Palaeogeography, Palaeoclimatology, Palaeoecology, Special Issue, volume 195 (issues 1-2)

An International Journal for the Geo-Sciences

Editors-in-Chief

D.J. Bottjer, Los Angeles, CA P. De Deckker, Canberra, A.C.T. F. Surlyk, Copenhagen

Editorial Board

E.J. Barron, University Park, PA D. Beerling, Sheffield W.A. Berggren, Woods Hole, MA H.J.B. Birks, Bergen J. Bloemendal, Liverpool A.J. Boucot, Corvalis, OR C.E. Brett, Cincinnati, OH R.G. Bromley, Copenhagen M.B. Cita, Milan T.J. Crowley, College Station, TX J.C. Duplessy, Gif-sur-Yvette A.A. Ekdale, Salt Lake City, UT N. Eyles, Scarborough, ON F.T. Fu«rsich, Wu«rzburg M. Gaetani, Milan F. Gasse, Aix-en-Province A. Hallam, Birmingham M.J. Head, Cambridge T.C. Johnson, Duluth, MN K. Kaiho, Sendai A.H. Knoll, Cambridge, MA A. Longinelli, Parma B.A. Maher, Lancaster M.O. Mancen‹ido, La Plata V. Markgraf, Boulder, CO J.I. Martinez, Medelin L.H. Nielsen, Copenhagen H. Okada, Sapporo Pinxian Wang, Shanghai I. Premoli Silva, Milano G. Retallack, Eugene, OR A.C. Scott, Egham, Surrey P. Swart, Miami, FL J. Syktus, Indooroopilly, QLD T.N. Taylor, Lawrence, KS H. Thierstein, Zu«rich E. Thomas, New Haven/Middletown, CT J.C. Tipper, Freiburg H. Visscher, Utrecht E.S. Vrba, New Haven, CT M.R. Walter, North Ryde, N.S.W. P.B. Wignall, Leeds J.A. Wolfe, Tucson, AZ J.C. Zachos, Santa Cruz, CA

VOLUME 195 (2003)

Amsterdam ^ London ^ New York ^ Oxford ^ Paris ^ Shannon ^ Tokyo

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) v

www.elsevier.com/locate/palaeo

Special Issue

Early Palaeozoic Palaeo(bio)geographies of Western Europe and North Africa Edited by

Thomas Servais, Jose¤ Javier Alvaro and Alain Blieck UMR 8014 CNRS, USTL, F-59655 Villeneuve d’Ascq Cedex, France

UMR 8014 CNRS - USTL

CONTENTS Foreword T. Servais, J.J. Alvaro and A. Blieck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Palaeogeographical controls on the Cambrian trilobite immigration and evolutionary patterns reported in the western Gondwana margin Ł lvaro, O. Elicki, G. Geyer, A.W.A. Rushton and J.H. Shergold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J. Javier A Global Ordovician vertebrate biogeography A. Blieck and S. Turner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cambroclaves from the Cambrian of Sardinia (Italy) and Germany: constraints for the architecture of western Gondwana and the palaeogeographical and palaeoecological potential of cambroclaves O. Elicki and T. Wotte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Palaeogeographical and palaeoecological aspects of the Cambro^Ordovician radiation of echinoderms in Gondwanan Africa and peri-Gondwanan Europe B. Lefebvre and O. Fatka . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Ireviken Event in the lower Silurian of Gotland, Sweden ^ relation to similar Palaeozoic and Proterozoic events A. Munnecke, C. Samtleben and T. Bickert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Armorica ‘microplate’: fact or ¢ction? Critical review of the concept and contradictory palaeobiogeographical data M. Robardet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ordovician organic-walled microphytoplankton (acritarch) distribution: the global scenario T. Servais, J. Li, S. Molyneux and E. Raevskaya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cosmopolitan arthropod zooplankton in the Ordovician seas J. Vannier, P.R. Racheboeuf, E.D. Brussa, M. Williams, A.W.A. Rushton, T. Servais and D.J. Siveter . . . . . . . . . . . Patterns of ostracod migration for the ‘North Atlantic’ region during the Ordovician M. Williams, J.D. Floyd, M.J. Salas, D.J. Siveter, P. Stone and J.M.C. Vannier . . . . . . . . . . . . . . . . . . . . . . . . . . . .

doi:10.1016/S0031-0182(03)00425-5

1

5 37

55

73 99 125 149 173 193

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 1^3 www.elsevier.com/locate/palaeo

Foreword Thomas Servais  , Jose¤ Javier Alvaro, Alain Blieck UMR 8014 CNRS, USTL, F-59655 Villeneuve d’Ascq Cedex, France

In September 2001, the meeting ‘‘Early Palaeozoic palaeogeographies and palaeobiogeographies of western Europe and North Africa’’ was organised in the Conference Centre of the University of Sciences and Technologies of Lille at Villeneuve d’Ascq. The congress was organised by the department of Palaeozoic Palaeontology and Palaeogeography (LP3, UMR 8014), a research unit associated with the Centre National de la Recherche Scienti¢que (CNRS). The congress was sponsored by several organisations, including the French Palaeozoic Working Group, the geological societies of France (SGF) and Belgium (Geologica Belgica), and IGCP project no. 410 ‘‘The Great Ordovician Biodiversi¢cation Event’’. This meeting was the second focused on the topic, the ¢rst conference being held at Lille in 1992. We hope that further meetings will be organised in order to better understand the palaeogeography of the mosaic area of western Europe and North Africa. Over 100 scientists from 16 countries attended the meeting and presented their results in over 70 talks and posters. Oral communications were presented during the 3 days of indoor sessions. Each afternoon a workshop was organised, chaired by two experts in the ¢eld, in order to discuss palaeogeographical maps and palaeobiogeographical scenarios in the Cambrian, Ordovician and Silurian, respectively. These workshops not only al-

lowed to clarify the palaeogeographical terminology, but also to compare the various models of palaeogeographical reconstructions presented during the scienti¢c sessions, that were presented by both palaeogeographers (including geophysicists) and palaeobiogeographers (including palaeontologists). The discussions showed the necessity of maintaining multidisciplinary projects between scientists from various disciplines, such as palaeomagnetism, palaeogeography, biogeography, biostratigraphy, chronostratigraphy, geochemistry, etc. Two geological ¢eld trips were organised in connection with the Lille conference. The pre-conference excursion in Belgium visited the southeastern part of the Avalonia continent, while the post-conference excursion to the Montagne Noire in southern France allowed discussions on the evolution of the north-western margin of Gondwana and its related microterranes. The nine papers presented in this volume are based on the results of the discussions during the workshops held at Lille. Following the meeting, we asked some specialists whether they were willing to present a review paper in order to document the palaeobiogeographical scenario of their fossil group. Some of the co-authors that contributed to this volume were able to meet together at Lille, in some cases for the ¢rst time, to exchange their palaeogeographical models and views. For

* Corresponding author. Tel.: +33-3-20 33 72 20; Fax: +33-3-20 43 69 00. E-mail addresses: [email protected] (T. Servais), [email protected] (J.J. Alvaro), [email protected] (A. Blieck).

0031-0182 / 03 / $ ^ see front matter C 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00299-2

PALAEO 3070 9-5-03

2

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 1^3

this reason the present issue includes a series of review papers of individual fossil groups, some of these including new palaeogeographical maps that were drawn following the discussions at Lille. Some of the contributions presented herein are thus the direct outgrowth of the workshop sessions of September 2001. In the ¢rst paper several specialists working on Cambrian trilobites of western Europe and North Africa brought their data together in order to document the evolutionary patterns related to the palaeogeographical evolution of the trilobites on the Western Gondwana margin (Alvaro et al.). Blieck and Turner review the global Ordovician vertebrate biogeography. They rediscuss various recent palaeogeographical models published for the Ordovician in order to document the radiation of the earliest vertebrates. In the next paper, Elicki and Wotte review the cambroclaves, a poorly known group of problematic microfossils among the small shelly fossils of the Early and Middle Cambrian. The new ¢ndings of cambroclaves in Sardinia and Germany point to a wider distribution than previously suggested and support the model of an existence of a widespread uniform facies belt (shelf) around parts of Gondwana during the Early^Middle Cambrian time interval. The palaeoecology and tempo of the Early Palaeozoic radiation of echinoderms are discussed in the paper of Lefebvre and Fatka. These authors compare the diversity patterns observed in Cambro^Ordovician echinoderm faunas from Laurentia and the northern Gondwana margin. Munnecke et al. discuss the relative timing of stable isotope development, extinctions, and facies development across the Llandovery/Wenlock boundary at the ‘‘Ireviken Event’’ of the Silurian of Gotland, Sweden. These results are compared with similar events in the Lower Palaeozoic, but also in the Proterozoic. The authors propose a palaeoceanographic/climatic model that might be applicable to older events. Robardet critically reviews the palaeogeography of the Armorica ‘‘microplate’’. Following a thorough revision of the literature, Robardet concludes that the concept of an Armorica microcontinent, although repeatedly maintained for more

than 20 years, can today only be considered as a ¢ction. The three last papers in this issue are complete palaeobiogeographical reviews of Ordovician fossil groups. Servais et al. review the Ordovician acritarch palaeobiogeography and conclude that some previously published acritarch ‘‘provinces’’ cannot be maintained. The plotting of some microphytoplankton assemblages on recent palaeogeographical maps indicate that at least two distinct geographical areas occur during the time of maximum continental separation that was the interval between the late Tremadocian and Arenig (Lower/Middle Ordovician). Vannier et al. review the global distribution of caryocaridid arthropods that constitute a zooplanktonic group of organisms in Ordovician marine ecosystems. Probably being distributed according to a cosmopolitan pattern, these arthropods represent a signi¢cant step in the post-Cambrian colonisation of midwater niches and in the construction of complex modern foodwebs. The last paper by Williams et al. is a detailed review of the complex migration pattern of neritic ostracods from the ‘North Atlantic’ region, i.e., between Gondwana and related microterranes, Baltica and Laurentia. The varying migration speed is related to ostracod taxonomic diversity, ostracod longevity, changing sea-level and varying palaeogeography, such as the spread of carbonate^mudstone shelf marine facies in Laurentia during the early and middle Caradoc, that resulted in the migration of up to 18 Baltic-origin genera to Laurentia.

Acknowledgements We would like to express our thanks to the organising institutions of the conference that took place in Villeneuve d’Ascq during September 2001: Universite¤ des Sciences et Technologies de Lille (USTL), Centre National de la Recherche Scienti¢que (CNRS), UMR 8014 du CNRS, Conseil Re¤gional Nord - Pas-de-Calais, Socie¤te¤ Ge¤ologique de France (Paris), Socie¤te¤ Ge¤ologique du Nord (Lille), Groupe FrancJais du Pale¤ozo|«que, Geologica Belgica, and the IGCP project no.

PALAEO 3070 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 1^3

410 ‘‘The Great Ordovician Biodiversi¢cation Event’’. We are particularly grateful to the following scientists (and two anonymous reviewers) who have generously assisted us by providing referee reports on manuscripts submitted for this special issue: P. Ahlberg, C.R. Barnes, E. Clarkson, S.

3

Conway Morris, W. Devos, R.A. Fortey, J. Hannibal, J. Li, G. Miller, A. Owen, A.R. Palmer, F. Paris, D.K. Elliott, D. Kaljo, C.R.C. Paul, A. Smith, P. Sheehan, R. Van der Voo, M. Vanguestaine, R. Wicander and G.C. Young.

PALAEO 3070 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 www.elsevier.com/locate/palaeo

Palaeogeographical controls on the Cambrian trilobite immigration and evolutionary patterns reported in the western Gondwana margin Ł lvaro a; , Olaf Elicki b , Gerd Geyer c , Adrian W.A. Rushton d , J. Javier A John H. Shergold e a

UMA 8014 CNRS, Universite¤ de Sciences et Technologies de Lille, Sciences de la Terre, 59655 Villeneuve d’Ascq Cedex, France b Institut fu«r Geologie, TU Bergakademie Freiberg, Bernhard-von-Cotta-Str. 2, 09596 Freiberg/Sachsen, Germany c Institut fu«r Pala«ontologie, Universita«t Wu«rzburg, Pleicherwall 1, 97070 Wu«rzburg, Germany d Palaeontological Department, The Natural History Museum, London SW7 5BD, UK e La Freunie, Benayes, 19510 Masseret, France Received 12 April 2002; received in revised form 18 July 2002; accepted 25 January 2003

Abstract Southward drifting of the western Gondwanan margin during the Cambrian has been demonstrated by means of both palaeomagnetic methods and lithological indicators of climate (such as carbonates and evaporites). Recent improvements in biostratigraphical correlations permit an enhanced understanding of the climatic and palaeobiogeographical constraints that controlled the distribution of Cambrian benthic communities. Palaeogeographical and biogeographical reconstructions based on trilobites are reported in this paper in order to test interaction between migration, speciation and extinction rates. The variability of the documented biogeographical patterns is directly related to species diversity, in which wider distribution coincides with transgressive trends and subsequent connection of neighbouring platforms. Early Cambrian trilobite faunas show a high degree of both substrate control and endemicity, although transgressions led to parallel shifts in faunal compositions. By contrast, Mid-Cambrian trilobite faunas are relatively uniform across western Gondwana, and latest Mid- and Late Cambrian associations document influence of an increased similarity with Asian trilobite faunas. 5 2003 Elsevier Science B.V. All rights reserved. Keywords: biogeography; palaeogeography; extinction; trilobites; western Gondwana; Cambrian

1. Introduction * Corresponding author. Tel.: +33-3-2033-6392; Fax: +33-3-2043-6900. E-mail addresses: [email protected] Ł lvaro), [email protected] (O. Elicki), (J. Javier A [email protected] (G. Geyer), [email protected] (A.W.A. Rushton), [email protected] (J.H. Shergold).

Four disciplines are the main sources for palaeogeographical reconstruction: palaeomagnetism, sedimentology, palaeoclimatology, and biogeography. The boundaries of the postulated plates and terranes (commonly discussed from the viewpoints of geophysics, tectonophysics and

0031-0182 / 03 / $ ^ see front matter 5 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00300-6

PALAEO 3071 9-5-03

6

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

structural geologists) do not always match the biogeographical boundaries worked out by palaeontologists and biostratigraphers. However, do the biogeographical boundaries that prevented reproductive communication necessarily coincide with those of the tectonic plates? In some cases they do; many plate and terrane boundaries, however, do not correspond with biogeographical barriers and, on the contrary, a single continental margin can display barriers, particularly during regressive conditions, which separate di¡erent groups of organisms that are signi¢cant biogeographically. As a result, a biogeographical unit contains similar biotas, but may or may not show physical continuity. One of the main projects of the International Subcommission on Cambrian Stratigraphy (ISCS) is the proposal of a global chronostratigraphical chart useful for international correlation (see Geyer and Shergold, 2000, for a recent statement of the problem). One of the major di⁄culties is the resolution of the drastic provincialism exhibited by the benthic fauna, such as the trilobites (Palmer, 1998). From the earliest occurrences of trilobite faunas there is clear evidence of biogeographical di¡erentiation into two main provinces: the olenellid province, comprising much of Baltica, Laurentia, and Siberia, and the redlichiid province of Gondwana (Palmer, 1972). Subsequently, an overlapping bigotinid province was distinguished (Pillola, 1991a). For the Mid- and Late Cambrian, Palmer (1972) reported a more complicated scheme involving four provinces for the continental seas, and three other ones for exposed shorelines. In addition, conodont distribution in the Late Cambrian indicates the presence of two faunal realms (Bergstro«m, 1990), a tropical Midcontinent region and a cool Atlantic region. One geographical area where the biostratigraphical correlations are still strongly debated is the western Gondwana margin, which includes the Mediterranean area and much of western and central Europe. Geophysical data that would indicate Cambrian plate boundaries are scarce within this area. However, biogeographical disparities based on detailed and extensive trilobite studies re£ect major palaeogeographical di¡erences and relative sea-level £uctuations. Although theoreti-

cally the reproductive communication of trilobites was maintained by means of larvae, the Cambrian trilobite-based biogeographical patterns of the western Gondwanan margin, as it drifted towards the South Cambrian pole, are not homogeneous, and the changes in patterns of dispersal and evolution of some trilobite families remain uncertain. This paper o¡ers an examination of the biogeographical patterns displayed by the Cambrian trilobites on the western margin of Gondwana, as a method of testing the established palaeogeographical models.

2. Basins and platforms: a nomenclatural approach The Moroccan and European Hercynian (or Variscan) massifs, the latter located south of the Rheno^Hercynian zone and north of the Alpine realm (Fig. 1), contain a mosaic of Cambrian successions assigned to the western Gondwanan margin, namely Morocco, the Iberian and Armorican massifs, Montagne Noire, Sardinia, Saxo^Thuringia and Bohemia. Associated with the same margin are other small cratonic domains recognised in the southern British Isles; some of their boundaries are less well-de¢ned (Cocks et al., 1997). 2.1. Morocco The Neoproterozoic( ?)^Cambrian successions of the Moroccan Atlas are located in the AntiAtlas and central High Atlas mountains, although some disconnected outcrops occur in the Jbilet and Rehamna regions, and the Meseta plateau. The axis of the Late Proterozoic to Early Palaeozoic Souss Basin (Geyer, 1989) roughly coincides with the modern trend of the Anti-Atlas (SW^NE). Common west-to-east facies changes throughout the Cambrian successions re£ect the eastern setting of proximal areas (Destombes et al., 1985; Geyer and Landing, 1995; Geyer et al., 1995). 2.2. The Iberian Massif The Cambrian tectonosedimentary outcrops of the Iberian Massif were subdivided by Lotze

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

TR AN S

SOUTH PORTUGUESE ZONE OSSA-MORENA ZONE

UR

OP

EA N

N - POLIS MA H GER IAN N CY ER -H NO E RH IA RING -THU O SAX U DAN MOL IAN ?-B ?

CANTABRIAN WEST ZONE ASTURIANLEONESE ARMORICAN ZONE MASSIF

CENTRALIBERIAN ZONE

-E

7

RTH NO

DEMANDA MOUNTAINS

FA UL T CA LE DO NID E

S

TEPLABARRANDIAN

PYRENEES

CENTRAL MASSIF ALPINE

REALM

MONTAGNE NOIRE

IBERIAN CHAINS

Western Plateau

High Atlas SW SARDINIA

Anti-Atlas

Pre-Hercynian outcrops of western and central Europe, and Morocco

Fig. 1. Main tectonostratigraphical units of the Variscan Belt (modi¢ed after Franke, 2000), and northwestern Africa.

(1961) into zones, viz. the Cantabrian (CZ), Western Asturian^Leonese (WALZ), Central Iberian (CIZ) and Ossa^Morena (OMZ) zones. Their evolution has been interpreted in terms of two distinct troughs: the Cantabro^Iberian and the Andalusian basins, the latter including the OMZ. The Cantabro^Iberian Basin comprises the CZ, WALZ, northern CIZ, and their eastern prolongation into the Demanda Mountains (DM) and the Iberian Chains (IC). It was limited to the NE by the Cantabro^Ebroan Land area, which constituted the main source of sediments for both the Cantabro^Iberian and the Pyrenean basins (Carls, 1983), and in the SW by some uplifted areas (or median highs; Lotze, 1961), which episodically supplied sediments (Aramburu et al., 1992). The geodynamic a⁄nity of the OMZ is still controversial. One hypothesis supposes its accretion to the Iberian Autochthon (which would include the rest of northern zones) during the Cadomian orogeny (Quesada, 1991), whereas an-

other hypothesis considers the Iberian Massif as a complex tectonic mosaic constructed from two distinct plates that collided during the Early Devonian (Acadian orogeny; Mart|¤nez-Garc|¤a and Rolet, 1991). Because of this discussion, and the displacement of the OMZ from NW to SE along the Badajoz^Co¤rdoba shear zone during the Hercynian orogeny, we will illustrate the OMZ as a neighbouring basin of the western Gondwanan margin. 2.3. The southern British Isles (‘eastern Avalonia’) The main structural elements are the Midland Platform, bounded to the NW and NE by depositional troughs, and truncated to the south by the Hercynian front. The trough to the NE is the concealed ‘Eastern England Caledonide Belt’, in which Cambrian rocks of the Tornquist Sea may be present, but have not been proven. To the NW of the Midland Platform lies the Welsh Trough, an elongated ensialic basin that experienced sub-

PALAEO 3071 9-5-03

8

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

sidence throughout much of the Cambrian Period. The Welsh Trough is limited to the NW by the tectonically active Monian Terranes that behaved as a positive area through the Cambrian, and which separate the Welsh Trough from the Leinster Basin of SE Ireland (Bluck et al., in Cope et al., 1992); the latter contains Cambrian rocks but has not yielded any trilobites. To the north, the Cambrian of Scotland is part of Laurentia, and is not discussed further here. The Cambrian of England and Wales is developed in three main facies: (1) basinal clastic deposits of the Welsh Trough, exposed in two main areas in northern Wales (Harlech Dome and Arfon) and two in the south (St. David’s and Llangynog areas); (2) outer-shelf clastics on the Midland Platform (Nuneaton); and (3) shallow-water clastics with very local limestones in the Welsh Borderland (Comley and Rushton areas), on the margin of the Midland Platform. The geology of these areas was reviewed by Rushton (1974) and Rushton et al. (1999), and the palaeogeography was summarised by Brasier et al. (in Cope et al., 1992). McKerrow et al. (1992) showed eastern Avalonia located close to Morocco at a latitude of about 40‡S in the Early Cambrian, drifting south to about 60‡S in the Late Cambrian (L.R.M. Cocks, pers. commun., 2002). 2.4. The Armorican Massif The Cambrian of the Armorican Massif consists dominantly of siliciclastic deposits. The facies strongly depends on the local position on the shelf and includes locally Lower Cambrian carbonates deposited in nearshore environments. Its apparent polar wander path can be superimposed across the Neoproterozoic^Early Palaeozoic interval on the Gondwana path. A counter-clockwise rotation of the massif has been proposed for Early Palaeozoic times (Young, 1990), based on sedimentological and palaeogeographical similarities with some Portuguese outcrops. 2.5. The Montagne Noire and southwestern Sardinia To the north of the Cantabro^Ebroan Land

area, a mosaic of platforms is represented in the Cambrian outcrops of the Pyrenees and Montagne Noire. The Cambrian stratigraphical and biogeographical patterns of southwestern Sardinia reveal a close similarity to the Montagne Noire so that this platform can be considered neighbouring the Montagne Noire segment at that time. 2.6. Saxo^Thuringia and Bohemia According to Havl|¤c›ek et al. (1994), both the German and Barrandian (an area situated in the central part of the Bohemian Massif) Cambrian deposits belong to the ‘Perunica terrane’. This crustal segment was de¢ned as a separate microplate occupying ‘the major part of the Bohemian Massif, and involving the Moldanubian, Barrandian and Saxothuringian (Saxothuringian^Lugian) Zones’ (Havl|¤c›ek et al., 1994). This interpretation was based on biogeographical a⁄nities of benthic faunas with those of Baltica and a northward drift of Perunica from high (57.9‡S for the Lower Cambrian Paseky Shale) to lower latitudes (31.1‡S for Middle Cambrian sediments and 28.6‡S for Upper Cambrian andesites; Krs et al., 1987; Vra¤na and SYtefldra¤, 1997). By contrast, Kukal (1971) assumed a semi-arid, rather warm climate of continental type for the Lower Cambrian. The Barrandian palaeolatitude setting has been widely debated. Recently, Linnemann et al. (2000) revised the concept of Perunica, eliminating the Saxo^Thuringia area; they distinguished the Brittany^Normandy, Perunica and Saxo^ Thuringia terranes, which would form the socalled ‘Armorican Terrane collage’. Discussion about the existence of these microplates is beyond the aims of this paper. Nonetheless, two tectonostratigraphical areas are considered here: the Barrandian and Saxo^Thuringian areas. The latter comprises a fault-bounded crustal fragment that contains Cambrian outcrops in the Doberlug Syncline, Thuringia and the Franconian Forest area. The palaeogeographical a⁄nity of the Go«rlitz Syncline o¡ers some uncertainties but is here considered as a relic of an overlapping succession on the Cadomian basement of Saxo^Thuringia. In the Barrandian area, a distinct Cambrian sedimentary trough, the Pr›¤|bram^Jince Basin, is

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

commonly recognised. It contains a thick Lower Cambrian continental succession covered by Middle Cambrian, transgressive marine fossiliferous shales (the Jince Formation), and Upper Cambrian volcaniclastic complexes. Laterally, the Jince Formation crops out in the Skryje^Tyr›ovice area. In the biogeographical discussions expressed below, the trilobites of the Skryje^Tyr›ovice area will be reported with those of the Pr›¤|bram^Jince Basin.

3. Cambrian palaeogeographical evolution: an overview Lack of uniformity in the successions of faunas and depositional environments in the western Gondwanan margin has led to the development of di¡erent regional biostratigraphies (Fig. 2). In order to avoid nomenclatural confusion in chronostratigraphical correlations, we here use a com-

9

bination of the Cambrian Moroccan and Iberian scales for the Lower Cambrian, and the Iberian^ Montagne Noire scale for the Middle Cambrian. The sparse Upper Cambrian faunas are related to the Australian succession, though in Britain there are su⁄cient olenid trilobites to correlate with the Scandinavian succession. For Britain, the stratigraphical classi¢cation used in Cowie et al. (1972) is employed because it supplies a single standard for England and Wales and facilitates reference to Rushton’s (1974) stratigraphical summary. It is recognised that the lower two of the British regional series divisions ^ Comley and St. David’s series ^ do not correspond precisely with the usage of Lower and Middle Cambrian in use elsewhere, the inter-series boundary lying rather higher than the Lower^Middle Cambrian boundary in Morocco (Geyer, 1990a). The base of the Merioneth Series corresponds closely to that of the Upper Cambrian as generally understood and its top lies very close to the internationally

Ł lvaro and Vizca|«no, Fig. 2. Chronostratigraphic correlation of regional charts (based on Thomas et al., 1984; Geyer, 1990a; A 1998; Sdzuy et al., 1999; Geyer and Shergold, 2000).

PALAEO 3071 9-5-03

10

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Fig. 3. Palaeogeographical evolution interpreted in the Cambrian basins and platforms of the western Gondwanan margin across eight intervals (not to scale).

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Fig. 3 (Continued).

PALAEO 3071 9-5-03

11

12

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

agreed base of the Ordovician (Landing et al., 2000). Fig. 3A,B provides graphic representation of the changing palaeoenvironmental relationships discussed on the following pages. 3.1. Morocco The Late Proterozoic (‘Panafrican’) orogeny in Morocco and northwestern Africa was followed by deposition of a Neoproterozoic(?)^Cambrian cover succession that illustrates a long-term transition in lithofacies associations, and includes a thick interval (locally more than 2000 m) of carbonate-dominated rocks followed by siliciclasticdominated rocks higher in the Cambrian in the western Anti-Atlas. The Neoproterozoic(?)^Cambrian transition in this region, viz the Adoudou, Lie de vin, and Igoudine Formations, is dominated by restricted marine platform limestones and dolostones, and is overlain by a long-term shale^ carbonate cycle in the Lower Cambrian, the Amouslek Formation. Uppermost Lower Cambrian units include up to 200 m of ¢ne-grained siliciclastics with relatively minor nodular and bedded limestones (Issafen Formation), overlain by the higher energy, shallow-marine, sandstonedominated facies of the Asrir and Tazlaft Formations, which characterises the Lower^Middle Cambrian transition. This period re£ects a signi¢cant change in sedimentation, related to an important phase of relative sea-level £uctuation which caused a major regression followed by an immediate transgressive event (Geyer, 1989; Geyer and Landing, 1995). Limited carbonates also characterise the lower Middle Cambrian. Fossil-hash limestone beds are regularly encountered only in the ‘Bre'che a' Micmacca’ Member (lower part of the Jbel Wawrmast Formation, up to 300 m thick). Progressively stronger, higher-latitude in£uences are shown by the thick shallow-marine deposits that accumulated during most of the Mid- and Late Cambrian; they are a monotonous succession, up to 700 m thick, of siliciclastic sediments representing sandy shoals and tidal-£at environments as well as £uvial to deltaic environments prograding on distal muddy substrates (Jbel Afraou, Rich Khlifa, Bailiella, Azlag, and Jbel Lmgaysmat Forma-

tions). The late Late Cambrian is unknown from the Moroccan Atlas but appears in part to be characterised by some 115 m of shallow-marine ¢ne-grained siliciclastics (Dar Bou Azza Formation) in the Moroccan Meseta (Mergl et al., 1998). In this area this unit overlies a signi¢cant unit of 100 to 200 m of sandstones representing another important regressive unit of apparently earliest Late Cambrian shoreline deposits, the El Hank Formation. 3.2. The Iberian Massif The Cambrian rocks rest unconformably on Neoproterozoic rocks in the whole massif except in the CIZ, where the Vendian^Cambrian transition is tentatively located in the Pusa Shales (Brasier et al., 1979). The transition contains megabreccias, conglomerates, quartzites, limestones and anoxic sediments deposited in a mixed shelf and slope apron (Valladares, 1995). The Cambrian commences with a dominantly terrigenous succession (or lower lithosome): the Herrer|¤a Formation, 900^1700 m thick, in the CZ, representing deltaic deposits with £uvial and intertidal episodes (Rodr|¤guez Ferna¤ndez et al., 1991); the Ca¤ndana Group, up to 2500 m, in the WALZ with shallowwater marine and continental facies (Crimes et al., 1977); the Tamames^Azorejo Formations in the CIZ, 500^600 m thick; the Torrea¤rboles Formation, 0^350 m thick, in the OMZ with shallow subtidal and intertidal deposits; and the Ba¤mbola and Embid Formations, 500^900 m thick, in the IC, interpreted as progradational deltaic systems and storm- and wave-dominated nearshore settings passing upward into tidally in£uenced subtidal and intertidal environments. During Early Cambrian times, the OMZ displayed the setting of rifting processes (Vegas, 1978; Mata and Munha¤, 1990). The second lithosome re£ects the diachronous establishment of mixed carbonate^siliciclastic platforms, ranging from Ovetian to early MidCambrian times: the lower member of the La¤ncara Formation (150^225 m thick) in the CZ, and the Vegadeo Formation (50^500 m) in the WALZ, displaying peritidal to shallow-water deposits with isolated microbial^archaeocyathan

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

buildups (Zamarren‹o, 1972; Russo and BechŁ lvaro et al., 2000b). Mixed platsta«dt, 1994; A forms were established across the late Ovetian^ Marianian interval in the IC (the Jalo¤n, Ribota and Hue¤rmeda Formations ; 300^500 m thick), Los Navalucillos and Tamames Limestones in the CIZ (120^600 m thick), and the Pedroche and Alconera Formations in the OMZ (350^900 m thick). These contain stromatolitic and archaeocyathan-microbial buildups, sandstones and variegated shales deposited in shallow subtidal to intertidal and supratidal environments ; locally they Ł lbear anhydrite or gypsum pseudomorphs (see A varo et al., 2000a, and references therein). By contrast, the mixed sedimentation of the second lithosome was interrupted in the IC, CIZ and OMZ by early Bilbilian regressive siliciclastic sediments Ł lvaro and Venrelated to the Daroca regression (A nin, 1998), lithostratigraphically recognised as the Daroca (IC), Los Cortijos^Endrinal (CIZ) and Castellar (OMZ) Formations (50^150 m thick). An early Bilbilian prograding shoreline is postulated in these areas, in which coastal and shoal complexes grew and migrated seawards during times of rapid sediment in£ux and regressive conditions. The Lower^Middle Cambrian transition is characterised by a signi¢cant change in the style of sedimentation, related primarily to an important phase of tectonic activity. As a result, an irregular topography was widely developed on the platforms, which were di¡erentiated into a mosaic of horsts and grabens. In situ carbonate production on relative topographic highs gave rise to the ‘griotte’ facies of the upper member of the La¤ncara (CZ) and Mansilla (IC) Formations. The third lithosome re£ects the disappearance of carbonate deposition in the whole Iberian Massif; it is interpreted as shallow platform deposits of intertidal and braid-plain deltaic environments in the Oville Formation (80^800 m) and the lower part of the Barrios Formation (300^400 m thick) in the CZ (Aramburu et al., 1992; Aramburu and Garc|¤a-Ramos, 1993), the shallow-marine deposits of the Cabos Series (up to 4000 m thick ; Baldwin, 1977; Marcos and Pe¤rez Estau¤n, 1981), and the Murero Formation, Aco¤n Group and Valconcha¤n Formation in the IC (600^1400 m thick)

13

representing storm-dominated o¡shore, progradŁ ling sandy shoals, and tidal-£at environments (A varo and Vennin, 2001). 3.3. The southern British Isles In the southern British Isles, Cambrian rocks occur in England and Wales in scattered inliers, the largest of which are: (1) the Harlech Dome (and the St. Tudwal’s Peninsula), and the Arfon area of northern Wales ; (2) Pembrokeshire in south Wales; (3) a number of small outcrops in Shropshire; (4) the Malvern Hills; and (5) the Nuneaton area in Warwickshire, British Midlands. The lithostratigraphical successions of these areas di¡er greatly. The relatively thick and comparatively complete succession of northern Wales includes Lower Cambrian deposits consisting of s 450 m of deltaic sandstones (Dolwen Grits) overlain by up to 200 m of prodeltaic shales (Llanbedr Slates), and a formation of up to 800 m of proximal turbiditic sandstones with shale intercalations (Rhinog Grits). Middle Cambrian deposits are generally basinal turbiditic sandstones and shales, together about 700 m thick, capped by 100 m of dark fossiliferous shales. Manganese-bearing rocks occur close to the base of the Middle Cambrian. The Upper Cambrian consists of 2000 m of ¢ne-grained shales (or slates) and feldspathic £ags, capped by a condensed unit of black mudstone, the Dolgellau Formation, about 100 m thick. The lower Upper Cambrian (Maentwrog Formation) is basinal, but the Festiniog and Dolgellau are open-shelf deposits. The successions in Pembrokeshire show a comparable depositional sequence, commencing with the Lower Cambrian Caerfai Group, about 300 m thick, consisting of sandstones and shales overlying a transgressive basal conglomerate deposited in a shallow-marine, continuously subsiding basin. A depositional break testi¢es a regressive trend near the top of the group. The lower Middle Cambrian Solva Group consists of roughly 500 m of sandstones and shales deposited in relatively high-energy waters, whereas the upper Middle Cambrian, about 250 m of pyritous mudstones,

PALAEO 3071 9-5-03

14

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

termed the Menevian Group, suggests quiet, oxygen-depleted environments. The Upper Cambrian ‘Lingula Flags’ (more than 600 m thick) are shallow-water storm-in£uenced deposits of a rapidly subsiding marine basin. The shelf deposits in the English Midlands commence with Early Cambrian sandstones, the Hartshill Formation (260 m thick), which shows progressive deepening from a basal shoreface sand. There follow 600 m of outer-shelf mudstone, the Stockingford Shales, which extend from the higher Lower Cambrian to the top of the Upper Cambrian. The upper Middle Cambrian and the uppermost Upper Cambrian show strong lithological and palaeontological similarities to the correlative rocks in Wales, though the formations are thinner. The upper Middle Cambrian to Tremadoc succession is relatively complete. The successions in Shropshire, on the margin of the Midlands Platform, partly deposited in fault-bounded depocentres (Smith and Rushton, 1993), suggest the in£uence of shallow water almost throughout. At the top of the Comley Series are the condensed Comley Limestones that represent shallow-marine carbonates containing the best preserved and most important late Early Cambrian (in the traditional British sense) trilobite faunas of the British Isles. In summary, the lowest Cambrian deposits were deposited unconformably on a variegated volcanic basement in both Wales and England, whereas younger Cambrian sediments in Wales formed in a subsiding ensialic basin, while their English equivalents formed on the subsiding Midland Platform. 3.4. The Armorican Massif The Cambrian rocks of the Armorican Massif unconformably overlie the Cadomian basement. The most representative and complete Lower Cambrian succession is that of the North Contentin Syncline. A SW^NE-trending trough was ¢lled by a succession ranging from 590 to 2100 m in thickness (Dore¤, 1994) with, from bottom to top: (1) the Couville Formation, up to 400 m thick, composed of conglomerates and sandstones interbedded with a volcanoclastic complex; (2) the

Carteret Formation, ca. 1000 m, consisting of littoral alternations of shales and silty sandstones rich in microfossils and ichnofossils of Cordubian age; and (3) the Saint-Jean-de-la-Rivie're Formation, ca. 200 m thick, made up of peritidal alternations of limestones and siltstones, containing salt pseudomorphs, microbial^archaeocyathan buildups and oolitic shoals. Other Lower Cambrian successions are known in central and southern Normandy and Maine, and comprise small shelly fossils, brachiopods and ichnofossils. A post-orogenic volcanism took place along two volcano^ tectonic structures : ignimbrite sequences underlying Lower Cambrian sediments, and the Maine graben interbedded with Lower Cambrian to Arenig? sediments. Other Cambrian fossiliferous rocks are incompletely known, such as a Middle Cambrian trilobite assemblage from the northeastern Vende¤an Massif and a suspected Late Cambrian brachiopod assemblage from the Chantonnay synclinorium. 3.5. The Montagne Noire and southwestern Sardinia Both areas display the same three-fold lithosome subdivision as that reported in the CZ and WALZ for Early to Mid-Cambrian times. The Neoproterozoic^Cambrian transition is not clearly identi¢ed in the former, where the lower terrigenously dominated lithosome begins with the Marcory (more than 1000 m thick) and Matoppa (300^600 m) Formations. The second, Ovetian^Bilbilian, lithosome represents the establishment of carbonate platforms; in the southern Montagne Noire these are subdivided into the Pardailhan, Lastours and Pont de Poussarou Formations (100^500 m thick), and in Sardinia the Punta Manna, Santa Barbara, Planu^Sartu and San Giovanni Formations (600^1200 m thick). These platforms were widely colonised by buildups of microbial-archaeoyathan communities, associated in Sardinia with a tensional tectonic activity re£ecting the establishment of an isolated carbonate platform (see a summary in Perejo¤n et al., 2000). In both areas the Lower^Middle Cambrian transition displays the tectonic instability already reported in the CZ and the IC, with de-

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

15

velopment of palaeotopographies and deposition of the ‘griotte’ facies (La Tanque and Campo Pisano Formations ; up to 80 m thick). After drowning of the platforms and burial of the palaeotopography by o¡shore shaly sediments (Coulouma and lower Cabitza Formations ; 30^80 m thick), the evolution in both areas di¡ered drastically. The southern Montagne Noire exhibits the progradation of sandy shoals (Ferrals Formation ; 130^200 m thick) and tidally induced systems (La Gardie Formation ; 0^500 m thick), followed by important carbonate developments leading to deposition of new mixed carbonate^siliciclastic sediments on intra-shelf ramps (Val d’Homs Formation ; 60^200 m thick). This episodic carbonate sedimentation is unique across the whole western Gondwanan margin for Late Cambrian times. By contrast, southwestern Sardinia shows a Middle^ Upper Cambrian shale-dominated deposition (Cabitza Formation), distinctly condensed and remarkably uniform (ca. 190 m thick ; Loi et al., 1995), which gradually passes into the Tremadocian.

deposited on an open siliciclastic platform. The Middle Cambrian of the Franconian Forest consists of isolated outcrops that represent at least ¢ve di¡erent formations, in ascending order the Galgenberg, Wildenstein, Triebenreuth, Lippertsgru«n and Bergleshof Formations, of unknown thickness, all characterised by signi¢cant trilobite faunas. The Early Cambrian of the Go«rlitz Syncline is very incomplete, but has yielded important trilobites. The Charlottenhof Formation is a fragment of a carbonate platform and represents the transition from rather shallow to deep subtidal conditions (Ludwigsdorf Member) and subsequently to deeper platform environments (Lusatiops Member). The age of the formation is late Marianian or late Banian (Elicki, 1994; Geyer and Elicki, 1995). Finally, the Cambrian (s.l.) of Thuringia (Berga Anticline) is known only from two boreholes: it is a slightly metamorphosed carbonate succession that has yielded a poorly preserved microfauna, with siliciclastic deposits at the base (Elicki, 1997).

3.6. Saxo^Thuringia

3.7. Bohemia

Lower and Middle Cambrian successions occur in the Doberlug^Delitzsch and Go«rlitz synclines, the Franconian Forest and Thuringia (Berga Anticline). Early Cambrian relationships between them are controversial because they are of di¡erent ages and the limited occurrence of fossils. Upper Cambrian successions are generally not preserved, as a result of uplift combined with denudation and weathering (Linnemann and Buschmann, 1995). The incomplete Early Cambrian succession of the Doberlug Syncline represents a carbonate ramp with archaeocyathan-microbial buildups (Zwethau Formation, more than 700 m thick) developed during early Ovetian times on the Cadomian basement of Saxo^Thuringia (Elicki, 1999). The Lower^Middle Cambrian transition is unknown because of the scarcity of outcrops and structural di⁄culties. The Middle Cambrian of the Doberlug area (Tro«bitz and Delitzsch Formations, ca. 600 m thick), early Leonian^Caesaraugustian in age, is relatively fossiliferous (Sdzuy, 1972). The sediments were

The Pr›¤|bram^Jince Basin of the Barrandian area contains about 2000 m of Lower Cambrian continental conglomerates and sandstones, punctuated by the deposition of sediments in a brackish environment (the Paseky Shale) containing endemic merostome( ?) arthropods and possibly phyllocarids (Chlupa¤c›, 1995). Deposition of Middle Cambrian transgressive marine fossiliferous shales was accompanied by two major palaeogeographical changes : (1) an eastward shift of the area of maximum subsidence, and (2) a change in orientation of the main structural directions and longitudinal basin axis from a NE^SW to a NW^SE trend (Havl|¤c›ek, in Chlupa¤c› et al., 1998). The Jince Formation (up to 400 m thick) consists of shales and ¢ne-grained sandstones containing a rich and diverse fossil fauna of trilobites, brachiopods and echinoderms. The unfossiliferous Upper Cambrian deposits consist of volcanic complexes with rhyolitic e¡usions and pyroclastic layers. In the Skryje^Tyr›ovice area, the Cambrian succession consists only of the lower part of the Jince

PALAEO 3071 9-5-03

16

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Formation, which rests unconformably on Upper Proterozoic rocks. These outcrops re£ect a Middle Cambrian deposition in a ‘channel-like’ trough a¡ected by synsedimentary deformation and partly eroded during Upper Cambrian regressive conditions. The thickness of the Middle Cambrian decreases from about 200 m in Skryje to several tens of metres towards the northeast.

4. Palaeooceanographic and palaeoclimatic constraints The western Gondwanan margin during the Cambrian is commonly portrayed as oriented approximately SW^NE in a position on the Southern Hemisphere (McKerrow et al., 1992). As a consequence of this orientation, the shape of the tropical zones was a¡ected by the course of major ocean currents transporting cool waters towards the tropics (Fig. 4). This would have caused the tropics to occupy a narrower latitudinal area on this western margin compared with the eastern margins (Briggs, 1995). This inferred palaeogeographical situation may be compared with that of the western Australian margin during the Eocene (McGowran et al., 1997), when Australia and Antarctica were close enough to prevent the development of a circum-Antarctica circulation. As a result, two main surface circulation patterns took place : a South Equatorial Current £owing (up to the Recent) as a low-salinity surface component of the global thermocline system, and a

western Australian Current (parallel to the western face of Australia) that disappeared with the Eocene opening of a gap between both continents. Apparently a similar western Gondwanan SW^ NE current rising in latitude from polar areas operated during Cambrian times, which obviously a¡ected faunal migrations and the palaeoclimates. The lack of extensive carbonate deposition across the Precambrian^Cambrian transition in the western Gondwanan margin, except the Adoudou and Lie-de-vin Formations in Morocco, can be explained by a combination of active Cadomian orogenic phases, and active source regions, which supplied coarse-grained siliciclastics and inhibited carbonate production. Much of our understanding of the Lower Cambrian palaeoclimate evolution on the western Gondwanan margin comes exclusively from lithological indicators of climate, such as limestone (mainly reefal) and Ł lvaro et al., 2000a). evaporite precipitation (A Lower Cambrian salinities seem to have been episodically and locally high enough to generate widespread deposition of evaporates, suggesting a generally subtropical arid climate. Although an Early Cambrian warm period has been claimed, global warming does not necessarily imply a poleward shifting of isotherms but might re£ect their spacing when thermal gradients became £atter. During Early and earliest Mid-Cambrian times, the western Gondwanan margin was rimmed by enormously extensive shallow-water subtropical platforms, where carbonate deposition was some-

Fig. 4. Global Cambrian palaeogeography (modi¢ed after McKerrow et al., 1992) and proposal of palaeocurrents discussed in the text.

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

times outpaced by local, craton-derived terrigenous clastic sediments. As a result, the proportion of carbonate and siliciclastic rocks varies both regionally and stratigraphically. Incursions of shale onto the platforms is commonly related to sea-level rises, when the remobilisation and transport seaward of ¢ne-grained siliciclastic sediments outpaced carbonate production. By contrast, regressive trends are represented by basinward shifts of peritidal deposits in carbonate and mixed platforms, and deltaic and coastal plain dilution related to terrigenous in£ux on platforms close to continental sources. For instance, the siliciclastic deposition related to the Daroca regression during the earliest Tissa¢nian or early Bilbilian age, affected the Souss Basin, the OMZ, CIZ and IC platforms, but was absent in the CIZ, WALZ, Montagne Noire and southwestern Sardinia platforms. Despite the lack of lithological indicators for the climate during the Mid- to Late Cambrian, a transition from carbonate-dominated rimmed platforms to siliciclastic temperate- to cool-water platforms is attributed to changing oceanographic conditions across the Lower^Middle Cambrian transition. The southward-moving warm tropical waters of western Gondwana progressively invaded cold seas; as a result, Middle Cambrian sediments re£ect a transition to temperate-water platform sediments. In a study of the Cambrian Ł lvaro and Vennin (2001) Iberian platform (IC), A described a succession of benthic communities interpreted to have developed during a cooling period. This cooling trend, from subtropical to temperate waters, was marked by the successive decline and disappearance of: (1) ooids, evaporite relics and microbial carbonates ; (2) eocrinoidsponge meadows; (3) hydrodynamic bioclastic carbonates related to the contraction of carbonate factories into increasingly narrow belts (‘griotte’ facies) ; and there are trends to (4) an increase of trilobite-dominant communities on muddy bottoms; and (5) to decreasingly fossiliferous communities, then dominated by linguliform brachiopods in sandy shoals. The Languedocian^Late Cambrian interval seems to mark the culmination of the cooling trend as shown by the pervasive decrease in diversity, associated with the drastic

17

disappearance of bioclastic carbonates. On the other hand, the development of Late Cambrian isolated carbonate platforms within temperatewater settings of the southern Montagne Noire (Val d’Homs Formation) signals short-lived warming episodes in meridional parts of the western Gondwana margin extending into Tremadocian times.

5. Biogeographical patterns 5.1. Lower Cambrian The Lower Cambrian has for a long time been based on the concept of being characterised by the earliest Phanerozoic fossils. Bro«gger (1886) and subsequently Walcott (1891) developed the idea of a Lower^Middle Cambrian boundary based on the ¢rst appearance of the trilobite genus Paradoxides (s.l.) in Baltica and the disappearance of Olenellus (s.l.) in Laurentia. The lower boundary of the Cambrian was not ¢xed with any precision but in early days was assumed to coincide roughly with the ¢rst appearance of trilobites. Since 1974 work by the International Subcommission on Cambrian Stratigraphy has changed this concept. Incorporation of strata with ‘primitive’ Metazoan sclerites of uncertain systematic a⁄nities (the so-called small shelly fossils) pushed down the level of the Cambrian lower boundary considerably. In the present concept, the earliest Cambrian strata are characterised by rocks with the ¢rst diversi¢ed trace fossil assemblages that underly strata with small shelly fossils. As a result, the Lower Cambrian was extended dramatically and the trilobite-bearing strata represent less than half of Early Cambrian time as presently understood (Landing, 1994; Landing et al., 1998). In the Lower Cambrian, the ideal of a global biostratigraphy and palaeobiogeography su¡ers from both a relatively limited diversity of trilobites and their pronounced endemism. Furthermore, the distribution of trilobites is, from the earliest times, strongly controlled by facies, so that even a precise interregional correlation is dif¢cult. The most complete Early Cambrian trilo-

PALAEO 3071 9-5-03

18

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

bite succession of western Gondwana, with relatively diverse trilobite faunas from various biofacies, comes from the Moroccan Atlas ranges (Fig. 5A), with more than 200 trilobite species described from Lower and lowermost Middle Cambrian strata. This area clearly is the key for our understanding of the Lower Cambrian palaeobiogeography of western Gondwana and the neighbouring regions. Historically, this is the region where the ¢rst elaborate Lower Cambrian biozonation was developed (Hupe¤, 1952). The Lower Cambrian to lowermost Middle Cambrian biostratigraphy of Morocco now comprises ten biozones assigned to three stages (Fig. 2), which are all characterised by signi¢cant trilobite faunas and separated by faunal turnovers (Geyer, 1990a; Geyer and Landing, 1995). Recent correlations suggest that the earliest trilobites from the lowermost Issendalenian Stage pre-date other trilobite occurrences in western Gondwana and are among the ¢rst trilobites known so far on a global scale (Sdzuy and Geyer, 1988; Geyer and Landing, 1995). These faunas of the Eofallotaspis Zone include several species of the fallotaspidid genus Eofallotaspis and the bigotinid genera Hupetina and Bigotinops, all of which are endemic, as are most trilobites of the overlying Fallotaspis tazemmourtensis, Choubertella and Daguinaspis Zones. The entire Issendalenian is characterised by fallotaspidid, redlichiid and bigotinid trilobites. Fallotaspidids (the genera Choubertella, Daguinaspis, Eofallotaspis, and Fallotaspis) appear to have preferred ¢ne-grained siliciclastic substrates and are now found almost exclusively in shales ; deposition took place on a slightly restricted carbonate platform with archaeocyathan-microbial buildups, apparently with limited connections to the open sea. Elsewhere fallotaspidids (in a broad sense) occur with certainty in the Lower Cambrian of the White^Inyo Mountains, California, adjacent Nevada, and Siberia. Redlichiid^saukiandid trilobites (such as Marsai-

19

sia, Pararedlichia, and Resserops) are also typical in ¢ne-grained siliciclastics. They share characters with genera from other Cambrian regions such as the South China (or Yangtze) Platform (Nangaoian Stage) and are sometimes erroneously synonynised with Eoredlichia because of the limited set of characters. Lemdadella and bigotinid trilobites are typical of well-bedded platform carbonates. They may be helpful for future correlations but are still imperfectly known because the past research concentrated largely on shale faunas. Similar bigotinids are known from the Lower Cambrian of the Armorican Massif and the Co¤rdoba platform of the OMZ (Bigotina) but also from the Lower Cambrian Pagetiellus anabarus Zone (middle Atdabanian) of the Siberian Platform, in both areas also in platform carbonates. Finally, Lemdadella has been reported from the Lower Cambrian of the OMZ (Lin‹a¤n and Sdzuy, 1978) and Antarctica (Palmer and Rowell, 1995). In summary, the Issendalenian shows a moderate diversi¢cation of trilobite faunas with high preference for ¢ne-grained siliciclastics and peritidal carbonate muds. Most species and the majority of the genera are endemic, which is probably due to a lack of knowledge of coeval faunas from other western Gondwanan areas. The Issendalenian^Banian transition in Morocco is marked by a dramatic decrease of the fallotaspidids, whereas primordial ellipsocephaloids (the antatlasiids) arise and increase impressively in number and diversity to form the most characteristic faunal element in the Banian. Fallotaspidids are replaced by neltneriids. The ¢rst eodiscoids (Delgadella, Hebediscus) occur in the lower Banian. All trilobites apparently have noticeable preferences for distinct lithofacies, but the framework is more complicated than the simple distinction between shale and carbonate facies in the Issendalenian. Banian rocks are dominated by ¢ne-grained siliciclastics with relatively minor

Fig. 5. Global distribution chart illustrating presence (black) and absence (white) of Lower Cambrian trilobite genera reported in the western Gondwanan margin (A), biodiversity patterns (B), and analysis of hierarchical Phi^Pearson similarity (C). Some of the reported genera could occur in Fig. 6 as Middle Cambrian trilobites according to the considered horizon of the Lower^Middle Cambrian boundary.

