Abstract
The phylogeographical evolution and the consequent changing distribution and diversity of rhynchonelliform brachiopods through the Ordovician are linked to the dynamic palaeogeography of the period. The Early Ordovician (Tremadocian and Floian) is characterized by globally low-diversity faunas with local biodiversity epicentres, notably on the South China Palaeoplate; low-latitude porambonitoid-dominated faunas with early plectambonitoid and clitambonitoid representatives, as well as high-latitude assemblages mostly dominated by orthoids, can be recognized, but many taxa are rooted in Late Cambrian stocks. The Early Ordovician displays a steady increase in rhynchonelliformean biodiversity, which was mostly driven by the increasing success of the Porambonitoidea and Orthoidea, but the billingsellids and early plectambonitoids also contributed to this expansion. During the Early to Mid Ordovician (Dapingian–Darriwilian), marine life experienced an unprecedented hike in diversity at the species, genus and family levels that firmly installed the suspension-feeding benthos as the main component of the Palaeozoic fauna. However, this may have occurred in response to an early Darriwilian annihilation of existing clades, some of which had been most successful during the Early Ordovician. New clades rapidly expanded. The continents were widely dispersed together with a large number of microcontinents and volcanic arcs related to intense magmatic and tectonic activity. Climates were warm and sea-levels were high. Pivotal to the entire diversification is the role of gamma (inter-provincial) diversity and by implication the spread of the continents and frequency of island arcs and microcontinents. The phylogeographical analysis demonstrates that this new palaeogeographical configuration was particularly well explored and utilized by the strophomenides, especially the Plectambonitoidea, which radiated rapidly during this interval. The porambonitoids, on the other hand, were still in recovery following the early Darriwilian extinctions. Orthides remained dominant, particularly at high latitudes. Biodiversity epicentres were located on most of the larger palaeoplates, as well as within the Iapetus Ocean. Provincial patterns were disrupted during the Sandbian and early Katian with the migration of many elements of the benthos into deeper-water regimes, enjoying a more cosmopolitan distribution. Later Katian faunas exhibit a partition between carbonate and clastic environments. During the latest Katian, biogeographical patterns were disrupted by polewards migrations of warm-water taxa in response to the changing climate; possibly as a consequence of low-latitude cradles being developed in, for instance, carbonate reef settings. Many clades were well established with especially the strophomenides beginning to outnumber the previously successful orthides, although this process had already begun, regionally, in the mid to late Darriwilian. At the same time, atrypoid and pentameroid clades also began to radiate in low-latitude faunas, anticipating their dominance in Silurian faunas. The Hirnantian was marked by severe extinctions particularly across orthide-strophomenide clades within the context of few, but well-defined, climatically controlled provincial belts.
Supplementary material: The individual localities and a reference list for the data sources are provided at: http://www.geolsoc.org.uk/SUP18667
Overview
A number of nineteenth-century palaeontologists, including Nicholson & Etheridge (1878), noted provincial differences between North American and European Early Palaeozoic faunas. It was not until the 1960s and early 1970s, however, that the biogeography of Early Palaeozoic faunas was analysed in some detail with a particular focus on the benthos such as the Brachiopoda (e.g. Spjeldnæs 1960; Williams 1969, 1973; Burrett 1973; Jaanusson 1973, 1979).
Historically many of the discussions related to Ordovician biogeography have focussed on the better-known, better-sampled and contrasting European and North American faunas. Tuzo Wilson (1966), in a benchmark paper, provided an explanation for these differences, based on the existence of a previous Atlantic Ocean, the proto-Atlantic that divided Europe from North America. Orthogonal opening and closure of this Early Palaeozoic oceanic system formed the basis for the Wilson Cycle. During the late 1960s, a thorough multivariate investigation, using cluster analysis, of all the then-known Ordovician brachiopod faunas confirmed differences between the North American and European biotas but also indicated important provincial differences within the European faunas themselves; for example, the Dapingian brachiopods of the island of Anglesey could be associated with those from Baltica, in contrast to those from Shropshire, which apparently had affinities with those from the Montagne Noire of southern France (Williams 1969). A later re-evaluation of these data (Williams 1973) indicated the presence of a well-defined cluster, separate from those of Scoto–Appalachia, northwestern North America and the Baltic, characterized by the presence of Rhynchorthis but demonstrating close ties with the Baltic province. Williams (1973) considered that this, his ‘Celtic province’ containing marginal areas of Ireland and Wales, was a distinct biogeographical entity not necessarily associated with a single continental palaeoplate. Williams (1973) demonstrated a reduction of provinces during the Ordovician, from five in the Arenig to three in the Ashgill. Moreover, the relative numbers of endemic genera in each province ranged from 80 to 90% during the Arenig and Llanvirn, falling to some 55% during the later Ashgill (Williams 1973, fig. 12). During the 1970s and 1980s Tuzo Wilson's model was modified to accommodate rates of opening and closure based on faunal data (e.g. McKerrow & Cocks 1976) and a more complex palaeoplate configuration (Cocks & Fortey 1982; Fortey & Cocks 1986) involved the interaction of the continents of Avalonia, Baltica, Gondwana and Laurentia. None of these models explicitly indicated the presence of islands; however, Neuman (1984), building on his earlier research on the brachiopod faunas within the Caledonian–Appalachian orogen (e.g. Neuman 1964, 1972), indicated the existence of a group of intra-Iapetus islands in his palaeogeographical reconstructions, now amalgamated within the Caledonian orogen. A large number of the faunas associated with these island terranes could now be tied to Williams's concept of the Celtic province, suggesting origins and positions seawards of the main continental palaeoplates. These faunas, however, presented other characteristic features: they contained a high proportion of endemics, taxa common to adjacent continents, and some taxa that are better known from younger rocks in the adjacent platform provinces. These patterns had also been established for Scandinavian faunas (Bruton & Harper 1981), and later confirmed by statistical analyses (Harper & Mac Niocaill 2002; Harper et al. 2008).
Biogeographical studies of Ordovician Brachiopoda have followed three parallel approaches. First, many analyses have focussed on the occurrences of key genera or families to help map provincial data (Cocks & Fortey 1982). Second and by way of contrast, a number of authors have used the whole-data approach, analysing statistically all the available data from given time slices using a range of multivariate methods (Williams 1969, 1973; Harper 1992). Third, the biogeographical distributions of taxa have been mapped onto phylogenetic trees for some areas of the phylum (Popov et al. 2001; Candela 2011). Moreover the concept and application of provinces have differed between author groups (see below), with some authors basing their provinces entirely on the distribution of taxa whereas others have tied them to continental palaeoplates.
