Reg. Charity No. 1168330
The Swedish Cambrian ‘Orsten’‐type fossil sites have yielded diverse secondarily phosphatized three‐dimensionally preserved microfossils, mainly of arthropod affinities. Similar material has also been recorded from Canada, the UK, Poland, Siberia, China and Australia. Only one other non‐arthropod group, the Cycloneuralia, is commonly reported from any of these sites, leading to the general assumption that ‘Orsten’‐type preservation is largely restricted to animals with a chitin‐containing cuticle. We describe here secondarily phosphatized, originally unmineralized, thread‐shaped fossils etched out of Cambrian ‘Orsten’‐type deposits from the Agnostus pisiformis Biozone of the Alum Shale Formation in Sweden. These fossils strikingly resemble specimens previously described from Precambrian deposits, with at least two different morphotaxa identified (Siphonophycus kestron Schopf and Oscillatoriopsis longa Timofeev & Hermann) as well as the modern Oscillatoria. This leads us to interpret the new fossils as unbranched, uniseriate filamentous cyanobacteria. Our morphological investigations, combined with morphometrics, allow grouping the specimens assigned to O. longa into two size classes, suggesting an even higher diversity within the ‘Orsten’ assemblages. The lack of cyanobacterial material in any sample younger than the A. pisiformis Biozone indicates that rather drastic changes occurred in the environment, that is, substrate conditions on the Alum Shale seafloor. This coincided with a significant change in the composition of the trilobite communities and onset of the globally recognized Steptoean Positive Isotope Carbon Excursion (SPICE) in Scandinavia.
Cyanobacteria arguably form one of the most widespread and successful systematic groups of prokaryotic organisms on Earth. Genetically, they comprise a very diverse group of organisms with a worldwide distribution in a broad range of habitats, being widespread in freshwater, marine and terrestrial ecosystems. Photoautotrophic, oxygen‐producing cyanobacteria are thought to have converted the early reducing atmosphere into an oxidizing one, causing the rusting of Earth (Schopf 2012) and the Great Oxygenation Event (e.g. Holland 2002, 2006). Therefore, they may even have created the conditions that directed the evolution of aerobic metabolism and eukaryotic photosynthesis.
Contrary to other modern bacteria, cyanobacteria have the capacity to perform oxygenic photosynthesis (Ormerod 1992). Cyanobacteria are even able to simultaneously perform photosynthesis and respiration within the same compartment (e.g. Shevela et al. 2013). Many representatives are also capable of using dinitrogen (N2) as the sole source of nitrogen for growth, either by means of specialized cell types or anaerobic and N2‐fixing processes (e.g. Stal 2000). This metabolic flexibility, together with a rather simple nutritional requirement for growth, has enabled cyanobacteria to thrive in a wide range of ecological niches (Whitton 1992). Cyanobacteria are basically unicellular, but many taxa possess a multi‐cellular organization of great structural diversity, similar to that of for example plants, fungi and metazoans.
Cyanobacterial cells may be rod‐like to spheroidal in shape and are either individual (e.g. in Prochlorales, Chroococcales and Pleurocapsales) or colonial, forming bundles (e.g. Oscillatoriales, Nostocales and Stigonematales) or spherical colonies of filamentous structures (e.g. Nostoc pruniforme C. Agardh ex Bornet & Flahault, 1888). The cyanobacterial wall is a complex multi‐layered envelope structure made of an outer membrane including lipopolysaccharides and proteins and an inner peptidoglycan (murein) layer (Stanier & Cohen‐Bazire 1977). Various fibril structures, involved in motility, may occur between these two layers or on top of the outer membrane (e.g. Hoiczyk & Baumeister 1995; Adams et al. 1999). Additional envelope material surrounding the cell wall may also occur, for example slime or mucilage, but not around individual cells, but rather enclosing whole strands of several cells (a trichome). The thread‐like morphology is the result of successive cell divisions occurring in a single plane at right angle to the main axis. Yet in some cases, cells have additional division plane(s) leading to more complex morphologies with true branching patterns.
Individual cells, cell strands and bundles of strands, may bear an additional envelope, or sheath, which can be distinguished from the cell wall and the slime layer. This varies from a thin and soft to a robust and relatively firm sheath. Such sheaths are often open‐ended tubes. The association of a trichome and a sheath is termed a filament. Among filamentous cyanobacteria, cells display great variability in size, shape and function. A cell chain comprises either exclusively vegetative cells of regular size, or it can be a mixture of vegetative cells and multiple specialized cell types known as heterocysts and akinetes, that is, thick‐walled resting cells that develop in particular environmental conditions (e.g. Adams & Duggan 1999).
The cyanobacterial fossil record is very extensive; microfossils with analogue morphologies to cyanobacteria have been recorded as far back as the Archean Eon, some 3.5 billion years ago (e.g. Schopf 1992a, b, c, 1993, 2000, 2006; Golubic & Seong‐Joo 1999; but see, e.g. Doolittle et al. 1996; Feng et al. 1997; Blank 2004; Dodd et al. 2017 for a discussion on the origin of Cyanobacteria). By the Archean, microbes were capable of forming highly organized communities, or biofilms. When interacting with their environment, such communities could produce organo‐sedimentary structures called microbialites, forming microbially induced sedimentary structures (MISS) and stromatolites (e.g. Pinckney et al. 1995; Grotzinger & Knoll 1999; Papineau et al. 2005; for a review see Stal 2000; Noffke & Awramik 2013, and references therein).
Putative and genuine cyanobacterial fossils are preserved in many Precambrian and Phanerozoic carbonate deposits, but they may also be permineralized in cherts (secondary silicification: Gunflint chert in Canada, e.g. Barghoorn & Tyler 1965, Awramik & Barghoorn 1977; Bitter Spring chert in central Australia, e.g. Schopf 1968, Schopf & Blacic 1971, Sergeev et al. 2012). They also occur in phosphorites (secondarily phosphatized) or can be preserved as organic compressions and impressions in shales. Yet due to their organic nature, the cyanobacterial fossil record, particularly from the early Cambrian and onward, is often restricted to sheath‐calcified taxa (Riding 1991, 1994), although specimens preserved either as compressions in shales or permineralized in cherts, are reported progressively more often (e.g. Volkova et al. 1983; Wang et al. 1983; Song 1984; Sergeev 1989; Yao et al. 2005; Dong et al. 2009; among many others).
Fossil cyanobacteria are usually rod‐like to spheroidal or filamentous, exhibiting uniseriate (e.g. Schopf 1968; Butterfield et al. 1994) to multiseriate (e.g. Seong‐Joo & Golubic 1998) aggregates of discoidal to cylindrical structures. It has been common practice to regard such fossils as prokaryotic cyanobacterial cells and/or filaments based on their close morphological resemblance to certain types of modern cyanobacteria. A phylogenetic affinity to Cyanobacteria is also supported by geological data pointing to the same environments as for modern counterparts (Schopf 1968). Due to the close similarity between Proterozoic and modern cyanobacteria, they are thought to have undergone very little or no evident morphological changes over many hundred millions of years (e.g. Schopf 1987, 1995).
Herein, we report and describe secondarily phosphatized filamentous ‘Orsten’ fossils from the Agnostus pisiformis Biozone (Guzhangian Stage; Cambrian Series 3) of the Alum Shale Formation in the Kinnekulle area, Västergötland, southern Sweden. After describing the main morphological characteristics, including a morphometric analysis, the systematic status of the filamentous forms, which we interpret as threads of filamentous cyanobacteria, is discussed on the basis of taphonomical, morphological and ecological criteria. The ecological significance of the presence of cyanobacteria in the Alum Shale biota is also discussed.
Material and methodThe Alum Shales Formation of Scandinavia, ranging from the Cambrian Series 3 to the early Ordovician (Tremadocian), is a highly fossiliferous condensed succession of fine‐grained, blackish bituminous mudstone and shale with very fine parallel lamination, dominated by polymerid trilobites, agnostoids (e.g. Clarkson & Taylor 1995; Terfelt et al. 2008, 2011; Clarkson 2011) and brachiopods (cf. Popov & Holmer 1994; see also Clarkson 2011, p. 15). In most parts of these strata, there is very little or no sign of disturbance of laminae due to burrowing. Intercalated thin limestone beds occur in some places throughout the formation (e.g. Martinsson 1974; Nicoll et al. 1992). Bituminous, early diagenetic limestone concretions, ‘Orsten’ (Henningsmoen 1974), are also common within these strata and form individual lenses and more or less continuous limestone beds (Martinsson 1974; Thickpenny 1984; Andersson et al. 1985; Bergström & Gee 1985; Buchardt et al. 1997). Klaus J. Müller (Bonn) began etching such nodules using weak acids in the mid‐1970s and yielded, in addition to a rich conodont fauna (e.g. Müller & Hinz 1991), large numbers of secondarily phosphatized and three‐dimensionally preserved fossils, now known as the ‘Orsten’‐type microfossils (Müller 1979, 1990; Müller & Walossek 1985a, 1991). Being mainly of arthropod affinity, these fossil assemblages turned out to be very diverse. They include various developmental stages with appendages of the most abundant and rock‐forming agnostoid euarthropod Agnostus pisiformis (Müller & Walossek 1987; Eriksson & Horn 2017), numerous developmental stages mainly of crustacean affinity (including the largest group, the phosphatocopines with more than 50 000 specimens), a larval pycnogonid chelicerate (Waloszek & Dunlop 2002) and even a minute representative of the early level of arthropod evolution, such as a ‘lobopodian’ (Maas et al. 2007a) and several species of the parasitic pentastomids (Walossek & Müller 1994; Castellani et al. 2011). Many of the ‘Orsten’ taxa seem to be representatives of meiofaunal communities that lived on or within an upper, soft sediment layer, the so‐called flocculent layer (Müller & Walossek 1991; Walossek 1993), but the morphology of some species like Dala peilertae Müller, 1982, or the later developmental stages of Rehbachiella kinnekullensis Müller, 1983 suggest good swimming abilities and a more epibenthic life above the bottom layer (Walossek 1993; unpubl. data).
