Monogeneric clusters of five acritarch genera (Adara, Cymatiosphaera, Eliasum, Synsphaeridium and Timofeevia) were discovered in palynological residues obtained from fine-grained siliciclastic samples from the middle Cambrian (Miaolingian) Jince Formation of the Příbram–Jince Basin (Czech Republic). The clusters consist of two to more than 100 individual specimens and lack a common pattern of organisation. Acritarchs within clusters fall within a similar size range, regardless of generic affinity. Proposed mechanisms for the formation of these acritarch clusters are discussed: for the Jince Formation clusters, aggregation within algal blooms or primary colonial behaviour are the most plausible explanations. In addition, a summary of reports of acritarch clusters from lower Palaeozoic strata is included: clusters of a variety of acritarch genera have been documented from rocks of early Cambrian to Late Devonian age.

Acritarchs are a type of fossil remains of uncertain systematic position (Evitt 1963). Although acritarchs as a group are demonstrably polyphyletic (e.g. Servais et al. 1997), the majority of acritarchs likely represent the remains of various types of algae (Strother 1996), especially the more recalcitrant stages of their life cycles (i.e. cysts; Servais et al. 2016). The fossil record of acritarchs spans from the Precambrian to Recent and reflects various evolutionary and ecological events (Stover et al. 1996; Strother 1996). The abundance, recalcitrance and diversity of acritarchs makes the group highly useful for palaeoecological analyses and biostratigraphical correlations of sedimentary successions, especially in Palaeozoic rocks (e.g. Armstrong and Brasier 2005), although our understanding of many aspects of acritarch biology and ecology is still limited.

Generally, acritarchs are extracted as individual isolated specimens; however, acritarch clusters represent an important and generally understudied phenomenon. Synsphaeridium Eisenack 1965 and other aggregated leiosphere-like spheroids have often been documented and discussed (e.g. Eisenack 1965; Butterfield 2009; Slater et al. 2017), but there have been fewer reports of aggregates of other acritarchs and such clusters have only rarely been analysed (e.g. Downie 1973; Mullins 2003). Modes of aggregation are a crucial issue in recent and subrecent phytoplankton biology (e.g. Mertens et al. 2009; Xiao et al. 2018) and marine ecology (e.g. Smetacek 1985). Aggregation often represents a reaction to changing ecological conditions (e.g. Burd and Jackson 2009; Xiao et al. 2018) or part of a life strategy (Smetacek 1985), and can also serve as a trigger for further ecological interactions including providing new niches and food sources for various organisms (e.g. Huntley et al. 1987; Kiørboe 2001; Simon et al. 2002), both of which can have a notable impact on the entire marine ecosystem. Herein, we present a detailed analysis of monogeneric clusters of five genera, Adara Fombella 1977 emend. Martin in Martin & Dean 1981, Cymatiosphaera Wetzel 1933, Eliasum Fombella 1977, Timofeevia Vanguestaine 1978 and Synsphaeridium. The analysed material originates from Drumian (middle Miaolingian, Cambrian) strata of the Jince Formation of the Příbram–Jince Basin (Barrandian area, Czech Republic; Figure 1). The goal of this contribution is to summarise research on lower Palaeozoic acritarch clusters, discuss the possible origin of the Jince Formation acritarch clusters, and assess their palaeobiological and palaeoecological significance.

The Příbram–Jince Basin is part of the Teplá–Barrandian Unit, the central segment of the Variscan Bohemian Massif (Figure 1B). The basin was formed on the margin of western Gondwana (Drost et al. 2004). The volcano–sedimentary infill of the basin was deposited during the Cambrian and consists of a sequence of predominantly clastic rocks with a maximum thickness of 2500 m. Havlíček (1971) distinguished eight formations capped by upper Cambrian to Lower Ordovician volcanites of the Strašice Volcanic Complex (Drost et al. 2004). The basin was interpreted as an intermontane-type depression by Havlíček (1971) and Kukal (1971). Most of the sedimentary succession is thought to have been deposited in a continental environment; however, at least two units – the Paseky Shale Member of the Holšiny-Hořice Formation and the Jince Formation – are of marine origin (Kukal 1971).

