Palaeoecology of Devonian sclerobionts and their brachiopod hosts from the Western Canadian Sedimentary Basin

Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

Contents lists available at SciVerse ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology journal homepage: www.elsevier.com/locate/palaeo

Palaeoecology of Devonian sclerobionts and their brachiopod hosts from the Western Canadian Sedimentary Basin Kristina M. Barclay ⁎, Chris L. Schneider, Lindsey R. Leighton University of Alberta, Earth and Atmospheric Sciences Department, Edmonton, Alberta T6G 2E3, Canada

a r t i c l e

i n f o

Article history: Received 26 January 2013 Received in revised form 3 May 2013 Accepted 7 May 2013 Available online 15 May 2013 Keywords: Devonian Givetian Frasnian Palaeoecology Brachiopoda Skeletobionts Host Western Canadian Sedimentary Basin Assemblages

a b s t r a c t The majority of the studies on Devonian epibiosis are limited to those encrusting one, or only a limited number of host taxa. We present not only the first detailed study of sclerobionts and brachiopod hosts from the Devonian of the Western Canadian Sedimentary Basin, but we also examine entire assemblages of potential brachiopod hosts for sclerobionts. Over 1280 brachiopods were collected from the lower Firebag Member of the Waterways Formation (Late Givetian) in northeastern Alberta, and both the lower Hay River and upper Twin Falls Formations (Frasnian) of the southern Northwest Territories. The three units do not represent a continuous sequence, but represent three successive intervals in the history of the basin. Brachiopods were identified to genus level and examined for sclerobionts. Sclerobionts were identified to the lowest taxonomic level possible (usually genus). Data analyses were subdivided into three categories: (1) comparison of host assemblages, (2) comparison of sclerobiont assemblages, and (3) interactions between host and sclerobiont assemblages. Brachiopod assemblages increased in richness, evenness, and diversity across the three units, which corresponds to a decrease in terrigenous mud observed between the lithology of the lower Firebag and Hay River assemblages, and the younger Twin Falls assemblage. Desquamatia was the most abundant brachiopod collected from all three stratigraphic units. Sclerobiont assemblages experienced similar trends to the host assemblages, with an increase in evenness and diversity across each assemblage. Richness remained less varied between sclerobiont assemblages, with only one new taxon appearing in the latest stratigraphic unit (upper Twin Falls Formation). However, because sclerobionts and their hosts operate on such different spatial scales, changes experienced by the hosts may not have affected sclerobionts in the same manner and therefore should not be interpreted as such. Interactions between host and sclerobionts were examined for: (1) sclerobiont preference for host size, (2) sclerobiont preference for host taxa, and (3) sclerobiont preference for host valve. While there was increased encrustation of larger hosts in all assemblages, the brachiopod, Desquamatia, was most encrusted by sclerobionts, regardless of size, in the younger Twin Falls Formation. Preferences by sclerobionts for particular valves differ between the host assemblages, even among the same host taxa. These results suggest that (a) the inclusion of all possible brachiopod hosts from a given assemblage greatly improves the understanding of sclerobiont–host relationships, and (b) encrustation patterns, even on the same host taxa, do not always remain static between assemblages. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Encrusting organisms that attach to biotic hosts offer unique insights into organism interactions that would otherwise not normally be preserved in the fossil record. This is due to the in situ preservation of encrusting organisms on the hard surfaces of their hosts that may even record detailed interactions such as succession between different encrusting organisms on one host (Alvarez and Taylor, 1987; Taylor and Wilson, 2003; Rodland et al., 2004), synchronous growth ⁎ Corresponding author at: 1-26 Earth Science Building, University of Alberta, Edmonton, Alberta T6G 2E3, Canada. Tel.: +1 780 492 3983; fax: +1 780 492 2030. E-mail address: [email protected] (K.M. Barclay). 0031-0182/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.palaeo.2013.05.007

of encrusting organisms and their host (Alvarez and Taylor, 1987; Bose et al., 2011), and evidence that the host was alive or dead during the time of encrustation (Schneider, 2003; Taylor and Wilson, 2003; Bose et al., 2011). As well, encrusting organisms will often show apparent “preferences” for a particular substrate, or location on a substrate. This study focuses on the comparisons of three assemblages of Devonian brachiopods from the Western Canadian Sedimentary Basin and their interactions with encrusting organisms. Encrusting organisms are those that must attach to a hard or firm substrate, whether biotic or abiotic. Therefore, these organisms are found to occur more commonly on biotic hosts when in environments with greater accumulations of soft sediments on which the only hard

80

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

substrate is often the shells of larger organisms such as brachiopods (Taylor and Wilson, 2003; Rodland et al., 2006; Bose et al., 2011; Brett et al., 2011, 2012). In this study, we will follow the terminology proposed by Taylor and Wilson (2002), and shall refer to encrusting organisms that attach to, or bore into, any hard substrate (in this case, brachiopods) as “sclerobionts”. Biotic substrates, regardless of whether organisms were alive or dead at the time of encrustation, are referred to herein as “hosts”. Fossil sclerobiont–host relationships provide information on palaeoecological phenomena, such as sclerobiont preference of hosts, sclerobiont preference for location on a host, sclerobiont competition, and even anti-predatory benefits of sclerobionts to the host organism (Kesling et al., 1980; Alvarez and Taylor, 1987; Bordeaux and Brett, 1990; Lescinsky, 1996; Taylor and Wilson, 2003; Rodland et al., 2004, 2006; Bose et al., 2011). One of the most well studied sclerobiont–host systems is that between brachiopod hosts and their sclerobionts (e.g. Alvarez and Taylor, 1987; Bose et al., 2011). Most studies that examine sclerobiont communities involve either a single host taxon or single sclerobiont taxon (e.g. Alvarez and Taylor, 1987; Rodland et al., 2006; Bose et al., 2011; Zatoń and Krawczyński, 2011), yet examining an entire assemblage of hosts and sclerobionts may provide a more complete representation of the original fossil community (e.g. Rodland et al., 2004). By examining three entire assemblages of brachiopod hosts and their sclerobionts, we aim not only to provide a more accurate, quantitative representation of the patterns present in these original fossil communities, but also to compare these patterns between assemblages across the Late Givetian to Early Frasnian of the Western Canadian Sedimentary Basin (a region whose sclerobiont communities have never been examined thoroughly). Comparison of three entire assemblages of brachiopods will illuminate any changes in encrustation habits through time and also will test for the generality of encrustation patterns within one extensive basin. Sclerobionts on Devonian brachiopod hosts have been extensively studied, mostly due to the high rates of encrustation during this time period compared to the rest of the Palaeozoic Era (Alvarez and Taylor, 1987; Taylor and Wilson, 2003; Brett et al., 2012). Despite this large amount of literature on Devonian host–sclerobiont interactions, encrusting communities from the Devonian of the Western Canadian Sedimentary Basin have received little attention. The only previous works on sclerobionts from Western Canada are fossil lists that include the encrusting taxa Aulopora and Microconchus (listed as Spirorbis) from Givetian and Frasnian brachiopods from northeastern Alberta (Crickmay, 1957; Norris, 1963), and a pilot study of sclerobionts on brachiopod hosts (Schneider and Leighton, 2010). In contrast, brachiopods from the Givetian and Frasnian of Western Canada have been well documented (Norris, 1965; Johnson, 1974, 1975; Day, 1998; Day and Copper, 1998; Ma and Day, 2000). This study not only provides further documentation of sclerobiont–brachiopod assemblages from the Devonian of the Western Canadian Sedimentary Basin, but is also one of the few Devonian studies to examine complete assemblages of both hosts and sclerobionts. 2. Geology During the Late Givetian and Early Frasnian, Western Canada was part of a passive margin covered by a shallow sea, although an offshore, active island arc has been suggested (Moore, 1988). During the time interval studied, the Western Canadian Sedimentary Basin lay in the tropics (Witzke and Heckel, 1988) (Fig. 1). The climate was humid (Loranger, 1965; Witzke and Heckel, 1988), and the resulting rainfall, combined with the uplift and erosion of an orogen to the present-day northeast, most likely from the Ellesmerian Fold Belt (Stoakes et al., 1992; Wendte, 1992), or the Caledonian or Franklinian orogenic belts (Moore, 1988; Wendte and Uyeno, 2005), led to a significant amount of terrigenous sediment in the basin (Wendte and Uyeno, 2005). The influx of terrigenous mud into the basin from this erosion of orogens to