The early projects on ‘Orsten’ fossils, in Sweden and elsewhere, focused primarily on phosphatized conodonts and arthropod fossil remains, and resulted in a bias of knowledge of other systematic groups that were represented in the material. More recent investigations of the acid‐resistant residues have not only increased the biodiversity of the ‘Orsten’ faunal associations considerably but also demonstrated that the preserved fossil material may not necessarily be secondarily phosphatic. Additional taxa described from the ‘Orsten’ of Sweden include brachiopods (e.g. Popov & Holmer 1994; Clarkson 2011, p. 15; Topper et al. 2013), echinoderms (Ahlberg et al. 2005; CC, pers. obs. 2009–2013), possible graptolites (Ahlberg et al. 2005) and sponges (Castellani et al. 2012). Such non‐arthropod fossils, together with other still unidentified organisms many of them also secondarily phosphatized (‘Problematica’; CC, pers. obs. 2009–2013) retrieved from ‘Orsten’ nodules, have expanded our view of life in the Alum Shale sea and provide a much better understanding of Cambrian biodiversity and ecosystems in general.
The ‘Orsten’ nodules yielding the study material were collected between 1975 and 1981 by Klaus J. Müller. A total of 16 rock samples from the Gum quarry at Kinnekulle, Västergötland, yielded 72 clusters of tubular microfossils (Table 1; Fig. 1), and collectively comprise more than 350 individual tubes. The tubes seem to consist of stacked cylindrical elements inferred to represent remnants of organismic cells. Their occurrence is restricted to the A. pisiformis Biozone. This zone has formerly been regarded as the lowest stratum (Zone 1) of the traditional Upper Cambrian of Sweden but is now considered to be the uppermost zone of the Guzhangian Stage of Cambrian Series 3. The subsequent part of the sequence (zones 2–6 according to traditional terminology) now belongs to the Furongian Series (see Peng et al. 2004, 2006; see also Ahlberg 2003; Terfelt et al. 2008, 2011; Nielsen et al. 2014 for a review of the Cambrian biostratigraphy in Scandinavia, and for an overview of the Alum Shale Formation see Bergström & Gee 1985; Buchardt et al. 1997). Insoluble residues of ‘Orsten’ nodules from biozones other than A. pisiformis were also screened, but did not yield any similar fossils. All specimens retrieved are three‐dimensionally preserved and apparently secondarily phosphatized in a mode similar to that of most other ‘Orsten’ microfossils.
Table 1. Distribution and relative frequency of clusters of cyanobacterial remains in rock samples retrieved at the Gum quarry at Kinnekulle, Västergötland, Sweden (Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone) Rock samples No. clusters Frequency (%) 6364 1 1.39 6404 6 8.33 6409 17 23.61 6410 5 6.94 6414 17 23.61 6415 3 4.17 6416 1 1.39 6417 10 13.89 6749 1 1.39 6760 1 1.39 6761 3 4.17 6772 1 1.39 6773 1 1.39 6783 1 1.39 6784 3 4.17 6787 1 1.39 Total 72 100 Figure 1 Open in figure viewerPowerPoint Locality maps and sampling areas. A, map of Scandinavia with the distribution of the major occurrences of the Alum Shale Formation (black; modified from Martinsson 1974); the sample area is marked with the letter B (Kinnekulle area, Västergötland, Sweden, GPS coordinates: 58.583 N, 13.383 E]. B, locality map of the Kinnekulle area with ‘Orsten’‐type deposits; 1, Gössäter; 2, Toreborg; 3, Österplana; 4, Haggården‐Marieberg; 5, Ödbogården; 6, Brattefors; 7, Sandtorp; 8, Stubbegården; 9, Gum (approx. GPS coordinates: 58.521 N, 13.346 E]; 10, Ekebacka; 11, Backeborg; 12, Klippan; 13, Kakeled; 14, Pusabäcken; 15, Trolmen (modified from Müller & Hintz 1991; Maas et al. 2003). Rock samples with remains of filamentous cyanobacteria derive exclusively from the Gum quarry (see also Table ). CaptionLocality maps and sampling areas. A, map of Scandinavia with the distribution of the major occurrences of the Alum Shale Formation (black; modified from Martinsson ); the sample area is marked with the letter B (Kinnekulle area, Västergötland, Sweden, GPS coordinates: 58.583 N, 13.383 E]. B, locality map of the Kinnekulle area with ‘Orsten’‐type deposits; 1, Gössäter; 2, Toreborg; 3, Österplana; 4, Haggården‐Marieberg; 5, Ödbogården; 6, Brattefors; 7, Sandtorp; 8, Stubbegården; 9, Gum (approx. GPS coordinates: 58.521 N, 13.346 E]; 10, Ekebacka; 11, Backeborg; 12, Klippan; 13, Kakeled; 14, Pusabäcken; 15, Trolmen (modified from Müller & Hintz ; Maas et al. ). Rock samples with remains of filamentous cyanobacteria derive exclusively from the Gum quarry (see also Table ).Acid etching was done in the late 1970s and early 1980s by Müller and his workgroup. Subsequently, the microscopic fossils (size below 2 mm) were sorted into small jars for future use, mainly using Zeiss binocular lenses; selected specimens were sorted into plastic micro slides or were glued onto stubs for inspection under a scanning electron microscope (SEM; for details see Müller 1979, 1982, 1985). This early work yielded some of the tubular material described herein. Many additional specimens (= clusters) were later picked by one of the authors (CC) from Müller's sample residues using a Leica MS 5 stereo microscope in the Workgroup Biosystematic Documentation at the University of Ulm, Germany. The newly recovered specimens were mounted onto SEM stubs for detailed investigation. All specimens were photographed using a Zeiss DSM 962 SEM at the Central Unit for Electron Microscopy, University of Ulm. Subsequently, the SEM images were processed with the image‐processing software GraphicConverter (https://www.lemkesoft.de) and Adobe Photoshop (CS4).
The specimens are labelled UB W followed by digits representing the repository number according to the original system developed by Müller. The material is currently housed at the University of Lund, but forms part of the Müller collection of the Steinmann Institute, Bonn, and will eventually be stored there.
Due to the rather simple morphology of the tubular fossils and the superficial resemblance between all specimens, a morphometric approach was used to characterize and identify differences (Fig. 2). This approach is similar to those carried out in previous studies dealing with fossil cyanobacteria (e.g. Peel 1988; Stockfors & Peel 2005).
Figure 2 Open in figure viewerPowerPoint Schematic line drawings of Lower Palaeozoic filamentous cyanobacteria with location of measurements made for morphometric analysis. Abbreviations: ED, external diameter; L, length; LA, annulus length; WA, annulus width. Not to scale. CaptionSchematic line drawings of Lower Palaeozoic filamentous cyanobacteria with location of measurements made for morphometric analysis. Abbreviations: ED, external diameter; L, length; LA, annulus length; WA, annulus width. Not to scale.Preservation of the original morphology of the tubes in three dimensions permitted the use of morphometric data and their acquisition from SEM photographs. However, despite the exceptional preservation of the tubes, the state of preservation varies from specimen to specimen and even along the same specimen. Thus specimens with obvious traces of secondary encrustation were excluded, and only the best‐preserved individuals were included in the morphometric analysis. The approximate length (L) of the tube and its external diameter (ED) were measured. When the tubes exhibited a distinct annulation, also annulus length (LA) and width (WA) were recorded. Cell parameters were measured on the inner filament surface; however, in some cases they were measured on the outer surface when the inner filament surface was not exposed. Characteristic annulus length and width values were determined for each tube, averaged on the basis of several cells if possible. If no annular details were preserved, only the external diameter was measured at multiple positions on the tube, then the mean was calculated. The internal diameter was not taken into account in this study, as the wall surrounding the internal void of the tube (if at all hollow) is relatively constant and thin (c. 2 μm). All data were compiled in a frequency histogram of the external cross diameter (n = 262), the cell length and width (n = 110). The size distribution was analysed, and morphometrically associated groups of tubes were determined graphically. Annulus length and width were plotted together in order to characterize the different size classes of tubes and the morphometric variability within each size class. The number of measured specimens (n) and the minimum and maximum values in each size class are listed. Standard statistical data were calculated as the mean (x), the standard deviation (s), and the range. Measurements were taken with image‐processing software ImageJ (https://imagej.nih.gov/) and recorded in micrometres.
Morphological variation of taxonomic significance constitutes the main basis for recognition of taxonomic units when dealing with fossils. However, taphonomic processes can profoundly affect and bias morphological variation, upon which taxonomic decision can be erroneously based. It has long been recognized from fossil assemblages of cyanobacteria and decay experiments on extant cyanobacteria that taxonomic diversity of various assemblages usually comprises a high amount of degradational variants, later called ‘form taxa’ (Hofmann 1976), or ‘taphotaxa’ (Spencer 2001), of one or several taxa (Bartley 1996 and references therein). Recently, microfossil taxa, particularly of Precambrian age, have been distinguished on informal (i.e. established for micro‐remains of various degrees of preservation) and formal (i.e. truly biological) basis (e.g. Lee Seong‐Joo & Golubic 1998; Golubic & Seong‐Joo 1999; Sergeev et al. 2012). Understanding such degradational taphonomic pathways is a key aspect in the analysis of fossil assemblages of cyanobacteria and their taxonomic diversity.
In addition, mineralization (e.g. sheath calcification) is a common feature found in many cyanobacteria. In modern forms, this process is mainly restricted to freshwater species (e.g. Riding 1977; Merz 1992; Merz‐Preiß & Riding 1999) but in the geological past, cyanobacteria were able to undergo sheath calcification in marine environments as well (Riding 2000, 2006, 2011; Arp et al. 2001; for a modern example of marine calcified cyanobacteria see Planavsky et al. 2009). Calcification as an extracellular process is intimately related to the external mucopolysaccharide sheath (but see Couradeau et al. 2012; Riding 2012) and is specific to some cyanobacteria but not obligate since it directly depends on suitable environmental conditions (e.g. Merz‐Preiß & Riding 1999; see Riding 2006 for a review of the calcification process in cyanobacteria). Nucleation of calcium carbonate crystals occurs either within the protective mucilaginous sheath (impregnation) or on or close to it (encrustation) (e.g. Riding 1977; Merz 1992). In addition to in‐vivo sheath calcification, post‐mortem degradation by heterotrophic bacterial activity may induce partial calcification of the external sheath (Schroeder 1972; Chafetz & Buczynski 1992; Pratt 2001). Thus, processes of calcification may also drastically modify the original filamentous morphology of cyanobacteria. Early diagenetic changes such as secondary phosphatization may efficiently mask the original composition of the filament.
