Ordovician conodont biogeography - reconsidered

Ordovician conodont biogeography ± reconsidered YONG-YI ZHEN AND IAN G. PERCIVAL Zhen, Y.Y. & Percival, I.G. 2003 12 12: Ordovician conodont biogeography ± reconsidered. Lethaia, Vol. 36, pp. 357±370. Oslo. ISSN 0024-1164. Review of the traditional separation of global Ordovician conodont distribution into the North American Midcontinent Province (NAMP) and North Atlantic Province (NAP) reveals a confusing variety of concepts and de®nitions that hinder biogeographic analysis. Use of this twofold scheme and its subsequent variants should be discontinued in favour of the more detailed divisions proposed here. Major biogeographical entities of the Shallow-Sea and Open-Sea Realms, separated by the shelf-slope break, are both further subdivisible into Tropical, Temperate and Cold Domains. In the Cold domains, faunal differences between the two Realms and their subdivisions are not easily discernible, since biofacies zones and different habitats were highly condensed. Faunal differences are ampli®ed in the tropical regions, where the North American Midcontinent Province and North Atlantic Province were originally de®ned. Recognition of endemic taxa is essential for ®ner classi®cation within domains of the Shallow-Sea Realm (SSR). Our preliminary analysis of Early Ordovician conodont distribution identi®es the Laurentian Province (in the Tropical Domain), Australian (Tropical Domain), North China (Tropical Domain), South China (Temperate Domain), Argentine Precordillera (Temperate Domain) and Balto-Scandian Province (in the Cold Domain). The Open-Sea Realm (OSR) is dominated by cosmopolitan and widespread taxa, and formal subdivision at provincial level is yet to be achieved. The North Atlantic Province encompasses both the Open-Sea Realm and the Temperate and Cold Domains of the Shallow-Sea Realm. The North American Midcontinent Province sensu stricto is more or less equivalent to the Laurentian Province, representing shallow-water regions fringing Laurentia; in a broader sense the North American Midcontinent Province includes all provinces of the Tropical Domain within the ShallowSea Realm. & Biogeography, conodonts, Domain, ecology, Ordovician, Province, Realm. Yong-Yi Zhen [[email protected]], Division of Earth & Environmental Sciences, The Australian Museum, 6 College Street, Sydney N.S.W. 2010, Australia; Ian G. Percival [[email protected]], Specialist Geological Services, Geological Survey of New South Wales, P.O. Box 76, Lidcombe N.S.W. 2141, Australia; 13th August 2002, revised 29th August 2003.

Biogeography is the study of distributional patterns of organisms at various levels (e.g. communities, populations, a group of taxa, or even a single species), enabling analysis of the origins, evolutionary histories and major controlling mechanisms of such patterns through the history of life. It is a multi-disciplined science, interacting between biology, geology, ecology and physical geography. Of the three major schools (ecological, historical and phenetic) in the ®eld of biogeography, interpretation of Ordovician conodont biogeography has to date been overwhelmingly dominated by ecological concepts. We propose here a new scheme identifying conodont provinces, which is well founded on principles inherent in modern ocean biogeographic studies. It recognizes ®rst-order and second-order divisions which are de®ned by ocean hydrologic and climatic factors (an ecobiogeographical approach), whereas lower hierarchical levels (provinces) are de®ned by distribution of endemic taxa (central to historical

biogeographic concepts). Recognition of endemic taxa and their distribution forms the core of the new Ordovician conodont biogeography. Cosmopolitan taxa, which have tended to dominate conodont research due to their biostratigraphic applications, are not informative biogeographically, and should be eliminated from any such analysis. A review of the development of models of Ordovician conodont distribution over the past four decades reveals that only a few studies have used recognition of endemic taxa as the basis for identi®cation of biogeographical provinces. The scheme we advocate applies a hierarchical scheme of biogeography to Ordovician conodont distribution by analogy with the distributional patterns of extant marine benthic, nektobethic and pelagic organisms. We strongly believe that the modern ocean biogeographical model is applicable to a revised interpretation of the distribution and relationships of conodont faunas in the Ordovician, although the con®guration of lands DOI 10.1080/00241160310006402 # 2003 Taylor & Francis

358

Yong-Yi Zhen & Ian G. Percival

and seas, and circulation systems in the Ordovician oceans and atmosphere, were remarkably different (Wilde 1991; Christiansen & Stouge 1999; Wright et al. 2002).

Ordovician conodont biogeography: the traditional model Recognition by Sweet et al. (1959) of temperaturedependent conodont faunas in Upper Ordovician platform successions of North America enabled differentiation of the North American Midcontinent Province (NAMP), characterized by a warm water fauna, from the Anglo-Scandinavian-Appalachian Province, including cold-temperate faunas. The latter became the North Atlantic Province (NAP), with identi®cation of conodont faunas from the BaltoScandinavian region and eastern North American offshore successions. This twofold biogeographical faunal division was con®rmed by subsequent conodont workers (BergstroÈm & Sweet 1966; LindstroÈm 1970; BergstroÈm 1971; Barnes et al. 1973; Fortey & Barnes 1977; Dzik 1983; Ethington & Repetski 1984; Sweet & BergstroÈm 1984; Bagnoli & Stouge 1991, 1997; Pyle & Barnes 2002). BergstroÈm & Sweet (1966) indicated that water temperature and depth both played a signi®cant role in distinguishing the two provinces. Speci®cally, the North American Midcontinent Province was restricted to warm water faunas within the equatorial regions, and the North Atlantic Province was characterized by cold water faunas in high latitudinal regions. Barnes & FaÊhraeus (1975) interpreted the North American Midcontinent Province as constrained within an epeiric sea with raised temperature and salinity, which was geographically restricted to a fairly narrow equatorial belt, whereas North Atlantic faunas (dominated by cosmopolitan elements) inhabited normal marine conditions. They also regarded temperature and salinity as the main factors controlling the spatial distribution of conodont provincialism. Other authors (e.g. LindstroÈm 1976) suggested the North Atlantic Province represented cold-water environments, or cold-temperate water habitats (Bagnoli & Stouge 1991). Miller (1984) also viewed temperature as the primary distinction between the two faunal groupings, and informally termed these the warm and cold faunal realms; more recently, Murray & Stewart (2001) have formalized these terms with the acronyms WFR and CFR. Based on studies of conodonts from deep-water siliceous-carbonate successions in Kazakhstan and central Asia, Dubinina (1991, 1998) identi®ed a third biogeographic unit in

LETHAIA 36 (2003)

the Early Ordovician, additional to those of Miller (1984). On palaeoenvironmental criteria, Dubinina recognized (1) a cold faunal realm to include faunas (both shallow and deep water) from high palaeolatitudes and deep-water faunas from the middle palaeolatitudes (=North Atlantic Province), (2) a warm faunal realm as including faunas from low-mid latitude shallow-water seas (=North American Midcontinent Province), and (3) a transitional realm as an entity encompassing shelf margins and continental slopes in the low-mid latitudes (=Appalachian Province of Pohler & Barnes 1990). Dzik (1983) equated the North American Midcontient Province with Ordovician epicontinental seas of Laurentia, also de®ning the Baltica Province to include faunas in the Baltic region, and interpreting the North Atlantic Province as a third transitional region associated with islands in the Iapetus Ocean between the other two provinces. Faunas with North American Midcontinent Province af®nities have been identi®ed from eastern Siberia (Moskalenko 1972, 1983), North China (An et al. 1983; Wang et al. 1996), in the cratonic region of central and western Australia, and in carbonate shelf deposits surrounding offshore islands, which are now incorporated in eastern Australian fold belts (Webby et al. 2000; Murray & Stewart 2001). It is widely accepted that these North American Midcontinent Province faunas inhabited shallow, warm-water environments, whereas conodont faunas assigned to the North Atlantic Province typically occur in high latitudes, in temperate, deeper water settings, and in low latitude offshore deep-water regions of the Ordovician oceans. Faunas with North Atlantic Province af®nities were recognized from the Argentine Precordillera (Serpagli 1974; Lehnert 1995; Albanesi et al. 1998), NW Argentina (Rao 1999; Albanesi et al. 2001), South China (Wang et al. 1996), Kazakhstan (Dubinina 1991, 1998; Zhylkaidarov 1998), and within the fold belts of eastern Australia in slope or basinal sediments (Webby et al. 2000; Murray & Stewart 2001; Zhen et al. 2003a).

