The Svanbergfjellet Formation: eukaryotic life in a Tonian Sea Open Access

The Svanbergfjellet Formation, part of the Akademikerbreen Group, is exposed across Svalbard's Olav V Land and Ny Friesland in northeastern Spitsbergen, as well as across western Nordaustlandet (although here partially under the name Hunnberg Formation, Wilson 1961; Flood et al. 1969; Harland 1997; Halverson et al. 2004, 2007) (Fig. 1a–c). It is split into four informal members (Lower Dolostone/Lower Dolomite, Lower Limestone, Stromatolitic Dolostone/Algal Dolomite, Upper Limestone; see Butterfield et al. 1994; Harland 1997) and ranges from 600 m thick in Olav V Land to 100 m in northern Ny Friesland (Wilson 1961; Knoll and Swett 1990). Strata are predominantly limestone and dolostone, with minor calcareous mudstone and shale (Wilson 1961; Flood et al. 1969). The prevalence of planar to wavy microbialites, diverse stromatolites, subaerial exposure surfaces, centimetre-scale alternations between nodular limestone and shales, and thicker accumulations of calcareous and non-calcareous mudstone suggest that the formation records an array of shallow-marine depositional environments, ranging from subtidal to supratidal settings (Knoll and Swett 1990; Butterfield et al. 1994; Halverson et al. 2007).

The base of the formation contains the c. 810–790 Ma Bitter Springs carbon isotopic excursion (Swanson-Hysell et al. 2015; Halverson et al. 2018; Shields et al. 2022) and a Re–Os depositional age of 791.1 ± 4.9 Ma has been reported from the base of the Algal Dolomite member (Zhang et al. 2023). Diamictites of the Polarisbreen Group, ascribed to the c. 720–635 Ma Snowball Earth events, overlie the formation (Millikin et al. 2022). A Bayesian age–depth model constrains the duration of Svanbergfjellet sedimentation between 802.74 + 10.88/−8.65 Ma and 783.97 + 7.29/−10.93 Ma in the Tonian (Zhang et al. 2023).

Organic-walled microfossils have been documented throughout the Svanbergfjellet Formation in mudstones and early diagenetic cherts within carbonate strata, with three ‘principal’ mudstone horizons or localities from Spitsbergen proving especially productive and biodiverse (Fig. 2) (Butterfield et al. 1994). P2945 is from the Lower Dolomite member along the Polarisbreen glacier (79°41′30″N, 18°12′E; Butterfield et al. 1988, 1994). 86-G-62 is from the Algal Dolomite member at De Geerbukta (79°35′20″N, 17°42′56″E; Butterfield et al. 1988, 1994). 99-L-18 is from the Algal Dolomite member between De Geerbukta and Faksevågen (79°35′N, 17°44′E; Butterfield 2004). Less diverse fossil assemblages have been described from ‘non-principal’ horizons (mudstones and cherts) at various localities, including Svanbergfjellet (78°41′30″N, 18°14′E; Butterfield et al. 1988, 1994), and at localities around Murchisonfjorden (79°57′N, 18°27′E; Knoll 1982, 1984; Knoll and Calder 1983).

We focus on the ‘principal’ mudstone horizons within the Algal Dolomite member (86-G-62 and 99-L-18). The Algal Dolomite member is dominated by six large (8–10 m thick) stromatolitic bioherms with interbedded mudstones (Knoll and Swett 1990) (Fig. 1b). Microfossils derive from mudstone samples that sit directly on top of the stromatolitic bioherms (Box 1). The fossiliferous mudstones are commonly fissile, medium grey or green, with darker millimetre-scale laminations (Butterfield 2004).

The Re–Os radiometric date from the base of the Algal Dolomite member provides an absolute maximum age constraint of 791.1 ± 4.9 Ma on all the fossil horizons in the member (Zhang et al. 2023). We provide estimates for the ages of the ‘principal’ horizons using the Bayesian model of Zhang et al. (2023) (Fig. 1d; see Methods and Supplementary material for details).

In 2018, to ensure detailed geological context for this review, we re-collected the 86-G-62 horizon at De Geerbukta as a 4.6 m thick decimetre-resolution sample suite (n = 25 samples) named RPA1802 (79°35′20″N, 17°42′56″E). Additionally, we sampled a 0.4 m thick suite (n = 9 samples), also at decimetre resolution, in a new Algal Dolomite member locality at Freken (79°25′26″N, 17°55′56″E), named RPA1801. Both suites begin with 0 m at their contacts with their underlying stromatolitic bioherms (Fig. 1b).