ResultsThe material at hand consists of numerous bundles or entangled colonies of thread‐like forms, inferred to represent cyanobacteria (Figs 3-6). Isolated, individual strands are rare. Despite generally good preservation, the thread‐like forms are not complete lengthwise. The irregular margins of various areas along the threads indicate that many specimens were partially decayed before phosphatization (e.g. Figs 4E, 5A, C, 6C, E). The three‐dimensionally preserved specimens comprise hollow, open‐ended threads with traces of annulation on the surfaces (e.g. Fig. 4D–G). Where preserved, the annulated pattern corresponds to a series of more or less regular compartments (e.g. Fig. 5). These compartments are also distinguishable on the outer surface of the tube by evenly spaced, very shallow constrictions, and on the inner surfaces of the tube by regularly spaced transverse ridges (Figs 4-6). The location and regularity of this annular pattern suggest that it represents the cells mirroring the repetitive structure caused by the stacked cells of a trichome of cyanobacterial origin.
Figure 3 Open in figure viewerPowerPoint Clusters of variably packed cyanobacterial remains of the annulated tubular microfossil Oscillatoriopsis longa Timofeev & Hermann, 1979 emend. Butterfield et al., 1994 (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone). A, UB W 459, loosely arranged cylindrical tubes (upwards arrows) set in a phosphatic matrix (downwards arrows). B, UB W 460, flexuous and sinuous specimens intertwined with one another; note the close association of filaments with significantly different size diameter (arrows). C, UB W 461, cyanobacterial remains embedded into a phosphatic matrix; upward arrow points to an empty semi‐circular sheath, whereas downwards arrow indicates a solid rod‐shaped microfossil interpreted as possible remains of a trichome. D, UB W 462, cluster with short fragments of closely packed and curved tubular microfossils. E, UB W 463, broad and straight tubes running subparallel to one another. F, UB W 464, comprising several flexuous fragments of variable length. G, UB W 465, broad and slightly curved tubes of variable length. H, UB W 466, 3 mm long flexuous tubes associated with much shorter fragments (arrows). I–J, clusters with enrolled specimens; I, UB W 467; J, UB W 468. Scale bars represent: 200 μm (A–D, F, H, I); 100 μm (E, G, J). CaptionClusters of variably packed cyanobacterial remains of the annulated tubular microfossil Oscillatoriopsis longa Timofeev & Hermann, emend. Butterfield et al., (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone). A, UB W 459, loosely arranged cylindrical tubes (upwards arrows) set in a phosphatic matrix (downwards arrows). B, UB W 460, flexuous and sinuous specimens intertwined with one another; note the close association of filaments with significantly different size diameter (arrows). C, UB W 461, cyanobacterial remains embedded into a phosphatic matrix; upward arrow points to an empty semi‐circular sheath, whereas downwards arrow indicates a solid rod‐shaped microfossil interpreted as possible remains of a trichome. D, UB W 462, cluster with short fragments of closely packed and curved tubular microfossils. E, UB W 463, broad and straight tubes running subparallel to one another. F, UB W 464, comprising several flexuous fragments of variable length. G, UB W 465, broad and slightly curved tubes of variable length. H, UB W 466, 3 mm long flexuous tubes associated with much shorter fragments (arrows). I–J, clusters with enrolled specimens; I, UB W 467; J, UB W 468. Scale bars represent: 200 μm (A–D, F, H, I); 100 μm (E, G, J). Figure 4 Open in figure viewerPowerPoint Strongly collapsed, torn and deformed originally tubular, microfossils assigned to Oscillatoriopsis longa Timofeev & Hermann, 1979 emend. Butterfield et al., 1985 (arrows in A–C). Annulated (D–G) and non‐annulated (H–I) tubular microfossils assigned to O. longa and Siphonophycus kestron Schopf, 1968, respectively (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone). A, C, UB W 469: A, overview; C, close‐up. B, UB W 470, overview. D–E, UB W 471; D, overview; downwards arrows point to tubes with large diameter (size class Cl3), whereas upwards arrows indicate tubes with smaller diameter (size class Cl2); E, close‐up; tubes of two different cross‐sectional diameters co‐occur (arrows). F, UB W 472, annulated tubes with the broadest observed diameter. G, UB W 463, close‐up of UB W 463 (overview in Fig. 2E); annuli are evenly distributed along the tube with only slight irregularities (arrows). H–I, UB W 473: H, overview. I, close‐up; tubes (arrows) are narrow, short and smooth, note the density of co‐occurring specimens. Scale bars represent: 200 μm (A, D, F, H); 100 μm (B, C); 50 μm (E, G, I). CaptionStrongly collapsed, torn and deformed originally tubular, microfossils assigned to Oscillatoriopsis longa Timofeev & Hermann, emend. Butterfield et al., (arrows in A–C). Annulated (D–G) and non‐annulated (H–I) tubular microfossils assigned to O. longa and Siphonophycus kestron Schopf, , respectively (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone). A, C, UB W 469: A, overview; C, close‐up. B, UB W 470, overview. D–E, UB W 471; D, overview; downwards arrows point to tubes with large diameter (size class Cl3), whereas upwards arrows indicate tubes with smaller diameter (size class Cl2); E, close‐up; tubes of two different cross‐sectional diameters co‐occur (arrows). F, UB W 472, annulated tubes with the broadest observed diameter. G, UB W 463, close‐up of UB W 463 (overview in Fig. E); annuli are evenly distributed along the tube with only slight irregularities (arrows). H–I, UB W 473: H, overview. I, close‐up; tubes (arrows) are narrow, short and smooth, note the density of co‐occurring specimens. Scale bars represent: 200 μm (A, D, F, H); 100 μm (B, C); 50 μm (E, G, I). Figure 5 Open in figure viewerPowerPoint Filaments of Oscillatoriopsis longa Timofeev & Hermann, 1979 emend. Butterfield et al., 1994 (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone) with associated sketches. A–B, UB W 471, filaments with an array of annuli; the annulated pattern is seen both on the outer and inner surfaces of the thread (upper part of the picture) or only on the inner surface (bottom of the picture); note the deformation of the thread (arrows) indicating that the thread was probably soft and unmineralized prior to phosphatization. C–D, UB W 471, tubes of two different cross‐sectional diameters co‐occur; the annulated pattern is seen both on the outer and inner surfaces of the thread (arrows). E–F, UB W 477, a series of well‐defined semi‐circular ridges (cross‐walls, arrow) divides the thread into small compartments (feature that cannot be distinguished on the outer surface); possibly suggesting that the cellular wall of the trichome was partially preserved. G–H, UB W 478, the annulated pattern is barely visible on the outer surface of the thread (transverse arrows), and grades longitudinally and laterally into a smooth surface; note the shallow furrow at the top of the thread (downwards arrow). I–J, UB W 464, threads with annuli of their outer surface; the annulated pattern is slightly irregular; note the variation in distance between consecutive cross‐walls and in constriction at cross‐wall (arrows on right hand specimen; compare with arrows on the specimen on lower left of the picture). All scale bars represent 50 μm. CaptionFilaments of Oscillatoriopsis longa Timofeev & Hermann, emend. Butterfield et al., (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone) with associated sketches. A–B, UB W 471, filaments with an array of annuli; the annulated pattern is seen both on the outer and inner surfaces of the thread (upper part of the picture) or only on the inner surface (bottom of the picture); note the deformation of the thread (arrows) indicating that the thread was probably soft and unmineralized prior to phosphatization. C–D, UB W 471, tubes of two different cross‐sectional diameters co‐occur; the annulated pattern is seen both on the outer and inner surfaces of the thread (arrows). E–F, UB W 477, a series of well‐defined semi‐circular ridges (cross‐walls, arrow) divides the thread into small compartments (feature that cannot be distinguished on the outer surface); possibly suggesting that the cellular wall of the trichome was partially preserved. G–H, UB W 478, the annulated pattern is barely visible on the outer surface of the thread (transverse arrows), and grades longitudinally and laterally into a smooth surface; note the shallow furrow at the top of the thread (downwards arrow). I–J, UB W 464, threads with annuli of their outer surface; the annulated pattern is slightly irregular; note the variation in distance between consecutive cross‐walls and in constriction at cross‐wall (arrows on right hand specimen; compare with arrows on the specimen on lower left of the picture). All scale bars represent 50 μm. Figure 6 Open in figure viewerPowerPoint Morphological and structural details of Oscillatoriopsis longa Timofeev & Hermann, 1979 emend. Butterfield et al., 1994 (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone). A, UB W 459, imprint of specimen onto the surface of the surrounding phosphatic matrix (left arrow); closely associated, a trichome‐like rod (upwards arrows). B, UB W 467, specimen with a very thin tube wall; the reticulate pattern inside the tube wall is probably a diagenetic artefact (arrows). C, UB W 466, compartmentalization of tube is recognizable on the internal surface by the presence of evenly spaced low‐relief ridges (arrows), but barely on the outer surface in this case (overview in Fig. 2H). D, UB W 475, division of the inner tube marked by the presence of septum‐like ridge (arrow on left); note also the inner space of the tube partially filled with phosphatic matter (arrow on right). E, UB W 477, tube with irregular margins indicating that it may have been only partially phosphatized. F, UB W 461, compartmented trichome‐like rod roughly matching the division of the empty tubes; stacked disc‐shaped structures partially joined and dislocated from one another (arrows); note that the trichome‐like rod does not seem to be embedded within an envelope‐like sheath. G, UB W 474, trichome‐like rod displaying abrupt variation in annulus length between consecutive disc‐like structures within the same tube; annuli, located on the right side (arrows), are half as long as the one located the furthest left of the rod (arrow). H, UB W 476, partially preserved tube (upwards arrow) with an inner trichome‐like core; the inner rod consists of a stacked disc‐shaped structures arranged in a serial sequence (downwards arrows). Scale bars represent: 20 μm (A, B, G); 50 μm (C–F, H). CaptionMorphological and structural details of Oscillatoriopsis longa Timofeev & Hermann, emend. Butterfield et al., (Gum quarry, Kinnekulle, Västergötland, Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone). A, UB W 459, imprint of specimen onto the surface of the surrounding phosphatic matrix (left arrow); closely associated, a trichome‐like rod (upwards arrows). B, UB W 467, specimen with a very thin tube wall; the reticulate pattern inside the tube wall is probably a diagenetic artefact (arrows). C, UB W 466, compartmentalization of tube is recognizable on the internal surface by the presence of evenly spaced low‐relief ridges (arrows), but barely on the outer surface in this case (overview in Fig. H). D, UB W 475, division of the inner tube marked by the presence of septum‐like ridge (arrow on left); note also the inner space of the tube partially filled with phosphatic matter (arrow on right). E, UB W 477, tube with irregular margins indicating that it may have been only partially phosphatized. F, UB W 461, compartmented trichome‐like rod roughly matching the division of the empty tubes; stacked disc‐shaped structures partially joined and dislocated from one another (arrows); note that the trichome‐like rod does not seem to be embedded within an envelope‐like sheath. G, UB W 474, trichome‐like rod displaying abrupt variation in annulus length between consecutive disc‐like structures within the same tube; annuli, located on the right side (arrows), are half as long as the one located the furthest left of the rod (arrow). H, UB W 476, partially preserved tube (upwards arrow) with an inner trichome‐like core; the inner rod consists of a stacked disc‐shaped structures arranged in a serial sequence (downwards arrows). Scale bars represent: 20 μm (A, B, G); 50 μm (C–F, H).The nature of the thread walls remains uncertain. Previous studies on secondarily phosphatized cyanobacteria have noted that the preservation of annuli on the surfaces of thread‐like forms could provide reliable information with respect to the nature and structure of the wall. For example, Peel (1988) described helically coiled cyanobacteria from the early Cambrian Portfjeld Formation of Greenland. Based on the surface pattern, he interpreted annulated threads of Jiangispirellus groenlandicus Peel, 1988 as naked cellular trichomes (i.e. without a sheath) whereas specimens of Spirellus shankari Singh & Shukla, 1981 with a smooth outer surface, were considered to be filaments. However, it should be noted that it is not clear whether a sheath was originally present (Peel 1988; but see Yi & Bengtson 1989).