Limitations of the traditional model As the foregoing brief review shows, the traditional twofold division (and subsequent tripartite variants) of global Ordovician conodont distribution into temperature-controlled and depth-related `provinces' or realms based on ecological biogeographical concepts was initially developed prior to the widespread acceptance of plate tectonic theory. The North American Midcontinent Province (=WFR) and North Atlantic Province (=CFR) were de®ned primarily with

LETHAIA 36 (2003)

respect to the North America and European regions because, until relatively recently, there was very little information available from other shallow-water tropical regions like Australia and North China. There is no dispute that the broad-scale faunal patterns designated as North American Midcontinent Province and North Atlantic Province do exist ± indeed, they form the basis of two parallel, widespread conodont biozonations throughout the Ordovician ± but do they have any signi®cance biogeographically? Dzik (1983, p. 60) noted that `one basic trouble in establishing a good scienti®c framework for a discussion of conodont provincialism is the lack of unequivocal meaning for units in paleobiogeography'. Biogeographical af®nities de®ning provinces (or realms) in the traditional model were mainly based on overall faunal similarities, which work well for biostratigraphic correlation purposes but fail to identify areally-restricted faunas. Hence the established models of Ordovician conodont distribution tend to be inconsistent in detail with our current understanding of relationships among different landmasses or major tectonic plates and terranes, derived from biogeographical studies of other contemporaneous faunas. Indeed, when plotted on Ordovician global reconstruction maps, the North Atlantic Province and North American Midcontinent Province faunal groups are seemingly represented in almost every major continental block, sometimes ± as is the case with island faunas in the Lachlan Orogen of eastern Australia ± being situated immediately adjacent to each other with disjunct North American Midcontinent Province (=WFR) assemblages in shallow marine limestone successions apparently surrounded by North Atlantic Province (=CFR) faunas from deeper water siliciclastics (Murray & Stewart 2001). Similarly, in a study of Ordovician conodonts from the southern Mackenzie Mountains, Canada, Tipnis et al. (1978) noted that platform carbonates yielded faunal elements almost entirely characteristic of the North American Midcontinent Province, whereas the clastic facies of adjacent shelf margin or continental slope environments were dominated by typical North Atlantic Province taxa. Spatial and temporal mixing of the faunas representing these two biogeographic divisions is also widely recorded. When establishing the two major conodont provinces, Sweet et al. (1959) noted the existence of a mixed `Anglo-Scandinavian-Midcontinent assemblage' in the Upper Ordovician of the Cincinnati region and regarded it as resulting from migration of Anglo-Scandinavian and Appalachian conodonts into the North American Midcontinent region. In many cases faunas representing the North American Midcontinent Province and North Atlantic

Ordovician conodont biogeography

359

Province are found inter®ngering in successions associated with transgressive-regressive cycles which are developed along outer-shelf to shelf margins. In Newfoundland, mainly North American Midcontinent Province faunas with periodic strong incursions of deeper water elements (presumably within those sediments representing deepening intervals) were recorded from the shallow water St. George Group (Ji & Barnes 1994a, b). Conodonts from the contemporaneous Cow Head Group (deeper slope environment) have mixed North American Midcontinent Province and North Atlantic Province af®nities (according to Pohler 1994) interpreted as due to intermingling of in situ slope sediments and allochthonous clasts derived from the shallow shelf area. These spatial replacements and overlapping or inter®ngering of deep and shallow-water faunas cannot, however, be interpreted in biogeographic terms like in¯ux or migration, as such replacements merely re¯ect biofacies variations caused by tectonics and eustatic ¯uctuations.

Problems with provincial concepts By focusing on faunal similarities, the traditional model has provided suf®cient data to recognize the ®rst-order division of biogeography (more or less equivalent to the Shallow-Sea Realm and Open-Sea Realm of the modern oceans). Several de®nitions of these two realms have been proposed (Ethington & Repetski 1984; Pohler & Barnes 1990; Ethington et al. 1995), being based variously on biofacies, inferred habitats, palaeoecology and faunal composition. As currently de®ned, the two major conodont provinces (North American Midcontinent Province and North Atlantic Province) more accurately represent ecoregions (or ecosystems), which differ mainly in water depth and temperature. In low latitudes, the boundary between the North American Midcontinent Province and North Atlantic Province coincides with that separating the ShallowSea and Open-Sea Realms in the terminology of modern ocean biogeography, essentially coincidental with the shelf-slope break, where a sharp drop in water temperature along the shelf margin is indicated (cf. Gage & Tyler 1991, p. 13, ®g. 2.5). This steep gradient becomes less prominent towards high latitudes, which explains why Early Ordovician shallow and deeper water conodont faunas from Balto-Scandia cannot be as easily differentiated (Rasmussen 2001) as those from Laurentia and Australia, which were situated in low latitudinal positions during this time. Faunas from a shallow-water shelf in the tropical regions can be regarded as belonging to the North American

360

Yong-Yi Zhen & Ian G. Percival

Midcontinent Province, whereas contemporaneous faunas from the deeper and cooler water of the shelf edge or slope settings, and faunas from cold water (both deep and shallow) in the high-latitude regions would both be assigned to the North Atlantic Province (Dubinina 1991, ®g. 6). In tropical regions like eastern North America in Ordovician time, the boundary between these two provinces coincides with the facies boundary between limestones and more shaly deposits (Ethington et al. 1995, ®g. 14), whereas in the highlatitude regions of the Ordovician this boundary becomes much less recognizable. Obviously, such a distribution of North American Midcontinent Province and North Atlantic Province faunas does not re¯ect a biogeographical distribution consistent with plate tectonics-based reconstructions of the Ordovician world. In reference to the modern oceans, the North American Midcontinent Province would be equivalent to shallow-water regions (depth of less than 200 m) in low-mid latitudes with water temperature higher than 10°C. The distribution of the North Atlantic Province faunas is more complex, comparable with that of the Open-Sea Realm plus the Temperate and Polar domains of the Shallow-Sea Realm of the modern oceans (Cox & Moore 2000). The term `province' has long been employed by conodont workers in an ecological sense to group faunas that inhabited similar environments. In emphasizing relationships between different faunas and the environments they inhabited, and in further using the biofacies concept to de®ne biogeographic units (realms, domains, provinces, subprovinces, areas), the currently accepted Ordovician conodont biogeographic model does not distinguish ecological units from truly biogeographic units. Although both deal with the distribution of taxa, associations or faunas, studies on conodont biogeography and biofacies have fundamentally different objectives. The essential question for the former is how different are the faunas from three or more regions, which were geographically separated from each other, and why are they different? Biofacies studies focus on interrelationships between faunal associations and the environments (typically con®ned to a smaller area showing environmental gradients) in which they lived. In their analysis of Ordovician conodont provincialism, Barnes & FaÊhraeus (1975, p. 135), and Pohler & Barnes (1990) followed the de®nition advocated by Valentine (1968, p. 257), who de®ned provinces as representing `collections of communities that are associated in space and time'. However, this is an ecological concept, more or less equivalent to biota (which is now much more widely used), and has little to do with provincialism in the biogeographic sense;

