036. boundary stratigraphy at Maldeota, Lesser Himalaya. Geological Magazine 124, Cohen,P.A.,Schopf,J.W.,Butterfield,N.J.,Kudryavtsev,A.B.& 323–345. Macdonald, F. A. (2011). Phosphate biomineralisation in mid-Neoproterozoic Brasier,M.D.&Sukhov, S. (1998). The falling amplitude of carbon isotopic protists. Geology 39, 539–542. oscillations through the Lower to Middle Cambrian: northern Siberian data. Condon,D.,Zhu,M.,Bowring,S.,Wang,W.,Yang,A.&Jin, Y. (2005). U-Pb Canadian Journal of Earth Sciences 35, 353–373. ages from the Neoproterozoic Doushantuo Formation, China. Science 308,95–98. Brasier,M.D.,Magaritz,M.,Corfield,R.,Luo,H.,Wu,X.,Ouyang,L., Corsetti,F.A.&Kaufman, A. J. (2005). The relationship between the Jiang,Z.,Hamdi,B.,He,T.&Frasier, A. G. (1990). The carbon- and oxygen- Neoproterozoic Noonday Dolomite and the Ibex Formation: new observations isotope record of the Precambrian-Cambrian boundary interval in China and Iran and their bearing on snowball Earth. Earth Science Reviews 73, 63–78. and their correlation. Geological Magazine 127, 319–332. Cowie,J.W.&Brasier, M. D. (1989). The Precambrian – Cambrian Boundary.Clarendon Brasier,M.D.,Corfield,R.M.,Derry,L.A.,Rozanov,A.Y.&Zhuravlev, Press, Oxford. A. Y. (1994a). Multiple δ13C excursions spanning the Cambrian explosion to the Daley,A.C.,Budd,G.E.,Caron, J.-B., Edgecombe,G.D.&Collins, D. (2009). Botoman crisis in Siberia. Geology 22, 455–458. The Burgess Shale anomalocaridid Hurdia and its significance for early euarthropod Brasier,M.D.,Cowie,J.W.&Taylor,M.E.(1994b). Decision on the Precambrian- evolution. Science 323, 1597–1600. Cambrian boundary stratotype. Episodes 17, 3–8. Debrenne,F.&Reitner, J. (2001). Sponges, cnidarians, and ctenophores. In The Brasier,M.D.,Rozanov,A.Y.,Zhuravlev,A.Y.,Corfield,R.M.&Derry, Ecology of the Cambrian Radiation (eds A. Y. Zhuravlev and R. Riding), pp. 301–325. L. A. (1994c). A carbon isotope reference scale for the Lower Cambrian succession Columbia University Press, New York. in Siberia: report of IGCP Project 303. Geological Magazine 131, 767–783. Debrenne,F.&Vacelet, J. (1984). Archaeocyatha: is the sponge model consistent Brasier,M.D.,Shields,G.,Kuleshov,V.&Zhegallo, E. A. (1996). Integrated with their structural organisation? Palaeontographica Americana 54, 358–369, 2 pl., 3 chemo- and biostratigraphic correlation of early animal evolution: neoproterozoic- Early Cambrian of southwest Mongolia. Geological Magazine 133, 445–485. tables. Brasier,M.D.,Green,O.R.&Shields, G. (1997). Ediacaran sponge spicule Debrenne,F.&Zhuravlev, A. Y. (1994). Archaeocyathan affinities: how deep can clusters from south-western Mongolia and the origins of the Cambrian fauna. we go into the systematic affiliation of an extinct group?. In Sponges in Time and Space Geology 25, 303–306. (eds R. W. M. van Soest,T.M.G.van Kempen and J. C. Braekman), pp. 3–12. Brasier,M.D.,McCarron,G.,Tucker,R.,Leather,J.,Allen,P.A.&Shields, A. A. Balkema, Rotterdam. G. (2000). New U-Pb zircon dates for the Neoproterozoic Gubrah glaciation and Debrenne, F., Zhuravlev,A.Y.&Kruse, P. D. (2012a). Part E, revised, volume 4, for the top of the Huqf Supergroup, Oman. Geology 28, 175–178. Chapter 18: general features of the Archaeocyatha. Treatise Online 38, 1–102, 34 fig. Brasier,M.D.,Green,O.R.,Jephcoat,A.P.,Kleppe,A.K.,Van Kranendonk, Debrenne, F., Zhuravlev,A.Y.&Kruse, P. D. (2012b). Part E, revised, volume 4, M. J., Lindsay,J.F.,Steele,A.&Grassineau, N. V. (2002). Questioning the Chapter 19: systematic descriptions: Archaeocyatha. Treatise Online 50, 1–186, 134 evidence for Earth’s oldest fossils. Nature 416,76–81. fig. Brasier,M.D.,McLoughlin,N.,Green,O.&Wacey, D. (2006). A fresh look Derry,L.A.,Brasier,M.D.,Corfield,R.,Rozanov,A.Y.&Zhuravlev,A.Y. at the fossil evidence for early Archaean cellular life. In Major Steps in Cell Evolution: (1994). Sr and C isotopes in Lower Cambrian carbonates from the Siberian craton: Palaeontological, Molecular and Cellular Evidence of their Timing and Global Effects, Philosophical a palaeoenvironmental record during the ‘Cambrian excursion’. Earth and Planetary Transactions of the Royal Society, Series B, Volume 361 (eds T. Cavalier-Smith,M.D. Science Letters 128, 671–681. Brasier and T. M. Embley), pp. 887–902, London. Dohrmann,M.,Janussen,D.,Reitner,J.,Collins,A.G.&Worheide¨ , G. (2008). Brasier,M.D.,Antcliffe,J.B.