The Jince Formation is composed of fine-grained siliciclastic rocks (shales, greywackes and sandstones) deposited in a shelf marine environment and is the most thoroughly studied unit of the Cambrian sequence (see Geyer et al. 2008; Fatka and Szabad 2014a). Deposition of the Jince Formation started during the Wuliuan (early Miaolingian) and finished during the Drumian (middle Miaolingian; see Fatka and Valent 2019, fig. 1D and Figure 1D herein).

The macrofossil record of the Jince Formation is highly diverse, consisting predominantly of trilobites, agnostids, brachiopods, hyoliths and echinoderms. Various other fossils have been described (for a summary see Fatka et al. (2004) and Fatka and Szabad (2014a)). A review of the ichnofossil record was provided by Mikuláš (2000). Based on the skeletal fauna, a succession of biozones has been distinguished (Fatka and Szabad 2014a; Figure 1D).

Stratigraphy of the studied levels

The studied samples were obtained from a section at the slope called Vinice near Jince, on the northern bank of the Litavka River, particularly from localities 18 and 19 of Fatka and Kordule (1992). A sequence ca. 80 m thick exposed in a series of natural outcrops belongs to the lower third of the Jince Formation (Figure 1C). The lower part of the accessible succession of greywackes and shales with thin layers of fine sandstones represents the upper levels of the Paradoxides (Eccaparadoxides) pusillus Biozone of Fatka and Szabad (2014b). These strata contain abundant trilobites (e.g. Paradoxides, Brunswickia, Conocoryphe, Lobocephalina and Ptychoparia) associated with occasional agnostids (Peronopsis and Phalagnostus) and rare brachiopods (Vandalotreta, Glyptacrothele and Acrothele), echinoderms (e.g. Lichenoides) and bradoriids (Konicekion and Emeiella). Two of the analysed samples (JS-V5 and JS-V6) were obtained from these fossiliferous greywackes (Figure 1C and D).

The other two samples (JS-V3 and JS-V4) were collected from stratigraphically slightly higher levels, particularly from the lower part of the overlying, ca. 25 m thick sequence of the Onymagnostus hybridus Biozone of Fatka and Szabad (2014b) which is characterised by shales and greywackes. The fauna is dominated by agnostids (Onymagnostus, Doryagnostus, Tomagnostus, Peronopsis, Phalacroma and Phalagnostus), with the small eodiscid trilobite Dawsonia, diverse other trilobites, generally rare echinoderms (Lichenoides, Ceratocystis, Vizcainoia, Stromatocystites and ?Ctenocystis), brachiopods (Luhotreta, Glyptacrothele and Lingulella), helcionellid molluscs and hyoliths (see Fatka et al. 2004; Fatka and Szabad 2014a, 2014b).

Micropalaeontological research on the Jince Formation

The first report of organic-walled microfossils from the Jince Formation was that of Slavíková (1968), who documented the occurrence of multiple acritarch genera in the upper part of the formation, more precisely from the Ellipsocephalus hoffi–Lingulella–Paradoxides (Rejkocephalus) Biozone (Figure 1D). Acritarchs from the P. (E.) pusillus Zone were reported by Vavrdová (1982). The unpublished study of Fatka (1987) contains a description of acritarch assemblages from samples across the entire succession of the Jince Formation (Figure 1D). The material from the O. hybridus Zone was studied by Fatka (1989), who reported a diverse acritarch assemblage (Figure 1D). The unpublished study by Tonarová (2006) focused on morphological analysis of four acritarch genera: Adara, Cristallinium Vanguestaine 1978, Eliasum and Timofeevia from the O. hybridus Biozone (Figure 1D). Aggregates of spherical fossils (classified as Symplassosphaeridium Timofeev 1959 ex Timofeev 1966 and Synsphaeridium) were reported by Slavíková (1968) and Fatka (1987, 1989). No other acritarch clusters have so far been documented from the Příbram–Jince Basin.