LETHAIA 36 (2003)

thus Valentine's usage of province in a biogeographic context is inappropriate. In the light of their often equivocal usage, we recommend abandoning the two traditional terms, North American Midcontinent Province and North Atlantic Province, in favour of the more consistent and detailed biogeographic scheme proposed here. Confusion resulting from this variation in de®nition of the term `province' is not restricted to conodont specialists, but extends to biogeographers attempting to make sense of the distribution of continental blocks and surrounding oceans in the Ordovician world. Clari®cation of this problem, by (1) simpli®cation of terminology and (2) revision of the concept, provides the impetus for our re-evaluation of Ordovician conodont biogeography. Comparison of faunas dominated by cosmopolitan forms populating different biogeographical provinces within different domains or even realms will give a high (though misleading) similarity index. We contend that provincialism can only be accurately assessed through analysis of endemism, rather than general faunal similarities, and hence cosmopolitan forms must be discounted (cf. Rasmussen 1998). Dispersal, phylogeny, spatial and temporal distributional patterns of endemic elements are the most important factors assisting de®nition of biogeographic divisions. Therefore, an integrated approach of ecological and historical methods should be used in reconstructing conodont biogeography. Whereas cosmopolitan taxa are crucial for biostratigraphic correlation, recognition of endemic taxa and their restricted distribution is the key to reconstruction of biogeographic provincial patterns. Endemic taxa are therefore the characteristic elements of a biogeographical province, which is de®ned as a region with coincident occurrence of relatively larger numbers of well-differentiated endemic elements (Brown & Lomolino 1999).

Endemic taxa and province recognition Although recognition of Ordovician conodont provinces de®ned on distribution of endemic species dates back some 30 years, such concepts have (in our opinion) rarely received the acceptance they deserve. BergstroÈm (1971) speculated on the existence of an Australian conodont province, and subsequently (BergstroÈm 1990) identi®ed a separate North China Province within the North American Midcontinent Realm during the Early Ordovician, based on a Serratognathus-type fauna. Bagnoli & Stouge (1991) broadened this concept to invoke a China province

LETHAIA 36 (2003)

de®ned by the presence of Serratognathus in early Arenig (mainly pre-evae conodont Zone) successions. An early attempt to analyse endemicity (using a punch-card data®le) in Palaeozoic conodont faunas (Charpentier 1984) proceeded along correct theoretical lines, but was ¯awed through use of form-species concepts. By applying a similarity coef®cient analysis to a large database of conodont occurrences through the Late Cambrian to Late Silurian, BergstroÈm (1990) was the ®rst to directly relate distribution of multielement conodont taxa to plate reconstructions. Very low similarity indices (i.e. high endemism) between nearby continental blocks supported the existence of six conodont provinces in the Early Ordovician, namely those of the North American Interior, Mediterranean, North China, Siberian, Baltic and Australian provinces. Most of these (together with a British Province earlier recognized by Sweet & BergstroÈm 1984) continue into the Late Ordovician, although data to substantiate the presence of a separate province in the Australian region during the Caradoc±Ashgill interval were not available at this time. Co-occurrence of Taoqupognathus and Yaoxianognathus in the Upper Ordovician of Australia and China subsequently prompted Nowlan et al. (1997) to propose an Australasian Province. Sweet & BergstroÈm (1984) analysed the distribution of 43 conodont taxa in the warm water pelagic realm (equivalent to the North American Midcontinent Province) and delineated two clusters of indigenous species which they named the Red River Province and Ohio Valley Province, based on their occurrence in the Late Ordovician of midcontinental Laurentia. Rasmussen (1998) utilized statistical methods similar to BergstroÈm's analysis to interpret Ordovician conodont biogeography, but correctly pointed out that inclusion of cosmopolitan taxa in the data-set would lead to distorted results. He showed that deepwater faunas from the Iapetus Ocean were dominated by cosmopolitan species; if these were excluded from the analysis, faunas from shelf margins and slopes of the Laurentian margin exhibited more species in common with the North American Midcontinent Province than the North Atlantic Province.

Ecology and major causes of provincialism Interpretation of conodont biogeography is also heavily dependent on an understanding of the life mode of conodont animals. Geographical barriers, such as deep oceans, are generally regarded as the primary causes of provincialism for marine benthic

Ordovician conodont biogeography

361

Fig. 1. Hypothetic ecological model for Ordovician conodonts (modi®ed from Seddon & Sweet (1971) and Barnes & FaÊhraeus (1975)), not to scale, vertically exaggerated.

and nektobenthic organisms inhabiting different shelf regions connected through open ocean surface waters. Pelagic forms were mostly cosmopolitan, and the only signi®cant barrier for them would be ecological ± the surface water temperature. Of the two ecological models (Fig. 1) proposed to explain the distribution of Ordovician conodonts, that of Seddon & Sweet (1971) basically supports MuÈller's (1962) conclusion that (like modern-day chaetognaths) conodont animals were pelagic, with different species segregated by vertical strati®cation. This model best explains the distribution of cosmopolitan or widely distributed conodont species. An alternative model proposed by Barnes & FaÊhraeus (1975) suggested that the majority of Ordovician conodonts were benthic or nektobenthic, and only simple cone genera, like Panderodus, Drepanoistodus, Drepanodus, Paroistodus and Paltodus, were pelagic forms. This latter model emphasizes lateral segregation of conodont faunas, regarding depth-related factors like temperature, salinity, circulation, energy, and substrate as the major constraints on their distribution, and is comparable with the seriative model for benthic or nektobenthic animals of Tipper (1980). Highly variable morphology and extremely high diversity of Ordovician conodonts strongly suggest that conodont animals at this time likely occupied most of the available marine niches in the Ordovician

362

Yong-Yi Zhen & Ian G. Percival

oceans and adopted a variety of life modes. We propose an ecological model combining attributes of the previous vertical and lateral strati®cation models to better explain the distributional patterns of pelagic and benthic or nektobenthic forms, respectively (Fig. 1). This model suggests that the shelfal carbonate facies are presumably dominated by shallow-water benthic (or nektobenthic) and epipelagic forms. Deeper water benthic or nektobenthic forms and mesopelagic forms occasionally found in distal platform sediments were mainly introduced by upwelling currents. Depositional regimes of the slope facies are much more complex, and often consist of allochthonous limestone debris derived from shelf carbonates intermingled with in situ siliciclastic deposits and minor thinly bedded carbonate lenses. Faunas from deeper water sediments or allochthonous clasts derived from shelf margin or upper slope facies are dominated by cosmopolitan and widespread species (Zhen et al. 2003a), and accordingly are recognized as belonging to the Open-Sea Realm (OSR). Some conodont faunas from shelf margin deposits in the Scandinavian Caledonides (Rasmussen 2001) might also be assigned to the Open-Sea Realm, particularly those recovered from limestone nodules and lenses in graptolitic shales such as the Lower Ordovician Tùyen Formation. They were low in diversity and are also dominated by cosmopolitan and widespread species. Most simple coniform taxa were cosmopolitan in distribution in both the Open-Sea Realm and the Shallow-Sea Realm (SSR) and were likely epipelagic, and comprise a much greater proportion of faunas recovered in samples from slope and deep basinal deposits. Widespread coniform taxa found only in the deeper water facies were likely mesopelagic, a typical example being Juanognathus variabilis in the Early Ordovician. The benthic or nektobenthic deeper water forms were relatively rarer, although generally widely distributed. Endemics in these environments are extremely rare. Most of the Ordovician deep-sea basinal deposits subsequently disappeared into subduction zones, and only remnants, composed of ®negrained, argillaceous and siliceous deposits, were fortuitously preserved in the fold belts. Conodont records from these deep-water deposits, mainly from cherts, have received relatively little attention thus far, but increasing reliance on their biostratigraphic utility will ensure that they become better known (e.g. Dubinina 1991, 1998; Murray & Stewart 2001; Tolmacheva et al. 2001; Lyons & Percival 2002). Circulation patterns of Ordovician ocean currents presumably played an important role in the dispersal of marine benthic and nektobenthic organisms, as shown for example by the Iapetus Ocean effectively