&Callow, R. (2011). Evolutionary trends in Phylogeny and evolution of glass sponges (Porifera, Hexactinellida). Systematic Biology remarkable fossil preservation across the Ediacaran–Cambrian transition and the 57, 388–405. impact of metazoan mixing. In Taphonomy: Bias and Process Through Time (eds P. A. Donoghue,P.C.J.&Antcliffe, J. B. (2010). The origins of multicellularity. Nature Allison and D. Bottjer), pp. 519–567. Springer, Dordrecht. 466, 41–42. Braun,A.,Chen,J.,Waloszek,D.&Maas, A. (2007). First early Cambrian El Albani,A.,Bengtson,S.,Canfield,D.E.,Bekker,A.,Macchiarelli,R., Radiolaria. In The Rise and Fall of the Ediacara Biota, Geological Society Special Publications Mazurier,A.,Hammarlund,E.U.,Boulvais,P.,Dupuy,J.J.,Fontaine, 286 (eds P. Vickers-Rich and P. Kamarower), pp. 143–149. Geological Society C., Fursich,F.T.,Gauthier-Lafaye, F., Janvier,P.,Javaux,E.,Ossa, of London, London. F. O., Pierson-Wickmann,A.-C.,Riboulleau,A.,Sardini,P.,Vachard, Brocks,J.J.&Butterfield, N. J. (2009). Early animals out in the cold. Nature 457, D., Whitehouse,M.&Meunier, A. (2010). Large colonial organisms with 672–673. coordinated growth in oxygenated environments 2.1 Gyr ago. Nature 466, 100–104. Budd, G. E. (2008). The earliest fossil record of the animals and its significance. Erwin,D.H.,Laflamme,M.,Tweedt,S.M.,Sperling,E.A.,Pisani,D.& Philosophical Transactions of the Royal Society B 363, 1425–1434. Peterson, K. J. (2011). The cambrian conundrum: early divergence and later Budd,G.E.&Jensen, S. (2000). A critical reappraisal of the fossil record of the ecological success in the early history of animals. Science 304, 1091–1097. bilaterian phyla. Biological Reviews 75, 253–295. Fedonkin, M. A. (1996). Ausia as an ancestor of archeocyathans, and other sponge- Butterfield, N. J. (2005). Reconstructing a complex early Neoproterozoic eukaryote, like organisms. In Enigmatic Organisms in Phylogeny and Evolution, pp. 90–91. Moscow, Wynniatt Formation, arctic Canada. Lethaia 38, 155–169. Paleontological Institute, Russian Academy of Sciences, Abstracts.

Fedonkin,M.A.&Waggoner, B. M. (1997). The late Precambrian fossil Kimberella Hoffmann,K.H.,Condon,D.J.,Bowring,S.A.&Crowley, J. L. (2004). U–Pb is a mollusc like bilatarian organism. Nature 388, 868–871. zircon date from the Neoproterozoic Ghaub Formation, Namibia: constraints on Fedonkin,M.A.,Vickers-Rich,P.,Swalla,B.J.,Trusler,P.&Hall,M. Marinoan glaciation. Geology 32, 817–820. (2012). A new metazoan from the Vendian of the White Sea, Russia, with possible Hofmann,H.J.,O’Brien,S.J.&King, A. F. (2008). Ediacaran biota on Bonavista affinities to the ascidians. Palaeontological Journal 46, 1–11. Peninsula, Newfoundland, Canada. Journal of Paleontology 82, 1–36. Finks,R.M.&Rigby, J. K. (2004). Paleozoic demosponges. In Treatise on Invertebrate Hooper,J.N.A.&Van Soest, R. W. M. (2002). Systema Porifera: A Guide to the Paleontology, Part E (Revised), Porifera 3 (ed. R. L. Kaesler). The Geological Society of Classification of Sponges. Plenum, New York. America and University of Kansas, Boulder, Lawrence. Ivantsov,A.Y.,Malahovskaya,Y.E.&Serezhnikova, E. A. (2004). Finks,R.M.,Reid,R.E.H.&Rigby, J. K. (2003). Treatise on Invertebrate Palaeontology, Some problematic fossils from the Vendian of Southeastern White Sea Region. Part E (Revised), Volume 2: Introduction to the Porifera. The Geological Society of America Paleontological Journal 38, 1–9 (translated from Paleontologicheskij Zhurnal 2004(1),3–9). and The University of Kansas, Boulder and Lawrence. Ivantsov,A.Y.,Zhuravlev,A.Y.,Leguta,A.V.,Krassilov,V.A.,Melnikova, Gandin,A.&Debrenne, F. (2010). Distribution of the archaeocyath-calcimicrobial L. M. & Ushatinskaya, G. T. (2005). Palaeoecology of the Early Cambrian Sinsk bioconstructionson the Early Cambrian shelves. Palaeoworld 19, 222–241. biota from the Siberian Platform. Palaeogeography, Palaeoclimatology, Palaeoecology 220, Gehling, J. G. (1991). The case for the Ediacaran fossil roots to the metazoan tree. 69–88. Geological Society of India Memoire 20, 181–224. Jenner, R. A. (2006). Unburdening evo-devo: ancestral attractions, models organisms, Gehling, J. G. (1999). Microbial mats in terminal Proterozoic siliciclastics: ediacaran and basal baloney. Development Genes and Evolution 216, 385–394. death masks. Palaios 14,40–57. Kaufman,A.J.,Knoll,A.H.&Awramik, S. M. (1992). Biostratigraphic and Gehling,J.G.