the northeast was sufficient to overprint the general rise in sea-level with the regressive, argillaceous sequences of the Waterways Formation (Schneider et al., in press). By the time of deposition of the youngest unit in this study, the Twin Falls Formation (Frasnian) (Fig. 2), the Western Canadian Sedimentary Basin had separated into distinct carbonate shelves with argillaceous infill of the deeper, central basin (Switzer et al., 1994). The oldest unit of interest in this study is the Firebag Member from the base of the Waterways Formation (Fig. 2), corresponding with transgressive–regressive (T–R) cycle IIb and the Skeletognathus norrisi conodont zone (Johnson et al., 1985; Day, 1998). The Firebag Member contains a lower shale and an upper shale separated by a middle argillaceous limestone. The middle limestone is an argillaceous mudstone to wackestone with few fossils. Both shales are generally not fossiliferous except for distinct horizons of fossils. Brachiopod samples for this study were collected from an approximately 20 cm-thick fossiliferous interval at the increasingly calcareous transition between the top of the lower shale and the base of the middle argillaceous limestone. Fossils were collected loose from a weathered surface. A second brachiopod assemblage was collected from member B of the Hay River Formation (Montagne Noire conodont zone 6) (Fig. 2), which marks the beginning of T–R cycle IIc (Day, 1998). Nomenclature and stratigraphic correlation of the Hay River Formation varies among authors (Jamieson, 1967; Hadley and Jones, 1990; MacNeil and Jones, 2006; Energy Resources Conservation Board, 2009). The present study follows the nomenclature and stratigraphy recognised in the Energy Resources Conservation Board (2009) Table of Formations for the Hay River Formation, and the members established by Jamieson (1967), and used in Day (1998), which splits the Hay River Formation into informal members A through F (Fig. 2). The Hay River Formation consists of approximately 400 m of shale with thin, laterally traceable limestone units (Williams, 1977). Fossils are usually associated with the limestone units. Some of the limestone beds, notably in a limestone in member C, are biostromal and contain corals and stromatoporoids. Member B is mostly shale that contains an argillaceous limestone bed, locally a floatstone to rudstone. Fossils were collected loose from a shale horizon immediately below the base of this member B limestone. The third brachiopod sample was collected from the upper member of the Twin Falls Formation (Fig. 2), which corresponds with T–R cycle IId-1 and the Montagne Noire conodont zone 9 (Day, 1998). T–R cycle IId marks a sea-level high for the Devonian (Johnson et al., 1985; Day, 1998). Nomenclature also varies between authors, with MacNeil and Jones (2006) splitting the lower Alexandra Member into a distinct formation. The present study uses the nomenclature of the Energy Resources Conservation Board (2009) Table of Formations for the Twin Falls Formation with the higher-resolution stratigraphy suggested by Day (1998), which separates the Twin Falls Formation into two members: the Alexandra Member in the lower portion and an informal upper member. The upper member is an open marine, non-argillaceous limestone. The fossils collected for this study had weathered loose from the surrounding gravel of spill piles (less than one metre thick). Near the Alberta–Northwest Territories border, the Waterways Formation is contiguous and thus, stratigraphically equivalent to the lower Hay River Formation, the latter extending further north into the Northwest Territories. In Alberta, the Waterways Formation is an argillaceous limestone and shale with occasional, non-argillaceous biostromal and carbonate sand facies (Loranger, 1965; Buschkuehle, 2003). The Hay River Formation in northern Alberta and the southern Northwest Territories is mostly shale, with a few laterally extensive limestone beds. Although the members of the Waterways and Hay River formations have not been correlated, the Hay River Formation contains equivalents of the Waterways through Ireton Formations (Fig. 2). Reefs from the Alexandra Member of the Twin Falls Formation of the southern Northwest Territories are equivalent to the reefs of the Grosmont Formation

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

81

Fig. 1. Palaeogeographic map of the Middle–Late Devonian of North America. The palaeoequator is indicated above Northern Canada by a solid black line. Land masses are represented by dark, shaded areas. The inset is a larger view of the study area (northern Alberta and the southern Northwest Territories. Sampled localities are indicated by stars. Modified from Fig. 1 of Day, 1998.

of northern Alberta (Fig. 2). The upper member of the Twin Falls Formation represents a return to non-reef, open marine conditions, but has little to no argillaceous sediment. Excepting biohermal intervals, brachiopods are the most common fossils encountered in the Firebag Member of the Waterways Formation, member B of the Hay River Formation, and the upper member of the Twin Falls Formation. Late Givetian and Frasnian units in North America yield abundant variatrypine atrypids (including Desquamatia [Fig. 3A]), as well as spinatrypides such as Spinatrypina (Day, 1998). Atrypides such as Pseudoatrypa and Spinatrypina are known from western Canada; their presence is the result of a radiation corresponding with the onset of transgression in T–R cycle IIb-1 (Day, 1998). Brachiopod assemblages (particularly the atrypide clade) from the three units have been studied extensively (Day, 1998; Day and Copper, 1998; Ma and Day, 2000), yet their sclerobiont communities have only been described in a pilot study (Schneider and Leighton, 2010). 3. Methods 3.1. Sample and data collection Over 1280 brachiopods were collected from the lower Firebag Member of the Waterways Formation in northeastern Alberta (599 specimens), member B of the Hay River Formation of the southern Northwest Territories (154 specimens), and the upper member of the Twin Falls Formation of the southern Northwest Territories (530 specimens) (Figs. 1, 2, and 3, Table 1). Assemblages of brachiopods collected from the units will be referred to as the Firebag, Hay River, and Twin Falls assemblages herein. All brachiopod samples were taken from thin horizons in their respective units. The sampled beds were chosen specifically due to the large number and excellent preservation of brachiopod specimens

required for a detailed sclerobiosis study. There may have been a sampling bias towards non-fragmented brachiopod specimens, but given the goals of this sclerobiosis study, the results should not have been affected. Preliminary inspection of the brachiopod samples revealed that all three assemblages were dominated by the brachiopod genus, Desquamatia (Fig. 3A), thereby potentially facilitating a comparison of encrustation among the relatively similar host assemblages. The samples included in this study may or may not be reflective of each of the three units as a whole; the sampled intervals are but a fraction of each respective stratigraphic unit. However, the goal of the study is not to characterize encrustation patterns of entire units at the member or formation scale, but rather, to present the first whole and multiple assemblage sclerobiosis analysis from the Western Canadian Sedimentary Basin. Specimens were generally well preserved, based on the detail of external ornament, and low levels of abrasion (Fig. 3, Table 1). Brachiopod shells were rarely compacted or crushed. Indeed, choice of the three brachiopod assemblages for inclusion in the present study was partly due to preservation of sufficient quality to examine sclerobionts. Of the total specimens collected, poorly preserved shells, and those that contained less that 50% of the original shell were discarded from the study. After culling poor specimens, 1111 remained for sclerobiont and host analyses. Each brachiopod specimen used in the study was identified to genus and measured to the closest millimetre for anterior–posterior length (the distance between the umbo and the commissure), hinge width (the length of the line-segment from one end of the hinge line to the other), and maximum width (the widest point on the brachiopod that was parallel to the hinge width, which was the same as hinge width for some taxa). An additional measurement called the “maximum surface area” was calculated by multiplying the length by the maximum width of the brachiopod. The “maximum surface area” calculation gives a rough approximation of the surface area available to encrusting sclerobionts on each brachiopod, assuming that one valve is resting

82

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

Fig. 2. Stratigraphic correlation of the Middle–Late Devonian of northeastern Alberta and the southern Northwest Territories. The three units of study are indicated by darker shading. Modified from Fig. 2 of Day, 1998; and Schneider et al., in press.

against the substrate, and was also used as a proxy for the size of the brachiopod host in all statistical analyses. Only specimens on which both the length and maximum width could be measured were included in

statistical analyses involving host size. Articulation percentages of brachiopod valves were above 92% for all three assemblages, with the greatest articulation percentage in the Twin Falls assemblage (Table 1). The great percentage of articulation observed in brachiopod hosts suggests very good preservation (as per Alexander and Gibson, 1993), and thus, the samples collected are appropriate for a detailed study of sclerobiont–host assemblages. Schizophoria (Fig. 3C) from the Firebag assemblage, which had an articulation percentage of 80%, was the worst of the study taxa in this regard, but 80% is actually a large value for this genus. Schizophoria, like other orthides, were likely more easily disarticulated because of the nature of their deltidiodont hinge teeth, which are prone to disarticulation after death compared to the teeth of most other Devonian brachiopods (Alexander and Gibson, 1993). When Schizophoria was removed from the articulation percentage calculations, articulation rates of the Firebag assemblage increased to 95% (550 of 581 brachiopods were articulated), which was comparable to the Twin Falls assemblage. The strophomenide taxa in this study, which generally entirely lack hinge teeth and usually only possess minute denticles, were also normally articulated. Each brachiopod was examined under a microscope (10–40 × power) for sclerobionts. Sclerobionts in this study were identified to genus (if possible), and otherwise identified to the lowest taxonomic group possible (Fig. 4). For example, in order to identify trepostome bryozoans (Fig. 4E) to lower taxonomic rankings, thin-sections of the colonies are required. This would have been logistically challenging due to the small size of most colonies as well as requiring destruction of the brachiopod specimens. Even thin sections of encrusting bryozoans often lack diagnostic features of the upright, erect bryozoans colonies. Therefore, a higher taxonomic rank was deemed appropriate for encrusting trepostome bryozoans. It is also important to note that the solitary sclerobiont genus, Microconchus, appears as the genus, Spirorbis, in many older papers (e.g. Alvarez and Taylor, 1987; Gibson, 1992) (see Fig. 4A). Fossil sclerobionts frequently display “preferences” for host taxa, host size, host valve, and particular locations on host valves (Ager, 1961; Hoare and Steller, 1967; Hurst, 1974; Thayer, 1974; Pitrat and Rogers, 1978; Kesling et al., 1980; Alvarez and Taylor, 1987; Alexander and Scharpf, 1990; Bordeaux and Brett, 1990; Gibson, 1992; Lescinsky, 1997; Bose et al., 2011; Zatoń and Krawczyński,