For examination via optical and electron microscopy, organic-walled microfossils were liberated from the samples using 48% hydrofluoric acid at the University of Oxford. All new fossils are reposited at the Museum of Natural History, University of Oxford. Full details of the methods used to extract, image and analyse new organic-walled microfossils, as well as details of their repository and a discussion of our age model for the Formation can be found in the Supplementary material.

Based on the taxonomic assignments by Butterfield et al. (1994) as well as taxonomic placements from recent literature (e.g. Porter and Riedman 2016), we categorized all Svanbergfjellet microfossils (new fossils and those reported in previous work) into probably eukaryotic v. prokaryotic affinities (Fig. 2a). Leiospheres are probably polyphyletic and may represent eukaryotes or prokaryotes (Butterfield et al. 1994; Butterfield 1997; Talyzina and Moczydowska 2000; Willman and Moczydłowska 2007; Javaux and Knoll 2017).

The Algal Dolomite member harbours at least 21 species with probable eukaryotic affinities and 16 with probable prokaryotic affinities (Fig. 2). Despite the ‘principal’ horizons or localities across Ny Friesland capturing most of the microfossil diversity within the formation, four probably eukaryotic and six probably prokaryotic species are present only in ‘non-principal’ horizons or localities (Butterfield et al. 1994).

Re-sampling horizon 86-G-62 in RPA1802 (>14 total species, 819 microfossil specimens, 25 samples) shows that microfossil diversity not only varies between localities and horizons, but also stratigraphically at a decimetre scale within the ‘principal’ horizons (Fig. 2). The samples with highest fossil diversity are RPA1802 0.1, 1.0 and 1.2 m, with up to 10 species documented in each. Notably, the 0.1 m sample hosts the delicate, thin-walled and multicellular Palaeastrum dyptocranum (10 specimens). Samples 1.0, 1.2 and 3.7 m contain the equally delicate Proterocladus major. Several samples (0.6, 0.9, 1.4, 3.1, 3.2, 3.9 and 4.6 m), however, contain no microfossils. Consistently present throughout fossiliferous RPA1802 samples are the acritarchs Cerebrosphaera globosa (=C. Buickii; Butterfield et al. 1994), Leiosphaeridia spp. and Trachyhystrichosphaera aimika, possibly indicating that either they were ecologically ubiquitous or they had higher preservation potential (Cornet et al. 2019; Pang et al. 2020). The variability in microfossil diversity and abundance at this decimetre stratigraphic scale is not surprising; Butterfield (1992) found variability between thin sections that were taken from different chips of a single hand sample of 86-G-62, and Butterfield et al. (1994) captured up to 22 fossil species in 86-G-62 absent from our RPA1802 re-sampling (Fig. 2). This variation in diversity could be due to finer scale mineralogical heterogeneities influencing fossil preservation (Anderson et al. 2020), or it could represent variation in productivity where the taxonomically richest rock chips capture highly productive palaeoenvironments (Butterfield et al. 1994).

The new locality at Freken, RPA1801, is taxonomically depauperate compared with 86-G-62 or RPA1802 (more than eight total species, 223 microfossil specimens, nine samples). The highest diversity (more than eight species) is recorded in the 30.0 cm sample, yet most samples preserve only Cerebrosphaera globosa and Leiosphaeridia spp. Delicate taxa such as Palaeastrum dyptocranum and Proterocladus major are absent. This locality yields the only taxon recovered in our work that was not recovered in previous work: Trachyhystrichosphaera botula (Tang et al. 2013; see Box 2), of which there is one specimen in sample RPA1801 30 cm (Fig. 3b).

Beyond the Algal Dolomite member, Butterfield et al. (1994) alluded to palaeoenvironmental controls on taxonomic distributions. Comparing the size-frequency of planktic leiospheres from 86-G-62 (Algal Dolomite member) and P-2945 (Lower Dolomite member), they argued the relatively more even distribution of sizes in P-2945 reflected a more distal palaeoenvironment, possibly mid-shelf (Butterfield and Chandler 1992), relative to 86-G-62. Additionally, few benthic taxa were reported from P-2945, further supporting a distal setting (e.g. lower relative abundances of filamentous microbial mat builders) (Butterfield et al. 1994).

