A cryptic record of Burgess Shale‐type diversity from the early Cambrian of Baltica

Exceptionally preserved ‘Burgess Shale‐type’ fossil assemblages from the Cambrian of Laurentia, South China and Australia record a diverse array of non‐biomineralizing organisms. During this time, the palaeocontinent Baltica was geographically isolated from these regions, and is conspicuously lacking in terms of comparable accessible early Cambrian Lagerstätten. Here we report a diverse assemblage of small carbonaceous fossils (SCFs) from the early Cambrian (Stage 4) File Haidar Formation of southeast Sweden and surrounding areas of the Baltoscandian Basin, including exceptionally preserved remains of Burgess Shale‐type metazoans and other organisms. Recovered SCFs include taxonomically resolvable ecdysozoan elements (priapulid and palaeoscolecid worms), lophotrochozoan elements (annelid chaetae and wiwaxiid sclerites), as well as ‘protoconodonts’, denticulate feeding structures, and a background of filamentous and spheroidal microbes. The annelids, wiwaxiids and priapulids are the first recorded from the Cambrian of Baltica. The File Haidar SCF assemblage is broadly comparable to those recovered from Cambrian basins in Laurentia and South China, though differences at lower taxonomic levels point to possible environmental or palaeogeographical controls on taxon ranges. These data reveal a fundamentally expanded picture of early Cambrian diversity on Baltica, and provide key insights into high‐latitude Cambrian faunas and patterns of SCF preservation. We establish three new taxa based on large populations of distinctive SCFs: Baltiscalida njorda gen. et sp. nov. (a priapulid), Baltichaeta jormunganda gen. et sp. nov. (an annelid) and Baltinema rana gen. et sp. nov. (a filamentous problematicum).

priapulids are the first recorded from the Cambrian of Baltica. The File Haidar SCF assemblage is broadly comparable to those recovered from Cambrian basins in Laurentia and South China, though differences at lower taxonomic levels point to possible environmental or palaeogeographical controls on taxon ranges. These data reveal a fundamentally expanded picture of early Cambrian diversity on Baltica, and provide key insights into high-latitude Cambrian faunas and patterns of SCF preservation. We establish three new taxa based on large populations of distinctive SCFs: Baltiscalida njorda gen. et sp. nov. (a priapulid), Baltichaeta jormunganda gen. et sp. nov. (an annelid) and Baltinema rana gen. et sp. nov. (a filamentous problematicum).
There is, however, an alternative means of capturing a measure of BST diversity, even in the absence of exceptional macrofossil preservation. The (mostly) disarticulated, taphonomically recalcitrant elements of these organisms are widely recovered as 'small carbonaceous fossils' or SCFs (Butterfield & Harvey 2012). Like their 'small shelly' counterparts, SCFs are deeply polyphyletic, united by little more than their size and a particular search image, in this case organic-walled fossils that are too small to be seen on bedding surfaces, but too large or delicate to be reliably recovered via conventional palynological processing. Despite their fragmentary nature, however, many SCFs can nonetheless be resolved to a taxonomically informative level. Importantly, what SCFs lose from disarticulation they gain in abundance and distribution, offering a fundamentally enhanced account of palaeobiogeographical and stratigraphical patterns. The key to recovering these large-scale patterns, of course, is global sampling, though most Cambrian SCF studies to date have been centred on relatively few localities in western Canada and South China (Butterfield 1994;Butterfield & Harvey 2012;Harvey et al. 2012a, b; but see Smith et al. 2015a). Here we expand this record to the palaeogeographically isolated palaeocontinent of Baltica.
Global palaeogeographical reconstructions of the early Cambrian vary considerably (e.g. McKerrow et al. 1992;Li et al. 2008;Meert & Lieberman 2008), but there is general agreement that the sedimentary depocentres of Scandinavia/Baltoscandia and the East European Platform belonged to a coherent craton, Baltica, separated from other contemporary landmasses by substantial expanses of ocean (

GEOLOGICAL SETTING
The File Haidar Formation (early Cambrian, Stage 4) is locally exposed at the surface in and around southern Sweden, and subcrops extensively beneath the Baltic Sea (Figs 1, 2;Hagenfeldt 1989Hagenfeldt , 1994Nielsen & Schovsbo 2007). Deposition took place in a shallow epicontinental sea, broadly mirroring the architecture of the modern Baltic and Bothnian seas, and separated from the Russian margin by a substantial hinterland (Flod en 1980;Lidmar-Bergstr€ om 1988;Nielsen & Schovsbo 2011, figs 82-86).
The File Haidar Formation consists of fine-medium sandstones interspersed with recurrent mudstones and siltstones, overlying a weathered gneissic basement (Holm 1901). Across much of its extent, the File Haidar Formation is terminated by a regional unconformity corresponding to the 'Hawke Bay event' (Fig. 2;see Nielsen & Schovsbo 2015) and is overlain by the predominantly middle Cambrian (Stage 5) Borgholm Formation (Nielsen & Schovsbo 2007). In places, however, this latter contact is gradational, and parts of the Borgholm Formation (e.g. the Gr€ otlingbo Member) are recognized as latest early Cambrian in age, with the diachronous boundary younging westwards (Nielsen & Schovsbo 2011). The subsurface portions of the File Haidar Formation are well known from drillcore, and have been extensively studied both sedimentologically (e.g. Nielsen & Schovsbo 2007, 2011 and for acritarch biostratigraphy (Hagenfeldt 1989(Hagenfeldt , 1994Eklund 1990).