Bubík (2001) described fragmentary remains of foraminifera tests assigned to the genus Thuramminoides. These foraminiferal remains, with fragments of trilobite exoskeletons and sponge spicules, were obtained via mechanical disintegration of a rock sample from the O. hybridus Zone.

Each of the four samples (JS–V3 to JS–V6) consists of 20 g of grey–green silty shale. The samples were processed via the ‘low-manipulation hydrofluoric acid (HF) extraction’ method of Butterfield and Harvey (2012). The obtained residues were passed through a 30-μm sieve. For samples JS–V5 and JS–V6, the smaller fraction (below 30 μm) was retained and passed through a 10-μm sieve. The organic residue was stored in ethanol.

Isolated acritarchs and acritarch clusters (as well as other fossil remains) were hand-picked by pipette, mounted on coverslips and attached to glass slides. Slides are deposited at the Institute of Geology and Palaeontology Charles University (Prague, Czech Republic). The mounted fossils were examined and photographed using an Olympus BX 51 light microscope. The 'Adobe Photoshop CS6’ software was used for focus stacking.

Both isolated and clustered acritarchs were measured. For each acritarch two measurements were made: the maximal length of the acritarch (MaxL) and the length in the direction normal to MaxL (MinL). Processes were not included in measurements of MaxL and MinL. Only specimens in which both MaxL and MinL could be measured with sufficient precision were considered. For Eliasum, the values of MaxL and MinL were recalculated to a nominal diameter (i.e. the diameter of a sphere with the same volume; Dn), to facilitate comparisons with other genera. For the purposes of this calculation, Eliasum was modelled as an ellipsoid with the x-axis corresponding to MaxL and both the y-axis and the z-axis corresponding to MinL. The nominal diameter was then calculated as
Dn = (MaxL × MinL × MinL)1/3
(1)

AvgL represents the average MaxL of specimens measured within one cluster. CV denotes the coefficient of variation of MaxL within a single cluster; CV was not calculated for clusters in which only one specimen was measured. Graphs (see Figure 2) were made using the PAST 4.04 software (Hammer et al. 2001).

The processed samples yielded both acritarch clusters and isolated specimens. The clusters are exclusively monogeneric in composition; however, occasionally, specimens of two different genera are attached together (e.g. Plate 2, figure 5; see Section 5), or an otherwise monogeneric cluster may rarely include an attached specimen of another genus. The state of preservation of clusters is highly variable. Clusters were comparatively abundant in all four samples.

The dimensions of 190 isolated acritarch specimens (including 73 leiospheres) and 596 specimens in 87 clusters were measured. The measured acritarchs belong to the genera Adara, Cymatiosphaera, Eliasum, Synsphaeridium and Timofeevia. Counts of isolated acritarchs and clusters are provided in Table 1. More than one hundred additional clusters of Synsphaeridium and Cymatiosphaera were observed, but these clusters were not analysed in detail because the number of measured specimens was considered sufficient. Furthermore, 14 sheets associated with clusters of spheroids attached to the sheet surface (e.g. Plate 5, figures 1, 3) and two specimens of Symplassosphaeridium were observed.

Timofeevia (Plate 1)

Timofeevia is present in samples JS-V3 and JS-V4 (i.e. samples from the O. hybridus Zone). Seven clusters were recovered, of which the smallest cluster consists of two acritarchs. The largest cluster (JS-V4a-038, Plate 1, figure 15) contains more than 40 specimens and is the only cluster of Timofeevia to contain more than 10 specimens. Owing to the generally poor preservation and a large degree of overlap, only about one quarter of the specimens in the largest cluster could be measured. In total, 45 clustered specimens and 10 isolated specimens were measured. The MaxL of individual specimens within clusters is 16.1–34.2 µm (Figure 2A) and the MaxL of isolated specimens is 24.1–36.3 µm (Figure 2A). AvgL is 21.7–29.2 µm and CV is 0.09–0.17.