The annulated pattern of the examined specimens varies from thread to thread but annulation is often best preserved on the inner surface of the wall. Some specimens show cellular structures on the inner surface of the wall whereas the outer surface lacks annulation (e.g. Figs 5A–B, E–F; 6C). The presence of well‐defined ridges (e.g. Figs 5E–F, 6D–E) suggests that the cellular wall of the trichome was at least partly intact at the time of phosphatization and that phosphate was deposited on the cell wall. In this case, the annulation may have resulted from replication of the cellular wall but the annulated pattern gives no reliable information on the presence or absence of a sheath surrounding the cellular trichome. Less frequently, the annulation may also be visible on either surface of the thread wall (e.g. Figs 4E, 5A–D). This may indicate that either annulated threads are the remains of naked cellular trichomes, or that a thin sheath, which was tightly wrapped around the cellular trichome, was originally present. In the latter case, the external wall compartmentation or annulation may be simply the result of an imprint left by a trichome onto the sheath during the fossilization process (i.e. pseudoseptate filament). Slight changes in the distance between following constrictions on the outer surface of the threads or irregularities in constriction depth (Fig. 5I, J) may favour this last hypothesis. In addition, where imprinted on the outer surface, the cellular pattern may fade longitudinally (and even laterally) into a smooth surface (Fig. 5G, H). Accordingly, the degree of visibility of annulation might be a matter of preservational variation. By comparison, the organic‐walled microfossils Tortunema Hermann in Timofeev et al., 1976 and Calyptothrix Schopf, 1968 (originally described as markedly annulated trichomes with prominent ring‐like ridges and sharply truncated apices, but see Sergeev et al. 2012) were described as pseudoseptate filaments, that is empty, cyanobacterial sheaths with ring‐like casts of trichome cells on their surface, and often show similar kinds of deformation (e.g. Butterfield et al. 1994, pp. 67–69). However, it cannot be excluded that observed irregularities might arise from post‐mortem degradation of the cellular strand, causing the cells to break off from the main strand, irregularly collapse and partly lose their structural integrity.
The filamentous specimens are also preserved as empty, smooth threads that do not display traces of annulation (Fig. 4A–C, H–I). Well‐preserved specimens are similar in shape to the annulated threads, whereas other threads are slightly shrunken, sheared and collapsed, having partially lost their structural integrity. These specimens probably represent degradational variants of well‐preserved specimens of similar diameter size.
Specimens are also occasionally preserved as solid rods. However, it is not clear whether these result from an early impregnation and replacement of the cellular material or infilling of the outer envelope followed by its complete degradation. Where preserved, these rods usually occur within a tube, yet sometimes without a tube (compare Fig. 6A, F with Fig. 6D, H). The rods may also be compartmentalized and roughly match the cellular pattern present in the empty tubular annulated threads. The units are arranged in a serial sequence of few cells, and they may be joined or slightly dislocated from one another.
The tubes vary from perfectly three‐dimensional to strongly collapsed (compare Fig. 4C with 4E). Threads can be variably distorted, being slightly bent (Fig. 3A), twisted along their main axis (Fig. 5A‐B), significantly shrunken, and also sheared (Fig. 4A–C). In addition, the external surface is occasionally covered with small wrinkles and shallow depressions (Fig. 5A–B, G–H), which indicate a rather soft and pliable surface prior to phosphatization. Many filamentous fossil cyanobacteria described possess a mineralized sheath, such as species of Girvanella Nicholson & Etheridge, 1878 (e.g. Nicholson & Etheridge 1878; Mamet & Roux 1975; Riding 1977; Danielli 1981), or were not obligate calcifiers such as Obruchevella (Reitlinger 1948; Riding & Vorona 1984; Peel 1988) and Spirellus Jiang in Luo et al., 1982. Girvanella bears broad similarities to our specimens. Preserved in a wide range of sedimentary deposits, ranging from the Neoproterozoic (e.g. Knoll et al. 1993) to the Miocene (Pratt 2001), Girvanella has been described as tubular sheaths of filamentous organisms with open opposite endings, variably packed and rarely branching, which were encrusted and/or impregnated by calcareous minerals during lifetime (Riding 1977) or upon death (Pratt 2001; but see also Arp et al. 2002; Pratt 2002). However, considering the more or less constant thickness of the wall, the degree of change in the three‐dimensional appearance of the filaments, as well as the apparent flexibility and softness of the external surface, the filamentous cyanobacteria described herein were probably not mineralized during life. Peel (1988, pp. 24–25, figs 11A, C; 12) illustrated and described a similar type of deformation on the surface of several three‐dimensional preserved, secondary phosphatized, filamentous specimens assigned to S. shankari. He interpreted those specimens as a degradational variant of S. shankari with an unmineralized sheath.
Collectively the material at hand is interpreted as filamentous cyanobacteria preserved as three main degradational variants: (1) empty tubular sheaths (Fig. 4A–C, H, I); (2) empty multicellular trichomes (e.g. Figs 4E, 5A–D); and possibly (3) pseudoseptate filaments, that is, tubular structures with an imprint of the trichome on the sheath (e.g. Figs 4G; 5G‐J). However, in rare cases, tubes may represent the partial remains of cellular material (Fig. 6D, H), and moulds of multicellular trichomes (Fig. 6A, F).
Systematic palaeontologyThe species concept in Cyanobacteria has remained notoriously problematic. A cyanobacterial species is traditionally regarded as ‘a group of population (strains), which belong to one and the same genotype (genus), characterized by stabilized phenotypic features (definable and recognizable, with distinct limit of variation) and having the same ecological criteria; they occur repetitively (in time) in a variety of ecologically similar localities’ (Komárek 2010). Their morphology is quite simple and variability could depend on environmental factors. Thus morphological criteria alone are hardly applicable for species delimitation and must be in agreement with molecular, biochemical, ecological and cytological markers (Johansen & Casamatta 2005; for a recent review, see Komárek 2010). In addition, the possibility of horizontal exchange of genetic material between cyanobacterial individuals and populations (e.g. Rudi et al. 1998), together with the ability of rapid adaptation to changes in environmental factors (e.g. Kohl & Nicklisch 1981), complicate the identification of unique criteria for species delimitation. The species concept for extant Cyanobacteria, is however not applicable in palaeontology; ecological (where specimens are preserved in situ) and morphological criteria are the only available tools for species discrimination. Yet, fossil cyanobacteria have modern counterparts that provide an opportunity for broader comparison (e.g. Knoll & Golubic 1979).
Unbranched and uniseriate filamentous microfossils assigned to Cyanobacteria are very common. To accommodate the wide morphological range of these thread‐like forms, numerous taxa have been erected on the basis of morphological characteristics but also on preservational criteria, that is, the preservation of the cellular trichome and/or sheath (e.g. Butterfield et al. 1994; Sergeev et al. 2012). For example, multicellular trichomes are usually identified as representatives of Cyanonema, Megathrix, Cephalophytarion, Filiconstrictosus, Partitiofilum, Oscillatoriopsis and Veteronostrocale Schopf & Blacic, 1971 according to trichome size, type of constriction at cross‐wall and shape and size of medial and terminal cells, whereas trichomes associated with a sheath should be best referred to as Palaeolyngbya, in most cases. By contrast, Siphonophycus, Tortunema and Calyptothrix have been interpreted to include species represented by degradational variants such as empty sheaths and pseudoseptate filaments (for a comprehensive review, see Sergeev et al. 2012).
Based on their close morphological resemblance to Oscillatoriopsis Schopf, 1968 emend. Butterfield et al., 1994 and Siphonophycus Schopf, 1968 emend. Knoll et al., 1991, the Gum quarry specimens are assigned to these two morphotaxa.