In this study we have access to the most comprehensive database of brachiopod occurrences through the Ordovician. The database (see Supplementary material), assembled in the University of Copenhagen by M. Liljeroth and C. M. Ø. Rasmussen and supplemented with data from all the other co-authors, is locality-based and grouped into seven time slices, and includes the most up-to-date generic assignments of taxa. These data have been converted to presence–absence matrices for the seven time slices: Tremadocian, Floian, Dapingian–Darriwilian (the late Arenig–early Lanvirn brachiopods are difficult to precisely date in some parts of the world, particularly within the marginal and oceanic terranes), Sandbian, Lower Katian, Upper Katian and Hirnantian. The database has also been used to access the diversity of the rhynchonelliformean brachiopods through the Ordovician (Fig. 11.1) to specifically unravel the phylogeographical evolution within the clade (Fig. 11.16).
Changing diversity through the seven time slices, indicated by maximum number of genera and maximum number of localities together with extinction and origination curves. For a more detailed diversity curve, see Figure 11.16.
The data have been mapped onto BugPlates (http://www.geodynamics.no/bugs/SoftwareManual.pdf), but these, like many reconstructions, have shortcomings. In particular, the ongoing work on the Australasian part of the Gondwanan margin together with the Central Asian orogenic belt is not integrated with the currently available software and will be subject to much revision over the coming years (see e.g. Popov et al. 2009).
Biogeographical provinces
The data are analysed by a number of multivariate techniques. There are, however, a number of challenges interrogating such substantial databases that by their nature contain a great deal of noise and are not free from errors. First, a number of the major continental blocks (e.g. Laurentia, Baltica and South China) that are particularly rich in data from many localities across a range of facies commonly present a range of confusing and sometimes different biogeographical signals. Even within single basins, for example the Appalachians and adjacent areas (Jaanusson & Bergström 1980) and the Baltic basin (Jaanusson 1976), there are many different biofacies ranging from shallow to deep water, typified by a spectrum of substrates from varieties of carbonates to siliciclastic sediments. For these larger, more complex continental units data have been combined from the large numbers of localities available in the database to provide a more average but nevertheless comprehensive biogeographical signal. For the purposes of biogeographical analyses this process is defensible; however, from the Sandbian onwards the diversification of deeper-water, marginal and more cosmopolitan faunas requires the partition of some continental areas into shallower- and deeper-water faunas.
Second, in grouping together faunas into possible provinces, endemic taxa (essentially singletons) from individual localities or suites of adjacent localities have been eliminated. Admittedly provinces are defined on the basis of their endemic taxa; however, the numbers of endemic taxa are discussed under each of the time slices. Third, use of data at this resolution will only delineate the major brachiopod provinces and we have not attempted in most cases to identify in detail marginal faunas associated with the major continents.
Here, we also introduce the term ‘species pump’, to describe regions with great species diversity, which have acted as centres that allowed the extra-regional spread of taxa. This term is adapted from its use in modern biogeographical studies, and is thus preferred over terms such as ‘hot spots’, which have a different meaning in present-day biogeography. Further, we use the term ‘biodiversity epicentre’. This term is applied to regions with high biodiversity, but without any evidence of faunal migrations to neighbouring regions. Lastly, the term ‘cradle’ is applied to regions where early signs of developing species pumps or biodiversity epicentres are located. This tri-partite division of evolving species-rich regions may thus be viewed as a progressive development of a faunal region, going from ‘cradle’, where key taxa emerge, through ‘biodiversity epicentre’ with high levels of α- and β-diversity to finally a fully developed ‘species pump’, resulting in the increase in γ-diversity.
Time slices
Tremadocian
There have been few attempts to assess articulate brachiopod biogeography during the Tremadocian. For example, brachiopods did not merit mention in Jaanusson's review of Ordovician biogeography (1979). In addition, a number of previous studies prior to the establishment of the current base of the system and international stadial divisions (e.g. Williams 1973) included the Tremadoc Series within the Cambrian, whereas some studies (e.g. Cocks 2001) included discussion of these faunas with those of the Arenig Series. The articulate brachiopod assemblages from this stage are relatively few in number and represented by low-diversity associations strongly rooted in the Late Cambrian. Overall, based on currently available data, the mid Cambrian to Tremadocian seems to have been an interval of evolutionary stasis within the rhynchonelliforms. Nevertheless pockets of endemism and regions of high local diversity suggest that the subsequent Great Ordovician Biodiversification Event (GOBE) was already firmly rooted. As noted above, previous authors have largely ignored this interval with the exception of Jaanusson (1973), who noted the remarkable similarity amongst many Tremadocian assemblages although a distinctive Southern Fauna (essentially the Mediterranean Province) characterized by Poramborthis can be recognized.
Fifty localities with some 40 genera have been recorded globally. There is marked endemicity at local levels with 20 genera reported from only one site whereas one genus, Nanorthis, is reported from 25 sites. One site, Guizhou (South China Palaeoplate), contains the most diverse assemblage with 10 genera, while 12 sites are represented by only one genus. The standardized diversity for the Tremadocian is 0.82. Sample sizes are thus small and biogeographical patterns are far from robust. Nevertheless a few comments are possible based on the biogeographical distribution of taxa (Figs 11.2 & 11.3). High-latitude (Southern Fauna), low-latitude (Northern Fauna) and Baltic provinces are recognized. The last provided the initial immigrants that established the high-latitude Mediterranean Province (Havlíček 1989). The low-latitude faunas are dominated by orthidine and syntrophoid brachiopods, whereas higher-latitude assemblages are characterized by an array of polytoechioids associated with a few orthidines. Bassett et al. (2002), in a detailed analysis of the biogeography and diversity of brachiopod faunas through the Cambrian–Ordovician transition, identified a high-latitude group of faunas within the siliciclastic facies of the Gondwanan margins; a number of polytoechioids, probably derived from the earlier Billingsella fauna, dominated the margins, later migrating to the Urals. Lower-latitude faunas were dominated by the syntrophoids and orthidines. This partially confirms Benedetto's (2001) identification of high-latitude orthidine-dominated assemblages and low-latitude pentameride faunas during the Early Ordovician.
Palaeogeographical map for the Tremadocian time slice based on BugPlates and using a Mollweide projection. The individual localities are listed in the supplementary material.
Cluster analysis for a selection of faunas for the Tremadocian time slice using the Raup–Crick (Hammer & Harper 2006) similarity coefficient and a neighbour-joining algorithm.