Eliasum (Plate 2)

Eliasum is present in all analysed samples. From the lower sequence, only one cluster consisting of two specimens was recovered (Plate 2, figure 13). Two clusters were obtained from the upper part of the sequence (Plate 2, figures 12, 14). In both of these clusters, only some of the specimens could be measured. In total, 12 clustered specimens and 15 isolated specimens were analysed. MaxL of specimens in clusters is 28.3–85.6 µm (Figure 2B); MaxL of the isolated specimens ranges is 46.3–152.3 µm. Dn (see Section 3) of clustered specimens is 17.2–39.9 µm and Dn of isolated specimens is 22.6–75.8 µm. AvgL is 37–66.2 µm and CV is 0.02–0.22.

Adara (Plate 3)

Specimens belonging to Adara are present in all samples. Eleven clusters were observed in samples from the lower part of the sequence (i.e. the upper part of the P. (E.) pusillus Zone), but no clusters were recovered from the upper part of the sequence (i.e. the lower part of the O. hybridus Zone). Clusters consist of 2 to ca. 15 specimens (Plate 3). The dimensions of 54 clustered specimens were measured – the other clustered specimens were either poorly preserved or overlapping and could not be measured. The MaxL of clustered specimens is 17.2–41.8 µm. Twenty isolated specimens were measured: the MaxL of these specimens is 19–51.2 µm (Figure 2C). The AvgL is 21.75–37.2 µm and the CV is 0.03–0.25.

Cymatiosphaera (Plate 4)

Specimens assigned to the genus Cymatiosphaera were observed in all samples; this taxon is the most common across all samples. A total of 72 isolated specimens were measured. Cymatiosphaera clusters consist of three to ca. 70 individual specimens (Plate 4). In nearly all clusters some specimens could not be measured. In total, 271 clustered specimens were measured. MaxL of Cymatiosphaera within all but one cluster is 9.7–39.1 µm (Figure 2D). The AvgL of these clusters is 20.4–29.3 µm and the CV is 0.06–0.26. One of the analysed clusters (Plate 4, figure 16) consists of specimens with MaxL of 45.8–60.1 µm (average 49.3 µm) and CV of 0.1; specimens within this cluster are also morphologically distinct from those of other Cymatiosphaera clusters. The MaxL of isolated Cymatiosphaera is 19.1–70.6 µm (Figure 2D).

Taxonomic note: A wide variety of morphological forms showing the surface divided into roughly polygonal fields delimited by muri rising up to several micrometers above the surface were observed. However, as pointed out by Downie (1982), such variability of the surface pattern can be produced by collapse of Retisphaeridium Staplin et al. 1965. In this study, all these forms are classified as Cymatiosphaera.

Synsphaeridium (Plate 5)

Clusters assigned to Synsphaeridium were recovered from both the lower and upper sequences. The clusters contain five to more than 80 spheres (Plate 5). In nearly all clusters (especially in the larger ones), some specimens could not be measured (e.g. Plate 5, figure 7). Two hundred and fourteen clustered spheres were measured. The MaxL of specimens within Synsphaeridium is 13.6–40.8 µm (Figure 2E). The AvgL is 17.3–34.5 µm and the CV is 0.06–0.28. The dimensions of isolated leiospheres are not included, because the lack of defining characteristics would preclude the selection of objects that are analagous to those contained within Synsphaeridium and elimination of the rest of the objects assigned to leiospheres.

Summary

The studied material includes 29 clusters classified as Synsphaeridium, 36 clusters of specimens classified as Cymatiosphaera, 11 clusters of specimens classified as Adara, and seven clusters of specimens classified as Timofeevia. Only three clusters of specimens classified as Eliasum were recovered, one of which is very poorly preserved.

Observed clusters differ markedly in size, because they consist of a highly variable number of specimens, from two to more than 100 acritarchs. In all genera, we found clusters with low, and clusters with high internal variation of individual specimen size (i.e. having a low or a high CV value – see Figure 3A).

The recovered clusters include massive, densely packed accumulations (e.g. Plate 4, figure 18), clusters consisting of specimens arranged in ‘filament-like’ rows (e.g. Plate 4, figure 11) and clusters of loosely aggregated specimens (e.g. Plate 1, figure 11). However, no common pattern of organisation was observed.

