Notes
The William Lake site within the Williams Member of the Stony Mountain Formation (Katian, Upper Ordovician) contains an unusual exceptionally preserved fossil fauna, including arthropods and cnidarian medusae. Several shells of the brachiopod that strongly resemble Holtedahlina paraprostrata Jin and Zhan, 2001 are described for the first time from this unit. Forty percent of the shells were found in an unusually close association with phosphatic tube-like fossils, perhaps indicating a symbiotic relationship, although there is not enough clear evidence to definitively determine the nature of this relationship. It was clearly advantageous to grow on or against skeletal material on a soft, muddy shallow marine seafloor, and both the brachiopod and the tube-like fossil may have benefited from the association at different times. Although other species of Holtedahlina have been found in a variety of depositional environments, Holtedahlina paraprostrata has thus far only been found in hypersaline or variable salinity settings, perhaps indicating an ability to tolerate unusual marine conditions, unlike most other rhynchonelliformean brachiopods that are typically stenohaline. Holtedahlina at the William Lake site, like other calcareous fossils, are rare and poorly preserved perhaps reflecting the unusual preservation at the site or representing a real ecological signal.
Introduction
Strata of the Williams Member of the Stony Mountain Formation near William Lake, north of Grand Rapids in central Manitoba, Canada, contain an exceptionally preserved fossil fauna of Late Ordovician (Richmondian, or late Katian) age (Young et al. 2007, 2012). Several unusual fossil groups not normally preserved in Ordovician strata have been documented from the locality, including cnidarian medusae (jellyfish) (Young and Hagadorn 2010, 2020), eurypterids, xiphosurids (Rudkin et al. 2008), pycnogonids (Rudkin et al. 2013), and other groups. Some of these, such as the cnidarian medusae and pycnogonids, constitute globally unique Ordovician occurrences. A few examples of more typical Ordovician fossil groups, such as gastropods and cephalopods, are also found at the William Lake site. Other marine shelly fossils such as corals, bryozoans, trilobites, and crinoids, are entirely absent, even though they are commonly found elsewhere in Ordovician strata of Laurentia and can be locally abundant in the Stony Mountain Formation in the southern part of the Paleozoic outcrop belt in Manitoba (Elias et al. 2013).
Much of the Stony Mountain Formation is interpreted to have been deposited in a relatively open shallow marine environment, but there is evidence of restriction and increasing or variable salinity in the upper interval from the upper Gunton Member through the Williams Member, with the latter including features such as halite crystal moulds, dewatering structures, and microbial mat textures (Young et al. 2007, 2012). This restricted, sometimes-hypersaline setting may have been difficult for most organisms to tolerate and excluded metazoans common in more normal marine settings in the early Paleozoic, such as corals and most rhynchonelliformean brachiopods. These were instead replaced by a more specialized fauna dominated by linguliformean brachiopods and unusual arthropods such as eurypterids, preserved under conditions that enabled preservation of more delicate forms such as cnidarian medusae.
Rhynchonelliformean brachiopods are rare in the Williams Member, but several moulds of brachiopods that strongly resemble Holtedahlina paraprostrata Jin and Zhan, 2001 have been collected from the William Lake locality. Although this represents an extremely low-diversity and low-abundance rhynchonelliformean brachiopod fauna (in comparison to those of the underlying members of the Stony Mountain Formation and much of the Red River Formation), these few shells still have substantial paleobiological and paleoecological significance. Several of the shells were found in close association with large phosphatic or chitinophosphatic fossil tubes. The affinities of these unusual fossils are still uncertain, but the tubes, preserved as flattened or partly flattened darkened subtriangular structures, seem to share a number of morphological characteristics with some forms referred to hydrozoans and scyphozoans (Van Iten et al. 1992).
These types of associations may have played an important role in the evolution of the shelly benthos. The Ordovician Radiation marked the rise of the filter feeding fauna that would dominate the remainder of the Paleozoic, but also an increase in the complexity of the benthic marine ecosystem. The co-evolution of shelly benthos must have been a driver of this event along with other factors such as the planktonic revolution (Servais et al. 2008), climate change (Rasmussen et al. 2016), and changing oxygenation levels (Edwards et al. 2017), amongst other potential triggers (e.g., Servais et al. 2008, 2009; Stigall et al. 2019).