Sediments of the File Haidar Formation record marine deposition in a storm-influenced nearshore sand belt. Proximal environments are sand-dominated, with distal regions of the inner-shelf grading into siltstones and shales (Nielsen & Schovsbo 2007). The formation is subdivided into five members based on lithology and thin conglomeratic marker beds (Figs 2, 3;Nielsen & Schovsbo 2007, 2011. Sediments of the lowermost Viklau Member consist of interbedded green-grey quartz sandstones, bioturbated siltstones and green shales. The overlying N€ ar Shale Member is dominated by quartz sandstone in proximal settings (e.g. in the File Haidar-1 core; Fig. 3), grading distally into successions of heavily bioturbated green-grey siltstones and shales with subordinate sandstones (e.g. in the Gr€ otlingbo-1 core; Fig. 3). The N€ ar Sandstone Member is dominated by fine-medium grained quartz sandstones with occasional subordinate siltstone beds. The proximal equivalents of these three members are the Mickwitzia Sandstone and overlying Lingulid Sandstone members; both of these are extensively bioturbated, though the Mickwitzia is more thinly bedded, contains numerous clay horizons, and preserves a substantially greater diversity of ichnotaxa, including both hunting and priapulid-like trace fossils (Bergstr€ om 1973;Jensen 1990Jensen , 1997Jensen & Bergstr€ om 1995;Rydell et al. 2001).
The Viklau, N€ ar Shale and N€ ar Sandstone members are all substantially developed in subsurface drillcores around € Oland and Gotland (Thorslund & Westerg ard 1938;Hessland 1955); the Viklau Member is absent in the southern Bothnian Sea, however, where the Cambrian succession begins with the N€ ar Shale (Fig. 2) surface exposures of these members include the distinctively red-striated and well-cemented Kalmarsund sandstone, which crops out on the southwest coast of mainland Sweden and is now interpreted as diagenetically altered Viklau, N€ ar Shale and N€ ar Sandstone strata (Nielsen & Schovsbo 2007). Further inland, the Mickwitzia Sandstone and Lingulid Sandstone members are locally exposed at the surface around V€ asterg€ otland, € Osterg€ otland and N€ arke, where they are also known from drillcore material (Figs 1, 2).

MATERIAL AND METHOD
The File Haidar Formation was sampled from seven widely separated drillcores in and around southern Sweden (Figs 1, 3); Gr€ otlingbo-1 (Gr-1), File Haidar-1 (FH-1), Hammnudden-1 (Ha-1), B€ oda Hamn (BH), B arstad-2 (B a-2), Bernstorp-2 (Be-2) and Finngrundet (Fi). All cores are housed in the drillcore repository of the Geological Survey of Sweden at Eggebyholm, near Knivsta, Uppland, Sweden. Our sampling protocol included all unoxidized mudstones, siltstones and sandstones, apart from pure quartz arenites. At least one sample was collected per metre of suitable lithology, yielding a total of 217 samples representing all five of the constituent members (Fig. 3). Approximately 50 g of each sample was digested and processed for SCFs using a low-manipulation hydrofluoric acid maceration procedure, and resultant organic residues were hand-picked for SCFs (see Butterfield & Harvey 2012).

SMALL CARBONACEOUS FOSSILS
A total of 3613 SCFs were mounted onto 261 slides from 66 productive samples. Although all seven boreholes yielded fossils, samples from Gr-1, FH-1, BH-1 and B a-2 were by far the most abundant and diverse, with fissile siltstones and fine-grained glauconitic sandstones producing the bulk of specimens (Fig. 3). Most of the recovered SCFs are recognizably metazoan/bilaterian and can be classified as having either ecdysozoan or lophotrochozoan affiliations; phylogenetically less resolved forms are assigned to 'unresolved metazoans' or 'non-metazoan problematica'.