Preservation and size variation

The preservation of individual clusters varies greatly, from well-preserved clusters with easily discernible specimens to clusters in which individual specimens were poorly distinguished (and thus their dimensions could not be measured). In approximately half of the measured clusters, the state of preservation and the tendency of acritarchs to overlap mean that only some specimens can be distinguished and measured; the remaining elements often form a poorly differentiated to completely undifferentiated organic mass.

The MaxL of individual acritarchs within clusters generally does not exceed 45 µm, with the MaxL of the majority of specimens being between 20 and 40 µm (Figure 2). Exceptions are Eliasum, of which the largest specimen within a cluster is roughly 85 µm long (Figure 2B), and a single cluster assigned to Cymatiosphaera?, in which the MaxL of individual elements is between 45.8 and 60.1 µm (green rhombi in Figure 2D). In contrast, the MaxL of some isolated acritarchs is greater than 45 µm. The difference is most pronounced in Cymatiosphaera (Figure 2D); and is also noticeable in Adara (Figure 2C). In Timofeevia, both isolated and clustered specimens measure less than 40 µm in diameter (Figure 2A).

In contrast to other authors (e.g. Deunff 1968; Cramer and Díez de Cramer 1972), we observed notable variation in the size of individual acritarchs within some clusters (e.g. Plate 2, figure 12; Plate 3, figure 20). In some clusters, all acritarchs are of comparable size (e.g. Plate 1, figure 11; Plate 4, figure 11). This variation is also apparent in the relatively wide CV range (see Figure 3A); however, this metric should be applied with caution, because its values are easily skewed by uneven deformation of individual specimens and (especially in smaller specimens) by even relatively small measurement imprecision. Furthermore, AvgL differs markedly between individual clusters; this difference is most pronounced in Adara and Eliasum; however, the ranges of AvgL in Adara, Cymatiosphaera, Synsphaeridium and Timofeevia are generally comparable with only a few outliers, most notably in Adara (Figure 3B).

Previous reports of acritarch clusters

Clustering has been reported in numerous Cambrian to Devonian acritarch taxa (Figure 4; Supplementary material, Attachment 1). Although some authors have claimed that clusters are generally common (e.g. Cramer and Díez de Cramer 1972), reports of clusters are rather sparse. The most commonly reported examples are Synsphaeridium and other clusters composed of leiosphere-like spheroids (e.g. Symplassosphaeridium and Satka Jankauskas 1979). Such clusters have been reported from Precambrian rocks (e.g. Babu et al. 2014; Riedman et al. 2014; Shukla et al. 2020) as well as from Phanerozoic strata (e.g. Eisenack 1965; Slater et al. 2017; Machado et al. 2020) and are known from numerous stratigraphical levels in many areas of the world. Noteworthy is the recently described Involusphaeridium Filipiak et al. 2021 represented by isolated leiosphere-like spheroids and clusters of those spheroids enveloped in a thin membrane (Filipiak et al. 2021).

For other types of acritarch, generally ‘non-sphaeromorph’ taxa, clustering has been less thoroughly studied. Usually, when clustering has been noted, there have been only one or two reports for a given genus. The most notable exceptions are the genera Asteridium Moczydłowska 1991 (e.g. Moczydłowska 2011; Jachowicz-Zdanowska 2013), Baltisphaeridium Eisenack 1958 (e.g. Cramer 1966a; Quadros 1999) and Comasphaeridium Staplin et al. 1965 (e.g. Cramer and Díez de Cramer 1977; Palacios et al. 2017), for which the clustering habit has been reported by several authors. For the genera Adara, Eliasum and Timofeevia, clusters have not previously been reported.

The majority of reports on acritarch clusters describe monospecific clusters, but some clusters containing two genera are known. Henry (1964, fig. 5) figured an accumulation of numerous specimens of Leiosphaeridia miloni Henry 1964 and Micrhystridium baciliferum Deflandre 1946 observed on a rock surface (d'ensemble de la colonie d'Acritarches”; Henry 1964, p. 1002), and Le Hérissé (1989, 123) reported clusters composed of acritarchs assigned to Dilatisphaera williereae (Martin 1966) Lister 1970 and Elektoriskos sp. A.