Geological setting
Lower Paleozoic carbonates form an outcrop belt on the eastern and northeastern edge of the Williston Basin (Baillie 1952; Stearn 1956; Johnson and Lescinsky 1986; Bannatyne 1988; Bezys and Bamburak 2004; Young et al. 2008; Elias et al. 2013; Desrochers et al. 2023). Ordovician strata are best exposed in quarries north of Winnipeg near Stonewall, Stony Mountain, and Garson, but can be found in outcrops along the western shoreline of Lake Winnipeg extending northward through Grand Rapids and across central Manitoba to the Pas and Cranberry Portage near the Manitoba/Saskatchewan border (Bannatyne 1988). These units continue as outcrop into eastern central Saskatchewan but have received comparatively little study there due to limited road access (although there are apparently exposures along the shores of lakes in the region, see Kupsch 1952).
The exposed Upper Ordovician succession in the Grand Rapids Uplands consists of limestone, dolomitic limestone, and dolostone, divided into the Red River, Stony Mountain, and Stonewall formations (Fig. 1). The Red River and Stony Mountain formations correspond to two transgressive–regressive cycles in which changing water depth was associated with variations in restriction and salinity (Elias 1982; Young et al. 2008). The Stonewall Formation forms the uppermost unit of the Ordovician in Manitoba and is bounded by a disconformity at the top separating it from the overlying Silurian Interlake Group. The Ordovician–Silurian boundary has been inferred to lie within this disconformity, based on geochemical data (Elias et al. 2013; Demski et al. 2014).
In the Grand Rapids Uplands, where the William Lake site is located, the Stony Mountain Formation consists of the Penitentiary, Gunton, and Williams members (in ascending stratigraphic order). The Penitentiary Member is composed of silty dolomitic limestones and dolostones. These are overlain by the pale massive dolostones of the Gunton Member, and the recessive dolomudstones of the Williams Member. The lower part of the Stony Mountain Formation succession represents an open marine shelf as shown by the diversity of the macrofossil assemblage (Young et al. 2008; Elias et al. 2013) and high degree of bioturbation (Zheng et al. 2018), but shelly fossils become generally less common up-section in the Gunton and Williams members.
Rhynchonelliformean brachiopods are common throughout most of the Ordovician outcrop belt of Manitoba and are sometimes the most abundant macrofossils in the Stony Mountain Formation (Jin and Zhan 2001), but they are almost entirely absent in the Williams Member. Large strophomenide shells are most diverse, although less common, in the underlying Selkirk Member of the Red River Formation, including Eochonetes clarksvillensis, Nasutimena fluctuosa, Strophomena vetusta, Megamyonia nitens, Tetraphalerella churchillensis, Oepikina limbrata, Kjaerina hartae, and Nasutimena undulosa, occurring with the orthides Diceromyonia storeya, Plaesiomys occidentalis, and Gnamptorhynchos manitobensis; rhynchonellides Hiscobeccus capax, Rhynchotrema iowense, Hypsiptycha occidens, Hypsiptycha anticostiensis, and Lepidocyclus laddi; and the pentameride Parastrophinella cirrita. In the overlying Fort Garry Member, diversity falls sharply and only Holtedahlina paraprostrata is known from the member. A less diverse fauna in the Gunn and Penitentiary members of the lower Stony Mountain Formation includes Oepikina limbrata, Rhynchotrema increbescens, and Hysterocrates gigas, with E. clarksvillensis, S. vetusta, Holtedahlina paraprostrata, M. nitens, P. occidentalis, D. storeya, R. iowense, and Hypsiptycha occidens continuing upward from the Red River Formation. No brachiopods have yet been described from the Gunton or Williams Member at the top of the Stony Mountain Formation in southern Manitoba, but in the Grand Rapids Uplands Elias et al. (2013) reported occurrences of Diceromyonia storeya, Plaesiomys occidentalis (reported as Dinorthis), and Oepikina limbrata in the Gunton Member, and Holtedahlina sp. in the Williams Member.
The unusual fossils at the William Lake site, including eurypterids, xiphosurids, pycnogonids, and cnidarian medusae (jellyfish) are preserved in thin-bedded dolomudstones within the Williams Member (Young et al. 2007, 2012). The interval within which the fossils occur spans a thickness of less than 2 m (Fig. 1). Above and below this interval, other fossils have been collected but the unusual biota is absent. The dolomudstones containing the lagerstätte biota exhibit features including parallel lamination, trough cross lamination, and channels (Fig. 2). Microbial mat features, dewatering structures, and salt crystal moulds can be observed in some parts of the interval. These features, and the unusual fossils, are consistent with marginal marine conditions where the water was shallow, circulation was generally restricted but with occasional intervals of disturbance, and salinity was sometimes high. Conodonts recovered from these strata indicate a latest Ordovician (late Katian) age (Young et al. 2007).