&Rigby, J. K. (1996). Long expected sponges from the Neoproterozoic chemostratigraphic correlation of Neoproterozoic sedimentary successions: upper Ediacara fauna of South Australia. Journal of Paleontology 70, 185–195. Tindir Group, northwestern Canada, as a test case. Geology 20, 181–185. Gehling,J.G.,Jensen,S.,Droser,M.L.,Myrow,P.M.&Narbonne,G.M. Kaufman,A.J.,Ganqing,J.,Christie-Blick,N.,Banerjee,D.M.&Rhai,V. (2001). Burrowing below the basal Cambrian GSSP, Fortune Head, Newfoundland. (2007). Stable isotope record of the Neoproterozoic Krol platform in the Lesser Geological Magazine 138, 213–218. Himalaya of northern India. Precambrian Research 147, 156–185. German,T.N.&Podkovyrov, V. N. (2012). Records of a new spongelike group in Kemp, T. S. (1999). Fossils and Evolution. Oxford University Press, Oxford. the Riphean biota. Paleontological Journal 46, 219–227. Kemp, T. S. (2004). The Origin and Evolution of Mammals. Oxford University Press, Glaessner, M. F. (1984). The Dawn of Animal Life: A Biohistorical Study. Cambridge Oxford. University Press, Cambridge. Khomentovsky,V.V.&Karlova, G. A. (1993). Biostratigraphy of the Vendian- Gradstein,F.M.,Ogg,J.G.&Smith, A. G. (2004). A Geologic Timescale 2004. Cambrian beds and the lower Cambrian boundary in Siberia. Geological Magazine Cambridge University Press, Cambridge. 130, 29–45. Gradstein,F.M.,Ogg,J.G.,Schmitz,M.D.&Ogg, G. M. (2012). The Geologic Khomentovsky,V.V.&Karlova, G. A. (2002). Granitsa Nemakit-Daldynskogo Timescale 2012. Elsevier, Oxford. i Tommotskoga yarusov (Vend-Kembrij) Sibiri. [The boundary between Nemakit- Grazhdankin, D. (2004). Patterns of distributions in the Ediacaran biotas: facies Daldynian and Tommotian stages (Vendian–Cambrian) of Siberia.]. Stratigraphy and versus biogeography and evolution. Paleobiology 30, 203–221. Geological Correlation 10, 13–34. Grazhdankin,D.&Gerdes, G. (2007). Ediacaran microbial colonies. Lethaia 40, Khomentovsky,V.V.,Val’Kov,A.K.&Karlova, G. A. (1990). Novye dannye po 201–210. biostratigrafii perekhodnykh vend-kembrijskikh sloev v bassejne srednego techeniya Grazhdankin,D.&Seilacher, A. (2002). Underground Vendobionta from r. Aldan. [New data on the biostratigraphy of transitional Vendian–Cambrian Namibia. Palaeontology 45, 57–78. strata in the middle reaches of the River Aldan.]. In Pozdnij dokembrij i rannijpaleozoj Grazhdankin,D.V.,Podkovyrov,V.N.&Maslov, A. V. (2005). Paleoclimatic Sibiri. Voprosyregional’noj stratigrafii (eds V. V. Khomentovsky and A. S. Gibsher), environments of the formation of Upper Vendian rocks on the Belomorian–Kuloi pp. 3–57. Institut Geologii i Geofiziki, Sibirskoe Otdelenie, Akademiya nauk SSSR, Plateau, Southeastern White Sea region. Lithology and Mineral Resources 40, 232–244 Siberian Branch, Novosibirsk (in Russian). (Translated from Litologiya i Poleznye Iskopaemye 3, 267–280). Kimura,H.&Watanabe, Y. (2001). Oceanic anoxia at the Precambrian-Cambrian Grotzinger,J.P.,Bowring,S.A.,Saylor,B.Z.&Kaufman, A. J. (1995). boundary. Geology 29, 995–998. Biostratigraphic and chronostratigraphic constraints on early animal evolution. Kimura,H.,Matsumoto,R.,Kakuwa,Y.,Hamdi,B.&Zibaseresht, H. (1997). Science 270, 598–604. The Vendian-Cambrian δ13C record, North Iran: evidence for overturning of the Grotzinger,J.P.,Watters,W.A.&Knoll, A. H. (2000). Calcified metazoans in ocean before the Cambrian explosion. Earth and Planetary Science Letters 147,1–7. thrombolite-stromatolite reefs of the terminal Proterozoic Nama Group, Namibia. King, R. J. (2002). Arsenopyrite. Geology Today 16, 72–75. Paleobiology 26, 334–359. Knoll, A. H. (2011). The multiple origins of complex multicellularity. Annual Review of Grotzinger,J.P.,Adams,E.&Schroder¨ , S. (2005). Microbial-metazoan reefs Earth and Planetary Sciences 39, 217–239. of the terminal Proterozoic Nama Group (ca. 550–543 Ma), Namibia. Geological Knoll,A.H.,Walter,M.R.,Narbonne,G.M.&Christie-Blick, N. (2004). A Magazine 142, 499–517. new period for the geologic time scale. Science 305, 621–622. Hahn,G.&Pflug, H. D. (1985). Polypenartige organismen aus dem Jung- Kochnev,B.B.&Karlova, G. A. (2010). New data on biostratigraphy of the Prakambrium¨ (Nama-Gruppe) von Namibia. Geologica et Palaeontologica 19, 1–13. Vendian Nemakit-Daldynian stage in the southern Siberian platform. Stratigraphy Halanych, K. M. (2004). The new view of animal phylogeny. Annual Review of Ecology, and Geological Correlation 18, 492–504. Evolution, and Systematics 35, 229–256. Kouchinsky,A.,Bengtson,S.,Pavlov,V.,Runnegar,B.,Val’Kov,A.