Ecdysozoa
The protostome superphylum Ecdysozoa includes both the Panarthropoda (euarthropods, onychophorans and tardigrades) and Cycloneuralia (scalidophorans and nematoids), all of which are characterized by a variably tanned/sclerotized cuticle periodically shed through ecdysis. The inherent recalcitrance of this extracellular material, combined with a propensity for body parts to disarticulate along lines of cuticular jointing or thinning, means that ecdysozoans are particularly conducive to the production of SCFs. Arthropods are by far the most abundant and diverse ecdysozoans today, and dominate a number of Cambrian SCF assemblages (e.g. Harvey et al. Priapulid-like scalids, teeth and hooks. Sub-triangular, thick-walled elements bearing denticles, projections and/ or polygonal surface micro-ornament are locally abundant in parts of the Viklau and N€ ar Shale Members of the Gr-1 and FH-1 cores (Figs 3, 4, 5A-Q). Considerable variation is evident among the recovered specimens, but all conform to the same basic architecture, consisting of a prominent conical 'prong' at the distal tip of a thickened chevron-shaped 'arch' which caps an ovoid 'basal pad' (Fig. 4A;Smith et al. 2015b) though the thinner-walled basal pad has been secondarily lost from all but a few specimens (e.g. Fig. 5B). In most specimens the arch is fringed by an array of smaller conical 'denticles' which flare at the base and taper to an acute point ( Fig. 5A-Q). Beyond this basic architecture, the specimens exhibit a considerable range of outlines and marginal ornamentations, forming a broad continuum encompassing three apparent end-members or 'types'. Type 1 specimens exhibit an elongate triangular outline with small, closelyspaced marginal denticles distributed along most of the length of the arch/prong complex (e.g. Fig. 5F, H). Type 2 specimens bear fewer, more robust denticles that increase in size towards the central prong, and have a broader, more equidimensional outline (e.g. Fig. 5A). Type 3 specimens are more elongate with an acute-angled arch that tapers to a pronounced prong, and exhibit a significant gap between the most distal denticles and the tip of the central prong (e.g. Fig. 5B-E, G). The denticles in type 3 specimens are sometimes branched at their tips (e.g. Fig. 5O, Q) and are frequently arranged in an asymmetrical pattern on either margin of the arch; in contrast to type 2 specimens, there is no clear size trend among the denticles (e.g. Fig. 5B).
Taken together, these denticulate File Haidar SCFs are reliably interpreted as the dispersed cuticular remains of priapulid-like scalidophoran worms. 'Sclerites' of the same fundamental construction to those found in the Gr-1 and FH-1 cores are found in a variety of modern priapulids (e.g. van der Land 1970;Wennberg et al. 2009), as well as Cambrian stem-priapulids such as Ottoia and the tubicolous genus Selkirkia (Smith et al. 2015b); they are also abundant in many previously-described SCF assemblages (Butterfield & Harvey 2012;Harvey et al. 2012a, b). In particular, it is the approximately bilaterally symmetrical outline combined with a flattened-oval or triangular basal attachment shape that distinguishes these priapulid-type sclerites from simpler, less phylogenetically informative conical structures and from annelid scolecodonts (see discussions in Smith et al. 2015b, p. 713;Eriksson et al. 2016); the absence of an inner groove, strongly developed ridges (rugae), or fine parallel growth lines rules out comparisons with superficially similar machaeridian annelid plates (e.g. Adrain et al. 1991, figs 1, 2). In modern priapulids, spinose cuticular projections of the anterior body regions are termed 'scalids' when they occur on the introvert or trunk, whereas 'teeth' are positioned on the pharynx and point forward when everted (van der Land 1970), a distinction that is usefully applied to various fossil scalidophorans (Harvey et al. 2012a;Smith et al. 2015b). The specimens in Fig. 5A-Q are likely to be 'teeth' based on their outline shapes and prominent fringing spines, whereas scalids tend to be more elongate and are often unadorned, or have spines in rows closer to the midline, rather than marginally (Smith et al. 2015b).
The type 2 teeth described here are broadly comparable to the type B teeth in Ottoia prolifica (cf. Smith et al. 2015b, fig. 3) unlike the more elongate outlines of the type 1 and type 3 specimens. Even so, the prong is substantially less clearly differentiated from the supporting arch than in Ottoia, and none of the particular morphologies in the File Haidar has been previously documented, either from BST macrofossils or among priapulid-like SCFs. Type 1 and type 3 specimens are also distinct from other known priapulid teeth, but given the consistent co-occurrence of all three tooth types in our samples, as well as the tooth zonation patterns The File Haidar samples that preserve priapulid teeth also yield recurved hooks with a pronounced 'spur' (Figs 4B,6). Broadly comparable forms have been reported from a number of SCF assemblages from Canada and South China (Butterfield 2008;Butterfield & Harvey 2012;Harvey et al. 2012a;Smith et al. 2015b), though these latter occurrences are typically distinguished by serially arranged denticles or fine projections along the inner curve of the hook (see Butterfield 2008, fig. 6); such elaboration is absent from all but one of the File Haidar examples (Fig. 6X). Because these hook-like SCFs consistently co-occur in samples with types 1, 2 and 3 priapulid teeth, and exhibit a morphological range consistent with the variation seen within individual fossil worms (cf. Smith et al. 2015b), we interpret them as scalid elements conspecific with Baltiscalida njorda gen. et. sp. nov. (see Systematic Palaeontology below).