Materials and methods
All specimens examined were deposited in the invertebrate paleontology collection at the Manitoba Museum (MM) in Winnipeg, Manitoba, Canada. In the catalogued material from the William Lake site, a total of 20 slabs contain shells that can be confidently identified as strophomenide brachiopods, of which 13 are identified as Holtedahlina cf. Holtedahlina paraprostrata. Of the 20 strophomenide-containing slabs, eight contain both the brachiopods and phosphatic or chitinophosphatic tube-shaped fossils. All strophomenide brachiopods are preserved as moulds and casts despite the excellent preservation of phosphatic fossils, including linguliformean brachiopods, perhaps reflecting the pervasive dolomitization of the Williams Member. The low abundance and limited diversity of calcitic shelly fossils is notable given that there are many hundreds of part–counterpart slabs in the Museum’s William Lake collection. Specimens I-4581 and I-5476 to I-5479 are considered in the current study.
Specimens were photographed under low-angle lighting to highlight features of the shell. Casts of selected specimens were formed using liquid latex and a two-part silicone putty to try to recreate the features of the original shells. These were subsequently coated with ammonium chloride before photographing to further try to enhance details on the surface. Ultraviolet illumination was used to light the best-preserved specimen showing a close association between the tube and brachiopod, to better reveal any residual organic matter that could indicate the presence of soft tissue forming an attachment between the two organisms.
Systematic paleontology
- Order
Strophomenida Öpik, 1934
- Superfamily
Strophomenoidea King, 1846
- Family
Strophomenidae King, 1846
- Subfamily
Strophomeninae King, 1846
- Genus
HoltedahlinaFoerste, 1924
remarks: Foerste (1924) established Holtedahlina to include brachiopods previously assigned to Strophomena that have a biconvex profile and a distinctive dorsal fold and ventral sulcus at the anterior margin. Strophomenides are commonly concavoconvex or convexoconcave in lateral profile but Infurca, Esillia, and Trotlandella all have biconvex lateral profiles. These have been distinguished from Holtedahlina on the basis of the shape and size and shape of the cardinal process, ventral muscle field configuration, and character of the hinge (Zhan and Cocks 1998).
Holtedahlina cf. Holtedahlina paraprostrata Jin and Zhan,2001
Holtedahlina paraprostrataJin and Zhan, 2001, p. 35, pl. 8
figs. 5−8
holotype: MM I-2543b (Jin and Zhan 2001, pl. 8, fig. 7) from the Fort Garry Member of the Red River Formation at the Wright site on the west bank of the Red River, southern Manitoba, Canada. Deposited with paratypes MM I-2543a and MM I-2543c (Jin and Zhan 2001, pl. 8, figs. 6 and 8, respectively).
occurrence: This is the first occurrence of this genus identified in the Stony Mountain Formation. Holtedahlina paraprostrata described in Jin and Zhan (2001) was from the Fort Garry Member of the Red River Formation and thus far, the species is only known from the Manitoba Williston Basin outcrop belt.
description: Shells small, ranging from 12.3 to 17.8 mm in length and 16.4 to approximately 19.8 mm in length. Semicircular outline with cardinal extremities forming near right angles, and low biconvex lateral profile. Ventral interarea seemingly apsacline (although compressed in studied material); delthyrium poorly preserved, but probably partially covered by pseudodeltidium. Dorsal interarea obscured. Shell covered by rounded, equal-sized costellae increasing through intercalation towards anterior, approximately 3 per mm near anterior margin. Costellae interrupted by two to three prominent growth lamellae in anterior half of shell in most specimens.
