Trepostome bryozoans buck the trend and ignore calcite-aragonite seas

Trepostome bryozoan skeletalisation did not passively respond to changes in seawater chemistry associated with calcite-aragonite seas. According to Stanley and others, trepostome bryozoans were passive hypercalcifiers. However, if this was the case, we would expect their degree of calcitic colony calcification to have decreased across the Calcite I Sea to the Aragonite II Sea at its transition in the Middle Mississippian. Data from the type species of all 184 trepostome genera from the Early Ordovician to the Late Triassic were utilised to calculate the Bryozoan Skeletal Index (BSI) as a proxy for the degree of calcification. BSI values and genus-level diversity did not decrease across the transition from the Calcite I Sea to the Aragonite II Sea. Nor were there any changes in the number of genus originations and extinctions. This suggests that trepostome bryozoans were not passive hypercalcifiers but active biomineralisers that controlled the mineralogy and robustness of their skeletons regardless of changes in seawater chemistry.


Introduction
Stanley and Hardie (1998Hardie ( , 1999 synthesised a theory of temporal geochemical variation in the oceans through the Phanerozoic (last~550 Myr), describing oscillations in the mineralogy of marine inorganic and biogenic carbonates, with intervals dominated by aragonite and/or high-Mg calcite (aragonite seas) and low-Mg calcite (calcite seas) (Sandberg 1975(Sandberg , 1983Hardie 1996). They argued that temporal variation in the Mg/Ca ratio of seawater was driven by changes in plate tectonic spreading rates at mid-ocean ridges. Faster seafloor spreading lowers the Mg/Ca ratio in the oceans which favours biomineralisation of low-Mg calcite. The opposite creates aragonite conditions.
Regardless of the causal process, if seawater chemistry changes are global in scope, then it might be expected that geochemical signals should be expressed beyond Stanley and Hardie's largely tropical examples and beyond these few biogenic carbonate producers. If bryozoans are passive hypercalcifying animals, they could provide an independent test of these hypothesised temporal variations in seawater chemistry. Bryozoans are mineralogically variable (Smith et al. 1998). They have an excellent fossil record from the Ordovician to the present (Taylor 1993;Ernst 2020). They have adapted to a cool-water temperate to polar palaeolatitudinal distribution through the Phanerozoic (Taylor and Allison 1998). As colonial animals, they have at times been important hypercalcifying carbonate producers (Stanley andHardie 1998, 1999;Stanley 2006). Stanley andHardie (1998, 1999), Stanley et al. (2002), and Montañez (2002) argue that seawater chemistry exerts a strong control on the biomineralogy of morphologically simple clades that have weak control over their calcification. In addition to algae, sponges, and corals, they explicitly mention bryozoans as an example. Stanley andHardie (1998, 1999) and Stanley (2006) suggest the fossil abundance and degree of calcification of 'stony' stenolaemate bryozoans (i.e. trepostomes) were controlled by seawater chemistry because they were considered to be passive hypercalcifiers.
All trepostomes were calcitic (Taylor et al. 2010;Taylor 2020). While the detailed mineralogy of most trepostomes remains unknown, of those whose mineralogy has been determined, all were low-Mg calcite (LMC) with the exception of one, Nicholsonella from the Ordovician of North America, which was high-Mg calcite (Tavener-Smith and Williams 1972;Taylor and Wilson 1999;Smith et al. 2006). Assuming all trepostomes were LMC, we would predict their degree of calcification to be higher during the Calcite I Sea than the Aragonite II sea. If they are not significantly different, that suggests that bryozoans are able to actively biomineralise even in adverse conditions, thus are not passive as suggested by Stanley and others (Stanley andHardie 1998, 1999;Stanley 2006). The fossil record of trepostomes overlaps with the Calcite I Sea-Aragonite II Sea transition which occurred in the Middle Mississippian of the Early Carboniferous. We will test the hypothesis that trepostomes were passive hypercalcifying animals by quantifying their degree of calcification across this transition.