Svanbergfjellet fossils are of critical importance to our understanding of early eukaryote evolution. The ‘principal’ horizons or localities in the Algal Dolomite member preserve species of one of the three known pre-Ediacaran fossils that, from species preserved elsewhere, has unambiguously been identified as a crown eukaryote: Proterocladus, a green alga. They also preserve Palaeastrum dyptocranum, probably a green alga, along with several fossils with complex morphologies and plausible claims on crown eukaryote affinities. Below, we discuss how each of these fossils has been interpreted and, for those with plausible but ambiguous affinities, highlight where further study might resolve their position on the tree of life.

Proterocladus, represented in the Algal Dolomite member by P. major (RPA1802; 86-G-62), P. minor (86-G-62) and P. hermannae (86-G-62), is an epibenthic, multicellular, filamentous branching form, with cells of unequal length, and constriction at the cell–cell septa (Fig. 4i–k). Eleven specimens of P. major were recovered in our RPA1802 re-collection. Originally identified as a siphonocladalean chlorophyte, similar to Cladophoropsis or Cladophora (Butterfield et al. 1988, 1994), the few morphological characters (e.g. cells of unequal length) of the Svanbergfjellet specimens were insufficient to support this phylogenetic placement unambiguously (Berney and Pawlowski 2006; Graham 2019; Del Cortona et al. 2020). Recent discovery of P. antiquus in North China (Tang et al. 2020; Li et al. 2023) provided further characters for the phylogenetic placement of Proterocladus, including a subdiscoidal holdfast, heteromorphic akinete-like cells, lateral pores in possible reproductive cells, a gradual increase of cell width distally and apical cells that are round or capitate with an occasional narrow extension at the distal end. Proterocladus antiquus is now confidently identified as a total group chlorophyte (or at the least total group green algae) and may even sit within a total group Cladophorales (Tang et al. 2020).

Because Svanbergfjellet Proterocladus specimens lack the additional identifying features of P. antiquus, the precise taxonomic relationship between the four species of Proterocladus remains to be determined (they may be synonymous; see Tang et al. 2020). The simplicity of the Svanbergfjellet specimens means they may even represent a eukaryote lineage unrelated to green algae that converged upon their simple form (Knoll 2014; Butterfield 2015). A fruitful avenue of future work is detailed comparison of the most complete Svanbergfjellet and North China material, placing each in their palaeoecological context (Tang et al. 2020).

Palaeastrum dyptocranum is a monostromatic (sheet-like), multicellular microfossil with numerous spheroidal cells connected by ellipsoidal plaques first reported from the Algal Dolomite member (Butterfield et al. 1994; Porter and Riedman 2016). Some examples show the sheets of cells arranged as enclosed vesicles (Butterfield 2009). The cellular organization of P. dyptocranum prompted initial comparisons with the coenobia of some extant freshwater chlorococcalean green algae; for example, modern Coelastrum (Scenedesmaceae), Pediastrum (Hydrodictyaceae) and Hyrdodictyon (Hydrodictyaceae) (Butterfield et al. 1994; Butterfield 2009), potentially adding to the diversity of green algae in Svanbergfjellet (Butterfield 2015). Palaeastrum dyptocranum is now also known from Canada (Loron et al. 2019b), Russia (Vorob'eva et al. 2015) and the USA (Porter and Riedman 2016).

In specimens (n = 23) from our RPA1802 re-collection, the opacity of cells varies between specimens, which may have histological significance (Butterfield et al. 1994). Cells are commonly attached via plaques to three further cells, but some connect with up to six. Between specimens, cell colonies have a multimodal distribution of cell maximum diameters. Four specimens have cells with mean size of 12 ± 2 $μ$m, n = 1130 cells (Mode 1), whereas four specimens have mean cell size of 18 ± 3 $μ$m, n = 487 cells (Mode 2). Cell size generally decreases with cell number within specimens (Mode 1: mean cell number per specimen is 282; Mode 2: mean cell number per specimen is 122).

The cell size v. cell number trend within colonies of Svanbergfjellet P. dyptocranum is inverse to that generally found in colonial Cambrian green algal microfossils from the Forteau Formation of Canada, although the trend can be variable between Cambrian specimens (Harvey 2023). Compared with these Cambrian microfossils as well as the modern Coelastrum and Pediastrum, P. dyptocranum does not have a fixed coenobium size (Butterfield 2009). Instead, cell colonies with >100 cells and hollow ellipsoids within its structures (Fig. 4a) have features akin to modern Hydrodictyon, which forms planar hexagonal colonies of elongated cells with two cellular attachments on each end (Butterfield 2009). However, 3D coenobia are not exclusive to Hydrodictyon, having evolved multiple times within Hydrodictyaceae (Buchheim et al. 2005), and the circular attachment plaques are more reminiscent of Coelastrum (Butterfield et al. 1994).