On the carbonate shelves along the margins of Gondwana, billingselloid, polytoechioid and syntrophoid associations developed in the equatorial zone and migrated later to Laurentia, Siberia and the Uralian margins of Baltica (Bassett et al. 2002). These and subsequent faunas dominating the Floian have parallels with the composition, distribution and structure of the Ibexian trilobite faunas, associated with late Cambrian–Early Ordovician carbonate facies. South China, situated in equatorial latitudes adjacent to the western-facing margins of Gondwana, was a local biodiversity epicentre, exhibiting unusually high levels of diversity (the Finkelnburgia and Tritoechia faunas; Wang & Xu 1966; Zhan & Harper 2006). These faunas are very similar to those from Laurentia and related regions. Overall, however, many of the taxa are linked directly with precursors in the later Cambrian (although in general the faunal compositions are quite different), consolidating the concept of an Ibexian brachiopod fauna, prior to the major change-over at the base of the Whiterock.
Floian
Some 30 localities have been identified globally for this stage containing 87 genera. One site, Guizhou, in South China, contains 38 genera whereas the remaining 29 sites have between one and 11 genera. Some 35 genera occur at only one site whereas two genera, Hesperonomia and Tritoechia, occur at 11 and 12 sites, respectively. The standardized diversity for the Floian is 2.90, an important increase from levels in the Tremadocian. The rhynchonelliform brachiopods of the Floian interval are to date poorly studied. These are, however, known from a large number of equatorial localities, dominated by orthidines such as Archaeorthis, Apheoorthis and Finkelnburgia together with camerelloids, associated with the isolation of Laurentia from the other major continents (Hansen & Holmer 2010). Higher latitudes were typified by a different group of orthidines, including Paurorthis, Prantlina and Ranorthis. South China remained a locus for high diversity (Zhan & Harper 2006), manifested by the Sinorthis fauna (a typical regional brachiopod fauna in late Early Ordovician representing the first radiation acme in South China; see Zhan & Jin 2008; Zhan et al. 2011). However, a major change in brachiopod faunal affinities occurred at the beginning of the Floian (Tetragraptus approximatus Biozone). Whereas Tremadocian brachiopods of South China show close linkages to those of Laurentia, during the Floian to Darriwilian interval, as the South China block drifted away from Gondwana, its faunal affinity gradually shifted to closer relationships with the terranes of Baltica, Avalonia, Sibumasu and southern Kazakhstan (Zhan et al. 2011); nevertheless it displayed strong endemism, marked by the presence of Pseudomimella, Pseudoporambonitoides, Sinorthis, Xinanorthis and Yangtzeella (Xu & Liu 1984). In this study (Figs 11.4 & 11.5) we recognize high- and low-latitude provinces together with a Baltic Province.
Palaeogeographical map for the Floian time slice based on BugPlates and using a Mollweide projection. The individual localities are listed in the supplementary material.
Cluster analysis for a selection of faunas for the Floian time slice using the Raup–Crick (Hammer & Harper 2006) similarity coefficient and a neighbour-joining algorithm.
Dapingian–Darriwilian
This is a key interval for brachiopod diversification and continental disparity (Harper et al. 2009). The Dapingian and Darriwilian have been grouped into one time slice since operationally it has been difficult to separate some of their respective brachiopod faunas, especially those within the late Arenig–early Llanvirn interval and particularly those from the world's mountain belts that lack other age constraints. Some 65 localities have been identified globally for this stage, containing over 230 genera. Two sites in the East Baltic have diversities approaching 40 genera, whereas some 35 sites have diversities ranging from one to seven genera. Over 100 genera occur at only one site, whereas two genera, Tritoechia and Paralenorthis, occur at over 15 and 20 sites respectively. The standardized diversity for the Dapingian–Darriwilian interval is 3.57, another significant increase from levels in the Floian.
A disparate group of continental fragments and island arcs loosely assigned to the Celtic province has been identified, containing a distinctive suite of shelly faunas that formed a testable biogeographical unit. The Celtic faunas are characterized by a large number of endemic brachiopod taxa, some cosmopolitan forms and taxa at the beginning or end of their stratigraphical ranges. Multivariate analyses of many of the Dapingian–early Darriwilian (late Arenig–early Llanvirn) brachiopod faunas (Neuman & Harper 1992; but see also Harper 1992) have confirmed the discrete and distinctive grouping of both the sites and taxa constituting the Celtic province (Figs 11.6 & 11.7). This biogeographical unit has been tested by both new data and modern statistical techniques, and its integrity remains consistently reproducible (Harper et al. 1996, 2008). The concept of the Celtic province was, however, challenged by McKerrow & Cocks (1993), who suggested that the term should be abandoned on the basis of the very widespread distribution of its brachiopod genera and the lack of an obvious pool of cross-province endemics. Implicit in their argument was the suggestion that some groups of brachiopods were better biogeographical indicators than others, a concept that has been applied to a number of other fossil groups (e.g. Fortey & Mellish 1992; Servais & Sintubin 2009). This argument was, however, not based on any particular taxonomic groups within the Brachiopoda, rather it was only those associated with the Celtic province that were clearly of limited value as they occurred along the same volcanic arc as localities with the marginal Laurentia Toquima–Table Head fauna. The single Bronson Hill–Tetagouche–Lushs Bight Island arc, however, proposed by McKerrow & Cocks (1993) is in fact at least two separate arcs: one part (with the Toquima–Table Head taxa) was associated with marginal Laurentia, whereas the other (associated with the Celtic province) developed at higher latitudes, within the Iapetus Ocean (Neuman et al. 1994). A more detailed analysis of this interpretation and an alternative model was provided by Harper et al. (1996) and Williams et al. (1996).
Palaeogeographical map for the Dapingian–Darriwilian time slice based on BugPlates and using a Mollweide projection. The individual localities are listed in the supplementary material.
Cluster analysis for a selection of faunas for the Dapingian–Darriwilian time slice using the Raup–Crick (Hammer & Harper 2006) similarity coefficient and a neighbour-joining algorithm.
More recently, support for the Celtic province has accumulated from three sources of data. First, new faunas particularly from South America, for example, Argentina (Benedetto & Sánchez 2003), Bolivia and Peru (Gutiérrez-Marco & Villas 2007), have provided new occurrences of taxa associated with the Celtic province, supporting its extension along a high- to mid-latitude belt. Second, new data from already well-documented sites such as Otta, central Norway (Harper et al. 2008), continue to provide taxa that anchor these faunas within the province. Third, new data from the platform provinces such as Baltica (Harper & Hints 2001; Sturesson et al. 2005; Rasmussen et al. 2007) and South China (Rong et al. 2005; Zhan & Harper 2006) confirm the differences between these faunas and those of the Celtic province. The brachiopod faunas of Dapingian and Darriwilian age are represented by the Yangtzeella and the Saucrorthis faunas, respectively, in South China, both of which are regional brachiopod faunas with very important palaeobiogeographical significance (Zhan et al. 2007, 2010). Multivariate analyses of the global dataset consistently identify the Celtic group as distinct from the platform provinces associated with Baltica and Laurentia (low latitude), and separate them from the marginal Laurentian Toquima–Table Head realm (e.g. Harper 2006a).
