Geologic setting and age

The Eramosa Lagerstätte is situated within the upper part of the Eramosa Formation, as defined by Brett et al. (Reference Brett, Tepper, Goodman, LoDuca and Eckert1995), in the Bruce Peninsula region of Ontario. Most of the Eramosa Formation in this area is composed of well-bedded dolostone with occasional argillaceous intervals, but stromatoporoid-rich bioherms are locally developed (Armstrong, Reference Armstrong1993; Tetreault, Reference Tetreault2001). A Wenlockian age for this interval was reported by Bancroft et al. (Reference Bancroft, Kleffner and Brunton2016) based on conodont evidence and carbonate carbon-isotope stratigraphy.

In an overview of the Eramosa Lagerstätte provided by von Bitter et al. (Reference von Bitter, Purnell, Tetreault and Stott2007), three biotas were recognized, each known only from a single limited geographic area in the southern part of the Bruce Peninsula. The macroalgal taxon described herein belongs to Biota 3 of that report, located in a quarry operated by Owen Sound Ledgerock Ltd. near Wiarton (locality D of von Bitter et al., Reference von Bitter, Purnell, Tetreault and Stott2007). All specimens of the new taxon occupy bedding plane surfaces of large blocks removed from the lowermost part of a small pit on the northeasternmost part of the quarry property, north of Oliphant Road. The stratigraphic section at this location and level comprises light gray dolostone beds 15 to 30 cm thick that alternate with laminated intervals 5 to 15 cm thick. Within the laminated intervals, individual beds range from less than 1 to nearly 10 mm in thickness and alternate between buff to light gray dolostone and dark gray to brown, organic-rich calcareous shale. All of the alga material described herein was recovered from the organic-rich calcareous shale layers. The thicker dolostone beds that separate the laminated intervals show graded bedding and include chaotically oriented brachiopods and coral heads (favositids). The entire section at this location has a gentle southward dip, the dip angle increasing updip and eventually flattening out downdip. This local irregularity likely reflects deposition along the flank of a large bioherm, in which case the alga-bearing sediments accumulated near the base of this structure because the blocks that produced the specimens were quarried from an area close to where the dip begins to level out. The bioherm itself at this location is not exposed, but similar flank beds and their associated bioherms are exposed in nearby roadcuts.

In the main pit on the south side of Oliphant Road and located slightly lower stratigraphically is a further occurrence of Biota 3 of the Eramosa Lagerstätte (von Bitter et al., Reference von Bitter, Purnell, Tetreault and Stott2007). Noncalcified algae occur within this interval as well (Tetreault, Reference Tetreault2001), in this case in association with a diverse array of nonbiomineralized animal taxa, but the algal taxon described herein has not been recovered from this occurrence, or from the other two biotas described by von Bitter et al. (Reference von Bitter, Purnell, Tetreault and Stott2007).

Paleoecology and taphonomy

The new macroalgal taxon comprises part of a distinctive, low-diversity assemblage. Associated fossils are limited to scolecodonts, which often occur as paired assemblages without associated soft parts, orthoconic cephalopods, rhynchonellid brachiopods, high- and low-spired gastropods, rare eurypterids, and several additional noncalcified macroalgal taxa (Tetreault, Reference Tetreault2001; Eriksson and von Bitter, Reference Eriksson and von Bitter2015). Shelly material is sparse, and trace fossils are lacking. Because noncalcified macroalgae dominate the nonbiomineralized component of the biota, the Eramosa Lagerstätte at this stratigraphic level and location is an algal-Lagerstätte (sensu LoDuca et al., Reference LoDuca, Melchin and Verbruggen2011).

Material of the new macroalgal taxon is the most abundant nonbiomineralized component of the biota and in most cases occurs as localized patches devoid of other taxa with the exception of brachiopods, which are preserved as faint, difficult to detect molds. Commonly, surfaces surrounding the algal patches are barren and, occasionally, specimens show alignment (Fig. 1.1, 1.2). Specimens with preserved rhizoids in some cases show this structure in contact with brachiopods (e.g., Fig. 2.1, 2.2), suggesting a preference for attachment to hard substrates. Small specimens typically retain the full complement of laterals, whereas large specimens are often surrounded by a halo of detached lateral segments (Fig. 1.1, 1.2).