Materials and methods
Trepostomes are an order of stenolaemate bryozoans which evolved in the early Ordovician and went extinct in the Late Triassic (Taylor 2020). Most abundant in the early Palaeozoic, they are informally known as 'stony bryozoans'. All species of trepostomes have long, tubular zooids. Although encrusting, massive, frondose, and bifoliate colonies occur, most are ramose with bifurcating cylindrical branches that form small bush-like colonies (Key et al. 2016). These branches have an inner thinner-walled endozone surrounded by a peripheral thicker-walled exozone (Fig. 1).
We used the Bryozoan Skeletal Index (BSI) developed by Wyse Jackson et al. (2020) as a proxy for the degree of calcification of the trepostomes. The BSI measures the relative proportion of skeletal carbonate to intrazoarial void space in stenolaemate bryozoan colonies. It is calculated from exozone width in longitudinal or transverse section (EW), zooecial wall thickness between adjacent autozooecial apertures in shallow  fig. 2). Abbreviations: EW, exozone width; MZD, autozooecial aperture diameter; ZWT, zooecial wall thickness tangential section (ZWT), and autozooecial aperture diameter in shallow tangential section (MZD) using the following equation: BSI = ((EW × ZWT)/MZD) × 100 (Fig. 1). The multiplication factor of × 100 ensures that BSI is a whole number. Lower numbers represent lower levels of calcification and higher numbers the converse. For circular autozooecial apertures, the maximum MZD was measured. For oval-shaped apertures, both maximum and minimum MZD were determined and the average of the means of these used in the equation. For multilaminar encrusting species in which successive layers overgrow those immediately beneath, the thickness of the exozone in one lamina was measured.
We calculated BSI for the type species of each trepostome genus included in the forthcoming chapter on Order Trepostomata for the Treatise on Invertebrate Paleontology, Part G, Bryozoa, Revised, Volume 2 (Boardman and Buttler, forthcoming). We included all genera, regardless of their zoarial habit (Wyse Jackson et al. 2020, fig. 1). We determined the zoarial habit (i.e. ramose, encrusting, massive, foliose, bifoliate, or mixed) for each type species. When possible, we determined the zoarial habit from the type species' original publication and plates. For many species erected in the nineteenth century for which illustrations are either lacking or uninformative, we used the descriptions from the forthcoming Treatise on Invertebrate Paleontology's trepostome volume (Boardman and Buttler, forthcoming). Our reliance on type specimens of type species of each genus assumes they are representative of the genus. They should be, as that is the fundamental purpose of type specimens and type species (ICZN 1999). But this approach requires the assumption that there is no significant intrageneric, intraspecific, or intracolony variation which is known to exist in bryozoans (Taylor 2020).
When possible, we acquired mean values for the three BSI morphometric parameters (EW, ZWT, MZD) from the type specimens of the type species of each genus. For more recently erected species, the morphometric data were extracted from published data tables. For older species descriptions lacking data tables, we used data from recent revisions of the taxa. Where means or ranges were not reported in the literature, replicate measurements for each parameter were taken directly from figured material accompanying the type descriptions or from the plates of Boardman and Buttler (forthcoming).
In some taxa, the dimensions of some parameters could not be determined from the type specimens of the type species. In these rare cases, it was measured from non-type material from the same species or from another species whose stratigraphic age and geographical location matched as closely as possible to those of the type species. This was mainly necessary for some taxa forming encrusting and massive zoaria in which the exozone width could not be accurately measured. In all such cases, the other parameter dimensions were compared with those of the type species to ensure that they were equivalent.
Type species ages were defined as the midpoint of the type locality stratigraphic age ranges from Shanan Peters' Macrostrat.org. Where type locality stratigraphic ages were not included in Macrostrat.org, we used updated stratigraphic concepts revised since the description of the species or local literature for some non-North American species. Genus stratigraphic ranges were based on the genus descriptions from Boardman and Buttler (forthcoming) and augmented with Phil Bock's Bryozoa.net. All numerical ages were based on the latest version of the ICS geologic time scale (Cohen et al. 2020).
We followed the methodology of Kiessling et al. (2008) and van Dijk et al. (2016) in using t tests to analyse for significant difference between the mean values before and after the Calcite I-Aragonite II transition. We used two approaches. The first examined BSI values across this transition using the midpoint of the age of the type species of each genus. The second used the stratigraphic range of each genus to look for a drop in the diversity of trepostomes across the transition. Hardie (1996) put the Calcite I-Aragonite II transition date at 335 Ma based on the 1983 DNAG geologic time scale (Palmer 1983). Porter (2010, fig. 3), using a more recent geologic time scale (i.e. the 2008 ICS version whose Carboniferous boundary ages have not changed since then; Cohen et al. 2020), showed the transition date at 342 Ma. According to van Dijk et al. (2016), there is some disagreement about when the Calcite I-Aragonite II transition occurred. Depending on if one uses Stanley andHardie's (1998, 1999)

Results
This study included all 184 trepostome genera (Table 1). They ranged from Orbiramus from the Tremadocian stage of the Early Ordovician of China to Styloclema from the Norian stage of the Late Triassic of New Zealand. That is an expanse of 265 Myr from 480 to 215 Ma (Fig. 2). The type species' BSI values ranged from 3 to 211 (mean: 44; standard deviation: 37) ( Table 1). As determined by two sample t tests, there is no significant (i.e. P > 0.05) change in BSI values from before to after the 350 Ma mark, 342 Ma mark, or 333 Ma mark ( Table 2). The mean BSI of all genera per 1 Myr bins across the ranges of all genera varied from 29 to 95 (mean: 53; standard deviation: 14) (Fig. 2). There is essentially no change in mean BSI values at either the 350 Ma mark (47.5 before and after), 342 Ma mark (45.6 before and after), or 333 Ma (45.6 before and after).
Does the lack of change in BSI values hold up for individual zoarial types? There are not enough foliose and bifoliate trepostomes (n = 6), so they were excluded from this additional analysis. Genera whose type species are ramose or most commonly ramose (n = 125) showed no significant change in BSI across the Calcite I Sea/Aragonite II Sea transition, regardless of which of the three transition ages were chosen (two sample ttests, P > 0.05, Table 2). Genera whose type species are encrusting or most commonly encrusting (n = 38) showed no significant change in BSI across the Calcite I Sea/Aragonite II Sea transition, regardless of which of the three transition ages were chosen (two sample t tests, P > 0.05, Table 2). Finally, genera whose type species are massive (n = 15) showed no significant change in BSI across the Calcite I Sea/Aragonite II Sea transition, regardless of which of the three transition ages were chosen (two sample t tests, P > 0.05, Table 2). Therefore, the pattern of lack of change in degree of calcification in response to changes in seawater chemistry is robust across each trepostome zoarial habit. Additionally, there is no change in genus-level trepostome diversity, originations, or extinctions across this transition (Fig. 3).