PALAEO 3071 9-5-03

20

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

nodular and bedded limestones deposited in a low-energy middle to upper platform environment. During the Banian, for the ¢rst time, strong similarities of trilobite faunas can be detected between the Moroccan Atlas region and the OMZ. Delgadella is found in the lower, middle and even in the upper part of the Banian (A. hollardi, A. guttapluviae and Sectigena Zones) and is also a frequent faunal element in the earliest assemblages of the OMZ. The Sectigena Zone of Morocco shares a number of genera and even species with the lower two assemblages of the OMZ, such as Andalusiana cornuta, Gigantopygus, Rinconia, Saukianda andalusiae, Termierella and Triangulaspis. This distinctive zone is represented in Morocco by several quite di¡erent local assemblages that portray a variety of coeval biofacies. Stratigraphical successions dominated by low-energy shales include the Ossa^Morena-type Saukianda^Termierella assemblage. Shales to ¢ne-grained sandstones deposited in more agitated shallow-marine environments tend to yield a Berabichia^Delgadella^ Pentagonalia fauna. Limestones and marls yield an assemblage with Hebediscus and Sectigena. Other trilobites are facies-independent and occur in di¡erent types of rocks. Issafeniella has been identi¢ed from the Marianian of the OMZ, which also shares a ¢rst and simple species Acanthomicmacca) with the Moroccan Sectigena Zone. Sectigena itself, Berabichia and Issafeniella, have their counterparts, and obviously close relatives, in the Siberian genera Charaulaspis and Chorbusulina. Both occur in the latest Atdabanian and early Botoman of the northern Siberian Platform, where they are accompanied by Hebediscus spp. and Triangulaspis. Two additional species of Chorbusulina were described from the Lower Cambrian of Antarctica, but in fact could belong to Berabichia. An important genus that has its ¢rst appearance in the Sectigena Zone is Serrodiscus. This genus has a long stratigraphical range, but it is remarkable that the same type of red shales of the Issafen Formation, in which Serrodiscus ¢rst occurs in the Anti-Atlas, also contains the ¢rst Serrodiscus species (S. silesius and other species) in the Marianian of the CIZ and OMZ, and in the

Charlottenhof Formation of the Go«rlitz area, Saxo^Thuringia (S. silesius). The German and Spanish Marianian assemblages also share the genera Ferralsia and Lusatiops. Ferralsia is known from the OMZ, but was originally described from the uppermost(?) Lower Cambrian of the Montagne Noire. An Acanthomicmacca, although dif¢cult to judge in terms of precise ontogeny, is also known from the Charlottenhof Formation. A probable species of Holmia is also identi¢ed from the Charlottenhof Formation, and species tentatively assigned to Holmia and Kjerul¢a were described from the lower? Sectigena Zone of the Anti-Atlas. Both genera are distinctive of the Holmia kjerul¢ Zone of Baltica. The faunas of the Moroccan Sectigena Zone and the coeval assemblages from Iberia, Saxo^ Thuringia, the southern British Isles, southeastern Newfoundland and other areas are characterised by an assemblage that include relatively widespread eodiscoids. Typical species of this assemblage include Calodiscus helena, C. lobatus (in strict sense), C. schucherti, Hebediscus (or Dipharus) attleborensis, Serrodiscus bellimarginatus, S. silesius, S. speciosus, Triangulaspis annio, T. vigilans, and other species of Triangulaspis (Fletcher, 1972; Robison et al., 1977; Geyer, 1990b; Geyer and Palmer, 1995). Most of these species are relatively endemic, but their ranges show some regional overlap and associated taxa permit a subglobal recognition, for example in Baltica (Sweden), Avalonia (southeastern Newfoundland, New Brunswick, Welsh Borderlands), western Gondwana (Spain, Morocco), Siberia, the Altay^Sayan Mountains, and Kazakhstan. In western Gondwana, Triangulaspis zirarii of Morocco has its counterpart in the extremely similar Triangulaspis fusca from the OMZ, which is associated with Calodiscus schucherti, a species that also occurs in the Moroccan Sectigena Zone. Such assemblages are the earliest to show clear faunal relationships between Avalonia and western Gondwana. The top of the Callavia broeggeri Zone in southeastern Newfoundland (regarded as a distinct Serrodiscus bellimarginatus Zone by Fletcher, 1972) includes not only Serrodiscus bellimarginatus and Hebediscus attleborensis, but also Acanthomicmacca species typical of the Sectigena

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Zone, the Marianian of the IC, and the Charlottenhof Formation in Saxo^Thuringia. It also includes Mallagnostus (formerly Ladadiscus ; see Whittington et al., 1997) llarenai, a species ¢rst described from the OMZ. Very similar trilobites, including Acanthomicmacca, Calodiscus lobatus, Hebediscus attleborensis, Serrodiscus bellimarginatus, and Triangulaspis annio, are known from various levels in the Comley Limestones in the Welsh Borderlands of England. Some eodiscoids, such as Calodiscus, Hebediscus, Serrodiscus, and Triangulaspis, are shared with assemblages from the top of the Judomia and B. micmacciformis^Erbiella Zones on the Siberian Platform. The Lower Cambrian of southwestern Sardinia has yielded a relatively diverse trilobite fauna with a pre-Sectigena age. Most of the species and even genera are endemic, probably because they are largely from lithologies incompletely studied in supposedly coeval successions from other parts of western Gondwana. Particularly well known are the genera Dolerolenus, Giordanella, and Metadoxides. The few eodiscoids in these assemblages are of little help for either biostratigraphical or biogeographical correlations because they are also endemic. Another so-called Dolerolenus fauna has been described from the Ovetian of the IC (Sdzuy, 1987), an assemblage representing the oldest diverse trilobite fauna known from Spain. However, the trilobites are mostly described under open nomenclature because of their poor preservation. Nevertheless, the taxonomic a⁄nities are su⁄cient to recognise that both areas share a similar trilobite fauna and obviously were biogeographically connected. Both may be regarded as faunas particularly adapted to relatively isolated, probably rimmed, carbonate platforms. In a monographic reappraisal of the Sardinian trilobites, Pillola (1991a) described a number of new, endemic species and revised older species, mainly redlichioids. He assigned many of them to genera known almost exclusively from the South China Platform and proposed a strong palaeobiogeographical connection to that Cambrian continent. However, trilobite faunas of similar age and from similar facies are unknown from the platforms supposed to have been situated between those regions during the Lower Cambrian.

21

Pillola’s (1991a) revision partly a¡ected the nomenclature of trilobites from the Lower Cambrian of the Montagne Noire. Galloredlichia noiri was assigned to the ‘Yangtze’ genus Eoredlichia, a genus that occurs in relatively early trilobite assemblages on the South China Platform such as the famous Chengjiang Fauna. This assemblage can be correlated with the Uper Issendalenian and Upper Ovetian. In the Montagne Noire, Galloredlichia occurs together with Granolenus midi, a species also known from the Upper Ovetian of the CIZ. A Banian^Tissa¢nian faunal turnover is marked by a relatively sharp change in the composition of benthic communities, exempli¢ed by: (1) the appearance and massive increase of protolenine trilobites, (2) the sudden but apparently stacked appearance of other new trilobite groups, and (3) the relatively abrupt disappearance of ellipsocephaloid lineages. These strata with Lower^ Middle Banian faunas are traditionally termed the ‘Protolenus Zone’ and regarded as of Early Cambrian age. Early Cambrian (Comley Series) trilobites are known from Comley and the Rushton area in the Welsh Borderland and at Nuneaton on the Midland Platform, together with a very few, including Hamatolenus and Pseudatops, from northern Wales. There has been little recent revision of British Early Cambrian trilobites, and in view of signi¢cant recent work in Morocco, Spain and Germany, the generic placement of certain British taxa, particularly of ellipsocephaloids, is now open to question. Thomas et al. (1984) listed some 36 taxa from the Comley Series. If we accept the names currently applied, their a⁄nities lie most strongly with trilobites from elsewhere in Avalonia. About 12 species are common to western Avalonia, including some eight of eodiscoids and species of Protolenus, Strenuella and possibly of Callavia ; and a further 12 represent genera known from Avalonia and the Taconic Belt of New York. In contrast, only two species and three or four genera can be compared with Baltic faunas. The a⁄nities with marginal Gondwana appear to be restricted to three species known from the OMZ ^ Calodiscus schucherti, Mallagnostus llarenai and possibly Serrodiscus serratus ^ and

PALAEO 3071 9-5-03

22

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Fig. 6. Global distribution chart illustrating presence (black) and absence (white) of Middle Cambrian trilobite genera reported in the western Gondwanan margin (A), biodiversity patterns (B), and analysis of hierarchical Phi^Pearson similarity (C).

there are another 12 genera in common. The British representatives of these genera are: Callavia callavei and other species, Cobboldites comleyensis, Comluella platycephala (Morris, 1988, has

Strenuella (Comluella)), Condylopyge amitina, Ellipsocephalus (or Ourikaia ?) heyi (see Geyer, 1990b), Hamatolenus douglasi, Hebediscus attleborensis, Latoucheia latouchei, Micmacca spp.,

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Protolenus paradoxoides, Serrodiscus bellimarginatus and other spp., and Triangulaspis annio. According to Sdzuy (1961), possibly Andalusiana ? (or Kjerul¢a?) granulata should also be included. The greatest resemblance is shown by faunas from the ‘Bre'che a' Micmacca’ in the basal Middle Cambrian in Morocco. Geyer and Malinky (1997) correlated this unit with the Protolenus Limestone of Comley, Shropshire (Cobbold’s unit Ac5; see Rushton et al., 1999), which in Britain was treated hitherto as late Early Cambrian. 5.2. Middle Cambrian As discussed above for the Lower Cambrian trilobites of the western Gondwana margin, the database from which the following analysis is made is heterogeneous. This is due to the existence of barren intervals, across the Leonian in the Montagne Noire and the Middle^Late Languedocian in Morocco and the OMZ, which prevent complete correlations (Fig. 6A). The Lower^ Middle Cambrian boundary is not uniform over the whole area, because it has been characterised by di¡erent bioevents based on trilobite and acritarch appearances. The Lower^Middle Cambrian transition contains numerous barren stratigraphical intervals and unconformities, so that the transition (and not its boundary) will be described below in two key areas with continuous sedimentation and fossil wealth: Morocco and the IC. In the present state of knowledge, two di¡erent biodiversity patterns are recognised during MidCambrian times: (1) a progressive decrease in diversity in Morocco, and (2) a composite increase^ decrease trend in the whole of southwestern Europe. Consequently, during Early to Middle Cambrian times, the area of maximum diversity migrated northeastwards ; a result that may be related to the southwestward drifting of the western Gondwana margin. In general, two Middle Cambrian trilobite diversi¢cations are recognised (Fig. 6B): (1) a major migratory radiation of trilobites during the Tissa¢nian/Upper Bilbilian to Caesaraugustian when most families appeared without obvious ancestors in the fossil record; and (2) a Late Languedocian trilobite diversi¢cation connected with transgres-

23

sive episodes and representing new migratory inputs of recurrent trilobite genera. Both episodes were punctuated by distinct faunal turnovers. Several lines of evidence have been put forward that imply extinction events towards the end of the Early Cambrian. They point to the concept of a widespread latest Early Cambrian biocrisis, as the culmination of a worldwide decline of trilobites and other taxa that commenced earlier in the Early Cambrian and led to a global diachronous extinction (Debrenne, 1991; Brasier, 1995; Ł lvaro et al., 1999). The termiZhuravlev, 1995; A nal Lower Cambrian strata display a succession of extinction events related to the disappearance in the western Gondwana margin of microbial^ archaeocyathan buildups and the establishment of regressive conditions. A local event has been identi¢ed in the IC at the end of this tendency, termed the ‘Valdemiedes event’. Data from the trilobites do not necessarily support a catastrophic extinction at this horizon, their record being imperfect on account of barren intervals prior to that event. The event represents the end of the relatively endemic, inner-platform survivors of the area, followed by an inshore migration of more pandemic, outer-platform faunas. It should be noted that during the Lower^Middle Cambrian transition the Moroccan succession shows a di¡erent pattern from that portrayed by southwestern Europe. In the Atlas ranges, the terminal Lower Cambrian^basal Middle Cambrian regressive^transgressive trends are represented by a marked coarse clastic input, regionally overprinted by volcanoclastic deposits. Two distinct unconformities exist, one within the sandstone^ volcanoclastic unit (which is accordingly divided into the Tazlaft and Tatelt Formations), and another one at its top. The fossil record, however, is rich across both formations and suggests that a punctuated extinction did not take place. Although the faunal turnover reported across the Banian^Tissa¢nian transition signals a strongly depauperate fauna in the basal Tissa¢nian Hupeolenus Zone, some species range across the boundary; this turnover pre-dates the ‘Valdemiedes event’ of the IC, which roughly coincides with the base of the Cephalopyge notabilis Zone. The fossil poverty of the Hupeolenus Zone fauna can

PALAEO 3071 9-5-03

24

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

largely be attributed to unfavourable living conditions during the pronounced input of terrigenous clastics and volcanogenic components. By contrast, the C. notabilis Zone includes one of the most diverse trilobite assemblages in the world known from this time. Its base is regionally placed within the clastic blanket so that the diversi¢cation can be proved to be a true radiation rather than merely created by a favourable diversity of ecological niches. Nevertheless, the peak diversity was not reached until the establishment of a number of diverse facies conditions on the basal Middle Cambrian platforms. As noted above, the strata with Lower^Middle Banian faunas are correlated with the ‘Protolenus Zone’ in areas such as eastern Avalonia (England and Wales), western Avalonia (southeastern Newfoundland and New Brunswick) and Baltica, and are regarded as of Early Cambrian age. The earliest Mid-Cambrian faunal recovery can be regarded as an adaptative radiation of an immigrant trilobite community that progressively invaded the western Gondwanan margin from deeper-water environments. The Tissa¢nian/Leonian immigration is characterised by the stepwise appearance of relatively cosmopolitan trilobites, such as the acrocephalitids, agraulids, corynexochids, conocoryphids, paradoxidids, solenopleurids and agnostoids (condylopygids, phalacromids and spinagnostids). The observed pattern of diversi¢cation was not facies-controlled because trilobites appeared in a wide diversity of facies and environments. Some trilobite families crossed the Lower^Middle Cambrian transition, although decreasing sharply in diversity, such as the ellipsocephaline and protolenine trilobites. Subsequently, the western Mediterranean area apparently operated as a centre of evolutionary radiation (or a site of speciation), from which the species were able to spread outward into adjacent platŁ lforms, in some cases by sympatric speciation (A varo and Vizca|«no, 2001). As a result, a relative biogeographical concordance of evolutionary patterns is virtually complete for some trilobite families, and there are no apparent extrinsic barriers separating the sibling species. Some Caesaraugustian benthic species evolved in this area and proved capable of migrating to colonise areas

from Newfoundland and Rhode Island to Turkey (e.g. Badulesia tenera). There were no physical barriers preventing trilobite dispersal, which was favoured by uniform sea£oor conditions (openplatform, muddy o¡shore substrates), and low temperature-gradients. Diversity increased and species proliferated during Caesaraugustian times. The acme is associated with transgressive conditions and subsequent connection of platforms, and represents the peak in diversity of trilobite families, genera and species during the Middle Cambrian. A major reduction in the number of trilobite genera occurred across the Caesaraugustian^Languedocian transition, at a time of widespread coarse-grained terrigenous input associated with a well-documented regressive trend. The disappearance of trilobite taxa is related to the disappearance of their habitats. This diachronous decline was catastrophic for some previously successful trilobites, such as the solenopleuropsinae (except the genus Sao). This biocrisis is not a major extinction at familial level for trilobites, but represents a signi¢cant generic turnover. The faunal turnover coincides with a rapid prograding and shoaling that produced widespread areas of coarse-grained sandstones and, therefore, seems to re£ect geographically extensive environmental changes. In Baltica there is a regression at the level of the late Middle Cambrian Andrarum Limestone (Nielsen, 1996), and this is recognised in the southern British Isles and Newfoundland as an ‘Andrarum Limestone regression’, with the laevigata Zone above. It is unlikely that the last Languedocian fauna is as young as the L. laevigata Zone (see Fig. 2); paradoxidids appear to be extinct by then in Baltica and a new fauna occurs with trilobites of Upper Cambrian aspect, such as Agnostus, Andrarina, and Proceratopyge. The Late Languedocian trilobite diversi¢cation is well documented in the southern Montagne Noire, Sardinia, and Saxo^Thuringia, where it is characterised by an abrupt increase in genera and families. The appearance of this new assemblage of trilobites represents an immigration event related to transgressive pulses. Episodically, o¡platform trilobite taxa migrated towards the inner platform, leading to a strict facies control on the

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

lithofacies di¡erentiation, and coinciding with the establishment of suitable muddy o¡shore substrates. However, the diversity of the new groups is less than in the previous radiation. Dispersal took place to some extent, but the presence of some provincialism suggests that species exchange was selective, and that the oceanic areas may have served as partial barriers to migration. Late Languedocian trilobites include typical genera of the Leonian^Caesaraugustian diversi¢cation (such as the conocoryphids and paradoxidids) and new Asian invaders, such as the genera Abharella, Chelidonocephalus, Derikaspis and Dorypyge (see Ł lvaro et al., 1999, and references therein). This A in£uence illustrates the immigration of faunas from the northern Gondwanan margin, which went on during Late Cambrian times (see below). However, the percentage of pandemic species decreased across the Languedocian, and geographic distribution also decreased because of successive regressive pulses. The Middle Cambrian of the British Isles (St. David’s Series) is in the lower part generally developed as arenaceous clastic deposits, and in Shropshire (and possibly elsewhere also, though the evidence is not so clear) they overlie a break in the succession. These deposits contain a restricted fauna of agnostoids, paradoxidids and a few other taxa such as Bailiella. Higher up, dark mudstones (the classical ‘Menevian Beds’) are widespread. They contain a richer fauna of agnostoids, paradoxidids and various polymerid trilobites, commonly including blind genera. These dark beds are overlain by a hiatus that in most places is evinced by erosion during the marine low-stand corresponding to the Solenopleura (now Erratojincella) brachymetopa Zone (Rushton, 1978; Nielsen, 1996). Thomas et al. (1984) listed 106 named St. David’s taxa, and more recent work has added a few more: in St. Tudwal’s Peninsula, northern Wales, Young et al. (1994) ¢gured forms derived from the brachymetopa Zone, namely Bailiaspis glabrata (as a senior synonym of B. nicholasi), Dolichometopus cf. svecicus, Linguagnostus aristatus, and fragments of Acrocephalites, Dorypyge and Centropleura, whilst Bridge et al. (1998) have ¢gured Luhops expect-

25

ans, L.? pugnax (Clarella in Morris, 1988) and Paradoxides abenacus. The faunas are most closely related to those of Baltica, with which 40 species are in common, and a further 14 at generic level, and with western Avalonia (mainly Newfoundland), with which 34 species are in common and a further fourteen at the generic level. Resemblance with the Gondwanan margin is considerably lower: there are eight species in common with Bohemia and Saxo^Thuringia, namely Condylopyge rex, Eccaparadoxides pusillus, Luhops expectans, Peronopsis integra(?), Phalacroma bibullatum, Phalagnostus nudus, Pleuroctenium granulatum granulatum, and Plutonides (Hydrocephalus) hicksii. Further genera in common appear to include Acadoparadoxides, Agraulos, Ctenocephalus, Dawsonia? (if Metadiscus is a synonym), Holocephalina, Hypagnostus, Paradoxides, Ptychagnostus, and Ptychoparia? There are about six species in common with the Iberian Peninsula, namely Agraulos longicephalus, Condylopyge carinata, C. rex, Paradoxides davidis ?, Peronopsis fallax?, Plutonides hicksii, and, from Sardinia, Anopolenus henrici ? (Loi et al., 1995). The only British member of the important peri-Gondwanan family Solenopleuropsidae is Solenopleuropsis (Manublesia) variolaris, which occurs in the punctuosus Zone. This assessment shows once again that the agnostoid trilobites are the most widely dispersed of taxa, with Condylopyge rex being recorded from each of the regions considered here. The resemblance of the British (eastern Avalonia), Newfoundland (western Avalonia) and Baltic faunas rests largely on the similarity of their agnostoids, and doubtless the relative paucity of periGondwanan agnostoids is a hindrance to interregional correlation. Avalonia and Baltica show agnostid faunas intermediate between those of Laurentia and parts of the Gondwanan margin (Conway Morris and Rushton, 1988). Some species of paradoxidids are also relatively widely distributed, although their larger size and commonly fragmentary preservation sometimes makes for di⁄culties in collection and identi¢cation. The polymerid trilobites such as conocoryphids and ptychopariids appear much more frequently to be endemic.

PALAEO 3071 9-5-03

26

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

5.3. Upper Cambrian Upper Cambrian trilobites are very rare in the western Gondwanan margin, compared with the wealth recorded in the Lower and Middle Cambrian, and occur in a variety of lithofacies associations ranging from inner platform sandstones and shales to outer platform and slope carbonates. Temperate-water faunas were less diverse and more impoverished than Lower^Middle Cambrian ones. In contrast, British faunas contain around 100 taxa that represent the whole Upper Cambrian. 5.3.1. The Iberian Massif Late Cambrian trilobites occur in the WALZ, DM and IC. The occurrence of Late Cambrian trilobites in the DM was ¢rst reported by Colchen (1967) who determined the presence of Chuangia and Prochuangia in the lower member of the Najerilla Formation. These trilobites were subsequently described and revised by Shergold et al. (1983), who recognised two trilobite associations : (1) an oldest one comprising four taxa (aphelaspidine, leiostegiid?, Maladioidella colcheni, and Langyashania felixi); and (2) a younger assemblage, from the second member of the Najerilla Formation, containing only two taxa (pagodiine and a solenopleuroidean ? a¡. Lajishanaspis). These determinations support Colchen’s contentions of an Asian in£uence, e.g. Maladioidella has a distribution that extends from Spain to northern China (Rushton and Hughes, 1996 ; Jago, in Brock et al., 2000). In the IC, the ¢rst Late Cambrian trilobites were determined by Sdzuy (in Josopait, 1972), and subsequently described by Shergold and Sdzuy (1991). These occur both in the middle part of the Valtorres Formation and the lower part of the overlying Valconcha¤n Formation. The faunal association includes agnostoid, aphelaspidine a¡. Aphelaspis rara (OrIowski) sensu Z]ylin¤ska (2001), Elegantaspis cf. beta, Parachangshania sp., Pseudagnostus sp., Punctaspis? schmitzi, solenopleuroidean, and Valtorresia volkeri. The Upper Cambrian assemblage is overlain in the uppermost Valconcha¤n Formation by a sandstone bed containing Pagodia (Wittekindtia) alarbaensis

and olenid fragments now thought to represent a species of Jujuyaspis that would indicate an earliest Ordovician age. It is possible that the occurrence of P. (Wittekindtia) in the DM is also of this age. Both there and in the IC, this subgenus is associated with the echinoderm-like Oryctoconus lobatus Colchen and Ubaghs, 1969. By virtue of the aphelaspidoid morphologies in the IC, and Maladioidella in the DM, the trilobites of northern Spain are considered to have an Iverian age according to the Australian standard and a Sunwaptan age on the Laurentian timescale. 5.3.2. Montagne Noire Late Cambrian trilobites occur in the Val d’Homs Formation (Ferrals-les-Montagnes). They were ¢rst described by Feist and Courtessole (1984), who recognised two species referred to Prochuangia gallica and Bergeronites latifalcatus. Subsequent collecting from a limestone lens close to the base of the formation has revealed the presence of further taxa (Shergold et al., 2000): Abharella sp., Ammagnostus (A.) a¡. sinensis, Kormagnostus? sp., Olentella cf. africana, Palaeadotes latefalcata, Paraacidaspis ultima, Proceratopyge (P.) spp., Prochuangia gallica, Shengia cf. spinosa, and Stigmatoa courtessolei. The age of this fauna is judged to be similar to, but slightly older than, that characterised by Maladioidella in the DM. Many of its elements have a widespread distribution along the Gondwanan margin (Shergold et al., 2000), and the agnostoids and Proceratopyge extend even to Laurentia. However, the Ferrals association shows greatest relationship with southern and central China, Australia and Antarctica. Elsewhere in the Montagne Noire, an assemblage containing species of Maladioidella, Onchonotellus, Probilacunaspis, Proceratopyge, Prochuangia, and a pseudagnostinid occurs in carbonates of the Val d’Homs Formation to the southwest of Coulouma. Once more, its composition is dominated by Chinese genera. This assemblage is the youngest Cambrian fauna found so far in the Montagne Noire. The earliest Ordovician is marked traditionally by the appearance of Proteuloma geinitzi described by Sdzuy (1958) from reddish limestones just below the base of

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

the La Dentelle Quartzite in the vicinity of Combes de Barroubio. 5.3.3. Sardinia Although they have not been systematically documented, Late Cambrian trilobites have been noted and illustrated in the upper part of the Cabitza Formation in the Cabitza area and north of Domusnova (Loi et al., 1995, 1996). They occur in ¢ne sandstones and siltstones probably representing a tide-dominated deltaic environment. Illustrated trilobites (Loi et al., 1995, 1996) are from faunal assemblage CAB 5b, which contains Macropyge sp., Maladioidella cf. colcheni, Micragnostus cf. haudei, Niobella cf. primaeva, Onchonotellus ? amsassensis, Proceratopyge sp., a leiostegiid a¡. Pagodia, a eulomid, and a calymenid. At Monte Cani this assemblage is succeeded by green siltstones containing (CAB 6) Rhabdinopora £abelliformis (Pillola and Gutie¤rrez-Marco, 1988; Pillola and Leone, 1993), Proteuloma geinitzi (Pillola, 1991b) and Oryctoconus cf. lobatus. The younger assemblage is clearly of earliest Tremadocian age. 5.3.4. Morocco Two Late Cambrian trilobite taxa have been documented from the upper member of the Jbel Lmgaysmat Formation, in the Foum Zguid region (Destombes and Feist, 1987). Olentella africana and Seletella latigena represent the ¢rst Upper Cambrian occurrence in North Africa. Both genera occur in the Sakian, Aphelaspis^Kujandaspis Zone, in central Kazakhstan (Ivshin, 1955, 1956, 1962). Olentella cf. africana has also been described from Ferrals-les-Montagnes in the Montagne Noire (Shergold et al., 2000). The fossiliferous Upper Cambrian is overlain by a discordance followed by shales containing Rhabdinopora £abelliformis. 5.3.5. British Isles The Late Cambrian (Merioneth Series) trilobites have been found in England and Wales wherever rocks of that age are exposed, and all the main zones are proved. In the lower Merioneth Series there are commonly muddy, silty and sandy beds with Homagnostus and a few species

27

of Olenus. The higher Merioneth is developed everywhere as black mudstones related to the olenid biofacies and contains a considerable variety of olenids, admixed with a small number of agnostoid and other non-olenid genera. A number of new records have appeared since the publication of Thomas et al. (1984) : Allen and Jackson (1985) recorded the presence of Niobella homfrayi preciosa, Parabolina acanthura, P. cf. angusta, and Parabolinella contracta ; Hughes and Rushton (1990) transferred ‘Dikelocephalus’ discoidalis to the ceratopygid genus Cermatops; Cope and Rushton (1992) described previously unknown faunas from Llangynog, south Wales, that include Ctenopyge (C.) linnarssoni ; Howells and Smith (1997) recorded the presence of Loganellus sp., Maladioidella abdita (‘Conokephalina’ in Morris, 1988), Olenus cf. solitarius, Parabolina a¡. mobergi and P. (Neoparabolina) lobata praecurrens; Bridge et al. (1998) noted Ctenopyge (C.) pecten tenuis from a borehole. Of these trilobites, the olenids are widely distributed where the appropriate dysaerobic facies is present, and some others, such as agnostoids (especially Glyptagnostus reticulatus), Proceratopyge and other ceratopygoids, Irvingella and Maladioidella (Rushton and Hughes, 1996) are widely distributed in a variety of shelf environments. In Britain other rare non-olenid, non-agnostoid taxa include species of Acanthopleurella, Aphelas] ylin¤ska, 2001), Araiopleura, Conophrys, Cypis (Z clolorenzella, Eoasaphus, Loganellus, Modocia, Niobella, Parabolinoides, Proteuloma, Richardsonella, and Schmalenseeia. Thomas et al. (1984) listed 86 named taxa and the above-mentioned works added nine more. Of these, 61 are identi¢ed with, or compared with, species from Baltica, and a further 10 are comparable at the generic level. The Upper Cambrian faunas of western Avalonia appear to be less fully known, but, out of a total of about 20 species recorded by Martin and Dean (1988), there are 13 species in common with British faunas and four or ¢ve additional shared genera. The Upper Cambrian in the peri-Gondwanan region contains a number of faunas (discussed above), but, presumably because the proli¢c olenid biofacies is not well developed, there are rel-

PALAEO 3071 9-5-03

28

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

atively few taxa in common with Britain. For example, there is nothing in British faunas to match the oriental genera recorded from the Montagne Noire, and with the transfer of Olentella rara of Rushton (in Allen and Jackson, 1985) to Aphelas] ylin¤ska, 2001), there is no genus in common pis (Z with the Moroccan fauna. Upper Cambrian outcrops from Spain (Shergold et al., 1983) have only yielded Maladioidella, and possibly an undetermined aphelaspidine that can be compared with taxa in British faunas, while a fauna from Sardinia only has a few more comparable taxa, namely Maladioidella, Proceratopyge, an early form of Niobella and some unspeci¢ed olenids (Loi et al., 1995). 5.3.6. Turkey Late Cambrian trilobites have been recorded from the Taurus and Amanos Mountains in south^central Turkey. In the northeast, they occur in the Seydis^ehir Formation of the Sultan Dagfl (Shergold and Sdzuy, 1984; Dean et al., 1993), on the Tu«lu«ce Tepe section. Trilobites described by Shergold and Sdzuy (1984) include Homagnostus? sp., Maladioidella kelteri, Pseudagnostus cf. cyclopyge, Pseudagnostus sp., and undetermined ptychoparioids and solenopleuroideans. Dean et al. (1993) also illustrated specimens of Homagnostus and M. kelteri. All these assemblages are Late Cambrian in age, which may well correlate with the Maladioidella-bearing horizons of Sardinia, Montagne Noire, and northeastern Spain. In the Sultan Dagfl, this fauna is succeeded by a younger association (Shergold and Sdzuy, 1984) containing Koldinioidia (K.) cf. sulcata, Macropyge cf. taurina, Micragnostus haudei, Niobella cf. primaeva, Onchonotellus a¡. amsassensis, Parakoldinioidia cf. gibbosa, Proteuloma cf. geinitzi, and Rhaptagnostus? sp. This was assumed by Shergold and Sdzuy (1984) to indicate an Early Tremadocian age since elements of the association are traditionally considered to be of this age and have a wide distribution in the Franconian Forest area (Germany), the Montagne Noire, the Czech Republic, southern Kazakhstan, southern Siberia, and the Xinjiang Province in China. However, all of the generic taxa have latest Cambrian origins, except perhaps Macropyge (although the genus is

noted to occur in the Sardinian CAB 6 fauna, though not illustrated). Micragnostus, Niobella and Parakoldiniodia, for example, are related to taxa described from the Acerocare Zone (Rushton, 1982), which Landing et al. (1978) regard as correlatable with the Cordylodus proavus conodont Zone. Given the current acceptance of the position of the Cambrian^Ordovician boundary (Cooper and Nowlan, 1999) at the ¢rst appearance of the conodont Iapetognathus £uctivagus Nicoll et al., 1999, associations correlated with the C. proavus Zone are terminal Cambrian age, or Datsonian in the Australian standard (see also the discussion of the Sardinian CAB 6 assemblage by Loi et al., 1995 ; and the correlations proposed by Geyer and Shergold, 2000). Accordingly, it may be necessary to reassess the Sultan Dagfl ‘Tremadocian’. Finally, in southeastern Turkey, Late Cambrian trilobites have been reported by Dean et al. (1981) and Dean and Monod (1997) from the Seydis^ehir Formation of the Samur Dagfl area, on the Yayla Tepe section. Trilobites reported from here include species of Chuangia, Drepanura, and Prochuangia, overlain by younger assemblages characterised by Alborsella, Niobella, Pagodia (Wittekindtia), and Saukia, all indicating a⁄nities with northern Iran.

6. Biogeographical cluster analyses A matrix tabulating all genera and species known from the Lower and Middle Cambrian of Morocco, Ossa^Morena, the Cantabro^Iberian Basin, Montagne Noire, Sardinia, Saxo^Thuringia, the Barrandian area and Turkey (the latter only for the Middle Cambrian) has provided signi¢cant information (Figs. 5C and 6C). Phylogenetic biogeography (Lieberman, 2002) is not available for the Cambrian trilobites of the western Gondwanan margin due to the scarcity of cladistic analyses. However, the data reported here (114 genera and 249 species for the Lower Cambrian, and 113 genera and 442 species for the Middle Cambrian) were analysed using a standard method of linkage cluster analysis, namely analysis of hierarchical Phi^Pearson similarity,

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

which produces dichotomising patterns that are easier to interpret than multivariate analyses. Percentage similarity in a single linkage cluster analysis provides an approximate key to robustness of branching nodes within the constraints of parametric analysis. Trilobite faunas from the di¡erent basins and platforms are associated in terms of their shared genera and species. The analysis of similarity produces bifurcating dendrograms readily interpreted in terms of biogeography. The disadvantage of this method is that biofacies patterns are not considered, and that the total diversity of each platform strongly a¡ects the percentage of shared taxa. Endemic taxa were eliminated from the database before they were entered in the data matrix because they may represent local ecological conditions, lack of marine connections to neighbouring platforms, or sampling biases. Further research might extend the known biogeographical distribution of some taxa but it seems unlikely that such new data would seriously modify the basic patterns discussed in this paper. 6.1. Lower Cambrian The Lower Cambrian shallow subtropical marine biota exhibits the greatest diversity patterns in the whole Cambrian. One of the most interesting features of the Lower Cambrian is the widespread development of endemic species. The Lower Cambrian palaeogeographical complexities of the carbonate and mixed platforms, episodically related to the development of microbial^archaeocyathan reefal complexes, o¡ered various opportunities for the isolation of populations and subsequent allopatric speciation. Many subtropical endemic trilobites assemblages are essentially related to reef communities. Trilobite similarity exhibits two major clusters (Fig. 5C): (1) the southern British Isles, Morocco, and the OMZ; and (2) Saxo^Thuringia, Montagne Noire, the CIZ, the Cantabro^Iberian Basin, and Sardinia. The close relationship of the OMZ with the entity ‘eastern Avalonia^Morocco’ is signi¢cant because it could re£ect easier reproductive communication in the southwesternmost part of the area studied due to open oceanographic conditions and a possible setting of the OMZ closer to Morocco than that

29

illustrated in palaeogeographical maps. Trilobite faunas show a marked change in distribution patterns across the Lower^Middle Cambrian transition; this is attributed to a major palaeogeographical reorganisation, in which the endemic patterns displayed by the Lower Cambrian trilobite fauna ended diachronously and were followed by a period of relative sedimentary homogeneity, with an open connection of platforms that allowed immigration of trilobites. 6.2. Middle Cambrian Trilobite similarity shows two main clusters (Fig. 6C): (1) Baltica^eastern Avalonia, and (2) the rest of the basins discussed here. This also shows strong support for the presence of two Mid-Cambrian biogeographical entities within the Mediterranean subprovince of Sdzuy (1972), distinctly di¡erentiated from eastern Avalonia and Baltica. Therefore the relative Baltic a⁄nity of the Middle Cambrian faunas from the Pr›¤|bram^Jince Basin (Havl|¤c›ek, 1999) does not support di¡erentiation of a Middle Cambrian Perunica terrane. Another important biogeographical relationship with Bohemia is indicated by a Middle Cambrian trilobite assemblage found in the Ptychagnostus atavus Zone of the Carolina Slate belt (eastern United States). The latter is considered as an exotic terrane likely accreted to North America during the Early or Middle Palaeozoic, although an alternative possibility has been proposed placing this fauna in cool-water environments along the periphery of Laurentia (Samson et al., 1990). The variability of the biogeographical distribution documented in the studied populations is directly related to species diversity. Decreasing provinciality of trilobites during the Caesaraugustian appears to have been accompanied by an increase in total diversity. The development of regional regressions induced habitat di¡erences on a local and regional scale favouring ‘microallopatric’ speciation. Regarding the British Cambrian faunas, they have their closest a⁄nity to those of the Gondwanan margin at the end of the Early Cambrian (or earliest Mid-Cambrian in Morocco) and this

PALAEO 3071 9-5-03

30

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

a⁄nity declines through the Middle and Late Cambrian. Through the same period the British Cambrian faunas retain considerable similarities with western Avalonian faunas (those of east maritime Canada and eastern United States, e.g. Massachusetts, New Brunswick, Newfoundland, etc.) and become progressively more similar to Baltic faunas. 6.3. Upper Cambrian A cluster analysis of the Upper Cambrian trilobite faunas is not signi¢cant because of their low diversity patterns in areas other than Britain. In addition, trilobite biofacies possibly occur at only three stratigraphical levels: (1) the Idamean or Steptoean (IC, Montagne Noire, Morocco), characterised by aphelaspidoid morphologies; (2) the Iverian or Sunwaptan (DM, Montagne Noire, Sardinia, western Turkey), characterised by the occurrence of Maladioidella; and (3) the Datsonian, of Cordylodus proavus Zone age. If the age of strata with Niobella, Macropyge, Pagodia (Wittekindtia), and Parakoldinioidia can be ¢rmly established as of Cordylodus proavus Zone age, then that biofacies becomes the third of latest Cambrian age. Some support for this assumption comes º zgu«l and Gedik from conodonts recovered by O (1973) from the Seydis^ehir Formation in the central Taurus Mountains, which by Australian standards are terminal Cambrian. Strata with Oryctoconus also support this assignment, since they occur in association with the Acerocare Regressive Event (Erdtmann, 1986; Nicoll et al., 1992; Loi et al., 1995, 1996), a global sea-level £uctuation close to the currently adopted Cambrian^Ordovician boundary. Most of the Upper Cambrian faunas described from western Gondwana contain elements interpreted as of central and southeastern Asian provenance.

7. Conclusions and further questions One process can be considered primarily responsible for the northeastward migration of the biodiversity centre (based on trilobites) on the western Gondwanan margin during the Cambri-

an: the cooling associated with the southwestward drift of the continental margin. Diverse Lower Cambrian trilobite assemblages are found in the shallow waters of subtropical areas, as studied in the Souss Basin of Morocco. The progressive change from relatively endemic to cosmopolitan trilobite assemblages across the Lower^Middle Cambrian transition can be related to three main events: (1) the disappearance of microbial^ archaeocyathan buildups and neighbouring perireefal environments, (2) the relative uniformity of muddy o¡shore substrates related to a regional (eustatic?) early Mid-Cambrian transgression, and (3) the disappearance of signi¢cant temperature gradients on the Southern Hemisphere. The peak of Middle Cambrian trilobite diversity progressively migrated from Morocco to southwestern Europe (Cantabro^Iberian, Montagne Noire, Saxo^Thuringia and Barrandian areas), although these subtropical- to temperate-water platforms did not reach the richness reported in the Lower Cambrian of the Souss Basin. Finally, Upper Cambrian temperate-water faunas are dramatically impoverished compared with the previous wealth, except in the southern British Isles, where the proli¢c olenid biofacies developed, exhibiting a strong biogeographical relationship with Baltica. Ocean currents played a signi¢cant role in the regional distribution of benthic communities. Two Cambrian unidirectional circulation trends seem to have been in operation: (1) in the Early to Mid^Cambrian, a western peri-Gondwanan current with a SW^NE trend extended from polar areas to lower latitudes and favoured interchange with benthic communities of Baltic a⁄nity; (2) in the Late Cambrian a NE^SW longshore current £owed towards higher latitudes and seem to have involved organisms of Asian a⁄nity. However, it remains a question whether the dispersal of subtropical, Australo^Sinian trilobites into temperate or cold seas of the Southern Hemisphere can be attributed to transport of southward-drifting currents ? Or did the migration take place in a direction opposite to that of the main northeastward £ow of surface waters? Numerous questions are still open for future discussion that will bene¢t from a multidiscipli-

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

nary approach. This paper constitutes only one new step in the establishment of Cambrian palaeobiogeographical models. The palaeogeographical maps and biogeographical interpretations expressed here are intended to stimulate critical remarks from specialists of other fossil groups and geological disciplines, and to be tested further in the light of new data.

Acknowledgements The authors thank warmly the proposals and discussions of Per Ahlberg (Lund), Pete Palmer (Boulder), and the editing work of Alain Blieck (Lille), which have helped to complete and improve a previous manuscript. This paper is a contribution to Project ATI 15-52. References Allen, P.M., Jackson, A.A., 1985. Geology of the country around Harlech. Mem. Brit. Geol. Surv., Sheet 135 with part of 149 (England and Wales), H.M.S.O., London, xii+111 pp. Ł lvaro, J.J., Vennin, E., 1998. Stratigraphic signature of a A terminal Early Cambrian regressive event in the Iberian Peninsula. Can. J. Earth Sci. 35, 402^411. Ł lvaro, J.J., Vennin, E., 2001. Benthic marine communities A recorded in the Cambrian Iberian platform, NE Spain. Palaeontographica A 262, 1^23. Ł lvaro, J.J., Vizca|«no, D., 1998. Re¤vision biostratigraphique A du Cambrien moyen du versant me¤ridional de la Montagne Noire (Languedoc, France). Bull. Soc. ge¤ol. Fr. 169, 233^ 242. Ł lvaro, J.J., Vizca|«no, D., 2001. Evolutionary trends in the A ornamentation of Cambrian solenopleuropsine trilobites. Palaeontology 44, 131^141. Ł lvaro, J.J., Vizca|«no, D., Vennin, E., 1999. Trilobite diversity A patterns in the Middle Cambrian of southwestern Europe: a comparative study. Palaeogeogr. Palaeoclimatol. Palaeoecol. 151, 241^254. Ł lvaro, J.J., Rouchy, J.M., Bechsta«dt, T., Boucot, A., Boyer, A F., Debrenne, F., Moreno-Eiris, E., Perejo¤n, A., Vennin, E., 2000a. Evaporitic constraints on the southward drifting of the western Gondwana margin during Early Cambrian times. Palaeogeogr. Palaeoclimatol. Palaeoecol. 160, 105^ 122. Ł lvaro, J.J., Vennin, E., Moreno-Eiris, E., Perejo¤n, A., BechA sta«dt, T., 2000b. Sedimentary patterns across the Lower^ Middle Cambrian transition in the Esla nappe (Cantabrian Mountains, northern Spain). Sediment. Geol. 137, 43^61.

31

Aramburu, C., Garc|¤a-Ramos, C., 1993. La sedimentacio¤n cambro-ordov|¤cica en la Zona Canta¤brica (NO de Espan‹a). Trab. Geol. Univ. Oviedo 19, 45^73. Aramburu, C., Truyols, J., Arbizu, M., Me¤ndez-Bedia, I., Zamarren‹o, I., Garc|¤a-Ramos, J.C., Sua¤rez de Centi, C., Valenzuela, M., 1992. El Paleozoico Inferior de la Zona Canta¤brica. In: Gutie¤rrez-Marco, J.C., Saavedra, J., Ra¤bano, I. (Eds.), Paleozoico Inferior de Ibero^Ame¤rica. UNEX Press, Me¤rida, pp. 397^422. Baldwin, C.T., 1977. The stratigraphy and facies associations of trace fossils in some Cambrian and Ordovician rocks of northwestern Spain. In: Crimes, T.P., Harper, J.C. (Eds.), Trace Fossils 2. Geol. J., Spec. Issue 9, 9^40. Bergstro«m, S.M., 1990. Relations between conodont provincialism and the changing paleogeography during the early Paleozoic. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeographies and Biogeographies. Geol. Soc. London Mem. 12, pp. 163^174. Brasier, M.D., 1995. The basal Cambrian transition and Cambrian bio-events. In: Walliser, O.H. (Ed.), Global Events and Event Stratigraphy in the Phanerozoic. Springer, Berlin, pp. 113^118. Brasier, M.D., Perejo¤n, A., San Jose¤, M.A., 1979. Discovery of an important fossiliferous Precambrian^Cambrian sequence in Spain. Estud. Geol. 35, 379^383. Bridge, D.McC., Carney, J.N., Lawley, R.S., Rushton, A.W.A., 1998. The geology of the country around Coventry and Nuneaton. Mem. Brit. Geol. Surv., Sheet 169 (England and Wales). The Stationery O⁄ce, London, x+185 pp. Briggs, J.C., 1995. Global Biogeography. Develop. Palaeontol. Strat. 14, 1^454. Brock, G.A., Engelbretsen, M.J., Jago, J.B., Kruse, P.D., Laurie, J.R., Shergold, J.H., Shi, G.R., Sorauf, J.E., 2000. Palaeobiogeographic a⁄nities of Australian Cambrian faunas. Mem. Assoc. Aust. Palaeontol. 23, 1^61. Bro«gger, W.C., 1886. Om alderen af Olenelluszonen i Nordamerika. Geol. Fo«ren. Stockholm Fo«rhandl. 101, 182^213. Carls, P., 1983. La Zona Asturoccidental^Leonesa en Arago¤n y el macizo del Ebro como prolongacio¤n del Macizo Canta¤brico. In: Comba, J. (Coord.), Contribuciones sobre temas generales. IGME, Libro Jubilar J.M. R|¤os 3, pp. 11^32. Chlupa¤c›, I., 1995. Lower Cambrian arthropods from the Paseky Shale (Barrandian area, Czech Republic). J. Czech Geol. Soc. 40, 9^36. Chlupa¤c›, I., Havl|¤c›ek, V.J., Kr›¤|z›, J., Kukal, Z., SYtorch, P., 1998. Paleozoic of the Barrandian (Cambrian to Devonian). Czech Geol. Surv. Publ., Prague, 183 pp. Cocks, L.R.M., McKerrow, W.S., van Staal, C.R., 1997. The margins of Avalonia. Geol. Mag. 134, 456^460. Colchen, M., 1967. Sur la pre¤sence du Cambrien supe¤rieur a' Prochuangia et a' Chuangia dans la Sierra de la Demanda (Logron‹o), Espagne. C.R. Acad. Sci. 264, 1687^1690. Colchen, M., Ubaghs, G.P., 1969. Sur les restes d’Echinodermes (?) du Cambro^Ordovicien de la Sierra de la Demanda (Burgos^Logron‹o, Espagne). Bull. Soc. ge¤ol. Fr. Se¤r. 7 11, 649^654. Conway Morris, S., Rushton, A.W.A., 1988. Precambrian to

PALAEO 3071 9-5-03

32

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Tremadoc biotas in the Caledonides. In: Harris, A.L., Fettes, D.J. (Eds.), The Caledonian^Appalachian Orogen. Geol. Soc. London Spec. Publ. 38, pp. 93^109. Cooper, R.A., Nowlan, G.S., 1999. Proposed global stratotype and point for base of the Ordovician System. Acta Univ. Carolinae Geol. 43, 61^63. Cope, J.C.W., Rushton, A.W.A., 1992. Cambrian and early Tremadoc rocks of the Llangynog Inlier, Dyfed, South Wales. Geol. Mag. 129, 543^552. Cope, J.C.W., Ingham, J.K., Rawson, P.F. (Eds.), 1992. Atlas of Palaeogeography and Lithofacies. Mem. Geol. Soc. Lond. 13, xiii+153 pp. Cowie, J.W., Rushton, A.W.A., Stubble¢eld, C.J., 1972. A correlation of Cambrian rocks in the British Isles. Spec. Rep. Geol. Soc. London 2, 1^42. Crimes, T.P., Legg, I., Marcos, A., Arboleya, M.L., 1977. ?Late Precambrian^low Lower Cambrian trace fossils from Spain. In: Crimes, T.P., Harper, J.E. (Eds.), Trace Fossils. Geol. J. Spec. Issue 9, pp. 91^138. Dean, W.T., Monod, O., 1997. Cambrian development of the Gondwanaland margin in Southeastern Turkey. Turk. Assoc. Petrol. Geol. Spec. Publ. 3, 61^74. Dean, W.T., Monod, O., Perinc^ek, D., 1981. Correlation of the Cambrian and Ordovician rocks in southeastern Turkey. Petroleum activities at the 100th year (100 Yilda Petrol Faaliyeti). T.C. Petrol I]sleri Genel Mu«du«rlu«gflu« Dergisi 25, Ankara, pp. 269^300. Dean, W.T., Martin, F., Monod, O., Gu«l, M.A., Bozdogflan, º zgu«l, N., 1993. Early Palaeozoic evolution of the N., O Gondwanaland margin in the western and central Taurids, Turkey. In: Turgut, S. (Ed.), Tectonics and Hydrocarbon Potential of Anatolia and Surrounding Regions. Ozan Sungflurlu Symposium Proceedings, November 1991, Ankara, pp. 262^272. Debrenne, F., 1991. Extinction of the Archaeocyatha. Hist. Biol. 5, 95^106. Destombes, J., Feist, R., 1987. De¤couverte du Cambrien supe¤rieur en Afrique (Anti-Atlas, Central Maroc). C.R. Acad. Sci. se¤r. 2 304, 719^724. Destombes, J., Hollard, H., Willefert, S., 1985. Lower Palaeozoic rocks of Morocco. In: Hollard, C.H. (Ed.), Lower Palaeozoic Rocks of the World, vol. 4. Lower Palaeozoic of North-Western and West Central Africa. Wiley, Chichester, pp. 157^184. Dore¤, F., 1994. Cambrian of the Armorican Massif. In: Keppie, J.D. (Ed.), Pre-Mesozoic Geology in France and Related Areas. Springer, Berlin, pp. 136^141. Elicki, O., 1994. Lower Cambrian carbonates from eastern Germany: palaeontology, stratigraphy and palaeogeography. N. Jb. Geol. Palao«nt. Abh. 191, 69^93. Elicki, O., 1997. Biostratigraphic data of the German Cambrian ^ present state of knowledge. Freiberger Forsch.-H. C 466, 155^165. Elicki, O., 1999. Palaeoecological signi¢cance of calcimicrobial communities during ramp evolution: an example from the Lower Cambrian of Germany. Facies 41, 27^40. Erdtmann, B.D., 1986. Early Ordovician eustatic cycles and

their bearing on punctuations in early nematophorid (planctic) graptolite evolution. In: Walliser, O.H. (Ed.), Global Bio-Events. Lecture Notes in Earth Sciences 8, Springer, Berlin, pp. 139^152. Feist, R., Courtessole, R., 1984. De¤couverte de Cambrien supe¤rieur a' trilobites de type est-asiatique dans la Montagne Noire (France me¤ridionale). C.R. Acad. Sci. se¤r. 2 298, 177^ 182. Fletcher, T.P., 1972. Geology and Lower to Middle Cambrian Trilobite Faunas of Southwestern Avalon, Newfoundland. Part Two. Unpubl. Ph.D. Thesis, Univ. of Cambridge, 295 pp. Franke, W., 2000. The mid-European segment of the Variscides: tectonostratigraphic units, terrane boundaries and plate tectonic evolution. In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenic Processes: Quanti¢cation and Modelling in the Variscan Belt. Geol. Soc. London Spec. Publ. 179, pp. 35^61. Geyer, G., 1989. Late Precambrian to early Middle Cambrian lithostratigraphy of southern Morocco. Beringeria 1, 115^ 143. Geyer, G., 1990a. Revised Lower to lower Middle Cambrian biostratigraphy of Morocco. Newslett. Stratigr. 22, 53^70. Geyer, G., 1990b. Die marokkanischen Ellipsocephalidae (Trilobita: Redlichiida). Beringeria 3, 1^363. Geyer, G., Elicki, O., 1995. Lower Cambrian trilobites from the Go«rlitz Synclinorium (Germany) ^ review and new results. Pala«ontol. Z. 69, 87^119. Geyer, G., Landing, E., 1995. The Cambrian of the Moroccan Atlas regions. Beringeria 2, 7^46. Geyer, G., Malinky, J.M., 1997. Middle Cambrian fossils from Tizi N’Tichka, the High Atlas, Morocco. Part 1. Introduction and trilobites. J. Paleontol. 71, 620^637. Geyer, G., Palmer, A.R., 1995. Neltneriidae and Holmiidae (Trilobita) from Morocco and the problem of Early Cambrian intercontinental correlation. J. Paleontol. 69, 459^474. Geyer, G., Shergold, J.H., 2000. The quest for internationally recognized divisions of Cambrian time. Episodes 23, 188^ 195. Geyer, G., Landing, E., Heldmaier, W., 1995. Faunas and depositional environments of the Cambrian of the Moroccan Atlas region. Beringeria 2, 47^120. Havl|¤c›ek, V., 1999. Perunica microplate: relation to Ukranian Shield, mid-Bohemian rift, and hypothetic large-scale Y esk. Geol. u¤st. 74, overthrusts in central Bohemia. Veflst. C 75^81. Havl|¤c›ek, V., Vanek, J., Fatka, O., 1994. Perunica microcontinent in the Ordovician (its position within the Mediterranean province, series division, bentic and pelagic associations). Sbor. geol. Vefld, Geol. 46, 23^56. Howells, M.F., Smith, M., 1997. Geology of the Country around Snowdon. Mem. British Geol. Surv., Sheet 119 (England and Wales), H.M.S.O., London, x+104 pp. Hughes, N.C., Rushton, A.W.A., 1990. Computer-aided restoration of a late Cambrian ceratopygid trilobite from Wales, and its phylogenetic implications. Palaeontology 33, 429^ 445.