Interior poorly preserved and largely obscured in current material. Ventral muscle field suboval in outline and open to anterior, measuring approximately one-quarter of shell length (Fig. 5).
materials: MM I-4581, MM I-5476, and MM I-5478 A (Table 1) all contain external moulds of Holtedahlina cf. Holtedahlina paraprostrata associated with phosphatic tubes with possible affinities to Sphenothallus. MM I-5477 contains a single partial external mould of a ventral valve.
remarks: Holtedahlina is a cosmopolitan genus but the specimens described herein represent only the third occurrence of this species documented in the Williston Basin thus far. Jin and Zhan (2001) described the type material from the Fort Garry Member of the Red River Formation in southern Manitoba and it has since been discovered in the Penitentiary Member of the Stony Mountain Formation. Despite a number of broad similarities between the Upper Ordovician brachiopod faunas of southern Manitoba (Jin and Zhan 2001) and the Hudson Bay Lowlands (Jin et al. 1997), Holtedahlina has not yet been found in the Hudson Bay Basin.
Although the limited number of specimens and poor preservation of the current material precludes a full systematic description, the even costellae and poorly preserved subelliptical ventral muscle field indicate affinities with Holtedahlina rather than other strophomenides that are common in the upper Katian of Laurentia. The biconvex, but relatively flat profile of these specimens fits the description of Holtedahlina paraprostrata well in comparison to other species of the genus that are typically more biconvex although some of the moulds and casts examined in this study show indications of compression.
When they initially described the species, Jin and Zhan (2001) compared Holtedahlina paraprostrata to Holtedahlina sp. from the Maquoketa Shale of Iowa (Wang 1949) and Holtedahlina sinicaZhan and Cocks, 1998 from the upper Katian Changwu Formation of South China. Holtedahlina paraprostrata differs from the former chiefly in being smaller with a thinner shell with lower dental plates in the interior, and from the latter in its internal features including less prominent cardinalia and dental plates that do not dip towards the anterior. The shell interior is largely obscured in the external moulds of Holtedahlina cf. Holtedahlina paraprostrata from the William Lake site, but the poorly preserved impression of a pseudodeltidium (Fig. 4.3) and the larger size of one of the specimens may indicate some gradation between the morphology of the Manitoba and midwestern shells. Holtedahlina sinica has a thin shell and largely open delthyrium and has dental plates that dip towards the centre and more prominent cardinalia. Externally, Holtedahlina sinica resembles Holtedahlina paraprostrata in its low fold and sulcus forming a relatively flat profile.
Holtedahlina sp. was described by Jin and Blodgett (2020) from an unnamed Upper Ordovician limestone in Alaska near the Yukon border. In that publication, they also figured a previously unpublished shell from the Penitentiary Member of the Stony Mountain Formation in Stony Mountain, southern Manitoba (lower in the Stony Mountain Formation below the Williams Member). Both illustrated shells strongly resemble both the Holtedahlina cf. Holtedahlina paraprostrata figured herein and Holtedahlina paraprostrata from the Fort Garry Member of the Red River Formation in southern Manitoba in their low deflection of the anterior commissure because of a relatively weak fold and sulcus and near-right angle cardinal extremities. It seems likely that these represent the same species although the limited number of specimens and poor preservation (especially in the case of the shell from Alaska) make comparisons difficult.
These species differ from the type species of the genus, Holtedahlina sulcata, in lacking a prominent fold and sulcus. Material illustrated by Foerste (1912, 1924) from Indiana and Manitoulin Island shows a more prominent deflection at the anterior and more acute cardinal extremities. Holtedahlina varsensisFoerste, 1924, from east of Ottawa, possesses significantly coarser and less numerous ribs than any other Holtedahlina described thus far.
Several Holtedahlina species have been described from Baltica. In comparison to Holtedahlina sakuensis Oraspõld, 1956 and Holtedahlina rakverensis Männil in Oraspõld, 1956, species from the Caradoc of Estonia described by Oraspõld, 1956, Holtedahlina paraprostrata shares a weak fold and sulcus and a less alate outline than Holtedahlina sulcata. The shell interiors of those species remain poorly known but the ventral muscle field in Holtedahlina rakverensis appears more circular than those illustrated by Jin and Zhan (2001) in Holtedahlina paraprostrata from elsewhere. Holtedahlina paraprostrata lacks the distinctive rugellae of Holtedahlina suedicaCocks, 2005, and has a weaker fold and sulcus.