Figure 1 Wiartonella nodifera n. gen. n. sp. from the Silurian Eramosa Formation, Wiarton, Ontario: (1) thalli showing current alignment and surrounding localized accumulations of detached second-order lateral segments (ROM 63801); (2) thalli showing current alignment, specimen at left is surrounded by detached second-order lateral segments (ROM 63797). All specimens photographed dry in direct light. Scale bars represent 5 mm.

Figure 2 Wiartonella nodifera n. gen. n. sp. from the Silurian Eramosa Formation, Wiarton, Ontario: (1, 2) thalli with holdfast (arrow) in contact with poorly preserved brachiopods (ROM 63793, ROM 63802). (3) Composite EDS elemental map of a part of ROM 63803 (arrow at lower left in 3.9). Scale bar is (1, 2) 5 mm; (3) 0.33 mm.

Characteristics of the algal material noted above indicate essentially in-situ burial, and the associated sedimentology points to obrution during small-scale storm events, the thin layers of entombing sediment perhaps delivered by sediment gravity flows generated along the flanks of nearby bioherms. Collectively, the lack of bioturbation, preservation of nonbiomineralized material, and biotic composition indicate a poorly oxygenated, low-energy, shallow marine depositional environment. The patchiness of the algal material may have been controlled by the availability of brachiopod valves to serve as attachment sites, and nearby bioherms may have restricted circulation, thereby promoting development of dysoxic bottom waters.

Specimens of the new algal taxon are jet black and, when freshly exposed, are difficult to detect against the nearly equally dark matrix. Upon weathering, however, the fossils become apparent as the tone of the matrix lightens. Energy-dispersive X-ray spectroscopy (EDS) analyses indicate a carbonaceous composition for the algal material (Fig. 2.3). Interestingly, main axes and laterals are often essentially circular in cross-section (Figs. 2.3, 4.1, 4.2). Such three-dimensional preservation indicates particularly rapid lithification of the surrounding sediment following deposition.

Systematic paleontology

Order Dasycladales Pascher, Reference Pascher1931

Family Triploporellaceae (Pia, Reference Pia1920) Granier and Bucur in Granier et al., Reference Granier, Dias-Brito, Bucur and Tibana2012

Tribe Triploporelleae (Pia, Reference Pia1920) Bassoullet et al., Reference Bassoullet, Bernier, Deloffre, Génot, Jaffrezo and Vachard1979

Subtribe Triploporellinae (Pia, Reference Pia1920) Bassoullet et al., Reference Bassoullet, Bernier, Deloffre, Génot, Jaffrezo and Vachard1979

Genus Wiartonella new genus

Ontogeny

Morphologic differences between the various specimens assigned to Wiartonella nodifera are interpreted herein as reflecting a complex ontogeny involving distinct stages. During the initial stage (stage 1) whorls of unbranched hairlike laterals were produced as the main axis commenced elongation (Figs. 2.2, 5.1). After at least five such whorls were produced, the thallus entered a markedly different growth stage (stage 2), during which numerous whorls of branched laterals were produced in succession, the hairlike second-order lateral segments emerging from distinctly expanded terminations of the first-order lateral segments (Figs. 3.1–3.3, 5.2). Finally, late in the ontogenetic sequence (late stage 2), the second-order lateral segments were shed, resulting in thalli largely or entirely devoid of these structures (Figs. 3.7–3.10, 5.3).

Figure 5 Schematic illustrations of Wiartonella nodifera thalli showing different ontogenetic stages: (1) stage 1 thallus with only unbranched hairlike laterals; (2) early stage 2 thallus with branched laterals; (3) late stage 2 thallus for which all second-order lateral segments have been shed. For simplicity, only two laterals per whorl are shown. See text for details.

Size data for the material is consistent with interpretation of stage 1 and stage 2 thalli as ontogenetic stages of a single species rather than distinct species. In particular, in terms of main axis length, the largest specimens interpreted as stage 2 thalli are larger than the largest of the stage 1 thalli by nearly a factor of two. That the second-order lateral segments of W. nodifera were shed as a natural part of the life cycle, as opposed to simply detached as a consequence of transport prior to burial, is indicated by two factors. First, thalli are known that retain all of the first-order segments but none of the second-order (e.g., Fig. 3.7–3.10). Second, some thalli are surrounded by haloes of detached second-order lateral segments (Fig. 1.1, 1.2), a pattern consistent with shedding followed by in-situ burial. Notably, shedding of lateral segments during the ontogenetic sequence is common for living Dasycladales, the cytoplasm from these deciduous segments being retracted into the thallus before abscission commences (Berger and Kaever, Reference Berger and Kaever1992).