Fig. 3. Common brachiopod taxa from the Firebag Member of the Waterways Formation, member B of the Hay River Formation, and the upper member of the Twin Falls Formation. (A) Desquamatia specimen from member B, Hay River Formation; (B) Radiatrypa specimen from the Firebag Member; (C) Schizophoria specimen from the Firebag Member; (D) Cyrtospirifer specimen from member B, Hay River Formation; (E) Stropheodonta specimen from the Twin Falls Formation; and (F) Regelia specimen from member B, Hay River Formation.

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

83

Table 1 Summary data of the brachiopods collected from the Firebag, Hay River, and Twin Falls assemblages. The values presented are raw numbers and percentages based on the actual number of specimens. Assemblage

Total specimens

Articulated brachiopods

Isolated valves

% Articulated brachiopods

Encrusted brachiopods

Unencrusted brachiopods

% Encrusted brachiopods

Firebag Member Member B Twin Falls Formation

565 120 426

525 111 402

40 9 24

93.0% 92.5% 94.4%

118 60 189

447 60 237

20.9% 50.0% 44.4%

2011). The term “preference” will be used frequently in this paper, and references to any bias for location or substrate by encrusting organisms, whether it represents a preferred settlement site actually chosen by the larvae, or whether the larvae are preferentially found due to other external factors, such as hydrodynamics of the surrounding water, grazing, predation, mechanical abrasion, or antifouling defenses, will be regarded as a “preference”. To examine patterns of encrustation within and between host assemblages, the sclerobionts attached to each brachiopod specimen were tabulated and classified as solitary or colonial. During abundance counts, solitary and colonial sclerobionts were treated equally. For example, each individual Microconchus (Fig. 4A) was given a count of ‘one’, and each distinct colony of Hederella (Fig. 4D) was also counted as ‘one’. “Preferred” host taxa would have an overall higher abundance of sclerobionts, as well as having a greater proportion of encrusted specimens. To study the coverage of the host shell by a colonial sclerobiont, the area for colonial taxa was approximated by multiplying the lengths of two roughly perpendicular lines across the colony (maximum length by width) which served as a proxy for the actual area of brachiopod shell covered by a sclerobiont colony. Solitary taxa were rarely substantially larger or smaller than 1 mm2 in this study, and so all solitary taxa were assigned an arbitrary surface area of 1 mm2 per individual for all analyses involving sclerobiont area. Measurements were

taken for those rare solitary sclerobionts (two occurrences) that were obviously larger than 1 mm 2 (see Fig. 4C). Area measurements for the sclerobiont taxon, Ascodictyon, were conservative. It was often impossible to generate accurate area measurements due to the unique preservation and fragility of this genus. For example, on many specimens, it was apparent that partial sclerobiont remains similar to Ascodictyon were present on the shell, but they had been subsequently eroded and were missing in some places. In other cases, skeletal remains of Ascodictyon were found to be sparsely or erratically covering the entire shell. Where possible, area measurements were taken, but if areas of the specimens were impossible to discern, Ascodictyon remains were conservatively assigned an area of 1 mm2. Preservation of Ascodictyon was poor in all assemblages, particularly in the Twin Falls assemblage. It is generally assumed that Ascodictyon would have had a greater “biomass” representation in life than what is preserved in the fossil record and therefore presented in these results. Measuring the area of the host shell covered by the sclerobiont provides an approximation of the percentage of calcified biomass each sclerobiont taxon contributes to the total sclerobiont assemblage. The host valve to which each sclerobiont was attached was also recorded. If sclerobionts exhibited valve “preferences,” then one valve of host taxa would have higher encrustation than the other. The use of the term “preference” does not necessarily imply active

Fig. 4. Common sclerobiont taxa on brachiopod hosts from the Middle to Late Devonian of the Western Canadian Sedimentary Basin. (A) Microconchus specimen; (B) Cornulites specimen; (C) rugose coral specimen; (D) Hederella specimen; and (E) trepostome bryozoan specimen.

84

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

indicated a linear curve, whereas greater values indicate more peaked and therefore a less convex curve, while negative kurtosis is indicative of more convex curves. Percent abundances of encrusting taxa were also calculated as an additional means of comparing encrustation patterns among the three brachiopod assemblages (Fig. 7). Sclerobiont “preferences” were analyzed for (1) host size, (2) host taxon, and (3) host valve. Analyses were conducted for each of the three assemblages, and the results were compared among assemblages. To test for sclerobiont preference of host size, two-tailed t-tests assuming unequal variances compared the size of encrusted versus unencrusted hosts (based on the maximum surface area measurements of the brachiopods) (Table 3). To examine sclerobiont preference for host taxa, 2 × 2 chi-square tests were conducted. One 2 × 2 chi-square test was performed to compare the number of sclerobionts encrusting the brachiopod Desquamatia (Fig. 3A) versus all other hosts (columns = Desquamatia vs. other hosts, rows = encrusted vs. unencrusted hosts) (Table 4). In all three assemblages, Desquamatia represented a significant portion of the brachiopod assemblage and generally was larger than the other brachiopod host taxa. By testing Desquamatia against all of the other host taxa, it could be determined if Desquamatia influenced the trends observed in the data set, and if sclerobionts preferred Desquamatia over other host taxa more than would be expected based on the abundance of Desquamatia. In a second 2 × 2 chi-square test, each sclerobiont taxon was compared against all the other sclerobiont taxa to test for the preference or lack of preference of each sclerobiont taxon for Desquamatia hosts. As adult Desquamatia is one of the largest brachiopods present in all three assemblages, and the only common large brachiopod, a further 2 × 2 chi-square test on large and small Desquamatia was conducted to test whether or not sclerobiont preferences were being driven by a preference for Desquamatia itself, or if the preference was merely for the size of the host. Brachiopod taxa were also divided into large atrypides (brachiopod taxa that had an average maximum surface area of more than 470 mm2), and all other taxa. A 2 × 2 chi-square test was performed on this data set to compare the number of sclerobionts encrusting the large atrypides (including Desquamatia and Variatrypa) versus “all other taxa” (columns = large atrypides vs. other hosts, rows = encrusted vs. unencrusted hosts) (Table 4). This analysis of “morphotaxa” was performed to control for the large portion of large atrypides in each of the three assemblages, as well as account for the greater size of brachiopods in the large atrypide “morphotaxon”; other taxa generally did not exceed an average maximum surface area of 470 mm2. One exception was the orthide Schizophoria (Fig. 3C), but this genus only occurred 20 times in the Firebag assemblage. Sclerobiont preference for host valves of Desquamatia was tested using a 2 × 2 chi-square test that compared the number of sclerobionts present on the ventral valve to the number present on dorsal valves (columns = number of encrusted valves observed vs.