Genus OSCILLATORIOPSIS Schopf, 1968 emend. Butterfield et al., 1994Oscillatoriopsis obtusa Schopf, 1968 emend. Butterfield et al., 1994.
See Butterfield et al. (1994).
Butterfield et al. (1994) assigned all unbranched, unsheathed, uniseriate cellular trichomes according to their cell length/diameter ratio to two different taxa. Cyanonema Schopf, 1968 emend. Butterfield et al., 1994, includes all trichomes with a cell length/diameter ratio greater than one, whereas trichomes of Oscillatoriopsis have a length/diameter ratio less than one. In both these taxa, trichomes consist of similar medial cell types with uniform cell size in individual strands, and are exclusively made of vegetative cells without specialized cells known as heterocysts and akinetes (Schopf 1968).
Oscillatoriopsis was originally described by Schopf (1968) to include naked cellular trichomes of oscillatoriacean cyanobacteria similar to the modern taxon Oscillatoria Vaucher ex Gomont, 1892. Oscillatoriopsis was subsequently emended twice: (1) by Mendelson & Schopf (1982) to include all Oscillatoria‐like and Lyngbya‐like trichomes with an indistinct sheath less than 1–2 μm; and (2) by Butterfield et al. (1994) to include all unbranched, uniseriate cellular trichomes that have a length/diameter ratio less than one. Butterfield et al. (1994) united more than 75 species of Oscillatoriopsis and other genera known into four species based on diameter or trichome cells only.
However, Sergeev et al. (2012) criticized the Cyanonema–Oscillatoriopsis split and noted that the way Oscillatoriopsis is currently characterized is oversimplified and includes several morphologically distinguishable species. Overlapping size ranges are relatively common but distinct medial and terminal cells ensure their status as independent species. Consequently, Sergeev et al. (2012, pp. 296–299) did not recognize either the synonymy of Oscillatoriopsis or the separation into four species by Butterfield et al. (1994).
It should also be noted that Oscillatoriopsis differs from Megathrix Yin, 1987 emend. Dong et al., 2009 by its flat and complete cross‐walls. Specimens assigned to Megathrix are typically less than 100 μm in diameter and characterized by a tube subdivided by regularly spaced transverse cross‐walls that are either complete or incomplete and relatively corrugated.
Oscillatoriopsis longa Timofeev & Hermann, 1979 emend. Butterfield et al., 1994 Figures 3, 4A–G, 5, 6See Butterfield et al. (1994) and Zhang et al. (1998).
Gum quarry at Kinnekulle in Västergötland, southern Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone (Fig. 1).
71 clusters of filamentous cyanobacteria. In total, the material comprises nearly 350 individual filaments (Table 1).
The material consists of microscopic thread‐like forms, considered to be tubular encrustations and/or impregnations. These elongated threads are inferred to represent secondarily phosphatized naked cellular trichomes. The threads are exclusively unbranched. Based on preservation, it is not clear if the trichome is enclosed within a thin tightly packed sheath. However, the filamentous structures were non‐mineralized during life (see above).
The thread‐like structures can be enclosed by a phosphatic matrix (Fig. 3A–C, E, G). Considerable variation exists in the mode of aggregation of the strands. Threads are randomly arranged but more or less oriented along the same axis. The threads can be loosely or tightly packed (compare Fig. 3A–D), irregularly coiled (Figs 3B, H–J; 4D), and may occasionally run subparallel to one another (Figs 3E, 4F). Cylindrical cellular trichomes are straight to highly flexuous and sinuous (Figs 3, 4), some even being enrolled (Fig. 3I, J). Generally, specimens with the widest diameter (Figs 3E, 4F) are less twisted and contorted than medium‐sized tubes (Figs 3A–B, H–J; 4D). The number of individuals per cluster is highly variable and ranges from less than 5 per cluster up to 20 individuals or more. However, specimens were never found as solitary filaments. Post‐mortem processes may have played a role in the position, distribution and density of the threads in each cluster. Although each individual is easily recognizable in a cluster, the walls of adjacent tubes may occasionally coalesce.
The tubes are fairly circular and equal in cross‐diameter along their entire lengths (Figs 3, 4). Whereas slight irregularities may occur, tapering toward the ends of the structures was not observed. In addition, every single thread has open opposite endings with a circular edge (open terminations). A few filamentous structures are strongly collapsed and sheared. These distorted smooth tubes have lost their structural integrity and their original topology; their cross‐diameter is highly irregular along the entire length (Fig. 4A–C).
Trichome length is highly variable and ranges from less than 100 μm to slightly less than 3000 μm. Most of the tubes are between 500 and 1000 μm long (Figs 3, 4). The external diameter of the tubes ranges from 14 to 37 μm, but more than 75% of the tubes fall within the size range 24–33 μm (Fig. 7A; Table 2). The original wall thickness appears to be constant along the entire tube length and measures c. 2 μm (Fig. 5I, J). Owing to secondary mineral encrustation and impregnation of the tube wall, in a very few specimens wall thicknesses can reach c. 5 μm.
Figure 7 Open in figure viewerPowerPoint Morphometric analyses of a sample population of Siphonophycus kestron Schopf, 1968 and Oscillatoriopsis longa Timofeev & Hermann, 1979 emend. Butterfield et al., 1994. A, size frequency distribution (number of filaments) of the external cross‐sectional diameter indicating three size classes (Cl1, Cl2, Cl3); mode values are determined graphically and indicated by arrows; size class Cl1 comprises individuals assigned to S. kestron, whereas the size classes Cl2 and Cl3 include filaments assigned to O. longa. B, cross‐plot and frequency distribution of the annulus length against width of a sample population of O. longa (i.e. comprises individuals of Cl2 and Cl3); the two size classes are highlighted with different shading (x‐axis: annulus length; y‐axis: annulus width). CaptionMorphometric analyses of a sample population of Siphonophycus kestron Schopf, and Oscillatoriopsis longa Timofeev & Hermann, emend. Butterfield et al., . A, size frequency distribution (number of filaments) of the external cross‐sectional diameter indicating three size classes (Cl1, Cl2, Cl3); mode values are determined graphically and indicated by arrows; size class Cl1 comprises individuals assigned to S. kestron, whereas the size classes Cl2 and Cl3 include filaments assigned to O. longa. B, cross‐plot and frequency distribution of the annulus length against width of a sample population of O. longa (i.e. comprises individuals of Cl2 and Cl3); the two size classes are highlighted with different shading (x‐axis: annulus length; y‐axis: annulus width). Table 2. Standard descriptive statistics of all three size classes regarding tube diameter/annulus length (ED/LA) and cross‐wall spacing (= annulus width, WA) Size class Tube diameter (ED/LA) Cross‐wall spacing (WA) s Vmax Vmin Range n Frequency (%) s Vmax Vmin Range n Frequency (%) Cl1 8.63 0.59 9.80 7.60 2.20 24 9.16 – – – – – – – Cl2 16.30 3.38 19.50 14.60 4.90 17 6.49 5.21 0.55 5.70 4.40 1.30 3 2.73 Cl3 28.93 1.33 37.10 20.00 17.10 221 84.35 11.29 0.72 15.80 7.50 8.30 107 97.27 Total 262 100 110 100The external surface of the threads is either annulated (Figs 4G; 5C, D, I, J) or smooth (Figs 5E–H; 6B, C, H). The annulation pattern divides the structure into a sequence of disc‐shaped compartments, evenly distributed along the main axis of the tube, of more or less equal dimension. Margins between each cavity are marked by faint constrictions and transverse ridges on the external and internal surface of the tube wall, respectively (e.g. Figs 4E, G; 5A, B, E, F; 6C–E).
Two types of ridges can be recognized on the internal surface of the tube wall. The first type is a very low‐relief ridge, slightly triangular in cross‐section, with a rather soft and smooth transition to the neighbouring annulus (e.g. Figs 4E, 5A–D, 6C). The second type consists of a septum‐like ridge with sharp edges and forms a right angle to the neighbouring annulus (Figs 5E–F; 6D–E). Constrictions and ridges are circular and run all around the surface (= complete). The single annuli are circular in cross‐section and at least twice as wide as they are long (= tabular or discoidal). Where preserved, the annulated pattern is rarely interrupted by breaks and can continuously follow the entire length of a thread (e.g. Fig. 6C); fragments with more than 200 cells in a row have been observed. In the material at hand, the void (= internal space) of the annuli ranges from c. 4–16 μm in length and c. 14–37 μm in width (see below; Table 2). Although slight irregularities in terms of size and shape may occur from annulus to annulus within the same tube, the length/width ratio of an annulus is constant along the main axis of a single tube and ranges between 0.35 and 0.50 among the investigated specimens. Yet, in a few specimens, the length between succeeding annuli may vary abruptly from one to another. This sudden morphometric change involves only a small number of annuli set in a row, and these cylindrical structures are usually approximately half as long as the others (Fig. 6G).
Within the same cluster of filamentous cyanobacteria obvious differences exist between associated tubes with regard to their diameter. Although part of the variability observed is the result of combined factors such as organic decay, mineral replication and diagenetic processes, these mechanisms cannot fully explain the size differences, particularly regarding the external diameter. Some tubes are much shorter and wider than others, yet they do not show any significant traces of post‐mortem modification (Figs 4D–E, 5C–D).
The size frequency distribution analysis for a sample population of cellular trichomes of O. longa reveals that the material consists of different size classes on the basis of their cross‐sectional diameter and annulus length and width (Fig. 7). Size frequency distribution of the external cross‐sectional diameter suggests the presence of two size classes with a statistical mode at 15–16 μm (mean = 16.30 μm, s = 3.38 μm, n = 17 measured specimens) and 29–30 μm (mean = 28.93 μm, s = 1.33 μm, n = 221 measured specimens) in cross‐sectional diameter (Fig. 7A). These size classes (Cl2, Cl3) comprise mostly annulated forms. These threads do not exhibit any specific morphological features that could clearly discriminate them from one another, except morphometric differences. Plotting annulus length against width confirms partitioning of the annulated tubes into two size classes (Fig. 7B; Table 2).