Holtedahlina orientalisPopov and Cocks, 2006, from the late Caradoc Dulankara Formation of the Chu-Ili Terrane (Kazakhstan), has a much more prominent fold and sulcus that forms a significantly more prominent tongue at the anterior margin in comparison to Holtedahlina cf. Holtedahlina paraprostrata. The cardinal angles are significantly more acute in comparison to the near right angles. Popov et al. (2003) also mention the presence of Holtedahlina in the Angrensor Formation at the Odak section in north-central Kazakhstan and later assigned the species questionably to Holtedahlina under open nomenclature (Nikitin et al. 2006), comparing it to Holtedahlina sulcata from eastern North America. Holtedahlina cf. Holtedahlina paraprostrata appears to have a slightly shallower sulcus and weaker fold than Holtedahlina sp. of Nikitin et al. (2006). Elsewhere in Asia, Holtedahlina sp. Cocks and Modzalevskaya, 1997, from the middle Ashgill of Taimyr, appears to be similar to Holtedahlina cf. Holtedahlina paraprostrata in outline but is somewhat more convex in lateral profile. The internal structures are difficult to compare given that they are poorly preserved in the Manitoban specimens but the cardinalia appear to be more prominent in their shells. Holtedahlina sp. of Nikiforova and Andreeva (1961), from the early Ashgill of Siberia, appears to have a much more prominent ventral interarea than Holtedahlina cf. Holtedahlina paraprostrata and may have slightly more acute cardinal extremities.
Two basic morphotypes of Holtedahlina appear to exist, in that some species have a prominent fold and sulcus and are generally more biconvex, while others are only weakly biconvex with only minor deflection of the commissure at the anterior. There does not seem to be any biogeographic significance to this, as species of both morphotypes are found in Laurentia and Baltica, although the examples from the Kazakh terranes seem to only match the more biconvex morphotype. It could, instead, represent ecophenotypic plasticity in the genus, as the two morphotypes are never found together in the same region. A more convex profile and larger deflection at the anterior may have implications for the shape and size of the lophophore and could have provided better separation of the feeding currents. The advantage of a more planar profile in some species is less apparent, though, and is especially unusual given an overall trend of increasing shell convexity in most strophomenide lineages through the Late Ordovician.
Discussion
Rarity and mouldic preservation of calcareous body fossils in the Williams Member
Although an analysis of the unusual taphonomy of the fossil assemblage at William Lake is beyond the scope of this work, it is significant that none of the shells examined in this study are preserved as the original calcite. The rhynchonelliformean brachiopods are rare, and are all preserved as moulds or casts in dolostone. To form these moulds some process acting on the shells in the environment before burial or during diagenesis caused dissolution of the original calcite. Given the lack of any indicators of high hydrodynamic energy, the shells are more likely to have been dissolved at some stage post-mortem.
Although other fossils are exceptionally preserved at William Lake, including many types of organisms that almost never fossilize elsewhere (e.g., cnidarian medusae), there is a conspicuous rarity of calcareous body fossils in general. As noted above, a variety of benthic organisms have been recovered, but most of these have phosphatic skeletons. Brachiopods are not rare overall, as phosphatic linguliformean brachiopods are abundant, but their calcareous counterparts are conspicuously sparse despite filling a similar niche as a low-level benthic filter feeder despite their abundance in other units of the Williston Basin Ordovician succession. The calcareous corals and echinoderms that are commonly well-preserved elsewhere in the Ordovician units (Young et al. 2008; Elias et al. 2013) are also notably missing, while nautiloid cephalopods are rare, preserved as moulds, and generally represent unusual thin-shelled forms that would have been readily transported by weak waves or currents.
Mouldic preservation of calcareous shells is not unique to the Williams Member in the Paleozoic of the northeastern Williston Basin, however, and is also characteristic of much of the Penitentiary Member lower in the Stony Mountain Formation (e.g., Elias 1982; Jin and Zhan 2001; Young et al. 2008; Elias et al. 2013). This provides some support for our interpretation that these organisms were, in fact, uncommon by the time that the Williams Member were being deposited rather than being apparently rare as an artifact of diminished preservation potential. The sometimes exceptionally saline environment, perhaps in concert with elevated water temperatures in a shallow, restricted, muddy setting may have excluded most other Paleozoic filter feeders that favoured near normal marine conditions. A more detailed analysis of the paleoecology and taphonomy of other fossil groups, in conjunction with diagenetic studies, may provide further insight into this unusual pattern of fossil preservation in the Williams Member.