choice on the part of sclerobionts, so much as indicate their ecological distributions. Valve “preferences” may differ between host taxa, between sclerobiont taxa, and between localities. 3.2. Analyses There were 1111 brachiopods that were sufficiently well-preserved to calculate the maximum surface area and were therefore included in statistical analyses (Table 1). For host communities, analyses were broken into two data sets: (1) analyses that included all brachiopods, and (2) analyses that included only those individual brachiopods that were encrusted. These analyses were performed to reveal differences between the entire assemblage of potential host brachiopods, and those individual hosts that were encrusted. Differences between the two data sets could indicate preferences by sclerobionts for particular hosts or host conditions. For both data sets, diversity metrics (richness, Shannon's index, Margalev's diversity, and Buzas and Gibsons' evenness) were calculated for each assemblage (Table 2). Rank-abundance curves (RACs) were generated for each of the three brachiopod assemblages, following the methodology proposed by Webb et al. (2009) (Fig. 5). Rank-abundance curves measure the diversity structure of the assemblages and serve as a means of indicating similarities or differences between the assemblages (Webb et al., 2009). Separate RACs were generated for all brachiopods and for only encrusted brachiopods. Diversity and rank-abundance measures were also calculated for all sclerobionts from each host assemblage in order to determine whether sclerobionts from each assemblage tracked the structure of the related host assemblage (Fig. 5, Table 2). Rank-abundance curves for sclerobionts were generated twice, once using sclerobiont abundances, and once using surface area of sclerobionts (Fig. 6). Abundance counts were used as a conservative measure of the actual settlement of larvae, or number of sclerobionts present on a host. Rank-abundance curves generated using surface area of sclerobionts accounted for colonial sclerobionts that were able to cover a larger portion of host shells and therefore would have had a greater “biomass”. Therefore, RACs generated using sclerobiont surface area could be different from RACs generated from sclerobiont abundance. Rank-abundance curves have been used previously to estimate the utilization of resources among competing taxa within an ecological guild (Bazzaz, 1975; Webb et al., 2009; Webb and Leighton, 2011). For example, sclerobionts could be competing for space on the host. Rank-abundance curves also have been used as an indicator of stress (Gray, 1981; Webb et al., 2009; Webb and Leighton, 2011) because convex-down curves (i.e., when the relative abundance of taxa is dominated by a single abundant taxon) suggest systems that are experiencing greater, or possibly more recent, rates of disturbance or stress, whereas convex-up curves (i.e., when several taxa are roughly equal in dominance) indicate environmentally stable systems. The kurtosis of each RAC was calculated to quantify convexity of the curve (Webb et al., 2009) (Figs. 5 and 6). Kurtosis values of zero

Table 2 Diversity metrics for all three brachiopod assemblages. Genus richness, Shannon's indices, Buzas and Gibsons' evenness, and Margalev's diversity indices were calculated for (1) all hosts, (2) only brachiopods that were encrusted, (3) sclerobiont assemblages based on the area (i.e. biomass) of each sclerobiont, and (4) sclerobiont assemblages based on abundance of individual entities only. Assemblage

Genus richness

Shannon's index

Buzas and Gibsons' evenness

Margalev's diversity

Firebag Member: All hosts Firebag Member: Encrusted hosts Firebag Member: Sclerobiont abundance Firebag Member: Sclerobiont area Member B: All hosts Member B: Encrusted hosts Member B: Sclerobiont abundance Member B: Sclerobiont area Twin Falls Formation: All hosts Twin Falls Formation: Encrusted hosts Twin Falls Formation: Sclerobiont abundance Twin Falls Formation: Sclerobiont area

6 6 6 6 9 7 6 6 12 10 7 7

0.85 0.80 1.21 1.05 1.45 1.51 1.55 1.49 2.01 1.87 1.48 1.34

0.39 0.37 0.56 0.48 0.47 0.84 0.78 0.74 0.62 0.65 0.63 0.55

0.78 1.05 0.89 0.82 1.59 1.39 1.08 0.99 1.71 1.59 1.16 0.95

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

85

Fig. 5. Rank-abundance curves with kurtosis values for each of the three host brachiopod assemblages, both those that include all brachiopod taxa and only those taxa that were encrusted. Desquamatia is the most abundant host taxa in all cases.

expected (based on the assumption that there should be an equal probability for each valve to have been encrusted), rows = ventral vs. dorsal valves) (Table 5). Only Desquamatia was used to test for valve preference to avoid any potential differences in life modes of the various brachiopod taxa. Articulated brachiopods were used to ensure that a preservational bias for one valve would not affect the results of the test. All column entries in each 2 × 2 chi-square test were standardized to the summation of the smaller column, so that the sums of each column were equal, regardless of the number of actual specimens; this approach is conservative. Suitable statistical tests do not exist to compare the summary results from the three brachiopod assemblages as there are only three data-points, but parallel tests performed on all three assemblages allowed the qualitative assessment of potential differences between the structure of the sclerobiont and host portions of each assemblage. Similarities or differences in the patterns of sclerobiont and hosts could be indicators of environmental similarities or differences between the Firebag, Hay River, and Twin Falls assemblages.

communities increased from the Firebag assemblage (richness = 6) to the Hay River assemblage (richness = 9), to the Twin Falls assemblage (richness = 12). However, all brachiopod taxa were encrusted in the Firebag assemblage, while only 7 taxa were encrusted in the Hay River assemblage and 10 taxa in the Twin Falls assemblage (Table 2). Diversity and evenness also increased in a manner similar to richness across the three assemblages of brachiopods (Table 2), with the smallest diversity and evenness in the Firebag assemblage and the greatest in the Twin Falls assemblage (Table 2). When examining only those brachiopods that were encrusted, the changes across the three assemblages were less marked, with more subtle increases in evenness and only a 0.01 increase in the evenness between the Hay River and Twin Falls assemblages (Table 2). Rank-abundance curves and kurtosis values also indicate an increase in evenness from the oldest to youngest assemblages (Fig. 5). The relative ranks in abundance of genera changed slightly when examining only those brachiopods that were encrusted, but Desquamatia was always the most abundant genus.

4. Results 4.2. Sclerobiont assemblages The proportion of encrusted brachiopods exceeded 20% for each assemblage, with the highest proportion of encrustation of host shells seen in the Hay River assemblage (50%), followed by the Twin Falls assemblage (44%), and the Firebag assemblage (21%) (Table 1). 4.1. Host assemblages Desquamatia was the most abundant brachiopod in all six data sets (Figs. 3A and 5). Richness (number of genera) of the brachiopod

Analyses were conducted for both sclerobiont abundances and the area of the host shell covered by sclerobionts. The area results were used as an approximation for the calcified biomass of the colonial taxa (bryozoans [Fig. 4E] and Hederella [Fig. 4D]), and produced slightly different results than those calculated using abundances (Figs. 6 and 7, Table 2). Trends between abundance and area measurements were similar (Fig. 6, Table 2), but larger differences existed among the RACs and kurtosis values for sclerobiont abundance versus area.

86

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

Fig. 6. Rank-abundance curves with kurtosis values of sclerobiont taxa for all three brachiopod assemblages. Rank-abundance curves were based on individual sclerobiont abundance (top row), and sclerobiont area (“biomass”).

Taxon richness of sclerobionts was nearly equal for all three assemblages of brachiopod hosts, with the same five major sclerobiont taxa appearing in each assemblage (trepostome bryozoans, Hederella, Microconchus, Cornulites, and Ascodictyon), and only very rare occurrences

of any other encrusting taxa (Figs. 4 and 7). Diversity and evenness of the sclerobionts were also similar across the three assemblages, unlike those of the host brachiopod taxa (Table 2). This was especially true of the diversity for the Hay River and Twin Falls assemblages, which

Fig. 7. Percent abundances for each of the major sclerobiont taxa in the three brachiopod assemblages. Percent abundances were separated based on individual sclerobiont abundance and sclerobiont area (“biomass”) in order to account for larger colonial sclerobionts that were able to cover larger areas of brachiopod hosts.

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91 Table 3 Two-tailed t-test results for each of the three brachiopod assemblages comparing the maximum surfaces area (length × maximum width) of encrusted and unencrusted brachiopods. Mean size of brachiopods is given in millimetres squared. Assemblage