& Halverson,G.N.,Hoffman, P. F., Schrag,D.P.,Maloof,A.C.&Rice,A. Young, E. (2005). Pre-Tommotian age of the lower Pestrotsvet Formation in H. N. (2005). Toward a Neoproterozic composite carbon-isotope record. Gelogical the Selinde section on the Siberian platform: carbon isotopic evidence. Geological Society of America Bulletin 117, 1181–1207. Magazine 142, 319–325. Hamdi,B.,Brasier,M.D.&Jiang, Z. (1989). Earliest skeletal fossils from Kouchinsky,A.,Bengtson,S.,Pavlov,V.,Runnegar,B.,Torssander, Precambrian-Cambrian boundary strata, Elburz Mountains, Iran. Geological Magazine P., Young,E.&Ziegler, K. (2007). Carbon isotope stratigraphy of the 126, 283–289. Precambrian–Cambrian Sukharikha River section, northwestern Siberian platform. Hammel,J.U.,Filatov,M.V.,Herzen,J.,Beckmann, F., Kaandorp,J.A. Geological Magazine 144, 1–10. & Nickel, M. (2011). The non-hierarchical, non-uniformly branching topology Kouchinsky,A.,Bengtson,S.,Runnegar,B.,Skovsted,C.,Steiner,M.& of a leuconoid sponge aquiferous system revealed by 3D reconstruction and Vendrasco, M. (2012). Chronology of early Cambrian biomineralisation. Geological morphometrics using corrosion casting and X-ray microtomography. Acta Zoologica Magazine 149, 221–251. 93, 160–170. Kruse,P.D.,Zhuravlev,A.Y.&James, N. P. (1995). Primordial metazoan- Harvey, T. H. P. (2010). Carbonaceous preservation of Cambrian hexactinellid calcimicrobial reefs: tommotian (Early Cambrian) of the Siberian platform. Palaios sponge spicules. Biology Letters 6, 834–837. 10, 291–321. Harvey,T.H.P.,Velez´ ,M.I.&Butterfield, N. J. (2012). Exceptionally preserved Landing,E.,Peng,S.,Babcock,L.E.,Geyer,G.&Moczydlowska-Vidal, crustaceans from western Canada reveal a cryptic Cambrian radiation. Proceedings of M. (2007). Global standard names for the lowermost Cambrian series and stage. the National Academy of Sciences of the USA 109, 1589–1594. Episodes 30, 287–289. Hill, D. (1972). Archaeocyatha. In Treatise on Invertebrate Paleontology, Part E, Revised Landing,E.,Geyer,G.,Brasier,M.D.&Bowring, S. (2013). Cambrian (Volume 1,ed.C.Teichert), p. 158 p., 107 fig., 3 tables. The Geological Society evolutionary radiation: context, correlation, and chronostratigraphy—overcoming of America & The University of Kansas, Boulder & Lawrence. deficiencies of the first appearance datum (FAD) concept. Earth Science Reviews 123, Hinz, I. (1987). The Lower Cambrian microfauna of Comley and Rushton, 133–172. Shropshire/England. Palaeontographica A 198, 41–100. Levin,L.A.&Dibacco, C. (1995). Influence of sediment transport on short-term Hod,I.M.,Schouten,S.,Jelleman,J.&Sinninghe Damste, J. S. (1999). recolonization by seamount infauna. Marine Ecology Progress Series 123, 163–175. Origin of free and bound mid-chain methyl alkanes in oils, bitumens and kerogens Levin,L.A.,Leithold,E.L.,Gross, T. F., Huggett,C.L.&Dibacco,C. of the marine, Infracambrian Huqf Formation (Oman). Organic Geochemistry 30, (1994). Contrasting effects of substrate mobility on infaunal assemblages inhabiting 1411–1428. two high-energy settings on Fieberling Guyot. Journal of Marine Research 52, 489–522.

Li, C.-W., Chen,J.-Y.&Hua, T.-E. (1998). Precambrian sponges with cellular Peterson,K.J.,Lyons,J.B.,Nowak,K.S.,Takacs,C.M.,Wargo,M.J.& structures. Science 279, 879–882. McPeek, M. A. (2004). Estimating metazoan divergence times with a molecular Li,D.,Ling, H. F., Jiang,S.Y.,Pan,J.Y.,Chen, Y. Q., Cai,Y.F.&Feng,H. clock. Proceedings of the National Academy of Sciences 101, 6536–6541. Z. (2009). New carbon isotope stratigraphy of the Ediacaran–Cambrian boundary Philippe,H.,Derelle,R.,Lopez,P.,Pick,K.,Borchiellini,C.,Boury- interval from SW China: implications for global correlation. Geological Magazine 146, Esnault,N.,Vacelet,J.,Renard,E.,Houliston,E.,Queinnec´ ,E.,Da 465–484. Silva,C.,Wincker,P.,Le Guyader,H.,Leys,S.,Jackson,D.J.,Schreiber, Liu,A.G.,McIlroy,D.,Antcliffe,J.B.&Brasier, M. D. (2011). Effaced F., Erpenbeck,D.,Morgenstern,B.,Worheide¨ ,G.&Manuel, M. (2009). preservation in the Ediacara biota and its implications for the early macrofossil Phylogenomics revives traditional views on deep animal relationships. Current Biology record. Palaeontology 54, 607–630. 19, 706–712. Logan,G.A.,Hayes,J.M.,Hieshima,G.B.&Summons, R. E. (1995). Terminal Philippe,H.,Brinkmann,H.,Lavrov,D.V.,Littlewood,D.T.J.,Manuel,M., Proterozoic reorganization of biogeochemical cycles. Nature 376, 53–56. Worheide¨ ,G.