Whereas the existence and integrity of the Celtic group of brachiopod faunas has been demonstrated by a range of multivariate statistical analyses, the precise pre-drift positions of the terranes, and by implication the geographical extent of the Celtic province, require a more multidisciplinary approach. Early to Mid Ordovician palaeogeography has been refined by a series of recent studies based on both palaeontological and palaeomagnetic data (Harper et al. 1996). More recent publications have emphasized again the role of shallow-water, marine benthos in defining provinces, for example, those by Cocks (2000, 2001) and Fortey & Cocks (2003). These workers, although recognizing the existence of marginal and peripheral sites to Avalonia, Baltica and Laurentia, opposed the concept of marginal or oceanic provinces based on a lack of province-wide endemics; nevertheless, they have accepted that these faunas cannot readily be accommodated within the platform provinces associated with major continental palaeoplates. Further, modern palaeogeographical analyses (e.g. for Baltica, Cocks & Fortey 1998; Cocks & Torsvik 2005; for Avalonia, Cocks et al. 1997; for Siberia, Cocks & Torsvik 2007) have consistently recognized the complexity of Ordovician geography and the existence and location of marginal terranes. Current palaeogeographical studies have also emphasized the existence and significance of other comparable archipelagos. The Toquima–Table Head province (Toquima–Table Head Realm of Ross & Ingham 1970) includes a number of Laurentian marginal terranes now part of the Caledonian Appalachian belt. Parts of western Ireland together with SW Scotland and a chain of localities along the eastern seaboard of the USA have also yielded a distinctive fauna with some endemics and very different from the carbonate platform faunas of the adjacent Laurentian craton (Neuman & Harper 1992; Harper & Mac Niocaill 2002).
Typical of the province are the genera Aporthophyla, Idiostrophia, Leptella, Leptellina, Neostrophia, Taphrodonta, Toquimia, Trematorthis and Trondorthis. However, Aporthophyla and a number of other typical members of the Toquima–Table Head province are known from outside the margins of Laurentia. Aporthophyla occurs together with Leptellina in South China, together with a number of endemic taxa such as Parisorthis (Zhan et al. 2005). Aporthophyla apparently had a widespread distribution across broadly low-latitude sites from Laurentia to equatorial peri-Gondwanan sites. On the other hand, Paralenorthis was widely distributed across higher latitudes as well as some localities of lower latitude (e.g. in southern Tibet and South China; Rong et al. 2005; Zhan et al. 2005). Both taxa occur in terranes from South Kazakhstan, providing an interface between the two provinces (Zhan et al. 2005; Nikitina et al. 2006). The many terranes associated with central Asia have their own distinctive faunas. For example, the early Darriwilian brachiopod faunas of the Chu-Ili range and the West Balkhash region of South Kazakhstan contain some 60 genera distributed across at least five palaeocommunity types. The faunas are highly endemic but show strong links with South China at this time. The early to mid–Darriwilian (late Arenig–early Llanvirn) was an interval of intense magmatic and tectonic activity. This is reflected in the presence of a large variety of island arcs and microcontinents dispersed across a spectrum of latitudes that interacted with a range of oceanic currents (Christiansen & Stouge 1999) associated with radiations (Droser & Sheehan 1997; Bottjer et al. 2001; Harper et al. 2004), causing γ-diversity to increase rapidly.
Many of the data evaluated here are derived from the distributions of articulated (rhynchonelliform) brachiopods. The intensive study of sites associated with the Celtic province is relatively recent, essentially since the early 1960s, compared with those of the coeval platform provinces, many of which have been documented since the late 1800s and early 1900s (Harper 1998). This, coupled with the relative rarity of localities and the imperfect preservation of material, has contributed to the difficulties in defining this biogeographical unit. Moreover, some of the new genera described from this province are difficult to relate to existing taxa because of their poor preservation, the lack of some taxonomic information and their different and distinctive morphologies. Nevertheless, since the 1990s a critical mass of data has been available for analysis. The majority of these earlier Ordovician faunas were developed in relatively shallow-water facies (e.g. Lockley 1983; Cocks 1996) and thus in most cases assemblages from similar depths are compared. Multivariate analyses of large datasets of brachiopod distributions (Harper 1992, 2006a, b; Neuman & Harper 1992; Harper et al. 1996, 2008) have consistently isolated a cluster of brachiopods, identifiable as the Celtic province (Figs 11.6 & 11.7). About 15 taxa occur at two or more of the Celtic sites; four taxa (Paralenorthis, Productorthis, Tritoechia and Rugostrophia) are relatively widespread, whereas some, such as Famatinorthis, Ffynnonia, Monorthis, Platytoechia, Rhynchorthis and Treiroria, are relatively restricted, occurring only at two or three sites. A number of genera, such as Fistulogonites, Ottadalenites, Rutrumella and Schedophyla, are reported only from single sites. Many of these sites have been subsequently amalgamated into allochthonous complexes and are thus located in Caledonian or coeval orogenic belts so that today, unlike the coherent platform provinces, they have a very disjunct distribution. Nonetheless, where assembled in their pre-drift positions, the majority of sites form a high- to mid-latitude belt marginal to or seawards of Gondwana. The position of North China here and in many reconstructions is anomalous. More data are required from this microcontinent to establish its exact positions through the Ordovician.
The brachiopods experienced a marked diversification at the species, genus and family levels during the early and mid Ordovician (Harper et al. 2004; Zhan & Harper 2006; Rasmussen et al. 2007). It is clear, however, that the distributional patterns and ranges of the two main informal groups within the Brachiopoda, the non-articulated (Bassett et al. 1999; Popov et al. 1999) and articulated stocks (Harper et al. 1999, 2004), developed independently during the biodiversification and, moreover, ecological events within the phylum were not always correlated with taxonomic expansions. These groups have different zoogeographical patterns based on the behaviour of their respective larvae. Nonarticulate brachiopod larvae have a well-developed lophophore that allows them to feed in the plankton and live as plankton for months, whereas rhynchonelliform brachiopod larvae are lecithotrophic, and they do not feed in the water column and settle out of the plankton in only a few days.
Globally, the order Rhynchonellida first appeared during the late Darriwilian (Llanvirn) and underwent a modest diversification represented by Rostricellula, Ancistrorhyncha, Dorytreta, Sphenotreta and Lenatoachia, widely distributed in Laurentia, Siberia, Avalonia and Kazakhstan (Jin 1996). These newly evolved taxa, being unspecialized small shells and without complications of hold-over stratigraphical ranges, are particularly significant for biogeographical analysis. Their distribution suggests a semi-cosmopolitanism of the brachiopod faunas among Laurentia and its adjacent palaeoplates.