Palaeoscolecids. Small (~40-100 lm diameter) discoidal elements occur locally in the Viklau Member of the GB-1 core ( Fig. 5R-V). These plate-like SCFs have a thin outer rim, and a darker, thicker central region that bears between three and five protuberances or 'nodes'. Mineralized (phosphatic) plates of this size and form are F I G . 4 . Schematic guide to File Haidar SCF terminology. A-B, priapulid teeth and scalids. C, palaeoscolecid plates. D-F, annelid chaetae: D, E, Baltichaeta jormunganda showing morphological end members; F, distal tip of Burgessochaeta setigera chaeta for comparison. G, wiwaxiid sclerite (ventrolateral sclerite). * indicates usually damaged or missing in recovered specimens.
conventionally assigned to the form-genus Hadimopanella which occurs widely in Cambrian assemblages of SSFs. Though originally described from isolated sclerites, Hadimopanella-type plates are now known to have adorned the surface of palaeoscolecids, an extinct group of early Palaeozoic worms with an annulated armoured trunk, a retractile introvert and tail hooks (Hinz et al. Along with a preliminary report from the Furongian of western Canada (Butterfield & Harvey 2012), the identification of Hadimopanella-type sclerites from siliciclastic sediments of the File Haidar Formation substantially expands the known biogeographical and taphonomical ranges of a prominent group of early Palaeozoic worms. Both occurrences are most parsimoniously interpreted as the taphonomically demineralized remains of an originally phosphatic form, though it is possible that they represent a distinct non-biomineralizing habit (see Butterfield & Nicholas 1996;Harvey 2010).
Cuticular 'spines'. Alongside the distinctively scalidophoran-type teeth, hooks and cuticular plates, the File Haidar assemblage preserves a diversity of more simple coniform and spinose SCFs. The most common of these are stout, conical spines with a flared, elliptical to circular base ( A significant subset of the shorter, thinner-walled File Haidar spines ( Fig. 7A-H, N-R) are indistinguishable from microfossils that have been assigned to the 'acritarch' form-taxon Ceratophyton (Kiryanov 1979in Volkova et al. 1979Hagenfeldt 1989Hagenfeldt , 1994Wright et al. 1993;Konzalov a 1995;Jachowicz & Moryc 1995;Zang et al. 2007;Palacios et al. 2011). Despite this implied protistan-grade affiliation, however, there is a strong case for recognizing such SCFs as metazoan (Steiner & Fatka 1996;Moczydłowska 2008;Agi c et al. 2016), and, more specifically, scalidophoran in origin (Butterfield & Harvey 2012;Smith et al. 2015b). Specimens with ovoid bases and the long axis of the cone oblique to the basal attachment, for example, or a polygonal micro-ornamentation of the basal cuticle, share these features with more complex and definitively priapulid-like scalids and teeth (cf. Smith et al. 2015b). Moreover, several of the File Haidar 'Ceratophyton' exhibit basal extensions suggesting they were once attached to a larger cuticular sheet (e.g. F I G . 6 . Scalids representing a mixture of tail and introvert hooks derived from stem priapulids. A-V, variety of scalids lacking any evident denticles; distal curling of some specimens (e.g. Q) is probably a taphonomic feature. W, cluster of scalids with extended portion of cuticle. X, probable introvert hook bearing short denticles along the arch. All slide numbers have the prefix NRM PZ X:
Annelid chaetae. Strap-shaped SCFs with distinctively bifid tips (Fig. 8) are particularly common in parts of the Viklau and N€ ar Shale Members (GB-1 and FH-1 cores; Fig. 3). Most occur as isolated shafts but one semi-articulated specimen is represented by five constituent elements (Fig. 8B, Z). Individual specimens are~35-210 lm wide with a maximum (incomplete) length in excess of a millimetre. Preserved tips consist of a pair of lateral prongs (thickened longitudinal margins) with thinner material in between forming a comb-like series of serrations (Figs 4D-E, 8W). The prongs always extend beyond the intervening serrations, though they vary considerably in size both within and between individual specimens. Individual serrations of the comb either taper to a point (e.g. Fig. 8X) or are somewhat rounded (e.g. Fig. 8Z), the latter type typical of specimens with more symmetrical termini and shorter serrations of consistent dimensions. As a population, the distal morphology varies along a continuum between two co-occurring end-members, one in which asymmetrical prongs project well beyond the comb tip (e.g. Fig. 8A), and the other with proportionately broader, shorter prongs of approximately equal length (e.g. Fig. 8B-E); the latter form has substantially wider combs and often exhibits two hyper-developed serrations that divide the comb approximately into thirds (Fig. 8X). Other features of these strap-shaped fossils include a pattern of regularly spaced transverse ridges with a~5-12 lm interval (Fig. 8Y), and, in well-preserved specimens, an internal microstructure of densely packed, <1 lm non-overlapping striations (Fig. 8W-Y) directly comparable to that of microvilli-secreted chaetal structures.
The combination of microvilli-like microstructure, complex tip morphology and bundled occurrence seen in these SCFs is directly comparable to that of annelid chaetae (cf. Hausen 2005;Merz & Woodin 2006;Tilic et al. 2015a). In particular, close morphological comparisons can be made with the 'brush-chaetae' of certain 'polychaete' annelids (Mettam 1971(Mettam , 1984Roy 1974;Knight-Jones 1981;Rouse & Pleijel 2001). The characteristically complex-tipped chaetae of annelids, particularly marine polychaetes, contrast with the simple spines or rods secreted by other groups, for example brachiopod setae or bryozoan teeth (cf. Gordon 1975;Zhang et al. 2007;Zhang & Holmer 2013;Topper et al. 2015). Additional support for an annelidan affinity comes from the cluster of five superimposed elements, which we interpret as a fragmentary portion of a single polychaete chaetigera fan of chaetae borne on the lateral margins of a body segment (Fig. 8B, Z).