Regarding the lack of diagnostic features (e.g. shape of terminal cells, presence of a sheath) that could help to precisely assign our specimens taxonomically, we tentatively follow the scheme proposed by Butterfield et al. (1994). Several different size classes have been proposed to define species of Oscillatoriopsis according to their diameter. Butterfield et al. (1994) recognized four species: O. vermiformis (Schopf, 1968) Butterfield in Butterfield et al., 1994, 1–3 μm; O. obtusa Schopf, 1968, 3–8 μm; O. amadeus (Schopf & Blacic, 1971) Butterfield in Butterfield et al., 1994, 8–14 μm; and O. longa Timofeev & Hermann, 1979, 14–25 μm.
The annulated specimens described herein have regular, closely spaced, and internal transverse partitions suggesting preservation of cell walls in a trichome of oscillatoriacean type. In our material, there is a broad recurring modality regarding thread diameter that may suggest a single underlying biological origin (Fig. 7A). Differences in preservation suggest that individuals assigned to O. longa are preserved as different degradational variants but shared morphological features indicate a common origin. Consequently, the material is treated as a single taxonomic unit. The specimens assigned to size class Cl2 and some individuals identified as forming a part of size class Cl3 fall within the size range recognized by Butterfield et al. (1994) and, thus, are regarded as O. longa. Regarding the size class Cl3, although some specimens recorded could be wider than 25 μm (the upper size limit recognized by Butterfield et al. 1994) we interpret our specimens as belonging to the same taxonomic unit and consider that there are no morphological criteria that clearly discriminate them from one another. At present, it appears to be unnecessary to erect new species until a more comprehensive revision of Oscillatoriopsis and other related groups is available. It should also be noted that Butterfield et al. (1994) defined the lower and upper size limit of O. longa from small sample sizes (see details in Butterfield et al. 1994). Therefore, this 14–25 μm size range awaits further support from a single large population. Additional diagnostic features were given by Timofeev & Hermann (1979; see also Sergeev et al. 2012) and they described O. longa as multicellular trichomes formed by wide discoidal cells with a cell width and length of c. 25 μm and c. 5–6 μm, respectively. Although differences in cell size are obvious, we still find it most parsimonious to interpret our specimens from Kinnekulle as representatives of this morphotaxon. Most other species of Oscillatoriopsis have shorter cell length and width ranges, and, thus our material cannot be considered as the remains of one of these other species (see Butterfield et al. 1994; especially Sergeev et al. 2012, table 13, p. 298). Knoll et al. (1988) described, from the Palaeoproterozoic Duck Creek Formation of Australia, O. majuscula Knoll et al., 1988 as a multicellular trichome consisting of stacked discoidal cells with a very broad width of 63.0 μm and 6.0–11.0 μm cell length. Trichomes of O. majuscula differ significantly from the individuals described herein by having a much larger cell width.
We are aware that the taxonomic status of the specimens described herein is directly related to the degree of visibility of annulation. Some specimens interpreted as O. longa might be regarded as representatives of Tortunema and/or Calyptothrix (Tortunema is a possible junior synonym of Calyptothrix; Sergeev et al. 2012). They are usually regarded as unbranched filamentous empty single‐layered sheaths with transverse annular ribs (i.e. pseudoseptate filaments) rather than naked trichomes (Butterfield et al. 1994; Sergeev et al. 2012). According to Butterfield et al. (1994), pseudoseptate filaments can often be identified on the basis of degradation patterns (see Butterfield et al. 1994, pp. 67–69, figs 24H, 27A–C; see also Hermann 1986, figs 7–8). Among the specimens assigned to O. longa, few individuals show a similar degradation pattern, that is an annulated pattern that grades longitudinally into Siphonophycus type sheaths (Fig. 5G, H) or degradational collapse of the threads rather than a disaggregation into separate cells (Figs 4A–C, 5A–B). Lastly, a few individuals could also be interpreted as representatives of Palaeolyngbya if sheaths and associated trichomes are equally preserved (see Butterfield et al. 1994; Sergeev et al. 2012). Palaeolyngbya differ from Oscillatoriopsis mainly by being surrounded by prominent uni‐ or multilayered smooth sheath(s), usually larger than the trichome width (see discussion on Palaeolyngbya in Schopf 1968; Butterfield et al. 1994; Sergeev et al. 2012). Possible candidates in our material could be the filamentous forms with traces of annulation on the inner surface but having a smooth outer surface (Figs 5E–F, 6C).
Genus SIPHONOPHYCUS Schopf, 1968 emend. Knoll et al., 1991Siphonophycus kestron Schopf, 1968.
See Knoll et al. (1991), Butterfield et al. (1994) and Moczydłowska (2008).
Described from the late Precambrian bedded carbonaceous cherts of the Bitter Spring Formation in the Ross River area of central Australia, the morphotaxon Siphonophycus Schopf, 1968 emend. Knoll et al. (1991) originally encompassed fragments of cylindrical, non‐septate, unbranched, straight to slightly bent thread‐like forms, of variable length and diameter, with no tapering toward filament termini. Schopf (1968) identified them as possibly mucilaginous, empty sheaths of Lyngbia‐like or Oscillatoria‐like cyanobacteria.
Although few morphotaxa of Siphonophycus may rather be considered as remains of eukaryotic microorganisms (Moczydłowska 2008; Sergeev et al. 2012), most Siphonophycus are definitely of cyanobacterial origin, and may form microbial mats (Knoll & Golubic 1979; Mendelson & Schopf 1982; Knoll et al. 1991; Sergeev 1992a, b, 2006). For example, some species like S. robustum from the Bitter Springs Formation have very close modern counterparts among Lyngbya or Phormidium mat‐forming cyanobacteria inhabiting the intertidal regions of the Persian Gulf (Knoll & Golubic 1979). Following several taxonomic revisions (e.g. Knoll et al. 1991; Butterfield et al. 1994) Siphonophycus now includes all filamentous microfossils with an unbranched, smooth‐walled and non‐septate cylindrical morphology with no tapering, occasionally solitary, but mostly gregarious in tangled masses. On the basis of the cross‐sectional diameter of the sheaths, several different size classes have been proposed to characterize the species of Siphonophycus (Butterfield et al. 1994). It should be noted that these classes have been arbitrarily chosen, are often overlapping in their ranges depending on the number of specimens measured and thus may not necessary reflect natural populations (see Moczydłowska 2008; Sergeev et al. 2012).
Siphonophycus kestron Schopf, 1968 Figure 4H, ISee Butterfield et al. (1994) and Moczydłowska (2008).
Gum quarry at Kinnekulle in Västergötland, southern Sweden; Cambrian Series 3, Guzhangian Stage, Agnostus pisiformis Biozone (Fig. 1).
One cluster comprising more than 50 individual sheaths.
The cluster consists of narrow 7.60–9.80 μm (mean = 8.63, s = 0.59, n = 24 measured filaments; Table 2) and short (c. 150–200 μm), prostrate and straight, exclusively smooth and unbranched cylindrical thread‐like forms, occurring in high density. Lack of annulation on outer and inner tube surfaces suggests that they represent empty open‐ended sheaths. Slight irregularities and distortions on the outer surface (e.g. shallow depressions) indicate that the sheath was originally non‐mineralized during life. No morphological features regarding the trichome are preserved.
These tubular sheaths differ significantly from O. longa with respect to their morphology and mode of aggregation (Fig. 4H, I). In addition, they cannot be considered as a degradational variant of the aforementioned species due to their size (Fig. 7A, Cl1). Butterfield et al. (1994) suggested that species‐level classification of Siphonophycus should be based on filament diameter, that is S. thulenema Butterfield in Butterfield et al., 1994, 1 μm; S. septatum (Schopf, 1968) Knoll et al., 1991, 1–2 μm; S. robustum (Schopf, 1968) Knoll et al., 1991, 2–4 μm; S. typicum (Hermann, 1974) Butterfield in Butterfield et al., 1994, 4–8 μm; S. kestron Schopf, 1968, 8–16 μm; S. solidum (Golub, 1979) Butterfield in Butterfield et al., 1994, 16–32 μm, and S. punctatum Maithy, 1975, emend. Buick & Knoll, 1999, 32–64 μm. Our material best fits the characterization of S. kestron. It should also be noted that few specimens interpreted as O. longa of the size class Cl3 (Fig. 7A) have a smooth outer surface with no obvious traces of annulation or internal cross‐walls (Fig. 4A–C). They may be regarded as Siphonophycus and could be identified as S. solidum (strand diameter 16–32 μm). However, they are considered herein to be degradational variants of O. longa since they co‐occur and fall within the same diameter size range.
DiscussionThe fossils described exhibit rather simple filamentous morphology, and their identification can be problematic. SEM examination revealed that the filaments comprise an empty core (lumen), which can be surrounded by either a well‐defined smooth or annulated cylindrical envelope (sheath/coat). Similar filamentous networks can be attributed to different organisms: prokaryotes (i.e. filamentous bacteria and cyanobacteria) and eukaryotes (i.e. fungi).
The filamentous forms broadly resemble hyphae (chains of multinucleate, tubular, filament‐like cells, 1–10 μm in diameter and up to several centimetres in length) of multicellular fungi and the various bundles could be the remains of fungal mycelia (a tangled mass of branching hyphae). However, morphological features typically associated with fungi, such as the branching pattern of hyphae and hyphal anastomosis (e.g. Harris 2008) or even spores (microscopic propagules), are lacking in the specimens at hand, making a fungal origin unlikely.
The specimens strikingly resemble those of the modern cyanobacterial group Oscillatoria. Oscillatoria has a worldwide distribution in freshwater, brackish and marine waters. Trichomes are uniseriate, straight to loosely sinuous, and flexible (cf. Figs 3; 4A, B, D, F). Trichomes are solitary, but if clustered or in fabric‐like mats they are not surrounded by a morphologically distinct common sheath (cf. Figs 3, 4). Trichome diameters range from about 1 μm to occasionally 100 μm. Trichomes are divided into a sequence of cells of similar size (cf. Figs 3E, 4E–G, 6C–H). Cells may be much shorter than broad to a few times longer than broad; in broader trichomes (> 15 μm in diameter) cells are shorter than long, as observed in our specimens (e.g. Figs 4E, 6C). Transverse septa are present (e.g. Fig. 4E) and constrictions may occur at cross‐walls (cf. Fig. 4E–G). All Oscillatoria produce exclusively vegetative (non‐specialized) cells. Filamentous cyanobacteria divide exclusively by binary fission in one plane (cf. Fig. 6G). Collectively, all structural features indicate that the specimens described herein are cyanobacteria and closely resemble such extant forms as Oscillatoria. Finally, the size range of filament diameters and cellular structures of the study specimens do not fall within the common size range of filaments of bacterial origin.