Sandbian
The basal Sandbian is signalled by a massive marine transgression that encouraged the migration of many deeper-water brachiopod taxa onto the cratons, for example the draboviid Oanduporella (Rasmussen 2011) and the strophomenoid Foliomena (Rong et al. 1999). The Sandbian also marks the first major migration of benthic shelly communities into deeper-water environments (Sepkoski & Sheehan 1983), reflected in a second peak in global diversification during the GOBE (Harper et al. 1999; Harper 2006a; Servais et al. 2010). In particular, the Foliomena fauna, characterized by minute-sized species, lived in relatively deep water, mainly outer-shelf, slope and basin, in many parts of the world (Rong et al. 1999). The advent of deeper-water faunas, with their own distinctive biogeographical signals, adds a further complexity to the distributional patterns of the global brachiopod fauna (Figs 11.8 & 11.9). Attempts at seriation of these data against a latitudinal gradient (Parkes et al. 1990) were unsatisfactory, indicating that the brachiopods were already participating in a wide range of complex community types across a range of depth zones and facies. Detrended correspondence analysis (Harper 1992) and cladistic analysis (Fortey & Mellish 1992) provided more consistent and conclusive results, although these studies largely excluded deeper-water assemblages such as the Foliomena fauna (Rong et al. 1999).
Palaeogeographic map for the Sandbian time slice based on BugPlates and using a Mollweide projection. The individual localities are listed in the supplementary material.
Cluster analysis for a selection of faunas for the Sandbian time slice using the Raup–Crick (Hammer & Harper 2006) similarity coefficient and a neighbour-joining algorithm.
Some 70 localities have been identified globally for this stage, containing over 240 genera. Two sites on the Laurentian margin have diversities approaching 70 genera, reflecting Cooper's (1956) monographic burst, whereas some 28 sites have diversities ranging from one to seven genera. Over 100 genera occur at only one site whereas one genus, Sowerbyella, occurs at over 30 sites. The standardized diversity for the Sandbian stage is 3.48, a slight drop from levels in the Dapingian–Darriwilian. Jaanusson (1973) recognized Northern and Southern faunas, corresponding to low- and high-latitude assemblages, but he also isolated a separate Scoto-Appalachian fauna, succeeding in part the earlier Toquima–Table Head fauna, around the margins of Laurentia; the Balto-Scandian fauna was now quite distinct. In the same time interval, Williams (1973) identified clusters relating to American, Baltic, Anglo-French and Bohemian faunas, whereas Havlíček (1989) noted that the Mediterranean Province is first unambiguously defined during the Sandbian.
The integrity and distinctive character of the high-latitude peri-Gondwanan faunas are confirmed, dominated by dalmanellidines such as Drabovia, Heterorthina, Svobodaina and Tafilaltia. There are now strong faunal links between the mid-latitude continents of Avalonia and Baltica, suggesting that well-developed species pumps were operating on both continents. These faunas were distinct from those of Laurentia and its margins. The deep-water Foliomena fauna was still limited to South China and parts of southern Kazakhstan (Rong et al. 1999).
Lower Katian
A recent detailed investigation of the biogeography and geography of this interval, essentially the late Caradoc (Candela 2006), interrogated a matrix of some 180 genera from nearly 30 localities around the world using cluster and detrended correspondence analyses. Key conclusions included the clear division of Kazakhstan into a number of independent terranes during this interval, with different faunal affinities and histories, the prevalence of deep-water Foliomena faunas in South China and a peri-equatorial location for the Gornoi-Altai terrane which would facilitate the westward migration of taxa. This migration route is in line with published palaeo-ocean current models for the late Ordovician (see Rasmussen 2011 and references therein).
More recently, Rasmussen et al. (2012) mapped out the relationships amongst, between and within; (a) the tropical platforms, terranes and temperate margins including Avalonia, deep-water faunas of Baltica, Laurentia and associated terranes, as well as Siberia and Altai; (b) tropical Gondwana and (c), high-latitude Gondwana. The Kazakh terranes are most similar to shallow-water facies in South China, and shallow-water assemblages in the Baltic form their own disparate branch signalling that although faunas are gradually becoming more cosmopolitan, there are still strongly endemic regions, primarily restricted by facies preferences. Earlier studies such as Williams (1973) identified American (mid-continent), circum-American, Baltic and Bohemian clusters, whereas Jaanusson (1973) noted his persistent Southern fauna together with American mid-continent and marginal Scoto-Appalachian faunas.
A palaeobiogeographical analysis of the early Katian (Trentonian) brachiopod faunas of Laurentia and its surrounding palaeoplates by Sohrabi & Jin (2013a, b) confirmed the existence of a unique Scoto-Appalachian fauna from the Sandbian to early Katian (see also Jaanusson & Bergström 1980); whereas the North American epicontinental brachiopod fauna, which started to become distinct during the early Katian, was most closely related to the shallow platform brachiopod fauna of Baltica (see also Harper & Hints 2001).
Some 70 localities have been identified globally for this stage containing some 240 genera. Two sites in the Kazakh arc have diversities approaching 50 whereas the majority of sites have diversities ranging from one to almost 25 genera. One genus, Sowerbyella, occurs at 25 sites whereas the majority of genera occur at between one and four sites. The standardized diversity for the Lower Katian interval is 3.43 (Figs 11.10 & 11.11).
During the Late Ordovician, the orthern path of South China is reflected by changing faunal affinities, with increasing linkages to Kazakhstanian terranes developing from the Sandbian, and strong affinities to eastern Gondwana (New South Wales) faunas becoming evident by the mid to late Katian as South China intersected migration pathways defined by surface currents. Also, throughout this interval, the Chu–Ili Terrane of Kazakhstan stands out as a cradle for a number of biogeographically significant genera, with the Chingiz–Tarbagatai, Boshchekul and Selety terranes providing secondary centres of origination (Percival et al. 2011).
Palaeogeographical map for the Lower Katian time slice based on BugPlates and using a Mollweide projection. The individual localities are listed in the supplementary material.
Cluster analysis for a selection of faunas for the Lower Katian time slice using the Raup–Crick (Hammer & Harper 2006) similarity coefficient and a neighbour-joining algorithm.