The chaetae of modern annelids exhibit an enormous diversity of form, which at the broadest scale fall into two categories: simple or compound, with the latter characterized by an internally jointed articulation and/or dentate terminal hooks. Like all other known Cambrian annelid chaetae (see Parry et al. 2014Parry et al. , 2016, the File Haidar specimens are simple; the transverse ridges (Fig. 8Y)  The closest comparison, however, is found in the Burgess Shale 'polychaete' Burgessochaeta setigera (Conway Morris 1979a, b), whose simple strap-shaped chaetae are similarly distinguished by a bifid tip with asymmetric prongs (Fig. 4F; see Eibye-Jacobsen 2004). Even so, the asymmetry of Burgessochaeta chaetae is conspicuously more pronounced than in the File Haidar forms, with the lateral prongs diverging only distally to yield a distinctively splayed 'Y'-shape tip (the 'lyrate-chaetae' condition; Rouse & Pleijel 2001); by contrast, the lateral/marginal prongs of the File Haidar specimens are uniformly parallel, leaving space for the expanded brush-like comb (the 'brush-chaetae' condition; Rouse & Pleijel 2001). Bedding-plane macrofossils do not permit microstructural analysis on a par with SCFs, but if any analogous fine brush is present in Burgessochaeta, it would necessarily be confined to the larger inclined prong as in some modern lyrate-chaetae (Fig. 4F). Nonetheless, the brush-tipped Baltic SCFs are sufficiently distinct to be recognized as a new genus of Burgessochaeta-like annelid, Baltichaeta jormungandr gen. et. sp. nov. (see Systematic Palaeontology below).
Wiwaxiid sclerites. Isolated sclerites of Wiwaxia were recovered from the N€ ar Sandstone Member of the B€ oda Hamn core (Figs 3, 9AC) and from the early-middle Cambrian boundary interval at the base of the Borgholm Formation in the B arstad-2 core (Figs 3, 9Z-AB). The majority of these specimens are asymmetrical, originating from the 'ventrolateral' region of the compound dorsal scleritome. One of the Boda Hamn (File Haidar) specimens (Fig. 9AC) is close to a millimetre long (0.8 mm), the largest Wiwaxia sclerite yet recovered as an SCF, roughly corresponding to a whole-organism body length of~3 cm (see Smith 2014). This same sclerite is further characterized by its blunt termination and associated collar-like thickening, suggesting a biologically mediated loss of the tip.
Most of the recovered Wiwaxia sclerites have prominently developed ribs and in this respect are indistinguishable from the Burgess Shale type species Wiwaxia corrugata. The single specimen with a thickened distal collar appears to be qualitatively different from all other occurrences, but will require a larger sample size to legitimize separate species recognition. Other named species of Wiwaxia have been established on the basis of distinctive rib thickness, sclerite outline, root morphology and/ or surface ornament (Smith et al. 2015a). Given their conspicuously chaeta-like habit and microvillar microstructure, all wiwaxiid sclerites can be confidently placed in total-group Lophotrochozoa, despite on-going debate regarding their phylum-level status (Butterfield 1990b(Butterfield , 2006Smith 2014;Zhang et al. 2015b;Smith et al. 2015a).

Unresolved metazoans
Beyond the recognizable ecdysozoan and lophotrochozoan elements, the File Haidar SCF assemblage includes a further range of forms that are clearly bilaterian, but are otherwise lacking taxonomically diagnostic features. The most distinctive of these are sporadically distributed multidenticulate structures represented by series of progressively larger conical-elements, usually connected via a basal membrane (Fig. 10). Such architecture is comparable to the denticulate mouthparts of various bilaterians, including, but not limited to, the jaws of eunicid and phyllodocid polychaete annelids (fossilized as 'scolecodonts'; see Eriksson et al. 2016, fig. 3), molluscan radulae (Butterfield 2008), the feeding apparatus of fossil Wiwaxia/Odontogriphus (Caron et al. 2006) and fossil conodonts. Although some of the File Haidar specimens (e.g. Fig. 10A) are comparable to 'possible wiwaxiid mouthparts' identified in a Cambrian SCF assemblage from Columbia (see Smith et al. 2015a, fig. 3), it is notable that these do not occur in the same samples yielding Wiwaxia sclerites here.

Non-metazoan problematica
There is nothing specifically metazoan about SCFs of course, the only qualifying criteria being acid resistance and sufficiently large dimensions to be seen and hand-picked (> 40-50 lm). It is certainly notable that a majority of the File Haidar SCFs are metazoan, but there is a conspicuous subset that are demonstrably not. We include these non-metazoan SCFs here because they represent a further range of forms that have been systematically overlooked/destroyed by conventional techniques, and for what they contribute to the larger scale palaeoenvironmental and taphonomic patterns. Apart from animals, carbonaceous fossils of early Cambrian age might conceivably include fungi, protistan-grade eukaryotes and prokaryotes, though these can often be difficult to distinguish solely on the basis of fossil morphology.