LETHAIA 36 (2003)

separating the Laurentian Bathyurid trilobite Province from the Balto-Scandian Asaphid trilobite Province during the Early and Middle Ordovician (Fortey & Barnes 1977). If, like trilobites, many conodont animals were dominantly benthic or nektobenthic, and their distribution was interpreted as temperature dependent (Fortey & Barnes 1977, p. 307, ®g. 6), they should be expected to have a similar distribution pattern. On the other hand, the majority of Ordovician conodont taxa were cosmopolitan or widespread, giving the appearance that provincialism was weak in comparison with other benthic groups like brachiopods and trilobites. Endemic forms have not been found to be dominant in the biodiversity of any faunas, except for those which lived in very specialized environments. Therefore, a sharp faunal difference observable in geographically closely linked areas is more likely to be attributed to either different realms or local environmental variations, rather than provincially important. Hence assessment of biogeographic af®nities between different provinces should be based on comparison among faunas inhabiting similar environments. Shallow-water faunas in different shelf regions are connected through open ocean surface waters in which their larvae are dispersed. Because faunal distribution is mainly controlled by surface water temperature gradients, they form different biogeographic regions in a general latitude-parallel pattern. Detailed studies of modern marine organisms have demonstrated that the deep ocean acts as an effective barrier between shelf faunas (Cox & Moore 2000). An ocean exceeding 1000 km wide is generally regarded as a major geographical barrier preventing the dispersal of benthic organisms (Cook & Taylor 1975; McKerrow & Cocks 1986; Mac Niocaill & Smethurst 1994; Christiansen & Stouge 1999). According to the latest global reconstruction maps (Webby et al. 2000; Scotese 2001), Ordovician land masses were distributed from the equatorial or subequatorial zone to the South Pole, with high latitudes of the northern hemisphere covered by the vast Paleo-Paci®c (Panthalassic) Ocean. Two other major oceans are recognized: the Iapetus Ocean which separated Balto-Scandia and Laurentia and northeastern Gondwana to the north, and the Paleo-Tethys Ocean bounded by Gondwana to the east and Laurentia and Siberia to the west. Ordovician conodont provinces should re¯ect this distribution of the continental blocks and their bounding ocean depths, rather than being globally dispersed. If typical elements of the Laurentian Province were successfully dispersed across ocean barriers (Paleo-Paci®c and Paleo-Tethys, which were inhabited by faunas belonging to the Open-Sea Realm) into similar habitats in the interior of the North China

LETHAIA 36 (2003)

Ordovician conodont biogeography

363

and the cratonic region of Australia, they should be regarded as widespread taxa, if not cosmopolitan.

Biogeographic hierarchy ± reconsidered Ecological biogeographers assert that ocean hydrology, more speci®cally the seasonal variation in temperature and salinity of the water, controls the distribution of oceanic life (Bailey 1998, p. 11). They subdivide the modern world's oceans into a hierarchical system of various hydrographic zones, which are closely related to factors like latitude (controlling temperature and vertical circulation), major wind systems (in¯uencing horizontal and vertical circulations) and precipitation and evaporation (controlling salinity and vertical circulation). Two realms are recognized, the Shallow-Sea and Open-Sea (Cox & Moore 2000), which are further subdivided into domains (tropical, temperate and polar) and then regional provinces. The Open-Sea Realm is also subdivided on the sea ¯oor pro®le (Slope or Bathyal Zone, Abyssal Zone and Hadal Zone) affecting distribution of benthic organisms, and on the water column depth relating to pelagic organisms (Epipelagic, Mesopelagic, Bathypelagic and Hadopelagic Zones). The Shallow-Sea Realm, occupying about 14.6% of the modern oceans (Bailey 1998, table 2.3), has a water depth from 0 to c. 200 m (approximating the shelf break), and forms the transitional zone between the Open-Sea Realm and the land. The most striking characteristic of the ShallowSea Realm is its substantial exposure to local environmental changes, due to variable topography of coasts and sea ¯oors, input of sediments and fresh water through drainage systems on lands, and the magnitude of tides and currents. More importantly, domains and provinces of the Shallow-Sea Realm are often separated from each other by deep oceans which may act as effective barriers to dispersal, resulting in a signi®cantly higher ratio of locally distributed endemic forms. We show here how this hierarchical scheme of modern ocean biogeography can be readily applied to Ordovician conodont distribution (Fig. 2). Divisions of the ®rst (realms) and second (domains) orders are mainly ecologically de®ned, based on ocean hydrology, more speci®cally re¯ecting a group of habitats or ecosystems delimited by similarity in overall faunal composition. The striking difference between the two ®rst-order divisions, namely the Shallow-Sea Realm and the Open-Sea Realm, is the much higher ratio of endemic elements in the former. Consequently, provincialism is best expressed in the shallow-water

Fig. 2. Hierarchical scheme of biogeographic subdivisions based on Ordovician conodont data; provinces shown pertain to the Early Ordovician. For further explanation, see text.

faunas. Diversity is highest in faunas occupying the Tropical Domain of the Shallow-Sea Realm (e.g. those from the shelf regions of Laurentia, North China and Australia). Faunas from the Cold Domain of the Shallow-Sea Realm, such as the Early Ordovician Balto-Scandian fauna which occupied high latitudinal positions at the time, were of relatively lower diversity, and were also dominated by cosmopolitan and widespread taxa. The third-order division (provinces), based on analysis of endemism, forms the core of the biogeographical reconstruction. Provinces within the same domains presumably had faunas inhabiting areas with similar environmental conditions, but were more or less isolated from each other geographically. Signi®cantly, each of the provinces is characterized by unique endemic elements. Analysis of endemism, using historical biogeographical methods, permits recognition of the distribution patterns of these taxa, their ancestors and descendants. Such distribution patterns, linked with their phylogenetic history, form the basis for interpretation of biogeographical relationships among the major landmasses at the provincial level during the Ordovician. Conodont data from basinal or open-ocean sediments are scarce due to the apparent absence of carbonate sediments deposited below the carbonate compensation depth in these environments, and conodont faunal records from deep-water cherts and argillaceous ®ne sediments remain incomplete. Known Open-Sea Realm faunas, including those from shelf margin, slope and basinal facies, are dominated by cosmopolitan and widespread taxa, and thus exhibit at best weak indications of provincialism. Therefore, formal subdivision of the Open-Sea Realm into area-speci®c provinces is not attempted at this point.

364

Yong-Yi Zhen & Ian G. Percival

Ordovician conodont provinces ± updated Suf®cient new data obtained over the past decade now permits revision of the provinces ®rst identi®ed by BergstroÈm (1990) and modi®ed by subsequent authors. Particularly noticeable are changes in provinces covering East Gondwana (including Chinese blocks and the Australian region) and South America following extensive studies in the Argentine Precordillera. The Laurentian Province represents the distribution of shallow, warm water faunas inhabiting shelf regions fringing Laurentia, which was within the tropical zones throughout the Ordovician. It is more or less equivalent to the North American Midcontinent Province sensu stricto, except that it may also include Siberia, which was located adjacent to Laurentia through this time. A number of well-recognized Laurentian endemics are also found in the Siberian Platform during transgressive episodes, especially in the Early Ordovician (Bagnoli & Stouge 1991). However, restricted distribution of some forms to Siberia indicates that the region may qualify as a subprovince of the Laurentian Province, or as an independent province (BergstroÈm 1990). The Australian Province (in part the Australasian Province of Nowlan et al. 1997) was initially recognized on occurrences of characteristic endemic forms in the Late Ordovician. Recent studies have con®rmed that this region remained distinctive through most of the Ordovician Period, with close biogeographic relationships between Australia and three major tectonic blocks of China (North, South and Tarim) indicated by sharing distinctive endemic elements, such as species of Serratognathus, Bergstroemognathus and Rhipidognathus, in the middle Early Ordovician (Zhen et al. in Webby et al. 2000), Tasmanognathus in the Gisbornian (early Late Ordovician) and Taoqupognathus and Yaoxianognathus in the Eastonian (middle Late Ordovician) (Zhen 2001). Based on the occurrence of a number of species restricted to the Australian cratonic region and North China (including Korea), these areas are treated herein as two separate provinces. Sharing of a number of regionally distributed taxa also indicates that they were more closely related to each other biogeographically than to Laurentia throughout the Ordovician. Tarim was likely within the North China Province, evidenced by the occurrence of some characteristic North China taxa. However, recent research on conodonts (mainly from core samples associated with petroleum exploration in the region) also indicated some ties of this block with Laurentia (Zhao et al. 2000). Conodont faunas from South China show a close