Discussion and conclusions
There was no significant decrease in BSI values as the calcitic trepostomes transitioned from the Calcite I Sea into the Aragonite II Sea (Table 2, Fig. 2). There was no change in diversity at either the 350 Ma mark (27 genera before and after), 342 Ma mark (26 genera before and after), or 333 Ma (26 genera before and after) (Fig. 3). In addition to the diversity graph being relatively flat during the transition, there was essentially no change in origination or extinction rate (Fig. 3). If trepostomes were responding to the changing ocean chemistry structurally, then one would expect new taxa to appear as trepostome systematics is based on skeletal structures. Trepostomes were structurally resilient to whatever was changing in seawater at the transition. If trepostomes were passive hypercalcifiers, then one would expect a drop in BSI values as the calcitic trepostomes transitioned from the Calcite I Sea into the Aragonite II Sea. If trepostomes were passive hypercalcifiers, then one could expect a macroevolutionary drop in diversity as the calcitic trepostomes transitioned from the Calcite I Sea into the Aragonite II Sea.
In contrast, two prominent diversity drops are visible in Fig. 3     Increased originations did lead to greater trepostome diversity and thus more scatter in BSI values (Fig. 2). This is best seen 460-455 Ma during the Great Ordovician Biodiversification   (Webby et al. 2004). The combination of increased diversity and BSI values may have contributed to bryozoans being the most diverse group of reef-building organisms during the GOBE (Ernst 2018;Servais and Harper 2018). The Ordovician diversification of trepostomes was a component of the GOBE (Taylor and Larwood 1990;Taylor and Ernst 2004), including as a substrate for the Ordovician Bioerosion Revolution (Mángano et al. 2016). The lack of change from the Calcite I Sea into the Aragonite II Sea suggests that the trepostomes were accommodating the change in seawater chemistry in some other way. Laboratory experiments on extant passive aragonite and calcite biomineralisers grown in the different seawater Mg/Ca ratios (Ries 2005;Stanley et al. 2010;Mewes et al. 2014) suggest that passive biomineralisers respond to unfavourable seawater conditions by changing skeletal composition. Thus, the calcitic trepostomes would have been using more energy to precipitate their skeletons in the less favourable seawater chemistry in the Aragonite II Sea, so they may have actively channelled more resources to calcification. Alternatively, they may have been dealing with the changing ocean chemistry metabolically or reproductively, not structurally as evidenced in the BSI. The fact that there was no change in BSI values, diversity, originations, or extinctions at the Calcite I Sea transition into the Aragonite II Sea provides strong evidence that trepostomes were not passive hyper-calcifiers but actively managed their calcification.
From a mineralogy perspective, modern bryozoans are not passive biomineralisers. They can precipitate a mix of LMC, HMC, and aragonite where needed in their skeletons tailored to the functional needs of the colonies (Smith et al. 2006;Taylor et al. 2009). Thus, they are best described as active biomineralisers. This has been previously argued for the post-Palaeozoic cheilostomes (Smith et al. 2006;Taylor et al. 2009) and the Palaeozoic trepostomes (Taylor and Kuklinski 2011).
In Taylor and Kuklinski's (2011) study, they used two independent proxies (i.e. branch diameter and ZWT) for degree of calcification in trepostomes. We used BSI which incorporates three parameters including exozone width instead of branch diameter which is composed of the highly calcified exozone and the minimally calcified endozone. They looked only at dendroid (i.e. ramose) forms, whereas we included all zoarial habits. They compiled their data from the literature for 188 species in 44 genera from the Ordovician (calcite sea), Devonian (calcite sea), and Permian (aragonite sea), whereas we included all genera and the entire trepostome stratigraphic range from the Ordovician to the Triassic and all periods in between. Despite these differences in sample size and range, our results were similar. Trepostome calcification did not alter in response to changes in the Mg/Ca ratio of seawater. Factors other than ocean chemistry control their calcification. Trepostomes, and likely all bryozoans, are active biomineralisers, not passive hypercalcifiers.
Acknowledgements Robert S. Nelson, III, collected some of this data as part of his senior thesis at Dickinson College. Caroline Buttler provided access to the genus descriptions and plates for the forthcoming Trepostome volume of the Treatise. This manuscript was greatly improved by the thoughtful reviews of Andrej Ernst and Abigail Smith.
Funding Open Access funding provided by the IReL Consortium.

Declarations
Conflict of interest The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.