The filamentous Gum quarry cyanobacteria are dominated by naked multicellular trichomes and, in rarer cases, empty sheaths. However, some individuals assigned to O. longa could also be the remains of cyanobacterial sheaths with ring‐like casts of trichome cells on their surface, rather than naked trichomes. Based on the annulated pattern and thread morphology, several conclusions about multicellular trichomes preserved in the ‘Orsten’ assemblages can be drawn.
Inferred cell structures are well known from Proterozoic and younger fossil cyanobacteria observed in thin sections of cherts and as phosphate replica. Similar structures in filamentous forms such as species of Oscillatoriopsis, Palaeolyngbya Schopf, 1968, Cyanonema, Megathrix, Cephalophytarion Schopf, 1968, Filiconstrictosus Schopf & Blacic, 1971, Partitiofilum Schopf & Blacic, 1971 emend. Sergeev et al., 1995, Obruchevella Reitlinger, 1948 (e.g. Song 1984) and Obruchevella‐like forms (e.g. Peel 1988; see also Sergeev et al. 2012, pp. 293–296) show closely similar thread‐like forms compartmented into arrays of regular stacked discoidal structures. In extant cyanobacteria, the cell shape in trichomes ranges from elongated cylindrical to discoid. The cell shape is often linked to filament diameter size; forms with a diameter greater than 10 μm possess discoid cells, whereas cells in forms with a smaller diameter are usually isodiametric or elongated cylindrical (Golubic & Focke 1978; Demoulin & Janssen 1981). Consequently, tube division indicates that the trichomes described herein were made up of a series of fairly regular cells with a discoidal shape (at least twice as wide as long). Compared to modern filamentous cyanobacteria, the trichomes comprise exclusively vegetative (medial) cells and were devoid of all specialized cell types known as heterocysts and akinetes (e.g. Adams & Duggan 1999).
Differences in trichome and cell morphologies among various individuals assigned to the same species are commonly observed in Precambrian organic‐walled cyanobacterial fossils (Bartley 1996, particularly p. 573, table 1; see also Awramik et al. 1972; Knoll & Barghoorn 1975; Golubic & Hofmann 1976; Golubic & Barghoorn 1977; Knoll & Golubic 1979; among others), thus the slight variation in constriction observed between consecutive cells in our material is most likely to be the result of inorganic processes. Based on the annulated surface pattern there is little constriction between each cell. This indicates that the outer membrane layer (located outside the cytoplasmic membrane, which is not capable of being preserved; see e.g. Awramik et al. 1972; Knoll & Barghoorn 1975; Westall 1997) participates little in cross‐wall formation. Support for this interpretation comes from modern filamentous cyanobacteria. In Oscillatoria, while the inner peptidoglycan layer runs parallel to the cytoplasmic membrane, the outer membrane runs continuously into the periphery of the filament. The outer membrane can be slightly indented, but it usually does not take part in cross‐wall formation between neighbouring cells (e.g. Ris & Singh 1961; Lamont 1969; Hoiczyk & Baumeister 1995; Flores et al. 2006). By contrast, in species of Pseudanabaena Lauterborn, 1915 cells are separated by a deep constriction, implying that the outer membrane layer invaginates to a greater extent inside the cross‐wall. In extreme cases (e.g. species of Fischerella, stigonematalean cyanobacteria; Thurston & Ingram 1971), the outer membrane and the external sheath also enter and extend into the cross‐wall.
Abrupt variation in cell length between consecutive cells was observed only in a few examined specimens and involved only a small number of cells within the same filament. In these cases, aberrant cells are half as long as the others. Several types of cell division have been described for extant cyanobacteria (e.g. Stanier & Cohen‐Bazire 1977). In unbranched non‐heterocystous filamentous cyanobacteria, cell division occurs exclusively by intercalary cell division only in a single plane at right angles to the long axis and involves cross‐wall (septum) formation. Upon cell division, two daughter cells are inserted into the filament at the parent cell location via invagination of the cell walls. Newly divided cells are all approximately the same size and half as long as the parent cell. In our view, the cellular size range observed in some filaments match this condition and may thus represent a case of binary fission (Fig. 6G).
It is difficult to unequivocally establish the irregularly coiled nature of O. longa as reflecting a natural growth habit, rather than being the result of physical processes operating during deposition and preservation of the organisms, but it seems quite unlikely that this organization could be anything other than biological in nature. This interpretation is consistent with the rapid mineralization of the filament (secondary phosphatization) and is furthermore supported by the lack of deformed specimens. Overall, straight threads are rare in the material, and most of the filaments are sinuous, showing smooth curves with a variable degree of curvature (Figs 3; 4D, F).
Filaments are usually surrounded or enclosed by a phosphatic matrix. The imprint of tubular sheaths onto the surface of the matrix indicates that the matter was soft prior to phosphatization (Figs 3A, C, E, G; 4D, F; 5; 6A, E, F). Microbial organisms, including cyanobacteria, produce and excrete high‐molecular‐weight extracellular polymeric secretions (EPS) and sometimes mucus into their close surroundings (e.g. Stal 2000, particularly pp. 88–90). The presence of a phosphatic matrix enclosing the cyanobacterial threads, together with occasional imprints of sheaths onto its surface (Fig. 6A) suggests that the matrix represents the remains of such extracellular polymeric substance. This substance, possibly microbial in origin, became partly secondarily phosphatized at the same time as the enclosed filamentous cyanobacteria.
On the basis of cross‐sectional diameter and size of the cellular pattern, the material assigned to O. longa can be further sub‐divided into two distinct size classes: Cl2 and Cl3 (Fig. 7A). Individuals of distinct size classes usually co‐occur in the same cluster (e.g. Fig. 4D, E). Except for cross‐section and cell division dimensions, no morphological features separate these size classes from each other (Fig. 7). Individuals with a cross‐diameter ranging from 15 to 20 μm (Cl2) are clearly separated from other individuals by having a trichome made of smaller cells (Table 2). However, there is a more significant size variation observed among individuals within size class Cl3. This variation may be best explained as a preservational artefact, probably caused by post‐mortem transformation. The annulated pattern on the outer surface is less regular than that located on the inner surface. Consequently, observed variation in cell size may have been artificially increased.
In addition, experimental studies on cyanobacterial decay, alongside comparison with individuals from the fossil record, have demonstrated that degradation of cyanobacterial filaments may significantly alter the original trichome and sheath before fossilization (Bartley 1996, and references therein). Yet, transport, environmental stress, crystallization, dissolution, as well as mineral replication, may also increase the apparent variability in specimens belonging to the same species. It should also be noted that size limits between size classes Cl2 and Cl3 have been chosen arbitrarily, based on the abundance of the specimens measured, and thus these classes do not necessary reflect natural populations. Sample size of class Cl2 (n = 24 in Fig. 7A; n = 3 in Fig. 7B) suffers from inadequate size in comparison to size class Cl3 (n = 221 in Fig. 7A; n = 107 in Fig. 7B).
Nonetheless, there is no clear overlap regarding cell length and width between trichomes of these two size classes, and thus these may well be considered as distinct biological units. This would need to be statistically tested, but is hampered by the relative rarity of well‐preserved individuals. Owing to the lack of morphological features discriminating between individuals of the different size classes of O. longa, it is not apparent whether each size class constitutes a distinct morphotaxon. The overall good preservation of individual filaments, as well as the absence of signs indicating drastic morphological changes from one size type to another during diagenesis, indicates that taphonomy was not a major cause of size differences between each class. In addition, the lack of an apparent mineralized sheath, together with a more or less constant sheath thickness, excludes the possibility that increase in filament diameter size results from the degree of mineralization of filament sheath, as exemplified by Peel (1988). Increase in filament diameter size by secondary overgrowth of calcium phosphate may have played a role in size variation in rare cases but it did not influence the distribution pattern observed.
The material consists exclusively of dislocated and fragmented filaments of variable length. Such fragmentation may be the result of post‐mortem transport, decay and mineralization. The Alum Shale depositional environment is generally thought to have been anoxic to dysoxic and fairly deep and stagnant (e.g. Bergström & Gee 1985; Dworatzek 1987; Buchardt et al. 1997). By contrast, Eklöf et al. (1999) and Terfelt (2003) concluded from investigations of the orientation of A. pisiformis shields from the Kakeled locality (Kinekulle, Västergötland, Sweden; locality 13 in Fig. 1B) that the shields were occasionally affected by currents, which may have induced pre‐depositional transport. In addition, they suggested that the bottom topography of the Alum Shale Sea was not flat but probably/possibly scattered with intra‐basinal highs (i.e. irregularities; Eklöf et al. 1999; Terfelt 2003). Maeda et al. (2011) documented the close association of cross‐lamination structures with ‘Orsten’‐type fossils from the early‐diagenetic limestone nodules (‘Orsten’ nodule) and suggested transport of the fossils over a short distance caused by low‐density sediment‐gravity flows. Evidence for post‐mortem transport from the ‘Orsten’‐type fossil assemblages is less evident, but points in the same direction. Müller (1990) characterized the ‘Orsten’ assemblages as thanatocoenoses, that is death assemblages, including both autochthonous (in situ) components as well as allochthonous material that may have been moved by weak currents. There is nothing to suggest that a complete life‐community on or within the flocculent layer has been retrieved from acid‐resistant residues from a single limestone nodule (Maas et al. 2006). Thus, transport probably played some role in the formation of the ‘Orsten’‐type assemblages, but must have had limited influence on the observed community composition (CC, pers. obs.), as also exemplified by the presence of very fine morphological details preserved on numerous ‘Orsten’ fossils (e.g. Walossek 1993; among many others), which could not have survived longer transport. The presence of benthic cyanobacteria in ‘Orsten’ deposits supports this conclusion. It is likely that they were transported from their growth sites before final deposition, though their overall excellent state of preservation suggests that transport was limited to short distances only, under low energy conditions.