&Baurain, D. (2011). Resolving difficult phylogenetic questions: Love,G.D.,Grosjean,E.,Stalvies,C.,Fike,D.A.,Grotzinger,J.P.,Bradley, why more sequences are not enough. PLoS Biology 9, e1000602. A. S., Kelly,A.E.,Bhatia,M.,Meredith,W.,Snape,C.E.,Bowring,S.A., Pick,K.S.,Philippe,H.,Schreiber, F., Erpenbeck,D.,Jackson,D.J.,Wrede, Condon,D.J.&Summons, R. E. (2009). Fossil steroids record the appearance of P., Wiens,M.,Alie´,A.,Morgenstern,B.,Manuel,M.&Worheide¨ ,G. Demospongiae during the Cryogenian period. Nature 457, 718–721. (2010). Improved phylogenomic taxon sampling noticeably affects nonbilaterian Macdonald,F.A.,Cohen,P.A.,Dudas,F.O.&Schrag,D.P.(2010).Early relationships. Molecular Biology and Evolution 27, 1983–1987. Neoproterozoic scale microfossils in the lower Tindir Group of Alaska and the Ponomarenko,A.G.,Ivantsov,A.Y.,Zhuravlev,A.Y.,Krasilov,V.A., Yukon Territory. Geology 38, 143–146. Ushatinskaya,G.T.&Malakhovskaya, Y. Y. (2005). Unikal’nye sinskiye Maliva,R.G.,Knoll,A.H.&Siever, R. (1989). Secular change in chert mestonakhozhdeniya rannekembriyskikh organizmov. [Unique Sinian lagerstatten¨ distribution: a reflection of evolving biological participation in the silica cycle. Palaios of Early Cambrian organisms.]. Trudy Paleontologicheskogo Instituta 284, 140 pp. 4, 519–532. Porter, S. M. (2007). Seawater chemistry and early carbonate biomineralization. Maliva,R.G.,Knoll,A.H.&Simonson, B. M. (2005). Secular change in the Science 316, 1302. Precambrian silicacycle: insights from chert petrology. Geological Society of America Porter, S. M. (2010). Calcite and aragonite seas and the de novo acquisition of Bulletin 117, 835–845. carbonate skeletons. Geobiology 8, 256–277. Maloof,A.C.,Rose,C.V.,Beach,R.,Samuels,B.M.,Calmet,C.C.,Erwin, Porter,S.M.,Meisterfeld,R.&Knoll, A. H. (2003). Vase-shaped microfossils D. H., Poirier,G.R.,Yao,N.&Simons, F. J. (2010). Possible animal-body fossils from the Neoproterozoic Chuar Group, Grand Canyon: a classification guided by in pre-Marinoan limestones from South Australia. Nature Geoscience 3, 653–659. modern testate amoebae. Journal of Paleontology 77, 409–429. Marshall, C. R. (2006). Explaining the ‘‘Cambrian Explosion’’ of animals. Annual Reitner,J.&Worheide¨ , G. (2002). Non-Lithistid fossil Demospongiae-origins of Review of Earth and Planetary Science Letters 34, 355–384. their palaeobiodiversity and highlights in history of preservation. In Systema Porifera: A Guide to the Classification of Sponges Mazumdar,A.&Banerjee, D. M. (1998). Siliceous sponge spicules in the Early (eds J. N. A. Hooper and R. Van Soest), pp. Cambrian chert-phosphate member of the Lower Tal Formation, Krol belt, Lesser 52–68. Kluwer, New York. ,J., ,N.-V.& , G. (2010). Advances in Stromatolite Geobiology, Lecture Himalaya. Geology 26, 899–902. Reitner Queric Arp Notes in Earth Sciences. Springer-Verlag, Berlin, Heidelberg. McCaffrey,M.,Moldowan,J.M.,Lipton,P.,Summons,R.E.,Peters,K.E., Riding,R.&Zhuravlev, A. Y. (1995). Structure and diversity of oldest sponge- Jegnathan,A.&Watt, D. S. (1994). Paleoenvironmental implications of novel microbe reefs: lower Cambrian, Aldan River, Siberia. Geology 23, 649–652. CM steranes in Precambrian to Cenozoic age petroleum and bitumen. Cosmochimica Rigby, J. K. (1986). Sponges of the Burgess Shale (Middle Cambrian), British et Geochimica acta 58, 529–532. Columbia. Palaeontographica Canadiana 2, 1–105. McIlroy,D.&Logan, G. A. (1999). The impact of bioturbation on infaunal ecology Rigby,J.K.&Hou, X. (1995). Lower Cambrian demosponges and hexactinellid and evolution during the Proterozoic-Cambrian transition. Palaios 14, 58–72. sponges from Yunnan, China. Journal of Paleontology 69, 1009–1019. McIlroy,D.,Green,O.R.&Brasier, M. D. (2001). Palaeobiology and evolution Rota-Stabelli, O., Campbell,L.,Brinkmann,H.,Edgecombe,G.D., of the earliest agglutinated Foraminifera: Platysolenites, Spirosolenites and related forms. Longhorn,S.J.,Peterson,K.J.,Pisani,D.,Philippe,H.&Telford,M. Lethaia 34, 13–29. (2011). A congruent solution to arthropod phylogeny: phylogenomics, microRNAs McLoughlin,N.,Wilson,L.A.&Brasier, M. D. (2008). Growth of synthetic and morphology support monophyletic Mandibulata. Proceedings of the Royal Society of stromatolites and wrinkle structures in the absence of microbes – implications for London B 278, 298–306. the early fossil record. Geobiology 6, 95–105. Rozanov,A.Y.