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A Hupe¤, P., 1952. Sur les zones de Trilobites du Cambrien infe¤rieur marocain. C.R. Acad. Sci. 235, 480^481. Ivshin, N.K., 1955. Vosem’novykh rodov trilobitov iz verkhnego kembriya tsentral’nogo Kazakhstana. Izvestiya Akademiya Nauk Kazakhskoy SSR, Syr’ya Geologiya 21, 106^ 123. Ivshin, N.K., 1956. Verkhnekembriskie trilobity Kazakhstana, Chast 1. Kuyandinskiy faunisticheskiy gorizont mezhdurechiya Olenty^Shiderty. Izdatel’stvo Institut Geologicheskikh Nauk, Akademiya Nauk Kazakhskoy SSR, Alma-Ata, pp. 1^98. Ivshin, N.K., 1962. Verkhnekembriskie trilobity Kazakhstana, Chast 2. Seletinskiy gorizont Kuyandinskogo yarusa tsentral’nogo Kazakhstana. Izdatel’stvo Institut Geologicheskikh Nauk, Akademiya Nauk Kazakhskoy SSR, Alma Ata, pp. 1^412. Josopait, V., 1972. Das Kambrium und das Tremadoc von Ateca (westliche Iberischen Ketten, NE-Spanien). Mu«nst. Forsch. Geol. Pala«ontol. 23, 1^121. Krs, M., Krsova¤, M., Pruner, P., Chvojka, R., Havl|¤c›ek, V., 1987. Palaeomagnetism, palaeogeography and the multicomponent analysis of Middle and Upper Cambrian rocks of the Barrandian in the Bohemian Massif. Tectonophysics 139, 1^20. Kukal, Z., 1971. Sedimentology of Cambrian deposits of the Barrandian area (Central Bohemia). Sbor. Geol. Vefld. Geol. 20, 53^100. Landing, E., 1994. Precambrian^Cambrian boundary global stratotype rati¢ed and a new perspective of Cambrian time. Geology 22, 179^184. Landing, E., Taylor, M.E., Erdtmann, B.D., 1978. Correlation of the Cambrian^Ordovician boundary between the Acado^ Baltic and North American faunal provinces. Geology 6, 75^78. Landing, E., Bowring, S.A., Davidek, K.L., Westrop, S.R., Geyer, G., Heldmaier, W., 1998. Duration of the Early Cambrian; U^Pb ages of volcanic ashes from Avalon and Gondwana. Can. J. Earth Sci. 35, 329^338. Landing, E., Bowring, S.A., Davidek, K.L., Rushton, A.W.A., Fortey, R.A., Wimbledon, W.A.P., 2000. Cambrian^Ordovician boundary age and duration of the lowest Ordovician Tremadoc Series based on U^Pb zircon dates from Avalonian Wales. Geol. Mag. 137, 485^494. Lieberman, B.S., 2002. Phylogenetic biogeography with and without the fossil record: gauging the e¡ects of extinction and paleontological incompleteness. Palaeogeogr. Palaeoclimatol. Palaeoecol. 2726, 1^14. Lin‹a¤n, E., Sdzuy, K., 1978. A trilobite from the Lower Cambrian of Co¤rdoba (Spain), and its stratigraphical signi¢cance. Senckenb. lethaea. 59, 387^399. Linnemann, U., Buschmann, B., 1995. Die cadomische Diskordanz im Saxothuringikum (oberkambrisch^tremadocische Overlap-Sequenzen). Z. geol. Wiss. 23, 707^727. Linnemann, U., Gehmlich, M., Tichomirowa, M., Buschmann, B., Nasdala, L., Jonas, P., Lu«tzner, H., Bombach, K., 2000. From Cadomian subduction to Early Palaeozoic rifting: the evolution of Saxo^Thuringia at the margin of

33

Gondwana in the light of single zircon geochronology and basin development (Central European Variscides, Germany). In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenic Processes: Quanti¢cation and Modelling in the Variscan Belt. Geol. Soc. London Spec. Publ. 179, pp. 131^ 153. Loi, A., Pillola, G.L., Leone, F., 1995. The Cambrian and Early Ordovician of southwestern Sardinia. In: Cherchi, A. (Ed.), Sardinia 95, 6th Paleobenthos International Symposium, Guide-book. Rend. Sem. Fac. Sci. Univ. Cagliari, Suppl. vol. 65, pp. 63^81. Loi, A., Pillola, G.L., Leone, F., 1996. La limite Cambrien^ Ordovicien dans le SW de la Sardaigne: relations avec des e¤ve¤nements eustatiques globaux. C.R. Acad. Sci. se¤r. 2 323, 881^888. Lotze, F., 1961. Das Kambrium Spaniens. Teil 1: Stratigraphie. Abh. Math. Naturwiss. Kl. 1961, 1^498. Marcos, A., Pe¤rez Estau¤n, A., 1981. La estratigraf|¤a de la Serie de los Cabos en la Zona de Vegadeo (Zona Astur^Occidental Leonesa, NW de Espan‹a). Trab. Geol. Univ. Oviedo 11, 89^94. Martin, F., Dean, W.T., 1988. Middle and Upper Cambrian acritarch and trilobite zonation at Manuels River and Random Island, eastern Newfoundland. Geol. Surv. Can. 381, 1^91. Mart|¤nez-Garc|¤a, E., Rolet, J., 1991. Correlations between the Palaeozoic orogens of Spain and France. New data from the Iberian, Armorican and Aquitanian Massifs. Project 233 TCAPO, Terranes in the Arctic Caledonides, Tromso, Norway, Abstr. vol. 4. Mata, J., Munha¤, J., 1990. Magmatoge¤nese de Metavulcanitos Ca“mbricos do Nordeste Alentejano: os esta¤dios iniciais de ‘rifting’ continental. Comun. Serv. Geol. Port. 76, 61^89. McGowran, B., Li, Q., Moss, G., 1997. The Cenozoic neritic record in southern Australia: the biogeohistorical framework. In: James, N.P., Clarke, A.D. (Eds.), Cool-Water Carbonates. SEPM, Spec. Publ. 56, pp. 185^203. McKerrow, W.S., Scotese, C.R., Brasier, M.D., 1992. Early Cambrian continental reconstructions. J. Geol. Soc. London 149, 599^606. Mergl, M., Geyer, G., El-Attari, A., 1998. The billingsellid genus Saccogonum (Brachiopoda) from the Moroccan Cambrian and its signi¢cance for the regional geology and stratigraphy. N. Jb. Geol. Pala«ontol. Abh. 209, 273^293. Morris, S.F., 1988. A Review of British Trilobites, Including a Synoptic Revision of Salter’s Monograph. Monogr. Palaeontogr. Soc. London Publ. 574, 316 pp. Nicoll, R.S., Miller, J.F., Nowlan, G.S., Repetski, J.E., Ethington, R.L., 1999. Iapetonudus (new genus) and Iapetognathus Landing, unusual earliest Ordovician multielement conodont taxa and their utility for biostratigraphy. Brigham Young Univ. Geol. Stud. 4, 27^101. Nicoll, R.S., Nielsen, A.T., Laurie, J.R., Shergold, J.H., 1992. Preliminary correlation of latest Cambrian to Early Ordovician sea level events. In: Webby, B.D., Laurie, J.R. (Eds.), Global Perspectives on Ordovician Geology. Balkema, Rotterdam, pp. 381^394.

PALAEO 3071 9-5-03

34

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A

Nielsen, A.T., 1996. Iltsvind, sort slam og trilobiter. Varv for 1996 (1), pp. 3^39. º zgu«l, N., Gedik, I., 1973. New data on the stratigraphy and O conodont faunas of the C^altepe Limestone and Seydis^ehir Formation lower Paleozoic of the central Taurus Range. Bull. Geol. Soc. Turk. 16, 39^52. Palmer, A.R., 1972. Cambrian trilobites. In: Hallam, A. (Ed.), Atlas of Palaeobiogeography. Elsevier, Amsterdam, pp. 3^ 11. Palmer, A.R., 1998. Why is the intercontinental correlation within the Lower Cambrian so di⁄cult? Rev. Esp. Paleontol., 17^21. Palmer, A.R., Rowell, A.J., 1995. Early Cambrian trilobites from the Shackleton Limestone of the Central Transantarctic Mountains. J. Paleontol. 45, 1^28. Perejo¤n, A., Moreno-Eiris, E., Fro«hler, M., Bechsta«dt, T., Boni, M., 2000. Archaeocyathan assemblages from the Gonnesa Group, Lower Cambrian (Sardinia, Italy) and their sedimentologic context. Boll. Soc. Paleontol. Ital. 39, 257^ 291. Pillola, G.L., 1991a. Trilobites du Cambrien infe¤rieur du SW de la Sardaigne, Italie. Palaeontographia Ital. 78, 1^174. Pillola, G.L., 1991b. Occurrence of Proteuloma (Trilobita) in the Cabitza Formation (Cambro^Tremadoc, SW Sardinia, Italy): palaeobiogeographic implications. In: Geologia del Basemento Italiano. Convegno in memoria di Tommaso Cocozza, Siena, 21^22 Marzo 1991, pp. 59^61. Pillola, G.L., Gutie¤rrez-Marco, J.C., 1988. Graptolites du Tremadoc du sud-ouest de la Sardaigne (Italie): pale¤oe¤cologie et contexte tectono^se¤dimentaire. Ge¤obios 21, 553^ 565. Pillola, G.L., Leone, F., 1993. Biostratigraphical and sedimentological data from the Cabitza Formation (Middle Cambrian^Tremadoc, SW Sardinia, Italy): palaeogeography and Cambrian^Ordovician boundary. IGCP 319, Liverpool, Abstracts, p. 22. Quesada, C., 1991. Geological constraints on the Paleozoic tectonic evolution of tectonostratigraphic terranes in the Iberian Massif. Tectonophysics 185, 225^245. Robison, R.A., Rosova, A.V., Rowell, A.J., Fletcher, T.P., 1977. Cambrian boundaries and divisions. Lethaia 10, 257^262. Rodr|¤guez Ferna¤ndez, L.R., Barba, P., Bardaj|¤, T., Silva, P.G., Sua¤rez Rodr|¤guez, P.L., Heredia, N., Gallastegui, G., Paniagua, A., Gala¤n, L., 1991. Mapa Geolo¤gico de Espan‹a, E. 1/50.000, 2a serie, 102 (Los Barrios de Luna). ITGE Press, Madrid, pp. 1^130. Rushton, A.W.A., 1974. The Cambrian of Wales and England. In: Holland, C.H. (Ed.), Cambrian of the British Isles, Norden and Spitsbergen. Wiley, London, pp. 42^122. Rushton, A.W.A., 1978. Fossils from the Middle^Upper Cambrian transition in the Nuneaton District. Palaeontology 21, 245^283. Rushton, A.W.A., 1982. The biostratigraphy and correlation of the Merioneth^Tremadoc boundary in North Wales. In: Bassett, M.G., Dean, W.T. (Eds.), The Cambrian^Ordovician Boundary: Sections, Fossil Distributions, and Correla-

tions. National Museum of Wales, Cardi¡, Geol. Ser. 3, pp. 41^59. Rushton, A.W.A., Hughes, N.C., 1996. Biometry, systematics and biogeography of the Late Cambrian trilobite Maladioidella abdita. Trans. R. Soc. Edinb. Earth Sci. 86, 247^256. Rushton, A.W.A., Owen, A.W., Owens, R.M., Prigmore, J.K., 1999 (2000). British Cambrian to Ordovician Stratigraphy. GCR Series 18, Joint Nature Conservation Committee, Peterborough, xxi+435 pp. Russo, A., Bechsta«dt, T., 1994. Evolucio¤n sedimentolo¤gica y paleogeogra¤¢ca de la formacio¤n Vegadeo (Ca¤mbrico Inferior^Medio) en la zona entre Visun‹a y Piedra¢ta do Caurel (Lugo, NO de Espan‹a). Rev. Geol. Esp. 7, 299^310. Samson, S., Palmer, A.R., Robison, R.A., Secor, D.T.Jr., 1990. Biogeographical signi¢cance of Cambrian trilobites from the Carolina slate belt. Geol. Soc. Am. Bull. 102, 1459^1470. Sdzuy, K., 1958. Fossilien aus dem Tremadoc der Montagne Noire. Senckenb. lethaea 39, 235^253. Sdzuy, K., 1961. Das Kambrium Spaniens. Teil II: Trilobiten, 1. Abschnitt. Akad. Wiss. Lit. (Mainz). Abh. Math.-Naturwiss. Kl., for 1961, 7, pp. 499^594 (217^312). Sdzuy, K., 1972. Das Kambrium der acadobaltischen Faunenprovinz ^ Gegenwa«rtiger Kenntnisstand und Probleme. Zent. Geol. Pala«ontol. II 1972, 1^91. Sdzuy, K., 1987. Trilobites de la base de la Formacio¤n Jalo¤n (Ca¤mbrico Inferior) de Arago¤n. Rev. Esp. Paleontol. 2, 3^8. Sdzuy, K., Geyer, G., 1988. The base of the Cambrian in Morocco. In: Jacobshagen, V.H. (Ed.), The Atlas System of Morocco. Studies on its Geodynamic Evolution. Lecture Notes in Earth Sciences 15, pp. 91^106. Sdzuy, K., Lin‹a¤n, E., Gozalo, R., 1999. The Leonian stage (early Middle Cambrian): a unit for Cambrian correlation in the Mediterranean subprovince. Geol. Mag. 136, 39^48. Shergold, J.H., Sdzuy, K., 1984. Cambrian and early Tremadocian trilobites from Sultan Dagfl, central Turkey. Senckenb. lethaea 65, 51^135. Shergold, J.H., Sdzuy, K., 1991. Late Cambrian trilobites from the Iberian Mountains, Zaragoza Province, Spain. Beringeria 4, 193^235. Shergold, J.H., Lin‹a¤n, E., Palacios, T., 1983. Late Cambrian trilobites from the Najerilla Formation, north-eastern Spain. Palaeontology 26, 71^92. Shergold, J.H., Feist, R., Vizca|«no, D., 2000. Early Late Cambrian trilobites of Australo^Sinian aspect from the Montagne Noire, southern France. Palaeontology 43, 599^632. Smith, N.J.P., Rushton, A.W.A., 1993. Cambrian and Ordovician stratigraphy related to structure and seismic pro¢les in the western part of the English Midlands. Geol. Mag. 130, 665^671. Thomas, A.T., Owens, R.M., Rushton, A.W.A., 1984. Trilobites in British stratigraphy. Geol. Soc. London Spec. Rep. 16, 1^78. Valladares, M.I., 1995. Siliciclastic^carbonate slope apron in an immature tensional margin (Upper Precambrian^Lower Cambrian), Central Iberian Zone, Salamanca, Spain. Sediment. Geol. 94, 165^186.

PALAEO 3071 9-5-03

Ł lvaro et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 5^35 J. Javier A Vegas, R., 1978. Sedimentation and tectonism in the Iberian massif prior to the Hercynian deformation (late Precambrian times). Cuad. Sem. Est. Cera¤m. Sargadelos 27, 271^ 286. Vra¤na, S., SYtefldra¤, V., 1997. Geological model of Western Bohemia related to the KTB borehole in Germany. Sbor. Geol. Vefld. Geol. 47, 1^240. Walcott, C.D., 1891. The Fauna of the Lower Cambrian or Olenellus Zone. Tenth Ann. Rep. U.S. Geol. Surv., Part 1, pp. 509^774. Whittington, H.B., Chatterton, B.D.E., Speyer, S.E., Fortey, R.A., Owens, R.M., Chang, W.T., Dean, W.T., Jell, P.A., Laurie, J.R., Palmer, A.R., Repina, L.N., Rushton, A.W.A., Shergold, J H., Clarkson, E.N.K., Wilmot, N.V., Kelly, S.R.A., 1997. Trilobita. Introduction, Order Agnostida, Order Redlichiida. Treatise on Invertebrate Paleontology, Part O, revised, vol. 1. University of Kansas and Geological Society of America, xxiv+530 pp.

35

Young, T.P., 1990. Ordovician sedimentary facies and faunas of southwest Europe: palaeogeographic and tectonic implications. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. Mem. 12, pp. 421^430. Young, T., Martin, F., Dean, W.T., Rushton, A.W.A., 1994. Cambrian stratigraphy of St Tudwal’s Peninsula, Gwynedd, northwest Wales. Geol. Mag. 131, 335^360. Zamarren‹o, I., 1972. Las litofacies carbonatadas del Ca¤mbrico de la Zona Canta¤brica (NW Espan‹a) y su distribucio¤n paleogeogra¤¢ca. Trab. Geol. Univ. Oviedo 5, 1^118. Zhuravlev, A.Yu., 1995. Reef ecosystem recovery after the Early Cambrian extinction. In: Hart, M.B. (Ed.), Biotic Recovery from Mass Extinction Events. Geol. Soc. Spec. Publ. 102, pp. 79^96. ] ylin¤ska, A., 2001. Late Cambrian trilobites from the Holy Z Cross Mountains, central Poland. Acta Geol. Pol. 51, 333^ 383.

PALAEO 3071 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54 www.elsevier.com/locate/palaeo

Global Ordovician vertebrate biogeography Alain Blieck a; , Susan Turner b;c a

Universite¤ des Sciences et Technologies de Lille, Sciences de la Terre, Laboratoire de Pale¤ontologie et Pale¤oge¤ographie du Pale¤ozo|«que (LP3), UMR 8014 et FR 1818 du C.N.R.S., 59655 Villeneuve d’Ascq Cedex, France b School of Geosciences, Monash University, Clayton, Vic. 3088, Australia c Queensland Museum, PO Box 3300, South Brisbane, Qld 4101, Australia Received 17 April 2002; received in revised form 22 October 2002; accepted 25 January 2003

Abstract Cambrian^Ordovician vertebrate and supposed vertebrate occurrences have been repeatedly claimed during recent decades, with confirmed taxa bearing mineralized tissues with a vertebrate histomorphology still relatively rare. The only biogeographic province that we can presently recognize is the Gondwana Endemic Assemblage (GEA) with possible Late Cambrian fragmentary remains from Australia but more definite Early Ordovician (Arenigian) to early Late Ordovician (Caradocian) arandaspids (i.e., Sacabambaspis, Arandaspis) and other taxa known from South America and Australia. Certain chondrichthyans (‘sharks’, at first without teeth and which might not constitute a monophyletic group) might have originated in East Gondwana province and then are found in the Late Ordovician and Early Silurian of Mongolia, Tarim, and South China. The GEA fauna proper disappears by middle^late Caradocian and vertebrates do not reappear in Gondwana until mid Late Silurian. Late Ordovician (3455 Myr or earlier) vertebrates are also known with certainty from Laurentia, viz., North America (pteraspidomorphs Astraspis, Eriptychius, and various gnathostome-like taxa including chondrichthyan-, placoderm- and acanthodian-like remains), and Siberia (astraspid-like microremains with an unusual histology, which might correspond to a new group of lower vertebrates) as well as scales from putative loganiid and thelodontidid thelodonts from North America and Russia (Timan^Pechora, the Severnaya Zemlya archipelago and Siberia). This is defined as the Laurentia^Baltica^Siberia Assemblage (LBSA). We also mention one enigmatic reference to a Late Ordovician anaspid in South Africa. There is no clear association of taxa between the GEA and LBSA despite a small overlap in time. Various recent palaeogeographic models published for the Ordovician are critically analyzed and considered within four groups: the archetypal palaeogeographic reconstructions, two alternative solutions, and a compact version. Habitats of vertebrates in mostly BA1 (marine intertidal) to BA3 (shallow subtidal) environments, and their dispersal capabilities are evaluated with regard to those models. The main feature of Ordovician vertebrate biogeography is endemism. Furthermore, the present lack of complete descriptions of most taxa, which are often represented only by isolated microremains, and the need for a thorough phylogenetic analysis preclude any phylogenetic palaeobiogeographic study. In such a framework, we also evaluate possible links between external, physical factors and the Ordovician radiation of vertebrates. The Late Proterozoic deposition of oceanic phosphate and the Early Cambrian increase in oxygen on Earth might have been the spur for vertebrate evolution before the phase when hard tissues appeared. The sharp decline of the marine strontium isotope ratio during the Middle to Late Ordovician transition, interpreted as having been controlled primarily by continental collisional tectonics and its associated erosion and weathering, has

* Corresponding author. Tel.: +33-3-20434140; Fax: +33-3-20436900. E-mail address: [email protected] (A. Blieck).

0031-0182 / 03 / $ ^ see front matter A 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00301-8

PALAEO 3072 9-5-03

38

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

been proposed as the consequence of a possible mantle superplume event which could have caused the prominent Caradocian transgressive phase. This might have been a factor in the changeover from a Gondwanan to a Laurentian focus for vertebrates. A 2003 Elsevier Science B.V. All rights reserved. Keywords: Agnatha; Gnathostomata; Gondwana; Laurentia; Baltica; Siberia

1. Introduction Ordovician vertebrates have been known since the late 19th century, with notable discoveries mostly from recent decades including articulated Ordovician ¢sh from western USA, Bolivia and central Australia (Turner et al., in press a,b). However, little work has been attempted to place these records in a global context or to assess their palaeobiogeographic signi¢cance (Blieck et al., 1991, 2001; Elliott et al., 1991; Webby et al., 2000). Only recently has this become possible through more detailed biostratigraphic work (e.g., Young, 1997; Sansom et al., 1996, 1997, 2001; Turner et al., in press a,b). What is most significant is the now-recognized widespread record of Ordovician vertebrates in a circum-equatorial location. Possible Late Cambrian, Early Ordovician to early Late Ordovician (Caradocian) taxa are known from Gondwana, and constitute

our Gondwana Endemic Assemblage (GEA, Fig. 1) (Blieck et al., 2001; Turner et al., in press a). This fauna disappears by middle^late Caradocian and vertebrates do not reappear in Gondwana until mid Late Silurian (e.g., Pickett et al., 2000). By 3455 Myr (or earlier) a wide range of taxa from several groups: pteraspidomorphs, putative loganiid and thelodontidid thelodonts, and gnathostomes including chondrichthyan-, placoderm- and acanthodian-like remains, occur in Laurentia (Colorado, Wyoming, and Ontario). Late Ordovician to earliest Silurian vertebrates (astraspid? Tesakoviaspis, thelodonts) are restricted to Laurentia (Arctic Canada, Wisconsin, and Que¤bec), Baltica (Timan^Pechora) and Siberia (including western Mongolia); we have proposed the name of Laurentia^Baltica^Siberia Assemblage (LBSA) for this series of faunas (Fig. 1) (Karatajute-Talimaa, 1998; Blieck et al., 2001; Turner et al., in press a). Almost everywhere a

Fig. 1. Stratigraphic distribution of the Gondwana Endemic Assemblage (GEA) and the Laurentia^Baltica^Siberia Assemblage (LBSA) of vertebrates in the Ordovician. Chronostratigraphic scale of Webby et al. (in press).

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

latest Ordovician to earliest Silurian (earliest Rhuddanian) gap in the vertebrate record exists, which we called the Talimaa’s Gap (Turner et al., in press a). Sadler and Cooper (in press) provided a calibration of global Series intervals of 17, 11.5, and 17.5 Myr for the Early, Middle, and Late Ordovician respectively. During this comparatively short time span true mineralized vertebrates appeared in the Arenigian (see Turner et al., in press a, for our de¢nitions of vertebrate status). From the Darriwilian onward, however, the record becomes ¢rmer and diversity increases rapidly especially on the LBS blocks. By the Caradocian it appears that all major vertebrate clades are present and the Darriwilian to Caradocian was the time of their major expansion. Signi¢cant is the disappearance of the GEA at ca. 3453 Myr with a slight overlap with the incoming of vertebrates in North America. The last 3 Myr of the Ordovician show an upsurge of new taxa and possibly new clades of thelodonts and pteraspidomorphs. All of these evolutionary events in the vertebrate record had an underpinning of the times the animals lived through. Here we have sought to analyze recent palaeogeographic models to test against the actual evidence of the vertebrate record. This paper is a partial answer to the challenge for ‘a new evolutionary synthesis’ put forward by Carroll (2000), which we stress is necessary for every geological period, and in particular for the great Ordovician biodiversi¢cation event. Carroll (2000, p. 27) de¢ned a program of research as follows : (1) To increase knowledge of the fossil record and get accurate geological dating: this is now more clearly demonstrated for Ordovician vertebrates (Turner et al., in press a,b). (2) To de¢ne plate tectonics/continental drift in£uence on climate and capacity of dispersal of organisms as major forces in driving evolutionary change : this is the major topic of the present paper. (3) To elaborate on phylogenetic systematics: this has still to be properly made as there is no consensus on basal vertebrate relationships (comments in Turner et al., in press a).

39

(4) To obtain the contribution of molecular developmental biology: we are partly concerned with this because most clades are extinct, and no molecular data are available on Ordovician taxa (‘ostracoderms’ sensu Janvier, 1996b, 2001, and gnathostomes). Carroll (2000, p. 28, and cited references) observed that ‘largest scale change T between the cephalochordates and early vertebrates’ occurred ‘when the number of Hox clusters duplicated twice, resulting in four clusters by the time early bony ¢sh (Osteichthyes) appeared some 415 million years ago’, i.e., in the Silurian according to Carroll’s (2000, ¢g. 1) geological time scale. This cannot be tested yet on fossil cephalochordates and ‘cyclostomes’ (hag¢shes and lampreys) which are virtually unknown in the Cambrian^Ordovician record, except for putative cephalochordates and allied basal chordates from the Early and Middle Cambrian of China and British Columbia (references and comments in Smith et al., 2001), and a putative lamprey from the Caradocian, Harding Sandstone of Colorado (new genus B of Sansom et al., 2001, ¢g. 10.4e^f). Neither can it be tested on ‘ostracoderms’ and earliest known gnathostomes (placoderm-, chondrichthyan- and acanthodian-like in Ordovician times, Turner et al., in press a,b, and references therein). As already assessed by various authors (among them, e.g., Carroll, 1997, pp. 349^350), the incompleteness of the early fossil record of vertebrates precludes any detailed view of their phylogenetic relationships, ¢rst adaptive radiation, and tempo and mode of evolution. Some recent analyses of molecular phylogeny of living vertebrates estimate the time of divergence between cephalochordates and early vertebrates at about 3751 R 30.9 Myr (Hedges, 2001), that is, nearly 262^277 Myr before the earliest veri¢ed fossil record of mineralized vertebrates, depending on whether the latest Cambrian to earliest Ordovician remains are considered as vertebrates or not, conodonts being excluded from the early vertebrate record (Turner et al., in press a,b); and nearly 221 Myr before the earliest putative vertebrates from the mid Lower Cambrian of China (Shu et al., 1999). The hypothesis of an early radiation of vertebrates in relation to the end-Proterozoic glacial events

PALAEO 3072 9-5-03

40

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

(the Neoproterozoic Refugia Model in Hedges, 2001) is clearly in contradiction with our hypothesis of early radiations in the Ordovician (¢rst in the Early^Middle Ordovician and mainly during the Darriwilian, second in the Late Ordovician, that is, mostly during the Caradocian: Turner et al., in press a). However, if Donoghue et al. (2000) and Smith et al. (2001) are right and certain conodonts are true vertebrates, then there is a much longer record to account for. This paper is complementary to Smith et al.’s (2002) and Young’s (in press) papers on Early Palaeozoic and Siluro^Devonian vertebrate biogeography respectively : what are the conditions needed during and at the end of the Ordovician, which could explain the situation at the beginning of the Silurian (see, e.g., Blieck and Janvier, 1991, 1999; Turner, 1999)? All three papers will thus give the ¢rst general overview of early vertebrate biogeographic dispersal.

2. Material studied The material on which our review is based is published in two separate papers, one on the interpretation of biodiversity (Turner et al., in press a) and one on the database of Cambrian^Ordovician vertebrates (Turner et al., in press b). As already said elsewhere about global palaeobiodiversity analyses (e.g., Foote and Sepkoski, 1999), our data can only be based on what is known (published or not), which is the result of ¢eld explorations. It appears that most data come from countries where ¢nancial means in terms of employed people and ¢eld expeditions could be mounted, so that the various data are clearly from restricted areas, mostly North America, former USSR (European and Asian, Siberian parts), and Australia. Some data also come from Gondwanan regions other than Australia, that is, South America (Bolivia and Argentina) and South Africa (a ¢rst and problematic record of an anaspid). Several vast areas are still totally devoid of information, i.e., Europe ^ except Timan^Pechora; nonRussian Asia ^ except perhaps China; Africa ^ except perhaps South Africa ; and Antarctica. This severely a¡ects any interpretation of the dy-

namics of origin and evolution of earliest vertebrates, and thus their palaeobiogeographic interpretation. The other point that we want to stress is that (palaeo)biogeography is a historical science, and that any biogeographic scenario is provisional, as pointed out by many other authors (e.g., Smith et al., 2002).

3. Habitat and dispersal of Ordovician ¢shes McKerrow et al. (2000, p. 10) have recently summarized various organisms commonly preserved as fossils into ¢ve categories (a^e) of increasing biogeographic utility in distinguishing regions which were approaching collision (also Young, in press). Fishes were included into categories ‘a’ [free swimming (or pelagic) nekton], possibly ‘b’ (benthic, but with pelagic larval stage), and ‘d’ [restricted to non-marine or brackish environments (or with reduced marine in£uence) such as the Old Red Sandstone sediments] (ibid.). However, as concerned with ¢sh assemblages of the Old Red Sandstone type, ‘analysis of empirical data for some well-known fossil ¢sh occurrences T has not produced unequivocal evidence of a strictly freshwater habitat’ (Young et al., 2000, p. 211). Category ‘a’ is of little help for studying the palaeogeographic relationships of palaeocontinents and their shelves (but of paramount importance in long distance biostratigraphic correlations). This category probably had a reduced capacity of fossilization because, inter alia, most Palaeozoic oceanic £oors have disappeared. We thus remain with two main categories of ¢shes: (I) benthic taxa on the continental shelves, but with pelagic larval stages allowing great dispersal capacity in marine environments, this characterizing very widespread taxa; these are not very di¡erent from pelagic taxa; (II) taxa with very limited distribution, apparently con¢ned by marine barriers (e.g., wide oceanic areas, longitudinal oceanic currents), either on continental shelves in true marine environments, or restricted to non-marine-brackish environments. I and II correspond respectively to the ‘oceanic’ and ‘continental’ groupings of Rosen (1974, p. 323; also Young, in press). Group II, ‘continen-

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

tal’ is the most appropriate to test the palaeoposition of continents through time. In Ordovician time, because of the very incomplete fossil record, most taxa seem endemic and linked to very restricted geographic areas, and would correspond to category II. All pteraspidomorph agnathans have indeed been collected in marine benthic assemblages BA1 (marine intertidal) to BA3 (shallow subtidal) sensu Boucot and Janis (1983), and they do not seem to have been capable of transoceanic migrations (Gagnier et al., 1986; Elliott et al., 1991; Gagnier and Blieck, 1992). Similarly, many thelodont taxa are typical of BA1 to BA3 (Turner, 1999). A few taxa only might correspond to category I, ‘oceanic’, viz., chondrichthyan-like, ?mongolepids and, perhaps, some thelodonts, but not in Ordovician (references in Turner et al., in press a,b); more likely their more widespread distribution is explained by their having pelagic larval dispersal.

4. A global palaeogeographic problem Ideas regarding the palaeogeographic relationships of the various ‘northern’ landmasses and the southern Gondwana supercontinent for Ordovician time are still in a state of £ux. This problem was treated brie£y in a previous review of Ordovician vertebrates by Elliott et al. (1991), who concluded that there was an inadequate ¢t between the known fossil record of vertebrates and proposed global palaeocontinental reconstructions. This same situation has prevailed in latest symposia on Early Palaeozoic palaeogeographies. The most crucial point that Elliott et al. (1991) observed was the inconsistency of the wide ocean hypothesized between Laurentia (for the North American Ordovician localities) and Gondwana (for the Bolivian and Australian localities), which mitigated against any direct migratory relation between the two palaeocontinents. This problem can also be stressed with regard to Baltica and Siberia, from where Ashgillian vertebrates are now known. Most Ordovician vertebrate-bearing localities are interpreted as corresponding to marine benthic assemblages BA1 to BA3 (see above), and Ordovician vertebrates do not seem to have

41

been capable of transoceanic migrations (references ibid.). So, to explain the occurrence of vertebrates on both extremities of Gondwana and other Ordovician landmasses, we need palaeogeographic reconstructions based upon geological and palaeomagnetic data which show a rather close connection between those palaeocontinents, at least during some time slices of the Ordovician Period. Or else we have to conclude in a diphyletic origin for vertebrates, which is against the current universally accepted concept of monophyly of the group. Reconstructions with too wide oceanic areas preclude direct biogeographic relations between the vertebrate localities of the continents. This is true in any biogeographic model. In a migratory model, we would hypothesize migrations of vertebrates (or their juveniles or larvae) from the Late Cambrian^Early Ordovician neritic platforms of Australia to the Middle^Late Ordovician ones of Argentina-Bolivia, then to the Late Ordovician (Caradocian) ones of N. America, and ¢nally to the latest Ordovician (Ashgillian) sites of Siberia and Baltica. (This model has been named the ‘out of Gondwana model’ by Smith et al., 2002.) In this case, we need a close relationship (even if not a direct contact) between Gondwana and Laurentia, which is not the case on some of the proposed reconstructions. If we do not hypothesize ancestor^descendant relationships between the various Ordovician vertebrates, but consider the problem within the framework of a cladistic analysis (Elliott et al., 1991, ¢g. 5A; Gagnier, 1995, ¢g. 7; Janvier, 1996a, ¢g. 9.1) (see Section 6), we need a generalized track covering Gondwana+Laurentia or Laurentia+Baltica+Siberia with a close relationship between those various blocks. (For alternative palaeocurrent scenarios, see below in Section 6.) Smith et al. (2002) did consider the biogeographic dispersal of early vertebrates under the support of a cladistic scheme, and concluded to what may be named an ‘out of Laurentia model’, where ‘the latest common ancestor of all ‘ostracoderms’ and jawed vertebrates was Laurentian rather than Gondwanan (contra Elliott et al., 1991)’ (Smith et al., 2002, pp. 76^77). This result, however, is largely dependent upon the fact that

PALAEO 3072 9-5-03

42

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

they consider (1) the conodonts as the sister group of ‘ostracoderms’+gnathostomes, and (2) the arandaspids as a derived (rather than basal) group of pteraspidomorphs (their ‘heterostracomorphs’), so that the cosmopolitan distribution of conodonts, and the Laurentian distribution of basal pteraspidomorphs, earliest thelodonts, and chondrichthyans would favor a Laurentian track for the early dispersal of mineralized ‘non-conodont vertebrates’ (sensu Smith et al., 2002). Nevertheless, several data and interpretations are against this model : conodonts may not be vertebrates (our hypothesis); arandaspids are generally considered as basal pteraspidomorphs; the possible Caradocian chondrichthyans of Australia, and the putative Ashgillian anaspid of South Africa do not ¢t the model ; the latest Cambrian possible vertebrate from the Gola Beds of Australia (Young et al., 1996) is not taken into consideration by Smith et al. (2002) ; the Cambrian^Ordovician taxon Anatolepis, even if endemic to Laurentia, is not placed within their cladistic scheme. So, we treat the problem quite di¡erently. Several di¡erent solutions have been proposed for Ordovician global palaeogeographies. Here we will consider four groups of solutions, viz. (1) what is now classically known as the ‘archetypal’ solution of, e.g., Scotese and others; (2) the ‘alternative palaeogeographic approach’ of Dalziel and others ; (3) another alternative solution of Lewandowski ; and (4) the ‘compact version’ of Boucot et al. In the archetypal palaeogeographic reconstruction of Scotese (1986; Scotese and McKerrow, 1990, 1991 ; also Trench and Torsvik, 1992; McKerrow and Cocks, 1995; Scotese, 1997; Lewandowski, 1998) (archetypal sensu Torsvik et al., 1995, p. 284), rather wide oceanic areas are hypothesized between western Gondwana and Laurentia on one hand, and between Siberia (+Kazakhstania) and eastern Gondwana on the other hand. This is true for both Early Ordovician reconstructions (Arenigian : Trench and Torsvik, 1992, ¢g. 1; McKerrow and Cocks, 1995, ¢g. 2; Torsvik and Rehnstro«m, 2003, ¢g. 6A; Tremadocian to Arenigian/Darriwilian: Tait et al., 1997, ¢gs. 3,4) and Late Ordovician reconstructions (‘Llandeilo’^Caradocian : Scotese and McKerrow,

1991, ¢g. 4; Trench and Torsvik, 1992, ¢gs. 1,2; Tait et al., 1997, ¢g. 5; Torsvik and Rehnstro«m, 2003, ¢g. 6B). A rather di¡erent solution has been proposed by Neuman (1984; also Nowlan and Neuman, 1995) who hypothesized numerous ‘insular terranes’ between the major landmasses (viz., Laurentia, Baltica, NW Gondwana) in Arenigian^ Darriwilian time. These intermediate elements (mostly volcanic) might have worked as ‘staging posts for the dispersal of shallow-marine biota’ (sensu Talent, 1985, ¢g. 1; Talent et al., 1987, p. 92 and ¢g. 1), and the vertebrates might have migrated under a process called ‘sweepstake migration’ by Simpson (1969, p. 70: ‘route de courses d’obstacles’ in French ; also Blieck and Janvier, 1991, p. 378; Blieck et al., 2002). Such a pattern with intermediate volcanic arcs in the Iapetus Ocean is also hypothesized by Mac Niocaill et al. (1997, ¢g. 2 and p. 161: ‘complex geometry, resembling, in many ways, the tectonic complexity of the modern southwest Paci¢c’; also Van Staal et al., 1998; the complex history of Avalonia, the Armorican Terrane Assemblage (ATA) and Moldanubia in Franke, 2000; and the numerous volcanic arcs in Cocks, 2001, ¢g. 1 for the Arenigian). This pattern would allow direct relationships between Laurentia, Avalonia, and NW Gondwana. However, on this model, Siberia is too far away from Laurentia. In another solution based on palaeomagnetic and palaeoclimatic data, Golonka et al. (1994, ¢gs. 10^12, ‘Llandeilo’) retain wide oceans between Gondwana and the ‘northern’ landmasses, but the latter are grouped together in an equatorial to southern tropical location. This could satisfy a generalized track for the LBSA of Late Ordovician vertebrates (Turner et al., in press a). Such a close connection between LBS is also reconstructed for Late Ordovician time by Tait et al. (2000, ¢g. 3b), based on palaeomagnetic data. In the alternative palaeogeographic approach of Dalziel (1997, ¢gs. 4,15,16), Laurentia is reconstructed in an equatorial location just north of the South American margin of Gondwana, with an intermediate palaeocontinental element in between, called the ‘Texas Plateau’. This model would solve some of the problems concerning

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

the vertebrates, that is, by providing a rather close contact between Gondwana and Laurentia. It would also be acceptable to explain other biogeographic relationships such as the distribution of Early Palaeozoic benthic foraminifers collected both in the SE USA and Bolivia (Vachard, in Gagnier et al., 1996). Erdtmann (1998, ¢g. 1) also hypothesized a drift of Laurentia along the western margin of Gondwana, with a rather close relation between Laurentia and S. America in Early Ordovician time. However, on Dalziel’s model, Siberia is once again too far away from Laurentia. This problem is perhaps solvable if we displace Siberia westwards in a closer relation to Laurentia (but of course with the same latitudinal position). This is the case in the model of Torsvik et al. (1995, ¢g. 11d^f) which shows a close relation between S. America, N. America, and Siberia. However, a comment has to be made about the conclusions of Torsvik et al. (1995). In their abstract, they say, ‘T a tight continental ¢t (between Laurentia and S. America) during the entire Ordovician is contradicted by biogeographic data’ although, in their text (p. 280), they say, this ‘T is apparently contradicted by some of the biogeographic evidence’. The biogeographic evidence is here taken from papers by McKerrow, Cocks and Fortey (references in Torsvik et al., 1995), and is used in the archetypal model explained here above. We submit here that other biogeographic evidence (vertebrates and foraminifers at least as outlined above) do support the alternative approach of Torsvik et al. (1995) and Dalziel (1997), the latter being the only one tested by Smith et al. (2002). We also place in this group of models that of Benedetto (1998) who reconstructed a series of insular, partly volcanic, terranes including the Precordillera terrane (Cuyania) of Argentina between Laurentia and S. America in Early Ordovician time (Arenigian/Llanvirnian). This Precordillera terrane is considered as having been drifting from a Laurentian location to a West Gondwanan location, based on mostly benthic faunal evidence (microplate hypothesis of Benedetto, 1998, ¢g. 6); it was thus accreted to S. America at least in the Late Ordovician (Benedetto et al., 1999, ¢g. 6).

43

Another alternative solution has been proposed by Lewandowski (1993, ¢g. 27) for the Tremadocian^Arenigian. Apart from the problem of the location of the Holy Cross Mountains in Poland, and of Cadomia (including the Armorican Massif of France), which is treated in Lewandowski’s paper, his model shows a rather close relationship between Laurentia, Siberia and the Australian margin of Gondwana, with narrow oceanic areas in between. This could well be another good solution for explaining the vertebrate distribution in Ordovician time, except that the migratory route from Australia, through Siberia, to Laurentia is not in agreement with the ages of the fossil record: Early to early Late Ordovician in Gondwana, Ashgillian in Siberia and Baltica, Caradocian in N. America (given that the fossil record is complete, which is probably not the case). Lewandowski’s (1993) model also requires a (much too) rapid clockwise rotation of the whole Gondwana for bringing West Gondwana towards Laurentia in post-Ordovician time. Finally, the palaeoclimatically based compact version of Boucot et al. (1995; also Scotese et al., 2001, ¢gs. 3,4) for Ibexian (Early Ordovician) and post-Ibexian^pre-Hirnantian (Middle^Late Ordovician) time puts the neritic platforms close together in a southern latitudinal position, enabling migratory relations between Laurentia, Baltica, West and East Gondwana, but excludes Siberia, which is reconstructed in rather low northern latitudes. However, this model does not take into account the traditionally reconstructed oceanic areas such as the Iapetus Ocean between Laurentia and Baltica (e.g., Vannier et al., 1989; Erdtmann, 1991; Oliver et al., 1993) or the Tornquist Sea between Baltica and NW Gondwana (‘Armorica’) (same references; also Servais and Fatka, 1997) (without entering the debate on the reality of the Tornquist Sea; see, e.g., Paris and Robardet, 1990; vs. Servais and Fatka, 1997.) At the same time, in the model of Boucot et al. (1995), no oceanic area seems to have been intercalated between the various terranes of East Asia, SE Asia and East Gondwana, in agreement with the solution of Metcalfe (1999, ¢g. 3). In such a confusing situation, which solution to

PALAEO 3072 9-5-03

44

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

select? A model that would associate close palaeogeographic relationships between the major continents (that is, Gondwana, Laurentia, Baltica and Siberia as concerned with Ordovician vertebrates) has to be preferred. A combination with intermediate smaller elements such as exotic terranes, drifting microplates, and/or volcanic arcs might also be favored. As none of the recently published reconstructions ful¢ll all requirements, we choose here one of the latest, that is, Li and Powell’s (2001) series of global Ordovician models (also in Webby et al., 2000) where Gondwana and Laurentia are rather closely related in the Early

Ordovician (their ¢g. 9), and LBS relatively grouped in the intertropical zone of the Late Ordovician (their ¢g. 10), with three major subduction zones in between and their probable associated insular terranes.

5. Fish distribution patterns in the Ordovician As stressed in our review of Cambrian^Ordovician vertebrate diversity (Turner et al., in press a), Ordovician vertebrates are highly endemic. For each selected time slice, only a few spots with

B

A Porophoraspis Arandaspis

PC

IV Po

F

LAU

LAU

SCB Pi

GON

GON

PC S

IAP

S AVA

BAL

C

Sacabambaspis Areyongalepis Apedolepis indet. vert.Assemblage 3 indet. vert.Assemblage 4

lamprey? (nov. gen.B) chondrichthyan A of. Sinacanthus thelodont (nov. gen. C) loganiid Astraspis

mongolepid? chondrichthyans? (nov. gen. A,F) Skiichthys thelodontid loganiid? Eriptychius

D "Tesakoviaspis" Astraspidida? Acanthodii n. gen.A

SIB

LAU GON Sa? PC GON

IAP

LAU Ac?

An?

IAP

S AVA

AVA

BAL

"Tesakoviaspis" Sandivia Stroinolepis

Fig. 2. Fish distribution pattern in four time slices of the Ordovician. (A) Lower Arenigian vertebrates on the reconstruction of Li and Powell (2001, ¢g. 9) at ca. 3480 Ma (Tremadocian). (B) Darriwilian vertebrates on Li and Powell’s reconstruction of Webby et al. (2000, ¢g. 7). (C) Caradocian vertebrates on the reconstruction of Li and Powell (2001, ¢g. 10) at ca. 3450 Ma (latest Caradocian). (D) Ashgillian vertebrates on the same latest Caradocian reconstruction. Abbreviations for palaeocontinental elements: AVA ^ Avalonia, BAL ^ Baltica, GON ^ Gondwana, LAU ^ Laurentia, PC ^ Precordillera, SCB ^ South China Block, SIB ^ Siberia; and for palaeocean: IAP ^ Iapetus. Abbreviations for taxa: F ^ Fenhsiangia, Po ^ Porophoraspis, IV1 ^ indet. vertebrate Assemblage 1, Pi ^ Pircanchaspis, S ^ Sacabambaspis, Sa? ^ cf. Sandivia, An? ^ anaspid?, Ac? ^ acanthodian?

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

Ordovician vertebrate localities are known, and each palaeocontinent bears a totally di¡erent taxic assemblage. In pre-Arenigian time, the record is very fragmentary and problematic, and is not examined here. In the early Arenigian (time slices 2a^2c of Webby et al., in press), Gondwana only is colonized by taxa with questionable vertebrate a⁄nities, viz., Porophoraspis and an ‘indet. vertebrate Assemblage 1’ in the Amadeus Basin of Australia, and Pircanchaspis in Bolivia. Additionally, we can mention the phosphatic fragments Fenhsiangia from the South China Block (time slices 1d^2a) which would represent the only non-Gondwanan record for both the Lower and Middle Ordovician, if we reject the fragment of ‘plate’ from the lower Middle Ordovician of Inner Mongolia, which probably corresponds to a non-vertebrate record (references in Turner et al., in press a,b) (Fig. 2A). The Darriwilian record is more abundant and also restricted to Gondwana: Porophoraspis, Arandaspis and Sacabambaspis in 4b; Sacabambaspis alone in 4c, and on to early Caradocian 5a, the two former genera being from the Amadeus Basin of Australia, and Sacabambaspis from Bolivia and Argentina in the Precordillera terrane (Fig. 2B). This record identi¢es what Young (in Webby et al., 2000, ¢g. 7) called a ‘Sacabambaspis fauna’, and Turner et al. (in press a) a ‘Gondwana clade assemblage’ based upon the arandaspidiform agnathans Sacabambaspis and Arandaspis. This is the major constituent of our GEA. Sacabambaspis is still recorded in the Caradocian (5b^5c) of the Amadeus Basin together with Areyongalepis, Apedolepis, an ‘indet. vertebrate Assemblage 3’, and an ‘indet. vertebrate Assemblage 4’. They represent the last record of vertebrates on the Gondwana supercontinent from where they disappear until they re-occur in Silurian time (Pickett et al., 2000) ^ unless we except the claimed anaspid recorded from the late Ashgillian (6c) of South Africa (references in Turner et al., in press a,b). In Caradocian time, the most diversi¢ed record is found on Laurentia with the bulk of the assemblage from the Harding Sandstone of Colorado (5a^5c) and its numerous equivalents all over USA and Canada. Nearly 13 di¡erent taxa are

45

now known there, among them the pteraspidomorph agnathans Astraspis and Eriptychius were the ¢rst to be described. Both genera have been recorded from 5a^5c to 5d. The other elements of the Harding Sandstone fauna (i.e., the Laurentian component of our LBSA) include several problematic taxa, i.e., a lamprey?, thelodonts, Skiichthys (osteostracan? or placoderm?), cf. Sinacanthus (acanthodian? or shark?), and chondrichthyans (Fig. 2C). This de¢nitely settles the occurrence of both agnathans (lamprey? and ostracoderms sensu Janvier, 1996b) and gnathostomes in the Ordovician. It also corresponds to a probable radiation event, as far as we can judge from the existing fossil record. Lastly, further uncertain thelodont and acanthodian records are known from 5c^6a. The Ashgillian record is poorer than the Caradocian, and mostly outside Laurentia, except the problematic acanthodian-like spine from Girvan, in the Midland Valley of Scotland (6b). This might signify a second possible extinction event on Laurentia where both agnathans and gnathostomes ‘come back’ in Silurian time (Blieck and Janvier, 1991, 1999; Turner, 1999). Other Ashgillian vertebrates are known from Baltica (mostly the Timan^Pechora region in the NE part of European Russia) with the enigmatic ‘Tesakoviaspis’, and thelodonts Sandivia and Stroinolepis (6cLlandovery), and from Siberia (including Tuva and Mongolia) with again ‘Tesakoviaspis’ and Astraspididae? (6c-Llandovery), and ‘Acanthodii n.g. A’ (6c) (Turner et al., in press a,b). This would identify a ‘Tesakoviaspis fauna’, i.e., the Baltica^Siberia component of our LBSA (Fig. 2D).

6. Problems and interpretation The Ordovician is characterized, inter alia, by (a) widespread epeiric seas with extensive carbonate accumulations in low latitude palaeocontinents, (b) a generally high sea level, particularly in the Early and Late Ordovician with a drastic drop in the latest Ashgillian, (c) wide dispersal of palaeocontinents, and (d) a greenhouse state climate during most of the period, with a deteriora-

PALAEO 3072 9-5-03

46

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

tion toward an icehouse state in the Ashgillian (e.g., Ross and Ross, 1992; Barnes et al., 1996; Morrow et al., 1996). In such a global context, all the vertebrate localities, as plotted on Li and Powell’s (2001) reconstructions, fall within the intertropical zone of the Ordovician (Fig. 2), that is certainly in warm water conditions (Spjeldnaes, 1979, ¢g. 10), even if the Earth was rotating faster in Ordovician times than today, with climatic zones nearer to the equator (references in Christiansen and Stouge, 1999). This being stated, how to interpret the strong endemism of vertebrate assemblages? First problem : Cambrian vertebrates (excluding conodonts once again) are very poorly known, and might be represented by ‘naked agnathans’ in China and a single mineralized taxon from Australia. The latter has super¢cial similarities with Porophoraspis, known later in the Ordovician of Australia as well (Fig. 2A,B). Does this mean that vertebrates may have originated in ‘eastern’ Gondwana (in fact northern Gondwana on the Cambrian palaeocontinental reconstruction of Eldridge et al., in Brock et al., 2000), or is it simply due to a strong sampling artefact? The latter seems more probable. Sampling of all potential Cambrian limestones in Australia is recommended. Acceptance of this hypothesis refutes the ‘out of Laurentia’ theory of Smith et al. (2002). Second problem : in our review of Cambrian^ Ordovician vertebrates (Turner et al., in press a), we considered some Anatolepis records as possible vertebrate-like occurrences from the latest Cambrian to late Early Ordovician of Wyoming, Utah and Texas (Nitecki et al., 1975, ¢g. 3; Repetski, 1980, ¢gs. e^f; Smith et al., 1996, ¢gs. 1^ 3). If really vertebrates, these ‘Anatolepis’ records could serve as a test for the Texas Plateau hypothesis of Dalziel (1997), as pointed out by Smith et al. (2002). The only problem is that no vertebrate-

47

like Anatolepis is presently known from Gondwana, where the GEA starts in the late Early Ordovician. Third problem : what connection/relation can we hypothesize between the GEA on Gondwana and the LBSA on LBS? Apparently none for the time being: there is no common taxon between GEA and LBSA. However, at the very least, somewhere either on Gondwana or on LBS (preferably Laurentia where the oldest record of the LBSA is known), at least one speciation event is needed as an origin for the LBSA taxa. Let us take the pteraspidomorph vertebrates as an example. They are known ¢rst in Australia (Late Cambrian ?^Early Ordovician), then in South America (Middle Ordovician), in North America (Caradocian), and ¢nally in Baltica^Siberia (Ashgillian). We might thus be tempted to imagine a faunal spreading for pteraspidomorphs from Gondwana as a center of origin, to Laurentia, and then Baltica and Siberia (the ‘out of Gondwana’ theory). But, what are the phylogenetic relationships between the GEA and LBSA pteraspidomorphs, that is, between arandaspids and Astraspis, Eriptychius, or between Astraspis and ‘Tesakoviaspis’? Most recent cladistic analyses, however, have resulted in no consensus concerning basal vertebrate relationships, and particularly pteraspidomorphs (see short comments in Turner et al., in press a). For instance, Eriptychius was supposed to be the sister group of Astraspis by Gagnier (1995, ¢g. 7) in the following topology [((((Pteraspidiformes, Cyathaspidiformes) (Eriptychius, Astraspis)) (Arandaspis, Sacabambaspis)) petromyzontids) myxinoids], rooted on cephalochordates (Fig. 3A); or the sister group of gnathostomes by Donoghue et al. (2000, ¢gs. 14,17) in the following topology [Myxinoidea (Petromyzontida (Conodonta ((Astraspis (Arandaspida, Heterostraci)) ((Anaspida (Jamoytius, Euphaner-

Fig. 3. Several recent schemes of relationships among early vertebrates, with details of pteraspidomorphs and their geographical distribution, from (A) Gagnier (1995, ¢g. 7); (B) Janvier (1996a, ¢g. 9.1); (C) Janvier (1996b, ¢g. 5B); (D) Janvier (1997, ¢g. 1A); (E) Janvier (1998, ¢g. 7); (F) Donoghue et al. (2000, ¢g. 14). Solutions A, B, D ¢t better both the geographical and stratigraphical records of pteraspidomorphs. Abbreviations for palaeocontinents as in Fig. 2. H is for Heterostraci, and P for Pteraspidomorphi (polyphyletic in solution F).