These shared characteristics of coenobial size and development among multiple potential modern green algal counterparts challenges the precise phylogenetic placement of P. dyptocranum (Porter and Riedman 2016). Given the simplicity of its construction, convergence on its morphology even by an unrelated early algal lineage cannot be excluded (Lewis and McCourt 2004; McManus and Lewis 2011; Butterfield 2015). Further complicating placement, cellular organization of colonial green algae may reflect biological factors (physiology, function, regulation of internal structural stress during growth) (Bell and Mooers 1997; Jacobeen et al. 2018; Day et al. 2022), but colony size can also be a function of environmental stressors such as temperature, salinity and pH, masking phylogenetic signal (e.g. González-Hourcade et al. 2023). Palaeastrum shows that colonial construction in eukaryotes was well established by the Tonian (building on discoveries in much older rocks, see Miao et al. 2024), but until the diagnostic characters between colonial green algal lineages are discerned, its precise phylogenetic placement is elusive (Butterfield et al. 1994; Butterfield 2009; Harvey 2023; Riedman et al. 2023).

Germinosphaera, represented by G. bispinosa, G. fibrilla and G. jankauskaii, has long process-bearing vesicles that easily distinguish it from other ornamented acritarchs within Svanbergfjellet (Butterfield et al. 1994) (Fig. 4h). These processes are non-septate and typically non-branching depending on species, with the number of processes per vesicle varying from one to six. Originally the random nature of the processes was thought to indicate that Germinosphaera was an ‘actively growing vegetative structure’ comparable with the modern germinating form of Vaucheria, a xanthophyte alga (Butterfield et al. 1994). However, the variation in Germinosphaera morphology requires further exploration on a population level as comparisons have also been made with fungal spores (e.g. those of Glomeromycota) based on the presence of the tubular processes (Butterfield 2005; Retallack 2015; Loron and Moczydłowska 2018; Miao et al. 2019). No Germinosphaera specimens were found from RPA1801 or the RPA1802 re-collection despite their abundance in the original 86-G-62 material.

Jacutianema solubila, a botuliform multicellular eukaryote, has tentatively been proposed as the Gongrosira-phase of a vaucheriacean xanthophyte alga (Butterfield 2004). It has been most intensely studied from 99-L-18 (Butterfield 2004), but is now documented from RPA1802 in two samples (n = 33 specimens) (Fig. 4n), as well as the Neoproterozoic of Arctic Canada (Loron et al. 2019b). It occasionally possesses secondary cysts and filamentous protrusions, combined with large size, distinguishing J. solubila from heterocystous cyanobacterial akinetes (Butterfield 2004). The multicellularity of Jacutianema is unconventional and distinct from other multicellular forms in the Svanbergfjellet Formation. The 99-L-18 specimens lack a direct cytoplasmic connection via plasmodesmata that would imply intercellular communication (Butterfield 2004). In contrast, Proterocladus and Palaeastrum both feature septae or ‘plaques’ between cells, which demonstrate direct cell–cell contact. Butterfield (2004) did, however, argue that cells exhibiting incomplete division suggest a channel for intercellular communication and that the nature of the intercellular constrictions can be compared with the Gongrosira-phase (a vegetative phase so called because it resembles Gongrosira, a branched filamentous chlorophyte) of the extant xanthophyte alga Vaucheria, despite little similarity between the 99-L-18 specimens and Vaucheria in its non-vegetative state. A tentative assignment of Jacutianema to the xanthophyte algae, or as a stem lineage, is consistent with molecular clock predictions of the maximum origin of stem ochrophytes (the phylum encompassing xanthophytes) at c. 1300 Ma (Strassert et al. 2021; Choi et al. 2024). However, Tang et al. (2020) used their co-occurrence in Svanbergfjellet and similar siphonous or siphonocladous construction to suggest J. solubila as the akinete of Proterocladus. This green algal hypothesis merits further investigation: in our RPA1802 dataset, P. major and J. solubila were almost always recovered from the same samples (Fig. 2a), possibly supporting a link, although their co-occurrence may be a result of shared taphonomy and habitat.