p-value

Mean unencrusted

Mean encrusted

Variance unencrusted

Firebag Member ≪0.01 Member B 0.58 Twin Falls 0.03 Formation

410.70 mm2 420.28 mm2 331.68 mm2

582.57 mm2 69408.34 445.86 mm2 48042.42 380.70 mm2 51261.41

Variance encrusted 51382.75 84597.31 52984.86

overlapped in results for both abundance and area measurements (Table 2). Evenness was greatest in the Hay River assemblage with the Twin Falls assemblages only slightly less (Table 2). The evenness of the sclerobionts was greater than the evenness of the brachiopods within each assemblage (Table 2). The RACs and kurtosis values for the sclerobiont assemblages varied between the abundance and area calculations, but generally followed the same, but somewhat less extreme, trends as the brachiopod RACs (Fig. 6). Unlike the brachiopod taxa in each of the three datasets, no single sclerobiont taxon was highly abundant compared to the rest, and the rank-abundance order of the five major sclerobiont taxa differed for each assemblage (Fig. 6). In the area RAC, Hederella (Fig. 4D) was the most abundant taxon across all of the assemblages, but was the first ranked in the abundance RAC only for the Firebag assemblage (Fig. 6). The proportion of Hederella in each sclerobiont assemblage decreased in both abundance and biomass (areal coverage) across the three assemblages (Fig. 7), even though it was still usually within the top two ranks in all RACs (Fig. 6). Bryozoan sclerobionts increased in both abundance and area proportions progressively through the three assemblages (Figs. 4E and 7). The increase in bryozoans was also very apparent from the area RAC for the Twin Falls assemblage and greatly affected the shape of the RAC (Fig. 6). 4.3. Sclerobiont and host interactions Encrusted brachiopods were larger than unencrusted brachiopods in the Firebag and Twin Falls assemblages (p ≪ 0.01, and p = 0.03, respectively), but the p-value was not significant for the Hay River assemblage (Table 3). Desquamatia (Fig. 3A) was analyzed separately from other brachiopod host taxa because this large atrypide comprised a large, though decreasing, proportion of each sample (Firebag assemblage = 77%, Hay River assemblage = 48%, Twin Falls assemblage = 27%). When comparing the number of encrusted Desquamatia to all other brachiopod taxa, the percentage of Desquamatia that were encrusted increased from the oldest Firebag to the youngest Twin Falls assemblages (Table 4). While the 2 × 2 chi-square test for encrusted Desquamatia versus all other host taxa was not significant for the Firebag assemblage, it was weakly significant for the Hay River assemblage (p = 0.05), and strongly significant for the Twin Falls assemblage (p b 0.01) (Table 4). Table 4 p-values of 2 × 2 chi-square tests comparing the proportions of encrusted brachiopod taxa. Analyses are separated into (1) the number of encrusted Desquamatia compared to all other encrusted taxa, (2) the number of encrusted large atrypides compared to all other encrusted “morphotaxa”, and (3) the number of small encrusted Desquamatia compared to large encrusted Desquamatia. Assemblage

p-value p-value Desquamatia p-value small vs. large vs. other hosts large atrypides vs. other “Morphotaxa” Desquamatia

Firebag Member 0.65 Member B 0.05 Twin Falls Formation b0.01

0.65 0.14 ≪0.01

≪0.01 0.13 0.80

87

Table 5 p-values of 2 × 2 chi-square tests comparing the number of encrusted ventral and dorsal valves from each assemblage. Assemblage

p-value dorsal vs. ventral

Firebag Member Member B Twin Falls Formation

≪0.01 (dorsal) 0.43 0.25

In the 2 × 2 chi-square test of the morphotaxon of large atrypides (which included Desquamatia), the results for both the Firebag and Hay River assemblages were not significant (the Firebag assemblage had no large atrypides other than Desquamatia, so the results remained the same), but that of the Twin Falls assemblage had a smaller p-value than that of only Desquamatia versus all other brachiopods (p ≪ 0.01) (Table 4). Sclerobionts did not consistently encrust one or the other valve of the host taxa (Table 5). While there was a strong preference for the dorsal valve in the Firebag Member (p ≪ 0.01), the results of the other two assemblages were not significant (Table 5). 5. Discussion 5.1. Host assemblages Overall, the observed trends were similar for analyses of encrusted taxa and all hosts. Although differences between the three assemblages of encrusted hosts are less than those observed for all brachiopods in each assemblage, there is still a clear increase in species richness, evenness, and diversity of brachiopod hosts through time in both data sets. These results may stem from the “preference” for larger, more common hosts by sclerobionts. Therefore, analyses including only encrusted hosts would tend to exclude the extremely small hosts. Smaller brachiopods may have been much less likely to be encrusted simply because such brachiopods presented a much smaller target on which sclerobiont larvae could land (Rodland et al., 2004, 2006). With the smaller, unencrusted hosts not included in the encrusted host analyses, richness and evenness were not as varied. The Twin Falls assemblage had increased richness from the addition of several new, smaller or highly ornamented taxa of brachiopod hosts, such as Spinatrypina. Although not all host taxa were encrusted, sclerobionts generally took advantage of most potential host taxa. Evenness was also greater and less varied between the three assemblages when including only encrusted hosts, due to the elimination of the very small potential hosts, or brachiopod taxa that were avoided by sclerobionts. Encrusted hosts from the Hay River and Twin Falls assemblages had very similar evenness. From the older Firebag and Hay River assemblages to the youngest Twin Falls Formation, there is a decrease in the proportion of terrigenous mud in the carbonate beds associated with the fossil assemblages. The Firebag and Hay River assemblages were both collected from similar environments at the base of argillaceous limestones, whereas the Twin Falls assemblage represents a cleaner, open marine environment free from most argillaceous sediment input. The differences in diversity metrics calculated from each assemblage may be a reflection of the amount of argillaceous sediment input. Brachiopods in the Firebag and Hay River assemblages were living during a period of substantial influx of terrigenous mud, which would have potentially disturbed suspension-feeding brachiopods. The Firebag and Hay River assemblages represent environments that had recently and repeatedly received a large amount of sediment input, which would have deterred many brachiopod taxa from inhabiting the environment. Particularly, the diversity metrics and extremely convex-down RAC in the Firebag assemblage are probably reflective of the great amount of sediment deposited in the lower Firebag Member. In the Twin Falls assemblage,

88

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

there is a significant decrease in the amount of argillaceous sediment, suggesting a more stable environment with fewer disturbances (e.g., fewer sediment pulses). Diversity metrics and RACs for the Twin Falls assemblage are consistent with the hypothesis of a more stable environment that received little to no terrigenous sediment influx. The general composition of these three brachiopod assemblages is quite similar; the less diverse Firebag assemblage is largely a taxonomic subset of the Hay River assemblage and both are subsets of the Twin Falls assemblage (Fig. 3). For example, the Twin Falls has several additional taxa, but the same taxa are also present in the Firebag assemblage. Many authors have suggested that the composition of brachiopod communities is reflective of the depth along the continental shelf (Zeigler, 1965; Johnson, 1974; De Keyser, 1977; Brett et al., 1993; Patzkowsky and Holland, 1999). The strongly nested overlap of taxa in the present study suggests broadly comparable water depth for the three assemblages of brachiopods in the present study. Furthermore, all three assemblages come from a shallow marine setting above storm wave base and below fair-weather wave base: the Firebag from a shallow marine portion of an argillaceous ramp (Schneider et al., in press), the Hay River from a ramp formed by a thick accumulation of terriginous mud, and the Twin Falls from a carbonate platform. Therefore, a decrease in the amount of terriginous sediment input may influence the increase in brachiopod species richness, evenness, and diversity observed between the Firebag and Hay River assemblages, and the Twin Falls assemblage. Past research indicates a potential trend in climate that may be consistent with the patterns in the brachiopod assemblages of the current study. A rise in magnetic-susceptibility values, correlated to depositional changes indicating sea level rise, suggests a warming trend throughout the Frasnian (Whalen and Day, 2010). Oxygen and carbon isotope analyses also indicate warming throughout the Frasnian until the Early Famennian (Joachimski et al., 2004; van Geldern et al., 2006; Simon et al., 2007). Magnetic-susceptibility values and oxygen and carbon isotope analyses correlate with transgression–regression models for the Givetian and Frasnian. Eustatic sea levels reached an all-time high in the Devonian during T–R cycle IId (Johnson et al., 1985), coinciding with the deposition of the upper member of the Twin Falls Formation (Day, 1998). Increases in both temperature and sea level are known to be associated with increases in biodiversity throughout the fossil record (Sepkoski, 1976; Cobianchi and Picotti, 2001; Ruban, 2007; Kröger and Landing, 2010). Changes in temperature and sea level have specifically been well documented for brachiopod communities in the Middle to Late Devonian in North America, and have often been attributed to species dispersal across North American basins (Day, 1996, 1998; Stigall Rode and Lieberman, 2005; Dominici and Zuschin, 2007). North American basins that were normally separated by tectonic structures during periods of lower sea level received an increase in the regional species pool during times of higher sea level (Day, 1996, 1998; Stigall Rode and Lieberman, 2005; Dominici and Zuschin, 2007). The increasing diversity among the three successive brachiopod assemblages, which only add taxa to the original subset, could be reflective of increases in the regional species pool in western North American basins, associated with basin connections at the time. The RACs and kurtosis values also provide a strong representation of the trend between the three assemblages. Rank-abundance curves may serve to indicate how much time has passed between community disturbances (Bazzaz, 1975; Webb et al., 2009; Webb and Leighton, 2011). Sharply convex-down curves may be indicative of an assemblage that has recently been under severe stress (Gray, 1981; Webb et al., 2009; Webb and Leighton, 2011), such as the disturbances caused by sediment influx in the Firebag and Hay River assemblages. When these assemblages are viewed as temporal “snapshots” rather than one continuous trend, it is important to consider that they represent assemblages that may have been deposited at different stages

of development. For example, the Twin Falls assemblage may not only represent a “cleaner” environment, but it may also be from a well established and/or stable environment that had time to accumulate shell material. The RAC may be artificially less peaked due to potential accumulation of shell material and time-averaging.