Some specific cases of aggregation of acritarchs and algal fossils have been documented, and can easily be distinguished from the clustering described herein. Clusters of spherical internal bodies have been reported in several acritarch genera, including Ancorosphaeridium Sergeev et al. 2011 emend. Moczydłowska & Nagovitsin 2012, Hoegklintia Dorning 1981, and specimens classified as Baltisphaeridium-like, Cymatiosphaera-like and Neoveryhachium-like (e.g. Kaźmierczak and Kremer 2009; Wood 2009; Moczydłowska 2016); these have been interpreted as developmental stages. Also clearly distinct are coenobial forms such as Deflandastrum Combaz 1962, Ericanthea Cramer & Díez de Cramer 1977, Kahfia Le Hérissé et al. 1995, Proteolobus Wood 1997 and Speculaforma Miller et al. 2017. Usually, well-organised Palaeozoic forms consisting of simple spheres (such as Grododowon orthogonalis Strother et al. 2017) have been interpreted as colonies of chlorophycean or zygnemaphycean algae or prasinophycean phycomata (see discussion in Navidi-Izad et al. 2020). These coenobial forms exhibit a highly organised structure, unlike the other clusters discussed herein.

Mechanisms of clustering

Possible mechanisms leading to the formation of acritarch clusters were discussed by Mullins (2003), who listed five possibilities:

  1. formation by chance during laboratory processing;

  2. clusters representing sporangia;

  3. formation of clusters via faecal pelletisation;

  4. formation of clusters as aggregates via coagulation in phytoplankton blooms; and

  5. aggregation as a defense mechanism.

A sixth option was proposed by Tappan (1980) – postmortem clumping on the sea bottom at burial. In general, under variable conditions, clusters may form by a variety of mechanisms. Below, Alternatives 1, 2, 3 and 6 are only briefly discussed as they are deemed less likely to represent the cause of clustering within our material. Alternatives 4 and 5 are more plausible explanations for the Jince Formation clusters and are discussed in detail.

Alternative 1: Formation by chance during laboratory processing. Most clusters from the Jince Formation clearly did not form via a random aggregation during laboratory processing or by any other random process during deposition and fossilisation, because of the co-occurrence of multiple types of exclusively monogeneric clusters as well as the obvious tendency of large clusters to generally show poorer preservation within our material. Nonetheless, random aggregation during laboratory processing should always be considered when studying clusters of microfossils and has been proposed as the formation mechanism for some clustered material (e.g. Agić 2016). In our material, some smaller accumulations of acritarchs (e.g. Plate 2, figures 5, 13) probably resulted from random aggregation during laboratory processing or subsequent handling of the sample (e.g. clumping during placement on a slide); however, elucidating the mode of origin of the majority of clusters is more challenging.

Alternative 2: Sporangia. Downie (1973) proposed a possible affinity with Parka Fleming 1831 for some aggregates of leiospheres. Spore masses broadly comparable in size and outline to larger acritarch clusters have been repeatedly reported (e.g. Wellman et al. 1998; Edwards et al. 1999) and interpreted as infills of sporangia (Wellman et al. 2003). However, interpretation of a wide variety of acritarchs as spores of multicellular algae or land plants does not seem to be well substantiated, especially for Cambrian taxa (e.g. Servais et al. 1997).

Alternative 3: Faecal pelletisation. Faecal pelletisation has been discussed as an important mechanism in clustering of primary producers, enhancing the vertical flux and establishing conditions for various ecological interactions (e.g. Turner 2015). Harvey and Butterfield (2011) demonstrated this process for the fossil record. However, no evidence supporting the formation of clusters via faecal pelletisation was discovered in our material.

Alternative 6: Postmortem clumping on the sea floor (Tappan 1980, 152). This process might form monogeneric clusters in the aftermath of large blooms; however, there would presumably be a high chance of random aggregation of multiple acritarch genera, unless the newly formed acritarch clusters were established under conditions in which the flux of one taxon to the seafloor dominated and the clusters were buried quickly.