&Varlamov, A. I. (2008). The Cambrian System of the Siberian Platform. , M. A. S. (1998). The Garden of Ediacara: Discovering the First Complex Life. McMenamin Part 1. The Aldan-Lena Region. PIN-RAS, Moscow and Novosibirsk. Columbia University Press, New York. Rozanov,A.Y.&Zhuravlev, A. Y. (1992). The Lower Cambrian fossil record of Meek, F. B. (1868). Preliminary notice of a remarkable new genus of corals, probably the Soviet Union. In Origin and Early Evolution of the Metazoa (eds J. H. Lipps and P. typical of a new family. American Journal of Science (series 2) 45, 62–64. W. Signor), pp. 205–282. Plenum, New York. Moldowan,J.M.,Fago,F.J.,Lee,C.Y.,Jacobson,S.R.,Watt,D.S., Rozanov,A.Y.,Missarzhevsky,V.V.,Volkova,N.A.,Voronova,L.C., Slougui,N.-E.,Jeganathan,A.&Young, D. C. (1990). Sedimentary 12-n- Krylov,I.N.,Keller,B.M.,Korolyuk,I.K.,Lendzion,K.,Michniak,R., propylcholestanes, molecular fossils diagnostic of marine algae. Science 247, 309–312. Pykhova,N.G.&Sidorov, A. D. (1969). Tommotskij jarus i problema nizhnej Narbonne,G.M.,Myrow,P.M.,Landing,E.&Anderson, M. M. (1987). granizty kembrija. (The Tommotian Stage and the Cambrian lower boundary A candidate stratotype for the Precambrian–Cambrian boundary, Fortune Head, problem). In Trudy Geologicheskogo instituta AN SSSR 206, p. 380 pp. (In Russian; Burin Peninsula, southeastern Newfoundland. Canadian Journal of Earth Sciences 24, English edition: 1981, 359 pp. Amerind Publishing Co., New Delhi. 1277–1293. Rozanov,A.Y.,Repina,L.N.,Apollonov,M.K.,Shabanov,Y.Y.,Zhuravlev, Narbonne,G.M.,Saylor,B.Z.&Grotzinger, J. P. (1997). The youngest A. Y., Pegel,T.V.,Fedorov,A.B.,Astashkin,V.A.,Zhuravleva,I.T., Ediacaran fossils from Southern Africa. Journal of Paleontology 71, 953–967. Egorova,L.I.,Chugaeva,M.N.,Dubinina,S.V.,Ermak,V.V.,Esakova, Noffke,N.,Gerdes,G.,Klenke,T.&Krumbein, W. E. (2001). Microbially N. V., Sundukov,V.V.,Sukhov,S.S.&Zhemchuzhnikov, V. G. (1992). induced sedimentary structures: a new category within the classification of primary Kembrij Sibiri. (The Cambrian of Siberia). Nauka, Novosibirsk. sedimentary structures. Journal of Sedimentary Research 71, 649–656. Rozanov,A.Y.,Khomentovsky,V.V.,Shabanov,Y.Y.,Karlova,G.A., Nosenko,T.,Schreiber, F., Adamska,M.,Adamski,M.,Eitel,M.,Hammel,J., Varlamov,A.I.,Luchinina,V.A.,Pegel,T.V.,Demidenko,Y.E.,Parkhaev, Maldonado,M.,Muller¨ ,W.E.G.,Nickel,M.,Schierwater,B.,Vacelet, P. Y., Korovnikov,I.V.&Skorlotova, N. A. (2008). To the problem of stage J., Wiens,M.&Worheide¨ , G. (2013). Deep metazoan phylogeny: when different subdivision of the Lower Cambrian. Stratigraphy and geological correlation 16, 1–19. genes tell different stories. Molecular Phylogenetics and Evolution 67, 223–233. Runnegar, B. (1992). Proterozoic fossils of soft-bodied metazoans (Ediacara Faunas). Ogg, J. G. (2004). Status of divisions of the international geologic timescale. Lethaia In The Proterozoic Biosphere: A Multidisciplinary Study (eds J. W. Schopf and C. Klein), 37, 183–199. pp. 999–1007. Cambridge University Press, Cambridge. Opik¨ , A. A. (1975). Cymbric Vale fauna of New South Wales and Early Cambrian Runnegar,B.N.&Fedonkin, M. A. (1992). Proterozoic metazoan body fossils. In biostratigraphy. Bureau of Mineral Resources, Geology and Geophysics, Australia, Bulletin The Proterozoic Biosphere: A Multidisciplinary Study (eds J. W. Schopf and C. Klein), 159,iv+ 78 p., 14 fig., 7 pl. pp. 369–388. Cambridge University Press, Cambridge. Orr,P.J.,Benton,M.J.&Briggs, D. E. G. (2003). Post-Cambrian closure of the Ruppert,E.E.,Fox,R.S.&Barnes, R. D. (2003). Invertebrate Zoology: A Functional deep-water slope-basin taphonomic window. Geology 31, 769–772. Evolutionary Approach. Brooks Cole, Thompson Learning, Belmont. Pel’Man,Y.L.,Ermak,V.V.,Fedorov,A.B.,Luchinina,V.A.,Zhuravleva, Sansom,R.S.,Gabbott,S.E.&Purnell, M. A. (2010). Non-random decay of I. T., Repina,L.N.,Bondarev,V.I.&Borodaevskaya, Z. V. (1990). Novye chordate characters causes bias in fossil interpretation. Nature 463, 797–800. dannye po stratigrafii i paleontologii verkhnego dokembriya i nizhnego kembriya r. Schroder¨ ,S.&Grotzinger, J. P. (2007). Evidence for anoxia at the Ediacaran- Dzhandy (pravyj pritok r. Aldan). In Biostratigrafiya i paleontologiya kembriya Severnoj Azii Cambrian boundary: the record of redox-sensitive trace elements and rare earth (ed. L. N. Repina), pp. 3–32. Nauka, Novosibirsk (in Russian). elements in Oman. Journal of the Geological Society 164, 175–187.