PALAEO 3072 9-5-03

48

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

ops)) (Loganellia ((Eriptychius, Gnathostomata) (Osteostraci, Galeaspida)))))))], rooted on Tunicata and Cephalochordata (Fig. 3F). Note, however, that in Donoghue et al. (2000) 85% of data was missing for Eriptychius, and a number of characters were miscoded. As for Janvier (1996a,b, 1997, 1998), nothing seems de¢nitely settled as he placed Eriptychius either as the sister group of Arandaspida+Heterostraci (Janvier, 1996b, ¢g. 5B; nearly similar to Smith et al., 2002, ¢g. 4) (Fig. 3C), of Heterostraci+?other ‘Thelodonti’ (Janvier, 1997, ¢g. 1) (Fig. 3D), or of Heterostraci alone (Janvier, 1996a, ¢g. 9.1; Janvier, 1998, ¢g. 7) (Fig. 3B,E). We can just note that some of these proposals are in disagreement with the fossil record, leading to long-standing ghost lineages as, e.g., supposed early astraspids, anaspids, and basal taxa of the clade (Osteostraci, Galeaspida, ?Pituriaspida) in Donoghue et al.’s (2000, ¢g. 17) hypothesis. On the contrary, the ghost lineage for basal chondrichthyans, as in one of Janvier’s (1998, ¢g. 7) hypotheses, is now partly ¢lled with the discovery of Ordovician taxa as early as the Caradocian, and the more classical hypothesis of Janvier (ibid.) may be considered as ¢tting better the stratigraphical record. (Except for myxinoids and lampreys, nearly unknown as fossils ^ except the Chinese, Early Cambrian taxa of Shu et al. (1999), and a Caradocian ?lamprey mentioned by Sansom et al. (2001), and much later Carboniferous taxa; and thus, with very long-lasting ghost lineages in all analyses.) [This problem of ghost lineages is more developed by Smith et al. (2002).] The problem is also that ‘scale taxa’ such as ‘Tesakoviaspis’ have still to be accurately described, and they have never been introduced in any cladistic analysis. In such conditions, any scenario of spreading of pteraspidomorphs, based upon stratigraphy and geography is nothing else than an author’s opinion, where exotic terranes and/or microcontinents such as the Texas Plateau of Dalziel (1997), drifting between or connected to Laurentia and Gondwana, might have played a signi¢cant role for vertebrate dispersal in Ordovician time. In another model, this role would be played by either west£owing equatorial ocean currents between Siberia,

Laurentia and Gondwana (in the apparent wrong direction as concerned with vertebrates), or by an east-£owing equatorial counter current between Gondwana, Laurentia and Siberia (this time in a perceived accurate direction for vertebrates), which would have transported pelagic larvae between the widely separated Ordovician palaeocontinents (Webby et al., 2000). This just means that we do not know. The palaeocurrent hypothesis is an ad hoc scenario, inferred from the supposed distribution of landmasses vs. oceans (e.g., Wilde, 1991 ; also Webby et al., 2000; vs. Christiansen and Stouge, 1999), and the distribution of mostly benthic marine faunas (trilobites, molluscs, gastropods, brachiopods, and echinoderms), including encrusting/reef building organisms (bryozoans, sponges, and stromatoporoids), and some nektonic taxa (conodonts and nautiloids) (Christiansen and Stouge, 1999; Webby et al., 2000), which cannot be tested by the currently known Ordovician pteraspidomorph zoogeographical pattern. Nevertheless, at least one regional problem seems to be solved. The Precordillera microplate model of Benedetto (1998; Benedetto et al., 1999) would explain the occurrence of the same arandaspid genus Sacabambaspis both in the Precordillera (Argentina) and West Gondwana (Bolivia) in the Darriwilian (references in Turner et al., in press a) during what Benedetto et al. (1999) called the pre-accretion stage (Llanvirnian^Caradocian) of the Precordillera terrane (Fig. 2B). Sacabambaspis would have been able to colonize the Precordillera from Gondwana at a time when both elements were not too far away from each other. Moreover, how to integrate the other Ordovician vertebrates (thelodonts, and gnathostomes) in this scheme? Siberia has traditionally been seen as a point of origin for thelodonts (Blieck and Goujet, 1978; Karatajute-Talimaa, 1978; Turner, 1999), which are, however, now known from Timan^Pechora (Baltica), with the earliest Silurian and earlier Ordovician taxa in Laurentia (Sansom et al., 2001; Turner et al., in press a). Thelodonts (Loganellia sp.) had even dispersed as far as Mongolia by late Llandovery (e.g., Turner, 1999). Once again, in absence of a detailed phy-

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

logenetic analysis of thelodonts, and taking into account stratigraphy and palaeogeography, thelodonts would have originated in Laurentia before spreading to Baltica and Siberia, thus paralleling the pteraspidomorph pattern. Fourth problem: what other kinds of external, physical factors can explain the strong radiation of Ordovician vertebrates? Can¢eld (1998; also Carroll, 2000) noted ‘substantial increases in the availability of atmospheric oxygen, which would have enabled the achievement of larger body sizes and the formation of calcareous or phosphatic skeletons in many lineages’ in the Early Cambrian. However, such an increase in oxygen values is not observed for Ordovician time by Berner et al. (2000) who identi¢ed a rather long stable oxygen curve from the Cambrian to the Devonian. Nevertheless, some comments must be made as concerned with oxygen. One problem is that the oceanic oxygen content can only be indirectly determined quantitatively by, e.g., the isotopic composition of sedimentary sul¢des which is in£uenced also by other factors. Especially for Palaeozoic oceans, the range of mechanisms in£uencing the oxygenation of ocean and atmosphere are not yet fully understood. Oceanic surface water was more or less directly connected to the atmosphere. The large reservoir of the deep ocean, however, remains unknown due to the subduction of Palaeozoic deep-sea sediments. For example, many authors suggest that during the uppermost Ordovician (Hirnantian glaciation) the global oceanic circulation was similar to the modern one (with an oxygenated deep ocean) (e.g., Brenchley et al., 1994, 1995). This assumption seems questionable because at that time the edges of Gondwana extended well into the subtropics, thereby prohibiting a deep-water convection as it is known in the modern ocean (Munnecke et al., 2003, this issue). So, the Early Cambrian increase in oxygen, plus the great deposition of phosphate in the Late Proterozoic (at ca. 3750 Myr, e.g., Cook and Shergold, 1986), which might or might not have been related to the breakup of the proposed supercontinent Rodinia (Li et al., 1996; Li and Powell, 2001), might have been the spur for vertebrate evolution before the phase when hard tis-

49

sues appeared. For example, Halstead (1974) postulated a link between the phosphate cycle, apatite-based skeletons and mitochondrial activity in vertebrates, but did not know of the evidence of past times of superabundance of phosphate in sea water which would satisfy many requirements. Nevertheless, this hypothesis is in contradiction with the fact that phosphatic skeletons occur in many invertebrate groups early in the Cambrian, and not only in the latest Cambrian^Ordovician as for vertebrates. Strontium isotope analysis of Ordovician and Silurian brachiopods and conodonts by Qing et al. (1998, ¢g. 6; also Holser et al., 1996, ¢g. 2, on apatite and carbonate) shows a gradual decrease in 87 Sr/86 Sr from Tremadocian to late Llanvirnian, a sharp decline during the late Llanvirnian^early Caradocian, little change during Caradocian^Ashgillian, and a steady rise through the Silurian. These variations have been interpreted by Qing et al. (1998) as controlled primarily by continental collisional tectonics and its associated erosion and weathering. In particular, they hypothesized that the rapid decline near the Llanvirnian/Caradocian boundary ‘suggests a strong hydrothermal £ux likely due to increased sea-£oor spreading and a possible superplume event’ which might ‘have caused the prominent transgressive phase, the largest in the Phanerozoic’ (also Barnes et al., 1996; Barnes, in press). This series of physical e¡ects might have been a factor in the changeover from a Gondwanan to a Laurentian focus for vertebrates.

7. Conclusion The evolutionary and biogeographic history of Ordovician vertebrates is perforce earlier than and di¡erent from that of Silurian^Devonian vertebrates, because in Ordovician times: (1) agnathans seem to predominate over gnathostomes in terms of species diversity; and (2) vertebrates were undergoing strong adaptive radiation, without apparent major extinction events, except perhaps in Gondwana during the Caradocian (where they only appear again in the Silurian), and in Laurentia during the late Ashgillian (where they

PALAEO 3072 9-5-03

50

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

occur again in the Silurian) (Pickett et al., 2000; Turner et al., in press a). Our review shows a strong endemic pattern for Ordovician vertebrates, with a GEA in Middle Ordovician time, and a LBSA in Late Ordovician time : a Laurentian fauna in the Caradocian, and Baltica^Siberia faunas in the Ashgillian, with no common genera between those palaeocontinental blocks. This pattern re£ects the currently proposed palaeogeographical reconstructions for the Ordovician, when Gondwana, Laurentia, Baltica and Siberia are supposed to have been widely separated (e.g., Li and Powell, 2001). This essay may be read as complementary to Young’s (in press) paper in the sense that it sets the scene at the end of the Ordovician, before (1) the enigmatic Talimaa’s Gap at the very beginning of the Silurian, and (2) the Silurian radiation events of vertebrates from either the ‘northern realm’ (LBS) (sensu Young, in press) and/or the southern supercontinent (Gondwana). It proposes also an interpretation which is di¡erent from that of Smith et al. (2002), who hypothesized that most Cambrian^Ordovician vertebrates, still unknown to science, were ecologically di¡erent from their Middle Palaeozoic relatives, and might also be found in deeper-water lithic facies. However, there are as yet too many uncertainties in the fossil record of Ordovician vertebrates and, therefore, our interpretations should be regarded with caution. As Lieberman (2002) says, after a simulation analysis of the e¡ects of palaeontological incompleteness on phylogenetic biogeographic analyses: ‘Phylogenetic biogeographic studies of fossil taxa should avoid groups with low diversity and a poor fossil record; these studies should also avoid time periods or regions with a poor fossil record’, which is exactly the present situation for Ordovician vertebrates. Publication of recently discovered taxa, and new discoveries will certainly modify the image proposed here. Nevertheless, what is absolutely needed is a new research program to discover and describe (publish) Early Palaeozoic macroremains and microremains altogether, as well as their morphology/anatomy and palaeohistology, the latter especially using and illustrated by the same techniques.

Acknowledgements This is a contribution to IGCP Project No. 410: The Great Ordovician Biodiversi¢cation Event. We are indebted to Drs. M. Legrand-Blain (Gradignan, France), O.H. Walliser (Go«ttingen University, Germany), J. Lazauskiene (Geological Survey of Lithuania, Vilnius), B. Meyer-Berthaud (Montpellier University, France), S. Steyer (while at Lille University, France), D.E. Can¢eld (Odense University, Denmark), C.R. Barnes (University of Victoria, Canada), T. Servais (Lille University), M.G. Carrera (Cordoba University, Argentina), and the librarians of the Socie¤te¤ Ge¤ologique de France in Paris and the Queensland Museum for help in ¢nding references. We thank the organizers of both the IGCP 421 meeting in Frankfurt am Main (May 2001) and the IGCP 410 meeting in Riverside (June 2001) for allowing us to meet and discuss with many colleagues, and in particular Dr. G.C. Young (Australian National University, Canberra). We bene¢ted from discussion on geochemistry with Dr. A. Munnecke (Erlangen University, Germany). We also o¡er this paper as a contribution to IGCP 440: Breakup of Rodinia, as a mark of respect and thanks to the Late Professor Chris Powell. A.B. thanks IGCP 410: The Great Ordovician Biodiversi¢cation Event, the French IGCP National Committee (PICG France), and the French National Committee of Geology (CNFG) for ¢nancial support. S.T. thanks the Australian IGCP Committee for ¢nancial support to attend IGCP 410 and 421 meetings in 2000^2001, and the CNRS for a three-month fellowship at USTL in late 2001. Our referees C.R. Barnes, D.K. Elliott, and T. Servais helped to improve our paper.

References Barnes, C.R., in press. Was there an Ordovician superplume? In: Webby, B.D., Droser, M.L., Paris, F., Percival, I.G. (Eds.), The Great Ordovician Biodiversi¢cation Event. Columbia Univ. Press, New York. Barnes, C.R., Fortey, R.A., Williams, S.H., 1996. The pattern of global bio-events during the Ordovician Period. In: Walliser, O.H. (Ed.), Global Events and Event Stratigraphy in

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54 the Phanerozoic. Springer, Berlin and Heidelberg, pp. 139^ 172. Benedetto, J.L., 1998. Early Palaeozoic brachiopods and associated shelly faunas from western Gondwana: their bearing on the geodynamic history of the pre-Andean margin. In: Pankhurst, R.J., Rapela, C.W. (Eds.), The Proto-Andean Margin of Gondwana. Geol. Soc. London Spec. Publ. 142, 57^83. Benedetto, J.L., Sanchez, T.M., Carrera, M.G., Brussa, E.D., Salas, M.J., 1999. Palaeontological constraints on successive palaeogeographic positions of Precordillera terrane during the early Palaeozoic. In: Ramos, V.A., Keppie, J.D. (Eds.), Laurentia-Gondwana Connections Before Pangea. Geol. Soc. Am. Spec. Paper 336, 21^42. Berner, R.A., Petsch, S.T., Lake, J.A., Beerling, D.J., Popp, B.N., Lane, R.S., Laws, E.A., Westley, M.B., Cassar, N., Woodward, F.I., Quick, W.P., 2000. Isotope fractionation and atmospheric oxygen: implications for Phanerozoic O2 evolution. Science 287, 1630^1633. Blieck, A., Goujet, D., 1978. A propos de nouveau mate¤riel de The¤lodontes (Verte¤bre¤s Agnathes) d’Iran et de Tha|«lande: apercZu sur la re¤partition ge¤ographique et stratigraphique des Agnathes des ‘re¤gions gondwaniennes’ au Pale¤ozo|«que moyen. Ann. Soc. Ge¤ol. Nord 97, 363^372. Blieck, A., Janvier, P., 1991. Silurian vertebrates. In: Bassett, M.G., Lane, P.D., Edwards, D. (Eds.), The Murchison Symp. Palaeontology Spec. Paper 44, 345^389. Blieck, A., Janvier, P., 1999. Silurian^Devonian vertebrate dominated communities, with particular reference to agnathans. In: Boucot, A.J., Lawson, J.D. (Eds.), Palaeocommunities: A Case Study from the Silurian and Lower Devonian. Cambridge Univ. Press, Cambridge, pp. 79^105. Blieck, A., Elliott, D.K., Gagnier, P.-Y., 1991. Some questions concerning the phylogenetic relationships of heterostracans, Ordovician to Devonian jawless vertebrates. In: Chang, M.-m., Liu Y.-h., Zhang, G.-r. (Eds.), Early Vertebrates and Related Problems of Evolutionary Biology. Science Press, Beijing, pp. 1^17. Blieck, A., Turner, S., Young, G.C., 2001. Cambrian^Ordovician vertebrate biogeography. In: Alvaro, J.J., Servais, T. (Eds.), Early Palaeozoic Palaeogeographies and Biogeographies of Western Europe and North Africa, Villeneuve d’Ascq, France, Sept. 24^26, 2001. Poster and Abstract 10. Blieck, A., Karatajute-Talimaa, V.N., Mark-Kurik, E., 2002. Upper Silurian and Devonian heterostracan pteraspidomorphs (Vertebrata) from Severnaya Zemlya (Russia): a preliminary report with biogeographical and biostratigraphical implications. Geodiversitas 24, 805^820. Boucot, A.J., Janis, C., 1983. Environment of the Early Palaeozoic vertebrates. Palaeogeogr. Palaeoclimat. Palaeoecol. 41, 251^287. Boucot, A.J., Chen, X., Scotese, C.R., 1995. Ibexian and postIbexian palaeogeography based on climatically sensitive sediments and biogeographic data. In: Cooper, J.D., Droser, M.L., Finney, S.C. (Eds.), Ordovician Odyssey: Short Papers for the Seventh Int. Symp. on the Ordovician Sys-

51

tem, Las Vegas, NE, June 1995. Paci¢c Sect. Soc. Sediment. Geol. (SEPM), Fullerton, CA, pp. 291^295. Brenchley, P.J., Marshall, J.D., Carden, G.A.F., Robertson, D.B.R., Long, D.G.F., Meidla, T., Hints, L., Anderson, T.F., 1994. Bathymetric and isotopic evidence for a shortlived Late Ordovician glaciation in a greenhouse period. Geology 22, 295^298. Brenchley, P.J., Carden, G.A.F., Marshall, J.D., 1995. Environmental changes associated with the ‘¢rst strike’ of the late Ordovician mass extinction. Mod. Geol. 20, 69^82. Brock, G.A., Engelbretsen, M.J., Jago, J.B., Kruse, P.D., Laurie, J.R., Shergold, J.H., Shi, G.R., Sorauf, J.E., 2000. Palaeobiogeographic a⁄nities of Australian Cambrian faunas. In: Wright, A.J., Young, G.C., Talent, J.A., Laurie, J.R. (Eds.), Palaeobiogeography of Australasian Faunas and Floras. Assoc. Australas. Palaeont. Mem. 23, 1^61. Can¢eld, D.E., 1998. A new model for Proterozoic ocean chemistry. Nature 396, 450^453. Carroll, R.L., 1997. Patterns and Processes of Vertebrate Evolution. Cambridge Univ. Press, Cambridge. Carroll, R.L., 2000. Towards a new evolutionary synthesis. Tree 15, 27^32. Christiansen, J.L., Stouge, S., 1999. Oceanic circulation as an element in palaeogeographical reconstructions: the Arenig (early Ordovician) as an example. Terra Nova 11, 73^78. Cocks, L.R.M., 2001. Ordovician and Silurian global geography. J. Geol. Soc. Lond. 158, 197^210. Cook, P.J., Shergold, J.H. (Eds.), 1986. Phosphate Deposits of the World 1. Proterozoic and Cambrian Phosphorites. Cambridge University Press, Cambridge. Dalziel, I.W.D., 1997. Neoproterozoic-Palaeozoic geography and tectonics: review, hypothesis, environmental speculation. Geol. Soc. Am. Bull. 109, 16^42. Donoghue, P.C.J., Forey, P.L., Aldridge, R.J., 2000. Conodont a⁄nity and chordate phylogeny. Biol. Rev. 75, 191^ 251. Elliott, D.K., Blieck, A.R.M., Gagnier, P.-Y., 1991. Ordovician vertebrates. In: Barnes, C.R., Williams, S.H. (Eds.), Advances in Ordovician Geology. Geol. Surv. Can. Paper 90^9, 93^106. Erdtmann, B.-D., 1991. The post-Cadomian Early Palaeozoic tectonostratigraphy of Germany (Attempt at an analytical review). Ann. Soc. Ge¤ol. Belg. 114, 19^43. Erdtmann, B.-D., 1998. Neoproterozoic to Ordovician Baltica and Laurentia interactions with Gondwana: Palaeobiological and tectonostratigraphic constraints. In: Linnemann, U., Heuse, T., Fatka, O., Kraft, P., Brocke, R., Erdtmann, B.-D. (Eds.), Int. Geol. Conf.: Pre-Variscan Terrane Analysis of ‘Gondwanan Europe’, Dresden, Germany, 28 April^4 May, 1998. Schr. Staatl. Mus. Miner. Geol. Dresden 9, 128^ 129 [abstract]. Foote, M., Sepkoski, J.J., Jr., 1999. Absolute measures of the completeness of the fossil record. Nature 398, 415^417. Franke, W., 2000. The mid-European segment of the Variscides: tectonostratigraphic units, terrane boundaries and plate tectonic evolution. In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenic Processes: Quanti¢cation

PALAEO 3072 9-5-03

52

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

and Modelling in the Variscan Belt. Geol. Soc. London Spec. Publ. 179, 35^61. Gagnier, P.-Y., 1995. Ordovician vertebrates and agnathan phylogeny. In: Arsenault, M., Lelie'vre, H., Janvier, P. (Eds.), Etudes sur les Verte¤bre¤s infe¤rieurs, VII Symp. Int., Miguasha, Que¤bec, 9^22 Juin 1991. Bull. Mus. Natl. Hist. Nat., 4e se¤r., 17 C (1^4), 1^37. Gagnier, P.-Y., Blieck, A., 1992. On Sacabambaspis janvieri and the vertebrate diversity in Ordovician seas. In: MarkKurik, E. (Ed.), Fossil Fishes as Living Animals. Academia 1, Tallinn, pp. 9^20. Gagnier, P.-Y., Blieck, A.R.M., Rodrigo-S., G., 1986. First Ordovician vertebrate from South America. Geobios 19, 629^634. Gagnier, P.-Y., Blieck, A., Emig, C.C., Sempere, T., Vachard, D., Vanguestaine, M., 1996. New palaeontological and geological data on the Ordovician and Silurian of Bolivia. J. S. Am. Earth Sci. 9, 329^347. Golonka, J., Ross, M.I., Scotese, C.R., 1994. Phanerozoic palaeogeographic and palaeoclimatic modeling maps. In: Embry, A.F., Beauchamp, B., Glass, D.J. (Eds.), Pangea: Global Environments and Resources. Can. Soc. Petrol. Geol. Mem. 17, 1^47. Halstead, L.B., 1974. Vertebrate Hard Tissues. The Wykeham Science Series, London. Hedges, S.B., 2001. Molecular evidence for the early history of living vertebrates. In: Ahlberg, P.E. (Ed.), Major Events in Early Vertebrate Evolution ^ Palaeontology, Phylogeny, Genetics and Development. Syst. Assoc. Spec. Vol. Ser. 61, 119^134. Holser, W.T., Magaritz, M., Ripperdan, R.L., 1996. Global isotopic events. In: Walliser, O.H. (Ed.), Global Events and Event Stratigraphy in the Phanerozoic. Springer, Berlin and Heidelberg, pp. 63^88. Janvier, P., 1996a. Early Vertebrates. Oxford Science Publishers and Clarendon Press, Oxford. Janvier, P., 1996b. Palaeontological Association 1995 Annual Address ^ The dawn of the vertebrates: characters versus common ascent in the rise of current vertebrate phylogenies. Palaeontology 39, 259^287. Janvier, P., 1997. Les classi¢cations phyloge¤ne¤tiques des Verte¤bre¤s actuels et fossiles. Bull. Soc. Zool. Fr. 122, 341^354. Janvier, P., 1998. Les Verte¤bre¤s avant le Silurien. Geobios 30, 931^950. Janvier, P., 2001. Ostracoderms and the shaping of the gnathostome characters. In: Ahlberg, P.E. (Ed.), Major Events in Early Vertebrate Evolution ^ Palaeontology, Phylogeny, Genetics and Development. Syst. Assoc. Spec. Vol. Ser. 61, 172^186. Karatajute-Talimaa, V.N., 1978. Silurian and Devonian Thelodonts of the USSR and Spitsbergen. Mokslas, Vilnius (in Russian). Karatajute-Talimaa, V., 1998. Determination methods for the exoskeletal remains of early vertebrates. Mitteilungen aus dem Museum fu«r Naturkunde in Berlin, Geowissenschaftliche Reihe 1, 21^52. Lewandowski, M., 1993. Palaeomagnetism of the Palaeozoic

rocks of the Holy Cross Mountains (Central Poland) and the origin of the Variscan Orgen. Publ. Inst. Geophys. Pol. Acad. Sci. A 23, 1^85. Lewandowski, M., 1998. Assembly of Pangea: combined palaeomagnetic and palaeoclimatic approach. In: Ginter, M., Wilson, M.V.H. (Eds.), Circum-Arctic Palaeozoic Faunas and Facies. Ichthyolith Issues Spec. Publ. 4, 29^32. Li, Z.-x., Powell, C.Mc.A., 2001. An outline of the palaeogeographic evolution of the Australasian region since the beginning of the Neoproterozoic. Earth Sci. Rev. 53, 237^ 277. Li, Z.-x., Zhang, L., Powell, C.McA., 1996. Positions of the East Asian cratons in the Neoproterozoic supercontinent of Rodinia. Aust. J. Earth Sci. 43, 593^604. Lieberman, B.S., 2002. Phylogenetic biogeography with and without the fossil record: gauging the e¡ects of extinction and palaeontological incompleteness. Palaeogeogr. Palaeoclimat. Palaeoecol. 178, 39^52. Mac Niocaill, C., Van der Pluijm, B.A., Van der Voo, R., 1997. Ordovician palaeogeography and the evolution of the Iapetus Ocean. Geology 25, 159^162. McKerrow, W.S., Cocks, L.R.M., 1995. The use of biogeography in the terrane assembly of the Variscan Belt of Europe. In: Gee, D.G., Beckholmen, M. (Eds.), EUROPROBE and the Trans-European Suture Zone. Stud. Geophys. Geod. 39, 269^275. McKerrow, W.S., Mac Niocaill, C., Ahlberg, P.E., Clayton, G., Cleal, C.J., Eagar, R.M.C., 2000. The Late Palaeozoic relations between Gondwana and Laurussia. In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenic Processes: Quanti¢cation and Modelling in the Variscan Belt. Geol. Soc. London Spec. Publ. 179, 9^20. Metcalfe, I., 1999. Gondwana dispersion and Asian accretion: an overview. In: Metcalfe, I. (Ed.), Gondwana dispersion and Asian accretion. Balkema, Rotterdam and Brook¢eld, pp. 9^28. Morrow, J.R., Schindler, E., Walliser, O.H., 1996. Phanerozoic development of selected global environmental features. In: Walliser, O.H. (Ed.), Global Events and Event Stratigraphy in the Phanerozoic. Springer, Berlin and Heidelberg, pp. 53^ 61. Munnecke, A., Samtleben, C., Bickert, T., 2003. The Ireviken Event in the lower Silurian of Gotland, Sweden ^ relation to similar Palaeozoic and Proterozoic events. Palaeogeogr. Palaeoclimat. Palaeoecol. S0031-0182(03)00304-3 Neuman, R.B., 1984. Geology and palaeobiology of islands in the Ordovician Iapetus Ocean: review and implications. Geol. Soc. Am. Bull. 95, 1188^1201. Nitecki, M.H., Gutschick, R.C., Repetski, J.E., 1975. Phosphatic microfossils from the Ordovician of the United States. Fieldiana Geol. 35, 1^9. Nowlan, G.S., Neuman, R.B., 1995. Palaeontological contributions to Palaeozoic palaeogeographic and tectonic reconstructions. In: Williams, H. (Ed.), Geology of the Appalachian/Caledonian Orogen in Canada and Greenland. Geol. Surv. Can., Geology of Canada 6, 815^842. Oliver, G.J.H., Corfu, F., Krogh, T.E., 1993. U-Pb ages from

PALAEO 3072 9-5-03

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54 SW Poland: evidence for a Caledonian suture zone between Baltica and Gondwana. J. Geol. Soc. Lond. 150, 355^369. Paris, F., Robardet, M., 1990. Early Palaeozoic palaeobiogeography of the Variscan regions. In: Matte, P. (Ed.), Terranes in the Variscan Belt of Europe and Circum-Atlantic Palaeozoic Orogens. Tectonophysics 177 (1^3), 193^213. Pickett, J.W. (Convenor), Burrow, C.J., Holloway, D.J., Munson, T.J., Percival, I.G., Rickards, R.B., Sherwin, L., Simpson, A.J., Strusz, D.L., Turner S., Wright, A.J., 2000. Silurian palaeobiogeography of Australia. In: Wright, A.J., Young, G.C., Talent J.A., Laurie, J.R. (Eds.), Palaeobiogeography of Australasian Faunas and Floras. Assoc. Australas. Palaeont. Mem. 23, 127^165. Qing, H.-r., Barnes, C.R., Buhl, D., Veizer, J., 1998. The strontium isotopic composition of Ordovician and Silurian brachiopods and conodonts: relationships to geological events and implications for coeval seawater. Geochim. Cosmochim. Acta 62, 1721^1733. Repetski, J.E., 1980. Fossil. In: Mc Graw-Hill Yearbook of Science and Technology (1979), pp. 202^204. Rosen, D.E., 1974. Phylogeny and zoogeography of salmoniform ¢shes and relationships of Lepidogalaxias salamandroides. Bull. Am. Mus. Nat. Hist. 153, 265^325. Ross, J.R.P., Ross, C.A., 1992. Ordovician sea-level £uctuations. In: Webby, B.D., Laurie, J.R. (Eds.), Global Perspectives on Ordovician Geology. Balkema, Rotterdam and Brook¢eld, pp. 327^335. Sadler, P.M., Cooper, R.A., in press. Calibration. In: Webby, B.D., Droser, M.L., Paris, F., Percival, I.G. (Eds.), The Great Ordovician Biodiversi¢cation Event. Columbia University Press, New York. Sansom, I.J., Smith, M.M., Smith, M.P., 1996. Scales of thelodont and shark-like ¢shes from the Ordovician of Colorado. Nature 379, 628^630. Sansom, I.J., Smith, M.P., Smith, M.M., Turner, P., 1997. Astraspis ^ the anatomy and histology of an Ordovician ¢sh. Palaeontology 40, 625^643. Sansom, I.J., Smith, M.M., Smith, M.P., 2001. The Ordovician radiation of vertebrates. In: Ahlberg, P.E. (Ed.), Major Events in Early Vertebrate Evolution ^ Palaeontology, Phylogeny, Genetics and Development. Syst. Assoc. Spec. Vol. Ser. 61, 156^171. Scotese, C.R., 1986. Phanerozoic reconstructions: a new look at the assembly of Asia. Univ. Texas Inst. Geophys. Tech. Rep. 66, 1^54. Scotese, C.R., 1997. Palaeogeographic Atlas ^ PALAEOMAP Progress Report 90-0497. PALAEOMAP Project, Univ. Texas Arlington, 20 colour maps, 18 b and w maps, 18 slides, 21 pp. Scotese, C.R., McKerrow, W.S., 1990. Revised World maps and introduction. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, 1^21. Scotese, C.R., McKerrow, W.S., 1991. Ordovician plate tectonic reconstructions. In: Barnes, C.R., Williams, S.H. (Eds.), Advances in Ordovician Geology. Geol. Surv. Can. Paper 90-9, 271^282.

53

Scotese, C.R., Boucot, A.J., Chen X., Ruan, Y.-p., Peng, S.-c., 2001. Reconstructions of the Cambrian to Permian geography and climatic belts. In: Chen, X., Ruan, Y.-p., Boucot, A.J. (Eds.), Palaeozoic Climatic Evolution of China. Science Press, Beijing, pp. 125^143. (in Chinese). Servais, T., Fatka, O., 1997. Recognition of the Trans-European Suture Zone (TESZ) by the palaeobiogeographical distribution pattern of early to middle Ordovician acritarchs. Geol. Mag. 134, 617^625. Shu, D.-g., Luo, H.-l., Conway Morris, S., Zhang, X.-l., Hu, S.-x., Chen, L., Han, J., Zhu, M., Li, Y., Chen, L.-z., 1999. Lower Cambrian vertebrates from South China. Nature 402, 42^46. Ł volution. Masson, Simpson, G.G., 1969. La ge¤ographie de l’E Paris. Smith, M.P., Sansom, I.J., Repetski, J.E., 1996. Histology of the ¢rst ¢sh. Nature 380, 702^704. Smith, M.P., Sansom, I.J., Cochrane, K.D., 2001. The Cambrian origin of vertebrates. In: Ahlberg, P.E. (Ed.), Major Events in Early Vertebrate Evolution ^ Palaeontology, Phylogeny, Genetics and Development. Syst. Assoc. Spec. Vol. Ser. 61, 67^84. Smith, M.P., Donoghue, P.C.J., Sansom, I.J., 2002. The spatial and temporal diversi¢cation of Early Palaeozoic vertebrates. In: Crame, J.A., Owen, A.W. (Eds.), Palaeobiogeography and Biodiversity Change: The Ordovician and Mesozoic^Cenozoic Radiations. Geol. Soc. London Spec. Publ. 194, 69^83. Spjeldnaes, N., 1979. The palaeoecology of the Ordovician Harding Sandstone (Colorado, USA). Palaeogeogr. Palaeoclimat. Palaeoecol. 26, 317^347. Tait, J.A., Bachtadse, V., Franke, W., So¡el, H.C., 1997. Geodynamic evolution of the European Variscan fold belt: palaeomagnetic and geological constraints. Geol. Rundsch. 86, 585^598. Tait, J., Scha«tz, M., Bachtadse, V., So¡el, H., 2000. Palaeomagnetism and Palaeozoic palaeogeography of Gondwana and European terranes. In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenic Processes: Quanti¢cation and Modelling in the Variscan Belt. Geol. Soc. London Spec. Publ. 179, 21^34. Talent, J.A., 1985. Provinciality as a mean for quali¢ed resolution of separation of continental blocks in the past: preliminary exempli¢cation from Western Paci¢c borderlands. In: Simakov, K.V. (Ed.), Ecostratigraphy, Palaeobiogeography and Stratigraphical Boundaries. Akad. Nauk, Magadan, pp. 54^90 (in Russian). Talent, J.A., Gratsianova, R.T., Yolkin, E.A., 1987. Prototethys: fact or phantom? Palaeobiogeography in relation to the crustal mosaic for the Asia-Australia hemisphere in Devonian^Early Carboniferous times. In: Mc Kenzie, K.G. (Ed.), Shallow Tethys 2. Balkema, Rotterdam, pp. 87^111. Torsvik, T.H., Rehnstro«m, E.F., 2003. The Tornquist Sea and Baltica-Avalonia Docking. Tectonophysics 362, 67^82. Torsvik, T.H., Tait, J., Moralev, V.M., McKerrow, W.S., Sturt, B.A., Roberts, D., 1995. Ordovician palaeogeography

PALAEO 3072 9-5-03

54

A. Blieck, S. Turner / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 37^54

of Siberia and adjacent continents. J. Geol. Soc. London 152, 279^287. Trench, A., Torsvik, T.H., 1992. The closure of the Iapetus Ocean and Tornquist Sea: new palaeomagnetic constraints. J. Geol. Soc. London 149, 867^870. Turner, S., 1999. Early Silurian to Early Devonian thelodont assemblages and their possible ecological signi¢cance. In: Boucot, A.J., Lawson, J.D. (Eds.), Palaeocommunities: A Case Study from the Silurian and Lower Devonian. Cambridge University Press., Cambridge, pp. 42^78. Turner, S., Blieck, A., Nowlan, G.S., in press a. Vertebrates (agnathans and gnathostomes). In: Webby, B.D., Droser, M.L., Paris, F., Percival, I.G. (Eds.), The Great Ordovician Biodiversi¢cation Event. Columbia Univ. Press, New York. Turner, S., Blieck, A., Nowlan, G.S., in press b. Cambrian^ Ordovician vertebrate database. Ann. Soc. Ge¤ol. Nord. Vannier, J.M.C., Siveter, D.J., Schallreuter, R.E.L., 1989. The composition and palaeogeographical signi¢cance of the Ordovician ostracode faunas of Southern Britain, Baltoscandia, and Ibero-Armorica. Palaeontology 32, 163^222. Van Staal, C.R., Dewey, J.F., Mac Niocaill, C., McKerrow, W.S., 1998. The Cambrian^Silurian tectonic evolution of the northern Appalachians and British Caledonides: history of a complex, west and southwest Paci¢c-type segment of Iapetus. In: Blundell, D.J., Scott, A.C. (Eds.), Lyell ^ The Past is the Key to the Present. Geol. Soc. London Spec. Publ. 143, 199^242. Webby, B.D., Percival, I.G., Edgecombe, G.D., Cooper, R.A., Vandenberg, A.H.M., Pickett, J.W., Pojeta Jr., J., Playford, G., Winchester-Seeto, T., Young, G.C., Zhen, Y.-y., Nicoll, R.S., Ross, J.R.P., Schallreuter, R., 2000. Ordovician pa-

laeobiogeography of Australasia. In: Wright, A.J., Young, G.C., Talent, J.A., Laurie, J.R. (Eds.), Palaeobiogeography of Australasian Faunas and Floras. Assoc. Australas. Palaeont. Mem. 23, 63^126. Webby, B.D., Cooper, R.A., Bergstro«m, S.M., Paris, F., in press. Stratigraphic framework and time slices. In: Webby, B.D., Droser, M.L., Paris, F., Percival, I.G. (Eds.), The Great Ordovician Biodiversi¢cation Event. Columbia University Press, New York. Wilde, P., 1991. Oceanography in the Ordovician. In: Barnes, C.R., Williams, S.H. (Eds.), Advances in Ordovician Geology. Geol. Surv. Can. Paper 90-9, 283^298. Young, G.C., 1997. Ordovician microvertebrate remains from the Amadeus Basin, Central Australia. J. Vert. Palaeont. 17, 1^25. Young, G.C., in press. North Gondwana mid-Palaeozoic connections with Euramerica and Asia ^ Devonian vertebrate evidence. In: Mid-Palaeozoic Bio- and Geodynamics: The North Gondwana ^ Laurussia Interaction, Proc. 15th Int. Senckenberg Conf., Frankfurt, 11^21 May 2001. Cour. Forsch.-Inst. Senckenberg. Young, G.C., Karatajute-Talimaa, V.N., Smith, M.M., 1996. A possible Late Cambrian vertebrate from Australia. Nature 383, 810^812. Young, G.C., Long, J., Burrow, C., 2000. Vertebrata. In: Talent, J.A., Mawson, R. et al., Devonian palaeobiogeography of Australia and adjoining regions; in: Wright, A.J., Young, G.C., Talent, J.A., Laurie, J.R. (Eds.), Palaeobiogeography of Australasian Faunas and Floras. Assoc. Australas. Palaeont. Mem. 23, 209^219 and 250.

PALAEO 3072 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71 www.elsevier.com/locate/palaeo

Cambroclaves from the Cambrian of Sardinia (Italy) and Germany: constraints for the architecture of western Gondwana and the palaeogeographical and palaeoecological potential of cambroclaves Olaf Elicki  , Thomas Wotte Freiberg University, Geological Institute, Bernhard von Cotta Street 2, 09596 Freiberg, Germany Received 22 April 2002; received in revised form 22 October 2002; accepted 25 January 2003

Abstract Cambroclaves represent a group of problematic microfossils previously known from strata close to the Early/ Middle Cambrian boundary of only a few palaeogeographic regions (Kazakhstan, China, Australia). Because of their frequent occurrence as disarticulated remains, they have hitherto been assumed to be provincially restricted microfossils of unclear palaeobiological affinity. Discoveries of cambroclaves from the Early to early Middle Cambrian of southern (Sardinia) and central (Germany) Europe point to a much wider distribution during their short stratigraphic range, and imply closer palaeogeographic relations between the European shelf of western Gondwana and the areas from which cambroclaves were previously known. These relations are also supported by the common occurrence of other small shelly fossils. These facts support the existence of a widespread uniform facies belt (shelf) around parts of Gondwana during the Early^Middle Cambrian time interval, and contradict the interpretation of the European depositional areas as isolated basins or as distinctly separate Cambrian terranes. The western Gondwana cambroclaves occur in carbonate successions indicative of special palaeoecological conditions. The specimens are limited to distinct layers formed during transgressive phases that opened inner and partly restricted platform areas to open-marine and more distal (deeper subtidal) environments, possibly accompanied by a transition from a rather arid to more humid climatic conditions. Because of the short stratigraphic window of occurrence and of distinctive facies characteristics, cambroclaves are palaeoecologically and palaeobiogeographically useful, and consequently contribute important evidence for both the reconstruction of the Perigondwana realm and the relations to other palaeocontinents in the Cambrian. 7 2003 Elsevier Science B.V. All rights reserved. Keywords: Cambrian; Gondwana; cambroclaves; small shelly fossils; palaeobiogeography; palaeoecology

1. Introduction * Corresponding author. Tel.: +49-3731-39-2435; Fax: +49-3731-39-3599. E-mail address: [email protected] (O. Elicki).

Cambroclaves are minute sclerites with a platelike basal shield bearing a prominent elongate spine. The shield and the spine may have smooth

0031-0182 / 03 / $ ^ see front matter 7 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00302-X

PALAEO 3073 9-5-03

56

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

or ornamented surfaces (Qian, 1978; Mambetov and Repina, 1979; Conway Morris and Chen, 1991). The original mineralogy was probably calcareous, but they were phosphatised during diagenesis. Possibly the hollow shield was originally ¢lled with soft tissue. Cambroclaves show a wide morphological variability that has led to many form-taxa, to an extensive number of synonyms, and to taxonomic confusion (Bengtson et al., 1990; Conway Morris et al., 1997). The latter authors have distinguished four morphological groups of cambroclaves : zhijinitids and paracarinachitids (oval base and prominent spine), cambroclavids (dumb-bell shaped base with spine at the anterior end), deiradoclavids (nearly circular base, spine may form a transverse ridge), and deltaclavids (tear-drop shaped base, short spine at the expanded anterior end). The ¢rst two groups are the most common and widespread. Cambroclaves are mostly known from disarticulated remains. There are very few examples of some elements connected with each other (Kazakhstan, South Australia, China). These, and some particular morphological features, have contributed to several reconstructions of scleritomes (Mambetov and Repina, 1979; Bengtson et al., 1990; Conway Morris and Chen, 1991). Nevertheless, the shape of the complete animal is still unknown. Zhijinitids (with a more circular shield) may have been positioned on the animal’s outer surface and separated from each other by soft tissue, or perhaps they were located in small interspaces within cambroclavid scleritomes. All the other cambroclave groups probably covered the animal totally (interconnected arrays or sheets ; for discussion see Bengtson et al., 1990; Conway Morris and Chen, 1991). Most authors agree in interpreting a protective function for the sclerites against predators or physical abrasion. Some special morphologies (e.g. curved spines) may also point to a grasping behaviour. The cambroclave-bearing animal is regarded as bilaterally symmetrical ^ maybe slug-like ^ with an unclear systematic position. Tentative interpretations range from protoconodont-related (Mambetov and Repina, 1979) to endoparasitic worms (Acanthocephala according to Qian and Yin,

1984) or to animals with a⁄nity to priapulid worms or to aschelmiths (Conway Morris et al., 1997). Dzik (1994) compared such remains with receptaculids (usually interpreted as algae, but seen as sponge-related by that author). A more reliable systematic decision will not be possible until discovery of articulated specimens. Cambroclaves are typical Cambrian, but otherwise rather unusual microfossils. They were ¢rst reported nearly simultaneously from South China and Kazakhstan (Qian, 1978; Mambetov and Repina, 1979), and they were subsequently found in North China (Tarim), South Australia, and Germany (for a summary of previously known occurences see Conway Morris et al., 1997; Fig. 2).

2. General characteristics of the Early/Middle Cambrian transition interval on the European shelf During the Cambrian period several major continents (Gondwana, Laurentia, Siberia, Baltica) and several minor landmasses existed (Brasier, 1992; Kirschvink, 1992; McKerrow et al., 1992; Torsvik et al., 1996; Seslavinsky and Maidanskaya, 2001; Smith, 2001). Many sections representing this time interval (especially those of Perigondwana origin) occur today on terranes (e.g. Matte et al., 1990; Erdtmann, 1991; Narebski, 1994; Oczlon, 1994; Linnemann, 1995; Tolluoglu and Su«mer, 1995; Tait et al., 1997; Debrenne et al., 1999; Zulauf et al., 1999; Belka et al., 2000; Cocks, 2000; Linnemann et al., 2000; Demange, 2001). During Early Cambrian time, the European shelf of the western Gondwana margin was marked by a generally transgressive trend accompanied by a warming of the climate to arid^subŁ lvaro and Vennin, 1998; A Ł ltropical conditions (A varo et al., 2000a,b,c). Thus, middle, southern and southwestern Europe are characterised by transition of the marine environments from siliciclastic (deeper and shallow shelf areas) to carbonate (archaeocyath-bearing shallow ramps and shelfs) under equatorial/subequatorial conditions (Debrenne, 1964; Perejo¤n, 1986; Gandin, 1987; BechŁ lvaro et al., sta«dt and Boni, 1994; Elicki, 1999; A

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

57

Fig. 1. Location map of the working areas in Sardinia (Italy) and Germany.

2000a,b; Ferna¤ndez-Sua¤rez et al., 2000; Seslavinsky and Maidanskaya, 2001). In contrast, Middle Cambrian sediments of the European shelf are predominantly siliciclastic. Only at the beginning some carbonates were deposited on deepened shelfs and ramps (Gandin, 1979; Courjault-Rade¤, 1990; Lin‹a¤n and Quesada, Ł lvaro et al., 2000b; A Ł l1990; Loi et al., 1995; A varo et al., 2001; Elicki, 2001). The reasons for these dramatic changes were complex (e.g. diachronous tectonics of di¡erent local intensity, changing climate, sea-level rise, palaeogeographic movements) and are not completely understood. This general sedimentological evolution during the Early/Middle Cambrian transition produced a very uniform stratigraphic pattern over most of the Mediterranean region, including the Cambrian deposits in Spain, France, Sardinia (Italy), and Germany, but also shows similarity to Morocco, Turkey and the Middle East (Haude, 1969; º nalan, 1986; Courjault-Rade¤ et Sdzuy, 1972; O Ł al., 1991; Alvaro et al., 1993; Lin‹a¤n and Ga¤mez-Vintaned, 1993; Pillola et al., 1994; Geyer and Landing, 1995; Elicki, 1997). Faunal similarŁ lvaro et al., ities between these areas (see also A 2003, this issue) are further support for such palaeogeographic relations. Many of these deposits

in Europe are incorporated in tectonostratigraphic terranes (e.g. the German successions). The incompleteness of the sections, their isolation and poor preservation have led to many problems in correlation and in palaeogeographic and tectonic reconstruction. In recent years an improved correlation of the European deposits has been worked out. In addition to intensive biostratigraphic work, the detection of sea-level-related regional and global events (e.g. Daroca event, Valdemiedes event) as well as provenance data from detrital minerals have helped to clarify the structural evolution of the Ł lvaro et al., 1993; Lin‹a¤n European shelf area (A and Ga¤mez-Vintaned, 1993; Pillola, 1993; Geyer Ł lvaro and Vizca|«no, 1997, and Elicki, 1995; A 1998; Chang, 1998a; Vidal et al., 1999; Belka et al., 2000; Cocks, 2000)

3. The cambroclaves from the western Gondwana European shelf After having been reported for the ¢rst time in 1978, for a long time cambroclaves were known only from Asia. Following discoveries in southern Australia (Bengtson et al., 1990), cambroclaves

PALAEO 3073 9-5-03

58

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

Fig. 2. Palaeogeographic reconstruction of Gondwana during the latest Early Cambrian (modi¢ed after McKerrow et al., 1992; Courjault-Rade¤ et al., 1992; Seslavinsky and Maidanskaya, 2001). Localities of cambroclaves are indicated by black triangles.

were ¢rst reported from western Perigondwana by Elicki and Schneider (1992). Later these fossils ^ on the basis of more specimens ^ were revised and described as Cambroclavus ludwigsdorfensis (Elicki, 1994). The most recent ¢ndings come from southwestern Sardinia (Elicki, 1998^2000). 3.1. Cambroclaves from southwestern Sardinia (Italy) The Cambrian succession of southwestern Sardinia (Fig. 1) has been investigated palaeontologically (trilobites, archaeocyaths) and sedimen-

tologically over many years, but especially extensively during the last decades (for results and summaries of research, see Bechsta«dt et al., 1985; Bechsta«dt et al., 1988; Pillola, 1991; Bechsta«dt and Boni, 1994; Loi et al., 1995; Pillola et al., 1995). The succession (Fig. 3) represents an evolution from a clastic tidal shelf via a mixed shelf stage (both Nebida Group, Ovetian to Marianian) to an isolated carbonate platform (Gonnesa Group, upper Marianian to Bilbilian) which was subsequently disrupted and £ooded (basal Iglesias Group, uppermost Bilbilian to Caesaraugustian) and then covered by clastic shelf deposits

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

59

Fig. 3. Generalised geological section of the Cambrian succession of SW Sardinia (Italy) and sketch of the bathymetric history. C.P. = Campo Pisano Formation. Whether a ‘Bithia Fm.’ is really existing or only a synonym for a local equivalent of the lower Matoppa Fm. is under discussion. The position of cambroclaves is indicated.

