Ramose branches of an Ordovician trepostome, Bythopora- gracilis (Nicholson), Cincinnatian, Cincinnati, Ohio, NHM London 90208. Scale bar 10 mm 

Context in source publication

Context 1

... when the Mg/Ca ratio is low (<2), aragonite seas when it is high (>2). This is because calcite is the less soluble and therefore the favoured precipitate of CaCO 3 except when a high Mg level impedes its growth (Berner 1975) and favours precipitation of the more soluble phase aragonite. The pattern emerging for the Phanerozoic is that aragonite seas existed in the earliest Cambrian, calcite seas from then until the mid-Carboniferous, aragonite seas from the mid-Carboniferous until the Triassic, calcite seas from the Jurassic probably until the end of the Eocene, and aragonite seas from the Oligocene until the present day (e.g. Lowenstein et al. 2001). The evolution of carbonate skeletons has taken place against this backdrop of alternations between calcite and aragonite seas. Stanley and co-workers (e.g. Stanley and Hardie 1998, 1999; Stanley et al. 2002; Stanley 2006) have championed the idea that secular variations in seawater chemistry responsible for calcite/aragonite sea alternations have been important in determining the biomineralogy of marine organisms with calcareous skeletons, with calcite biomineralization being favoured during times of calcite seas but aragonite biomineralization during aragonite seas. Such biomineralogical bias can be manifested in four main ways (Stanley 2006; Taylor 2008): (1) newly evolved carbonate skeletons should have the same mineralogy as the sea type at the time of their origin (e.g. Porter 2007, 2010); (2) in organisms already possessing calcareous skeletons, evolutionary switchovers from calcite to aragonite skeletons will occur during times of aragonite seas, and from aragonite to calcite skeletons during calcite seas (e.g. Stolarski et al. 2007); (3) mineralogy may change phenotypically according to seawater type (see Ries 2010); and (4) organisms with calcite skeletons may construct larger or thicker skeletons in calcite seas, those with aragonite skeletons during aragonite seas (Stanley 2006). The last of these phenomena has been termed ‘ hypercalcification ’ . The present paper contributes to the debate about whether hypercalcification does indeed occur when carbonate biomineralogy matches seawater chemistry. Stanley (2006) pointed out that various groups of marine animals and plants with massive calcareous skeletons seem to have become hypercalcified when their mineralogy corresponded with that favoured by the ambient conditions. This correlation appears particularly strong for reef builders. For example, the main reef builders in the calcite seas of the Ordovician to Devonian were tabulate corals and stromatoporoids, both having calcite skeletons. These were succeeded in the Late Carboniferous to Permian by chaetetid, inozoan, sphintozoan and ‘ sclerosponges ’ with aragonitic skeletons. Organisms with dominantly calcite skeletons, notably rudistid bivalves, constructed reefs during the calcite seas of the late Mesozoic, followed by aragonitic scleractinian corals in the aragonite seas of the later Cenozoic. Trepostome bryozoans are one of the calcitic groups considered by Stanley (2006, p. 218) to have become hypercalcified in the extreme calcite seas of the Ordovician. Stanley pointed to the rock-forming ability of trepostomes in the Ordovician. Trepostomes did indeed produce very large colonies in Ordovician times (e.g. Cuffey et al. 2000), occasionally constructing small reefs (e.g. Cuffey and Fine 2005) and often being present in rock-forming abundances (Taylor and Sendino 2010). However, equally large, if not larger, trepostome colonies are reported from the Permian (e.g. Håkansson and Madsen 1991; Reid 2003) where they can also be rock-formers, despite the fact that this was a period of aragonite seas when hypercalcification of such exclusively calcitic bryozoans would not be expected. Our aim here is to test the hypothesis of trepostome hypercalcification in calcite seas by comparing the values of two quantitative proxies for hypercalcification (branch diameter and exozonal skeletal wall thickness) in trepostomes from the Ordovician, a time of calcite seas, with the Permian, a time of aragonite seas. To control for possible temporal evolutionary effects, data from trepostomes of intermediate age (Devonian), another period of calcite seas, has also been compiled. If trepostome hypercalcification did occur in the Ordovician then branch diameter and wall thickness should be significantly greater in species from this period. Trepostomes are an order of stenolaemate bryozoans ranging from Lower Ordovician to Upper Triassic. Partic- ularly common in the Early Palaeozoic, they are sometimes referred to informally as ‘ stony bryozoans ’ . All species of trepostomes have long, tubular zooids with calcitic walls. Although encrusting, dome-shaped and foliaceous trepostomes can be found, most species are dendroid, comprising bifurcating, cylindrical branches that