LETHAIA 36 (2003)

biogeographic linkage with those of North China, indicated by the occurrence of some regionally restricted taxa, like Serratognathus in the Early Ordovician and Taoqupognathus (only found in the northeastern part) in the Late Ordovician. However, unlike North China, typical shallow water, tropical forms are not recognized in the South China faunas. Faunal composition indicates that South China was a relatively deeper water setting (Yangtze Platform) and probably lay within the Temperate Zone during most of the Ordovician, except for its northeastern corner (western Zhejiang Province and vicinity). In western Zhejiang, carbonate build-ups and coralstromatoporoid reefs (Sanjushan Formation) developed during the Late Ordovician. Although South China was geographically adjacent to North China and Australia, domination of cooler water habitats suggests its recognition as a separate province, probably within the Temperate Domain of the Shallow-Sea Realm. As in South China, the Argentine Precordillera faunas were also characterized by the occurrence of a so-called `temperate water group', and were apparently lacking in typical shallow water, tropical forms. Recognizing this, Bagnoli & Stouge (1991) proposed the Precordilleran Province to include faunas which occupied intermediate palaeolatitudes, and also faunas recognized in shelf margin and slope settings in the tropical zone. Noticeably, most species of the Argentine Precordillera faunas are cosmopolitan or widespread. They show lower endemism in comparison with those faunas from the Tropical Domain. This was presumably the reason why Bagnoli & Stouge (1991) grouped them together with the open-sea faunas along the margins of Laurentia. These faunas are tentatively regarded herein as representing a separate province within the Temperate Domain in the Shallow-Sea Realm. Although both the South China and Argentine Precordillera provinces might be situated within the Temperate Domain during the Early Ordovician, the former had closer biogeographic links with the Australian and North China provinces, whereas the latter exhibited closer ties with the Laurentian Province. The Balto-Scandian Province is based on faunal studies of the eponymous region (van Wamel 1974; LoÈfgren 1978; Bagnoli & Stouge 1997), which was located in high latitudes during the Early Ordovician and separated from Laurentia by the Iapetus Ocean. Most of the species and genera de®ned by early workers as characteristic elements for this province during the Early Ordovician have subsequently been demonstrated to be either cosmopolitan or widespread taxa. Typical Early Ordovician shallow-water tropical forms were absent from this region. Along

LETHAIA 36 (2003)

Ordovician conodont biogeography

365

Fig. 3. Distribution of selected biogeographically important conodont taxa in the Early Ordovician (Prioniodus elegans and Oepikodus evae zones). Domains are abbreviated in right-hand columns as follows: Te = Temperate, Co = Cold.

with the closure of the Iapetus towards the end of the Ordovician, Baltica drifted northward towards the equator, and provincialism became less evident due to the in¯ux of tropical forms.

Early Ordovician conodont provincialism Faunas from the late Early Ordovician (Bendigonian

and Chewtonian time interval, approximately elegans to evae conodont zones) serve to demonstrate the applicability of revised biogeographic concepts advocated in this paper and are discussed in more detail below. During the mid±late Early Ordovician, conodonts experienced their highest provincialism and diversity of the period. Along with the closure of the Iapetus Ocean towards the end of Ordovician, the Balto-Scandian Province shifted from the Cold Domain in the Early Ordovician to the Tropical Domain in the Late Ordovician, when the landmass

366

Yong-Yi Zhen & Ian G. Percival

drifted towards the Equator. Each province in the Tropical Domain has faunas inhabiting various biofacies along a pro®le from nearshore to offshore open shelf, and is characterized by the presence of its own unique endemic elements (Fig. 3). By analysing these biogeographic data in reference to Ordovician global reconstructions, it becomes clear that although the Laurentian, North China and Australian continental blocks were situated more or less in the equatorial or sub-equatorial zones, the distinct provinces characterizing these regions were separated by vast ocean expanses of the Paleo-Paci®c and PaleoTethys. Typical Laurentian endemic elements during this time interval, including various species of Parapanderodus and Cristodus, and a number of other species (Fig. 3), are widely distributed in North American platform successions but have not been recorded in the other provinces. During the mid-Early Ordovician, Serratognathus bilobatus (species group) and Rhipidognathus? yichangensis had a wide distribution con®ned to the shelf regions fringing the Australian craton, North China Platform, Korea, Tarim Terrane, South China, and possibly Iran and Kazakhstan. These elements were only recorded from the shelf successions on the cratonic region in central and western Australia (Zhen et al. 2001; Webby et al. 2000; R. Nicoll, pers. comm. 2002). Elements of the Serratognathus fauna were also reported to occur in Kazakhstan (Bagnoli & Stouge 1991). Wang et al. (1996) showed that there were signi®cant dissimilarities between the faunas of North and South China through the later Early to Middle Ordovician interval (evae conodont Zone and younger). South China faunas are similar to those from Argentine Precordillera in lacking typical warmwater taxa, whereas cooler water species were common to both successions. Faunal similarities between North China and South China are more apparent in Tremadocian±earliest Arenig time, prior to the dramatic deepening event that is recorded on the Yangtze Platform (South China block). Signi®cant faunal differences recorded between these two Chinese blocks might be related to their relative palaeogeographic positions (palaeolatitudes), and probably in part were caused by subsidence of South China or by a major eustatic sea level rise ± the Oepikodus evae transgression (Fortey 1984; Lehnert 1995; Barnes et al. 1996). The deepening event is not apparently recorded in North China (Zhou & Fortey 1986; Webby et al. 2000). Rhipidognathus laiwuensis, R. maggolensis, Serratognathus extensus, Erraticodon tangshanensis, Eoserratognathus ovatus and Bergstroemognathus pectiniformis are characteristic endemics con®ned to the North China Province. Several species of Paraserratognathus were also recorded from North China. One

LETHAIA 36 (2003)

species of this genus (P. costatus) was also widely distributed in the Laurentian Province (Fig. 3). Bergstroemognathus kirki is endemic to the Australian cratonic regions. A closely related species, B. hubeiensis, is only recorded in South China, North China and Iran (Zhen et al. 2001). The possible ancestor of these two species, B. extensus, had a wide distribution in the Australian, South China, Laurentian and Argentine Precordilleran provinces, but has not been reported from the Balto-Scandian Province. It also spread to the Tropical Domain of the Open-Sea Realm (Zhen et al. 2001). Other possible endemics con®ned to the Australian Province include Triangulodus larapintinensis, Erraticodon patu and Acodus emanuelensis (Zhen et al. 2003b). Very few species from the Temperate Domain are recognized as provincial, whereas the Cold Domain is overwhelmingly dominated by cosmopolitan and widespread species. Signi®cantly, none of the 37 biogeographically important species and species groups in this time interval (listed in Fig. 3) has been reported from the Cold Domain, except for Oepikodus evae, which is common in the Open-Sea Realm and the Temperate Domain of the Shallow-Sea Realm (Johnston & Barnes 1999; Stewart & Nicoll 2003), but absent from the Tropical Domain of the Shallow-Sea Realm. Balto-Scandian mid±Early Ordovician faunas are well represented in platform rocks, which were deposited in cold shallow-water environments (LindstroÈm 1955, 1971; van Wamel 1974; LoÈfgren 1978). These cold-water successions are highly condensed with the P. elegans and O. evae zones represented by only 0.1 m and 1.5 m thick carbonates, respectively, in the Sjurberg section, central Sweden (LoÈfgren 1994). In contrast, in the Australian Province of the Tropical Domain, the Tabita Formation in western New South Wales has a maximum thickness of 182 m entirely within the evae Zone (Zhen et al. 2003b). Bagnoli & Stouge (1991) suggested that Baltoniodus? deltatus (LindstroÈm, 1955) sensu stricto (Bagnoli et al. 1988) and Protopanderodus rectus (LindstroÈm, 1955) were typical endemic elements con®ned to the BaltoScandian Province in this time interval. However, species de®nition for these two species is still poorly understood. Their precise distribution can only be con®rmed with future taxonomic revision of these species. LoÈfgren (1994) recorded 18 species in this time interval (elegans and evae zones) from central Sweden, all of which were either cosmopolitan or widespread. Conodont faunas recorded in the graptolitic shales at Hunneberg in Sweden may represent a deeper water fauna (LoÈfgren 1993). Late Tremadocian deeper water faunas from Scandinavia share more species in common with deeper water faunas of the