Reproductive functional morphology

In comparing Wiartonella with living dasycladalean algae, it is immediately obvious that the expanded terminations of the first-order lateral segments, referred to in this discussion generically as nodes to avoid unwarranted functional connotations, bear a resemblance to gametophores, at least superficially. In particular, the gametophores of Neomeris compare well in terms of size (diameter of 0.075–0.195 mm) and position along the first-order lateral segment (terminal). The earliest dasycladalean with structures described as gametophores, and therefore proposed as choristospore according to the terminology of Pia (Reference Pia1920), is Kalania from the Llandoverian of Estonia (lower Silurian; Tinn et al., Reference Tinn, Mastik, Ainsaar and Meidla2015). A gametophore identity for the illustrated structures, however, is equivocal. A stronger argument can be made for Uncatoella from the Lower Devonian of China (Kenrick and Li, Reference Kenrick and Li1998). Here, structures similar in size, shape, and position to the gametophores of the extant dasycladalean Batophora encapsulate unambiguous clusters of spherical objects comparable to the reproductive cysts of extant dasycladaleans, including the presence of an operculum.

For the nodes of Wiartonella, evidence of the sort necessary to definitively confirm or reject a gametophore identity is lacking. In particular, reproductive cysts are not evident anywhere within the thallus, even in scanning electron microscopy images taken in backscattered-electron mode, so that it cannot be determined that gametes occupied the nodes to the exclusion of other parts of the thallus. This is a key point because many extinct dasycladalean taxa lacked gametophores and, of these, some are known with certainty to have sequestered gametes within the main axis, such as Diplopora phanerospora (see Pia, Reference Pia1920), or within the first-order laterals, such as Triploporella remesi and other triploporellids (reviewed in Barattolo et al., Reference Barattolo, Bucur, Kołodziej, Hoffmann and Skupien2013), categorized, respectively, as endospore and cladospore by Pia (Reference Pia1920). In addition, although living dasycladalean algae are not known to produce laterals directly from the gametophores, an observation in keeping with the view that gametophores have a morphogenetic identity independent of the laterals (so that they are not simply modified lateral segments) (see Dumais and Harrison, Reference Dumais and Harrison2000), current knowledge of dasycladalean morphogenesis is sufficiently incomplete that production of secondary lateral segments from the nodes of Wiartonella does not unequivocally eliminate the nodes-as-gametophores hypothesis. Finally, because thallus morphology for Wiartonella differs markedly from that of all well-established choristospore dasycladalean genera, it is not possible to establish a gametophore identity for the nodes on the basis of phylogenetic arguments.

Although limitations noted above preclude definitive confirmation or rejection of the nodes-as-gametophores hypothesis, it should be possible to apply functional morphology, specifically biophysical modeling, to evaluate the effectiveness of these structures in such a role and thereby gain insights in this regard. Such an approach in this context is feasible because dasycladalean algae reproduce through holocarpy, whereby essentially all of the cytoplasm from various parts of the thallus (main axis and lateral segments) streams into the reproductive structures and is devoted to reproduction, the thallus subsequently perishing. Detailed photographic documentation of this process is provided in Berger and Kaever (Reference Berger and Kaever1992). Thus, for a given thallus, it is reasonable to infer that reproductive volume (R), defined as the collective volume of thallus regions utilized for gamete containment, will approximately equal cytoplasm volume (C). The latter, like reproductive volume, is a small subset of total interior thallus volume, in this case because most of the thallus interior in life is occupied by a large central vacuole, the cytoplasm forming only a thin film along the inner surface of the cell wall (Ngo et al., Reference Ngo, Garland and Mandoli2005) (Fig. 6.1). In basic outline this approach is similar to that used by Maksoud et al. (Reference Maksoud, Granier and Azar2014), but a key difference is that their study did not incorporate the central vacuole in calculations.