(upper Iglesias Group, middle Caesaraugustian up to Lower Ordovician; Bechsta«dt et al., 1988; Pillola et al., 1995). Despite the extensive knowledge of the Sardinian Cambrian, the drowning stage (basal Iglesias Group, Campo Pisano Formation, condensed nodular limestones) has until recently not been investigated in detail. Sedimentologic and facies data have been published by Gandin (1987), Cocozza and Gandin (1990) and Elicki (2001), while some palaeontologic work comes from Cherchi and Schroeder (1984), Mostler (1985), Pillola (1991) and Elicki (1998^2000). Following Pillola et al. (1995), the Campo Pisano Formation starts stratigraphically (trilobites) within the highest Lower Cambrian (upper Bilbilian) and extends

to the mid-Middle Cambrian (early/middle Caesaraugustian). The Sardinian cambroclaves were found in the Su Corovau section, which is situated about 6 km NE of the town Iglesias. They occur in a relatively thin (0.15 m) and distinct interval, about 1.8 m above the base of the Campo Pisano Formation. About 40 cambroclavid morphs have been identi¢ed (Fig. 7; special publication in progress). Whereas the biofacies of the directly underlying carbonate succession (basal part of the Campo Pisano Formation) is characterised by a predominance of poriferids (a phenomenon often recognised in Campo Pisano sections), the cambroclave-bearing level shows a more diverse fossil content that includes disarticulated echinoderm

PALAEO 3073 9-5-03

60

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

Fig. 4. Geological section of the Cambrian succession of the Go«rlitz Syncline (Germany) and sketch of the bathymetric history. The position of cambroclaves is indicated. For legend see Fig. 3.

remains, lingulate brachiopods, small trilobites, archaeogastropods (Pelagiella), and rare hyoliths (among others Microcornus) and chancelloriids. The sponge remains are nearly completely absent from this level upward for the rest of the section. After investigation of nearly 20 sections of the Campo Pisano Formation it seems that there is a more or less regular vertical sedimentological and biofacies pattern that re£ects the evolution of the Campo Pisano environment : thus, the sometimes nodular limestones of the lower Campo Pisano Formation follow the shallow isolated platform stage of the Gonnesa Group (Fig. 3). This transition represents a brief subsidence pulse (local breccias, abrupt changing of the biocoenoses, sudden onset of terrigenous in£ux) below the carbonate production optimum zone (probably middle subtidal). The fauna here is rather monotonous

(mainly sessile^benthic ¢lter-feeding poriferids). After that, the biofacies succession indicates a slight deepening to a greater subtidal depth which is the primary depositional regime for most of the Campo Pisano Formation. The fauna of this facies is typically open marine. Only toward the top of the formation can a second important subsidence impulse be observed (biota smaller and transported, decrease of carbonate content, alternation of carbonate and shaly layers), which led to a drop of the platform into deeper water. At this point, deposition became completely siliciclastic (Cabitza Formation; Fig. 3). The Sardinian cambroclaves occur only in the transition zone from the very shallow to the presumed middle subtidal, and from a rather restricted to an open marine and terrigenous-in£uenced environment (more complex shelly-fossil

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

61

Fig. 5. Geological column of the Cambrian succession of the Doberlug Syncline (Leipzig area, Germany) and sketch of the bathymetric history. The position of cambroclaves is indicated. For legend see Fig. 3.

biocoeonoses; distinct decrease of passive ¢lter feeders : no archaeocyaths, loss of the other poriferids ; no photosynthetic organisms). This dramatic reorganisation of the depositional area may have had di¡erent causes: Bechsta«dt et al. (1985), Gandin (1987), Bechsta«dt and Boni Ł lvaro and Vennin (1996) have sug(1989), and A gested tensional tectonics. However, when other sequences of this time interval in western Gondwana are taken into account, a general rise in relative sea level must also be considered. These processes may have been accompanied by general climate changes. Thus, arid conditions are assumed for the isolated platform stage of the Gonnesa Group, whereas more humid conditions may

be assumed for near the end of the group (Gandin, 1987; but see Bechsta«dt and Boni, 1989). A further source of change is the movement of this palaeogeographic area to higher southern latitudes (e.g. Courjault-Rade¤ et al., 1992; Kirschvink, 1992; Seslavinsky and Maidanskaya, 2001). Therefore, the strong sedimentologic and palaeoecologic changes that are expressed by the beginning of the Campo Pisano Formation may have been caused by a complex of tectonic, climatic, palaeogeographic, and sea-level events. The biostratigraphic position of the Sardinian cambroclaves can be constrained by correlation to the Campo Pisano type section (about 10 km away). From here Pillola (1991) described Proto-

PALAEO 3073 9-5-03

62

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

lenus (Protolenus) sp. ^ later identi¢ed as P. cf. pisidianus (compare Loi et al., 1995) ^ in the lower part of the section. According to these authors, the occurrence of this taxon (where Paradoxides is missing ^ as in Sardinia) is indicative of the Early/ Middle Cambrian boundary in the Mediterranean area. Because of the poor exposures, the basal spiculite facies of the Campo Pisano Formation (occurring in many sections) could not be positively identi¢ed in the type section. But there, immediately above the P. cf. pisidianus-bearing layer, the distinctive echinoderm-rich biofacies, which normally directly overlies the spiculite facies, starts. Thus, it seems that the transition from spiculite facies to echinoderm facies (as an ecostratigraphic boundary) in this local area corresponds with the biostratigraphic Early/Middle Cambrian boundary. Consequently, although P. cf. pisidianus has not yet been observed in the Su Corovau section, the base of the horizon containing the cambroclaves (which represents the ¢rst echinoderm-rich layer) can be roughly correlated with the Early/Middle Cambrian boundary interval. Therefore, the stratigraphic position of the Sardinian cambroclaves is probably lowest Middle Cambrian (early Leonian; Fig. 6).

3.2. Cambroclaves from eastern Germany Cambrian successions are very rare in Germany. A compilation of the occurrence, faunal content, and biostratigraphy is given in Elicki (1997). There is one short report on cambroclaves from two German regions (Fig. 1): (1) the Go«rlitz Syncline area (near the German^Polish boundary) and (2) the Leipzig area ( = Doberlug Syncline; Elicki, 1994). Both regions belong to the Mediterranean facies realm of the Cambrian European shelf of western Gondwana (Sdzuy, 1972) and represent the most northerly sites of this realm (Fig. 2). 3.2.1. Go«rlitz Syncline The Go«rlitz section (Lower Cambrian) is characterised by a succession of shallow-water massive dolostones and overlying bedded (and near the top partly nodular) limestones (Ludwigsdorf Member). This carbonate succession is overlain by a suite of claystone and siltstone with some sandstone intercalations (Lusatiops Member). The thickness of the whole section is about 250^ 300 m (Fig. 4). Due to the complex tectonic sit-

Fig. 6. Sketch of the stratigraphy of the cambroclave-bearing formations and members using the Iberian scale. For comparison, the Siberian scale for the Early Cambrian and the Baltic scale for the Middle Cambrian are also given. The position of cambroclaves is indicated.

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

uation, the stratigraphic continuations above and below are unknown. The poorly exposed outcrops have been studied for nearly 100 years ; the most recent sedimentologic, facies, palaeontologic and stratigraphic work comes from Elicki and Schneider (1992), Elicki (1994, 1998, 2000) and Geyer and Elicki (1995). A rich microfauna comes from the higher part of the limestone succession. The cambroclaves (Cambroclavus ludwigsdorfensis Elicki, 1994 ^ about 100 cambroclave morphs were recovered ; Fig. 7) are restricted here to the horizon with the richest biodiversity. The accompanying fauna consists of echinoderms, hyoliths and chancelloriids, poriferids, trilobites (e.g. Calodiscus cf. lobatus (Hall, 1847), Ferralsia saxonica (Geyer and Elicki, 1995), Lusatiops sp.), molluscs (pelagiellids, pelecypods, monoplacophorans) and problematica (e.g. Coleoloides typicalis (Walcott, 1890); Microcoryne cephalata (Bengtson, 1990); Aetholicopalla adnata (Conway Morris, 1990); Halkieria sp.). The litho- and ecostratigraphic level of occurrence of the cambroclaves is remarkable (Fig. 4). They are limited to a layer on top of which the transition from the shallow and open lagoonal carbonate environment to the siliciclastic outershelf conditions starts (for details see Elicki, 2000). The sedimentation is characterised by an increase in siliciclastic in£ux. Small channels ¢lled by phosphatic pebbles, mud pebbles and accumulations of fossil hard-parts may indicate the beginning of the deepening trend. In the immediately overlying stratum the siliciclastic content is much higher, the carbonates decrease, the texture of the carbonates is rather nodular and for the ¢rst time Rhombocorniculum cancellatum (Cobbold, 1921) occurs. This transition from carbonate to siliciclastic deposition is gradual. The most likely reason for this change is the general transgressive trend of this time interval. The fauna of the cambroclave-bearing horizon is dominated by suspension feeders and ¢lter feeders (hyoliths, chancelloriids, echinoderms, poriferids, brachiopods, helcionellids). Deposit feeders also occur (pelecypods, trilobites). Grazers and predators (e.g. early gastropods, halkieriids) were rather rare. Archaeocyaths (non-spicular

63

sponges typical of many Early Cambrian carbonate deposits) are absent. The favoured mode of life within this rather eutrophic environment was semi-infaunal (most of the hyoliths, some helcionellids), but includes also infaunal (pelecypods) and sessile and mobile epifaunal elements (poriferids, echinoderms, trilobites, brachiopods, ‘small shelly fossils’). The depositional area was a muddy carbonate soft ground. Fossils for which a nektonic behaviour can be assumed (Rhombocorniculum) occur ¢rstly directly above the cambroclaves and indicate (together with other faunal and sedimentological data) a deepening trend. The biostratigraphic position of these cambroclaves is determined by the co-occurrence of Ferralsia saxonica (Geyer and Elicki, 1995), which Ł lindicates a middle to upper Marianian age (A varo et al., 1998) and the onset of the ‘small shelly fossil’ R. cancellatum (Cobbold, 1921) immediately above, which is correlative to the start of the upper Atdabanian R. cancellatum zone (Rozanov and Zhuravlev, 1992) and to the Chinese lower Qiongzhusian (Jiang, 1992). The trilobites of the overlying Lusatiops Member belong to the early Botoman (Bergeroniellus micmacciformis level), which in turn correlates to the Microcornus parvulus zone in terms of the small-shelly-based biostratigraphy (Rozanov and Zhuravlev, 1992). 3.2.2. Doberlug Syncline The subsurface Cambrian succession of the Doberlug Syncline (Leipzig area; Early and Middle Cambrian) is a¡ected by complex tectonics, which have often led to problems in correlation. The Early Cambrian suite (Zwethau Formation) is represented by archaeocyath-bearing shallow marine carbonates and siliciclastics with some thin volcanic intercalations (Fig. 5). The contacts with the underlying Neoproterozoic volcano-sedimentary succession (Buschmann et al., 1995) and probably also the overlying Middle Cambrian siliciclastic succession (Brause, 1970) are disconformable. The thickness of the whole Cambrian is estimated as 1500^2000 m (Lower Cambrian 700^?1000 m). More recent research (facies, palaeontology, stratigraphy, tectono-facies) has come from Sdzuy (1972), Elicki and Debrenne (1993), Buschmann

PALAEO 3073 9-5-03

64

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

et al. (1995), Elicki (1999) and Jonas et al. (2000). The cambroclaves (Cambroclavus sp.) have been observed in only a single level of one drilling core (no. 1614/79). They are very rare (only two poorly preserved specimens ^ one camboclavid and one zhijinitid morph were found), probably re£ecting the limited quantity of sample material (diameter of drilling core: 4 cm). The accompanying fauna consists of very frequent tintinnids (Tintinnoidella praecursa (Elicki, 1994)), some chancelloriids (Archiasterella pentactina (Sdzuy, 1969); A. hirundo (Bengtson, 1990); Allonia tripodophora (Dore¤ and Reid, 1965); A. tetrathallis (Jiang, 1982)), and very rare Halkieria sp. Neither archaeocyaths nor calcimicrobes (usually common in the Doberlug Syncline) were observed. The palaeoecological position of this section is di⁄cult to determine, partly because of the diagenetic overprint. However, regarding the lithologic evolution, and comparing the characteristics of this siliciclastic/carbonate mixed succession with neighbouring and better preserved drilling cores, an open marine, deeper environment (approaching the outer mixed ramp) can be assumed. The cambroclaves occur within bioclastic wackestones dominated by tintinnids (Elicki, 1994). The morphology of the tintinnids points to a £oating mode of life in an open marine and a not too strongly agitated environment. The biostratigraphic position of the cambroclaves (Fig. 6) can also be ¢xed by comparison with the neighbouring drilling cores, which are dated by archaeocyaths (Elicki and Debrenne, 1993) as early Ovetian (equivalent to lower to middle Atdabanian).

4. Discussion The biostratigraphic position of the cambroclaves in the di¡erent regions of the world is lo-

65

cally more or less well known. The major, and until now unresolved, problem is to correlate these local scales regionally and intercontinentally (for discussion see Palmer, 1998). Cambroclaves occur in Kazakhstan in upper Atdabanian strata (Maly Karatau, Talassky Karatau; Mambetov and Repina, 1979). In China they are known from the Tarim region (North China), from Central (Hubei) and South China, all from beds of comparable age (Bengtson et al., 1990; Jiang, 1992). The South Australian specimens have been reported from the upper Atdabanian Ajax and Parara Limestones (Bengtson et al., 1990). Slightly younger cambroclaves have been reported from Hainan Island (latest Early Cambrian; Jiang and Huang, 1986). As described above, the cambroclaves from Germany are lower Ovetian, equivalent to lower to middle Atdabanian (based on archaeocyaths, Doberlug Syncline) and middle to upper Marianian, equivalent to upper Atdabanian (Go«rlitz Syncline). In contrast, the specimens from Sardinia are stratigraphically lowest Middle Cambrian (see above). So it seems that there are di¡erent levels of occurrence of cambroclaves: a ¢rst middle Atdabanian level is documented by only one record (Doberlug); a second level is higher Atdabanian (Kazakhstan, North, Central and South China, South Australia, Go«rlitz area); a third level is slightly younger Lower Cambrian (Hainan), and the highest level is basal Middle Cambrian (Sardinia). Taking into account that the biostratigraphic age determinations have been done on the basis of di¡erent taxa and further taking into account the signi¢cant problems in intercontinental correlation as mentioned above, the different stratigraphic levels of cambroclaves may only be an artefact of the current state of knowledge. Nevertheless, the very distinct stratigraphic window is remarkable. All cambroclaves on all

Fig. 7. Some representative specimens of the cambroclaves from Sardinia (panels A^D) and Germany (panels E^H). Each scale bar is 100 Wm. (A^D) cambroclaves indet.; disarticulated remains from chemical preparation; Su Corovau section (SW Sardinia, Italy); lower Campo Pisano Formation; basal Leonian. (E^H) Cambroclavus ludwigsdorfensis (Elicki, 1994); note the hollow interior (e.g. panels E and H) and the star-like cross-section of the vertical spine (panel E). Go«rlitz Syncline (Germany); Upper Ludwigsdorf Member; upper Marianian.

PALAEO 3073 9-5-03

66

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

palaeocontinents occur strictly between middle Atdabanian and most basal Middle Cambrian, a time span of only a few million years. So, this microfossil group represents a good index taxon for this short time interval. The reconstruction of migration paths of cambroclaves is not yet possible. Jiang (1992) discussed a diachronous migration of this fossil group across central and south Asia, from Kazakhstan through NW China to South China and to Hainan Island. However, the biostratigraphic position of the specimens from Kazakhstan is not as old as assumed by Jiang (the occurrence with R. cancellatum in the R. cancellatum zone indicates a late Atdabanian age). Cambroclaves also seem to be very characteristic for a special palaeogeographic region. Their more or less synchronous appearance along the European and the Asian shelf of western Gondwana as well as in South Australia and South China (Fig. 2) points to palaeogeographical connections between these areas (see also Pillola, 1990; Chang, 1998b; Debrenne et al., 1999; Brock et al., 2000; Shergold et al., 2000). This is con¢rmed for several of these regions by other Cambrian fossils also, such as trilobites, archaeocyaths and small shelly fossils (Kerber, 1988 ; Brasier, in Cowie and Brasier, 1989; Pillola, 1991; Courjault-Rade¤ et al., 1992; Elicki, 1994; Debrenne et al., 1999; Brock et al., 2000; and others). It may also point to clearer relations of the Gondwana/Perigondwana areas to Kazakhstan (a rather problematic palaeogeographic area in the Cambrian), to the Acado^Baltic region (sensu Sdzuy, 1972), and to Brasier’s Palaeotethyan belt (sensu Brasier, in Cowie and Brasier, 1989). Together with the accompanying fauna, the regional distribution of the cambroclaves has relevance for tectono-sedimentary reconstructions. Based on lithological and geochemical investigations and on the present-day geographical distribution of the Cambrian deposits, the existence of isolated pull-apart basins for this time on the European shelf was suggested by Buschmann et al. (1995). However, the very wide regional distribution of the characteristic Cambrian shelly faunas (e.g. Brasier, in Cowie and Brasier, 1989;

Elicki, 1994) clearly indicates a £ourishing faunal exchange not only between these ‘basins’ but also on an intercontinental scale. Further consideration of the palaeontologic and also lithologic data suggests, rather, the existence of a more or less uniform facies belt (which included the whole Mediterranean) in the late Early to early Middle Cambrian, without distinctly separated basins. The Cambrian faunal characteristics of the German and Sardinian sections do not point to their deposition on terranes, although these areas were terranes later in Palaeozoic time (e.g. Narebski, 1994; Oczlon, 1994; Dallmeyer et al., 1995; Linnemann, 1995; Cymerman et al., 1997; Zulauf et al., 1999; Belka et al., 2000; Cocks, 2000; Linnemann et al., 2000). There are clear relations of the German Early Cambrian archaeocyaths to Morocco and Spain (Elicki and Debrenne, 1993), of the Sardinian archaeocyaths to Spain (but not to Germany because of their younger age) and of Early Cambrian trilobites to Spain and France (Geyer and Elicki, 1995; Pillola, 1991). Nevertheless, Middle Cambrian trilobites from Germany show di¡erences in some time slices. Thus, the German taxa are related to Scandinavian faunas at the beginning of the Middle Cambrian and show Mediterranean a⁄nity only after the middle Middle Cambrian (Sdzuy, 1957, 1966, 1972). The neighbouring Middle Cambrian trilobite fauna in Bohemia (which does not contain basal Middle Cambrian taxa) is consistent with this pattern and indicates that it is a distinct but related depositional area. Further, the small shelly faunas from the German Early Cambrian and from the Sardinian Early and Middle Cambrian (which are both restricted to carbonate environments) indicate clear relations to the Palaeotethyan Belt (sensu Brasier, in Cowie and Brasier, 1989). The seemingly convincing palaeogeographic a⁄nities of the small shelly fossils to regions both near and far may, however, be a further artefact, due to the comparatively poor level of knowledge of such fossils in the Mediterranean. The existence of a rather continuous, uniform and little-di¡erentiated Cambrian European shelf is strongly supported by the palaeontological data. The model of a terrane-assemblage, includ-

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

ing the Armorican terrane-collage [Saxo-Thuringia terrane, Brittany/Normandy terrane = Avalonia sensu stricto, Perunica terrane ; Havlic›ek et al., 1994; Linnemann and Schauer, 1999; Linnemann et al., 2000], southern France/Sardinia and Spain, however, cannot be surely ruled out. If so, these terranes must have been very well connected. All these di¡erent regions show distinct similarity in their general sedimentary, biofacies, and structural patterns but also have special regional characteristics. These special di¡erences, however, can be explained e.g. by distinct patterns in palaeooceanographic circulation (e.g. the relations of the Middle Cambrian faunas), by regional faunal endemisms (generally usual for the Cambrian), and by divergent palaeogeographic positions on the shelf. Thus, the palaeontological, biostratigraphic and lithologic data of the Mediterranean Cambrian areas imply that in the Early Cambrian the Normandy and Doberlug regions were situated in a distal position on this structurally uniform shelf, adjacent to the continent including the southern France/Sardinia/Go«rlitz areas (compare also Courjault-Rade¤ et al., 1991); Bohemia was still terrestrial (£uvial). This distribution pattern persisted more or less into the Middle Cambrian (but Bohemia became marine). During the Late Cambrian/Early Ordovician, exposure and erosion of large areas is assumed by several authors (Chlupa¤c›, 1993; Linnemann et al., 2000). This evolution was succeeded by separation into terranes (e.g. Saxo-Thuringia and Perunica) from the Gondwana European shelf later in the Ordovician (e.g. Matte et al., 1990; Narebski, 1994; Belka et al., 2000; Cocks, 2000; Linnemann et al., 2000; Demange, 2001). Palaeoecologically, cambroclaves are indicative of shallow, normal marine, eutrophic conditions. They occur in deeper subtidal environments. The specimens from western Gondwana (as well as from Go«rlitz, Doberlug and Sardinia) mark special palaeogeographic positions : outer ramp deposits during a transgressive phase (Figs. 3^5). Often they occur within horizons with a high biotic diversity (this phenomenon is also observed from South China). The duration of appearance during the deepening process is mostly short, which may suggest that the cambroclave animal

67

was specialised to well de¢ned palaeoecologic conditions and was unable to migrate into neighbouring habitats.

5. Conclusions (1) Cambroclaves are of still unknown systematic a⁄nity. Their regional distribution is now extended from the northern Gondwana area (Asian shelf and Australia) and Kazakhstan to synchronous sections of western Gondwana (Sardinia/Italy, Germany). (2) The appearance of this remarkable fossil group is restricted to a small biostratigraphic window (lower Ovetian to lowermost Leonian; equivalent respectively to lower to middle Atdabanian and lowermost Middle Cambrian). Both the oldest (Germany) and youngest (Sardinia) occurrences are on the western Gondwana European shelf. Because of the problems of precise intercontinental correlation for this time interval and because of the use of di¡erent groups for biostratigraphy in each region, a more detailed time resolution within this interval of occurrence cannot yet be given. However, cambroclaves seem to be a good stratigraphic index taxon for the late Early to early Middle Cambrian time interval. (3) The occurrence of this fossil group in Germany and Sardinia points to clear palaeogeographic relations to the Palaeotethyan belt, especially to the Mediterranean region, to China and Australia, and to Kazakhstan. The late Early to early Middle Cambrian palaeogeographic position of Kazakhstan was possibly closer to the western Gondwana margin than previously assumed. (4) Together with other palaeontological and lithological data, cambroclaves indicate the existence of a rather uniform and continuous facies belt over the whole Mediterranean and so contradict the model of isolated local depositional basins or signi¢cantly separated terranes during that time. The palaeontologic data also point to the existence of a little-di¡erentiated shelf or (at least) of an assemblage of very closely related terranes (Avalonia terrane-collage [consisting of Saxothuringia terrane, Brittany/Normandy terrane, Perunica terrane], Spain, southern France/Sardinia) on

PALAEO 3073 9-5-03

68

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

or adjacent to the western Gondwana European shelf. (5) Because of poor biostratigraphic resolution within the time of occurrence of the cambroclaves, the evolutionary migration path of this fossil group cannot yet be reconstructed. Earlier attempts to do this must be rejected because of revision of the biostratigraphical database. There is still no real evidence regarding the location of the centre of origin and distribution. (6) The palaeoecological position of this fossil group is strongly restricted. Cambroclaves were typical for eutrophic, open normal marine, and rather deeper subtidal carbonate environments without strong agitation. The western Gondwana specimens accompany short transgressive phases during which the depositional area deepened and was distinctly in£uenced by increased siliciclastic input. These processes were related to a continuous rise of relative sea level, perhaps due to local tectonics and/or climate change, or to movement of the Gondwana landmass to higher southern latitudes. Further research on Cambrian small shelly fossils in the comparatively poorly investigated Mediterranean will lead to a better understanding of palaeogeographic relations within the western Gondwana shelf areas and to other Cambrian palaeocontinents.

Acknowledgements We are grateful for the assistance of many enthusiastic students who have helped us in ¢eld and laboratory work. Many thanks for good and helpful discussions go to Thilo Bechsta«dt (Heidelberg, Germany), Bernd Buschmann (Freiberg, Germany), Francoise Debrenne (Paris, France), Anna Gandin (Siena, Italy), Alfredo Loi and Gian Luigi Pillola (both Cagliary, Sardinia, Italy), Jo«rg Schneider (Freiberg, Germany), to A.R. (Pete) Palmer (Boulder, CO, USA) for Ł lvaro his very kind linguistic help, and to Javier A (Lille, France) and Simon Conway Morris (Cambridge, UK) for helpful comments on the manuscript. The work was signi¢cantly supported by

the German Research Foundation (DFG Research Projects EL 144/5 and Schn 408/1).

References Ł lvaro, J.J., Vennin, E., 1996. Tectonic control on Cambrian A sedimentation in south-western Europe. Eclogae Geol. Helv. 89, 935^948. Ł lvaro, J.J., Vennin, E., 1998. Stratigraphic signature of a A terminal Early Cambrian regressive event in the Iberian Peninsula. Can. J. Earth Sci. 35, 402^411. Ł lvaro, J.J., Vizca|«no, D., 1997. Re¤vision des trilobites SoleA nopleuropsinae du Cambrien moyen de la Montagne Noire (France). Geobios 30, 541^561. Ł lvaro, J.J., Vizca|«no, D., 1998. Re¤vision biostratigraphique A du Cambrien moyen du versant me¤ridional de la Montagne Noire (Languedoc, France). Bull. Soc. Ge¤ol. Fr. 169, 233^ 242. Ł lvaro, J.J., Gozalo, R., Lin‹a¤n, E., Sdzuy, K., 1993. The paA laeogeography of the northern Iberia at the Lower-Middle Cambrian transition. Bull. Soc. Ge¤ol. Fr. 164, 843^850. Ł lvaro, J.J., Lin‹a¤n, E., Vizca|«no, D., 1998. Biostratigraphical A signi¢cance of the genus Ferralsia (Lower Cambrian, Trilobita). Geobios 31, 499^504. Ł lvaro, J.J., Rouchy, J.M., Bechsta«dt, T., Boucot, A., Boyer, A F., Debrenne, F., Moreno-Eiris, E., Perejo¤n, A., Vennin, E., 2000a. Evaporitic constraints on the southward drifting of the western Gondwana margin during Early Cambrian times. Palaeogeogr. Palaeoclimatol. Palaeoecol. 160, 105^ 122. Ł lvaro, J.J., Vennin, E., Moreno-Eiris, E., Perejo¤n, A., BechA sta«dt, T., 2000b. Sedimentary patterns across the LowerMiddle Cambrian transition in the Esla nappe (Cantabrian Mountains, northern Spain). Palaeogeogr. Palaeoclimatol. Palaeoecol. 137, 43^61. Ł lvaro, J.J., Vennin, E., Mun‹oz, A., Sa¤nchez-Valverde, B., A Ojeda, J.L., 2000c. Interplay of orbital forcing and tectonic pulses in the Cambrian Iberian platform, NE Spain. Int. J. Earth Sci. 89, 366^376. Ł lvaro, J.J., Debrenne, F., Vizca|«no, D., 2001. The Lower A Cambrian of the southern Montagne Noire. Ann. Soc. Ge¤ol. Nord 8, 201^204. Ł lvaro, J.J., Elicki, O., Geyer, G., Rushton, A.W.A., SherA gold, J.H., 2003. Palaeogeographical controls on the Cambrian trilobite immigration and evolutionary patterns reported in the western Gondwana margin. Palaeogeogr. Palaeoclimatol. Palaeoecol., S0031-0182(03)00300-6. Bechsta«dt, T., Boni, M., 1989. Tectonic control on the formation of a carbonate platform: the Cambrian of southwestern Sardinia. SEPM Spec. Publ. 44, 107^122. Bechsta«dt, T., Boni, M., 1994. Sedimentological, stratigraphical and ore deposits ¢eld guide of the autochthonous Cambro-Ordovician of southwestern Sardinia. Memoire descrittive della carta geologica d’Italia, vol. 48. Servizio Geologico Nazionale, 434 pp.

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71 Bechsta«dt, T., Boni, M., Selg, M., 1985. The Lower Cambrian of SW-Sardinia: From a clastic tidal shelf to an isolated carbonate platform. Facies 12, 113^140. Bechsta«dt, T., Schledding, T., Selg, M., 1988. Rise and fall of an isolated, instable carbonate platform: The Cambrian of south-western Sardinia. Geol. Rundsch. 77, 389^416. Belka, Z., Ahrendt, H., Franke, W., Wemmer, K., 2000. The Baltica-Gondwana suture in central Europe: evidence from K-Ar ages of detrital muscovites and biogeographical data. In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenic processes: quanti¢cation and modelling in the Variscan belt. Geol. Soc. London Spec. Publ. 179, 87^102. Bengtson, S., 1990. Spicules. In: Bengtson, S., Conway Morris, S., Cooper, B.J., Jell, P.A., Runnegar, B.N. (Eds.), Early Cambrian fossils from South Australia. Assoc. Aust. Palaeontol. Mem. 9, 24^36. Bengtson, S., Conway Morris, S., Cooper, B.J., Jell, P.A., Runnegar, B.N., 1990. Early Cambrian fossils from South Australia. Assoc. Aust. Palaeontol. Mem. 9, 364 pp. Brause, H., 1970. Ein neuer wichtiger Aufschluss im Kambrium von Doberlug-Kirchhain. Geologie 19, 1048^1065. Brasier, M.D., 1992. Paleoceanography and changes in the biological cycling of phosphorus across the PrecambrianCambrian boundary. In: Lipps, J.H., Signor, P.W. (Eds.), Origin and Early Evolution of the Metazoa. Plenum Press, New York, pp. 483^523. Brock, G.A., Engelbretsen, M.J., Jago, J.B., Kruse, P.D., Laurie, J.R., Shergold, J.H., Shi, G.R., Sorauf, J.E., 2000. Palaeogeographic a⁄nities of Australian Cambrian faunas. Assoc. Aust. Palaeontol. Mem. 23, 1^61. Buschmann, B., Linnemann, U., Schneider, J., Su«ss, T., 1995. Die cadomische Entwicklung im Untergrund der TorgauDoberluger Synklinale. Z. Geol. Wiss. 23, 729^749. Chang, W.T., 1998a. Cambrian correlation within the Perigondwana faunal realm. Rev. Esp. Paleontol., no. extr. Homenaje al Prof. Gozalo Vidal, 23^34. Chang, W.T., 1998b. Cambrian biogeography of the Perigondwana faunal realm. Rev. Esp. Paleont., no. extr. Homenaje al Prof. Gozalo Vidal, 35^49. Cherchi, A., Schroeder, R., 1984. Middle Cambrian foraminifera and other microfossils from SW Sardinia. Boll. Soc. Paleontol. Ital. 23, 149^160. Chlupa¤c›, I., 1993. Geology of the Barrandian. Verlag Waldemar Kramer, Frankfurt a.M., 163 pp. Cobbold, E.S., 1921. The Cambrian horizons of Comley and their Brachiopoda, Pteropoda, Gasteropoda etc. Quart. J. Geol. Soc. London 76, 325^387. Cocks, L.R.M., 2000. The Early Palaeozoic geography of Europe. J. Geol. Soc. London 157, 1^10. Cocozza, T., Gandin, A., 1990. Carbonate deposition during early rifting: the Cambrian of Sardinia and the Triassic-Jurassic of Tuscany, Italy. Spec. Publ. Int. Ass. Sed. 9, 9^37. Conway Morris, S., 1990. Halkieriids. In: Bengtson, S., Conway Morris, S., Cooper, B.J., P. A. Jell, P.A., Runnegar, B.N. (Eds.), Early Cambrian fossils from South Australia. Assoc. Aust. Palaeontol. Mem. 9, 69^103. Conway Morris, S., Chen, M., 1991. Cambroclaves and Para-

69

carinachitids, early skeletal problematica from the Lower Cambrian of south China. Palaeontology 34, 357^397. Conway Morris, S., Crampton, J.S., Xiao, B., Chapman, A.J., 1997. Lower Cambrian cambroclaves (incertae sedis) from Xinjiang, China, with comments on the morphological variability of sclerites. Palaeontology 40, 167^189. Courjault-Rade¤, P., 1990. La transition Cambrien infe¤rieur/ Cambrien moyen dans le versant sud de la Montagne Noire (Nappes de Pardailhan et du Minervois). Evolution tectonose¤dimentaire et contexte ge¤odynamique. C.R. Acad. Sci. 310, 1101^1107. Courjault-Rade¤, P., Gandin, A., Debrenne, F., Dore¤, F., 1991. Geodynamic control on Lower Cambrian sedimentation in Normandy, Montagne Noire (France) and Sardinia (Italy). In: Geologia del Basamento Italiano, Convegno in memoria di Tommaso Cocozza, Siena, pp. 145^147. Courjault-Rade¤, P., Debrenne, F., Gandin, A., 1992. Palaeogeographic and geodynamic evolution of the Gondwana continental margins during the Cambrian. Terra Nova 4, 657^667. Cowie, J.W., Brasier, M.D., 1989. The Precambrian-Cambrian Boundary. Clarendon Press, Oxford, 213 pp. Cymerman, Z., Piasecki, M.A., Seston, R., 1997. Terranes and terrane boundaries in the Sudetes, northeast Bohemian Massif. Geol. Mag. 134, 717^725. Dallmeyer, R.D., Franke, W., Weber, K. (Eds.), 1995. PrePermian Geology of Central and Western Europe. Springer, Berlin, 604 pp. Debrenne, F., 1964. Archaeocyatha. Contribution a¤ l’e¤tude des faunes cambriennes du Maroc, de Sardaigne et de France. Notes Me¤m. Serv. Ge¤ol. Maroc. 179, 1^371. Debrenne, F., Maidanskaya, I.D., Zhuravlev, A.Yu., 1999. Faunal migrations of archaeocyaths and early Cambrian plate tectonics. Bull. Soc. Ge¤ol. Fr. 170, 189^194. Demange, M., 2001. Tectonostratigraphic setting of the Minervois and Pardhailan nappes of the southern Montagne Noire. Ann. Soc. Ge¤ol. Nord 8, 191^200. Dore¤, F., Reid, R.E., 1965. Allonia tripodophora nov. gen, nov. sp., nouvelle e¤ponge du Cambrien infe¤rieur de Carteret (Manche). C.R. Som. Se¤ances Soc. Ge¤ol. Fr., 20^21. Dzik, J., 1994. Evolution of ‘small shelly assemblages’. Acta Palaeontol. Pol. 39, 247^313. Elicki, O., 1994. Lower Cambrian carbonates from eastern Germany: palaeontology, stratigraphy and palaeogeography. N. Jb. Geol. Pala«ontol. Abh. 191, 69^93. Elicki, O., 1997. Biostratigraphic data of the German Cambrian - recent state of knowledge. Freiberg. Forsch.heft. C466. Pala«ontol. Stratigr. Fazies 4, 155^165. Elicki, O., 1998. First report of Halkieria and enigmatic globular fossils from the Central European Marianian (Lower Cambrian, Go«rlitz Syncline, Germany). Rev. Esp. Paleontol., no. extr. Homenaje al Prof. Gonzalo Vidal, 51^64. Elicki, O., 1998^2000. Vergleichende Fazies und Pala«oo«kologie im Finalstadium einer kambrischen Plattformentwicklung die small shelly Fauna der Campo Pisano Fm. SW-Sardiniens. Unpublished research reports to the German Research Foundation.

PALAEO 3073 9-5-03

70

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71

Elicki, O., 1999. Palaeoecological signi¢cance of calcimicrobial communities during ramp evolution: an example from the Lower Cambrian of Germany. Facies 41, 27^40. Elicki, O., 2000. Die kambrische ‘Go«rlitz-Fauna’: Charakteristik und Bedeutung fu«r die stratigraphische und pala«ogeographische Rekonstruktion Mitteleuropas. Z. Geol. Wiss. 28, 11^32. Elicki, O., 2001. Fazies und Genese kambrischer Knollenkalke des no«rdlichen Iglesiente (Gutturu Pala, SW-Sardinien, Italien). Zbl. Geol. Pala«ontol., Teil I (1/2), 33^54. Elicki, O., Schneider, J., 1992. Lower Cambrian (Atdabanian/ Botomian) shallow-marine carbonates of the Go«rlitz Synclinorium (Saxony/Germany). Facies 26, 55^66. Elicki, O., Debrenne, F., 1993. The Archaeocyatha of Germany. Freiberg. Forsch.heft. C 450. Pala«ontol. Stratigr. Fazies 1, 3^40. Erdtmann, B.-D., 1991. The post-Cadomian early Palaeozoic tectonostratigraphy of Germany (attempt at an analytical review). Ann. Soc. Ge¤ol. Belg. 114, 19^43. Ferna¤ndez-Sua¤rez, J., Gutie¤rrez-Alonso, G., Jenner, G.A., Tubrett, M.N., 2000. New ideas on the Proterozoic-Early Palaeozoic evolution of NW Iberia: insights from U-Pb detrital zircon ages. Prec. Res. 102, 185^206. Gandin, A., 1979. Analisi sedimentologica preliminare della parte superiore della Formazione di Gonnesa e della parte inferiore della Formazione de Cabitza (Cambriano inferiore e medio) della Sardegna. Mem. Soc. Geol. Ital. 20, 395^404. Gandin, A., 1987. Depositional and paleogeographical evolution of the Cambrian of southwestern Sardinia. Newsl. 5 IGCP 7, 151^166. Geyer, G., Elicki, O., 1995. Lower Cambrian trilobites from the Go«rlitz Synclinorium (Germany) - review and new results. Pala«ontol. Z. 69, 87^119. Geyer, G., Landing, E., 1995. The Cambrian of the Moroccan Atlas region. Beringeria, Spec. Issue 2, 7^46. Hall, J., 1847. Natural history of New York. Paleontology: Volume 1. Containing descriptions of the organic remains of the lower division of the New-York System. C. van Benthuysen, Albany, NY, 338 pp. Haude, H., 1969. Das Alt-Pala«ozoikum-Pra«kambrium bis Silurium in der Tu«rkei. Zbl. Geol. Pala«ontol. I, 702^719. Havlic›ek, V., Vanek, J., Fatka, O., 1994. Perunica microcontinent in the Ordovician (its position within the mediterranean province, series division, bentic and pelagic associations). Sbor. Geol. Ve›d. 46, 23.56. Jiang, Z., 1992. The Lower Cambrian fossil record of China. In: Lipps, J.H., Signor, P.W. (Eds.), Origin and Early Evolution of the Metazoa. Plenum Press, New York, pp. 311^ 333. Jiang Z., 1982. In: Luo, H., Jiang, Z., Wu, X., Song, X., Ouyang, L. (Eds.), The Sinian-Cambrian Boundary in Eastern Yunnan, China. PR China Publishing House, Yunnan, 265 pp. Jiang, Z., Huang, Z., 1986. Middle Cambrian small shelly fauna in Yaxian County, Hainan Island. Geol. Rev. 32, 317^324. Jonas, P., Buschmann, B., Gaitzsch, B., 2000. Unterkam-

brischer und unterkarbonischer Vulkanismus der Doberlug Synklinale (NE Saxothuringische Zone). Z. Geol. Wiss. 28, 157^175. Kerber, M., 1988. Mikrofossilien aus unterkambrischen Gesteinen der Montagne Noire, Frankreich. Palaeontogr. A 202, 127^203. Kirschvink, J.L., 1992. A paleogeographic model for Vendian and Cambrian time. In: Schopf, J.W., Klein, C. (Eds.), The Proterozoic Biosphere - A Multidisciplinary Study. Cambridge University Press, pp. 569^581. Lin‹a¤n, E., Quesada, C., 1990. Rift-Phase. In: Dallmeyer, R.D., Mart|¤nez Garcia, E. (Eds.), Pre-Mesozoic Geology of Iberia. Springer, Berlin, pp. 259^266. Lin‹a¤n, E., Ga¤mez-Vintaned, A., 1993. Lower Cambrian palaeogeography of the Iberian Peninsula and its relations with some neighbouring European areas. Bull. Soc. Ge¤ol. Fr. 164, 831^842. Linnemann, U., 1995. The Neoproterozoic terranes of Saxony (Germany). Prec. Res. 73, 235^250. Linnemann, U., Schauer, M., 1999. Die Entstehung der Elbezone vor dem Hintergrund der cadomischen und variszischen Geschichte des Saxothuringischen Terranes - Konsequenzen aus einer abgedeckten geologischen Karte. Z. Geol. Wiss. 27, 529^561. Linnemann, U., Gehmlich, M., Tichomirowa, M., Buschmann, B., Nasdala, L., Jonas, P., Lu«tzner, H., Bombach, K., 2000. From Cadomian subduction to early Palaeozoic rifting: the evolution of Saxo-Thuringia at the margin of Gondwana in the light of single zircon geochronology and basin development (Central European Variscides, Germany). In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenetic processes: quanti¢cation and modelling in the Variscan belt. Geol. Soc. London Spec. Publ. 179, 131^153. Loi, A., Pillola, G.L., Leone, F., 1995. The Cambrian and early Ordovician of south-western Sardinia. In: Cherchi, A. (Ed.), ‘Sardinia 95’, 6th Paleobenthos International Symposium, guidebook. Rendiconti del Seminario della Facolta die Scienze dell’Universita di Cagliari, suppl. vol. 65, pp. 63^ 81. Mambetov, A.M., Repina, L.N., 1979. Nizhniy kembriy Talasskogo Alatau I ego korrelyatsiya s razrezami Malogo Karatau I Sibirskoy platformy. In: Zhuravleva, I.T., Meshkova, N.P. (Eds.), Biostratigra¢ya i Paleontologiya Nizhnogo Kembriya Sibiriy. Nauka, Novosibirsk, pp. 98^138 (in Russian). Matte, P., Maluski, H., Rajlich, P., Franke, W., 1990. Terrane boundaries in the Bohemian Massif: result of large-scale Variscan shearing. Tectonophysics 177, 151^170. McKerrow, W.S., Scotese, C.R., Brasier, M.D., 1992. Early Cambrian continental reconstruction. J. Geol. Soc. London 149, 599^606. Mostler, H., 1985. Neue heteractinide Spongien (Calcispongea) aus dem Unter- und Mittelkambrium Su«dwestsardiniens. Ber. Naturwiss.-med. Ver. Innsbruck 72, 7^32. Narebski, W., 1994. Lower to Upper Paleozoic tectonomagmatic evolution of NE part of the Bohemian Massif. Zbl. Geol. Pala«ontol. I, 961^972.

PALAEO 3073 9-5-03

O. Elicki, T. Wotte / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 55^71 Oczlon, M.S., 1994. North Gondwana origin for exotic Variscan rocks in the Rhenohercynian zone of Germany. Geol. Rundsch. 83, 20^31. º nalan, M., 1986. Amanos daglaridaki Alt Palezoyik co«kellerO inin co«kelme ortamlari ve bo«lgenin paleocogra¢k evrimi. Bull. Geol. Soc. Turkey 29, 49^63. Palmer, A.R., 1998. Why is intercontinental correlation within the Lower Cambrian so di⁄cult? Rev. Esp. Paleontol. no. Extr. Homenaje Prof. Gonzalo Vidal, 17^21. Perejo¤n, A., 1986. Biostratigraf|¤a de los arqueociatos en Espan‹a. Cuad. Geol. Ibe¤r. 9, 213^265. Pillola, G.L., 1990. Lithologie et trilobites du Cambrien infe¤rieur du SW de la Sardaigne (Italie): implication pale¤obioge¤ographiques. C.R. Acad. Sci. 310, 321^328. Pillola, G.L., 1991. Trilobites du Cambrien infe¤rieur du SW de la Sardaigne, Italy. Palaeontogr. Ital. 78, 1^174. Pillola, G.L., 1993. The Lower Cambrian trilobite Bigotina and allied genera. Palaeontology 36, 855^881. Pillola, G.L., Leone, F., Ga¤mez-Vintaned, A., Lin‹an, E., Dabard, M.P., Chauvel, J.-J., 1994. The Lower Cambrian ichnospecies Astropolichnus hispanicus: palaeoenvironmental and palaeogeographic signi¢cance. Boll. Soc. Ital. spec. vol. 2, 253^267. Pillola, G.L., Leone, F., Loi, A., 1995. The lower Cambrian Nebida Group of Sardinia. In: Cherchi, A. (Ed.), ‘Sardinia 95’, 6th Paleobenthos International Symposium, guidebook. Rendiconti del Seminario della Facolta die Scienze dell’Universita di Cagliari, suppl. vol. 65, 27^63. Qian, Y., 1978. The early Cambrian hyolithids in central and southwestern China and their stratigraphical signi¢cance (in Chinese with English abstract). Mem. Nanjing Inst. Geol. Palaeontol. 11, 1^38. Qian, Y., Yin, G., 1984. Zhijinitidae and its stratigraphical signi¢cance (in Chinese with English abstract). Acta Palaeontol. Sin. 22, 82^94. Rozanov, A.Yu., Zhuravlev, A.Yu., 1992. The Lower Cambrian fossil record of the Soviet Union. In: Lipps, J.H., Signor, P.W. (Eds.), Origin and Early Evolution of the Metazoa. Plenum Press, New York, pp. 205^309. Sdzuy, K., 1957. Alter und tiergeographische Stellung des Mittelkambriums von Doberlug. Geologie 6, 465^475. Sdzuy, K., 1966. Das Kambrium des Frankenwaldes. 2: Die

71

Bergleshof-Schichten und ihre Trilobiten-Fauna. Senckenb. Lethaea 47, 57^86. Sdzuy, K., 1969. Unter- und Mittelkambrische Porifera (Chancelloria und Hexactinellida). Pala«ontol. Z. 43, 115^147. Sdzuy, K., 1972. Das Kambrium der acadobaltischen Faunenprovinz - Gegenwa«rtiger Kenntnisstand und Probleme. Zbl. Geol. Pala«ontol. II, 1^91. Seslavinsky, K.B., Maidanskaya, I.D., 2001. Global facies distributions from Late Vendian to Mid-Ordovician. In: Zhuravlev, A.Yu., Riding, R. (Eds.), The Ecology of the Cambrian Radiation. Columbia University Press, pp. 47^68. Shergold, J.H., Feist, R., Vizca|«no, D., 2000. Early late Cambrian trilobites of Australo-Sinian aspect from the Montagne Noire, southern France. Palaeontology 43, 599^632. Smith, A.G., 2001. Paleomagnetically and tectonically based global maps for Vendian to Mid-Ordovician time. In: Zhuravlev, A.Yu., Riding, R. (Eds.), The Ecology of the Cambrian Radiation. Columbia University Press, pp. 11^46. Tait, J.A., Bachtadse, V., Franke, W., So¡el, H.C., 1997. Geodynamic evolution of the European Variscan fold belt: palaeomagnetic and geological constraints. Geol. Rundsch. 86, 585^598. º ., 1995. An evolutionary model º ., Su«mer, E.O Tolluoglu, A.U on Early Paleozoic of Anatolian microcontinent, northern margin of Gondwanaland. Geol. Bull. Turkey 38, 1^10. Torsvik, T.H., Smethurst, M.A., Meert, J.G., Van der Voo, R., McKerrow, W.S., Brasier, M.D., Sturt, B.A., Walderhaug, H.J., 1996. Continental break-up and collision in the Neoproterozoic and Palaeozoic - A tale of Baltica and Laurentia. Earth Sci. Rev. 40, 229^258. Vidal, G., Palacios, T., MoczydIowska, M., Gubanov, A., 1999. Age constraints from small shelly fossils on the early Cambrian terminal Cadomian phase in Iberia. Geol. Fo«ren. Stockholm Fo«rh. 121, 137^143. Walcott, C.D., 1890. Descriptive notes of new genera and species from the Lower Cambrian or Olenellus Zone of North America. U.S. Natl. Mus. Proc. 12 (1889) 33^46. Zulauf, G., Schitter, F., Riegler, G., Finger, F., 1999. Age constraints on the Cadomian evolution of the Tepla¤ Barrandian unit (Bohemian Massif) through electron microprobe dating of metamorphic monazite. Z. Dtsch. Geol. Ges. 150, 627^639.

PALAEO 3073 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97 www.elsevier.com/locate/palaeo

Palaeogeographical and palaeoecological aspects of the Cambro^Ordovician radiation of echinoderms in Gondwanan Africa and peri-Gondwanan Europe Bertrand Lefebvre a; , Oldrich Fatka b b

a UMR CNRS Bioge¤osciences, Universite¤ de Bourgogne, 6 bld. Gabriel, 21000 Dijon, France Charles University, Institute of Geology and Paleontology, Albertov, 128 43 Prague 2, Czech Republic

Received 10 April 2002; received in revised form 18 July 2002; accepted 15 January 2003

Abstract Ecology and tempo of the Lower Palaeozoic radiation of echinoderms are discussed in this paper based on comparison of the diversity patterns observed in Cambro^Ordovician faunas from Laurentia and the northern Gondwana margin. The Cambrian ‘agronomic revolution’ triggered a global radiation of echinoderms, with the progressive disappearance of biomat-related lifestyles, and the colonisation of new environments. Both in Laurentia and on the northern Gondwana margin, soft-substrate echinoderm assemblages related to cold and/or deep environments were dominated by blastozoans and stylophorans. These assemblages show a pattern of continuous diversification from the Middle Cambrian to the Middle Ordovician. The major radiation of crinoids in the Lower to Middle Ordovician of Laurentia is only a local diversification correlated with the development of hardgrounds in warm, shallow environments. No comparable major Ordovician radiation event occurred in the cooler environments of the northern Gondwana margin. 5 2003 Elsevier Science B.V. All rights reserved. Keywords: Palaeoecology; Cambrian; Ordovician; Echinoderms; Gondwana; Laurentia

1. Introduction Echinoderms have represented a signi¢cant component of marine assemblages for more than 500 Ma. The oldest indisputable echinoderm remains known are from the middle Lower Cambrian (Durham, 1971; Ubaghs, 1975; Derstler, 1981;

* Corresponding author. Tel.: +33-238-25-5277; Fax: +33-238-63-1234. E-mail address: [email protected] (B. Lefebvre).

Sprinkle, 1981), but some uppermost Neoproterozoic body fossils (Gehling, 1987) and phosphatised embryos (Xiao and Knoll, 2000; Chen et al., 2000) have been attributed to echinoderms, which suggests that the origin of the phylum most probably lies deeper in the Late Neoproterozoic (Mooi and David, 1998). Like most other major metazoan phyla, the Echinodermata underwent a rapid and important radiation in the Lower Palaeozoic, with the appearance of about twenty classes between the Lower Cambrian and the Middle Ordovician. A single class (Echinoidea) has appeared since this major radiation (Upper Ordovician),

0031-0182 / 03 / $ ^ see front matter 5 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00303-1

PALAEO 3074 9-5-03

74

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

but possible sister-groups of echinoids are present in the Lower Ordovician (ophiocistioids), and possibly the Middle Cambrian (holothurians). Most authors consider that the Palaeozoic diversi¢cation pattern of echinoderms agrees with the model of evolutionary faunas, as established by Sepkoski (1979, 1981, 1991), with the succession of two distinct faunal assemblages designated ‘Cambrian Evolutionary Fauna’ (CEF) and ‘Palaeozoic Evolutionary Fauna’ (PEF). These two faunas have been distinguished by two successive radiations: the Lower Cambrian and the Lower Ordovician, respectively (Sprinkle, 1981, 1992; Guensburg and Sprinkle, 1992, 2001a; Droser and Sheehan, 1997; Sumrall et al., 1997; Gil Cid and Dom|¤nguez-Alonso, 1999; Dornbos and Bottjer, 2000; Zhuravlev, 2001). Under this model, most echinoderm classes showing their peak diversity during the Cambrian are simply incorporated into the CEF (e.g. cinctans, ctenocystoids, eocrinoids, and helicoplacoids), whereas classes the diversity of which peaks later in the Palaeozoic are assigned to the PEF (e.g. crinoids, edrioasteroids, diploporans, and rhombiferans ; Sepkoski, 1981; Sprinkle, 1981, 1992; Sprinkle and Guensburg, 1995, 1997; Sumrall et al., 1997). However, this evolutionary pattern has been severely questioned by some authors who consider that the diversi¢cation of echinoderms approximates to a more or less continuous and exponential expansion from the Cambrian to the Lower Ordovician (Derstler, 1981; Paul and Smith, 1984; Smith, 1988; Smith and Jell, 1990). For these authors, the Ordovician onset of the Palaeozoic fauna represents an apparent radiation resulting from taxonomic biases, and from the poor record of Upper Cambrian non-trilobite fossils. Smith (1988) rightfully argues that identical taxonomic ranks (e.g. classes) have been attributed for a variety of arbitrary reasons to very disparate groups of echinoderms (e.g. paraphyletic Eocrinoidea, polyphyletic Rhombifera), and that the description of the evolution of echinoderm diversity should rely on monophyletic groups rather than on arbitrary classes. The evolutionary signi¢cance of the extinction of a monophyletic class (e.g. cinctans) is clearly distinct from the progressive ‘extinction’ of a paraphyletic

class (e.g. eocrinoids) giving rise to several more derived groups (e.g. diploporans, parablastoids, rhombiferans, and solutes). In summary, most previous studies on the early radiation of echinoderms have been hampered by three major problems: (1) the extremely fast initial diversi¢cation of the phylum, (2) the di⁄culty of comparing the frequently puzzling, asymmetrical morphologies of Palaeozoic forms (e.g. helicoplacoids and solutes) with those of their modern, pentameral relatives (e.g. echinoids and sea-stars) and of incorporating them into a common systematic hierarchy, and (3) the absence of a global, detailed scheme of skeletal homologies for the phylum that would enable the de¢nition of a solid phylogenetic framework (see Smith, 1984, 1988; Sumrall, 1997). Recently, the Extraxial^Axial Theory (EAT) of body-wall homologies for echinoderms has dramatically changed and in£uenced our conception of echinoderm anatomy (Mooi et al., 1994; Mooi and David, 1997, 1998; David and Mooi, 1998; David et al., 2000). Based on embryological and morphological grounds, the EAT o¡ers a key to the identi¢cation of homologous skeletal elements in the various echinoderm groups, both fossil and extant. Consequently, this model provides a new perspective on the opportunity to perform solid phylogenetic relationships of echinoderms, based on well-established homologies (Mooi and David, 1997, 1998, 2000; David and Mooi, 1998, 1999; David et al., 2000). The EAT identi¢es two major components in the body wall of all echinoderms: axial and extraxial regions. The axial region derives from the larval rudiment and is closely associated with the mouth and the water-vascular system. During the ontogeny, each new axial element is formed at the distal end of a growing ray, following an alternating pattern known as the ‘Ocular Plate Rule’ (OPR ; David and Mooi, 1998; Mooi and David, 1998). The extraxial region is inherited from the non-rudiment part of the larval body, and it comprises all skeletal elements that are not developed according to the OPR. The extraxial region can be subdivided into perforate extraxial and imperforate extraxial. The perforate extraxial region exhibits several openings such as epispires, the hydropore, gonopores, and the

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

periproct. The imperforate extraxial region is not pierced by any ori¢ce. It forms the lower surface of most early echinoderms (e.g. edrioasteroids), the stalks and stems or various echinoderm groups (e.g. crinoids, eocrinoids, and solutes), and it is virtually absent in several derived forms (e.g. asterozoans, echinoids, and stylophorans; David and Mooi, 1998; Mooi and David, 1998). Most studies devoted to the Cambro^Ordovician radiation of echinoderms have focused on the description of the wide array of morphologies exhibited by early representatives of the phylum (Ubaghs, 1975; Derstler, 1981; Sprinkle, 1981, 1992; Paul and Smith, 1984; Mooi and David, 1998), but little attention has been given to the possible role of extrinsic factors (e.g. global environmental changes) in the diversi¢cation of Early Palaeozoic echinoderms. The ecological aspect of the echinoderm radiation has been documented in some recent contributions, but all of them concern the abundant and diverse Laurentian faunas (Guensburg and Sprinkle, 1992, 2001a; Sprinkle and Guensburg, 1995, 1997; Sumrall et al., 1997). The aims of this paper are: (1) to survey the data available on Lower Cambrian to Middle Ordovician echinoderms from North Africa ( = Gondwana) and Southwestern Europe ( = peri-Gondwanan regions) and to discuss their mode of life, (2) to compare the evolutionary pattern and palaeoecology of echinoderms from the Gondwana and peri-Gondwanan regions with those of other areas (Laurentia), and (3) to discuss the possible impact of extrinsic factors on the radiation of echinoderms in the light of the two main rival scenarios (evolutionary fauna model vs. continuous diversi¢cation). Following the phylogenetic framework provided by the EAT, all Lower Cambrian to Middle Ordovician echinoderms from the northern Gondwana margin can be assigned to three main monophyletic groups: (1) edrioasteroids, (2) blastozoans or ‘brachiole-bearing echinoderms’ (cinctans, ctenocystoids, diploporans, eocrinoids, parablastoids, rhombiferans, and solutes), and (3) ‘arm-bearing echinoderms’ (asterozoans, crinoids, and stylophorans ; David and Mooi, 1998; Mooi and David, 1998; David et al., 2000).