LETHAIA 36 (2003)

Tropical Domain within the Open-Sea Realm, than with adjacent shallow-water faunas of the BaltoScandian Province (LoÈfgren et al. 1999), leading to the suggestion that the Iapetus Ocean effectively restricted conodont faunal contact between these two provinces. Recent study of the conodont faunas from deep-water deposits of the Scandinavian Caledonides revealed low diversity faunas consisting of cosmopolitan and widespread species in this time interval (Rasmussen 2001). LindstroÈm's (1976) Juanognathus fauna, which was originally restricted to the interval of the evae conodont Zone of the Argentine Precordillera Province, is characteristic of the Tropical Domain of the Open-Sea Realm and the Temperate Domain of the Shallow-Sea Realm in this time interval. This fauna has been recognized in the Temperate Domain of the Shallow-Sea Realm (Argentine Precordillera and South China) as well as in the Tropical Domain (Open-Sea Realm), as evidenced in the deeper water facies of platform margins or slope settings disposed in low palaeolatitudes, such as the Cow Head Group fauna (Stouge & Bagnoli 1988; Pohler 1994; Johnston & Barnes 1999, 2000) of western Newfoundland, the Deep Kill Shale of eastern New York (Landing 1976) and the Hensleigh Siltstone of central New South Wales (Zhen et al. 2003a). However, the Juanognathus fauna has not been reported from the Cold domains of either realm. Although this fauna is dominated by cosmopolitan forms, others, including Reutterodus, Bergstroemognathus and Jumudontus, have a more restricted distribution. They are common in the Open-Sea Realm fringing the Australian and Laurentian provinces of the Tropical Domain as well as in the South China and Argentine Precordillera provinces which belong to the Temperate Domain (ShallowSea Realm), but also have not been recorded from the Cold domains of either Realm.

Conclusions Our review has shown that the traditional twofold model of Ordovician conodont distribution cannot be used as the basis for biogeographic analysis. Preponderance of cosmopolitan and widespread taxa in the North American Midcontinent Province and North Atlantic Province faunas is, however, very useful biostratigraphically, ensuring high global correlation potential. Even within the North American (Laurentian) region, confusion over concepts and de®nitions of the two `provinces' has lessened their utility. Continued application of the traditional model is at variance with current reconstructions of the Ordo-

Ordovician conodont biogeography

367

vician world, and con¯icts with examples of provincialism exhibited by other biotas of the time. Accordingly, we recommend that the terms North American Midcontinent Province and North Atlantic Province be abandoned in favour of the terminology and concepts advocated in this article. Much of the focus of the traditional model has been on the ®rst-order division (realms), but only the second-order (domain) and third-order (province) subdivisions proposed herein are signi®cant in illuminating the biogeographic histories and relationships of major landmasses during the Ordovician. Whereas faunas from different domains and provinces may once have inhabited more or less similar environments, over time they became geographically or ecologically isolated from each other resulting in various degrees of endemism. Physical barriers separating provinces within a given domain are probably the most crucial factor limiting the distribution of endemic taxa. Recognition of endemic taxa, and analysis of their distribution by using an integrated approach of ecological and historical methods, is essential to reconstructing conodont biogeography, especially at the provincial level of the Shallow-Sea Realm. On this basis, at least six conodont biogeographical provinces, namely Balto-Scandian (in the Cold Domain), Laurentian (Tropical Domain), Australian (Tropical Domain), North China (Tropical Domain), South China (Temperate Domain) and also Argentine Precordillera (Temperate Domain) can be recognized in the Early Ordovician Shallow-Sea Realm. These were separated by the Ordovician oceans, which acted as effective barriers to isolate endemics. The Open-Sea Realm is dominated by cosmopolitan and widespread taxa, and further formal subdivision is yet to be achieved through detailed studies. We predict that analysis of conodont faunas preserved in deep-water cherts will provide the key to distinguishing domains and even provinces within this Realm. We have demonstrated qualitatively that the more detailed biogeographical subdivisions here identi®ed in Early Ordovician conodont distributions are useful and valid. Our analysis has been restricted to a preliminary review of the literature, combined with our detailed knowledge of local faunas. There are signi®cant opportunities for quantitative studies to re®ne the de®nition of the provinces we have recognized, with potential to identify others. Furthermore, the concepts we espouse can be readily extended to other time intervals (and biotas) where provincialism is apparent. However, we caution that statistical methods are only valid when comparing faunas from similar environments (preferably faunas from the same domain), as indiscriminate analysis on faunas

368

Yong-Yi Zhen & Ian G. Percival

from different realms and domains could lead to signi®cantly distorted results. Acknowledgements. ± This study was supported by a Research Fellowship provided by Sydney Grammar School to YYZ. We thank Chris Barnes and Guillermo Albanesi for their constructive reviews of the manuscript, improving the original version presented at the First International Palaeontological Congress held at Macquarie University in July 2002. IGP publishes with permission of the Director-General, N.S.W. Department of Mineral Resources.

References Albanesi, G.L., HuÈnicken, M.A. & Barnes, C.R. 1998: BioestratigrafõÂa, biofacies y taxonomia de conodontes de las secuencias ordovõÂcicas del Cerro Porterillo, Precordillera central de San Juan, R. Argentina. Actas de la Academia Nacional de Ciencias 12, 1±249. Albanesi, G.L., Ortega, G. & Zeballo, F. 2001: Late Tremadocian conodont-graptolite biostratigraphy from NW Argentine basins. The Guide Book, Mongolian Ordovician and Silurian Stratigraphy and Abstracts for the Joint Field Meeting of IGCP 410-IGCP 421 in Mongolia, 125±127. An, T.X., Zhang, F., Xiang, W.D., Zhang, Y.Q., Xu, W.H., Zhang, H.J., Jiang, D.B., Yang, C.S., Lin, L.D., Cui, Z.T. & Yang, X.C. 1983: The Conodonts in North China and Adjacent Regions. 223 pp. Science Press, Beijing (in Chinese with English abstract). Bagnoli, G. & Stouge, S. 1991: Paleogeographic distribution of Arenigian (Lower Ordovician) conodonts. Anais da Academia Brasiliera de Ciencias 63, 171±183. Bagnoli, G. & Stouge, S. 1997: Lower Ordovician (Billingenian ± Kunda) conodont zonation and provinces based on sections È land, Sweden. Bolletino della SocietaÁ from Horns Udde, north O Paleontologica Italiana 35, 109±163. Bagnoli, G., Stouge, S. & Tongiorgi, M. 1988: Acritarchs and conodonts from the Cambro/Ordovician FuruhaÈll (KoÈpingsÈ land, Sweden). Rivista Italiana di Paleontologia e klint) section (O Stratigra®a 94, 163±248. Bailey, R.G. 1998: Ecoregions: the Ecosystem Geography of the Oceans and Continents. 176 pp. Springer, New York. Barnes, C.R. & FaÊhraeus, L.E. 1975: Provinces, communities, and the proposed nektobenthic habit of Ordovician conodontophorids. Lethaia 8, 133±149. Barnes, C.R., Fortey, R.A. & Williams, S.H. 1996: The pattern of global bioevents during the Ordovician Period. In Walliser, O.H. (ed.): Global Events and Event Stratigraphy, 139±172. Springer, New York. Barnes, C.R., Rhodes, C.B. & Miller, J.F. 1973: Lower Paleozoic provincialism. Geological Society of America, Special Paper 141, 157±190. BergstroÈm, S.M. 1971: Conodont biostratigraphy of the Middle and Upper Ordovician of Europe and Eastern North America. Geological Society of America, Memoir 127, 83±161. BergstroÈm, S.M. 1990: Relations between conodont provincialism and the changing palaeogeography during the Early Palaeozoic. In McKerrow, W.S. & Scotese, C.R. (eds): Palaeozoic palaeogeography and biogeography. Geological Society of London, Memoir 12, 105±121. BergstroÈm, S.M. & Sweet, W.C. 1966: Conodonts from the Lexington Limestone (Middle Ordovician) of Kentucky and its lateral equivalents in Ohio and Indiana. Bulletins of American Paleontology 50, 271±441. Brown, J.H. & Lomolino, M.V. 1999: Biogeography. 2nd edn. 691 pp. Sinauer Associates, Massachusetts. Charpentier, R.R. 1984: Conodonts through time and space: studies in conodont provincialism. In Clark, D.L. (ed.): Conodont biofacies and provincialism. Geological Society of America, Special Paper 196, 11±32. Christiansen, J.L. & Stouge, S. 1999: Using palaeo-oceanographical