Figure 6 Models applied to investigate reproductive functional morphology for Wiartonella nodifera: (1) primary lateral in pre-reproductive state showing thin layer of cytoplasm (gray) along inner surface of cell wall and large central vacuole (white); (2) choristospore model, with node treated as a gametophore and filled with cytoplasm; (3) cladospore model, with entire primary lateral treated as a reproductive structure and filled with cytoplasm. Details of lateral morphology and cytoplasm thickness roughly to scale, but less than half of total length of primary lateral shown. R=reproductive volume; c.v.=central vacuole.

To evaluate whether the predicted relationship outlined above holds true for dasycladalean algae, biophysical modeling was applied to the extant species Dasycladus vermicularis, Neomeris dumetosa, and Cymopolia vanbosseae. For each, cytoplasm volume for a pre-reproductive whorl, including the gametophores, was modeled from aspects of external thallus morphology following the method described in LoDuca and Behringer (Reference LoDuca and Behringer2009), and reproductive volume for the same whorl was calculated as the total interior volume of the gametophore, modeled as a sphere, multiplied by the number of gametophores developed on the whorl. Applying this approach to morphometric data for a single specimen of each species obtained from images in Berger and Kaever (Reference Berger and Kaever1992) yielded C/R values for Dasycladus vermicularis, Neomeris dumetosa, and Cymopolia vanbosseae of 1.4, 1.8, and 1.7, respectively (Table 2). Note that for the latter species, one-third of the cytoplasm volume from an associated “sterile” whorl was included in the total cytoplasm volume, since there are approximately three fertile whorls (i.e., whorls that develop gametophores) for every sterile whorl on the thallus of this species of Cymopolia (Berger and Kaever, Reference Berger and Kaever1992). Without cytoplasm input from the sterile whorls, each of which yielded a calculated cytoplasm volume roughly three times that of a fertile whorl, the C/R for this species falls to 0.8, meaning that more reproductive volume would be produced than could be supplied by cytoplasm. For D. vermicularis and N. dumetosa, the same analyses were performed utilizing morphometric data presented in Maksoud et al. (Reference Maksoud, Granier and Azar2014). Using data from this source, the C/R value calculated for N. dumetosa is comparable to that reported above (1.6 vs. 1.8; Table 2). For D. vermicularis, however, the resulting C/R value differed by a factor of two (0.67 vs. 1.4). To explore this discrepancy, we collected morphometric data from an image of D. vermicularis in Berger (Reference Berger2006, fig. 82), the source utilized by Maksoud et al. (Reference Maksoud, Granier and Azar2014). Our measurements agree with those of Maksoud et al. (Reference Maksoud, Granier and Azar2014, table 8) in all regards except for width values associated with the second- and third-order laterals, which were larger by a factor of about two (Table 2). Incorporating our revised width measurements resulted in a C/R value of 1.5 (Table 2). Collectively, the results provide compelling evidence that reproductive volume closely approximates cytoplasm volume for living dasycladalean algae, as predicted on theoretical grounds. That the calculated C/R values are slightly above unity in each case is reasonable given that it is unlikely a thallus could stream all cytoplasm into the gametophores, a proposition supported by images in Berger and Kaever (Reference Berger and Kaever1992) which show some remnant cytoplasm adhering to the inner cell walls of all reproductively mature specimens. In general, there is no reason to think that a similar C/R relationship should not apply to extinct taxa and, crucially, it is possible to apply this approach to well-preserved fossil specimens.

To evaluate the nodes of Wiartonella as gametophores (choristospore hypothesis; Fig. 6.2), as a first pass, R was modeled as total node volume within a whorl, with nodes approximated as spheres, and C was modeled as cytoplasm volume exclusively for the first-order lateral segments of the same whorl. Results indicate a C/R value of 3.5 for a whorl with values of lateral length, lateral diameter, and node diameter equivalent to those of the holotype, ROM 63795 (Fig. 3.3) (model 1 in Table 3). For five other specimens modeled in this way, C/R values are similar, ranging from a low of 3.5 to a high of 4.7 (avg. 3.95) (Table 3). Note that although modeling of C and R was conducted using 12 laterals per whorl, the number of laterals per whorl is irrelevant from the perspective of the C/R value in this case, because node number and first-order lateral number per whorl are equivalent. Thus, the fact that C is substantially greater than R for Wiartonella when modeled in this way is entirely a consequence of the node being quite small relative to the preceding cylindrical portion of the lateral, the latter being many times longer and over half as wide as the node in all cases (Table 1). Considered more holistically, the difference between C and R is even greater, because C must also include cytoplasm volume from the associated second-order lateral segments (two for each first-order segment). When this factor is included, the resulting cytoplasm volume calculated for the various specimens is roughly an order of magnitude greater than the associated reproductive volume (model 2 in Table 3). Such a relationship casts considerable doubt on the possibility of homology between gametophores and the nodes of Wiartonella, because it would entail that most of the thallus cytoplasm (i.e., >90%) was not used for reproduction.