Alternatives 4 (coagulation in phytoplankton blooms) and 5 (aggregation as a defense mechanism) need to be addressed in more detail. One of the very conspicuous features of the Jince Formation clusters is the size range of individual elements: generally, the diameter (MaxL) of clustered specimens does not exceed 45 µm and at least some isolated specimens of the same taxa exhibit a MaxL larger than that of clustered specimens. Eliasum is exceptional in that clustered specimens have a MaxL of 85 µm; however, this observation can be explained by the elongate habit of Eliasum. Furthermore, the calculated nominal diameter (Dn) for each of the clustered specimens of Eliasum does not exceed 45 µm (see Section 4.2); the nominal diameter combined with the general shape and length/width ratio implies similar hydrodynamic properties in terms of sinking (see Mcnown and Malaika 1950), even though additional factors (such as the impact of ornamentation and variations in density) are also important. This similarity could imply that the underlying mechanism for clustering is physical and not primarily biological in its character. A comparable phenomenon was reported by Moczydłowska (2011), who noted that clustered specimens of Asteridium from the Lükati Formation are notably smaller than their isolated counterparts.

A possible explanation for the occurrence of clusters is coagulation of acritarchs during algal blooms. In an environment containing a sufficient concentration of sticky particles (in this case, acritarch specimens), the particles have a high probability of collision and subsequent coagulation (see Jackson and Lochmann 1992; Burd and Jackson 2009). Collisions are generally driven by three main causes (Jackson and Burd 1998):

  1. Brownian motion;

  2. differential settling; and

  3. laminar and turbulent shear.

The coagulation rate is further influenced by the ‘stickiness’ of particles, i.e. the probability of two particles remaining attached after a collision. The ‘stickiness’ can be affected by the presence/absence of processes (e.g. Butterfield 1997) as well as by various biological mechanisms, including excretion of exocellular polymeric substances (Burd and Jackson 2009). Other factors, such as the presence of various clay particles, can also have an impact on the formation of aggregates (Hamm 2002). Coagulation has been cited as a crucial mechanism enhancing vertical transport (especially in the aftermath of algal blooms; e.g. Smetacek 1985; Mertens et al. 2009), because the particle size positively influences the sinking rate. This mechanism was proposed by Mullins (2003) as a possible explanation for some acritarch clusters and was discussed as a probable major factor in vertical transport in past environments by Butterfield (1997).

The process of coagulation would explain the random structure of clusters. Furthermore, the similarity of the size distribution pattern of the Jince Formation clusters (Figure 2) (except for clusters in sheets and in Symplassosphaeridium) might be attributed to an underlying set of environmental conditions that caused the aggregation. Because the observed clusters are monogeneric, a temporal/spatial separation of acritarch taxa is most probably required, e.g. consecutive monogeneric blooms. Such blooms would result in the high concentration of acritarchs that would be necessary for coagulation, and would require notable shifts in the phytoplankton composition over time, e.g. seasonality associated with different ecological needs of individual taxa. This explanation is speculative; however, a comparable phenomenon (i.e. a causal relationship between the succession of species abundances in the water column and their succession in aggregates) has been demonstrated in recent diatoms (see Riebesell 1991), and various types of phytoplankton (most notably including many dinoflagellates) do form monospecific blooms (e.g. Smayda and Reynolds 2003; Sunda et al. 2006).

Colonial behaviour is another possible explanation for the formation of acritarch clusters. In extant ecosystems, formation of aggregates occurs in various types of phytoplankton. The formation of colonies can be induced or influenced by a variety of triggers, such as shifts in salinity (Wei et al. 2017), raised levels of pollutants (Sztrum et al. 2012), or even the presence of predators (Lurling and Beekman 2006). In the context of acritarch clusters, the formation of palmelloids, i.e. generally unorganised ‘colonies’ that are variable in size and composed of a diverse number of unicells bound together by secreted organic matter, is potentially relevant. Such palmelloids are known in numerous groups of algae (see e.g. Starr 1984; Rosovski 2003; Sahoo and Baweja 2015). Alternatively, unorganised monogeneric clusters of acritarchs have previously been interpreted as possible remains of colonies of Chlorococcales (e.g. Moczydłowska 2011). However, applying this explanation to the material from the Jince Formation is problematic, because fossils of several morphologically distinct, likely only distantly related taxa exhibit common clustering patterns, particularly the comparable sorting by size (Figure 2). The two processes (i.e. clustering via colonial behaviour and coagulation) are also obviously not separate, because algal colonies can become parts of larger aggregates (e.g. Wassmann et al. 1990).