LETHAIA 36 (2003) modelling in reconstructing Early Ordovician palaeogeography. Acta Universitatis Carolinae ± Geologica 43, 515±518. Cook, H.E. & Taylor, M.E. 1975: Early Palaeozoic margin sedimentation, trilobite biofacies, and the thermocline, western United States. Geology 3, 559±562. Cox, C.B. & Moore, P.D. 2000: Biogeography: an Ecological and Evolutionary Approach. 6th edn. 298 pp. Blackwell Science, Oxford. Dubinina, S.V. 1991: Upper Cambrian and Lower Ordovician conodont associations from open ocean paleoenvironments, illustrated by Batyrbay and Sarykum sections in Kazakhstan. In Barnes, C.R. & Williams, S.H. (eds): Advances in Ordovician Geology. Geological Survey of Canada, Paper 90-9, 107±124. Dubinina, S.V. 1998: Conodonts from the Early Ordovician (midArenig) deep-water deposits of central Asian Paleobasins. In Szaniawski, H. (ed.): Proceedings of the Sixth European Conodont Symposium (ECOS VI). Palaeontologica Polonica 58, 79±86. Dzik, J. 1983: Relationship between Ordovician Baltic and North American Midcontinent faunas. Fossils and Strata 15, 59±85. Ethington, R.L., Finney, S.C., Miller, J.F., Ross, R.J. Jr. & ValdesCamin, C. 1995: Pre-meeting trip ± Central Great Basin Transect. In Cooper, J.D. (ed.): Ordovician of the Great Basin: Fieldtrip Guidebook and Volume for the Seventh International Symposium on the Ordovician System, Book 78, 1±50. Paci®c Section SEPM, Fullerton, California. Ethington, R.L. & Repetski, J.E. 1984: Paleobiogeographic distribution of Early Ordovician conodonts in central and western United States. In Clark, D.L. (ed.): Conodont biofacies and provincialism. Geological Society of America, Special Paper 196, 89±101. Fortey, R. 1984: Global earlier Ordovician transgressions and regressions and their biological implications. In Bruton, D.L. (ed.): Aspects of the Ordovician System. Paleontological Contributions from the University of Oslo 295, 37±50. Fortey, R. & Barnes, C.R. 1977: Early Ordovician conodont and trilobite communities of Spitsbergen: in¯uence on biogeography. Alcheringa 1, 297±309. Gage, J.D. & Tyler, P.A. 1991: Deep-Sea Biology: a Natural History of Organisms at the Deep-Sea Floor. 504 pp. Cambridge University Press, Cambridge. Ji, Z.L. & Barnes, C.R. 1994a: Lower Ordovician conodonts of the St. George Group, Port au Port Peninsula, western Newfoundland, Canada. Palaeontographica Canadiana 11, 1±149. Ji, Z.L. & Barnes, C.R. 1994b: Conodont palaeoecology of the Lower Ordovician St. George Group, Port au Port Peninsula, western Newfoundland. Journal of Paleontology 68, 1368±1383. Johnston, D.I. & Barnes, C.R. 1999: Early and Middle Ordovician (Arenig) conodonts from St. Paul's Inlet and Martin Point, Cow Head Group, western Newfoundland, Canada. 1. Biostratigraphy and paleoecology. Geologica et Palaeontologica 33, 21±70. Johnston, D.I. & Barnes, C.R. 2000: Early and Middle Ordovician (Arenig) conodonts from St. Paul's Inlet and Martin Point, Cow Head Group, western Newfoundland, Canada. 2. Systematic paleontology. Geologica et Palaeontologica 34, 11±87. Landing, E. 1976: Early Ordovician (Arenigian) conodont and graptolite biostratigraphy of the Taconic allochthon, eastern New York. Journal of Paleontology 50, 614±646. Lehnert, O. 1995: Ordovizische Conodonten aus der PraÈkordillere Westargentiniens: Ihre Bedeutung fuÈr Stratigraphie und PalaÈogeographie. Erlanger Geologische Abhandlungen 125, 1±193. LindstroÈm, M. 1955: Conodonts from the lowermost Ordovician strata of south-central Sweden. Geologiska FoÈreningens i Stockholm FoÈrhandlingar 76, 517±604. LindstroÈm, M. 1970: Faunal provinces in the Ordovician North Atlantic areas. Nature 225, 1158±1159. LindstroÈm, M. 1971: Lower Ordovician conodonts of Europe. In Sweet, W.C. & BergstroÈm, S.M. (eds): Symposium on conodont biostratigraphy. Geological Society of America, Memoir 127, 21± 61. LindstroÈm, M. 1976: Conodont palaeogeography of the Ordovician. In Bassett, M.G. (ed.): The Ordovician System: Proceedings of a Palaeontological Association Symposium, Birmingham,