Association and possible attachment of Holtedahlina cf. Holtedahlina paraprostrata to phosphatic tube-like fossils
Brachiopods are known to have been able to attach to other organisms since the Cambrian (see Topper et al. 2018 for a review of brachiopod attachment strategies). Some productides used curled posterior spines to wrap around crinoid stems, raising their shells above the seafloor (Etheridge 1876; Unklesbay and Niewoehner 1959; Grant 1963, 1966) enabling the brachiopod to filter feed higher in the water column than others on the seafloor. Other brachiopods likely had modified pedicles adapted to attaching to other organisms to similarly reach higher filter feeding tiers (Wright 1967; Harper and Pickerill 1996; Sandy 1996; Wang et al. 2012). Some used their pedicle to attach to mobile benthic organisms, permitting some degree of secondary mobility (Schneider 2003; Topper et al. 2014) unlike most other brachiopods that were generally sessile (although see Dattilo 2004, 2009, for evidence that similar strophomenide lineages were capable of limited movement). Some brachiopods may have even evolved a highly branching form of pedicle similar to the byssal threads that mussels use to attach to substrates (Schumann 1969; Gaspard 1997).
Early strophomenide brachiopods generally have a delthyrium either partially or completely covered by a pseudodeltidium, with a supra-apical foramen. The pedicle passed through this foramen in larval forms, but the foramen then closed in mature shells (Rong and Cocks 1994), preventing development of a robust pedicle that could be adapted to attach to other organisms. These large early strophomenides, unlike some of the later productides, are generally interpreted as having lived unattached (using their large and often flattened valves to lie flat on the seafloor; Thayer 1975) or perhaps semi-infaunally (Stanley 2020). The posterior of Holtedahlina has been poorly illustrated but is generally described as possessing a pseudodeltidium covering the delthyrium in the ventral interarea, including Holtedahlina paraprostrata (Jin and Zhan 2001, pl. 8, figs. 5–8). The ventral interarea is only very poorly impressed in a single mould here (Fig. 3) and the configuration of the interarea is not clear in these specimens.
A single specimen of Holtedahlina sp. from the Maquoketa Shale in Iowa shows a very prominent pedicle opening, however (Wang 1949, pl. 9, fig. E4), indicating that at least some species of Holtedahlina may have maintained a functional pedicle later in ontogeny. If the ventral interarea was at least partially open in these shells and occupied by the pedicle, and the pedicle was large enough to remain functional later in ontogeny in Holtedahlina cf. H. paraprostrata found in the Williams Member, the pedicle may have been used to attach to a firm surface. The fine-grained dolostones of the Williams Member apparently represent a very restricted environment with little if any input of allochthonous sediment, and with sediment being deposited under minimal hydrodynamic energy. A soft, muddy seafloor may have been difficult for sessile epifaunal filter feeding organisms to colonize, and a brachiopod pedicle may not have been able to securely attach to such a substrate (unless it was highly modified as suggested for other lineages, above). Instead, brachiopods that required such an attachment point may have preferentially attached to other organisms, such as fallen phosphatic or chitinophosphatic tubes. These tubes were the only common benthic organisms at the William Lake locality that are considerably larger than the strophomenide brachiopods, making them perfect sites of attachment for these brachiopods. Other than the tubes, the only other substantial skeletal material would have consisted of the occasional shells of nautiloid cephalopods or the exoskeletons of the larger arthropods. If H. cf. H. paraprostrata did not maintain a functional pedicle as the shell grew larger, it would not preclude the possibility of juvenile shells initially settling on the phosphatic tubes following the planktonic larval stage, and then losing their direct connection to the tubes as the pedicle became dysfunctional. This would be expected if these shells lost their pedicle opening through ontogeny like many other strophomenides in the early Paleozoic.
In light of this possible association between Holtedahlina cf. H. paraprostrata and the phosphatic/chitinophosphatic tube-shaped fossils, it is notable that, of the 20 strophomenide-containing slabs, eight (40%) contain both the brachiopods and tube-shaped fossils. In several of these instances, the brachiopod is within several centimetres of a tube (Figs. 3A–3D), and in two instances the brachiopod and tube are superimposed (Figs. 4A–4D). Given the relative rarity of the strophomenide brachiopods, and given that the tubes are sparsely distributed through a substantial area of outcrop, this proximity is remarkable and suggests that this association is more than happenstance. Further, on the remaining 12 strophomenide-containing slabs in which tubes are absent, nearly all of the brachiopods occur in shell lags associated with other shelly fossils of similar size, suggesting that they have been subject to some degree of sorting. All of the strophomenide specimens from the lagerstätte interval are from horizons where the tube-like fossils are also known to occur, so it is quite possible that the proportion that lived in proximity to tubes was higher than indicated by the basic numerical data.