Upper Katian
This interval, essentially the pre-Hirnantian Ashgill, is marked by widespread development of carbonate mudmound facies and the marked migration of tropical faunas into the temperate and higher-latitude climatic belts (Boucot et al. 2003). Williams (1973) recognized a diversity of provinces, including North European, North American and Bohemian clusters. Jaanusson (1973), in addition to his Early Katian provinces, added the Hiberno-Salairian fauna that essentially included the belt of mudmounds which stretched at this time across northern Europe. Some 70 localities have been identified globally for this interval containing nearly 250 genera. Four sites, one in southeastern South China (Rong & Zhan 2004), one in Avalonia and two on the Laurentian margin (e.g. Harper 2006b), have diversities exceeding 55 genera whereas the majority of sites have diversities ranging from one to 20 genera. One genus, Christiania, occurs at over 20 sites. The standardized diversity for the Upper Katian is 3.49, a slight increase from levels in the Lower Katian. A recent analysis by Rasmussen et al. (2012) indicated that later Katian faunas had a more definite cosmopolitan stamp, although tropical shelf margins, tropical platforms and associated terranes, together with Gondwana and adjacent terranes, are still clearly differentiated (Rasmussen et al. 2012, fig. 6).
The late Katian was also a time when cradles for many of the post-extinction forms were being developed. Particularly, early virgianid brachiopods, such as Brevilamnulella, appear to have evolved from deep-water mound settings at lower latitudes, such as the Boda Mounds in Sweden, to become very important in the aftermath of the extinctions (Wright & Rong 2008; Rasmussen et al. 2010). Thus, the basic characters of Brevilamnulella probably gave rise to the rest of the pentameroid group, which became very successful in shallow-water settings. At the same time, the clorindoids – possibly also derived from a Brevilamnulella-like ancestor closely related to Clorilamnulella – maintained a deeper water preference. Another virgianid of late Katian age, Deloprosopus (very similar to Tcherskidium) was evolved in southeastern South China, and may be the local ancestor for the Silurian pentameroids in South China (Jin et al. 2006). Thus, habitats like the Boda Mounds seem to have been of profound importance as cradles and subsequently species pumps for opportunistic taxa that were to play a pivotal role in restoring the biodiversity levels in the wake of the Hirnantian extinctions.
Pronounced brachiopod provincialism is well known for the middle–late Katian (e.g. Sheehan & Coorough 1990). This is well demonstrated by the newly evolved rhynchonellide fauna of Laurentia during this time interval. For example, the ubiquitous North American brachiopod fauna represented by Hiscobeccus, Lepidocyclus and Hypsiptycha in the Maysvillian–Richmondian was virtually confined to Laurentia, except for a few isolated, putative occurrences elsewhere (see Jin 1996, 2001; Rasmussen 2013; Sohrabi & Jin 2013a, b; Figs 11.12 & 11.13).
Palaeogeographical map for the Upper Katian time slice based on BugPlates and using a Mollweide projection. The individual localities are listed in the supplementary material.
Cluster analysis for a selection of faunas for the Upper Katian time slice using the Raup–Crick (Hammer & Harper 2006) similarity coefficient and a neighbour-joining algorithm.
Hirnantian
Since recognition of the widespread distribution of the Hirnantia brachiopod fauna in the mid-1960s (e.g. Temple 1965; Marek & Havlíček 1967; Bergström 1968; Wright 1968), there have been a number of analyses of the global distributions of the Hirnantian brachiopod fauna (e.g. Rong 1984; Rong & Harper 1988; Harper et al. 2013). There is a clear division between the low-latitude Edgewood Province (Amsden 1974) and those of the higher-latitude Bani and Kosov provinces dominated by variants of the Hirnantia fauna, although clear elements of the Edgewood province are present in, for example, Estonia and the Oslo Region, arriving to colonize suitable warm-water carbonate facies. The differentiation between the Bani and Kosov provinces is more subtle, the former characterized by low-diversity faunas and differences at the species level from comparable taxa in the more-diverse Kosov Province (Villas et al. 1999). Some 50 localities have been identified globally for this stage, containing over 140 genera. One site, Estonia, has a diversity approaching 40 genera whereas the remaining sites have diversities ranging from one to 25 genera. Seven key genera, Cliftonia, Dalmanella, Eostropheodonta, Hindella, Hirnantia, Leptaena and Plectothyrella, each occur at over 15–25 sites. The standardized diversity for the Hirnantian Stage is 2.8; a drop from levels in the Upper Katian.
Rasmussen & Harper (2011b) pinpointed a number of regions and settings that acted as safe havens for various groups of brachiopods. The Yangtze Platform, for instance, played a pivotal role in preserving the mid-shelf stocks, whereas the Oslo Region in Baltica apparently housed an unusually large number of deep-water taxa during the crisis interval. This may explain why a large number of holdover or Lazarus taxa are found in these regions during the earliest Silurian Rhuddanian Stage (Baarli & Harper 1986; Rong et al. 2006; Figs 11.14 & 11.15).
Palaeogeographical map for the Hirnantian time slice based on BugPlates and using a Mollweide projection. The individual localities are listed in the supplementary material.
Cluster analysis for a selection of faunas for the Hirnantian time slice using the Raup–Crick (Hammer & Harper 2006) similarity coefficient and a neighbour-joining algorithm.
Brachiopod biodiversity and phylogeography during the Ordovician
The rhynchonelliformean brachiopods comprise a major constituent of the GOBE (Harper 2006a). Their diversity through the period has previously been assessed in some detail (e.g. Harper et al. 2004 and references therein). The current study, based on a new, much more comprehensive, dataset, roughly mirrors the diversity curve produced by previous studies (Fig. 11.16a), although the current dataset reveals an overall higher diversity level throughout the period, possibly reflecting the more recent intense sampling and research being conducted within this field.
(a) Ordovician generic brachiopod diversity and phylogeographical distribution through the Ordovician Period. Genera are grouped into superfamilies, and in some cases higher groupings. Related phylogenetic groups share the same colour shades. These shades are also used in the phylogeographical time-slice maps (b–d). Radial colourings indicate dominance of each superfamily in a given region, going from least dominant (centre of circles) to most dominant (outermost part of circles). Abbreviations: Dap., Dapingian; Hir., Hirnantian. See legend for names of superfamilies not written on the biodiversity graph.