5.2. Sclerobiont assemblages The encrusting taxa present in each of the three sclerobiont assemblages are generally consistent with those typical of the Middle– Late Devonian (Alvarez and Taylor, 1987; Bordeaux and Brett, 1990; Webb et al., 2009; Bose et al., 2011; Zatoń and Krawczyński, 2011; Brett et al., 2012) (Fig. 4). Other than variations observed in richness and evenness of the brachiopod assemblages, the composition of the brachiopod assemblages consists of similar taxa, all dominated by Desquamatia. Due to this similarity between host assemblages, it is assumed that sclerobiont assemblages would be accordingly similar between the three assemblages. As well, the three assemblages generally consist of the same sclerobionts, and are also very similar to a particular “sclerobiofacies” documented in Brett et al. (2012), typical of high sedimentation rates and shallow waters that are observed in the Middle Devonian of New York. Overall “biomass” proxies are also fairly consistent between the assemblages. Although there is a trend of increased encrustation rates and sclerobiont evenness, it is important to note that the sample size of the Hay River assemblage was much smaller than that of both the Twin Falls and Firebag assemblages. The smaller sample size of the Hay River assemblage affected the strength of the results obtained (Margalev's diversity, which controls for sample size, is very similar for sclerobionts in each assemblage [Table 2]). Each individual brachiopod host effectively forms a substrate for an entire community of potentially interacting sclerobionts. Within the three stratigraphically distinct assemblages of brachiopod hosts, there are as many communities of sclerobionts as there are individual hosts in each assemblage. Evenness and diversity could be moderated simply by the large sample size of individual sclerobiont communities (i.e. each brachiopod host). Space was not limiting to sclerobiont growth, as there was never an instance where sclerobionts were growing directly beside one another. Furthermore, colonial and solitary sclerobionts may have avoided direct competition, simply by exploiting different means of growth and reproduction, allowing for greater differences in resource acquisition and partitioning. Proportions of Hederella (Fig. 4D) in the sclerobiont assemblages decrease through time with a concurrent increase in bryozoans (Fig. 7). Brett et al. (2012) observe a similar trend, or change in “sclerobiofacies”, associated with a change from muddy to “cleaner” substrates, which would support the results and interpretations of a decrease in terriginous mud observed across the three assemblages. Taylor and Wilson (2003) observed that hederellids and microconchids are more common sclerobionts in the Devonian than bryozoans, but perhaps the shift towards increasing dominance of bryozoans occurs locally during the Frasnian. By the Carboniferous, trepostome bryozoans are one of the most abundant sclerobionts (Lescinsky, 1997), with hederelloids becoming more rare (Taylor and Wilson, 2008). Qualitative observations indicate that the Hay River assemblage had the most well preserved host specimens and sclerobiont assemblages, which may have accounted for the increased encrustation and evenness of sclerobiont assemblages. Preservation may also play a large role in the observed evenness and diversity of each assemblage. For example, due to its fragility, the encrusting taxon, Ascodictyon, is the least likely to be preserved of all sclerobionts encountered in this study. It is most abundant in the Hay River assemblage, and only rarely noted in the Twin Falls assemblage, which appeared to have the worst preservation of the three assemblages based on lack of some external ornamental detail on both brachiopods and sclerobionts. Preservation is therefore

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

an important consideration when examining sclerobiont assemblages, particularly for comparisons between localities.

5.3. Sclerobiont and host interactions Sclerobiont assemblages have often been the subject of palaeoecological studies due to the preservation of spatial relationships on potentially live hosts (Alvarez and Taylor, 1987; Rodland et al., 2004; Bose et al., 2011). By examining sclerobionts present on three entire assemblages of brachiopods, the present study explored the relationships between sclerobionts and their brachiopod hosts, and how such relationships may change between assemblages. There have been many studies in the past that have looked at sclerobiont preference for a particular host taxon (e.g. Thayer, 1974; Bordeaux and Brett, 1990; Bose et al., 2011). Prior research suggests that spiriferides are often the preferred taxon for encrustation with atrypides less so, even when atrypides are more abundant (Ager, 1961; Hoare and Steller, 1967; Thayer, 1974). Although spiriferides do not make up a very large proportion of the brachiopod taxa present in any of the assemblages, individual encrustation abundances for each brachiopod taxon do not demonstrate any significant preference for spiriferides. Desquamatia was the most encrusted brachiopod and was preferred by all sclerobionts, except a few instances of an encrusting rugosan coral, which preferred strophomenides. However, the relative proportion of Desquamatia is greatest in the Firebag and smallest in the Twin Falls assemblage. “Preference” for large atrypides became even more evident when considering the “morphotaxon”. Despite the marked “preference” for Desquamatia as a host, it is important to note that Desquamatia was also the largest of all potential hosts. Sclerobiont occurrence on different sizes of brachiopod hosts (surface area) has been examined extensively in the past (e.g. Thayer, 1974; Alexander and Scharpf, 1990; Rodland et al., 2004; Bose et al., 2011). Not only do larger hosts present a larger target to sclerobiont larvae during settlement, but larger hosts of any given taxon are also older and have therefore had more time to accumulate sclerobionts. While taphonomic removal and degradation of older sclerobionts may counteract the longevity of a larger host (see Brett and Bordeaux, 1991), some authors argue that taphonomic erasure is negligible (Rodland et al., 2006). Sclerobionts in the Firebag assemblage are significantly more abundant on large Desquamatia, but there were almost as many sclerobionts on small Desquamatia as there were on large Desquamatia in the Twin Falls assemblage (Table 4). The significant preference for Desquamatia, regardless of size, by sclerobionts in the Twin Falls assemblage, suggests that there is a true biological preference for Desquamatia that develops in the Twin Falls assemblage as opposed to a strong preference only for size, which is demonstrated in the Firebag assemblage. Desquamatia is not only a large host that would present a big target for sclerobiont larvae, but it may be that sclerobionts prefer particular types of ornamentation (Gibson, 1992; Taylor and Wilson, 2003). For example, Spinatrypina, which was not present in the Firebag assemblage, is a small to medium-sized brachiopod with closely spaced spines that form frill-like ornamentation on both valves (Copper, 1967). Even though it was fairly abundant in the Twin Falls assemblage (sample size of more than 50 specimens), only 26% of Spinatrypina were encrusted, whereas at least 47% of other brachiopod taxa were encrusted. In particular, the atrypide taxon Radiatrypa (Fig. 3B) in the Twin Falls assemblage was comparable in size to Spinatrypina and was encrusted at 47%. While Desquamatia also had ornamentation in the form of large, frill-like lamellae, its lamellae were spread out and not nearly as abundant as those present on Spinatrypina. Other studies have also cited the decreased encrustation of particularly densely frilled species of brachiopod hosts (Bordeaux and Brett, 1990). The interlamellar spaces between the frills of Desquamatia may provide an advantageous position for sclerobionts, but large size is a more parsimonious explanation.

89

Live/dead associations between sclerobionts and encrusted hosts were difficult to ascertain in this study as there was a lack of evidence indicating these types of live/dead host associations (cf. Alvarez and Taylor, 1987; Bose et al., 2011). Only a handful of specimens combined from all three assemblages clearly indicated that encrustation occurred after the host was dead (at least in part). This occurred when the sclerobiont was growing over the commissure of the brachiopod host, or when the sclerobiont was growing in the pedicle foramen of a brachiopod. Live associations are usually indicated by parallel growth of the sclerobiont and its host, but this can be challenging to quantify with certain taxa of sclerobionts, particularly colonial taxa such as bryozoans and Hederella, which do not necessarily restrict themselves to unidirectional growth (Alvarez and Taylor, 1987; Taylor and Wilson, 2003). Cornulities (Fig. 4B), occasionally appeared to be growing synchronously with the host, but Cornulites was fairly rare. Many sclerobiont studies have also examined sclerobiont preference for brachiopod host valve (e.g. Thayer, 1974; Kesling et al., 1980; Alexander and Scharpf, 1990, Bordeaux and Brett, 1990), but the results of this study indicate that valve preference may not be consistent, even across sclerobiont taxa or with the same host taxon, across space and time. Sclerobionts on the brachiopods of the Firebag assemblage, comprised largely of Desquamatia, were significantly more abundant on the dorsal valve. Other studies have also found increased abundance on the dorsal valve in atrypides (Kesling et al., 1980; Bose et al., 2011). Bose et al. (2011) attributed this result to the life orientation of the host, Pseudoatrypa, in which the authors inferred that the dorsal valve would have been exposed. It is important to note that this preference does not continue into the two stratigraphically younger assemblages, which emphasizes that sclerobiont preference for host valve should never be assumed to stay consistent between assemblages, even if the host is the same. An alternative explanation is that encrustation may have occurred at different times during the exposure of the brachiopod assemblage, in one case during life and in another unit after the death of the hosts. However, Rodland et al. (2006) demonstrated that there exists a narrow window for encrustation before burial. This narrow window of encrustation before burial would likely inhibit the development of large colonies, as observed by the small size of bryozoans colonies in the Twin Falls assemblage. It is also possible that brachiopod hosts, such as Desquamatia, do not maintain the same position after death (i.e. different valves may be exposed in life position and after death), and that one of these assemblages represents an assemblage where encrustation occurred during the life of the host, whereas encrustation may have occurred after host death in the other assemblage. Another possibility is a change in the direction of water currents. Kesling et al. (1980) found that sclerobionts encrusted hydrodynamically advantageous positions on the sides of the hosts. Depending on the direction of the current and the life orientation of the host, sclerobionts may therefore have been more abundant on different valves to take advantage of water currents, although this implies a predominant direction to currents, for which we have found no independent supporting evidence. It is plausible that decreased sediment influx and time since disturbance played an important role in shaping sclerobiont patterns observed between the three assemblages. The most important observation one can make from the results of this study is that sclerobiont preferences do not remain static from assemblage to assemblage, even when the host assemblages are similar. Therefore, when examining sclerobionts in the fossil record, it should never be assumed that sclerobiont patterns will remain the same through time, or across similar environments. Further investigations of the factors affecting these changes in sclerobiont patterns are necessary to understand the interactions between sclerobionts and hosts more fully, both in the fossil record and in the modern. The observations made in this study provide a starting point and protocol for many other ecological studies, and reinforce the importance of examining several entire sclerobiont and host assemblages.