An alternate explanation could be that the brachiopods served as anchors for the (presumably) filter feeding phosphatic or chitinophosphatic tubes. Bolton (1994) suggested that Sphenothallus, a somewhat similar tube-like fossil, may have used brachiopods as hard substrates as anchors on an otherwise soft seafloor to explain the unusual association of Sphenothallus with Rafinesquina (brachiopod) shells in the Collingwood Member at Bowmanville Quarry in Ontario (GSC 108060). In that specimen, tubes appear to pass both directly beneath and above several shells (e.g., pl. 1.1, fig. 2). This could, instead, be interpreted as the brachiopod attaching to the tube but it is much more ambiguous given the position of the brachiopod apparently near the pointed end of the tube. Although the tubes do not terminate under the brachiopod shells, they do pass beneath the brachiopods or adjacent to the shells in a way that they plausibly could have been attached in life by soft tissue that is no longer preserved (Figs. 3 and 4).
A third possibility is that there is no real biological association at all between the Holtedahlina cf. H. paraprostrata in the Williams Member and the tube-like fossils. Very few of the moulds and casts collected from the William Lake locality show definitive evidence of being conjoined shells that would provide further evidence for being preserved in life position. Of our figured material, only the shell illustrated in Fig. 4 shows evidence of both valves, and the ventral valve is only represented by a poorly preserved impression of the posterior. If these are instead disarticulated specimens, they could represent bioclastic material accumulated postmortem through sedimentary processes such as winnowing or transport. This would suggest that they are not in life position, and that this is just a coincidental association. Unlike many of the linguliformean brachiopods that are typically better preserved in this unit, we lack complete shells that would allow us to rule out this hypothesis completely, but the common association of Holtedahlina with phosphatic tubes makes this scenario less likely, especially when at least some shells appear to have clustered around the phosphatic tubes (Fig. 4B).
Holtedahlina paraprostrata as a euryhaline brachiopod in the epicontinental seas North America
Thus far, Holtedahlina paraprostrata has only been documented in the Penitentiary Member of the Stony Mountain Formation (Jin and Blodgett 2020, fig. 5.5–5.8) and in the Fort Garry Member of the underlying Red River Formation (Jin and Zhan 2001, p. 35, pl. 8, figs. 5–8) in the Williston Basin of Laurentia. Both members are thought to represent stressed, restricted environments (Elias 1991) or were deposited under conditions of elevated temperature (Pratt and Haidl 2008). Both the Williams and Fort Garry members host fossil suites representing notably low diversity ecosystems (Elias 1991; Jin and Zhan 2001; Young et al. 2008), where the typical normal shallow marine assemblage of tabulate and rugose corals, brachiopods, bryozoans, and stromatoporoids was replaced by unusual faunas dominated by linguliformean brachiopods, cnidarian medusae, and unusual arthropod groups (Young et al. 2007, 2012).
Given its affinity for these types of settings, Holtedahlina paraprostrata (including the Holtedahlina cf. H. paraprostrata documented herein) seems to be an unusual rhynchonelliformean brachiopod that was able to tolerate, if not thrive, in stressed living environments. Rhynchonelliformean brachiopods tend to favour open marine conditions and are generally not found associated with salt crystal moulds, such as can be seen in the Williams Member. There is no evidence of any type of high energy intervals or event deposits that might have enabled these shells to be carried into the lagoonal setting by a tidal current, storm deposit, or wave activity, indicating that these shells were only minimally transported. There also does not appear to be any damage or notable abrasion to the shells, although this is somewhat difficult to determine definitively from the moulds in this collection.
The characteristics that would have enabled Holtedahlina paraprostrata and Holtedahlina cf. H. paraprostrata to tolerate this stressed environment nevertheless remain unclear. Only the biconvex lateral profile of both the ventral and dorsal valves (many strophomenides have at least one flattened valve), and the slightly more pronounced fold and sulcus that form a uniplicate anterior margin (although this is more pronounced in other species of Holtedahlina), differ from features of strophomenides that favoured normal marine conditions. Neither of these characteristics were likely key to survival in the William Lake paleoenvironment, except in perhaps allowing the brachiopod to better prevent fouling of the gape in a generally soft, soupy substrate of carbonate mud by raising the anterior of the shell further above the seafloor. If the species were to have a more robust pedicle that allowed for more secure attachment to other organisms in the environment, this may have been helpful, but we see only the peculiar propensity of Holtedahlina cf. H. paraprostrata to be found in association here with the phosphatic tubes as evidence for that here, and that is circumstantial at best. Brachiopod pedicles are almost never preserved in the fossil record, so studying their direct associations with other organisms continues to be a challenge unless they are cemented to other fossils (as in many craniiformean brachiopods) or attached via calcified spines (such as in the productides above).