Accumulations of spheres attached to organic sheets are also present in the Jince Formation samples; these accumulations are distinctly different from the clusters. These sheets resemble an object figured by Butterfield (2009, fig. 1C) and referred to as a ‘simple clonal colony’; however, the poor ‘impression-like’ preservation of these sheets (see Plate 5, figures 1, 3) prevents further interpretation, because it is not even possible to determine whether all specimens associated with sheets are simple spheres, or whether some are ‘non-sphaeromorph’ acritarchs. Moreover, most of the recovered sheets are relatively small fragments with only a few attached spheres.

The Jince Formation clusters were obtained from a stratigraphical level that has previously been studied for organic-walled microfossils (see Section 2.2 and Figure 1D). However, acritarch clusters have not been documented from this unit, except for Synsphaeridium and Symplassosphaeridium. This lack can most likely be attributed to differences in the laboratory processing and focus of previous studies; nonetheless, standard methods of palynological maceration can yield acritarch clusters, as demonstrated by studies listed in Attachement 1 (Supplementary material).

Coagulation of primary producers is known to markedly increase the downward flux of organic carbon and other nutrients in marine environments (Jackson and Lochmann 1992; Butterfield 1997), thus influencing the carbon cycle (e.g. Smith et al. 1991). The existence of such aggregates facilitates ecological interactions within the water column (e.g. Simon et al. 2002). Similarly, colonial stages in algae have been shown to play important roles in numerous ecological interactions (e.g. Hansen et al. 1993; Smith et al. 2014). Therefore, further research on the processes leading to formation of aggregates in the past and comparison of fossil clusters to extant analogs could contribute to a better understanding of the development of the carbon cycle and of trophic interactions.

  1. Various forms of aggregation in primary producers are of high importance for marine ecosystems and carbon cycling. Surprisingly, acritarch clusters are a generally underevaluated phenomenon.

  2. Clusters of Adara, Eliasum and Timofeevia are reported for the first time.

  3. Monogeneric clusters of Adara, Cymatiosphaera, Eliasum and Timofeevia alongside Synsphaeridium, Symplassosphaeridium, and sheets with attached clusters of highly degraded spheroids were discovered co-occurring in palynological residues from the Miaolingian Jince Formation.

  4. In all samples, clustered acritarchs of Adara, Cymatiosphaera, Eliasum and Timofeevia exhibit similarities in their size ranges across taxa. The same pattern of size distribution is also observed in individual spheres constituting Synsphaeridium.

  5. Colonial behaviour or physical aggregation in the aftermath of blooms is the most likely interpretation for the origin of the Jince Formation clusters.

The authors thank Marta Kerkhof, Monika Uhlířová and Jaime Yesid Suárez-Ibarra (Institute of Geology and Palaeontology, Charles University) for their help with obtaining literature, Yvonne Němcová for providing a consultation on specific aspects of extant phytoplankton biology, Lucy Muir for language editing and James B. Riding and two anonymous referees for providing valuable feedback that significantly improved the manuscript.

Vojtěch Kovář is a PhD student at the Institute of Geology and Palaeontology, Faculty of Science, Charles University, Czech Republic. His research is focused on the Lower Palaeozoic palynology, mainly acritarchs and small carbonaceous fossils.

Oldřich Fatka is a professor of palaeontology at the Institute of Geology and Palaeontology, Faculty of Science, Charles University, Czech Republic. His research is focused on the biostratigraphy, palaeobiogeography and palaeoecology of Palaeozoic fossils.

Jakub Vodička is a PhD student at the Institute of Geology and Palaeontology, Faculty of Science, Charles University, Czech Republic. His research is focused on chitinozoans, mainly of Ordovician and Silurian age.

No potential conflict of interest was reported by the authors.

This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.