LETHAIA 36 (2003) September 1974, 501±522. University of Wales Press and National Museum of Wales, Cardiff. LoÈfgren, A. 1978: Arenigian and Llanvirnian conodonts from JaÈmtland, northern Sweden. Fossils and Strata 13, 129 pp. LoÈfgren, A. 1993: Conodonts from the Lower Ordovician at Hunneberg, south-central Sweden. Geological Magazine 130, 215±232. LoÈfgren, A. 1994: Arenig (Lower Ordovician) conodonts and biozonation in the eastern Siljan District, central Sweden. Journal of Paleontology 68, 1350±1368. LoÈfgren, A., Repetski, J.E. & Ethington, R.L. 1999: Some transIapetus conodont faunal connections in the Tremadocian. In Serpagli, E. (ed.): Studies on conodonts ± Proceedings of the Seventh European Conodont Symposium (1998). Bolletino della SocietaÁ Paleontologica Italiana 37, 159±173. Lyons, P. & Percival, I.G. 2002: Middle to Late Ordovician age for the Jindalee Group of the Lachlan Fold Belt, New South Wales: conodont evidence and some tectonic implications. Australian Journal of Earth Sciences 49, 801±808. Mac Niocaill, C. & Smethurst, M.A. 1994: Palaeozoic palaeogeography of Laurentia and its margins: a reassessment of palaeomagnetic data. Geophysical Journal International 116, 715±725. McKerrow, W.S. & Cocks, L.R.M. 1986: Oceans, island arcs and olistostromes: the use of fossils in distinguishing sutures, terranes and environments around the Iapetus Ocean. Journal of the Geological Society, London 143, 185±191. Miller, J.F. 1984: Cambrian and earliest Ordovician conodont evolution, biofacies, and provincialism. In Clark, D.L. (ed.): Conodont biofacies and provincialism. Geological Society of America, Special Paper 196, 43±68. Moskalenko, T.A. 1972: Ordovician conodonts of the Siberian Platform and their bearing on multielement taxonomy. Geologica et Palaeontologica 1, 47±56. Moskalenko, T.A. 1983: Conodonts and biostratigraphy in the Ordovician of the Siberian Platform. Taxonomy, ecology and identity of conodonts (ECOS III). Fossils and Strata 15, 87±94. MuÈller, K.J. 1962: Taxonomy, evolution and ecology of conodonts. In Moore, R.C. (ed.): Treatise on Invertebrate Paleontology, part W, Miscellanea, W83±91. Geological Society of America, New York and University of Kansas Press, Lawrence. Murray, S.I. & Stewart, I.R. 2001: Palaeogeographic signi®cance of Ordovician conodonts from the Lachlan Fold Belt, southeastern Australia. Historical Biology 15, 145±170. Nowlan, G.S., McCracken, A.D. & McLeod, M.J. 1997: Tectonic and paleogeographic signi®cance of Late Ordovician conodonts in the Canadian Appalachians. Canadian Journal of Earth Sciences 34, 1521±1537. Pohler, S.M.L. 1994: Conodont biofacies of Lower to lower Middle Ordovician megaconglomerates, Cow Head Group, Western Newfoundland. Geological Survey of Canada, Bulletin 459, 1±71. Pohler, S.M.L. & Barnes, C.R. 1990: Conceptual models in conodont paleoecology. In Ziegler, W. (ed.): 1st International Senckenberg Conference and 5th European Conodont Symposium (ECOS V). Contributions IV: papers on conodonts and Ordovician to Triassic conodont stratigraphy. Courier Forschungsinstitut Senckenberg 118, 409±440. Pyle, L.J. & Barnes, C.R. 2002: Taxonomy, Evolution and Biostratigraphy of Conodonts from the Kechika Formation, Skoki Formation, and Road River Group (Upper Cambrian to Lower Silurian), Northeastern British Columbia. NRC Research Press, Ottawa, Ontario, Canada, 227 pp. Rao, R.I. 1999: Los conodontes cambro-ordovõÂcicos de la sierra de Cajas y del Espinazo del Diablo, Cordillera Oriental, RepuÂblica Argentina. Revista EspanÄola de MicropaleontologõÂa 31, 23±51. Rasmussen, J.A. 1998: A reinterpretation of the conodont Atlantic Realm in the late Early Ordovician (early Llanvirn). In Szaniawski, H. (ed.): Proceedings of the Sixth European Conodont Symposium (ECOS VI). Palaeontologica Polonica 58, 67±77. Rasmussen, J.A. 2001: Conodont biostratigraphy and taxonomy of the Ordovician shelf margin deposits in the Scandinavian Caledonides. Fossils and Strata 48, 1±180.

Ordovician conodont biogeography

369

Scotese, C.R. 2001: Digital Paleogeographic Map Archive on CDrom, Paleomap Project, Arlington, Texas. Seddon, G. & Sweet, W.C. 1971: An ecologic model for conodonts. Journal of Paleontology 45, 869±880. Serpagli, E. 1974: Lower Ordovician conodonts from Precordilleran Argentina (Province of San Juan). Bolletino della SocietaÁ Paleontologica Italiana 13, 17±98. Stewart, I. & Nicoll, R.S. 2003: The 15 element, septimembrate, apparatus structure of the Early Ordovician conodont Oepikodus evae LindstroÈm from Australia and Sweden. Courier Forschungsinstitut Senckenberg. In press. Stouge, S. & Bagnoli, G. 1988: Early Ordovician conodonts from Cow Head Peninsula, western Newfoundland. Palaeontographica Italica 75, 89±179. Sweet, W.C. & BergstroÈm, S.M. 1984: Conodont provinces and biofacies of the Late Ordovician. In Clark, D.L. (ed.): Conodont biofacies and provincialism. Geological Society of America, Special Paper 196, 69±87. Sweet, W.C., Turco, C.A., Warner, E. & Wilkie, L.C. 1959: The American Upper Ordovician standard. I. Eden conodonts from the Cincinnati region of Ohio and Kentucky. Journal of Paleontology 33, 1029±1068. Tipnis, R.S., Chatterton, B.D.E. & Ludvigsen, R. 1978: Ordovician conodont biostratigraphy of the southern district of Mackenzie, Canada. In Stelck, C.R. & Chatterton, B.D.E. (eds): Western and Arctic Canadian biostratigraphy. Geological Association of Canada, Special Paper 18, 39±91. Tipper, J.C. 1980: Some distributional models for fossil animals. Paleobiology 6, 77±95. Tolmacheva, T., Danelian, T. & Popov, L. 2001: Evidence for 15 m.y. of continuous deep-sea biogenic siliceous sedimentation in early Paleozoic oceans. Geology 29, 755±758. Valentine, J.W. 1968: The evolution of ecological units above the population level. Journal of Paleontology 42, 253±267. van Wamel, W.A. 1974: Conodont biostratigraphy of the Upper È land, Cambrian and Lower Ordovician of north-western O south-eastern Sweden. Utrecht Micropalaeontological Bulletins 10, 125 pp. Wang, Z.H., BergstroÈm, S.M. & Lane, H.R. 1996: Conodont provinces and biostratigraphy in Ordovician of China. Acta Palaeontologica Sinica 35, 26±59. Webby, B.D., Percival, I.G., Edgecombe, G.D., Cooper, R.A., VandenBerg, A.H.M., Pickett, J.W., Pojeta, J. Jr., Playford, G., Winchester-Seeto, T., Young, G.C., Zhen, Y.Y., Nicoll, R.S., Ross, J.R.P. & Schallreuter, R. 2000: Ordovician palaeobiogeography of Australasia. Association of Australasian Palaeontologists, Memoir 23, 63±126. Wilde, P. 1991: Oceanography in the Ordovician. In Barnes, C.R. & Williams, S.H. (eds): Advances in Ordovician Geology. Geological Survey of Canada, Paper 90-9, 283±298. Wright, C.A., Barnes, C.R. & Jacobsen, S.B. 2002: Neodymium isotopic composition of Ordovician conodonts as a seawater proxy: testing paleogeography. Geochemistry Geophysics Geosystems 3 10.1029/2001GC000195, 12 pp. Zhao, Z.X., Zhang, G.Z. & Xiao, J.N. 2000: Paleozoic Stratigraphy and Conodonts in Xinjiang. 340 pp. Petroleum Industry Press, Beijing. Zhen, Y.Y. 2001: Distribution of the Late Ordovician conodont Taoqupognathus in eastern Australia and China. Acta Palaeontologica Sinica 40, 351±361. Zhen, Y.Y., Nicoll, R.S., Percival, I.G., Hamedi, M.A. & Stewart, I. 2001: Ordovician rhipidognathid conodonts from Australia and Iran. Journal of Paleontology 75, 186±207. Zhen, Y.Y., Percival, I.G. & Webby, B.D. 2003a: Early Ordovician (Bendigonian) conodonts from central New South Wales, Australia. Courier Forschungsinstitut Senckenberg. Zhen, Y.Y., Percival, I.G. & Webby, B.D. 2003b: Early Ordovician conodonts from far western New South Wales, Australia. Records of the Australian Museum 55, 169±220. Zhou, Z.Y. & Fortey, R.A. 1986: Ordovician trilobites from North and Northeast China. Palaeontographica A 192, 157±210. Zhylkaidarov, A. 1998: Conodonts from Ordovician ophiolites of central Kazakhstan. Acta Palaeontologica Polonica 43, 53±68.