90

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91

6. Summary/conclusions While the literature contains numerous studies examining the interactions between hosts and sclerobionts, this study is an attempt to demonstrate the need to examine not only sclerobiont assemblages present on “entire” host assemblages, but to also examine multiple assemblages of those hosts. There are several important trends that emerge from the results of this study that have potential impact, not only for future research on communities of brachiopod hosts and sclerobionts of the Western Canadian Sedimentary Basin, but also for sclerobiont/host studies in general. 1) Richness, diversity, and evenness of brachiopods increase through time from the Firebag to the Hay River and Twin Falls assemblages. This is most likely due to a decrease in the amount of terrigenous mud influx and changes associated with sediment influx between the lower two assemblages and the Twin Falls assemblage. Diversity trends are also slightly less pronounced when including only encrusted hosts, and evenness of encrusted hosts is higher in each assemblage. 2) Even though there is a stronger preference for large hosts in the Firebag assemblage, the assemblage which was most heavily dominated by Desquamatia, there is a stronger preference for Desquamatia, regardless of size, as a host for sclerobionts in the Twin Falls assemblage. This may represent a true biological preference for Desquamatia on the part of the sclerobionts in the Twin Falls assemblage. 3) Valve preference is not consistent across each assemblage, even though the assemblages are basically subsets of one another, largely comprised of Desquamatia. While this makes it impossible to infer life orientation of Desquamatia, it does emphasize that sclerobiont patterns do not remain static between assemblages. Therefore, the consideration of multiple assemblages of hosts and sclerobionts is of great importance to future palaeoecological studies. Acknowledgements This research was funded by an Undergraduate Student Research Award (first author), and a Discovery Grant (third author) received from the Natural Sciences and Engineering Research Council of Canada. We would like to thank Simen Linge-Johnsen for his assistance with the identification and measurement of brachiopod specimens from the Twin Falls and Hay River assemblages. The authors are very grateful to Darrin Molinaro for donating his time and skills to help with the photography of specimens. We would also like to thank Amelinda Webb for her advice and help with the kurtosis analyses. The authors are also grateful to Jed Day of Illinois State University for his guidance with stratigraphy and field identification of brachiopod specimens, as well as his assistance with field collection of brachiopods. We also thank Carl Brett, an anonymous reviewer, and Editor David Bottjer for their thoughtful reviews. References Ager, D.V., 1961. The epifauna of a Devonian spiriferid. Quarterly Journal of the Geological Society 117, 1–10. Alexander, R.R., Gibson, M.A., 1993. Paleozoic brachiopod autecology based on taphonomy: example from the Devonian Ross Formation of Tennesee (USA). Palaeogeography, Palaeoclimatology, Palaeoecology 100, 25–35. Alexander, R.R., Scharpf, C.D., 1990. Epizoans on late Ordovician brachiopods from Southeastern Indiana. Historical Biology 4, 179–202. Alvarez, F., Taylor, P.D., 1987. Epizoan ecology and interactions in the Devonian of Spain. Palaeogeography, Palaeoclimatology, Palaeoecology 61, 17–31. Bazzaz, F.A., 1975. Plant species diversity in old-field successional ecosystems in southern Illinois. Ecology 56, 485–488. Bordeaux, Y.L., Brett, C.E., 1990. Substrate specific associations of epebionts on Middle Devonian brachiopods: implications for paleoecology. Historical Biology 4, 203–220. Bose, R., Schneider, C.L., Leighton, L.R., Polly, P.D., 2011. Influence of atrypid morphological shape on Devonian episkeletobiont assemblages from the Lower Genshaw

Formation of the Traverse Group of Michigan: a geometric morphometric approach. Palaeogeography, Palaeoclimatology, Palaeoecology 310, 427–441. Brett, C.E., Bordeaux, Y.L., 1991. Taphonomy of brachiopods from a Middle Devonian shell bed; implications for the genesis of skeletal accumulations. Proceedings of the Second International Brachiopod Congress 2, 219–226. Brett, C.E., Boucot, A.J., Jones, B., 1993. Absolute depths of Silurian benthic assemblages. Lethaia 26, 25–40. Brett, C.E., Parsons-Hubbard, K.M., Walker, S.E., Ferguson, C., Powell, E.N., Staff, G., Ashton-Alcox, K.A., Raymond, A., 2011. Gradients and patterns of sclerobionts on experimentally deployed bivalve shells: synopsis of bathymetric and temporal trends on a decadal time scale. Palaeogeography, Palaeoclimatology, Palaeoecology 312, 278–304. Brett, C.E., Smrecak, T., Parsons-Hubbard, K., Walker, S., 2012. Marine Sclerobiofacies: encrusting and endolithic communities on shells through time and space. In: Talent, J.A. (Ed.), Earth and Life: International Year of Planet Earth. Springer Science + Business Media, New York, pp. 129–157. Buschkuehle, B.E., 2003. Sedimentology and stratigraphy of Middle and Upper Devonian carbonates in northern Alberta: a contribution to the carbonate-hosted Pb–Zn (MVT) targeted geoscience initiative. EUB-AGS Geo-Note 2002–14, 1–14. Cobianchi, M., Picotti, V., 2001. Sedimentary and biological response to sea-level and palaeoceanographic changes of a Lower–Middle Jurassic Tethyan platform margin (Southern Alps, Italy). Palaeogeography, Palaeoclimatology, Palaeoecology 169, 219–244. Copper, P., 1967. Spinatrypa and Spinatrypina (Devonian Brachiopoda). Palaeontology 10, 489–523. Crickmay, C.H., 1957. Elucidation of Some Western Canadian Devonian Formations, Self Published. Imperial Oil Limited, Calgary. Day, J., 1996. Faunal signatures of Middle–Upper Devonian depositional sequences and sea-level fluctuations in the Iowa Basin: U. S. Midcontinent. Geological Society of America Special Papers 306, 277–300. Day, J., 1998. Distribution of latest Givetian-Frasnian Atrypida (Brachiopoda) in central and western North America. Acta Palaeontologica Polonica 43, 205–240. Day, J., Copper, P., 1998. Revision of latest Givetian-Frasnian Atrypida (Brachiopoda) from central North America. Acta Palaeontologica Polonica 43, 155–204. De Keyser, T.L., 1977. Late Devonian (Frasnian) brachiopod community patterns in Western Canada and Iowa. Journal of Paleontology 51, 181–196. Dominici, S., Zuschin, M., 2007. Sea-level change and the structure of marine ecosystems. Palaios 22, 225–227. Energy Resources Conservation Board, 2009. Table of Formations, 1 Sheet. Gibson, M.A., 1992. Some sclerobiont–host and sclerobiont–sclerobiont relationships from the Birdsong Shale Member of the Lower Devonian Ross formation (west-central Tennessee, U.S.A.). Historical Biology 6, 113–132. Gray, J.S., 1981. Detecting pollution induced changes in communities using the log-normal distribution of individuals among species. Marine Pollution Bulletin 12, 173–176. Hadley, M.G., Jones, B., 1990. Lithostratigraphy and nomenclature of Devonian strata in the Hay River area, Northwest Territories. Bulletin of Canadian Petroleum Geology 38, 332–356. Hoare, R.D., Steller, D.L., 1967. A Devonian brachiopod with epifauna. The Ohio Journal of Science 67, 291–297. Hurst, J.M., 1974. Selective epizoan encrustation of some Silurian brachiopods from Gotland. Palaeontology 17, 423–429. Jamieson, E.R., 1967. Upper Devonian of the Hay River area. International Symposium on the Devonian System, Calgary, Guidebook, Field Trip A-11. (39 pp.). Joachimski, M.M., van Geldern, R., Breisig, S., Buggisch, W., Day, J., 2004. Oxygen isotope evolution of biogenic calcite and apatite during the Middle and Late Devonian. International Journal of Earth Sciences 93, 542–553. Johnson, J.G., 1974. Devonian brachiopod biofacies of western and Arctic North America. Journal of Paleontology 48, 809–819. Johnson, J.G., 1975. Speciation in fossil brachiopods. Journal of Paleontology 49, 646–661. Johnson, J.G., Klapper, G., Sandberg, C.A., 1985. Devonian eustatic fluctuations in Euramerica. Geological Society of America Bulletin 96, 567–587. Kesling, R.V., Hoare, R.D., Sparks, D.K., 1980. Epizoans of the Middle Devonian brachiopod Paraspirifer bownockeri: their relationships to one another and to their host. Journal of Paleontology 54, 1141–1154. Kröger, B., Landing, E., 2010. Early Ordovician community evolution with eustatic change through the middle Beekmantown Group, northeast Laurentia. Palaeogeography, Palaeoclimatology, Palaeoecology 294, 174–188. Lescinsky, H.L., 1996. Don't overlook the epibionts! Palaios 11, 495–496. Lescinsky, H.L., 1997. Epibiont communities: recruitment and competition on North American Carboniferous brachiopods. Journal of Paleontology 71, 34–53. Loranger, D.M., 1965. Devonian paleoecology of northeastern Alberta. Journal of Sedimentary Petrology 35, 818–837. Ma, X., Day, J., 2000. Revision of Tenticospirifer Tien, 1938, and similar spiriferid brachiopod genera from the Late Devonian (Frasnian) of Eurasia, North America, and Australia. Journal of Paleontology 74, 444–463. MacNeil, A.J., Jones, B., 2006. Sequence stratigraphy of a Late Devoniam ramp-situation reef system in the Western Canadian Seidmentary Basin: dynamic responses of sea-level change and regressive reef development. Sedimentology 53, 321–359. Moore, P.F., 1988. Devonian geohistory of the western interior of Canada. Devonian of the world: proceedings of the Second International Symposium of the Devonian System 1, 67–83. Norris, A.W., 1963. Devonian stratigraphy of northeastern Alberta and northwestern Saskatchewan. Memoir 313, Geological Survey of Canada, Ottawa. Norris, A.W., 1965. Stratigraphy of Middle Devonian and older Palaeozoic rocks of the Great Slave Lake region, Northwest Territories. Memoir 322, Geological Survey of Canada, Ottawa.