75

2. Cambro^Ordovician echinoderms from African Gondwana and European peri-Gondwanan domains During Lower Palaeozoic times, the northern Gondwana margin (NGM) comprised several regions of recent North Africa (e.g. Morocco), Southwestern and Central Europe (e.g. Armorican massif, Bohemia, Germany, Montagne Noire, Sardinia, Shropshire, Spain, Wales), Near East (Turkey), and North America (e.g. Nova Scotia and New Brunswick). A low palaeolatitude (?30‡S to ?40‡S) of the NGM in the Lower to lower Middle Cambrian (Fig. 1) is strongly supported by the presence of evaporites, carbonate sedimentation, reefs, and high-diversity faunal assemblagŁ lvaro et al., 2000a; A Ł lvaro es in these regions (A and Vennin, 2001; Smith, 2001). Clastic-dominated sedimentation, low-diversity faunal assemblages, palaeomagnetic data, as well as the disappearence of evaporites, carbonates and reefs, all suggest a progressive and relatively briskly southward drift of the NGM towards higher palaeolatitudes from the uppermost Middle to the Upper Cambrian. Due to the supposed opening and widening of the Rheic Ocean, some parts of the NGM (e.g. Avalonia) began to split o¡ from Gondwana in the Early Ordovician (Fig. 1) and to drift northwards more quickly, compared to Gondwana, colliding with Baltica at the end of the Ordovician (Cocks et al., 1997; Prigmore et al., 1997; Smith, 2001). Avalonia included several regions of Western Europe (e.g. Ardennes, England, Southeast Ireland, and Wales) and North America (e.g. New Brunswick, New England, Nova Scotia, and part of Newfoundland). Similar but slower transfer across the Tornquist ocean has also been recently proposed for some other European peri-Gondwanan terranes. Detailed reconstruction of this region for the pre-Variscan time interval is still in progress (Crowley et al., 2000). 2.1. Lower Cambrian echinoderm faunas As elsewhere in the world, the fossil record of Early Cambrian echinoderms is extremely poor on the NGM and consists mainly of commonly abundant, but undeterminable isolated skeletal el-

PALAEO 3074 9-5-03

76

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Fig. 1. Palaeogeography of Gondwanan North Africa and peri-Gondwanan Europe in the Middle Cambrian (A^B) and in the Lower Ordovician (C^D). Numbers on the maps refer to the study areas listed in the table. The table expresses the time intervals considered for each study area.

ements, usually referred to as ‘eocrinoid plates’. Such accumulations of echinoderm fragments have been reported from the uppermost Lower Cambrian of Montagne Noire, Sardinia, Shropshire, and Spain (see Appendix). Echinoderm fragments are frequently associated with micro-

Ł lvaro and Vennin, bial matgrounds, as in Spain (A Ł lvaro et al., 1999, 1997, 2001), Montagne Noire (A 2001a), and Sardinia (Perejo¤n et al., 2000). This suggests that Late Proterozoic-like biomats could have provided primitive echinoderms a ¢rm substrate for attachment during this time interval.

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

77

Fig. 2. Composition of Middle Cambrian echinoderm faunas from several regions of the northern Gondwana margin, distinguishing the three main clades (E, edrioasteroids; B, blastozoans; A, arm-bearing echinoderms). Abbreviations: SP, Spain; MN, Montagne Noire; BO, Bohemia; GB, Great Britain; MO, Morocco; SA, Sardinia; GE, Germany).

Complete specimens of Lower Cambrian echinoderms are extremely rare. A single species, the primitive blastozoan Gogia (Alanisicystis) andalusiae, has been described in the Lower Cambrian of the NGM (Spain; Ubaghs and Vizca|«no, 1990). A second undescribed species of Gogia is also present in Spain (Gil Cid and Dom|¤nguez-Alonso, 1999). In the taphonomic classi¢cation proposed by Brett et al. (1997), most early echinoderms correspond to ‘type 1’ forms, which rapidly ‘disintegrate into individual ossicles’. The rapid postmortem disarticulation of the multiplated, poorly organised thecae of primitive echinoderms generates numerous, tiny and loosely articulated elements. Consequently, the poor knowledge of Lower Cambrian echinoderms on the NGM is more the result of taphonomic bias than scarcity of these animals in the original assemblages. 2.2. Middle Cambrian echinoderm faunas The fossil record of echinoderms on the NGM is much richer in the Middle Cambrian than in the Lower Cambrian, with abundant isolated skeletal elements, but also with numerous complete specimens documented from several regions. This situation probably results from a greater number of ‘taphonomic windows’ (echinoderm Lagersta«tten) in the Middle Cambrian, but also from the progressive organisation of extraxial elements into more resistant stems and calyces (i.e. evolution of a rigid marginal frame delimiting the theca in cinctans and ctenocystoids) combined

with higher frequency of more diverse and better adapted echinoderm assemblages. Abundant and diverse Middle Cambrian echinoderm faunas have been recorded on the NGM. They are largely dominated by blastozoans (cinctans, ctenocystoids, and eocrinoids), associated with rare edrioasteroids and stylophorans (Fig. 2; see Appendix). Eocrinoids were attached, epibenthic suspension feeders. The morphology of several uppermost Lower to lowermost Middle Cambrian eocrinoids suggests that they could attach to ¢rm, ‘gelatinous’ substrates resulting from microbial activity (Dornbos and Bottjer, 2000). Attachment was realised either by a poorly di¡erentiated stalk (e.g. Gogia), or directly by the lower surface of the theca (e.g. Lichenoides ; Parsley and Prokop, in press). During the Middle Cambrian, the progressive transformation of the primitive stalk into more organised stems, frequently with a terminal holdfast (e.g. Akadocrinus), allowed some eocrinoids not only to colonise ¢rm substrates, but also to extend onto soft substrates, where they could attach to algae (Turek, 1990) or to various skeletal fragments, such as trilobite exoskeletons (Fatka and Kordule, 1990) or shells. However, the absence of any anchorage organ at the distal end of the stem in some eocrinoids (e.g. the gogiid Acanthocystites or some lepidocystids) raises the possibility that these forms rested on soft substrates with the theca held upright above the sea£oor (Paul, 1968). The possession of a £attened theca and a reduced number of ambulacra (two)

PALAEO 3074 9-5-03

78

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

in cinctans and ctenocystoids strongly suggest that these blastozoans were epibenthic, unattached suspension feeders (Fatka and Kordule, 1985; Friedrich, 1993; Dom|¤nguez-Alonso, 1998; Guensburg and Sprinkle, 2001a). Cinctans and ctenocystoids were probably mostly sessile organisms, as suggested by the absence of any locomotory device in ctenocystoids, and by the rigidity of the cinctan ‘appendage’ (stele), which was probably used to anchor the animal into the substrate (Ubaghs, 1975; Parsley, 1998). Edrioasteroids were disk-shaped sediment stickers, which could attach by their basal surface to ¢rm substrates, skeletal fragments or living organisms (Prokop, 1965; Sumrall, 2000). Stylophorans were the only arm-bearing Middle Cambrian echinoderms recorded on the NGM. Like cinctans and ctenocystoids, their £attened body and extremely reduced number of ambulacra (one) can be interpreted as derived characters suggesting an adaptation to an epibenthic, free mode of life (Dzik, 1999; David et al., 2000). Primitive stylophorans (Ceratocystis, Protocystites) were probably mostly sessile, as suggested by their massive, asymmetrical theca, with strong protuberances on the lower surface. The delicate and brittle feeding appendage (aulacophore) of stylophorans could hardly have been used to propel the organism (Ubaghs, 1967b ; Lefebvre et al., 1998a). Stylophorans were low-level suspension feeders, with their aulacophore probably extending over the substrate and facing the current, in life position (Parsley et al., 2000). Contrary to the situation in Laurentia, neither homoiostelean blastozoans (solutes) nor crinoids have been reported so far in Gondwanan Africa and peri-Gondwanan Europe. At least three di¡erent Middle Cambrian echinoderm assemblages existed on the NGM. In high-energy, shallow environments with ¢rm substrates, echinoderm assemblages are largely dominated by eocrinoids, associated with edrioasteroids and rare cinctans. Eocrinoid-sponge meadows corresponding to such assemblages have been reported in the uppermost Lower to lowermost Middle Cambrian of the Iberian Chains Ł lvaro and Vennin, 1997, (Northeast Spain; A 2001). A stylophoran-dominated community

has been reported in one locality from Bohemia (Skryje area; Ubaghs, 1967d; Je¡eries, 1969; Parsley and Prokop, in press). This second assemblage is characterised by abundant monospeci¢c remains of large, heavily plated primitive stylophorans (Ceratocystis), and it occurs in association with the trilobite Ctenocephalus in sandy lithologies, corresponding to shallow environments (below storm wave base). In low-energy, more distal environments, soft-substrate echinoderm assemblages are largely dominated by cinctans and ctenocystoids, associated with eocrinoids, with rarer stylophorans and edrioasteroids. Such softsubstrate cinctan-dominated communities have been reported from Bohemia (Fatka and Kordule, 1985; Friedrich, 1993; Parsley and Prokop, in press), Germany (Friedrich, 1993), Montagne Noire (Termier and Termier, 1973; Friedrich, Ł lvaro et 1993; Vizca|«no and Lefebvre, 1999; A al., 2001b), Morocco and Wales (Friedrich, 1993, 1995), Sardinia (Friedrich, 1995; Loi et al., 1995), and Spain (Friedrich, 1993; Sdzuy, Ł l1993; Gil Cid and Dom|¤nguez-Alonso, 1999; A varo and Vennin, 2001). 2.3. Upper Cambrian echinoderm faunas The fossil record of Upper Cambrian echinoderms is extremely poor on the NGM, with a unique fauna described so far from the Montagne Noire, France (Ubaghs, 1998; see Appendix). This situation probably re£ects taphonomic bias (disarticulated specimens may not easily be identi¢ed) and from the scarcity of Upper Cambrian exposures and echinoderm Lagersta«tten (sampling e¡ect; Sprinkle, 1981; Smith, 1988; Ubaghs, 1998), combined with the short duration of the Upper Cambrian (McKerrow and Van Staal, 2000). The Montagne Noire fauna (Fig. 3) comprises a single specimen of edrioasteroid (?Stromatocystites) and dozens of probably isolated edrioasteroid fragments (identi¢ed as Scoteinocystis by Ubaghs, 1998), ¢ve specimens of stylophorans (two cornutes and three examples of the primitive mitrate Lobocarpus), and several hundreds of specimens assigned to two genera of Macrocystella-like rhombiferan blastozoans (Velieuxicystis and Barroubiocystis). These blastozoans are char-

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

acterised by a theca organised into ¢ve regular rows of plates, and a well-de¢ned bipartite stem made of cylindrical columnals. The di¡erentiation of the rhombiferan stem into a short and broad, highly £exible proximal region, and a long and narrow, more rigid distal portion is mirrored in other unattached, epibenthic blastozoans (e.g. solutes) which used their stem to crawl or rest on soft substrates. Consequently, the possession of such a bipartite stem, as well as the absence of any anchorage structure (holdfast, root) at its distal end, along with the possession of a non-£attened theca suggest that Macrocystella-like rhombiferans were free, epibenthic suspension feeders, which used their stem to rest on soft substrates, and whose life position was probably with the theca held upright above the sea£oor (Paul, 1968). The Montagne Noire fauna is a soft-substrate assemblage. The preservation of articulated skeletons in most specimens suggests low-energy conditions and rapid burial (Brett et al., 1997). In composition this fauna (Fig. 3) is comparable to that of Middle Cambrian soft-substrate echinoderm assemblages, as it is largely dominated by blastozoans, associated with rare edrioasteroids and stylophorans. However, important di¡erences can be noticed within blastozoans : (1) cinctans and ctenocystoids have disappeared, and (2) stemmed blastozoans, which constituted a minor component of the Middle Cambrian assemblages, are the dominant Upper Cambrian group. The expansion of stemmed blastozoans into soft-substrate environments in the Middle Cambrian was probably limited by their need to attach to a ¢rm substrate (e.g. algae, skeletal fragments; Sprinkle and Guensburg, 1995; Guensburg and Sprinkle, 2001a). The di¡erentiation of a bipartite stem in

Fig. 3. Composition of the Upper Cambrian echinoderm fauna from Montagne Noire (MN), distinguishing the three main clades (E, edrioasteroids; B, blastozoans; A, arm-bearing echinoderms).

79

Macrocystella-like rhombiferans allowed a major expansion and the extensive colonisation of soft substrates by these blastozoans in the Upper Cambrian. Important accumulations of the enigmatic fossil Oryctoconus have been reported in high-energy, shallow environments across the Cambro^Ordovician transition in Spain, Sardinia, and Germany (see Appendix). New evidence based on collections of abundant new material from the Iberian Ł lvaro, pers. commun. 2001) lend strong Chains (A support to the interpretation of Oryctoconus as a grapnel-like structure belonging to the distal end of an eocrinoid stem (Colchen and Ubaghs, 1969). If this interpretation is correct, it would suggest the presence of large late Late Cambrian to early Tremadocian eocrinoid meadows in shallow, high-energy environments in several regions of the NGM. Contrary to the situation in the Middle Cambrian, these eocrinoid meadows developed on soft, siliciclastic substrates, and not on ¢rm substrates. The grapnel-like structure at the distal end of the stem was an anchorage device representing an adaptation to the life on soft substrates. 2.4. Lower Ordovician echinoderm faunas Abundant and diverse echinoderm faunas have been recorded in the Tremadocian and Arenig of several regions of the NGM (Fig. 4; see Appendix). They are all soft-substrate assemblages and comprise representatives of all three echinoderm clades. Edrioasteroids are extremely rare, and they are reported so far only from the lower Arenig of Montagne Noire and Morocco (Hemicystites, ?Pyrgocystis ; see Appendix). On the NGM, blastozoans were more diverse in the Lower Ordovician than in the Upper Cambrian. They comprised eocrinoids and rhombiferans, but also representatives of three new groups: diploporans, parablastoids, and solutes. As in the Cambrian, eocrinoid meadows have been described in shallow, high-energy Lower Ordovician environments in several regions of the NGM. These meadows were generally dominated by a single genus: Rhopalocystis in the Tremadoc of Morocco, and Lingulocystis in the Arenig of Montagne Noire (see Appendix). The lower Are-

PALAEO 3074 9-5-03

80

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Fig. 4. Composition of Lower Ordovician echinoderm faunas from several regions of the northern Gondwana margin, distinguishing the three main clades (E, edrioasteroids; B, blastozoans; A, arm-bearing echinoderms). Abbreviations: MN, Montagne Noire; GB, Great Britain; MO, Morocco; BO, Bohemia.

nig eocrinoid Balantiocystis occurs in more distal, low-energy environments in Montagne Noire and Morocco (see Appendix). The possession of a grapnel-like structure at the distal end of its stem allowed Balantiocystis to grip soft muddy substrates. As in the Upper Cambrian, Lower Ordovician rhombiferans (e.g. Macrocystella, Cheirocrinus, and Cheirocystella) were epibenthic, unattached suspension feeders. Macrocystella was the most common Lower Ordovician rhombiferan on the NGM. This genus has been reported in Bohemia, England, Germany, Montagne Noire, Morocco, Spain, and Wales (see Appendix). Diploporans were highly derived eocrinoids, characterised by the possession of poorly-organised thecae, the frequent absence of a stem, and the presence of typical perforations (diplopores). On the NGM, Lower Ordovician diploporans have been recorded from Bohemia, Morocco, and Germany (see Appendix). Diploporans were epibenthic, attached suspension feeders that lived on soft substrates, either stuck to skeletal fragments (e.g. Aristocystites), or with the lower portion of the organism ¢rmly inserted into the mud (e.g. Calix; Chauvel, 1941; Gutie¤rrez-Marco et

al., 1984). The parablastoids are a small group of derived eocrinoids characterised by a highly di¡erentiated, blastoid-like theca, internal respiratory structures (cataspires), and a well-de¢ned stem consisting of cylindrical columnals. The mode of life of the parablastoids is poorly known, but they were probably epibenthic, attached suspension feeders. Very few specimens of parablastoids are known on the NGM, and all of them are from the early to late Arenig of Wales (Paul and Cope, 1982). The solute Minervaecystis represents the oldest (lower Arenig) and only known occurrence of this class of blastozoans in the Lower Ordovician of the NGM. Minervaecystis has been reported from the Montagne Noire, France (Thoral, 1935; Ubaghs, 1969a), but is also present in Morocco (Lefebvre, pers. observations). Solutes probably originated in Laurentia, where they are recorded from the Middle Cambrian onwards (Ubaghs and Robison, 1985; Daley, 1995, 1996). The late appearance of this class on the NGM is possibly connected to the important Lower Ordovician transgressions. Ordovician solutes were free, unattached, epibenthic, low-level suspension or detritus feeders, which used their highly di¡erentiated stem to crawl or rest on soft substrates (Kolata et al., 1977; Daley, 1992a). Lower Ordovician arm-bearing echinoderms were much more abundant and diverse than in the Cambrian, with the appearance of two new groups (asterozoans and crinoids), along with stylophorans. Lower Ordovician stylophorans were characterised by an important diversi¢cation in most regions of the NGM, with the ¢rst appearance in the fossil record of representatives of the two cornute suborders Amygdalothecida (e.g. Amygdalotheca, Galliaecystis) and Cothurnocystida (e.g. Cothurnocystis, Phyllocystis), and of the three mitrate suborders Lagynocystida (Lagynocystis), Peltocystida (e.g. Anatifopsis, Peltocystis), and Mitrocystitida (e.g. Chinianocarpos, Ovocarpus). Lower Arenig cornutes and mitrates from Montagne Noire constitute the most diverse stylophoran assemblage recorded in the world (26 species), and represent more than 20% of all described stylophoran species (see Appendix). As in the Cambrian, Lower Ordovician stylophorans were mostly sessile, epibenthic, low-level suspen-

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

sion feeders. They were particularly abundant on soft substrates, in low-energy distal environments. However, some of them (kirkocystid mitrates) were apparently environmental generalists that could be also locally abundant in more proximal, shallower, high-energy environments, as suggested by the accumulation of several hundreds of isolated kirkocystid elements observed in some tempestites in the Montagne Noire (Vizca|«no et al., 2001). On the NGM, two genera of primitive Lower Ordovician crinoids are known from rare specimens : Ramseyocrinus in Montagne Noire, Morocco, and Wales and Aethocrinus in Montagne Noire and Shropshire, England (see Appendix). Possible crinoid columnals have been reported in the late Arenig of Bohemia, associated with ¢rm volcanic substrates (tu¡s; Zel|¤zko, 1911, Prokop and Petr, 1999). However, like solutes, older and more primitive crinoids (e.g. Echmatocrinus) are known since the Middle Cambrian in Laurentia (Sprinkle, 1973; Smith, 1984, 1988; Sprinkle and Collins, 1998; Guensburg and Sprinkle, 2001b; for di¡ering interpretations of Echmatocrinus, see Conway Morris, 1993; Ausich and Babcock, 1998). This suggests that the late appearance of crinoids on the NGM is possibly related to Lower Ordovician transgressions. Lower Ordovician crinoids were sessile, epibenthic, lowto high-level suspension feeders that needed a solid substrate to attach to (Guensburg and Sprinkle, 2001a). Because of the scarcity of suitable surfaces for attachment (e.g. hardgrounds, reefs) in the cold environments of the NGM, Lower Ordovician crinoids were not abundant and attached to skeletal fragments. On the NGM, a new group of arm-bearing echinoderms, the asterozoans (asteroids, ophiuroids, and somasteroids), appeared for the ¢rst time in the fossil record in the Lower Ordovician. Asterozoans are known in the lower Arenig of the Montagne Noire, France (Ampullaster, Chinianaster, Pradesura, and Villebrunaster), Shropshire, England (Palaeura), and Wales (Petraster ; see Appendix). Most primitive asterozoans were infaunal and they were suspension feeders, deposit feeders, and predators (Spencer, 1951). Like the crinoids, solutes, and parablastoids, asterozoans possibly represent Laurentian invaders that reached Gond-

81

wanan shores during the Lower Ordovician transgressions. As all Lower Ordovician echinoderm faunas of the NGM were soft-substrate assemblages, forms that needed hard or ¢rm substrates for attachment were poorly represented (crinoids, edrioasteroids, and parablastoids), whereas organisms adapted to soft substrates were much more abundant and diverse (asterozoans, diploporans, eocrinoids, rhombiferans, and stylophorans). In most regions of the NGM, shallow, high-energy environments were dominated by eocrinoids, associated with stylophorans (cornutes, kirkocystids) and rare crinoids. In more distal, deeper, low-energy environments, the situation was more varied. In Bohemia, echinoderm faunas were comparable to those of the Upper Cambrian, with low-diversity assemblages dominated by blastozoans (Fig. 4 ; see Appendix). The Upper Cambrian-like composition of echinoderm faunas and the extreme scarcity of groups with Laurentian a⁄nities (e.g. crinoids) in Bohemia suggest that this region was more or less isolated during the Lower Ordovician. The situation was clearly di¡erent in France (Montagne Noire), Morocco, or Great Britain (Shropshire, Wales), with more diverse echinoderm faunas, largely dominated by stylophorans, associated with asterozoans and blastozoans (Fig. 4). The signi¢cant increase in diversity in these regions resulted mainly from the important diversi¢cation of stylophorans. The arrival of several groups with Laurentian a⁄nities (crinoids, parablastoids, solutes, and possibly asterozoans) in Montagne Noire, Morocco and Great Britain suggests that these regions were palaeogeographically close so that faunal and £oral exchange was easily maintained compared to Bohemia in the Lower Ordovician. 2.5. Middle Ordovician echinoderm faunas Abundant and diverse echinoderm faunas, comprising members of the three main echinoderm clades, are recorded in the Llanvirn ( = Darriwillian = Abereiddian (Oretanian) plus Llandeilian) of the NGM (Fig. 5 ; see Appendix). As in the Cambrian and the Lower Ordovician, edrioasteroids were not common and formed a minor com-

PALAEO 3074 9-5-03

82

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Fig. 5. Composition of Middle Ordovician echinoderm faunas from several regions of the northern margin of Gondwana, distinguishing the three main clades (E, edrioasteroids; B, blastozoans; A, arm-bearing echinoderms). Abbreviations: BO, Bohemia; SP, Spain; AM, Armorican massif; MO, Morocco.

ponent of echinoderm faunas. On the NGM, blastozoans were characterised by the disappearance of small groups of Lower Ordovician Laurentian invaders (parablastoids, solutes), the scarcity of representatives of Cambrian groups such as eocrinoids (e.g. Ascocystites) and rhombiferans (e.g. Echinosphaerites), and the important diversi¢cation of diploporans (e.g. Aristocystites, Calix, and Phlyctocystis). Stylophorans were locally abundant (e.g. several hundreds of specimens sampled from the localities of Osek and Sa¤rka, in Bohemia, and Traveusot in the Armorican massif), but their diversity was less important than in the Lower Ordovician, with stylophoran communities frequently dominated by a single species : Mitrocystites in Osek, Lagynocystis in Sa¤rka, and Mitrocystella in Traveusot. Most Middle Ordovician stylophorans were epibenthic suspension feeders (e.g. all cornutes, the mitrate Mitrocystites). However, the presence of cuestashaped ribs on the theca of some mitrates (e.g. Mitrocystella) suggests the adoption of an infaunal mode of life (Je¡eries, 1984; Lefebvre, in press). As in the Lower Ordovician, other armbearing echinoderms were crinoids (e.g. Coralcrinus, Heviacrinus, and Ramseyocrinus), and asterozoans (e.g. Archegonaster, Eophiura, and Palaeura).

All Middle Ordovician echinoderm faunas recorded on the NGM were soft-substrate assemblages, because of the absence of ¢rm ones (e.g. hardgrounds, reefs) in these regions. Echinoderm faunas were thus largely dominated by organisms well-adapted to soft sandy and/or muddy sea£oors (asterozoans, diploporans, stylophorans), whereas the echinoderm groups that required hard surfaces for attachment (crinoids, edrioasteroids) were poorly represented. The expansion of crinoids and edrioasteroids was strongly limited by the few attachment sites available on soft substrates: for example, crinoid holdfasts attached to various skeletal fragments (e.g. pieces of trilobites, other echinoderms, shells of brachiopods or molluscs) have been reported in Spain (Acen‹olaza and Gutie¤rrez-Marco, 1998) and Bohemia (Prokop and Turek, 1997). Large meadows of stemmed blastozoans spreading in proximal, shallow, high-energy environments are poorly known in the Middle Ordovician of the NGM. Such assemblages, dominated by the single genus Ascocystites have been documented in the Moitiers d’Allonne Formation (Abereiddian) of Normandy, France (Re¤gnault, 1990). Most Middle Ordovician echinoderm assemblages ocurred in more distal, deeper, low-energy environments. The distribution of several echinoderms (e.g. stylophorans) in these distal environments was chie£y controled by palaeobathymetry, and correlatively, water temperature (Henry et al., 1997). For example, in shelf environments below storm wave base, high-diversity stylophoran assemblages were largely dominated by mitrocystitids, associated with cornutes, kirkocystids and rare lagynocystids (e.g. Osek in Bohemia, Traveusot in the Armorican massif), whereas in deeper environments (outer shelf, slope?), low-diversity stylophoran assemblages were largely dominated by lagynocystids, associated with rare kirkocystids and cornutes (e.g. SZa¤rka in Bohemia, Ancenis in the Armorican massif, synclinal del Valle in Spain; Henry et al., 1997; Lefebvre et al., 1998b; Parsley, 2000). In Bohemia, echinoderm faunas were diverse and largely dominated by stylophorans, associated with asterozoans and blastozoans (Fig. 5). The composition of the Middle Ordovician Bohemian assemblage thus strongly recalls that of Lower

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Ordovician faunas from France (Montagne Noire), Great Britain or Morocco (see Appendix). The conservative composition of the Bohemian echinoderm fauna could suggest that this region was still more or less isolated during the Middle Ordovician, but this situation could also be due to the presence of a very steep bathymetrical gradient in this basin (absence of the shallowest benthic communities; see Havl|¤c›ek, 1982; Havl|¤c›ek and Vanek, 1990, Havl|¤c›ek and Fatka, 1992). The composition of echinoderm faunas is clearly di¡erent in the Armorican massif, Morocco, Portugal, and Spain (see Appendix). These regions have yielded very similar, highly diversi¢ed assemblages, largely dominated by diploporans, associated with asterozoans and stylophorans (Fig. 5). The high-diversity patterns recorded in these regions are mainly due to the important diversi¢cation of diploporans. 2.6. Diversity pattern of echinoderm faunas on the NGM The fossil record of Middle Cambrian and Lower to Middle Ordovician echinoderms is rich on the NGM, thanks to more than 150 years of intense collecting and the existence of several echinoderm Lagersta«tten. By contrast, Lower and Upper Cambrian echinoderm faunas of the NGM are much more poorly known, mainly because of the extreme scarcity of taphonomic windows for these time intervals in Central and Southwestern Europe and North Africa. Firm substrate echinoderm communities were restricted to the Lower to lowermost Middle Cambrian (eocrinoid meadows associated with sponges, and/or biomats). Firm substrate environments (e.g. reefs, algal and/or microbial mounds) were absent from the Middle Cambrian to the Middle Ordovician, because of the southward drift of the NGM from temperate (?30‡S) to high (?60‡S) palaeolatitudes and the resulting cooling of sea water. Soft-substrate echinoderm communities are known from the Middle Cambrian to the Middle Ordovician. Shallow, high-energy communities dominated by eocrinoids are known from the Middle Cambrian to the Middle Ordovician. More distal, low-energy echinoderm

83

assemblages are dominated successively by cinctans (Middle Cambrian), rhombiferans (Upper Cambrian to Tremadocian), stylophorans (Arenig), and diploporans (Middle Ordovician). The development of these successive echinoderm faunas is apparently delayed in Bohemia, suggesting that this region was more or less isolated, or that the faunal exchange with other peri-Gondwanan areas was limited, at least from the Lower to Middle Ordovician. Edrioasteroids were present on the NGM, from the Middle Cambrian (at least) to the Middle Ordovician, but they remained rare and poorly diversi¢ed, because of the few suitable ¢rm surfaces for attachment. A continuous trend of diversi¢cation can be observed for blastozoans on the NGM, with successive radiations from a primitive Lower Cambrian eocrinoid stock: stalked eocrinoids, cinctans and ctenocystoids (Middle Cambrian), rhombiferans (Upper Cambrian), and diploporans (Lower Ordovician). A similar pattern can be observed for stylophorans (arm-bearing echinoderms), with a progressive diversi¢cation during the Middle and Upper Cambrian, and an important radiation in the Lower Ordovician. In summary, Early Cambrian to Middle Ordovician echinoderm faunas of the NGM show a pattern of continuous diversi¢cation in soft-substrate environments. This pattern was perturbed by the arrival of new groups with Laurentian a⁄nities (crinoids, parablastoids, solutes, and possibly asterozoans) in favour of Lower Ordovician transgressions and the progressive change in palaeogeographic position.

3. Discussion Most recent contributions to palaeoecological aspects of the Cambro^Ordovician radiation of echinoderms have been devoted to Laurentian echinoderm faunas (Guensburg and Sprinkle, 1992, 2001a; Sprinkle and Guensburg, 1995; Sumrall et al., 1997; Dornbos and Bottjer, 2000, 2001). Lower Palaeozoic Laurentian echinoderm faunas are abundant and well known, due to more than 150 years of active sampling and to the presence of numerous echinoderm Lagersta«t-

PALAEO 3074 9-5-03

84

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

ten, even in the Lower and Upper Cambrian (Sumrall et al., 1997; Dornbos and Bottjer, 2000). During Cambro^Ordovician times, Laurentia (much of North America and Greenland) was a continent in equatorial position separated from Gondwana by the Iapetus ocean (Scotese and McKerrow, 1990; Smith, 2001; Cocks, 2001). Consequently, contrary to the situation on the NGM, warm to temperate palaeoenvironmental conditions prevailed on Laurentian shores during this time interval. Early Cambrian Laurentian echinoderms were abundant and diverse on Late Proterozoic-like, soft substrates sealed by microbial mats (Dornbos and Bottjer, 2000, 2001). Two biomat-related lifestyles can be recognised in Lower Cambrian Laurentian echinoderms, following the de¢nitions of Seilacher (1999) : mat encrusters and mat stickers. The disk-shaped morphology and £at base of edrioasteroids support their interpretation as originally mat encrusters. Other Early Cambrian Laurentian echinoderms, such as Camptostroma, helicoplacoids and Lepidocystis, were probably mat stickers, as suggested by their cone-shaped morphology and the absence of anchorage structures (Dornbos and Bottjer, 2000). Continuous increase in bioturbation during Lower to Middle Cambrian times led to the progressive disappearance of Proterozoic-like microbial substrates in favour of modern soft substrates with a mixed layer. This ‘agronomic revolution’ (sensu Seilacher and P£u«ger, 1994) triggered a major radiation of echinoderms and involved dramatic anatomical transformations. The evolution of an extraxial pouch (stalk) was an important innovation allowing several Middle Cambrian echinoderms (e.g. Coleicarpus, Echmatocrinus, and Gogia) to attach to skeletal fragments on soft substrates. In Laurentia, Middle Cambrian soft-substrate echinoderm assemblages were largely dominated by unattached, epibenthic organisms (ctenocystoids, solutes, and stylophorans), generally characterised by a £attened theca and a drastic reduction of the number of ambulacra (Ubaghs, 1963b; Robison and Sprinkle, 1969; Ubaghs and Robison, 1985; Daley, 1995; Sumrall and Sprinkle, 1999). Middle Cambrian attached echinoderms (crinoids, edrioasteroids, and eocrinoids) were rare, because of

the few surfaces available for attachment (e.g. skeletal fragments ; Guensburg and Sprinkle, 1992, 2001a; Sprinkle and Guensburg, 1995). The development of carbonate hardgrounds on Laurentian shores during the Upper Cambrian triggered an important diversi¢cation of echinoderms in these new, shallow, high-energy environments, and the appearance of hard-substrate communities, largely dominated by eocrinoids associated with edrioasteroids (Brett et al., 1983; Rozhnov, 2001). In deeper, low-energy environments, Upper Cambrian soft-substrate echinoderm assemblages were comparable to those of the Middle Cambrian, and were largely dominated by stylophorans and solutes, associated with rare edrioasteroids and eocrinoids (Bell and Sprinkle, 1980; Sumrall et al., 1997). An important radiation of crinoids ocurred in the Lower Ordovician of Laurentia (Guensburg and Sprinkle, 1992, 2001a,b; Sprinkle and Guensburg, 1995). By the Arenig, crinoids were the dominant group in the hard-substrate echinoderm faunas, associated with edrioasteroids and eocrinoids. The Lower Ordovician soft-substrate echinoderm communities in Laurentia were largely dominated by rhombiferans, associated with stylophorans (Guensburg and Sprinkle, 1992; Sprinkle and Guensburg, 1995). The Middle Ordovician hardsubstrate communities in Laurentia were even more dominated by crinoids (decline of eocrinoids), associated with edrioasteroids. The development of several attachment structures (rootlike holdfasts, coiled stems) allowed the expansion of crinoids onto soft substrates during the Middle Ordovician. Distal, low-energy, soft-substrate, Middle Ordovician echinoderm faunas were dominated by crinoids, associated with asterozoans and rhombiferans (Sprinkle and Guensburg, 1995). In summary, the diversity patterns of Cambro^ Ordovician echinoderms recorded in two palaeogeographically distinct regions (NGM and Laurentia) show some similarities, but also important di¡erences. Both regions share comparable Early Cambrian echinoderm faunas well adapted to matgrounds, consisting of mat-encrusters and mat-stickers. In both regions, the ‘agronomic revolution’ triggered a major diversi¢cation event

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

from the uppermost Lower to lowermost Middle Cambrian, with the rapid colonisation of new, soft-substrate environments. However, di¡erent diversi¢cation histories in the two regions are observed in the Upper Cambrian^Middle Ordovician time interval. On soft substrates of the cool, high-palaeolatitude seas of the NGM, a continuous diversi¢cation of blastozoans and stylophorans occurred, with the succession of several low-diversity faunas dominated by cinctans, less frequently by stylophorans and/or eocrinoids (Middle Cambrian), rhombiferans (Upper Cambrian^Tremadocian), stylophorans (Arenig), and diploporans (Middle Ordovician). In Laurentia, a comparable pattern of continuous diversi¢cation of blastozoans and stylophorans occurred on soft substrates, in distal, low-energy, deep (and probably cool ^ below the picnocline between thermospheric and psychrospheric waters?) environments, with the succession of several lowdiversity faunas, dominated by ctenocystoids, solutes and stylophorans (Middle Cambrian), solutes and stylophorans (Upper Cambrian), and rhombiferans and stylophorans (Lower Ordovician). However, on the shallow, high-energy, temperate to warm seashores of Laurentia, the development of carbonate hardgrounds in the latest Cambrian to earliest Ordovician triggered a major diversi¢cation of attached echinoderms, and especially crinoids. Crinoids became the dominant group on hard substrates by the Arenig, and also on soft substrates during the Middle Ordovician. The progressive replacement of blastozoan^stylophoran-dominated faunas by crinoid-dominated faunas in the Lower to Middle Ordovician of Laurentia has been interpreted as evidence supporting the evolutionary fauna model of Sepkoski (1979). In this interpretation, blastozoan^stylophoran-dominated faunas correspond to the CEF, and crinoid-dominated faunas to the PEF (Sprinkle, 1981, 1992; Guensburg and Sprinkle, 1992, 2001a; Sprinkle and Guensburg, 1995; Sumrall et al., 1997). Examination of the diversity patterns for echinoderms in Laurentia and on the NGM suggests that Sepkoski’s model is probably mostly correct for Laurentia, where two major radiations, triggered by important environmental shifts, can be distinguished. However, this model

85

cannot be fully applied to the NGM, since no obvious Ordovician radiation can be observed there, and CEFs continued to dominate from the Middle Cambrian to, at least, the Middle Ordovician. Interestingly, a comparable discrepancy between ‘Laurentian’ and ‘Gondwanan’ patterns has been also documented in the Ordovician of Argentina, based on the comparison of contemporaneous invertebrate faunas (mostly trilobites and brachiopods) from the Precordillera (Laurentian a⁄nities) and the Northwest Basin (Gondwanan a⁄nities ; see Waisfeld and Sa¤nchez, 1996). The major radiation of crinoids in the Lower to Middle Ordovician of Laurentia thus represents not such a general event as supposed, triggered by the development of carbonate hardgrounds in shallow, warm environments. The diversity pattern observed in deeper and cooler environments, both in Laurentia and on the NGM, is clearly that of continuous diversi¢cation from the Middle Cambrian to the Middle Ordovician. Comparison of northern Gondwanan and Laurentian diversity patterns also suggests that the most dramatic diversi¢cation event for echinoderms did not occur in the earliest Cambrian, but probably later, from the late Early to early Middle Cambrian. The ‘agronomic revolution’ triggered the replacement of biomat-related echinoderm faunas probably existing since the Late Proterozoic (e.g. Arkarua), by new echinoderm faunas adapted to modern substrate conditions. Comparison of the diversity patterns in Laurentia and NGM suggests that the diversi¢cation of echinoderms was mostly continuous and exponential during Cambro^Ordovician times. However, severe environmental shifts leading to dramatic changes in substrate consistency may have triggered rapid and major diversi¢cations of echinoderms on a global scale (Cambrian), as well as on a local scale (Ordovician).

Note added in proof Crinoid remains are present in the Middle Ordovincian (Llandeilian) of Brittany, France (Claverie, pers. commun.).

PALAEO 3074 9-5-03

86

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Acknowledgements This paper greatly bene¢ted from exciting disŁ lvaro, C. Babin, J. Destombes, cussions with J.J. A S. Re¤gnault, T. Sa¤nchez, D. Vizca|«no, and B. Waisfeld during the conference on ‘Early Palaeozoic palaeogeography and palaeobiogeography of Western Europe and North Africa’, held at Lille in September 2001. The authors are particularly grateful to C.R.C. Paul, A.B. Smith, and the guest editors for reviewing the manuscript and making many helpful remarks, and to M. Prost for her invaluable help in drafting the ¢gures. This paper is a contribution of the UMR CNRS Bioge¤osciences, Dijon.

Appendix List of Lower Cambrian to Middle Ordovician echinoderm genera recorded in Gondwanan Africa and peri-Gondwanan Europe

Lower Cambrian Andaluc|¤a (South Spain) Eocrinoids (blastozoans): isolated ?eocrinoid plates (Richter and Richter, 1940; Henningsmoen, 1958) Gogia (Alanisicystis) (Ubaghs and Vizca|«no, 1990) Gogia (Gil Cid and Dom|¤nguez-Alonso, 1999) Arago¤n, Iberian Chains (Northeast Spain) Edrioasteroids : edrioasteroid indet. (Lin‹a¤n et al., 1996) Eocrinoids (blastozoans): Ł lvaro and Venisolated ?eocrinoid plates (A Ł nin, 1997, 2001; Alvaro et al., 2000b) Montagne Noire (South France) Eocrinoids (blastozoans): isolated ?eocrinoid plates (Courjault-Rade¤, Ł lvaro et al., 1999; Vizca|«no and Lefebvre, 1988; A 1999)

Sardinia (South Italy) Eocrinoids (blastozoans): isolated ?eocrinoid plates (Selg, 1988) Shropshire (West England) Eocrinoids (blastozoans): isolated ?eocrinoid plates (Donovan and Paul, 1982)

Middle Cambrian Andaluc|¤a and Extremadura (Ossa Morena, South Spain) Eocrinoids (blastozoans): Eocystites (Gil Cid and Dom|¤nguez-Alonso, 1998) Cinctans (blastozoans): Gyrocystis (Friedrich, 1995; Gil Cid and Dom|¤nguez-Alonso, 1998) Stylophorans (arm-bearing echinoderms): Ceratocystis (Gil Cid and Dom|¤nguez-Alonso, 1998) Anti-Atlas (Morocco) Cinctans (blastozoans): Sucocystis (Chauvel, 1971a; Friedrich, 1993) cincta gen. nov. (Friedrich, 1995) Arago¤n, Asturias and Leo¤n (Iberian Chains and Cantabrian Mountains, North Spain) Eocrinoids (blastozoans): Gogia (Gil Cid and Dom|¤nguez-Alonso, 1999) ?Lichenoides (Vizca|«no, pers. commun.) eocrinoids indet. (Gil Cid and Dom|¤nguezAlonso, 1999) Cinctans (blastozoans): Asturicystis (Sdzuy, 1993) Decacystis (Gisle¤n, 1927; Ubaghs, 1967c; Schroeder, 1973; Friedrich, 1993) Gyrocystis (Schroeder, 1973; Gil Cid and Dom|¤nguez-Alonso, 1995; Friedrich, 1993) Progyrocystis (Friedrich, 1993) Sotocinctus (Sdzuy, 1993) Sucocystis (Friedrich, 1993) Trochocystoides (Sdzuy, 1993) cincta gen. nov. (Friedrich, 1993)

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Ctenocystoids (blastozoans) : Ctenocystis (Fatka, pers. observations) Bohemia Edrioasteroids : Stromatocystites (Pompeckj, 1896; Termier and Termier, 1969) Eocrinoids (blastozoans): Acanthocystites (Barrande, 1887; Ubaghs, 1967b; Sprinkle, 1973; Fatka and Kordule, 1984) Akadocrinus (Prokop, 1962; Sprinkle, 1973; Fatka and Kordule, 1991) Lichenoides (Barrande, 1887; Ubaghs, 1953, 1967b; Sprinkle, 1973) Luhocrinus (Prokop and Fatka, 1985) Vyscystis (Fatka and Kordule, 1990) eocrinoid gen. nov. (Fatka, pers. observations) Cinctans (blastozoans): Asturicystis (Fatka and Kordule, 2001) Trochocystites (Barrande, 1887; Jaekel, 1918; Ubaghs, 1967c; Friedrich, 1993) Trochocystoides (Jaekel, 1918; Ubaghs, 1967c; Friedrich, 1993) Ctenocystoids (blastozoans) : Etoctenocystis (Fatka and Kordule, 1985) ?Jugoszowia (Fatka, pers. observations) unassigned ?blastozoans : Cigara (Barrande, 1887; Ubaghs, 1967a; Sprinkle, 1973) Stylophorans (arm-bearing echinoderms): Ceratocystis (Pompeckj, 1896; Ubaghs, 1967d; Je¡eries, 1969) Germany Eocrinoids (blastozoans): isolated ?eocrinoid plates (Sdzuy, 2000) Cinctans (blastozoans): Ludwigicinctus (Friedrich, 1993) Stylophorans (arm-bearing echinoderms): ?Ceratocystis (Gil Cid and Dom|¤nguez-Alonso, 1998) Montagne Noire (South France) Edrioasteroids : Cambraster (Jaekel, 1918; Termier Termier, 1969; Smith, 1985) Eocrinoids (blastozoans):

and

87

Eocystites (Miquel, 1912; Thoral, 1935; Ubaghs, 1987) Gogia (Ubaghs, 1987) Cinctans (blastozoans): Elliptocinctus (Bergeron, 1889; Termier and Termier, 1973; Friedrich, 1993) Gyrocystis (Jaekel, 1918; Friedrich, 1993) Sucocystis (Cabibel et al., 1959; Termier and Termier, 1973; Friedrich, 1993) Ctenocystoids (blastozoans): Ctenocystis (Ubaghs, 1987) Stylophorans (arm-bearing echinoderms): Ceratocystis (Ubaghs, 1987) Sardinia (South Italy) Eocrinoids (blastozoans): Eocystites (Loi et al., 1995) Cinctans (blastozoans): Sucocystis (Friedrich, 1995) Stylophorans (arm-bearing echinoderms): ?Ceratocystis (Loi et al., 1995; Gil Cid and Dom|¤nguez-Alonso, 1998) Wales Cinctans (blastozoans): Davidocinctus (Friedrich, 1993) Elliptocinctus (Friedrich, 1995) Ctenocystoids (blastozoans): Ctenocystis (Je¡eries et al., 1987, p. 438; Smith, 1988) Stylophorans (arm-bearing echinoderms): Protocystites (Hicks and Jones, 1872; Jefferies et al., 1987)

Upper Cambrian Montagne Noire (South France) Edrioasteroids: ?Stromatocystites (Ubaghs, 1998) Rhombiferans (blastozoans): Barroubiocystis (Ubaghs, 1998) Velieuxicystis (Ubaghs, 1998) Stylophorans (arm-bearing echinoderms): Lobocarpus (Ubaghs, 1998) cornutes indet. (Ubaghs, 1998)

PALAEO 3074 9-5-03

88

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Lower Ordovician (Tremadocian^Arenig) Anti-Atlas (Morocco) Edrioasteroids : Hemicystites (Chauvel, 1978) Eocrinoids (blastozoans): Balantiocystis (Chauvel, 1966a, 1978) Rhopalocystis (Ubaghs, 1963a; Chauvel, 1971b, 1978; Chauvel and Re¤gnault, 1986) Diploporans (blastozoans): Aristocystites (Chauvel, 1966a) Palaeosphaeronites (Chauvel, 1966a) Rhombiferans (blastozoans): Macrocystella (Chauvel, 1966a, 1969a) Solutes (blastozoans): Minervaecystis (Lefebvre, pers. observations) Crinoids (arm-bearing echinoderms): Ramseyocrinus (Donovan and Savill, 1988) Stylophorans (arm-bearing echinoderms): Anatifopsis (Chauvel, 1966b) Chauvelicystis (Chauvel, 1966a) Phyllocystis (Gigout, 1954) Thoralicystis (Chauvel, 1971a) cothurnocystid cornute gen. nov. (Lefebvre, pers. observations) hanusiid cornute gen. nov. (Lefebvre, pers. observations) mitrate gen. nov. (Lefebvre, pers. observations) Arago¤n (North Spain) ?Eocrinoids (blastozoans) : Oryctoconus (Colchen and Ubaghs, 1969; Wolf, 1980; Gil Cid and Dom|¤nguez-Alonso, 1999) Rhombiferans (blastozoans): Macrocystella (Josopait, 1970) Bohemia Diploporans (blastozoans): Glyptosphaerites (Havl|¤c›ek and Vanek, 1966; Prokop and Petr, 1999) Palaeosphaeronites (Havl|¤c›ek and Vanek, 1966; Prokop and Petr, 1999) ?Pyrocystites (Barrande, 1887; Prokop and Petr, 1999) Rhombiferans (blastozoans): ?Echinosphaerites (Prokop and Petr, 1999)

Macrocystella (Ubaghs, 1983; Smith, 1988; Prokop and Petr, 1999) Crinoids (arm-bearing echinoderms): isolated columnals (Prokop and Petr, 1999) Germany Diploporans (blastozoans): Aristocystitidae gen. et sp. indet. (Sdzuy et al., 2001) ?Eocrinoids (blastozoans): Ł lvaro, pers. Oryctoconus (Sdzuy, 1955; J.J. A commun.) Rhombiferans (blastozoans): Echinosphaerites (Sdzuy et al., 2001) Macrocystella (Sdzuy, 1955; Sdzuy et al., 2001) pelmatozoan columnals gen. et sp. indet. (Sdzuy et al., 2001) Montagne Noire (South France) Edrioasteroids: ?Pyrgocystis (Vizca|«no and Lefebvre, 1999; Vizca|«no et al., 2001) Hemicystites (Thoral, 1935; Ubaghs, 1983) Eocrinoids (blastozoans): Balantiocystis (Ubaghs, 1972, 1983) Lingulocystis (Thoral, 1935; Ubaghs, 1960, 1994) Rhombiferans (blastozoans): Cheirocystella (Thoral, 1935) Macrocystella (Thoral, 1935; Ubaghs, 1983) Solutes (blastozoans): Minervaecystis (Bather, 1913; Thoral, 1935; Ubaghs, 1969a) Asterozoans (arm-bearing echinoderms): Ampullaster (Fell, 1963a,b; Ubaghs, 1983) Chinianaster (Thoral, 1935; Spencer, 1950, 1951; Fell, 1963b; Ubaghs, 1983) Pradesura (Spencer, 1950; Ubaghs, 1983) Villebrunaster (Spencer, 1951; Fell, 1963b; Ubaghs, 1983) Crinoids (arm-bearing echinoderms): Aethocrinus (Ubaghs, 1969b) Ramseyocrinus (Ubaghs, 1983) Stylophorans (arm-bearing echinoderms): Ampelocarpus (Lefebvre and Vizca|«no, 1999) Amygdalotheca (Ubaghs, 1969a)

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Anatifopsis (Thoral, 1935; Vizca|«no and Lefebvre, 1999; Vizca|«no et al., 2001) Arauricystis (Thoral, 1935; Ubaghs, 1969a, 1994; Lefebvre and Vizca|«no, 1999) Balanocystites (Vizca|«no and Lefebvre, 1999; Vizca|«no et al., 2001) Chauvelicystis (Ubaghs, 1969a, 1983; Daley, 1992b) Chinianocarpos (Thoral, 1935; Chauvel, 1941; Ubaghs, 1961, 1969a; Je¡eries, 1986) Cothurnocystis (Ubaghs, 1969a) Galliaecystis (Ubaghs, 1969a, 1983) Lagynocystis (Ubaghs, 1991) Lyricocarpus (Ubaghs, 1994) Nanocarpus (Ubaghs, 1991) Ovocarpus (Ubaghs, 1994; Lefebvre, 2000a; Lefebvre and Gutie¤rrez-Marco, in press) Peltocystis (Thoral, 1935; Ubaghs, 1969a; Je¡eries, 1986) Phyllocystis (Thoral, 1935; Ubaghs, 1967b, 1969a) Proscotiaecystis (Ubaghs, 1983, 1994) Thoralicystis (Ubaghs, 1969a, 1983) Trigonocarpus (Ubaghs, 1994) Vizcainocarpus (Ruta, 1997; Lefebvre, 2000a) Sardinia (South Italy) ?Eocrinoids (blastozoans) : Oryctoconus (Loi et al., 1995) Shropshire (West England) Eocrinoids (blastozoans): eocrinoid gen. nov. (Fortey and Owens, 1991) Rhombiferans (blastozoans): Macrocystella (Paul, 1968, 1984) cheirocrinid isolated plates (Paul, 1984) Asterozoans (arm-bearing echinoderms): Palaeura (Spencer, 1950) Crinoids (arm-bearing echinoderms): Aethocrinus (Donovan, 1986) Stylophorans (arm-bearing echinoderms): Anatifopsis (Fortey and Owens, 1991) Prochauvelicystis (Daley, 1992b) Vizcainocarpus (Lefebvre, 2000a) cornute gen. nov. (Marti-Mus, pers. commun.)