Despite their problematic affinities, some of the nonmetazoan SCFs in the File Haidar assemblage are sufficiently distinctive to be recognized as biologically meaningful form-taxa. Among the most common are filamentous fossils recovered from the N€ ar Shale and Viklau members (Figs 3, 11A-M). Individual specimens typically comprise a thick (~70-130 lm wide) sinuous main 'filament,' from which a variety of smaller filaments emerge and may themselves branch further (e.g. Fig. 11D). Where preserved, the tips of the main filament and sub-branches taper to a blunt terminus (Fig. 11C). The fossils are further distinguished by the presence of multiple spheroidal to flask-shaped vesicles, anatomically connected to the main filament ( Fig. 11E-K, M). Carbonaceous fossils of this morphology have not previously been described from the Cambrian, but a loose comparison can be drawn with certain Ediacaranage 'vendotaenids' (Gnilovskaya et al. 1988). We establish a new taxon for these enigmatic organisms: Baltinema rana gen. et. sp. nov. (see Systematic Palaeontology below).
Notably, our processing did not recover any of the spiny/acanthomorphic acritarchs expected in early Cambrian mudstones, and previously reported from the File Haidar Formation (Hagenfeldt 1989(Hagenfeldt , 1994, presumably because they fall below the size range detectable using our hand-picking technique. The only vesicular acritarchs recovered in our samples were clusters of smooth-walled forms (Fig. 11N-S) which are locally abundant throughout the File Haidar Formation, but become concentrated around horizons yielding metazoan SCFs (Fig. 3). Individual clusters often exceed 100 lm (rarely > 500 lm) in maximum dimension. The absence of consistent cell counts or observable inter-cell connections argues against a green-algal-type coenobial habit (cf. Butterfield 2009) and limits taxonomic resolution; comparable clusters from Proterozoic and Palaeozoic assemblages are usually referred to as Synsphaeridium (e.g. Riedman et al. 2014, fig. 2) but remain of unknown biological affinity.

DISCUSSION
The array of SCFs described here presents a fundamentally new picture of early Cambrian palaeobiology in the File Haidar Formation and, by extension, Baltica. Although a non-biomineralizing fauna has long been apparent in the File Haidar from its trace fossils (e.g. Jensen 1997), there is a conspicuously limited record of associated body fossils, with the rare though notable records of the arthropod Paleomerus and lobopodian Xenusion (Størmer 1956;Bergstr€ om 1971;Krumbiegel et al. 1980;Dzik & Krumbiegel 1989). Our data provide the first direct record of priapulids, annelids and wiwaxiids in Baltica, along with a range of other metazoans and nonmetazoan forms. At the same time, the broad lithostratigraphic and geographic coverage of our dataset provides a novel view of the local to regional controls on fossil distribution (Figs 1, 3).

Stratigraphical and geographical occurrence
All stratigraphical members of the File Haidar Formation yielded SCFs, but with significant internal partitioning. Wiwaxiid and protoconodont remains were only found in samples from proximal environments (the Lingulid Sandstone and N€ ar Shale of the B a-2 and BH-1 cores; Fig. 3), while samples bearing annelid and priapulid remains were confined to outboard settings and tended to co-occur with concentrations of Baltinema rana gen. et sp. nov. and acritarch clusters (the Viklau and N€ ar Shale of the FH-1 and Gr-1 cores; Fig. 3). Potentially, these patterns reflect regional palaeobathymetric controls on distribution (see Nielsen & Schovsbo 2011). No SCFs were recovered from coarser sandstone units, but even within silt-and mud-dominated facies, productive samples were limited to stratigraphically discrete horizons separated by extended intervals noticeably devoid of fossils (Fig. 3). One of the most recurrent features shared by fossil-bearing samples is the localized abundance of cross-cutting and vertical burrows, often accompanied by the presence of sand-sized glauconite (typically in mudstone-hosted lenses), phosphorite clasts, and syneresis or dewatering cracks filled with coarser sand (Figs 3, 12; see Nielsen & Schovsbo 2011). Termed Kr aksten or 'crow-rock' in the regional literature (e.g. Hadding 1924), these conspicuously bioturbated and glauconitic sediments have been interpreted as sub-sequence drowning surfaces (cf. Possamentier & Allen 1999), where local sediment supply is outstripped by sea level rise (Nielsen & Schovsbo 2011;Alvaro et al. 2016). Time-averaging during such episodes of sediment-starvation could conceivably account for the increased bioturbation as successive communities of burrowers colonized the same sediment tier (Bromley 1996;McIlroy 2004).