form small bush-like colonies, commonly broken into short branch lengths during fossilization (Fig. 1). The branches of dendroid trepostomes characteristically have an inner endozone surrounded by an exozone. The endozone is occupied by the immature, thin-walled parts of zooids that are oriented parallel to the branch axis. Zooidal tubes bend through up to 90° into the exozone which is made up of the mature, thick-walled parts of the zooids. Polymorphic zooids (mesozooids and exilazooids) may become intercalated between the feeding zooids (autozooids) of the exozone, and rod-like acanthostyles often develop within the skeletal walls. All trepostomes appear to have possessed calcitic skeletons (Smith et al., 2006); suspected aragonitic trepostomes with ‘ recrystallized ’ walls were probably high-Mg calcite instead (Taylor and Wilson 1999). Hypercalcification refers to the formation of larger or thicker skeletons in carbonate-biomineralizing organisms. Ideally, levels of hypercalcification should be assessed by quantifying relative rates of carbonate skeleton formation, but this is not possible in most fossils where growth rates are unknown. Therefore, it is necessary to identify proxies for hypercalcification. In the case of trepostome bryozoans, colony size is unknown in most non-encrusting species because colonies are invariably broken into branch fragments during fossilization and there have been very few attempts to reconstruct colonies from fragments. Therefore, skeletal robustness must be used as a hypercalcification proxy. Skeletal wall thickness and branch diameter (Fig. 2) ade- quately quantify the robustness of trepostomes at the zooidal and colonial levels respectively. Data on the exozonal wall thicknesses and branch diameters of dendroid trepostomes were taken from taxonomic literature published between 1960 and the present day. Older references rarely provided the necessary measurements. A total of 14 papers were used for the Ordovician, 4 for the Devonian and 7 for the Permian, dealing with 107, 30 and 55 species respectively (Appendix). The same species was occasionally described in more than one paper, in which case duplicate records were culled on the basis of containing less precise information. While many of the references utilised describe North American species, efforts were made to minimise any palaeogeographical biases by including records from other parts of the world such as China, Europe and Australia. For each species, mean values of exozonal wall thickness and branch diameter given in the reference were noted. In cases where the mean values were not stated, the median was calculated from the observed range. Means, ranges and variances were calculated for the two parameters in the Ordovician, Devonian and Permian. Distributions of these values were tested for statistically significant differences between the three periods using analysis of variance (ANOV A) followed by log (x+1) data transformation to improve normality and homogeneity. The relationship between exozonal wall thickness and branch diameter was tested using linear regression for the full dataset. No significant correlation was found between wall thickness and branch diameter among the trepostome species used (Fig. 3), i.e. trepostomes with thick branches do not necessarily have thick skeletal walls. Therefore, these two metrics can be used independently as hypercalcification proxies. Figures 4 and 5 compare the frequency distributions of branch diameters and exozonal wall thicknesses respectively in trepostome species from the Ordovician, Devonian and Permian. For branch diameter, the Ordovician mean value is 5.88 mm, the Devonian 4.89 mm, and the Permian 5.72 mm (Table 1). Maximum values for the three periods are 24 mm, 10 mm and 24.1 mm respectively. ANOV A ( F 2,189 =0.42, p =0.655) shows no significant differences in branch diameters between the Ordovician, Devonian and Permian. For exozonal wall thickness, the Ordovician mean value is 0.07 mm, the Devonian is also 0.07 mm, and the Permian 0.10 mm. Maximum values for the three periods are 0.16 mm, 0.13 mm and 0.14 mm respectively. ANOV A ( F 2,189 =17.91, p <0.001) reveals significant differences in exozonal wall thicknesses between the data from the Ordovician, Devonian and Permian. These differences are due to the higher values of wall thickness in Permian trepostomes. There was no difference in wall thickness between species from the Ordovician and Devonian. Neither exozonal wall thickness nor branch diameter are significantly greater in trepostomes from the Ordovician than they are in species from the Permian, thus providing no support for the hypothesis of trepostome hypercalcification in the calcite seas of the Ordovician. In addition, these hypercalcification proxies do not show higher values in the Devonian, another period of calcite seas, than in the Permian when aragonite seas pertained. Either the two chosen proxies were inadequate as indicators of hypercalcification, or Ordovician trepostomes were not hypercalcified. Although geo- chemical ...