The Ordovician Period commenced with a relatively slow, but steady increase in generic diversity which continued through the Floian Stage and into the Dapingian. The persistence and slight diversification of Cambrian stocks were responsible for this slow rise in diversity, which occurred despite the demise of the billingselloids. It was primarily driven by radiations within the Orthoidea and Porambonitoidea. However, from the Floian onwards, the strophomenides gradually became more and more important, notably as the Plectambonitoidea was established. The phylogeographical map for the Tremadocian Stage (Fig. 11.16b) shows that the orthoids were present as a dominant constituent of the faunas in all regions at this time, notably, around high- and low-latitude Gondwana, as well as on the more southerly located continents of Avalonia and Baltica. Adjacent to Gondwana, Perunica hosted a somewhat more phylogenetically diverse fauna where orthoids were associated with billingsellids. In the Northern Hemisphere, however, the faunas were more complex. The South China Palaeoplate seems to have been a cradle for many clades, such as the clitambonitoids, which had a strong component together with porambonitoid taxa mixed with orthoids. Moreover, some of the earliest occurrences of the strophomenide clades were also present, whose cradles seem to have been located in equatorial settings. Porambonitoid and clitambonitoid genera are also found in parts of Gondwana (Tasmania) at this time. Along with the Orthoidea, the porambonitoids were important constituents throughout lower latitudes, commonly along with the billingselloids, or, as seen on the Salair Terrane and in Laurentia, together with plectambonitoids. In Laurentia, the pentameride taxa, again characterized by the porambonitoids, dominated the faunas, with the orthoids being less important. Billingselloids were also present as a minor part of the fauna. Laurentia and South China are apparently two of the cradles for the Strophomenoidea and some of the main biodiversity epicentres for the subsequent radiation of the superfamily. In Salair and South China, the plectambonitoids were firmly established in otherwise porambonitoid-dominated low-latitude regions.
This earliest Ordovician phylogeographical distribution continued into the Floian. The porambonitoids continued to radiate and thus reached a generic diversity almost at the same level as the orthoids. However, this diversification ceased within the Dapingian when one-third of the porambonitoid stocks disappeared. However, the Orthoidea continued to radiate. In addition the Clitambonitoidea began to radiate. However, the trajectories of the orthoids and porambonitoids were subsequently modified by extinctions within the early Darriwilian. The overall generic diversity reached a plateau, which primarily was caused by a generic loss within these two superfamilies. This was an interval when only the clitambonitoid and strophomenide clades displayed a net diversification. Although the strophomenoids continued to radiate, they could not compensate for the major loss of taxa within the porambonitoids and the orthides, as well as the demise of the Billingselloidea, which went extinct by the Dapingian. Based on the phylogeographical maps for the Tremadocian (Fig. 11.16b) and Dapingian–Darriwilian (Fig. 11.16c) intervals, it is evident that there was a drastic change in the phylogeographical composition in especially the low-latitude regions. The porambonitoids clearly suffered a large loss on Laurentia, but the demise of the orthoids is somewhat more difficult to locate geographically.
The Darriwilian extinctions apparently helped pave the way for an explosion in strophomenide clades. In many of the major extinctions clearing ecospace was clearly important for some subsequent radiations. The Plectambonitoidea and Strophomenoidea doubled their generic diversity by the mid Darriwilian and this continued through the rest of the Darriwilian, a trend that was mirrored by the dalmanelloids, a plexus that constituted the backbone of the brachiopod radiation during the GOBE. By the late Darriwilian more and more clades increased in generic diversity and geographical distribution, probably derived from low- to mid-latitude species pumps situated on the more diverse habitats of the larger continental palaeoplates, such as Baltica (see Rasmussen et al. 2007), but also on the emerging intra-oceanic palaeoplates. The Dapingian–Darriwilian phylogeographical map clearly demonstrates the shift from orthoid/porambonitoid dominated Tremadocian world (dark orange and purple colours) to a strophomenoid-dominated Mid Ordovician brachiopod fauna (green colours). Although orthide and pentameride losses were stabilized, mostly owing to the ever more successful dalmanelloids and camerelloids, it was the strophomenides that underpinned the main diversity peak. These taxa were now the main constituents in most regions, except for mid- and high-latitude Gondwana. On the larger continents, particularly Avalonia and Laurentia, orthoid biodiversity epicentres persistently dominated along with the new clades. Cradles for rhynchonellides and triplesiids, for example, were almost exclusively located on Avalonia and Laurentia as important parts of the overall phylogenetic composition of these continents. The clitambonitoids reached the acme of their distribution, constituting large proportions of the genera in Avalonia and Baltica and possibly acting as species pumps for the neighbouring regions, where the clitambonitoids to a lesser extent were recruited within the remainder of low- to mid-latitude regions. The dalmanelloids, in contrast, seem to have been quite successful in high-latitude Gondwana, which may have acted as a species pump, possibly facilitating their spread to the nearby palaeoplates of Avalonia and South China and then farther north to the Northern Precordilleran terrane and Perunica, as well as the intra-Iapetus terranes, such as the Leinster and Midland Valley terranes. Baltica and the North China Palaeoplate, however, were not characterized by the same dalmanelloid radiation during the Darriwilian. The most phylogenetically diverse regions were Avalonia, Laurentia and South China, but particularly the intra-Iapetus island arcs and microcontinents also held a wide range of clades, many carrying the new dominant strophomenide and clitambonitoid groups mixed with the older orthoid and porambonitoid stocks.
The GOBE reached its plateau during the early Sandbian, mainly constructed by the overwhelmingly successful strophomenides. However, a number of new clades contributed to this diversity spike, together with the older Cambrian stocks, such as the orthides, with radiations within all major clades, recovering from the early Darriwilian extinctions. Within the previously successful pentamerids, the camerelloids also managed to stage a notable recovery. By the earliest Sandbian, rhynchonelliformean diversity constituted more than 200 genera, of which about 80% belonged to either the Orthida or Strophomenida. This expansion in biodiversity continued well into the Katian, with some fluctuations, before one final hike during the latest Katian. This was bolstered by a final radiation of the orthoids and strophomenides, new pentameroids, and the atrypoids.
The Katian phylogeographical map (Fig. 11.16d) clearly shows this wealth of new clades, which occupied both the equatorial and boreal belts. By the late Katian, Baltica, with more than 20 superfamilies present, surpassed Laurentia and South China as the most phylogenetically diverse region. Strophomenoids, accounting for more than 40 genera, were overwhelmingly dominant with more than twice as many genera than the second largest group, the plectambonitoids. Compared with the Darriwilian map for Baltica, the atrypides, rhynchotrematoids and pentamerides radiated, whereas clitambonitoids and orthoids became less important. This appears to have been the more general picture globally. Although still dominated by strophomenides, many other clades gained a permanent foothold around the larger continents, which thus acted as large biodiversity epicentres and subsequent species pumps. This included new pentameroid taxa, the atrypids and athyridids, as well as a whole suite of other clades. Orthoid-dominated regions were now exclusively located in low-diversity, high latitudes, alongside the dalmanelloid and strophomenoid clades that remained; these were cosmopolitan at this time. Avalonia was still, with 16 superfamilies, a very phylogenetically diverse region, as elsewhere, mostly dominated by the strophomenoids and plectambonitoids together with dalmanelloids and orthoids. The eastern part of the continent had a somewhat different phylogenetic signature, lacking rhynchotrematoids and meristelloids.