Specimens of Valkyria borealis represent the most complex multicellularity within Svanbergfjellet with at least six different cell types, a diversity of cell types similar to that of some modern algae and fungi (Butterfield et al. 1994). Butterfield et al. (1994) speculated that if these cell types were specialized, V. borealis could represent the oldest fossil with tissue grade organization, a level of complexity not known until the Ediacaran Biota, over 200 myr later. Further analysis of V. borealis is required, using a wider variety of tools than was readily available in the 1990s, to reveal the nature of its cell types and whether they may have been specialized; for example, electron microscopy and Raman, IR and X-ray absorption spectroscopy (Javaux et al. 2004; Marshall et al. 2005; Loron et al. 2019a, 2022; Sforna et al. 2022). Valkyria borealis is the rarest taxon in our RPA1802 re-collection, with only a single specimen found in sample 3.7 m, adding to the 86 specimens from 86-G-62 (Fig. 4f). Our specimen is assigned based on its long tomaculate morphology, at least three cellular attachment points from the central body indicating multicellularity and a potential medial stripe. Fieldwork would be required to collect new population-scale material.

The eukaryotic taxonomic richness of the Svanbergfjellet Formation (>21 distinct taxa) is c. 50% greater than the average Proterozoic fossil assemblage (about eight taxa) (Cohen and Macdonald 2015). Its fossils are preserved with a morphological fidelity that is rivalled by few Proterozoic mudstones (Butterfield 1995), with detailed taphonomic comparisons undertaken with the Mesoproterozoic–Tonian Lakhanda Group and the Tonian Wynniatt Formation (Anderson et al. 2020). Whether there is any secular pattern to the distribution of Proterozoic mudstone-hosted Lagerstätten, and whether recently investigated biodiverse Proterozoic mudstone units (e.g. Porter and Riedman 2016; Loron et al. 2019a, b ; Tang et al. 2020) can also be considered Lagerstätten remains to be determined. Regardless, the Lagerstätte provides a key snapshot of how crown eukaryotic life had radiated by the Tonian. Indeed, its taxonomic richness may bias our view of whether and when Tonian eukaryotic biodiversity peaked prior to the Cryogenian Snowball Earth glaciations (e.g. Riedman and Sadler 2018). The high preservation quality of Svanbergfjellet microfossils means that several taxa have potential as bone fide representatives of crown eukaryotes and merit further investigation of their morphology as well as their organic composition using geochemical or spectroscopic techniques (Box 3). Resolving their affinities would substantially add to the known diversity of crown eukaryotes in the Tonian.

This research was inspired by the pioneering work of N. Butterfield, A. Knoll and K. Swett. We thank A. Knoll for field notes, T. Aasvoll for safe passage around Svalbard on Arctica I, G. Wedlake for support in fossil documentation, G. Hughes and P. Karamched for electron microscopy support, and K. Clayton, O. Green and S. Wyatt for support in fossil extraction. We are grateful to N. Butterfield for kindly providing fossil images for this review. Reviewers N. Butterfield and L. A. Riedman, as well as Editor F. Dunn, provided comments that greatly improved this paper.

SM: data curation (equal), methodology (equal), writing – original draft (lead), writing – review & editing (lead); AEGM: investigation (supporting), writing – review & editing (equal); TZ: investigation (supporting), writing – review & editing (equal); TMG: investigation (supporting), writing – review & editing (equal); ADR: conceptualization (equal), funding acquisition (equal), investigation (supporting), writing – review & editing (equal); NJT: conceptualization (equal), investigation (supporting), writing – review & editing (equal); KDB: conceptualization (equal), investigation (supporting), writing – review & editing (equal); JVS: conceptualization (equal), funding acquisition (equal), investigation (supporting), writing – review & editing (equal); RPA: conceptualization (equal), funding acquisition (equal), investigation (supporting), supervision (lead), writing – original draft (lead), writing – review & editing (equal).

All Souls College (RPA), the National Geographic Society (C-129R-17, J.V.S.), the National Science Foundation (EAR-1929593 and EAR-165012, J.V.S.; EAR-1929597, A.D.R.), and the Royal Society (RSG\R1\180077, URF\R1\221220 and RF\ERE\221057, R.P.A.) provided funding. Fieldwork was conducted through the Research in Svalbard programme (RiS ID 6867, 10295, 11035).

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All data generated or analysed during this study are included in this published article (and supplementary files). Newly collected fossils are currently held at the Oxford University Museum of Natural History. Details of the collection numbers are provided in the Supplementary material.