Seilacher, A. (1977). Evolution of trace fossil communities. In Patterns of Evolution (ed. Tiwari,M.,Pant,C.C.&Tewari, V. C. (2000). Neoproterozoic sponge spicules A. Hallam), pp. 359–376. Elsevier, Amsterdam. and organic walled microfossils from the Gangolihat Dolomite, Lesser Himalaya, Seilacher, A. (1992). Vendobionta and Psammocorallia: lost constructions of India. Current science 79, 651–654. Precambrian evolution. Journal of the Geological Society 149, 607–613. Vacelet, J. (1977). Une nouvelle relique du Secondaire: un representant´ actuel des Serezhnikova, E. A. (2007a). Palaeophragmodictya spinosa sp. nov., a bilateral benthic eponges fossiles Sphinctozoaires. Académie des Sciences, Paris, Comptes Rendus (sérieD) organism from the Vendian of the southeastern White Sea region. Paleontological 285, 509–511, 1 pl. Journal 41, 360–369. Vickers-Rich,P.(2007). The Nama Fauna of Southern Africa. In: The Rise of Animals: Serezhnikova, E. A. (2007b). Vendian Heimalora from Arctic Siberia reinterpreted Evolution and Diversification of the Kingdom Animalia.M.A.Fedonkin,J.G.Gehling, as holdfasts of benthic organisms. In The Rise and Fall of the Ediacaran Biota, K. Grey,G.M.Narbonne &P.Vickers-Rich (eds.). pp. 68–87. Johns Hopkins Geological Society of London Special Publications 286 (eds P. Vickers-Rich University Press, Baltimore. and P. Komarower), pp. 331–337. Geological Society of London, London. Wacey,D.,Kilburn,M.,McLoughlin,N.,Parnell,J.,Stoakes,C., Serezhnikova,E.A.&Ivantsov, A. Y. (2007). Fedomia mikhaili—A new spicule- Grosvenor,C.&Brasier, M. D. (2008a). Use of nanoSIMS in the search bearing organism of sponge grade from the Vendian (Ediacaran) of the White Sea, for early life on Earth: ambient inclusion trails in a c, 3400 Ma sandstone. Journal of Russia. Palaeoworld 16, 319–324. the Geological Society 165, 43–53. Shabanov,Y.Y.,Korovnikov,I.V.,Pereladov,V.S.&Fefelov, A. F. (2008). Wacey,D.,Kilburn,M.,Stoakes,C.A.,Aggleton,H.&Brasier,M.D. Excursion 1a. The traditional Lower-Middle Cambrian boundary in the Kuonamka (2008b). Ambient inclusion trails: their recognition, age range and applicability to Formation of the Molodo River section (the southeastern slope of the Olenek uplift early life on Earth. In Links Between Geological Processes, Microbial Activities and of the Siberian Platform) proposed as a candidate for GSSP of the lower bopundary Evolution of Life (ed. Y. Dilek), pp. 113–134. Springer, Berlin. of the Middle Cambrian and its basal (Molodian) stage, define by the FAD of Wacey,D.,Kilburn,M.,Saunders,M.,Cliff,J.&Brasier, M. D. (2011). Ovatoryctocara granulata.InThe Cambrian System of the Siberian Platform. Part 2: North-east Microfossils of sulphur –metabolizing cells in 3.4-billion-year-old rocks of Western of the Siberian Platform (eds A. Y. Rozanov and A. I. Varlamov. pp.), p. 140. PIN, Australia. Nature Geoscience 4, 698–702. RAS, Moscow, Novosibirsk. Walcott, C. D. (1894). The Fauna of the Lower Cambrian or Olenellus zone. United States Siegl,A.,Kamke,J.,Hochmuth,T.,Piel,J.,Richter,M.,Liang,C.,Dandekar, Geological Survey, Washington, 10th Annual Report. T. & Hentschel, U. (2011). Single-cell genomics reveals the lifestyle of Poribacteria, Walter,M.R.,Oehler,J.H.&Oehler, D. Z. (1976). Megascopic algae 1,300 a candidate phylum symbiotically associated with marine sponges. The International million years old from the Belt supergroup, Montana: a reinterpretation of Walcott’s Society for Microbial Ecology Journal 5,61–70. Helminthoidichnites. Journal of Paleontology 50, 872–881. Siveter,D.J.,Williams,M.&Waloszek, D. (2001). A phosphatocopid crustacean Wang,Y.&Wang, X. (2011). New observations on Cucullus Steiner from the with appendages from the Lower Cambrian. Science 293, 479–481. Neoproterozoic Doushantuo Formation of Guizhou, South China. Lethaia 44, Siveter,D.J.,Waloszek,D.&Williams, M. (2003). An early Cambrian 275–286. phosphatocopid crustacean with three-dimensionally preserved soft parts from Wood , R. A. (1998). The ecological evolution of reefs. Annual Review of Ecology and Shropshire, England. In Trilobites and Their Relatives, Special Papers in Palaeontology, Systematics 29, 179–206. Volume 70 (eds P. D. Lane,D.Siveter and R. A. Fortey), pp. 9–31. Blackwell Wood,R.A.,Grotzinger,J.P.&Dickson, J. A. D. (2002). Proterozoic modular Publishing. biomineralized metazoan from the Nama Group, Namibia. Science 296, 2383–2386. Sokolov,B.S.