In Laurentia, the plectorthoids were nearly as diverse as the dalmanelloids and orthoids. In addition, rhynchotrematoids, camerelloids and pentameroids were similarly diverse. For the Katian map, the phylogenies for the Anticosti Basin, Newfoundland and the Franklinian Basin are also shown, to give an impression of which clades were present on the Laurentian margins. Whereas the Franklinian fauna clearly demonstrates a strong pentameroid component (purple), within the strophomenide and orthide dominated fauna, the more southerly basins have only rare pentameroids, instead being dominated by rhynchotrematoids, atrypids and athyridids.
Off the western margin of Laurentia, the Klamath terrane was heavily dominated by the Orthida with many orthoids and plectorthoids, together with some dalmanelloids and enteletoids. Plectambonitoidea was less dominant here than was usual for this interval. Thus, in general, this distribution is likely to reflect temperature, with the pentameroids preferring the tropical settings, and other clades being more adaptable to temperate regions. Interestingly, the Eastern Klamath terrane appears to have acted as a museum, retaining clades that were dominant earlier in the Ordovician.
In the Iapetus Ocean, terranes such as the Midland Valley Terrane exhibit a substantial phylogenetical diversity, possibly conveying clades from either side of the extremely diverse surrounding continents, or providing stepping stones to migrations. The Midland Valley Terrane exhibits an unusually high number of remnants from the earlier Ordovician (Clitambonitoidea) and a greater orthide component dominated by Orthoidea than seen elsewhere. However, these were mixed with new clades, such as the Rhynchotrematoidea and Triplesioidea, which also were unusually diverse here. Also several atrypid and athyridid superfamilies were present, but only few pentamerides (Camerelloidea).
The regions that were the most conspicuous, during this time slice, were the Farewell and Kolyma terranes, with their unusually high concentration of pentameride stocks (Camerelloidea and Pentameroidea). Notably, for the Farewell Terrane, Anazygoidea was also unusually well represented. Plectambonitoidea, however, was the most dominant group, although with only a few more genera than the pentameroids. Orthoids were not particularly important and strophomenoids were represented by only one genus. Kolyma also had a very unusual phylogenetic composition. It was almost a ‘Silurian fauna’, being dominated by pentameroids, atrypids and athyridids and with a total absence of orthides. In Siberia, an odd mix of strophomenides, rhynchotrematoids and atrypids co-occurred with a few orthoids and plectorthoids. Dalmanelloids were lacking – to some extent this composition is also seen in the nearby, very phylogenetically diverse, Altai Terrane. Here, however, the camerelloids and pentameroids were also important constituents. Interestingly, on both the Siberia and Altai terranes, dalmanelloids seem to have been rare, having been replaced by orthoid and plectorthoid clades. In the Taimyr region of Siberia, orthoids were also rare, with dalmanelloids, the most common orthides, marked by only four genera – the same number as both the plectambonitoids and strophomenoids. Atrypoids were unusually diverse, as were the pentameroids and triplesioids.
Farther east from the Altai terrane, the Kazakh terranes, along with the North China Palaeoplate, were dominated by the strophomenides. In particular, the Tien–Shan terrane was dominated by plectambonitoids, with atrypoids and strophomenoids as the second, most-diverse constituents. Pentameroids were also important but only one orthoid genus was present. The Chu–Ili terrane displayed quite different phylogenies, including large pentameroid, meristelloid and enteletoid components mixed with the plectambonitoids. In between the Tien–Shan and Chu–Ili terranes, the Chinghiz Terrane was very diverse, phylogenetically, with many clades derived from the nearby North China Palaeoplate.
During the rest of the Ordovician Period, the South China Palaeoplate was one of the most diverse regions containing many unusual stocks and notably a high number of atrypoid genera represented by the Altaethyrella fauna (a regional brachiopod fauna of late Katian age mainly in South China, North China and southern Kazakhstan terranes; Rong & Zhan 2004).
In Gondwana, the low-latitude, equatorial regions stand out as being most phylogenetically diverse. These regions were strongly dominated by the Strophomenida, but with a rare large component of atrypides, camerelloids and rhynchotrematoids, although possibly linked to the siliciclastic facies dominant here, not as important constituents, as seen elsewhere in the tropics. Plectorthoids are the most dominant orthides. In high-latitude Gondwana, Orthoidea and Dalmanelloidea still dominated, whereas strophomenoids and triplesiids were minor constituents and some camerelloids, porambonitoids and plectorthoids were present only regionally.
Following this late Katian explosion in numbers of clades, the Hirnantian extinctions excised a significant number of the ancient stocks, clearing ecospace for new clades, such as the pentameroids and the atrypids, to expand rapidly during the subsequent Llandovery. Having mostly been located on relatively isolated intra-oceanic, low-latitude terranes during the latest Katian, these clades suddenly were able to gain a foothold on, for instance, the Laurentian margins once the recumbent faunas had been stressed, if not removed entirely (Sheehan 1975; Rasmussen & Harper 2011a, b).
Conclusions
Low global diversities of Tremadocian and Floian faunas were typical: the main biodiversity epicentres were located in South China, where the first spike of the GOBE manifested in the lower Floian.
Bursts of diversity within the Dapingian–Darriwilian interval were associated with dispersed island arcs, microcontinents and continents but retarded by subsequent regional extinctions in the early Darriwilian.
There was a mid-Darriwilian onset of a massive radiation within the strophomenide and dalmanelloid clades, constituting the main phylogenetical component of the GOBE. Main species pumps were located on Avalonia and Baltica.
Sandbian offshore and other migrations of faunas and the development of the Foliomena faunas occurred: many new clades had now emerged.
Katian development of cradles in carbonate mound environments contributed to migrations of fauna poleward. Particularly during the latest Katian, this caused a great expansion in pentameroid clades, as well as the atrypids and athyridids.
End-Ordovician (Hirnantian) glaciations coincided with well-defined, climatically controlled provinces and massive extinctions in the ‘Ordovician brachiopod fauna’.
Acknowledgments
David A. T. Harper, Christian M. Ø. Rasmussen thank the Danish Council for Independent Research, and Natural Sciences for financial support. Ian G. Percival publishes with permission of the Executive Director, NSW Trade and Investment, Department of Resources and Energy. J.-Y. Rong and R.-B. Zhan thank the National Natural Science Foundation of China (41221001, 41290260). This paper is also a contribution to the IGCP project 591 ‘Early to Middle Paleozoic Revolution’. We gratefully acknowledge the help of R. Cocks in providing us with the brachiopod data of Lees et al. (2002). Insightful and thoughtful reviews by L. Popov and P. Sheehan considerably improved the manuscript. M. McCorry helped consolidate our reference lists and the supplementary data. The reconstructions are based on BugPlates with some minor modifications.
- © The Geological Society of London 2013
References
Chapter 11 Biodiversity, biogeography and phylogeography of Ordovician rhynchonelliform brachiopods