89

Wales Parablastoids (blastozoans): Blastoidocrinus (Paul and Cope, 1982; Smith, 1988) Rhombiferans (blastozoans): Macrocystella (Paul, 1968, 1984) Asterozoans (arm-bearing echinoderms): Petraster (Spencer, 1950; Smith, 1988) Crinoids (arm-bearing echinoderms): Ramseyocrinus (Bates, 1968; Donovan, 1984; Cope, 1988; Smith, 1988) Stylophorans (arm-bearing echinoderms) Anatifopsis (Je¡eries, 1987) Balanocystites (Je¡eries, 1987) Cothurnocystis (Je¡eries, 1987; Woods and Je¡eries, 1992) Lagynocystis (Je¡eries, 1987) hanusiid cornute gen. nov. (Je¡eries, 1987) mitrocystitid mitrate gen. A (Je¡eries, 1987) mitrocystitid mitrate gen. B (Je¡eries, 1987)

Middle Ordovician (Llanvirn) Algeria Diploporans (blastozoans): Lepidocalix (Termier and Termier, 1950a) Sinocystis (Termier and Termier, 1950a) Anti-Atlas (Morocco) Diploporans (blastozoans): Aristocystites (Termier and Termier, 1950b; Chauvel, 1966a) Asteroblastus (Termier and Termier, 1950b) Calix (Termier and Termier, 1950b; Chauvel, 1966a, 1978) Isadalocystis (Chauvel, 1978) Phlyctocystis (Chauvel, 1966a, 1978) Protocrinites (Chauvel, 1978) Sinocystis (Termier and Termier, 1950b; Chauvel, 1966a) Sphaeronites (Termier and Termier, 1950b; Chauvel, 1966a, 1978) diploporans indet. (Chauvel, 1966a, 1980) Rhombiferans (blastozoans): Echinosphaerites (Chauvel, 1966a) Asterozoans (arm-bearing echinoderms): asteroids indet. (Destombes, pers. commun.)

PALAEO 3074 9-5-03

90

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

ophiuroids indet. (Destombes, pers. commun.) Crinoids (arm-bearing echinoderms): Botriocrinus (Termier and Termier, 1950b) fragments of stems (Termier and Termier, 1950b) Stylophorans (arm-bearing echinoderms): Anatifopsis (Cripps, 1990; Beisswenger, 1994) Aspidocarpus (Chauvel, 1971a; Cripps, 1990) Eumitrocystella (Chauvel, 1971a; Beisswenger, 1994) ?Lyricocarpus (Chauvel, 1971a, ¢g. 3g) ?Scotiaecystis (Chauvel, 1971a, ¢g. 2b^d) Bohemia Edrioasteroids : ?Argodiscus (Plas and Prokop, 1979; Prokop and Petr, 1999) Hemicystites (Havl|¤c›ek and Vanek, 1966) Eocrinoids (blastozoans): Archaeocystites (Barrande, 1887) Diploporans (blastozoans): Archegocystis (Spencer, 1951; Havl|¤c›ek and Vanek, 1966) Calix (Havl|¤c›ek and Vanek, 1966) Pyrocystites (Barrande, 1887; Havl|¤c›ek and Vanek, 1966) Asterozoans (arm-bearing echinoderms): Archegonaster (Spencer, 1950; Havl|¤c›ek and Vanek, 1966; Prokop and Petr, 1999) Asterias (Havl|¤c›ek and Vanek, 1966) Eophiura (Spencer, 1950, 1951; Havl|¤c›ek and Vanek, 1966; Prokop and Petr, 1999) Hypophiura (Prokop and Petr, 1999) Palaeura (Spencer, 1950, 1951; Havl|¤c›ek and Vanek, 1966; Prokop and Petr, 1999) asteroids indet. (Prokop and Petr, 1999) Crinoids (arm-bearing echinoderms): Caleidocrinus (Prokop and Petr, 1999) Ramseyocrinus (Prokop and Petr, 1999) crinoid stems and holdfasts (Prokop and Turek, 1997; Prokop and Petr, 1999) Stylophorans (arm-bearing echinoderms): Anatifopsis (Barrande, 1872; Parsley et al., 2000) Balanocystites (Barrande, 1872, 1887; Chauvel, 1941; Parsley et al., 2000) Hanusia (Cripps, 1989a)

Lagynocystis (Barrande, 1887; Jaekel, 1918; Chauvel, 1941; Je¡eries, 1973; Parsley, 2000) Mitrocystella (Barrande, 1887; Chauvel, 1941) Mitrocystites (Barrande, 1887; Chauvel, 1941; Je¡eries, 1968; Parsley, 1994) Prokopicystis (Cripps, 1989b) Promitrocystites (Jaekel, 1918; Chauvel, 1941; Parsley, 1994) Reticulocarpos (Je¡eries and Prokop, 1972) Thoralicystis (Ubaghs, 1967b) Brittany and Normandy (West France) Edrioasteroids: ?Pyrgocystis (Lefebvre, pers. observations) Eocrinoids (blastozoans): Ascocystites (Chauvel, 1941; Re¤gnault, 1990) Diploporans (blastozoans): Aristocystites (Kerforne, 1901; Chauvel, 1941) Calix (Rouault, 1851; Chauvel, 1941, 1977, 1980) Codiacystis (Chauvel, 1980) ?Lepidocalix (Lefebvre, pers. observations) Oehlerticystis (Chauvel, 1980) Pachycalix (Chauvel, 1941) Phlyctocystis (Chauvel, 1969b, 1980) Tholocystis (Chauvel, 1941, 1980) diploporans indet. (Chauvel, 1980) Asterozoans (arm-bearing echinoderms): ophiuroids indet. (Lefebvre, pers. observations) Stylophorans (arm-bearing echinoderms): Anatifopsis (Kerforne, 1901; Chauvel, 1941, 1981; Lefebvre, 2000b) Aspidocarpus (Chauvel, 1981; Lefebvre, 2000b) Balanocystites (Chauvel, 1941, 1981; Lefebvre, 2000b) Beryllia (Cripps and Daley, 1994) Diamphidiocystis (Chauvel, 1981; Lefebvre, 2000b) Domfrontia (Chauvel and Nion, 1977; Cripps and Daley, 1994) Lagynocystis (Chauvel and Nion, 1977; Henry et al., 1997) Milonicystis (Chauvel, 1986; Cripps and Daley, 1994)

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Mitrocystella (Chauvel, 1941, 1981; Je¡eries, 1968) Scotiaecystis (Lefebvre and Vizca|«no, 1999) Portugal and Spain Edrioasteroids : edrioasteroid indet. (Gutie¤rrez-Marco et al., 1984) Diploporans (blastozoans): Aristocystites (Gutie¤rrez-Marco and Baeza, 1996; Gutie¤rrez-Marco et al., 1999) Batalleria (Chauvel and Mele¤ndez, 1978, 1986) Calix (Gutie¤rrez-Marco et al., 1984; Gutie¤rrez-Marco and Acen‹olaza, 1999) Codiacystis (Gutie¤rrez-Marco et al., 1984, 1999) Destombesia (Gutie¤rrez-Marco et al., 1999) Oretanocalix (Gutie¤rrez-Marco, 2000) Phlyctocystis (Gutie¤rrez-Marco et al., 1984; Chauvel and Mele¤ndez, 1986) diploporans indet. (Chauvel and Mele¤ndez, 1978; Gutie¤rrez-Marco et al., 1984) Asterozoans (arm-bearing echinoderms): Palaeura (Gutie¤rrez-Marco et al., 1984) ?Urosoma (Chauvel and Mele¤ndez, 1978) encrinasterid indet. (Gutie¤rrez-Marco et al., 1984) stenurid indet. (Gutie¤rrez-Marco et al., 1999) Crinoids (arm-bearing echinoderms): Coralcrinus (Gutie¤rrez-Marco, 1999) Heviacrinus (Gil Cid et al., 1996a) Ramseyocrinus (Gutie¤rrez-Marco et al., 1984) crinoids indet. (Gutie¤rrez-Marco et al., 1984, 1999) crinoid holdfasts and stems (Gutie¤rrez-Marco et al., 1984; Acen‹olaza and Gutie¤rrez-Marco, 1998) Stylophorans (arm-bearing echinoderms): Anatifopsis (Dom|¤nguez and Gutie¤rrez-Marco, 1990; Gutie¤rrez-Marco et al., 1999) Lagynocystis (Gutie¤rrez-Marco et al., 1992) Mitrocystella (Chauvel and Mele¤ndez, 1978; Gutie¤rrez-Marco and Mele¤ndez, 1987) Ovocarpus (Lefebvre and Gutie¤rrez-Marco, in press) Scotiaecystis (Gil Cid et al., 1996b) cornute indet. (Lefebvre, pers. observations)

91

mitrocystitid gen. indet. 1 (Gutie¤rrez-Marco et al., 1984; Lefebvre and Gutie¤rrez-Marco, in press) mitrocystitid gen. indet. 2 (Lefebvre and Gutie¤rrez-Marco, in press)

References Acen‹olaza, G.F., Gutie¤rrez-Marco, J.C., 1998. Estructuras de ¢jacio¤n de pelmatozoos (equinodermos) en el Ordov|¤cico Medio de la Zona Centroibe¤rica espan‹ola. Col. Paleontol. 49, 23^40. Ł lvaro, J.J., Vennin, E., 1997. Episodic development of CamA brian eocrinoid-sponge meadows in the Iberian Chains (NE Spain). Facies 37, 49^64. Ł lvaro, J.J., Vennin, E., 2001. Benthic marine communities A recorded in the Cambrian Iberian platform, NE Spain. Palaeontographica A 262, 1^23. Ł lvaro, J.J., Vennin, E., Vizca|«no, D., 1999. Depositional conA trols on Early Cambrian microbial carbonates from the Montagne Noire, southern France. Trans. R. Soc. Edinb. Earth Sci. 89, 135^143. Ł lvaro, J.J., Rouchy, J.M., Bechsta«dt, T., Boucot, A., Boyer, A F., Debrenne, F., Moreno-Eiris, E., Perejo¤n, A., Vennin, E., 2000a. Evaporitic constraints on the southward drifting of the western Gondwana margin during Early Cambrian times. Palaeogeogr. Palaeoclimatol. Palaeoecol. 160, 105^ 122. Ł lvaro, J.J., Vennin, E., Moreno-Eiris, E., Perejo¤n, A., BechA sta«dt, T., 2000b. Sedimentary patterns across the Lower^ Middle Cambrian transition in the Esla nappe (Cantabrian Mountains, northern Spain). Sediment. Geol. 137, 43^61. Ł lvaro, J.J., Debrenne, F., Vizca|«no, D., 2001a. The Lower A Cambrian of the southern Montagne Noire. Ann. Soc. Ge¤ol. Nord 2, 201^204. Ł lvaro, J.J., Lefebvre, B., Shergold, J.H., Vizca|«no, D., 2001b. A The Middle^Upper Cambrian of the southern Montagne Noire. Ann. Soc. Ge¤ol. Nord 2, 205^211. Ausich, W.I., Babcock, L.E., 1998. The phylogenetic position of Echmatocrinus brachiatus, a probable octocoral from the Burgess Shale. Palaeontology 41, 193^202. Barrande, J., 1872. Syste'me Silurien du centre de la Bohe“me, Supple¤ment au Volume I. Trilobites, crustace¤s divers et poissons. Prague, 647 pp. Barrande, J., 1887. Syste'me Silurien du centre de la Bohe“me, Volume VII. Classe des Echinodermes, Ordre des Cystide¤es. Rivnac, Prague/Gerhard, Leipzig, 233 pp. Bates, D.E., 1968. On Dendrocrinus cambriensis Hicks, the earliest known crinoid. Palaeontology 11, 406^409. Bather, F.A., 1913. Caradocian Cystidea from Girvan. Trans. R. Soc. Edinb. 49, 359^529. Beisswenger, M., 1994. A calcichordate interpretation of the new mitrate Eumitrocystella savilli from the Ordovician of Morocco. Pala«ont. Z. 68, 443^462.

PALAEO 3074 9-5-03

92

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Bell, G.L., Jr., Sprinkle, J., 1980. New homoiostelean echinoderms from the Late Cambrian of Alabama. Geol. Soc. Am. Abstr. Programs 12, 385. Bergeron, J., 1889. Etude ge¤ologique du Massif ancien situe¤ au sud du plateau central. Ann. Soc. Ge¤ol. Nord 22, 1^362. Brett, C.E., Liddell, W.D., Derstler, K., 1983. Late Cambrian hard substrate communities from Montana/Wyoming: the oldest known hardground encrusters. Lethaia 16, 281^289. Brett, C.E., Mo¡at, H.A., Taylor, W.L., 1997. Echinoderm taphonomy, taphofacies, and Lagersta«tten. In: Waters, J.A., Maples, C.G. (Eds.), Geobiology of Echinoderms. Paleontol. Soc. Pap. 3, pp. 147^190. Cabibel, J., Termier, H., Termier, G., 1959. Les e¤chinodermes me¤socambriens de la Montagne Noire (Sud de la France). Ann. Paleontol. 44, 281^294. Chauvel, J., 1941. Recherches sur les cysto|«des et les carpo|«des armoricains. Me¤m. Soc. ge¤ol. mine¤r. Bretagne 5, 1^286. Chauvel, J., 1966a. Echinodermes de l’Ordovicien du Maroc. Cahiers de pale¤ontologie, Ed. du CNRS, Paris, 120 pp. Chauvel, J., 1966b. Pre¤sence du genre Anatifopsis Barrande (Crustace¤s) dans l’Ordovicien du Maroc. C. R. somm. Se¤ances Soc. ge¤ol. Fr. 8, 301^302. Chauvel, J., 1969a. Les e¤chinodermes macrocystellides de l’Anti-Atlas marocain. Bull. Soc. ge¤ol. mine¤r. Bretagne C 1, 21^ 32. Chauvel, J., 1969b. Sur la pre¤sence de Phlyctocystis Chauvel, 1966, e¤chinoderme cysto|«de, dans l’Ordovicien de la presqu’|“le de Crozon. Bull. Soc. ge¤ol. Fr. 7, 443^446. Chauvel, J., 1971a. Les e¤chinodermes carpo|«des du Pale¤ozo|«que infe¤rieur marocain. Notes Serv. ge¤ol. Maroc 237, 49^60. Chauvel, J., 1971b. Rhopalocystis Ubaghs: un e¤chinoderme e¤ocrino|«de du Tre¤madocien de l’Anti-Atlas marocain. Coll. Ordovicien^Silurien, Brest 1971, Me¤m. BRGM 73, pp. 43^ 49. Chauvel, J., 1977. Calix sedgwicki Rouault (e¤chinoderme cysto|«de, Ordovicien du Massif armoricain) et l’appareil ambulacraire des diploporites. C. R. Soc. ge¤ol. Fr. 6, 314^317. Chauvel, J., 1978. Comple¤ments sur les e¤chinodermes du Pale¤ozo|«que infe¤rieur marocain (diploporites, e¤ocrino|«des, e¤drioaste¤ro|«des). Notes Me¤m. Serv. ge¤ol. Maroc 272, 27^78. Chauvel, J., 1980. Donne¤es nouvelles sur quelques cysto|«des diploporites (e¤chinodermes) du Pale¤ozo|«que armoricain. Bull. Soc. ge¤ol. mine¤r. Bretagne 12, 1^28. Chauvel, J., 1981. Etude critique de quelques e¤chinodermes stylophores du Massif armoricain. Bull. Soc. ge¤ol. mine¤r. Bretagne C 13, 67^101. Chauvel, J., 1986. Milonicystis kerfornei n. gen. n. sp. un nouvel e¤chinoderme homalozoaire de l’Ordovicien armoricain. Hercynica 2, 79^81. Chauvel, J., Mele¤ndez, B., 1978. Les e¤chinodermes (cysto|«des, aste¤rozoaires, homalozoaires) de l’Ordovicien moyen des Monts de Tole'de (Espagne). Est. geol. 34, 75^87. Chauvel, J., Mele¤ndez, B., 1986. Note comple¤mentaire sur les e¤chinodermes ordoviciens de Sierra Morena. Est. geol. 42, 451^459. Chauvel, J., Nion, J., 1977. Echinodermes (Homalozoa: Cor-

nuta et Mitrata) nouveaux pour l’Ordovicien du Massif armoricain et conse¤quences pale¤oge¤ographiques. Geobios 10, 35^49. Chauvel, J., Re¤gnault, S., 1986. Variabilite¤ du genre Rhopalocystis Ubaghs, e¤ocrino|«de du Tre¤madocien de l’Anti-Atlas marocain. Geobios 19, 863^870. Chen, J.Y., Oliveri, P., Li, C.W., Zhou, G.Q., Gao, F., Hagadorn, J.W., Davidson, K., Davidson, E.H., 2000. Precambrian animal diversity: putative phosphatic embryos from the Doushantuo Formation of China. Proc. Natl. Acad. Sci. USA 97, 4457^4462. Cocks, L.R.M., 2001. Ordovician and Silurian global geography. J. Geol. Soc. Lond. 158, 197^210. Cocks, L.R.M., McKerrow, W.S., Van Staal, C.R., 1997. The margins of Avalonia. Geol. Mag. 134, 627^636. Colchen, M., Ubaghs, G., 1969. Sur des restes d’e¤chinodermes(?) du Cambro^Ordovicien de la Sierra de la Demanda (Burgos^Logrono, Espagne). Bull. Soc. ge¤ol. Fr. 7, pp. 649^654. Conway Morris, S., 1993. The fossil record and the early evolution of the Metazoa. Nature 361, 219^225. Cope, J.C.W., 1988. A reinterpretation of the Arenig crinoid Ramseyocrinus. Palaeontology 31, 229^235. Courjault-Rade¤, P., 1988. Analyse se¤dimentologique de la Formation de l’Orbiel (‘alternances calcaires’ auct, Cambrien infe¤rieur). Evolution tectono^se¤dimentaire et climatique (versant sud de la Montagne Noire, Massif Central, France). Bull. Soc. ge¤ol. Fr. 8, 1003^1013. Cripps, A.P., 1989a. A new genus of stem chordate (Cornuta) from the Lower and Middle Ordovician of Czechoslovakia and the origin of bilateral symmetry in the chordates. Geobios 22, 215^245. Cripps, A.P., 1989b. A new stem-group chordate (Cornuta) from the Llandeilo of Czechoslovakia and the cornute^mitrate transition. Zool. J. Linn. Soc. 96, 49^85. Cripps, A.P., 1990. A new stem craniate from the Ordovician of Morocco and the search for the sister group of the Craniata. Zool. J. Linn. Soc. 100, 27^71. Cripps, A.P., Daley, P.E.J., 1994. Two cornutes from the Middle Ordovician (Llandeilo) of Normandy, France, and a reinterpretation of Milonicystis kerfornei. Palaeontographica A 232, 99^132. Crowley, Q.G., Floyd, P.A., Winchester, J.A., Franke, W., Holland, J.G., 2000. Early Palaeozoic rift-related magmatism in Variscan Europe: fragmentation of the Armorican terrane assemblage. Terra Nova 12, 171^180. Daley, P.E.J., 1992a. The anatomy of the solute Girvanicystis batheri (?Chordata) from the Upper Ordovician of Scotland and a new species of Girvanicystis from the Upper Ordovician of South Wales. Zool. J. Linn. Soc. 105, 353^375. Daley, P.E.J., 1992b. Two new cornutes from the Lower Ordovician of Shropshire and southern France. Palaeontology 35, 127^148. Daley, P.E.J., 1995. Anatomy, locomotion and ontogeny of the solute Castericystis vali from the Middle Cambrian of Utah. Geobios 28, 585^615. Daley, P.E.J., 1996. The ¢rst solute which is attached as an

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97 adult: Mid-Cambrian fossil from Utah with echinoderm and chordate a⁄nities. Zool. J. Linn. Soc. 117, 405^440. David, B., Mooi, R., 1998. Major events in the evolution of echinoderms viewed by the light of embryology. In: Mooi, R., Telford, M. (Eds.), Echinoderms: San Francisco. Balkema, Rotterdam, 21^28. David, B., Mooi, R., 1999. Comprendre les e¤chinodermes: la contribution du mode'le extraxial^axial. Bull. Soc. ge¤ol. Fr. 170, 91^101. David, B., Lefebvre, B., Mooi, R., Parsley, R., 2000. Are homalozoans echinoderms? An answer from the extraxial^axial theory. Paleobiology 26, 529^555. Derstler, K., 1981. Morphological diversity of Early Cambrian echinoderms. In: Taylor, M.E. (Ed.), Short Papers 2nd Int. Symp. Cambrian System, USGS, 1981, pp. 71^75. Dom|¤nguez-Alonso, P., 1998. The early evolution of echinoderms: the class Ctenocystoidea and its closest relatives revisited. In: Candia Carnevali, M.D., Bonasoro, F. (Eds.), Echinoderm Research. Balkema, Rotterdam, pp. 263^268. Dom|¤nguez, P., Gutie¤rrez-Marco, J.C., 1990. Primeros representantes ibe¤ricos del ge¤nero Anatifopsis Barrande, 1872 (Homalozoa, Stylophora; Ordov|¤cico) y su posicio¤n sistema¤tica. Acta geol. Salmaticensia 168, 121^131. Donovan, S.K., 1984. Ramseyocrinus and Ristnacrinus from the Ordovician of Britain. Palaeontology 27, 623^634. Donovan, S.K., 1986. Pelmatozoan Columnals from the Ordovician of the British Isles, Part 1. Palaeontol. Soc. Monogr., pp. 1^68. Donovan, S.K., Paul, C.R.C., 1982. Lower Cambrian echinoderm plates from Comley, Shropshire, England. Geol. Mag. 119, 611^614. Donovan, S.K., Savill, J.J., 1988. Ramseyocrinus (Crinoidea) from the Arenig of Morocco. J. Paleontol. 62, 283^285. Dornbos, S.Q., Bottjer, D.J., 2000. Evolutionary paleoecology of the earliest echinoderms: helicoplacoids and the Cambrian substrate revolution. Geology 28, 839^842. Dornbos, S.Q., Bottjer, D.J., 2001. Taphonomy and environmental distribution of helicoplacoid echinoderms. Palaios 16, 197^204. Droser, M.L., Sheehan, P.M., 1997. Palaeoecology of the Ordovician radiation: resolution of large-scale patterns with individual clade histories, palaeogeography and environments. Geobios Me¤m. Spe¤c. 20, 221^229. Durham, J.W., 1971. The fossil record and the origin of the Deuterostomata. Proc. N. Am. Paleontol. Conv. 2, 1104^ 1132. Dzik, J., 1999. Evolutionary origin of asymmetry in early metazoan animals. In: Palyi, G., Zucchi, C., Caglioti, L. (Eds.), Advances in Biochirality. Elsevier Science, pp. 153^ 190. Fatka, O., Kordule, V., 1984. Acanthocystites Barrande, 1887 (Eocrinoidea) from the Jince Formation (Middle Cambrian) Ł str. u¤st. geol. 59, 299^302. of the Barrandian area. Ve¤st. U Fatka, O., Kordule, V., 1985. Etoctenocystis bohemica gen. et sp. nov., new ctenocystoid from Czechoslovakia (EchinoderŁ str. u¤st. geol. 60, 225^229. mata, Middle Cambrian). Ve¤st. U Fatka, O., Kordule, V., 1990. Vyscystis ubaghsi gen. et sp.

93

nov., imbricate eocrinoid from Czechoslovakia (EchinoderŁ str. u¤st. geol. 65, 315^ mata, Middle Cambrian). Ve¤st. U 320. Fatka, O., Kordule, V., 1991. Akadocrinus knizeki sp. nov., gogiid eocrinoid from Czechoslovakia (Echinodermata, Ł str. u¤st. geol. 66, 239^244. Middle Cambrian). Ve¤st. U Fatka, O., Kordule, V., 2001. Asturicystis Havl|¤ceki sp. nov. (Echinodermata, Homostelea) from the Middle Cambrian of Bohemia. J. Czech Geol. Soc. 46, 189^193. Fell, H.B., 1963a. A new family and genus of Somasteroidea. Trans. R. Soc. N. Z. 3, 143^146. Fell, H.B., 1963b. The phylogeny of sea-stars. Philos. Trans. R. Soc. Lond. 246, 381^435. Fortey, R.A., Owens, R.M., 1991. A trilobite fauna from the highest Shineton Shales in Shropshire, and the correlation of the latest Tremadoc. Geol. Mag. 128, 437^464. Friedrich, W.P., 1993. Systematik und Funktionsmorphologie mittelkambrischer Cincta (Carpoidea, Echinodermata). Beringeria 7, 3^190. Friedrich, W.P., 1995. Neue Nachweise mittelkambrischer Cincta (Carpoidea, Echinodermata) aus Marokko, Sardinien und Su«d-Wales. Beringeria 2, 255^269. Gehling, J.G., 1987. Earliest known echinoderm ^ a new Ediacaran fossil from the Pound Subgroup of South Australia. Alcheringa 11, 337^345. Gigout, M., 1954. Sur un he¤te¤roste¤le¤ de l’Ordovicien marocain. Bull. Soc. Sci. Nat. Phys. Maroc 34, 3^7. Gil Cid,.M.D., Dom|¤nguez-Alonso, P., 1995. Gyrocystis cruzae, una nueva especie de Cincta (Echinodermata Carpoidea) del Ca¤mbrico Medio de El Ferredal de Quintana (Asturias, Espan‹a). Bol. Geol. Min. 106, 517^531. Gil Cid, M.D., Dom|¤nguez-Alonso, P., 1998. ‘Carpoidea’ and Pelmatozoa from the Middle Cambrian of Zafra (SW Spain). In: Mooi, R., Telford, M. (Eds.), Echinoderms: San Francisco. Balkema, Rotterdam, pp. 93^98. Gil Cid, M.D., Dom|¤nguez-Alonso, P., 1999. Diversidad de equinodermos y carpoideos en el Ca¤mbrico Inferior y Medio de la Pen|¤nsula Ibe¤rica. Col. Paleontol. 50, 105^116. Gil Cid, M.D., Dom|¤nguez-Alonso, P., Silva¤n-Pobes, E., 1996a. Reconstruccio¤n y modo de vida de Heviacrinus melendezi nov. gen. nov. sp. (Disparida Iocrinidae), primer crinoide descrito para el Ordov|¤cico medio de los Montes de Toledo (Espan‹a). Rev. Soc. Geol. Esp. 9, 19^27. Gil Cid, M.D, Dom|¤nguez-Alonso P., Silva¤n-Pobes, E., Escribano-Rodenas, M., 1996b. Bohemiaecystis je¡eriesi n. sp.: primer Cornuta para el Ordov|¤cico espan‹ol. Est. geol. 313^ 326. Gisle¤n, T., 1927. A new Spanish carpoid. Arkiv Zoologi 19B, 1^3. Guensburg, T.E., Sprinkle, J., 1992. Rise of echinoderms in the Paleozoic evolutionary fauna: signi¢cance of paleoenvironmental controls. Geology 20, 407^410. Guensburg, T.E., Sprinkle, J., 2001a. Ecologic radiation of Cambro^Ordovician echinoderms. In: Zhuravlev, A.Y., Riding, R. (Eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York, pp. 428^444. Guensburg, T.E., Sprinkle, J., 2001b. Earliest crinoids: new

PALAEO 3074 9-5-03

94

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

evidence for the origin of the dominant Paleozoic echinoderms. Geology 29, 131^134. Gutie¤rrez-Marco, J.C., 1999. ‘Coralcrinus sarachagae gen. nov., primer crinoide (Disparida, Inadunata) descrito en el Ordov|¤cico medio de Sierra Morena’, original de M.D. Gil Cid, P. Dom|¤nguez-Alonso y E. Silva¤n-Pobes. Discusio¤n. Col. Paleontol. 50, 209^215. Gutie¤rrez-Marco, J.C., 2000. Revisio¤n taxono¤mica de ‘Echinosphaerites’ murchisoni Verneuil y Barrande, 1855 (Echinodermata, Diploporita) del Ordov|¤cico Medio centroibe¤rico (Espan‹a). Geogaceta 27, 83^86. Gutie¤rrez-Marco, J.C., Acen‹olaza, G.F., 1999. Calix inornatus (Mele¤ndez, 1958) (Echinodermata, Diploporita): morfolog|¤a de la regio¤n oral de la teca y revisio¤n bioestratigra¤¢ca. Temas Geol. Min. 26, 557^565. Gutie¤rrez-Marco, J.C., Baeza, E., 1996. Descubrimiento de Aristocystites metroi Parsley y Prokop, 1990 (Echinodermata, Diploporita) en el Ordov|¤cico medio centroibe¤rico (Espan‹a). Geogaceta 20, 225^227. Gutie¤rrez-Marco, J.C., Mele¤ndez, B., 1987. Nuevos hallazgos de Estilo¤foros (Homalozoos) en los materiales ordov|¤cicos de la zona Centroibe¤rica. Col. Paleontol. 41, 41^50. Gutie¤rrez-Marco, J.C., Aramburu, C., Arbizu, M., Berna¤rdez, E., Hacar-Rodr|¤guez, M.P., Me¤ndez-Bedia, I., MontesinosLo¤pez, R., Ra¤bano, I., Truyols, J., Villas, E., 1999. Revisio¤n bioestratigra¤phica de las pizarras del Ordov|¤cico Medio en el noroeste de Espan‹a (zonas Canta¤brica, Asturoccidental^leonesa y Centroibe¤rica septentrional). Acta Geol. Hisp. 34, 3^ 87. Gutie¤rrez-Marco, J.C., Chauvel, J., Mele¤ndez, B., Smith, A.B., 1984. Los equinodermos (Cystoidea, Homalozoa, Stelleroidea, Crinoidea) del Paleozoico Inferior de los Montes de Toledo y Sierra Morena (Espan‹a). Estud. geol. 40, 421^453. Gutie¤rrez-Marco, J.C., Mele¤ndez, B., Parsley, R.L., Prokop, R.J., Marek, L., 1992. Equinodermos (Cystoidea, Homalozoa, Asterozoa) de a¢nidades bohe¤micas en el Ordov|¤cico de las Zonas Centroibe¤rica y Ossa Morena, Espan‹a. Publ. Mus. Geol. Extremadura 1, 79^81. Havl|¤c›ek, V., 1982. Ordovician in Bohemia: development of the Prague Basin and its benthic communities. Sbor. Geol. Ved. G. 37, 103^136. Havl|¤c›ek, V., Fatka, O., 1992. Ordovician of the Prague Basin (Barrandian area, Czechoslovakia). In: Webby, B., Laurie, J. (Eds.), Global Perspectives on Ordovician Geology. Balkema, Rotterdam, pp. 461^472. Havl|¤c›ek, V., Vanek, J., 1966. The biostratigraphy of the Ordovician of Bohemia. Sbor. Geol. Ved. Paleontol. 8, 7^69. Havl|¤c›ek, V., Vanek, J., 1990. Ordovician invertebrate communities in black-shale lithofacies (Prague basin, CzechosloŁ str. u¤st. Geo. 65, 223^236. vakia). Ve¤st. U Hennigsmoen, G., 1958. Los trilobites de las capas de Saukianda, Ca¤mbrico inferior, en Andalusia. Est. geol. 14, 251^271. Henry, J.L., Lefebvre, B., Chauvin, D., 1997. Strati¢cation thermique probable des eaux marines sur la marge gondwanienne (Massif armoricain) pendant l’Ordovicien (Llanvirn): implications pale¤oge¤ographiques. N. Jb. Geol. Pala«ont. Abh. 205, 373^392.

Hicks, H., Jones, T.R., 1872. On some undescribed fossils from the Menevian Group. By Henry Hicks. With a note on the Entomostraca from the Cambrian rocks of St. David’s. By Prof. Rupert Jones. Quat. J. geol. Soc. Lond. 28, 173^185. Jaekel, O., 1918. Phylogenie und System der Pelmatozoen. Pal. Zeit. 3, 1^128. Je¡eries, R.P.S., 1968. The subphylum Calcichordata (Je¡eries 1967) primitive fossil chordates with echinoderm a⁄nities. Bull. Br. Mus. (Nat. Hist.) Geol. 16, 243^339. Je¡eries, R.P.S., 1969. Ceratocystis perneri Jaekel ^ A Middle Cambrian chordate with echinoderm a⁄nities. Palaeontology 12, 494^535. Je¡eries, R.P.S., 1973. The Ordovician fossil Lagynocystis pyramidalis (Barrande) and the ancestry of Amphioxus. Philos. Trans. R. Soc. B 265, 409^469. Je¡eries, R.P.S., 1984. Locomotion, shape, ornament, and external ontogeny in some mitrate calcichordates. J. Vert. Paleontol. 4, 292^319. Je¡eries, R.P.S., 1986. The Ancestry of the Vertebrates. Br. Mus. (Nat. Hist.)/Cambridge University Press, London/ Cambridge, 376 pp. Je¡eries, R.P.S., 1987. The chordates ^ a preliminary note. In: Fortey, R.A, Owens, R.M., (Eds.), The Arenig Series in South Wales: Stratigraphy and Palaeontology. Bull. Br. Mus. (Nat. Hist.) Geol. 41, 285^290. Je¡eries, R.P.S., Prokop, R.J., 1972. A new calcichordate from the Ordovician of Bohemia and its anatomy, adaptations and relationships. Biol. J. Linn. Soc. 4, 69^115. Je¡eries, R.P.S., Lewis, M., Donovan, S.K., 1987. Protocystites menevensis ^ a stem-group chordate (Cornuta) from the Middle Cambrian of South Wales. Palaeontology 30, 429^ 484. Josopait, V., 1970. Trilobiten-Funde im Tremadoc der Iberischen Ketten (NE Spanien). N. Jb. Geol. Pala«ontol., Mh., 127^128. Kerforne, F., 1901. Etude de la re¤gion silurique occidentale de la presqu’|“le de Crozon. Thesis, Rennes, 234 pp. Kolata, D.R., Strimple, H.L., Levorson, C.O., 1977. Revision of the Ordovician carpoid family Iowacystidae. Palaeontology 20, 529^557. Lefebvre, B., 2000a. A new mitrate (Echinodermata, Stylophora) from the Tremadoc of Shropshire (England) and the origin of the Mitrocystitida. J. Paleontol. 74, 890^906. Lefebvre, B., 2000b. Les e¤chinodermes stylophores du Massif armoricain. Bull. Soc. Sci. Nat. Ouest Fr. 22, 101^122. Lefebvre, B., in press. Functional morphology of stylophoran echinoderms. Palaeontology. Lefebvre, B., Gutie¤rrez-Marco, J.C., in press. New Ordovician mitrocystitidan mitrates (Echinodermata, Stylophora) from the Central Iberian zone (Spain). N. Jb. Geol. Pala«ont. Abh. Lefebvre, B., Vizca|«no, D., 1999. New Ordovician cornutes (Echinodermata, Stylophora) from Montagne Noire and Brittany (France) and a revision of the order Cornuta Jaekel 1901. Geobios 32, 421^458. Lefebvre, B., RachebEuf, P., David, B., 1998a. Homologies in stylophoran echinoderms. In: Mooi, R., Telford, M. (Eds.),

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97 Echinoderms: San Francisco. Balkema, Rotterdam, pp. 103^109. Lefebvre, B., Vidal, M., Henry, J.L., 1998b. Pale¤oenvironnements et biofacie's a' inverte¤bre¤s marins sur la marge nordgondwanienne pendant l’Ordovicien. Re¤sume¤s 17e'me RST Brest, 147 pp. Ł lvaro, J.J., GoLin‹a¤n, E., Villas, E., Ga¤mez-Vintaned, J.A., A zalo, R., Palacios, T., Sdzuy, K., 1996. S|¤ntesis paleontolo¤gica del Ca¤mbrico y Ordov|¤cico del Sistema Ibe¤rico (Cadenas Ibe¤ricas y Cadenas Hespe¤ricas). Rev. Esp. Paleontol., pp. 21^32. Loi, A., Pillola, G.L., Leone, F., 1995. The Cambrian and Early Ordovician of south-western Sardinia. Ren. Sem. Fac. Sci. Univ. Cagliari 65, 63^81. McKerrow, W.S., Van Staal, C.R., 2000. The Palaeozoic time scale reviewed. In: Franke, W., Haak, V., Oncken, O., Tanner D. (Eds.), Orogenic Processes: Quanti¢cation and Modelling in the Variscan Belt. Geol. Soc. Lond., Spec. Publ. 179, pp. 5^8. Miquel, J., 1912. Nouvel essai sur la ge¤ologie des terrains primaires du de¤partement de l’He¤rault, Classi¢cation des terrains siluriens. Bull. Soc. Et. Sci. Nat. Be¤ziers 33, 1^48. Mooi, R., David, B., 1997. Skeletal homologies of echinoderms. In: Waters, J.A., Maples C.G. (Eds.), Geobiology of Echinoderms. Paleontol. Soc. Pap. 3, pp. 305^335. Mooi, R., David, B., 1998. Evolution within a bizarre phylum: homologies of the ¢rst echinoderms. Amer. Zool. 38, 965^ 974. Mooi, R., David, B., 2000. What a new model of skeletal homologies tells us about asteroid evolution. Amer. Zool. 40, 326^339. Mooi, R., David, B., Marchand, D., 1994. Echinoderm skeletal homologies: classical morphology meets modern phylogenetics In: David, B., Guille, A., Feral, J.P., Roux M. (Eds.), Echinoderms through Time (Echinoderms Dijon). Balkema, Rotterdam, pp. 87^95. Parsley, R.L., 1994. Mitrocystitid functional morphology, evolution and their relationships with other primitive echinoderm classes. In: David, B., Guille, A., Fe¤ral, J.-P., Roux, M. (Eds.), Echinoderms through Time (Echinoderms Dijon). Balkema, Rotterdam, pp. 167^172. Parsley, R.L., 1998. The Cincta (Homostelea) as blastozoans. In: Candia Carnevali, M.D., Bonasoro, F. (Eds.), Echinoderm Research. Balkema, Rotterdam, pp. 369^375. Parsley, R.L., 2000. Morphological and paleoecological analysis of the Ordovician ankyroid Lagynocystis (Stylophora: Echinodermata). J. Paleontol. 74, 254^262. Parsley, R.L., Prokop, R.J., in press. Functional morphology and paleoecology of Middle Cambrian echinoderms from marginal Gondwana basins in Bohemia. 2001 Geol. Soc. America Abstracts with Programs. Parsley, R.L., Prokop, R.J., Derstler, K., 2000. Kirkocystid ankyroids (Stylophora: Echinodermata) from the Sa¤rka Ł str. u¤st. geol. Formation (Ordovician) of Bohemia. Ve¤st. U 75, 37^47. Paul, C.R.C., 1968. Macrocystella Callaway, the earliest glyptocystitid cystoid. Palaeontology 11, 580^600.

95

Paul, C.R.C., 1984. British Ordovician Cystoids, Part 2. Monogr. Palaeont. Soc. Lond., pp. 65^152. Paul, C.R.C., Cope, J.C.W., 1982. A parablastoid from the Arenig of South Wales. Palaeontology 25, 499^507. Paul, C.R.C., Smith, A.B., 1984. The early radiation and phylogeny of echinoderms. Biol. Rev. 59, 443^481. Perejo¤n, A., Fro«hler, M., Bechsta«dt, T., Moreno-Eiris, E., Boni, M., 2000. Archaeocyathan assemblages from the Gonnessa Group, Lower Cambrian (Sardinia, Italy) and their sedimentologic context. Boll. Soc. Paleont. Ital. 39, 257^291. Plas, V., Prokop, R.J., 1979. ?Argodiscus rarus sp. n. (Edrioasteroidea) from the Sa¤rka Formation (Llanvirn) of Ł str. u¤st. geol. 54, 41^43. Bohemia. Ve¤st. U Pompeckj, J.F., 1896. Die Fauna des Cambrium von Tejrovic und Skrej in Bo«hmen. J. k.-k. geol. Reich. 45, 495^614. Prigmore, J.K.A., Butler, A.J., Woodcock, N.H., 1997. Rifting during separation of Eastern Avalonia from Gondwana: evidence from subsidence analysis. Geology 25, 203^206. Prokop, R.J., 1962. Akadocrinus nov. gen., nova¤ lilijice z Ł str. u¤st. geol. 27, jinecke¤ho kambria (Eocrinoidea). Sbor. U 31^39. Prokop, R.J., 1965. Argodiscus hornyi gen. n. et sp. n. (Edrioasteroidea), Z Ceske¤ho Stredn’ho Ordoviku. Cas. Na¤rod. Muz. Odd|¤l Pr|¤rodovedny 134, 65^152. Prokop, R.J., Fatka, O., 1985. Luhocrinus monicae gen. et sp. nov. (Eocrinoidea) from the Middle Cambrian of Bohemia. Ł str. u¤st. geol. 60, 231^234. Ve¤st. U Prokop, R.J., Petr, V., 1999. Echinoderms in the Bohemian Ordovician. J. Czech Geol. Soc. 44, 63^68. Prokop, R.J., Turek, V., 1997. The crinoid holdfast Lichenocrinus Hall, 1866 in the Ordovician of Bohemia. Bull. Czech Geol. Surv. 72, 307^310. Re¤gnault, S., 1990. Morphologie et ontogene'se des Ascocystitidae (Eocrinoidea): donne¤es fournies par l’e¤tude de faunes de l’Ordovicien du Massif armoricain (France). Rev. Pale¤obiologie 9, 349^358. Richter, R., Richter, E., 1940. Die Saukianda-Stufe von Andalusien, eine fremde Fauna im europa«ischen Ober-Kambrium. Abh. Senckenberg. Natur. Gesell. 450, 1^88. Robison, R.A., Sprinkle, J., 1969. Ctenocystoidea: new class of primitive echinoderms. Science 166, 1512^1514. Rouault, M., 1851. Me¤moire sur le terrain pale¤ozo|«que des environs de Rennes. Bull. Soc. ge¤ol. Fr. 2, 358^399. Rozhnov, S.V., 2001. Evolution of the hardground community. In: Zhuravlev, A.Y., Riding, R. (Eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York, pp. 238^253. Ruta, M., 1997. A new mitrate from the Lower Ordovician of southern France. Palaeontology 40, 363^383. Schroeder, R., 1973. Carpoideen aus dem Mittelkambrium Nordspaniens. Palaeontographica A 141, 119^142. Scotese, C.R., McKerrow, W.S., 1990. Revised world maps and introduction. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. Mem. 12, pp. 1^21. Sdzuy, K., 1955. Cystoideen aus den Leimitz-Schiefer (Tremadoc). Senck. Lethaea 35, 269^276.

PALAEO 3074 9-5-03

96

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97

Sdzuy, K., 1993. Early Cincta (Carpoidea) from the Middle Cambrian of Spain. Beringeria 8, 189^207. Sdzuy, K., 2000. Das Kambrium des Frankenwaldes, 3. Die Lippertsgru«ner Schichten und ihre Fauna. Senck. Lethaea 79, 301^327. Sdzuy, K., Hammann, W., Villas, E., 2001. The Upper Tremadoc fauna from Vogtendorf and the Bavarian Ordovician of the Frankenwald. Senck. Lethaea 81, 207^261. Seilacher, A., 1999. Biomat-related lifestyles in the Precambrian. Palaios 14, 86^93. Seilacher, A., P£u«ger, F., 1994. From biomats to benthic agriculture: a biohistoric revolution. In: Krumbein, W.E., Paterson, D.M., Stal, L.J. (Eds.), Biostabilization of Sediments. Bibliotheks und Informations system der Carl von Ossietzky, Univ. Oldenburg, pp. 97^105. Selg, M., 1988. Origin of peritidal carbonate cycles: Early Cambrian, Sardinia. Sediment. Geol. 59, 115^124. Sepkoski, J.J., Jr., 1979. A kinetic model of Phanerozoic taxonomic diversity II. Early Phanerozoic families and multiple equilibria. Paleobiology 5, 222^251. Sepkoski, J.J., Jr., 1981. A factor analytic description of the Phanerozoic marine fossil record. Paleobiology 10, 246^ 267. Sepkoski, J.J., Jr., 1991. A model of onshore^o¡shore change in faunal diversity. Paleobiology 17, 58^77. Smith, A.B., 1984. Classi¢cation of the Echinodermata. Palaeontology 27, 431^459. Smith, A.B., 1985. Cambrian eleutherozoan echinoderms and the early diversi¢cation of edrioasteroids. Palaeontology 28, 715^756. Smith, A.B., 1988. Patterns of diversi¢cation and extinction in Early Palaeozoic echinoderms. Palaeontology 31, 799^828. Smith, A.B., Jell, P.A., 1990. Cambrian edrioasteroids from Australia and the origin of star¢shes. Mem. Queens. Mus. 28, 715^778. Smith, A.G., 2001. Paleomagnetically and tectonically based global maps for Vendian to Mid-Ordovician Time. In: Zhuravlev, A.Y., Riding, R. (Eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York, pp. 11^46. Spencer, W.K., 1950. Asterozoa and the study of Palaeozoic faunas. Geol. Mag. 87, 393^408. Spencer, W.K., 1951. Early Palaeozoic star¢sh. Philos. Trans. R. Soc. Lond. B 235, 12^129. Sprinkle, J., 1973. Morphology and evolution of blastozoan echinoderms. Mus. Comp. Zool. Spec. Publ., Cambridge, 284 pp. Sprinkle, J., 1981. Diversity and evolutionary patterns of Cambrian echinoderms. In: Taylor M.E. (Ed.), Short Papers 2nd Int. Symp. Cambrian System, USGS, 1981, pp. 219^221. Sprinkle, J., 1992. Radiation of Echinodermata. In: Lipps, J.H., Signor, P.W. (Eds.), Origin and Early Evolution of the Metazoa. Plenum, New York, pp. 375^398. Sprinkle, J., Collins, D.E., 1998. Revision of Echmatocrinus from the Middle Cambrian Burgess Shale of British Columbia. Lethaia 31, 269^282. Sprinkle, J., Guensburg, T.E., 1995. Origin of echinoderms in

the Paleozoic Evolutionary Fauna: the role of substrates. Palaios 10, 437^453. Sprinkle, J., Guensburg, T.E., 1997. Early radiation of echinoderms. In: Waters, J.A., Maples, C.G. (Eds.), Geobiology of Echinoderms. Paleontol. Soc. Pap. 3, pp. 205^224. Sumrall, C.D., 1997. The role of fossils in the phylogenetic reconstruction of Echinodermata. In: Waters, J.A., Maples C.G. (Eds.), Geobiology of Echinoderms. Paleontol. Soc. Pap. 3, pp. 267^288. Sumrall, C.D., 2000. The biological implications of an edrioasteroid attached to a pleurocystitid rhombiferan. J. Paleontol. 74, 67^71. Sumrall, C.D., Sprinkle, J., 1999. Ponticulocarpus, a new cornute-grade stylophoran from the Middle Cambrian Spence Shale of Utah. J. Paleontol. 73, 886^891. Sumrall, C.D., Sprinkle, J., Guensburg, T.E., 1997. Systematics and paleoecology of Late Cambrian echinoderms from the western United States. J. Paleontol. 71, 1091^1109. Termier, G., Termier, H., 1950a. Contribution a' l’e¤tude des faunes pale¤ozo|«ques de l’Alge¤rie. Bull. Serv. carte ge¤ol. Alge¤rie, 1e're se¤r. 11, 1^49. Termier, G., Termier, H., 1950b. Pale¤ontologie marocaine ^ Tome 2 ^ Inverte¤bre¤s de l’Ere primaire. fasc. 4 ^ Anne¤lides, Arthropodes, Echinodermes, Conularides et Graptolithes. Notes Me¤m. Serv. ge¤ol. Maroc 79, 1^149. Termier, H., Termier, G., 1969. Les Stromatocystito|«des et leur descendance. Essai sur l’e¤volution des premiers e¤chinodermes. Geobios 2, 131^156. Termier, H., Termier, G., 1973. Les e¤chinodermes Cincta du Cambrien de la Montagne Noire (France). Geobios 6, 243^ 266. Thoral, M., 1935. Contribution a' l’e¤tude pale¤ontologique de l’Ordovicien infe¤rieur de la Montagne Noire et re¤vision sommaire de la faune cambrienne de la Montagne Noire. Imprimerie de la Charite¤, Montpellier, 362 pp. Turek, V., 1990. Z novych p`|¤ru'stku' paleontologicke¤ho odd|'len|¤ Na¤rodn|¤ho Muzea v Praze. Cas. Na¤rod. Muz. Odd|¤l Pr|¤rodovedny 155, 136^140. Ubaghs, G., 1953. Notes sur Lichenoides priscus Barrande, e¤ocrino|«de du Cambrien moyen de la Tche¤coslovaquie. Bull. Inst. r. Sci. nat. Belg. 29, 1^24. Ubaghs, G., 1960. Le genre Lingulocystis Thoral (Echinodermata, Eocrinoidea). Ann. Pale¤ontol. 46, 81^116. Ubaghs, G., 1961. Un e¤chinoderme nouveau de la classe des carpo|«des dans l’Ordovicien infe¤rieur du de¤partement de l’He¤rault (France). C. R. Acad. Sci. 253, 2565^2567. Ubaghs, G., 1963a. Rhopalocystis destombesi n.g., n.sp., e¤ocrino|«de de l’Ordovicien infe¤rieur (Tre¤madocien infe¤rieur) du Sud marocain. Notes Serv. ge¤ol. Maroc 23, 25^45. Ubaghs, G., 1963b. Cothurnocystis Bather, Phyllocystis Thoral, and an undetermined member of the order Soluta (Echinodermata, Carpoidea) in the Uppermost Cambrian of Nevada. J. Paleontol. 37, 1133^1142. Ubaghs, G., 1967a. Eocrinoidea. In: Moore, R.C. (Ed.), Treatise on Invertebrate Paleontology, Part S, Echinodermata 1. Geol. Soc. Am./University of Kansas Press, Lawrence, pp. S455^S495.

PALAEO 3074 9-5-03

B. Lefebvre, O. Fatka / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 73^97 Ubaghs, G., 1967b. Stylophora. In: Moore, R.C. (Ed.), Treatise on Invertebrate Paleontology, Part S, Echinodermata 1. Geol. Soc. Am./University of Kansas Press, Lawrence, pp. S495-S565. Ubaghs, G., 1967c. Homostelea. In: Moore, R.C. (Ed.), Treatise on Invertebrate Paleontology, Part S, Echinodermata 1. Geol. Soc. Am./University of Kansas Press, Lawrence, pp. S565-S581. Ubaghs, G., 1967d. Le genre Ceratocystis Jaekel (Echinodermata, Stylophora). Univ. Kansas Paleontol. Contrib. 22, 1^16. Ubaghs, G., 1969a. Les e¤chinodermes ‘carpo|«des’ de l’Ordovicien infe¤rieur de la Montagne Noire (France). Cahiers de pale¤ontologie, Ed. CNRS, Paris, 110 pp. Ubaghs, G., 1969b. Aethocrinus moorei Ubaghs, n. gen., n. sp., le plus ancien crino|«de dicyclique connu. Univ. Kansas Paleontol. Contrib. 38, 1^25. Ubaghs, G., 1972. Le genre Balantiocystis Chauvel (Echinodermata, Eocrinoidea) dans l’Ordovicien infe¤rieur de la Montagne Noire (France). Ann. Pale¤ontol. 58, 3^26. Ubaghs, G., 1975. Early Paleozoic echinoderms. Ann. Rev. Earth Planet. Sci. 3, 79^98. Ubaghs, G., 1983. Echinodermata. Notes sur les e¤chinodermes de l’Ordovicien infe¤rieur de la Montagne Noire (France). In: Courtessole, R., Marek, L., Pillet, J., Ubaghs, G., Vizca|«no, D. (Eds.), Calymenina, Echinodermata et Hyolitha de l’Ordovicien infe¤rieur de la Montagne Noire. Me¤m. Soc. Et. Sci. Aude 3, pp. 33^35. Ubaghs, G., 1987. Echinodermes nouveaux du Cambrien moyen de la Montagne Noire (France). Ann. Pale¤ontol. 73, 1^27. Ubaghs, G., 1991. Deux Stylophora (Homalozoa, Echinodermata) nouveaux pour l’Ordovicien infe¤rieur de la Montagne Noire (France me¤ridionale). Pala«ontol. Z. 65, 157^171. Ubaghs, G., 1994. Echinodermes nouveaux (Stylophora, Eocrinoidea) de l’Ordovicien infe¤rieur de la Montagne Noire (France). Ann. Pale¤ontol. 80, 107^141.