&Zhuravleva, I.T. (1983). Yarusnoe raschlenenie nizhnego kembriya Sibiri. Worheide¨ , G. (2008). A hypercalcified sponge with soft relatives: Vaceletia is a keratose Atlas okamenelostej. (Stage subdivision of the Lower Cambrian of Siberia. Atlas of demosponge. Molecular Phylogenetics and Evolution 47, 433–438. fossils.). Trudy Instituta Geologii i Geofiziki SO AN SSSR 558, 216 pp. Worheide¨ ,G.,Dohrmann,M.,Erpenbeck,D.,Larroux,C.,Maldonado,M., Sperling,E.A.,Pisani,D.&Peterson, K. J. (2007). Poriferan paraphyly and its Voigt, O., Borchiellini,C.&Lavrov, D. V. (2012). Deep phylogeny implications for Precambrian palaeobiology. In The Rise and Fall of the Ediacaran Biota, and evolution of sponges (phylum Porifera). Advances in Marine Biology Geological Society of London Special Publications 286 (eds P. Vickers-Rich and 61, 1–78. P. Komarower), pp. 355–368. Geological Society of London, London. Xiao,S.,Yuan,X.,Steiner,M.&Knoll, A. H. (2002). Macroscopic carbonaceous Sperling,E.A.,Robinson,J.M.,Pisani,D.&Peterson, K. J. (2010). Where’s compressions in a terminal Proterozoic shale: a systematic reassessment of the the glass? Biomarkers, molecular clock, and microRNAs suggest a 200-Myr missing Miaohe biota, South China. Journal of Paleontology 76, 347–376. Precambrian fossil record of siliceous sponge spicules. Geobiology 8, 24–36. Yin,L.,Xiao,S.&Yuan, X. (2001). New observations on spicule-like structures from Sperling,E.A.,Peterson,K.J.&Laflamme, M. (2011). Rangeomorphs, Thectardis Doushantuo phosphorites at Weng’an, Guizhou Province. Chinese Science Bulletin 46, (Porifera?), and dissolved organic carbon in Ediacaran oceans. Geobiology 9, 24–33. 1828–1832. Srivastava,M.,Simakov, O., Chapman,J.,Fahey,B.,Gauthier,M.E.A., Yuan,X.L.,Xiao,S.H.,Yin,L.M.,Knoll,A.H.,Zhou,C.M.&Mu,X.N. Mitros,T.,Richards,G.S.,Conaco,C.,Dacre,M.,Hellsten,U.,Larroux, (2002). Doushantuo Fossils: Life on the Eve of Animal Radiation. University of Science and C., Putnam,N.H.,Stanke,M.,Adamska,M.,Darling,A.,Degnan,S.M., Technology of China Press, Hefei. Oakley,T.H.,Plachetzki,D.C.,Zhai,Y.,Adamski,M.,Calcino,A., Zhang,Y.,Yin,L.&Xiao, S. (1998). Permineralized fossils from the terminal Cummins, S. F., Goodstein,D.M.,Harris,C.,Jackson,D.J.,Leys,S. Proterozoic Doushantuo Formation, South China. The Paleontological Society, Memoir P., Shu,S.,Woodcroft,B.J.,Vervoort,M.,Kosik,K.S.,Manning,G., 50, 1–12. Degnan,B.M.&Rokhsar,D.S.(2010).TheAmphimedon queenslandica genome Zhang,S.,Jiang,G., Zhang,J.,Biao,S.,Martin,K.J.&Nicholas,C.B. and the evolution of animal complexity. Nature 466, 720–727. (2005). U-Pb sensitive high resolution ion microprobe ages from the Doushantuo Steiner, M. (1994). Die Neoproterozoishen Megaalgen Sudchinas. Berliner Formation in South China: constraints on Late Neoproterozoic glaciations. Geology Geowisernschaftliche Abhandlungen 15, 1–146. 33, 473–476. Steiner,M.&Reitner, J. (2001). Evidence of organic structures in Ediacara-type Zhao,Z.,Xing,Y.,Ding, Q., Liu,G.,Zhao,Y.,Zhang,S.,Meng,X.,Yin, fossils and associated microbial mats. Geology 29, 1119–1122. C., Ning,B.&Han, P. (1988). The Sinian System of Hubei. China University of Steiner,M.,Mehl,D.,Reitner,J.&Erdtmann, B.-D. (1993). Oldest entirely Geosciences Press, Wuhan. preserved sponges and other fossils from the lowermost Cambrian and a new Zhiwen,J.,Brasier,M.D.&Hamdi, B. (1989). Correlation of the Meishucunian facies reconstruction of the Yangtze platform (China). Berliner Geowissenschaftlich stage in South Asia. Acta Geologica Sinica 2, 1–9. Abhandlungern, E 9, 293–329. Zhou,C.,Yuan,X.&Xue, Y. (1998). Sponge spicule-like pseudofossils frm Tang,T.,Zhang,J.&Jiang, X. (1978). Discovery and significance of the Late Sinian the Neoproterozoic Doushantuo Formation in Weng’an, Guizhou, China. Acta fauna from western Hunan and Hubei. Acta stratigraphica Sinica 2, 32–45. Micropalaeontologica Sinica 15, 380–384 (in Chinese). Tiwari, M. (1999). Organic-walled microfossils from the Chert-Phosphate Member, Zhuravleva, I. T. (1970). Porifera, Sphinctozoa, Archaeocyathi– Their connections. Tal Formation, Precambrian–Cambrian boundary, India. Precambrian Research 97, Symposia of the Zoological Society, London 25, 41–59. 99–113.

(Received 20 June 2012; revised 16 January 2014; accepted 30 January 2014 )