However, several experimental studies have shown that fragmentation of soft‐bodied organisms may result from the simple action of organic decay and disturbance by surrounding organisms (Allison 1986; Plotnick 1986). Thus that fragmentation and distortion essentially depend upon physical properties of the tissues and their degree of decay before authigenic mineralization sets in (Sagemann et al. 1999; Allison 2001; Briggs 2003 and references therein). Secondary phosphatization triggered the ‘Orsten’‐type preservation, and apparently resulted from a fine balance between organic decay, authigenic mineralization, abiotic and biological factors (Sagemann et al. 1999; Briggs 2003; for the ‘Orsten’‐type preservation specifically see Maas et al. 2006; Eriksson et al. 2012), all acting at the microscale level. As a consequence, the morphological details preserved vary between specimens and within one and the same individual (e.g. Maas et al. 2006; Eriksson et al. 2012, 2016). Also the material studied here shows signs of taphonomic processes variably affecting the specimens, as is evident by the different appearances of tube morphology, trichome articulation, cell morphology and surface texture.
Study of modern cyanobacteria from the springs of the Kamchatka Peninsula has revealed that living cyanobacterial communities are unable to be properly permineralized (= silicified) (e.g. Krylov et al. 1983; Krylov & Tikhomirova 1988; Westall et al. 1995). Therefore, all organism remains pass through a post‐mortem degradation phase prior to fossilization (for information on post‐mortem alteration and fossilization of cyanobacteria see Sergeev et al. 2012). Thus, even limited and gentle transport prior to phosphatization could have enhanced fragmentation of the cyanobacterial filaments.
Due to the lack of distinct morphological criteria, the nature of the filamentous cyanobacterial remains equivocal; the strands may either represent multicellular trichomes, pseudoseptate filaments, or a combination of a sheath and its associated trichome. Experimental studies on cyanobacterial decomposition (e.g. Golubic & Hofmann 1976; Francis et al. 1978a, b; Westall et al. 1995; Bartley 1996; Westall 1997), but also observations from both modern (e.g. Horodyski et al. 1977) and fossil assemblages (e.g. Hofmann 1976) show that cyanobacterial assemblages are often dominated by sheaths rather than cellular trichomes.
The overrepresentation of sheaths may be attributed to one or both of the following factors:
‘Orsten’‐type preservation is caused by secondary mineralization by calcium phosphate and early sealing of the host sediment porosity by calcium carbonate (impeding extensive diagenesis and compaction; for information on the formation of mudrock‐hosted carbonate concretions in Scandinavia see e.g. Henningsmoen 1974; Israelson et al. 1996; for a broader perspective see Raiswell 1971; Selles‐Martinez 1996; Raiswell & Fisher 2000). Early diagenetic phosphatization may have been facilitated and enhanced by high local phosphorus levels in close vicinity of dead carcases assembled by low sediment‐gravity flows rich in faecal pellets (Maeda et al. 2011). Early diagenetic mineralization is an important process which, under favourable conditions, can lead to exceptional preservation even of soft tissues (Allison & Pye 1994). Maas et al. (2006, p. 268) described the ‘Orsten’‐type preservation as the result of an ‘early three‐dimensional stabilizing preservation’ by incrustation and impregnation of even softly sclerotized surfaces by means of secondary mineralization by calcium phosphate. Yet, so far ‘Orsten’‐type preservation sensu stricto, that is the secondary phosphatization of non‐mineralized structures, is restricted to very specific taxa among cuticle‐bearing organisms such as arthropods (e.g. Müller & Walossek 1985b, 1987, 1988; Walossek 1993; Maas et al. 2003, 2007a) and cycloneuralian nemathelminths (e.g. Maas et al. 2007b, 2009; Haug et al. 2009; Zhang et al. 2015), but in both cases it is the external, non‐chitin‐bearing layer that is preserved. Other cuticle‐bearing animals, with or without chitin, such as gastrotrichs, sipunculids, annelids, kamptozoans, molluscs, tunicates or chaetognaths have yet to be discovered in this type of preservation. If not, it suggests that their different types of cuticle cannot be preserved in this setting.
The mode of preservation of the ‘Orsten’ arthropods suggests that the phosphatization process is not related to the occurrence of chitin because phosphatization is concentrated on and within the outermost cuticular layer, that is the lipoprotein‐rich (hydrophobic) but chitin‐less epicuticle (Maas et al. 2006, p. 268). This layer may be secondarily thickened by a precipitation of phosphate on the surface. Evidence for this is provided by the preservation of fine structures such as denticles, setulae and pores, all between 0.2 and 1 μm, which are exclusively epicuticular. Exceptions in the ‘Orsten’ material are the preservation of muscle tissue in pentastomids (Andres 1989) and, putatively, remains of muscular and digestive systems in Skara and phosphatocopines (e.g. Müller & Walossek 1985b, pl. 17.3; Eriksson et al. 2012, 2016). The sheaths of filamentous cyanobacteria consists largely of cross‐linked polysaccharides and may contain variable amounts of phenolic elements (Garcia‐Pichel & Castenholz 1991; Robbins 1992), whereas the cellular material of the trichome is composed of peptidoglycans, lipopolysaccharides and various proteins (Brock & Madigan 1991). The former compounds are, indeed, similar in composition to the epicuticle of arthropods. Evidence from the fossil record shows that cyanobacterial filaments may have great potential of fossilization under exceptional conditions (silification, phosphatization, carbonization; e.g. Sergeev et al. 2002 and references therein).
ConclusionsIn this study, we describe an assemblage of filamentous cyanobacteria from the Agnostus pisiformis Biozone (Guzhangian Stage, Cambrian Series 3) of the Alum Shale Formation of Sweden. A few specimens are assignable to the well‐known taxon Siphonophycus kestron but most are identified as Oscillatoriopsis longa. Individuals of the latter species can be further divided into two size classes, differentiated on cell size but not morphology. Each of these classes may represent a distinct morphotaxon, but this remains equivocal pending recovery of more material and revision of Oscillatoriopsis and related taxa. Nonetheless, the taxonomic richness of Cyanobacteria in the ‘Orsten’ assemblages might be underestimated due to lack of distinguishable diagnostic morphological features.
‘Orsten’‐type fossils represent death assemblages including both autochthonous and allochthonous organism remains (Müller 1990), and it remains uncertain if the cyanobacterial material was preserved in situ. However, the lack of microbial mats in the surrounding shales and the fact that all specimens studied are slightly fragmented suggest that the cyanobacterial filaments were probably subjected to limited post‐mortem transportation. A period of decay prior to final deposition and differential mineralization of filaments may have enhanced fragmentation. Different preservational conditions may explain the absence of bacterial mats in the surrounding shales (cf. Schovsbo 2001).
The filaments recorded may be remains of larger patches of microbial mats that covered portions of the Alum Shale seafloor (for examples of modern cyanobacterial mats, see Jøgensen & Revsbech 1983). The presence of microbial mats is also supported by the dense mass of Siphonophycus tubes retrieved from the ‘Orsten’ nodules. Background sedimentation must be low enough to enable microbial colonization, combined with an environment suitable growth and sustainability of cyanobacteria. The environmental conditions prevailing in the shallow epicontinental Alum Shale Sea, alongside limited activity of grazing and burrowing organisms (Thickpenny 1984, 1987; Buchardt et al. 1997; Schovsbo 2000, 2001) seemingly facilitated the settlement and development of benthic cyanobacterial populations.
After detailed reinvestigation of several acid‐resistant ‘Orsten’ residues, ranging from the A. pisiformis Biozone to the Trilobagnostus holmi Biozone (see Terfelt et al. 2008, 2011), the cyanobacterial material appears to be restricted to the A. pisiformis Biozone. The lack of such fossils in surrounding strata suggests that cyanobacteria had already disappeared from the Alum Shale seafloor by the upper part of the A. pisiformis Biozone. However, the lack of cyanobacteria outside the A. pisiformis Biozone may also be related to different preservational and/or taphonomic conditions, given that the bulk of the exceptionally preserved Swedish ‘Orsten’ material derives from the A. pisiformis Biozone. This also suggests a change in substrate and environmental conditions at the boundary between the Cambrian Series 3 and the Furongian (Buchardt et al. 1997; Schovsbo 2000; Eriksson & Terfelt 2007). This interval also bears witness of major changes in the euarthropod communities (from an agnostoid to an (olenid) polymerid‐trilobite dominated community; e.g. Terfelt et al. 2011), and the base of the Furongian coincides with the extinction event at the Marjumiid–Pterocephaliid biomere boundary in Laurentia (Palmer 1965, 1984) and the onset of the globally recognized Steptoean Positive Carbon Isotope Excursion (SPICE) in Scandinavia (Ahlberg et al. 2009 and references therein).
AcknowledgementsFirst, we thank the late Klaus J. Müller, discoverer of the ‘Orsten’ fossils, for his extensive fieldwork, etching of the material, permission to continue his ‘Orsten’ research in Ulm in the workgroup of Dieter Waloszek and his continuous and generous support. Thanks are also due to Stefan Bengtson (Swedish Museum of Natural History, Stockholm, Sweden) and Per Ahlberg (Lund University, Sweden) for discussions of geological aspects of the Alum Shale Formation and the material. Special thanks to John S. Peel (Uppsala University, Sweden), and Robert E. Riding, (University of Tennessee, USA), for their extremely valuable comments on the first version of the manuscript. Georg Heumann and Martin Langer (University of Bonn, Germany) kindly provided access to additional parts of the Müller collection, part which is still housed in Bonn at the Steinmann Institute of Palaeontology (part on loan to Dieter Waloszek, to be returned eventually). We are also grateful to the Central Unit for Electron Microscopy, University of Ulm, and their team for their support and providing the SEM and associated equipment. The Deutsche Forschungsgemeinschaft DFG supported CC between 2009 and 2013 (grant no. Wa/754/18‐1). MEE was funded by the Swedish Research Council (grant no. 2015‐05084).