The occurrence of organic-walled fossils in bioturbated sediments is unexpected, particularly in light of the strong negative correlation between Burgess Shale-type preservation and bioturbation (Orr et al. 2003). Recognizable ichnogenera associated with SCF-bearing horizons in the File Haidar Formation cores include Teichichnus, Treptichnus and Gyrolithes. Insofar as at least some treptichnid ichnofossils may be the traces of priapulid-like worms (Orłowski & _ Zyli nska 1996;Dzik 2005Dzik , 2007, and Cambrian priapulids are known to be generalist predators/scavengers (Huang et al. 2004;Vannier 2012), it is notable that all of the priapulid sclerites and annelid chaetae in this study were recovered from burrowed sediments (see Fig. 3). Unlike their macrofossil counterparts, SCFs are relatively immune to physical disturbance, moderate levels of time averaging and even gut-passage (e.g. Vannier 2012). As such, the enhanced recovery of SCFs in bioturbated horizons may well derive from their localized concentration by scavenging metazoans, with or without the compounding effects of low sediment input and associated time averaging. To this extent, the distribution of SCFs is often more comparable to that of biostratinomically recalcitrant shelly fossils than to delicate articulated carcasses or moults, a factor with important implications for future studies.

Global significance
With a Cambrian palaeolatitude of 30-60°and clear separation from any other contemporaneous cratons, Baltica offers an important new view of early palaeobiogeographical partitioning. Our recovery of diverse SCFs from the File Haidar represents the first significant occurrence of non-biomineralizing 'Burgess Shale-type' metazoans from this palaeocontinent. New sclerites from Baltica confirm the presence of Wiwaxia on every major Cambrian craton (Fig. 13). With the addition of Baltiscalida njorda gen. et sp. nov., the record of Cambrian priapulids also acquires a fully global coverage (Fig. 13). In contrast to the cosmopolitan wiwaxiids and priapulids (Figs 13, 14), the chaetae of annelids have not featured in dispersed SCFs assemblages before now, despite their localized extraction from Burgess Shale macrofossils (Butterfield 1990b) and their demonstrable recalcitrance in actualistic taphonomic studies (Briggs & Kear 1993). Indeed, Baltichaeta jormunganda gen. et sp. nov. represents one of only eight known Cambrian polychaete taxa, and the only one known from beyond palaeotropical, macroscopic BST-Lagerst€ atten (Figs 13, 14).
Surprisingly, the File Haidar assemblage is broadly comparable to lower palaeolatitude BST 'worm'-biotas from Laurentia and South China, which also exhibit a predominance of priapulid, palaeoscolecid, protoconodont, wiwaxiid and polychaete 'worms ' (cf. Conway Morris 1989a, b;Han et al. 2008). At least at these coarse taxonomic levels, there appears to be little indication of palaeobiogeographical or palaeolatidudinal partitioning; in marked contrast to the pronounced endemism displayed by many contemporaneous shelly taxa (cf. Samson et al. 1990;Meert & Lieberman 2008). The File Haidar SCFs do, however, exhibit significant differences at lower taxonomic levels. Although priapulids and polychaetes are present, none of these is directly comparable to previously described species or genera. Both Baltiscalida and Baltichaeta could well be endemic to Baltica and/or high palaeolatidudes (cf. Smith et al. 2015b), though limited sampling leaves this open to further validation. Certainly the macrofossil record of polychaetous annelids in the Cambrian, currently known from just four biotas worldwide, is too sparse to derive a reliable biogeographic signal (Figs 13, 14). Even so, there are strong grounds for expecting inter-regional distinctions in species distributions (Meert & Lieberman 2008;Alroy 2010), and the fine level of detail provided by SCFs has the potential to illuminate such patterns (cf. Smith et al. 2015b). F I G . 1 2 . Microfacies typical of fossiliferous/SCF-productive horizons. A-C, core material from Gr-1 core: A, 645.65-645.70 m, containing syneresis cracks and Teichichnus burrows; B-C, close-ups of A, showing syneresis cracks in finely laminated mudstone (green-grey) filled with coarser sandstone (pale yellow), these have subsequently been deformed through compaction. D, plan view of bioturbated core from 479.03 m of the FH-1 core. Scale bars represent: 1 cm (A, D); 4 mm (B, C).
Despite the predominance of priapulids (including palaeoscolecids) in almost all Burgess Shale-type 'worm' biotas (Conway Morris 1979a, b;Zhang et al. 2006;Harvey et al. 2012a;Smith et al. 2015b) it is notable that both priapulids and polychaete annelids have been recovered together from the File Haidar sediments; indeed, B. jormunganda gen. et sp. nov. chaetae are locally far more abundant, and often better preserved, than the associated priapulid and wiwaxiid sclerites. Such expression rules out differential histology as a significant factor in the global expression of Cambrian polychaetes, pointing instead to their general marginalization in priapulid-dominated Cambrian oceans (Conway Morris 1979a, 1989b. At another level, however, it is clear there are fundamental differences between the 'typical' Burgess Shale-type biotas found in deeper water, low-oxygen (dysaerobic) environments, and those from shallower more fully oxygenated facies (Butterfield & Harvey 2012). The epicratonic Mount Cap and Deadwood biotas of western Canada, for example, include a surprisingly modern diversity of crustacean and molluscan SCFs, none of which appear in contemporaneous deeper water biotas (Harvey & Butterfield 2008;Harvey et al. 2012b). Intriguingly, both the Mount Cap and Deadwood successions contain bioturbated and syneresis-cracked glauconitic sandstone horizons, directly comparable to the fossiliferous 'Kr aksten' facies in the Baltic Basin. Once again, it is too soon to draw broad conclusions, but the discovery of a new polychaete taxon in well oxygenated/bioturbated horizons of the File Haidar Formation suggests that the current record is steeply biased in favour of dysaerobic, priapulid-dominated environments conducive to the preservation of Burgess Shale-type macrofossils.