Although Holtedahlina paraprostrata and Holtedahlina cf. H. paraprostrata are seemingly only known from low-diversity fossil assemblages likely reflecting difficult living conditions, other species of Holtedahlina occurred in more normal marine settings, even within North America. Holtedahlina sulcata, for example, has been reported from the Richmondian (upper Katian) of Manitoulin Island and the Ottawa Valley in Ontario, which contain a typical normal marine late Katian fossil assemblage with no indication of elevated salinity (Foerste 1924). Similarly, Holtedahlina sp. (Wang 1949) was described from the Fort Atkinson Member in Iowa which was largely deposited under normal marine conditions (Kolata and Graese 1983), indicating that the genus was not confined to hypersaline settings. The relative rarity of Holtedahlina in most fossil assemblages during the Ordovician might indicate that it was often outcompeted by other brachiopods under normal marine conditions, but was able to survive when other groups struggled with fluctuations in salinity. The rarity of Holtedahlina shells at William Lake shows that the species was, however, not thriving in these conditions in the manner demonstrated by the numerous associated linguliformean brachiopods.
Conclusions
The association of Holtedahlina cf. H. paraprostrata with the problematic phosphatic or chitinophosphatic tube-like fossils in the Williams Member of the Stony Mountain Formation at the William Lake site, Manitoba, may represent the use of the tubes as a hard substrate on the part of the brachiopods; 40% of strophomenide brachiopods collected from this site occur in proximity to the tubes. It is also possible, however, that the organism producing a tube-shaped skeleton used the hard-shelled brachiopod as an anchorage in an environment with few hard substrates for attachment. There are relatively few specimens, but the close association of these fossils implies that there was likely some type of in vivo association rather than chance co-occurrence due to post mortem transport.
The occurrence of Holtedahlina cf. H. paraprostrata in the finely laminated argillaceous mudstones of this unit could reflect an unusual niche for this brachiopod. Although the large, flat strophomenide brachiopods of the Ordovician are not unusual in depositional settings dominated by carbonate mud, it is more unusual that these organisms in particular were apparently able to survive in an environment otherwise dominated by a low diversity benthic assemblage in a restricted, sometimes exceptionally saline setting. This has implications for our understanding of the paleoecology of brachiopods, in that it is apparently not enough to simplify the niche of brachiopod lineages down to the family or genus level. There is clearly still significant paleoecological diversity even at the species level, which might be overlooked in more broad-scale analyses.
Acknowledgements
Ed Dobrzanski, David Rudkin, Sean Robson, Michael Cuggy, Deborah Thompson, Matt Demski, Robert (Bob) Elias, Lori Stewart, Norman Aime, Marion Foster, Marjorie Turton, Vicki Young, Erica Young, and Juliana Young are thanked for their collaboration or assistance in collecting specimens at the William Lake site. David Rudkin and Michael Cuggy are thanked for many ongoing discussions of this unusual locality. Joseph Moysiuk provided assistance in accessing the collections at the Manitoba Museum and assisted with some of the photography. This manuscript benefitted from the constructive comments of Jorge Colmenar and two anonymous reviewers. This is a contribution to IGCP 735: Rocks and the Rise of Ordovician Life.
Data availability
Data generated or analyzed during this study are provided in full within the published article. For further information on the collection examined, please contact the authors or Dr. Joseph Moysiuk ([email protected]) at the Manitoba Museum in Winnipeg.
Author contributions
Conceptualization: CDS, GAY
Formal analysis: CDS, GAY
Funding acquisition: CDS, GAY
Investigation: CDS, GAY
Methodology: CDS, GAY
Resources: GAY
Writing – original draft: CDS, GAY
Writing – review & editing: CDS, GAY
Funding information
This work was funded by a Natural Sciences and Engineering (NSERC) Discovery Grant to Sproat and grants from the Manitoba Museum Foundation to Young.