K.M. Barclay et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 383–384 (2013) 79–91 Patzkowsky, M.E., Holland, S.M., 1999. Biofacies replacement in a sequence stratigraphic framework: Middle and Upper Ordovician of the Nashville Dome, Tennessee, USA. Palaios 14, 301–323. Pitrat, C.W., Rogers, F.S., 1978. Spinocyrtia and its sclerobionts in the Traverse Group (Devonian) of Michigan. Journal of Paleontology 52, 1315–1324. Rodland, D.L., Kowalewski, M., Carroll, M., Simões, M.G., 2004. Colonization of a ‘Lost World’: encrustation patterns in modern subtropical brachiopod assemblages. Palaios 19, 381–395. Rodland, D.L., Kowalewski, M., Carroll, M., Simões, M.G., 2006. The temporal resolution of sclerobiont assemblages: are they ecological snapshots or overexposures? Journal of Geology 114, 313–324. Ruban, D.A., 2007. Jurassic transgressions and regressions in the Caucasus (northern Neotethys Ocean) and their influences on the marine biodiversity. Palaeogeography, Palaeoclimatology, Palaeoecology 241, 422–436. Schneider, C.L., 2003. Hitchhiking on Pennsylvanian echinoids: Sclerobionts on Archaeocidaris. Palaios 18, 435–444. Schneider, C.L., Leighton, L.R., 2010. Epizoan and predation traces of Devonian Hay River Formation brachiopods: indicators of complex ecosystems and ties to the Iowa Basin. GeoCanada 2010—Working with the Earth, Conference Abstracts. Schneider, C.L., Hauck, T., Grobe, M., 2013. Sequence stratigraphy and architecture of the Beaverhill Lake Sequence, Western Canada Sedimentary Basin: the influences of changing sedimentological and climatological regimes. Invited paper, SEPM Special Publication, Deposits, Architecture and Controls of Carbonate Margin, Slope, and Basin Systems (in press). Sepkoski Jr., J.J., 1976. Species diversity in the Phanerozoic: species-area effects. Paleobiology 2, 298–303. Simon, L., Goddéris, Y., Buggisch, W., Strauss, H., Joachimski, M.M., 2007. Modeling the carbon and sulfur isotope compositions of marine sediments: climate evolution during the Devonian. Chemical Geology 246, 19–38. Stigall Rode, A.L., Lieberman, B.S., 2005. Paleobiogeographic patterns in the Middle and Late Devonian emphasizing Laurentia. Palaeogeography, Palaeoclimatology, Palaeoecology 222, 272–284. Stoakes, F.A., Wendte, J.C., Campbell, C.V., 1992. Summary. In: Wendte, J.C., Stoakes, F.A., Campbell, C.V. (Eds.), Devonian—Early Mississippian Carbonates of the Western Canadian Sedimentary Basin: A Sequence Stratigraphic Framework: Society for Sedimentary Geology Short Course, 28, pp. 215–255. Switzer, S.B., Holland, W.G., Christie, D.S., Graf, G.C., Hedinger, A.S., et al., 1994. Devonian Woodbend-Winterburn strata of the Western Canadian sedimentary basin. Canadian Society of Petroleum Geologists. Calgary, Alberta, Canada.

91

Taylor, P.D., Wilson, M.A., 2002. A new terminology for marine organisms inhabiting hard substrates. Palaios 17, 522–525. Taylor, P.D., Wilson, M.A., 2003. Palaeoecology and evolution of marine hard substrate communities. Earth-Science Reviews 62, 1–103. Taylor, P.D., Wilson, M.A., 2008. Morphology and affinities of hederelloid “bryozoans”. Virginia Museum of Natural History, Special Publication 15, 301–309. Thayer, C.W., 1974. Substrate specificity of Devonian epizoa. Journal of Paleontology 48, 881–894. van Geldern, R., Joachimiski, M.M., Day, J., Jansen, U., Alvarez, F., Yolkin, E.A., Ma, X.-P., 2006. Carbon, oxygen, and strontium isotope records of Devonian brachiopod shell calcite. Palaeogeography, Palaeoclimatology, Palaeoecology 240, 47–67. Webb, A.E., Leighton, L.R., 2011. Exploring the ecological dynamics of extinction. In: Laflamme, M., Schiffbauer, J.D., Dornbos, S.Q. (Eds.), Quantifying the Evolution of Early Life: Numerical Approaches to the Evaluation of Fossils and Ancient Ecosystems. Springer Science + Business Media, New York, pp. 185–220. Webb, A.E., Leighton, L.R., Schellenberg, S.A., Landau, E.A., Thomas, E., 2009. Impact of the Paleocene–Eocene thermal maximum on deep-ocean microbenthic community structure: using rank-abundance curves to quantify paleoecological response. Geology 37, 783–786. Wendte, J.C., 1992. Cyclicity of Devonian strata in the Western Canada sedimentary basin. SEPM Short Course Notes 28, 25–39. Wendte, J., Uyeno, T., 2005. Sequence stratigraphy and evolution of Middle to Upper Devonian Beaverhill Lake strata, south-central Alberta. Bulletin of Canadian Petroleum Geology 53, 250–354. Whalen, M.T., Day, J.E., 2010. Cross-basin variations in magnetic susceptibility influenced by changing sea level, paleogeography, and paleoclimate: Upper Devonian, Western Canada Sedimentary Basin. Journal of Sedimentary Research 80, 1109–1127. Williams, G.K., 1977. The Hay River Formation and its relationship to adjacent formations, Slave River map-area, N.W.T. Geological Survey of Canada Paper 75—12, pp. 1–17. Witzke, B.J., Heckel, P.H., 1988. Paleoclimatic indicators and inferred Devonian paleolatitudes of Euramerica. Devonian of the World: Proceedings of the Second International Symposium of the Devonian System 1, 49–63. Zatoń, M., Krawczyński, W., 2011. Microconchid tubeworms across the Upper Frasnian–Lower Famennian interval in the Central Devonian Field, Russia. Palaeontology 54, 1455–1473. Zeigler, A.M., 1965. Silurian marine communities and their environmental significance. Nature 207, 270–272.