SYSTEMATIC PALAEONTOLOGY
We establish three new monospecific genera based on SCF material recovered from the File Haidar Formation. All specimens are permanently mounted on glass slides and are housed in the Swedish Museum of Natural History, Stockholm (NRM). Although these taxa are established on the basis of disarticulated elements, they are sufficiently distinctive to be used, in practice, as natural biological taxa. Like Halkieria (Conway Morris & Peel 1990) and the conodont animal (Briggs et al. 1983), we expect the whole-organism morphology associated with these SCFs to be discovered in due course. Our approach reflects prevailing practices in the taxonomy of SSFs derived from multi-element scleritomes (e.g. Skovsted & Peel 2007;Topper et al. 2009 LSID. urn:lsid:zoobank.org:act:DE398F3F-38A7-4D54-B7AA-70666CDB590E Derivation of name. With reference to the Baltic Basin.

Diagnosis. As for type species.
Baltiscalida njorda sp. nov. Derivation of name. After the sea deity, Njord, in Norse mythology.

Diagnosis.
A priapulid or priapulid-like scalidophoran with pharyngeal teeth that have an acute arch tapering to a central prong which is aligned with the angle of the arch; at least three tooth morphotypes present. In all teeth the arch is denticulate, with hollow distally-projecting stout marginal denticles which taper to a point and occasionally branch. Type 1 teeth with a narrow elongate outline and numerous short, closely-spaced denticles extending almost to the tip. Type 2 teeth with an equant outline and medium to long denticles arranged along a size gradient with the longest denticles nearest the prong. Type 3 teeth with an elongate outline and sparse denticles of moderate length with no size gradient in their distribution, and typically asymmetrically arranged with a gap between the distalmost denticles and the arch apex/prong. Remarks. The three morphotypes of isolated pharyngeal teeth of B. njorda gen. et sp. nov. commonly co-occur in samples and occupy a range of morphologies that is typically observed within a single species of priapulid. In addition, the three principle tooth morphotypes commonly co-occur with other tooth-and scalid-like morphologies (Figs 6, 7) which potentially belong to the same species; these other sclerites, however, lack distinctive features and are not formally included in the description. Occasionally smaller sclerites (e.g. Fig. 5M) display an intermediate morphology between types 2 and 3 teeth, probably reflecting ontogeny or variation along the introvert. The basal pad in B. njorda teeth is usually poorly preserved, suggesting it was comparatively labile in contrast to the heavily sclerotized spinose portions. A possible 'spur' may be developed in some specimens and is usually visible in co-occurring scalids; this is commonly adpressed to the remainder of the sclerite, however, obscuring its original morphology and orientation. LSID. urn:lsid:zoobank.org:act:C082DEC4-7DAB-4C75-8649-23FEE6E1135B Derivation of name. With reference to the Baltic Basin.
Diagnosis. As for type species.
Baltichaeta jormunganda sp. nov. Diagnosis. An annelid with simple comb chaetae. Bladelike shafts terminate distally in extended lateral prongs, separated by an interval of fine serrations. Prongs always project beyond the brush tip and may be either symmetrically or asymmetrically developed to different degrees.
Serrations of the comb have acute tips in larger specimens or sub-rounded tips in smaller forms. Comb may be laterally divided into thirds by two prominent serrations approximately twice the size of the average for the comb. Densely-packed linear striations run the internal length of the chaetae, perpendicular to this is a series of equally spaced horizontal ridges, spaces between ridges increasing with respect to the size of the chaetae.
Diagnosis. As for type species.
Baltinema rana sp. nov. Diagnosis. Filamentous organisms consisting of a sinuous primary axis bearing medially distributed secondary sinuous branches and globose lateral outgrowths. Length highly variable but always considerably exceeds the width. Main axis has an irregular 'pock-marked' appearance and is typically darker than the secondary structures. Apical terminations of primary axis taper to a rounded tip. Globose outgrowths roughly the width of the parent axis occur either in isolation or in pairs on opposing sides of the main axis. Branches resemble smaller versions of the primary axis and occasionally sub-branch but do not bear globose extensions.
Remarks. The biological affinity of these forms is unresolved. The thick, wrinkled, central axis of B. rana does not appear to be composed of a compound cluster of filaments, but instead has conspicuous surficial and internal pock-marked cavities throughout. In overall architecture B. rana bears similarities to certain Ediacaran 'vendotaenids' such as Eoholynia (see Urbanek & Rozanov 1983, fig. XLV-1;Gnilovskaya et al. 1988) and other Neoproterozoic microfossils such as Pseudodendron (see Butterfield et al. 1994, fig. 28); however, it is significantly smaller than the dichotomously branching Eoholynia, and Pseudodendron lacks globose outgrowths.