Interestingly, modeling of the same nature, but treating the entire first-order lateral segment of Wiartonella as a reproductive structure (cladospore hypothesis; Fig. 6.3), yields C/R values comparable to those calculated herein for extant taxa. For example, when R is modeled in this way for the holotype, the C/R value is 1.7, where C includes cytoplasm from the second-order lateral segments, and for the other specimens, C/R values modeled in the same way fall within a narrow range between 1.6 and 2.0 (model 3 in Table 3). Such an outcome provides strong support for the cladospore hypothesis. Nonetheless, cladospore reproduction for Wiartonella cannot be definitively established. Apart from lacking evidence of reproductive cysts (to mark the location of gamete containment), uncertainty arises because modeling indicates C/R values for endospore reproduction (where the entire main axis interior serves as R and C is determined by total thallus cytoplasm volume) not much different from those for cladospore reproduction. For example, for the holotype, the modeled C/R value for endospore reproduction is 1.8. Thus, results of biophysical modeling provide no clear distinction between cladospore and endospore reproduction for Wiartonella. Beyond the distinctive form of the first-order laterals, however, the ontogeny of Wiartonella provides additional support for the cladospore hypothesis, because late in thallus ontogeny the second-order lateral segments were shed but the first-order laterals remained attached. It is also notable in this regard that a few main axes have been found with few or no first-order laterals (Fig. 2.11, 2.12). This suggests that these, too, were eventually shed, during a separate and later shedding event. Such a fate is also known for the reproductive structures of some living dasycladalean taxa, perhaps to aid in dispersal of the gametes while still in a protected state.

Evolutionary significance

Regarded as either endospore or cladospore, Wiartonella is accommodated within the family Triploporellaceae, which comprises euspondyl taxa with either endospore or cladospore reproduction (LoDuca, Reference LoDuca1997; Granier et al., Reference Granier, Dias-Brito, Bucur and Tibana2012). If the euspondyl condition evolved only once in the evolutionary history of Dasycladales, which seems a reasonable proposition given the available evidence, then all euspondyl taxa comprise a single clade and the extinct Triploporellaceae is paraphyletic to extant Dasycladaceae and Polyphysaceae (Kenrick and Li, Reference Kenrick and Li1998; LoDuca et al., Reference LoDuca, Kluessendorf and Mikulic2003). Within the Triploporellaceae, the lack of a cortex and the presence of two orders of laterals indicate placement within the tribe Triploporelleae, and the inflated nature of the first-order lateral segments together with a likely reproductive role for these segments, as indicated by modeling and ontogeny, points to placement within the subtribe Triploporellinae (see Deloffre, Reference Deloffre1988). The phylogenetic significance of these taxa remains to be investigated, but so considered Wiartonella becomes the oldest member of both the tribe and subtribe, the next oldest members being calcified taxa of late Paleozoic age, as well as an early member of a putative lineage that includes the Mesozoic cladospore taxon Triploporella. It is possible that Kalania should be placed in the Triploporellinae as well, given that all aspects of lateral morphology and arrangement are consistent with such an assignment and the ovate objects labeled as gametophores in Tinn et al. (Reference Tinn, Mastik, Ainsaar and Meidla2015) would seem to leave room for alternative interpretations, in part because they are not shown to be in unequivocal anatomical contact with other thallus elements. Perhaps most importantly in evolutionary terms, the development of laterals with distinctly inflated terminations and euspondyl arrangement in both Wiartonella and cyclocrinitids suggests the possibility that these early Paleozoic macroalgae are rooted in a close common ancestor.