97

Ubaghs, G., 1998. Echinodermes nouveaux du Cambrien supe¤rieur de la Montagne Noire. Geobios 31, 809^829. Ubaghs, G., Robison, R.A., 1985. A new homoiostelean and a new eocrinoid from the Middle Cambrian of Utah. Univ. Kansas Paleontol. Contrib. 115, 1^24. Ubaghs, G., Vizca|«no, D., 1990. A new eocrinoid from the Lower Cambrian of Spain. Palaeontology 33, 249^256. Vizca|«no, D., Lefebvre, B., 1999. Les e¤chinodermes du Pale¤ozo|«que infe¤rieur de Montagne Noire: biostratigraphie et pale¤odiversite¤. Geobios 32, 353^364. Ł lvaro, J.J., Lefebvre, B., 2001. The Lower OrVizca|«no, D., A dovician of the Southern Montagne Noire. Ann. Soc. ge¤ol. Nord 2, 213^220. Waisfeld, B.G., Sa¤nchez, T.M., 1996. ‘Fauna Ca¤mbrica’ versus ‘Fauna Paleozoica’ en el Ordov|¤cico temprano del Oeste de Argentina: interaccio¤n entre provincialismo y ambiente. Geobios 29, 401^416. Wolf, R., 1980. The lower and upper boundary of the Ordovician system of some selected regions (Celtiberia, Eastern Sierra Morena) in Spain. N. Jb. Geol. Pala«ontol. Abh. 160, 1^118. Woods, I.S., Je¡eries, R.P.S., 1992. A new stem-group chordate from the Lower Ordovician of South Wales, and the problem of locomotion in boot-shaped cornutes. Palaeontology 35, 1^25. Xiao, S., Knoll, A.H., 2000. Phosphatized animal embryos from the Neoproterozoic Doushantuo Formation at Weng Oan, Guizhou, South China. J. Paleontol. 74, 767^788. Zel|¤zko, J.V., 1911. Zaj|¤mave¤ zbytky crinoidu' ze spodn|¤ho siluru od Ejpovic (Interesting crinoidal rests in the Lower Silurian at Ejpovice). Sborn. M|'st. Hist. Mus. v Plzni 2, 1^3. Zhuravlev, A.Y., 2001. Biotic diversity and structure during the Neoproterozoic^Ordovician transition. In: Zhuravlev, A.Y., Riding, R. (Eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York, pp. 173^ 199.

PALAEO 3074 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124 www.elsevier.com/locate/palaeo

The Ireviken Event in the lower Silurian of Gotland, Sweden ^ relation to similar Palaeozoic and Proterozoic events Axel Munnecke a; , Christian Samtleben b , Torsten Bickert c b

a Institut fu«r Pala«ontologie, Universita«t Erlangen, LoewenichstraMe 28, D-91054 Erlangen, Germany Institut fu«r Geowissenschaften (Geologie), Universita«t Kiel, OlshausenstraMe 40-60, D-24118 Kiel, Germany c Fachbereich Geowissenschaften, Universita«t Bremen, D-28334 Bremen, Germany

Received 29 March 2002; received in revised form 26 August 2002; accepted 15 January 2003

Abstract For a long time, the Silurian was thought to represent a time of stable environmental conditions in the greenhouse period that followed the Late Ordovician glaciation. During the past decade, knowledge about the Silurian has increased markedly and today it is known that the conditions in the Silurian were much more variable than previously assumed. Detailed isotopic investigations have revealed several distinct positive excursions in both carbon and oxygen isotope values. In low latitudes, these periods of high C- and O-isotope values are in many cases characterised by the growth of reefs and the formation of extended carbonate platforms. The sediments deposited during these excursions contain impoverished fossil assemblages, especially with respect to conodonts, graptolites, and trilobites. A conspicuous isotope excursion coincident with facies changes and a marked mass extinction is observed near the Llandovery/Wenlock boundary. This event is called the ‘Ireviken Event’ after its type locality on the island of Gotland, Sweden (Jeppsson, L., 1987. In: Palaeobiology of Conodonts. Ellis Horwood Ltd., Chichester, pp. 129^145). Here, isotope data from nine sections at the NW coast of Gotland are presented that cover the time interval of the Ireviken Event. The N13 C mean values rise from +1.4x to +4.5x, and the N18 O values increase from 35.6x to 35.0x. The relative timing of stable isotope development, extinctions, and facies development is discussed. It is shown that first extinctions precede the isotope excursion. This indicates that extinction events and stable isotope development are only indirectly connected but might reflect the same causes. Other events characterised by similar relationships between positive isotope excursions, mass extinctions, and facies development are found in younger parts of the Silurian (late Wenlock and Ludlow), in the late Ordovician, the late Cambrian, and, with some reservations, in the Proterozoic. The similarities between these events indicate analogous controlling mechanisms. For the Silurian, climatic changes between humid and arid conditions in low latitudes were postulated in an earlier study of the authors. A palaeoceanographic/climatic model was postulated which is consistent with most of the sedimentological, palaeontological, and geochemical data. Here we demonstrate that this model may be applicable also for the older events. B 2003 Elsevier Science B.V. All rights reserved. Keywords: Cambrian; Ordovician; Silurian; stable isotope; mass extinction; bioevents

* Corresponding author. Tel.: +49-9131-8526957; Fax: +49-9131-8522690. E-mail addresses: [email protected] (A. Munnecke), [email protected] (C. Samtleben), [email protected] (T. Bickert).

0031-0182 / 03 / $ ^ see front matter B 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00304-3

PALAEO 3075 9-5-03

100

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

1. Introduction Previously, the Silurian was thought to be an epoch characterised by high global sea level (Johnson et al., 1991), stable climatic greenhouse conditions (Fischer, 1983), cosmopolitan faunas, and by an absence of signi¢cant orogenies (for a review see Bassett and Edwards, 1991). Large shallow epicontinental seas were widely distributed, and the continents had a low relief (Scotese and McKerrow, 1990; Cocks and Scotese, 1991). Terrestrial plants were nearly absent and thus had no in£uence on the global carbon cycle. Major extinction events comparable to those of Ordovician or Devonian times were unknown (Boucot, 1991), and, except for the Malvinoka¡ric realm, reefs were widely distributed (Copper and Brunton, 1991; Copper, 2002; Kiessling, 2002). With increasing knowledge, this view of the Silurian changed. Major progress was achieved, e.g., by detailed conodont stratigraphy allowing platform^basin correlations. Today, several extinction events are known from the Silurian, in particular a¡ecting conodonts, graptolites, trilobites, acritarchs, chitinozoans, ostracods, brachiopods, and corals (Le He¤risse¤, 1989; Melchin, 1994; Jeppsson, 1997a,b,c, 1998; Jeppsson et al., 1994, 1995; Kaljo et al., 1995; Berry, 1998; Nestor, 1998; Mikulic and Kluessendorf, 1999; Dorning, 1999; Jeppsson and Calner, 2003). Jeppsson (1984, 1990) observed distinct conodont cycles in Silurian strata which he related to alternating periods of humid climate in low latitudes and cold polar regions (P episodes) and periods of arid climate in low latitudes and somewhat warmer polar regions (S episodes). Global oceanic circulation patterns were proposed to change accordingly. Based on isotope measurements of brachiopod shells the authors of the present study similarly postulated di¡erent climatic periods in the Silurian characterised by low or high isotope values (Samtleben et al., 1996), and reconstructed the palaeoceanographic conditions during humid ‘H-periods’ and arid ‘A-periods’ (Bickert et al., 1997). Similar to the Silurian, coincidence between stable isotope development, facies changes, and bioevents were observed in the late Ordovician (Ash-

gill), in the late Cambrian, and, less unequivocally, also in the Proterozoic (Brenchley et al., 1994, 1995; Berry et al., 1995; Samtleben et al., 1996, 2000; Wenzel and Joachimski, 1996; Glumac and Walker, 1998; Saltzman et al., 1998, 2000; Finney et al., 1999; Kump et al., 1999; Melezhik et al., 1999). The late Ordovician event represents one of the ‘big ¢ve’ mass extinctions in Earth history (Raup and Sepkoski, 1982). One conspicuous Silurian event located close to the Llandovery/Wenlock boundary (Jeppsson, 1987; Talent et al., 1993; Samtleben et al., 1996) is termed the ‘Ireviken Event’ after its type locality on Gotland (Jeppsson, 1987). During this time interval, 80% of the conodont species and 50% of the trilobite species became extinct (Jeppsson, 1987, 1997a,c). Acritarchs, graptolites, chitinozoans, corals, and brachiopods were also a¡ected (Bassett and Cocks, 1974; Laufeld, 1974; Helfrich, 1980; Le He¤risse¤, 1989; Melchin, 1994; Kaljo et al., 1995; Jeppsson, 1998; Nestor, 1998; Gelsthorpe, 2002). Some of these groups show a stepwise extinction. The conodont fauna development at the Llandovery/Wenlock boundary, for example, shows 10 separate conodont extinction steps (‘datum points’), and can be subdivided into ¢ve conodont zones. Most of the zones can be observed world-wide and be used for detailed stratigraphic correlations especially in carbonate platform environments (Jeppsson et al., 1994; Jeppsson, 1997a, 1998). In order to better understand the climatic and palaeoecologic processes during the transition from H-periods (with low C/O isotope values) to A-periods (with high C/O isotope values) it is necessary to determine the exact relative timing of extinction events, isotopic changes, and facies development. One of the thickest and stratigraphically most complete sections in the world spanning the Llandovery/Wenlock boundary is located at the NW coast of Gotland, Sweden, where the Lower and Upper Visby Beds are exposed for nearly 60 km along the coastal cli¡ (Jeppsson, 1997a). In this paper isotope data from nine sections (Fig. 1) and composite N13 C- and N18 O-curves covering the time interval under consideration are presented, and the relationships between isotopic development, facies

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

101

changes, and palaeontological events are examined. Similar relationships between positive isotope excursions, palaeontological events (extinctions), and facies development are found in younger parts of the Silurian, in the early Palaeozoic (Cambrian, Ordovician), and in the Proterozoic, indicating similar controlling mechanisms. The similarities as well as possible causes of the extinction events are discussed.

2. Palaeontological events at the Llandovery/ Wenlock boundary on Gotland Fig. 1. Sample localities on Gotland.

At the Llandovery/Wenlock boundary on Gotland, the most detailed biostratigraphic framework is based on investigations of conodonts

Fig. 2. Synoptic compilation of lithostratigraphy, biostratigraphy, and climatic stages (after Jeppsson and Ma«nnik, 1993; Aldridge et al. (1993), Jeppsson, 1997a,c; this work).

PALAEO 3075 9-5-03

Upper Visby Beds

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

Phaulactis layer

Phaulactis layer

a

Lower Visby Beds

102

Phaulactis layer

b

c

d

e

f

g

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

(Jeppsson, 1984, ¡). During the Ireviken Event an abundant and diverse pre-event conodont fauna of the late Telychian was reduced to an impoverished and low-diverse post-event fauna in the early Sheinwoodian. This change is interpreted to re£ect a decrease in hemi-pelagic productivity (Jeppsson, 1990, 1997a,c, 1998; Jeppsson et al., 1994). In most sections in the world the stepwise reduction (datum 1 to 8) is documented in strata of less than 1 m thickness (Jeppsson, 1997c). On Gotland, this time interval is represented by about 13 m in the type locality ‘Ireviken 3’ (Fig. 2): the uppermost 4.5 m of the Lower Visby Beds and the entire Upper Visby Beds (about 8.5 m) (Jeppsson and Ma«nnik, 1993; Jeppsson, 1997c). Due to minor faulting in the type locality of the Llandovery/ Wenlock boundary (Hughley Brook, Shropshire, UK; Aldridge et al., 1993) the exact position of the boundary on Gotland cannot be determined. Most probably its position is slightly above datum 2 (Fig. 2; Jeppsson, 1997c). The strongest reduction of conodont faunas is observed at datum 2 when common and widespread species became extinct (Jeppsson, 1997a,c). Trilobites were also strongly a¡ected at or close to this level (Aldridge et al., 1993). The second-most intense reduction occurred at datum 4 at the boundary between Lower and Upper Visby Beds. But, as only conodont species were affected which are rare in the palaeontological record, this event is less conspicuous (Jeppsson, 1997c). Brachiopods, ostracods, and corals were also slightly reduced at datum 4. Graptolites show a signi¢cant extinction at the Llandovery/Wenlock boundary (Melchin, 1994). Due to the sparse occurrence of graptolites on Gotland the exact position of the extinction level is di⁄cult to determine. Probably, it corresponds to the Upper Pseudooneotodus bicornis Zone (Jeppsson, 1997a). Acritarch data indicate a signi¢cant turnover in the phytoplankton communities during the Ireviken Event (Le He¤risse¤, 1989; Gelsthorpe, 2002).

103

Most of the extinctions occurred at the end of the event (in the upper part of the Upper Visby Beds), when many of the conodont extinctions took place. Impoverished acritarch communities are observed in the overlying Ho«gklint Beds.

3. Material and methods In the present study brachiopod samples for isotope measurements (N13 C, N18 O) were collected from the Lower and Upper Visby Beds in nine sections along the NW coast of Gotland (Fig. 1). Due to the gentle dip of the strata (less than 1‡) some outcrops are accessible for long horizontal distances (up to 200 m), but vertical accessibility is limited. The thickness of the individual strata varies considerably, generally with increasing values from NE to SW (Jeppsson, 1997c). For example, the thickness of the Upper Ps. bicornis Zone increases from 1.77 m at Ireviken 3 to 4.20 m at Buske which is located 34 km SW of Ireviken. The thickness of the Lower Pterospathodus procerus Zone increases from 1.33 m to 3.80 m between those two localities (Jeppsson, 1997a). Frequently, the boundary between Lower and Upper Visby Beds is characterised by abundant specimens of the large, horn-shaped rugose coral Phaulactis sp. (Plate II) (Samtleben et al., 1996; Jeppsson, 1997a,c). The positions of the sampled brachiopods are indicated relative to the base of this so-called Phaulactis layer. In Nyrevsudde, the southwestern-most locality (Fig. 1), the base of the Phaulactis layer was not determined since it is situated below sea level. Instead, a distinct limestone bed at the top of the Nyrevsudde pro¢le was chosen as the reference horizon. In a total of 253 diagenetically unaltered brachiopods stable isotopes were measured. Of these, 233 samples were taken from the species group of Atrypa reticularis. Brachiopods were chosen for

Plate I. (a) Stavklint: boundary of Lower and Upper Visby Beds; (b,c) Ygne: boundary of Lower and Upper Visby Beds, discernible by the change from regular to irregular limestone^marl alternation; (d) Buske: marl-dominated, regular limestone^marl alternation of the Lower Visby Beds; (e) Fridhem: small reef mound in the Upper Visby Beds; (f) Ygne: detail of the Lower Visby Beds; (g) Hallshuk: irregular limestone^marl alternation of the Upper Visby Beds.

PALAEO 3075 9-5-03

104

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

growth reaction

b

1 cm

a 2 cm

growth banding (years?)

Phaulactis sp.

Phaulactis layer

d

c

e

PALAEO 3075 9-5-03

f

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

isotope measurements because they are known to be the most suitable fossils for stable isotope analyses. They are assumed to record carbon and oxygen isotopic composition of sea water at the time of shell growth with an accuracy of U 0.4x (Samtleben et al., 2001). In order to remove micritic material attached to the tests, the brachiopods were cleaned manually and in an ultrasonic bath. Samples for isotope analyses were scratched o¡ with a scalpel under a binocular microscope. Material from the secondary shell layer was taken for isotope analysis. Between 60 and 80 Wg of homogenised carbonate material from each sample was measured using a Finnigan MAT 251 micromass-spectrometer coupled with a Finnigan automated carbonate preparation device at the University of Bremen. The carbonate was reacted with orthophosphoric acid at 75‡C. The reproducibility of the measurements, as referred to an internal carbonate standard (Solnhofen limestone), is U 0.07x for oxygen and U 0.05x for carbon isotopes (1c over a 1 year period). The conversion to the VPDB scale is performed using the international standard NBS 19.

4. Results 4.1. Lower and Upper Visby Beds The Lower Visby Beds are the oldest strata exposed on Gotland. They consist of at least 12 m of fossil-poor, regular alternations of 2^5-cmthick, wavy-bedded to nodular argillaceous limestones (predominantly mudstones) and about 10cm-thick marls (Plate Ib^d,f). The base of the Lower Visby Beds is not exposed. On the average, the carbonate contents of the marls scatter around 20%, those of the limestones around 70% (Munnecke, 1997). In some areas, up to

105

1-m thick Halysites-biostromes are observed that spread to a few hundred square metres in area. Thin layers of brachiopod and bryozoan debris are intercalated irregularly in the Lower Visby Beds. The sequence was deposited below storm wave base and below the photic zone in a distal shelf environment. The upper boundary of the Lower Visby Beds is marked by the extinction of the small rugose coral Palaeocyclus porpita. This index fossil for the Lower Visby Beds usually is not very abundant, therefore an accurate de¢nition of the boundary in the ¢eld is di⁄cult. However, at this boundary, which sometimes shows a thin layer enriched in oxidised pyrite, the large, solitary rugose coral Phaulactis sp. is very abundant (Samtleben et al., 1996; Jeppsson, 1997a,c). At several localities a conspicuous Phaulactis layer is present with locally in excess of 10 specimens per square metre (Plate IId^f). Phaulactis sp. is a horn-shaped coral, generally 10^30 cm in length (maximum 50 cm). The corals apparently grew in a reclining position, as indicated by their asymmetrical elliptic cross-section and by two specimens that grew against one another clearly in such a position (Plate IIa). Many specimens show growth banding most likely indicating rapid growth of up to 2 cm per year (Plate IIa). In the present work, the base of the Phaulactis layer was used as reference level for the isotope samples. The thickness of the Phaulactis layer increases from NE to SW, for example, from about 10 cm at the locality Ha«ftingsklint (Plate IId) to about 20^25 cm at Nygardsba«ckspro¢len (Plate IIf ; Fig. 1). In the southwesternmost section at Nyrevsudde the Phaulactis layer is at least 1.6 m thick, but specimens of this coral occur only sporadically (Plate IIc). The base of the layer is situated below sea level. Therefore its exact thickness could not be determined.

Plate II. (a) Two specimens of Phaulactis sp. from Ygne grown together (showing distinct growth banding). The growth reaction of the right specimen indicates that it has grown into the calyx of the dead left specimen in a reclining position. (b) Palaeocyclus porpita, index fossil of the Lower Visby Beds; (c) Nyrevsudde: single specimen of Phaulactis sp. (arrow). Here, in the southwestern-most locality (cf. Fig. 1), the Phaulactis layer is at least 1.6 m thick but is only sporadically occupied with corals (arrow). (d) Ha«ftingsklint: thin Phaulactis layer with high abundance of corals; (e) Lickershamn: top view of the Phaulactis layer with high abundance of irregularly distributed corals; (f) Nygardsba«ckspro¢len: Phaulactis layer, 20^30 cm thick.

PALAEO 3075 9-5-03

106

(m)

Storbrut Lundsklint .. Ronnklint

11

Upper Visby Beds

10 9 8 7 6

Buske 1 Nygardsbäcksprofilen 2

Nyrevsudde

5

Häftingsklint 5b 4

Ireviken1

(m)

(m)

-1

-3 2,0

2,5 3 ,0 3,5 (‰)

a/b Phaulactis layer

1

1

0

0

0

0

-1

-1

-2

-2

-3

-3

-3

-4

-4

-4

-4

-5

-5

-5

-5

-6

-6

-7

-7

1,5 2,0

e

d c

2,5

(‰)

1,0

1,5 2,0 (‰)

stratigraphic division after Jeppsson

e

-1

-8

-8

d c

(m)

2,5

3,0

2

(m)

2

1

(m)

Lower Visby Beds

PALAEO 3075 9-5-03

2

-2

3

c

Stavklint 1

0

-1

d c

-2

-2

d

-3

b

-6

c a/b e

-7

-8

-8

c

e

-2 2,0

2,5

3,0

3,5

(‰)

b

-9

-9

-10

-10

-11

-11 1,0

-12

0 -1

1,5

-6

-7

1

1,5 2,0

2,5

(‰) estimated stratigraphic division

1,0

1,5

2,0

2,5

3,0

3,5

4,0

4,5

(‰)

Fig. 3. N13 C values from brachiopod shells of the nine pro¢les investigated (stratigraphic division after Aldridge et al. (1993) and Jeppsson (1997a,c); cf. Fig. 2). Except for pro¢le 1 the height is given as vertical distance from the base of the Phaulactis layer (shaded). In Nyrevsudde (pro¢le 1) a distinct bioclastic limestone bed is taken as reference because the base of the Phaulactis layer is not exposed.

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

δ13 C

Storbrut Lundsklint Rönnklint (m) 11

9 8 7 6

Buske 1 Nygardsbäcksprofilen 2

Nyrevsudde (m) 0

Stavklint 1

(m)

-6,0 -5,5 -5,0 -4,5 (‰)

Phaulactis layer

0

0

0

-2

-3

-3

-3

-4

-4

-4

-5

-5

-5

-6

-6

-6

-7

-7

e

d

-8 -6,5

e

-1

-1

c -6,0 -5,5 -5,0 (‰)

d c

1

-2

(‰)

(m) (m)

1

-1

-8 -6,5 -6,0 -5,5 -5,0

Ireviken1

2

a/b

1

(m) 0

-3

Häftingsklint 5b

3

c

2

-2

5 4

-1

Lower Visby Beds

PALAEO 3075 9-5-03

Upper Visby Beds

10

d c

-2

2

c a/b e

-1 -2

d

-3

1 0

e

-1 -2

-6,5 -6,0 -5,5 -5,0 -4,5

c

-4

(‰)

-5

b

-7

-8

-8

-9

-9

-10

-10

-11

-11

stratigraphic division after Jeppsson

b

-6

-7

- 6 , 5 -6,0 -5,5 -5,0 -4,5 -4,0

-12

(‰)

estimated stratigraphic division

-6 , 5 -6,0 -5,5 -5,0 -4,5 -4,0 (‰)

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

δ18 O

Fig. 4. N18 O values from brachiopod shells of the nine pro¢les investigated (stratigraphic division after Aldridge et al. (1993) and Jeppsson (1997a,c); cf. Fig. 2). Except for pro¢le 1, the height is given as vertical distance from the base of the Phaulactis layer (shaded). In Nyrevsudde (pro¢le 1) a distinct bioclastic limestone bed is taken as a reference because the base of the Phaulactis layer is not exposed. 107

108

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

The boundary to the overlying Upper Visby Beds is characterised by a weak but abrupt facies change at the base of the Phaulactis layer (Plate Ia^c), marked by a distinct angle in weathered sections (Plate Ia,b). The Upper Visby Beds reach 12 m in thickness. Bedding is not as regular as in the Lower Visby Beds, and single beds cannot be traced over more than a few metres (Plate Ic). The limestone^marl ratio increases, and detritic limestones (wacke- to grainstones) become more abundant (Plate Ig), especially in the upper part of the Upper Visby Beds. Erosional surfaces and ripple marks point to increased water energy. Carbonate content is considerably higher than in the Lower Visby Beds, averaging 80% for limestones and 40% for marls (Munnecke, 1997). The abundance of sessile shells increases markedly, especially of brachiopods, bryozoans, crinoids, tabulate corals, and stromatopores. For the ¢rst time in the stratigraphic record of Gotland calcareous algae appear. Reef mounds of di¡erent sizes and shapes are intercalated in the limestone^marl alternation of the Upper Visby Beds (Plate Ie). The reef mounds are composed mainly of tabulate corals and stromatopores (Manten, 1971; Samtleben and Munnecke, 1999). In the entire outcrop area of the Upper Visby Beds two to three zones of predominantly bioclastic limestones occur, each up to 1 m thick. These bioclastic sediments were deposited in increased water energy conditions, and in many cases form the base of reef mounds. The Upper Visby Beds are overlain by crinoidal limestones and reefs of the Ho«gklint Beds. The base of the Ho«gklint Beds is not synchronous since it is de¢ned by the ¢rst occurrence of a thick crinoidal limestone bed (Hede, 1925), and thus depends on local environmental conditions like, e.g., distance from the next reef. The facies changes from the Lower Visby Beds, deposited below wave base, to the Ho«gklint Beds which exhibit widespread algal limestones in the upper part (Riding and Watts, 1981), clearly shows decreasing water depth during this time interval. This is in accordance to most global sea-level reconstructions showing a low stand in the early Sheinwoodian (Leggett et al., 1981; Lenz, 1982; Johnson et al., 1991; Johnson, 1996).

4.2. Development of C- and O-isotope values C- and O-isotope values are generally low in the Lower Visby Beds and have progressively higher values in the Upper Visby Beds (Figs. 3 and 4). Following the terminology of isotope excursions in the upper Silurian of Gotland (Bickert et al., 1997; Samtleben et al., 2000) the corresponding time periods are named ‘upper Telychian H-period’ and ‘lower Sheinwoodian A-period’, respectively. 4.2.1. Carbon isotopes The development of carbon isotope values shows three stages (Figs. 3 and 5): In the lower part of the sequence (Lower Visby Beds) the mean values rise gradually from 1.4x to nearly 2x with a slight decrease below the base of the Phaulactis layer (Fig. 5). In the lowermost 25 cm of the Upper Visby Beds the values increase abruptly to 3x. The lower part of this rapid transition corresponds to the Phaulactis layer. Above, in the Upper Visby Beds N13 C values rise steadily to 4.5x. The rapid change of the carbon isotope values coincides with datum 4. Datum 5 to 8 are within the range of continuously increasing values in the Upper Visby Beds. Presumably, the rapid rise of N13 C values at the base of the Phaulactis layer (datum 4) is caused by a condensed sediment accumulation or a short-term break in sedimentation. This is indicated by the thin pyrite seam which locally is observed at the base of the Phaulactis layer. This pyrite seam is interpreted as a poorly developed hardground and, consequently, as a hiatus. In contrast to the other localities, in the southwestern-most pro¢le at Nyrevsudde (Fig. 1), the facies of the Upper Visby Beds indicates deposition in a distal shelf area similar to the Lower Visby Beds (Plate IIc). Here, the Phaulactis layer is at least 1.6 m thick, indicating a more continuous sedimentologic record, and reveals N13 C values steadily increasing from 2.5 to 3.3x (Fig. 3). While elsewhere N13 C values at the base of the Phaulactis layer do not exceed 2x, the lowermost samples in the pro¢le at Nyrevsudde show values of about 2.5x. The base of the Phaulac-

PALAEO 3075 9-5-03

10 m

Upper Visby Beds

8 7

?

5m

6.2 6 5 4 3.3 3

0m

corals, brachiopods ostracods graptolites

Lower Visby Beds

trilobites

2

Reef mound

irregular limestone/ marl alternation

bioclastic limestone

regular limestone/ marl alternation

0

1

2

3

δ13C (‰)

4

5

6

-7

-6

δ18O (‰)

-5

-4

Gotland

-10 m

global

1

-5 m

conodont extinction steps (datum points)

PALAEO 3075 9-5-03

Phaulactis layer

diversity

Conodonts

Fig. 5. Combined diagram of schematic weathering pro¢le of the Lower and Upper Visby Beds (after Samtleben and Munnecke, 1999), N18 O and N13 C values of the nine pro¢les, and conodont extinction events (cf. Fig. 2). Vertical height is based on the conodont stratigraphy in the type section in Ireviken (Jeppsson and Ma«nnik, 1993; Jeppsson, 1997a,c).

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

Högklint Beds

109

110

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

tis-bearing bed is located probably 1 m below the present sea level. Consequently, the carbon isotope curve at Nyrevsudde does not show the abrupt change observed at the base of the Phaulactis layer from the more proximal pro¢les. Instead, the steady increase of carbon isotope values in the distal pro¢le at Nyrevsudde re£ects a continuous sedimentation during the formation of the Phaulactis layer without hiatus, with a slope of the isotope values similar to that in the Upper Visby Beds at Storbrut (Fig. 3). Similarly, in distal borings in the East Baltic Silurian (Estonia, Latvia), which were correlated to our samples from Gotland, N13 C values show a continuous rise at the Llandovery/Wenlock boundary (Kaljo et al., 1997, 1998). 4.2.2. Oxygen isotopes The development of oxygen isotope values shows a di¡erent picture (Figs. 4 and 5). The Lower Visby Beds are characterised by relatively constant mean values of about 35.6x. In the topmost metre of this series and in the lowest 50 cm of the Upper Visby Beds some values lower than 36x occur. Between the base of the Phaulactis layer ( = 0 m) and +0.25 m the average N18 O values change to about 35.0x. In contrast to the N13 C values, the oxygen isotope values reveal no further increase in the Upper Visby Beds, but show a slight increase in variability (Figs. 4 and 5), probably caused by the higher susceptibility of oxygen isotope values to local changes in temperature and salinity (Samtleben et al., 2000).

extinction event with respect to conodonts and trilobites (Jeppsson, 1997c). A very slight decrease in mean N13 C values is observed in the topmost metre of the Lower Visby Beds between datum 3 and 4, still without a facies shift. At datum 4 (base of the Phaulactis layer) for the ¢rst time organisms of the shallower shelf regions (corals, brachiopods, ostracods) were also a¡ected by extinction events. However, the impact was less severe than it had been earlier to the organisms of the distal facies areas. Above in the Upper Visby Beds both facies and isotope values begin to change, and the N13 C values rise continuously. From here on the facies shows a shallowing caused by the expanding productivity of sessile organisms on the sea £oor. The carbonate content of the sediments increased and small reefs began to grow. As a result, deposition of sediments took place under increasing energy in progressively shallower waters. The conodonts show some minor extinction events (datum 6 to 8). Sessile benthos (brachiopods, corals, stromatopores, bryozoans), in contrast, experienced favourable living conditions compared to the Lower Visby Beds. The benthic assemblages in the Upper Visby Beds are diverse and highly abundant. Reef formation reached its climax with the formation of the Ho«gklint Beds carbonate platform. It is composed of large patch reefs, biostromes, and biodetritic limestones of a shoal facies. It revealed the highest N13 C values (around 5x) of this A-period (Samtleben et al., 1996).

4.3. Phasing of isotope-value changes, facies development, and bioevents

5. Comparison with similar Palaeozoic and Proterozoic events from other areas

The chronological order of extinction events, changes of isotope values, and facies shifts at the boundary between Llandovery and Wenlock on Gotland is as follows : Within the Lower Visby Beds the ¢rst extinction step (datum 1) a¡ected graptolites, conodonts and trilobites, with isotope values remaining almost stable and without any observable facies change. Although these groups of organisms had di¡erent modes of life and ecological a⁄nities, they all lived preponderantly in hemipelagic distal areas. Datum 2 is the strongest

The event at the Llandovery/Wenlock boundary shows clear similarities with other events in the Precambrian (Melezhik et al., 1999), late Cambrian (Saltzman et al., 1998, 2000), late Ordovician (Hirnantian ; Brenchley et al., 1994, 1995), late Homerian, late Gorstian, and middle to late Ludfordian (Samtleben et al., 1996, 2000; Wenzel and Joachimski, 1996; Bickert et al., 1997; Saltzman, 2001) (for a summary see Fig. 6 and Table 1). Events in the Aeronian (Early Silurian; Azmy et al., 1998; Kaljo et al., 1998; Cop-

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

111

Fig. 6. Compilation of similar positive carbon isotope excursions in the Palaeozoic and Proterozoic (cf. Table 1) (after Andrew et al., 1994; Brenchley et al., 1994; Samtleben et al., 1996, 2000; Kaljo et al., 1997, 1998, 2001; Marshall et al., 1997; Underwood et al., 1997; Glumac and Walker, 1998; Heath et al., 1998; Melezhik et al., 1999; Wigforss-Lange, 1999; Saltzman et al., 2000).

per, 2002) and in the Caradocian (Middle Ordovician ; Patzkowsky et al., 1997; Ainsaar et al., 1999a,b; Meidla et al., 1999) may also belong to this pattern, but there is insu⁄cient current knowledge about global isotope development, facies development, and bioevents. 5.1. Silurian 5.1.1. Llandovery/Wenlock boundary (Ireviken Event) in places other than Gotland The Ireviken Event at the Llandovery/Wenlock boundary re£ects global environmental changes. In several regions world-wide, a rapid rise of marine N13 C values in the early Wenlock has been observed : (a) in the East Baltic cores ‘Ohesaare’, ‘Priekule’, ‘Ventspils’, and ‘Ruhnu’ (Kaljo et al., 1997, 1998); (b) from Estonian shelf-carbonate

sequences (Heath et al., 1998); (c) in the Carnic Alps, Austria, pro¢le ‘Oberbuchach 1’ (Wenzel, 1997); (d) in New South Wales, Australia, Borenore Caves (Talent et al., 1993); (e) from North America (Pete Hanson Creek II in Nevada; Highway 77 section, Oklahoma) (Saltzman, 2001) and (f) England (own data). Additionally, a succession of extinction events nearly identical to that on Gotland is documented in several successions, excluding a Signor Lipps e¡ect as possible cause (Jeppsson and Ma«nnik, 1993; Jeppsson, 1997c). A change to carbonaterich sediment series is found in many regions world-wide (Jeppsson, 1987; Jeppsson et al., 1995; Kaljo et al., 1998; Le He¤risse¤, 2000). Even in Bolivia, which in the early Silurian was situated at latitudes higher than 60‡S, warm-water carbonates of up to 10 m in thickness (Sacta

PALAEO 3075 9-5-03

112

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

SPICE event

Hirnantian event

lower Sheinwoodian A-period

upper Homerian A-period

upper Gorstian A-period

positive δ13 C excursion

x

x

x

x

x

x

x

positive δ18 O excursion

?

?

x

x

x

-

x

carbonate factories and reefs on shallow shelves

x

x

x

x

x

x

x

abundant stromatolites / oncolites

x

x

x

x

x

-

x

oxygenated sediments in deep shelf environments

?

?

x

x

x

?

x

anoxic sediments below and above in deep shelf environments

x

x

x

x

x

?

x

hiatus near the base in shallow shelf environments

?

x

x

x

x

-

x

intercalated siltstones and sandstones on proximal shelves

x

-

x

x

x

?

x

glacial sediments in high latitudes

-

-

x

?

-

-

-

extinction events in the marine environment

-

x

x

x

x

x

x

low diverse plankton community

-

?

x

x

x

x

x

short duration

?

x

x

x

x

x

x

high atmospheric CO2

x

x

x

x

x

x

x

low permanent/long-term ecological impact

-

x

x

x

x

x

x

upper Ludfordian A-period

Paleoproterozoic excursions

Table 1 Simpli¢ed compilation of common features of seven di¡erent Palaeozoic and Proterozoic time periods characterised by distinct N13 C excursions

See Fig. 6. Data after Leggett et al., 1981; Berner, 1990; Jeppsson, 1990, 1997a-c, 1998; Grahn and Caputo, 1992; Talent et al., 1993; Brenchley et al., 1994; Jeppsson et al., 1994, 1995; Melchin, 1994; Berry et al., 1995; Kaljo et al., 1995, 1997, 1998, 2001; Samtleben et al., 1996, 2000; Bickert et al., 1997; Underwood et al., 1997; Wenzel, 1997; Nestor, 1998; Saltzman et al., 1998, 2000; Glumac and Walker, 1998, 2000; Calner and Sa«ll, 1999; Finney et al., 1999; Kump et al., 1999; Melezhik et al., 1999; Mikulic and Kluessendorf, 1999; Paris et al., 2000; Zhang et al., 2000; Bourque et al., 2001; Copper, 2001, 2002; Saltzman, 2001; Calner, 2002; Muir, 2002.

Member) overlie glaciomarine sediments of the Cancan‹iri and Zapla Formations of late Ordovician and Llandovery age (D|¤az-Mart|¤nez, 1997, 1998). 5.1.2. Late Wenlock and Ludlow In the Silurian on Gotland, periods characterised by increased N13 C values occurred not only in the early Sheinwoodian but also in the late Homerian, in the late Gorstian, and in the middle to late Ludfordian (Samtleben et al., 1996, 2000; Wenzel and Joachimski, 1996). During these A-periods there was an intense growth of coralstromatopore reefs, and extensive carbonate platforms formed. With the exception of a newly dis-

covered, rather weakly developed A-period in the late Gorstian (Samtleben et al., 2000), these periods with positive N13 C excursions are also found in other regions of the world. In spite of di¡erences in amplitude of isotope values and from minor stratigraphical uncertainties, these N13 C shifts are correlated with high con¢dence. The excursion in the late Homerian has been described from the Eastern Baltic (Kaljo et al., 1995, 1997, 1998), from Great Britain (Cor¢eld et al., 1992), and from North America (Saltzman, 2001). The isotope excursion in the late Ludfordian, which reaches N13 C values of 8.5x on Gotland, is described with values of up to 12x from Australia (Andrew et al., 1994), 11x from southern Swe-

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

den (Wigforss-Lange, 1999), 8x from Que¤bec (Bourque et al., 2001), about 5x from the Eastern Baltic (Kaljo et al., 1995, 1997, 1998), and nearly 4x from North America (Saltzman, 2001). This event is the strongest N13 C excursion during the entire Phanerozoic. Its values are exceeded only by values from the Precambrian (see below). The Silurian A-periods coincide with changes in facies development (Jeppsson, 1987; Samtleben et al., 2000; Jeppsson et al., 2002). Whether the deposition of evaporites in North America (Michigan Basin) and Australia (New South Wales) during the Wenlock and Ludlow can be correlated with the A-periods on Gotland is yet unknown because of the lack of appropriate index fossils (Mesolella et al., 1974; Jones et al., 1987; Wilde et al., 1991). The onset of the A-periods is always marked by extinction events in the marine realm (Kaljo et al., 1995; Jeppsson, 1998; Dorning, 1999). Similar to the A-periods in the early Sheinwoodian, in the late Homerian A-period, and apparently also in the two later A-periods, ¢rst extinctions of hemi-pelagic organisms occurred earlier than the changes of isotope values (Samtleben et al., 2000). After that, N13 C and N18 O values began to rise, facies changed, and reefs started to grow. During the ascent of isotope values, a short-term regression occurred which produced hiatuses in shallow marine areas. After the regression, reef formation reached its maximum and additional species became extinct. The observation that brachiopods found favourable conditions during A-periods does not only apply to the lower Sheinwoodian A-period. Also the A-periods in the late Homerian (Mulde Beds) and the late Ludfordian (Eke Beds) are characterised by diverse and very abundant brachiopod assemblages (Bassett and Cocks, 1974; Samtleben et al., 2000). 5.2. Ordovician The best known A-period coincides with the Hirnantian glaciation in the late Ordovician (Brenchley, 1989; Brenchley et al., 1994, 1995; Underwood et al., 1997; Vennin et al., 1998; Finney et al., 1999; Rong and Harper, 1999; Kump

113

and Arthur, 1999; Kump et al., 1999; Droser et al., 2000; Sou¢ane and Achab, 2000; Sutcli¡e et al., 2000; Zhang et al., 2000; Sheehan, 2001). It shows clear similarities with the lower Sheinwoodian A-period (Table 1). Positive isotope excursions have been reported from North and South America, Baltica, and China (e.g., Long, 1993; Wang et al., 1993; Brenchley et al., 1994, 1997; Marshall et al., 1997; Finney et al., 1999; Kump et al., 1999; Hints et al., 2000; Kaljo et al., 2001). During the extraordinarius Subzone, which corresponds to the glaciation in the lower Hirnantian, N13 C values rose to 7x and, similarly, N18 O values increased by 4x. In the upper Hirnantian post-glacial persculptus Subzone both oxygen and carbon isotope values declined to pre-Hirnantian values (Brenchley et al., 1994, 1997; Marshall et al., 1997). During Hirnantian times glacial sediments were deposited on Gondwana, and simultaneously a world-wide regression occurred (Gagnier et al., 1996; D|¤az-Mart|¤nez, 1997, 1998). Caused by this regression, sediment sequences in shallow-marine facies areas show hiatuses (Brenchley, 1989). In Australia, evaporites accumulated (Williams, 1991). At the northern margin of Gondwana crinoid^bryzoan mounds and oolites were formed (Vennin et al., 1998; Copper, 2001). During the maximum regression sandstones and siltstones were widespread (Brenchley, 1989; Kaljo et al., 2001). Many deep-water deposits that had been oxygen poor during most of the Ordovician became aerated in the Hirnantian (Sheehan, 2001). At the stratotype locality of the Ordovician^Silurian boundary, Dob’s Linn (Scotland), black shales are overlain by bioturbated siltstones that correlate with the Hirnantian glaciation, that in turn are overlain by graptolitic shales (Hallam and Wignall, 1999). A sea-level drop reduced the extent of epicontinental seas, and reef distribution was more restricted than in the preceding Ashgill and Caradoc (Copper, 2001; Webby, 2002). However, reefs of Hirnantian age are still widespread and have been reported from Baltica, Kazakhstan, Siberia, arctic North America and Greenland, and in the northeastern parts of Eurasia (Copper, 2001). Oolites and oncolites were abundant (Brenchley, 1989;

PALAEO 3075 9-5-03

114

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

Underwood et al., 1997; Hallam and Wignall, 1999; Tuuling and Flode¤n, 2000; Zhang et al., 2000; Kaljo et al., 2001; Copper, 2001). The reefs are usually overlain by bituminous limestones or shales of post-extraordinarius and (or) Silurian age (Copper, 2001). Palaeontologically, the Hirnantian is characterised by ubiquitous species adapted to cold-water (Hirnantia-fauna). The name-giving brachiopod genus Hirnantia is not only typical for the cool temperate zone but also for areas with a high input of sand or silt (Copper, 2001). Almost all groups of organisms were a¡ected by the extinction event in the late Ordovician, which is the second-strongest mass extinction in the Phanerozoic (Droser et al., 2000; Sheehan, 2001). Similar to the Llandovery/Wenlock boundary, the organisms show a stepwise pattern of extinction with graptolites and radiolarians being a¡ected by the perturbations clearly before conodonts and chitinozoans (Finney et al., 1999; Chen et al., 2002). Among the benthos, deep-water communities were more a¡ected than those of shallower water (Brenchley, 1989). At the beginning of the event the graptolites became nearly extinct and about 40% of the trilobite genera were lost (Brenchley, 1989; Brenchley et al., 1994; Adrain et al., 2000). Of the trilobites, the pelagic forms or those with pelagic larval stages became totally extinct (Chatterton and Speyer, 1989), while trilobites adapted to reef environments were less a¡ected (Brenchley, 1989). Numerous species and genera of reef fauna became extinct at or near the top of the extraordinarius graptolite Subzone (Copper, 2002), but principally reef communities were less oppressed than level bottom communities (Copper, 1997, 2001, 2002; Flu«gel and Kiessling, 2002). The second phase of extinction that a¡ected mainly brachiopods, corals, chitinozoans, and acritarchs occurred at the end of the Hirnantian glaciation, coincident with a strong decrease in N13 C and N18 O values (Brenchley et al., 1995; Paris et al., 2000). Falling sea level and resulting loss of shelf habitat are unlikely to be the reason for the ¢rst extinctions because especially zooplanktonic organisms were severely a¡ected (Brenchley, 1989). Furthermore, many organisms became extinct at

the very beginning of the isotope excursion, i.e. prior to sea-level drop and related ecological changes (Brenchley, 1989; Brenchley et al., 1995). Also temperature changes are questionable as causes for the ¢rst extinctions. ‘‘Although there was climatic cooling at the start of the Hirnantian the maximum cooling appears to have been later, so it is debatable whether the initial fall in ocean water temperature was su⁄ciently large to be the ‘¢rst strike’’’ (Brenchley et al., 1995). Despite the intensity of the extinction events, there was only little ecological e¡ect related to the Ordovician glaciation with respect to larger time scales (Brenchley, 1989; Sheehan, 1996, 2001; Droser et al., 2000). ‘‘In a statistical sense this was a mass extinction, but ecologically it was less severe than the end Permian and end Cretaceous events; there was no major loss of higher taxa, nor was the ecological structure permanently disrupted’’ (Brenchley, 1989). Brenchley et al. (1995) assume that the extinction events of the Hirnantian were related to changes in ocean circulation. They assume a deeply ventilated ocean during glaciations similar to that of the modern ocean conditions. During interglacials, the ocean basins should have been ¢lled with warm saline deep water. As a consequence, the extinction event at the beginning of the Hirnantian would have been coincident to the upwelling of nutrient-rich deep water masses that might have been toxic to some near-surface biota. Brenchley et al. (1995) interpret the high N13 C values during the period of glaciations as the result of high plankton productivity, although there is no indication for this in the fossil record. 5.3. Cambrian Within the late Cambrian (Steptoean), sediments characterised by high N13 Ccarb values were deposited world-wide, e.g. in North America, Kazakhstan, China, and Australia (Saltzman et al., 2000). These deposits were related to an A-period known as the SPICE event (Steptoean positive carbon isotope excursion). This event coincides with a dramatic extinction of trilobites at the base of the Glyptagnostus reticulatus Zone at the

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

very beginning of the carbon isotope excursion. ‘‘Thus, (...) this Upper Cambrian faunal crisis appears to have resulted from the cause of the carbon cycle perturbation (...) rather than a deterioration of oceanic conditions which may have developed during the SPICE excursion’’ (Saltzman et al., 2000). During the A-period, i.e. during the interval of high N13 C values, a radiation of benthic organisms occurred at least in Laurentia. The related shelf deposits are mainly carbonates, i.e. oolites, micritic and sparitic limestones, dolomites, and algal limestones. Generally, the late Cambrian period is characterised by extensive black shale deposits (Leggett et al., 1981). The storage of 12 C-enriched organic material leaves the 4CO2 of the global ocean water enriched in 13 C and, thus, might be responsible for the positive isotope excursion. However, there is no clear stratigraphical evidence for black shale deposition during the SPICE event. 5.4. Proterozoic The strongest positive N13 Ccarb excursions ever described (up to 15x) occurred after the Precambrian Huronian glaciation between 2.40 and 2.06 Ba (Melezhik et al., 1999). As in other A-periods described above, these excursions are related to extensive carbonate platforms which consist of carbonates poor in organic carbon, i.e. stromatolites, dolomites, evaporites, and oolites. Occasionally, sand- and siltstones are intercalated. These deposits are generally overlain by black shales. Up to now, such Precambrian carbon isotope excursions are known from sites in North America, Europe, Africa, Asia, and Australia (Melezhik et al., 1999). Apart from the similarities, the comparability of this event with the Palaeozoic events is limited by the fact, that the environmental conditions during the Proterozoic N13 C excursion di¡er from those of the Palaeozoic A-periods. This is indicated by repeated deposition of banded iron formations. Their occurrence reveals that the atmosphere at that time was barren of free oxygen. The ¢rst red beds occurred at 2.2 Ba, that is 200 million years after the ¢rst N13 C excursion (Melezhik et al., 1999).

115

6. Discussion Although each of the events described above has its own characteristics, their conspicuous similarities (Table 1) indicate similar controlling mechanisms. This is corroborated by the fact that all of these events occur at times characterised by very high atmospheric CO2 concentrations (Berner, 1990), and by the absence of vegetation on land. The positive N13 C excursion in the middle to late Ludfordian (Silurian) with up to 12x is the strongest of the entire Phanerozoic (Andrew et al., 1994; Samtleben et al., 1996, 2000; Wenzel and Joachimski, 1996; Bickert et al., 1997; Wigforss-Lange, 1999). Several other carbon isotope excursions related to A-periods also exceed 5x, a shift too high to be explained by fractionation due to changes in oceanic productivity (Bickert et al., 1997). Furthermore, the excursions are not associated with widespread deposition of black shales in shelf environments like many other such events in the younger Earth history (Berger and Vincent, 1986; Weissert and Mohr, 1996; Jenkyns and Clayton, 1997; Kump and Arthur, 1999). Instead, these intervals are characterised by sediments depleted in organic matter (Leggett et al., 1981; Wenzel, 1997). The Hirnantian positive N18 O excursion is in accordance with the glaciation. However, Brenchley et al. (1994, 1995) noticed that the increase of about 4x requires a sea-level drop of about 100 m in addition to a temperature decrease in the tropics by about 10‡C, which is unrealistic for this period. The N13 C excursion is not easily interpreted either, because the Hirnantian is characterised by shelf sediments which are poor in Corg (Leggett et al., 1981; Wenzel, 1997; Finney et al., 1999; Zhang et al., 2000). Kump et al. (1999) proposed a ‘‘weathering hypothesis’’ as explanation for the glaciation. According to this hypothesis, late Ordovician mountain building resulted in exposure and weathering of silicate terrains, which resulted in the consumption of atmospheric CO2 . The falling pCO2 initiated the growth of polar ice caps. The growing ice caps began to cover high-latitude silicate rocks, resulting in lowered silicate weathering and lowered

PALAEO 3075 9-5-03

116

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

CO2 consumption. The pCO2 increased, and the glaciation ended. Kump et al. (1999) assume an increase of carbonate platform weathering resulting from the regression as the cause for the N13 C excursion of the Hirnantian Event. This interpretation seems to be unlikely at least for the other A-periods because of the preponderance of carbonate accumulation during A-periods. From the similarity between the Hirnantian glaciation and the Silurian events, Kump et al. (1999) infer the existence of glaciations also in the Silurian. However, to date, indications for a glaciation are restricted to the early Silurian (Grahn and Caputo, 1992; D|¤az-Mart|¤nez et al., 2001). In Bolivia, where the Hirnantian glaciation is represented by widespread glacial deposits, warm water carbonates with tabulate corals were formed in the early Wenlock (Sheinwoodian) (Gagnier et al., 1996; D|¤az-Mart|¤nez, 1997, 1998). Even during the strongest N13 C (and N18 O) excursion in the late Ludfordian there are no indications of glacial sediments, and the Precambrian positive N13 C excursions clearly occur later than the Huronian glaciation (Melezhik et al., 1999). Therefore, glaciations as the reason for the events described are unlikely. Nevertheless, in sequences with isotope records based on measurements in diagenetically unaltered brachiopod shells, A-periods show high N18 O values (Brenchley et al., 1994; Samtleben et al., 1996, 2000; Marshall et al., 1997; Heath et al., 1998). These values might indicate not only higher salinity under arid conditions but also lower temperatures in surface waters, and accordingly periods of cooler climates. 6.1. Oceanographic circulation in A- and H-periods Based on investigations of Silurian conodont assemblages, Jeppsson (1990, 1998) developed a palaeoceanographic model which is based on changes between humid and arid climatic conditions (Fig. 7). This model o¡ers an explanation for the facies shifts without need of signi¢cant sea-level changes. The low input of terrigenous material during A-periods led to a formation of reefs and carbonate platforms. During H-periods, an increased terrigenous input lowered the depo-

sition of carbonate-rich sediments. The diverse and abundant conodont assemblages in sediments from H-periods can be explained by the terrigenous in£ux which increased the fertility of surface waters in tropical shallow marine regions. Jeppsson (1990) postulated lower N13 C values in times of arid climate caused by low planktonic production. The authors of the present study modi¢ed the model of Jeppsson (1990) with respect to ocean circulation and stable isotope geochemistry (Bickert et al., 1997) (Fig. 7). We infer a shift between estuarine and anti-estuarine circulation in shallow seas caused by precipitation changes as the main driving mechanism. Because of permanent euxinic conditions below the surface waters in the ocean during the entire Silurian (Wilde et al., 1991) there was a strong fractionation in carbon isotopic composition between surface and deep waters produced by the settlement and deposition of 12 C-rich organic material in deep-sea sediments. Today a similar fractionation is observed in the Black Sea (Fry et al., 1991). For the Silurian ocean, Wilde et al. (1991) reconstructed a strati¢ed ocean with a thick oxygen minimum layer as shallow as 100 m water depth. The shift from humid to arid climates led to changes in ocean circulation. During A-periods the formation and downwelling of saline surface water caused an anti-estuarine circulation pattern in shallow seas, and O2 -rich but 12 C-depleted open ocean surface water reached the shelf areas, resulting in oxygenated deep shelf sediments observed for most A-periods (Table 1; Wenzel, 1997; Sheehan, 2001; Muir, 2002). The low N13 C values in the H-periods were produced by upwelling of 12 Crich deep water (Fig. 7). The assumption of a strati¢ed Palaeozoic ocean even during the Hirnantian contradicts the widespread opinion of a deeply ventilated ocean during the glaciation. However, it seems reasonable, because at that time the margins of Gondwana extended well into the subtropics thereby prohibiting a deep-water convection as is known from the modern ocean. The model of the oceanographic circulation of Jeppsson (1990) is questionable for the following reasons: (1) Jeppsson’s model should result in an

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

PALAEO 3075 9-5-03 Fig. 7. Comparison of di¡erent palaeoceanographic/climatic models of humid and arid periods in the Silurian; simpli¢ed after Jeppsson (1990) (left) and Bickert et al. (1997) (right).

117

118

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

accumulation of black shales on deep shelves during A-periods caused by the anoxic bottom water. Instead, these intervals are characterised by bioturbated, well-oxygenated sediments. (2) Following Jeppsson’s model, the Hirnantian glaciation, which in many aspects corresponds to other A-periods (Table 1), would have been an S-episode characterised by warmer temperatures in high latitudes (Fig. 7) (Jeppsson did not apply his model to this time interval). In some A-periods a depth-dependent gradient of N13 C values has been observed: b Lower Sheinwoodian A-period : In the Carnic Alps the sections Rauchkofelboden, Cellon, and Oberbuchach, which represent a transition from shallow to deep shelf environments of the Palaeotethys, exhibit a decrease in N13 C values (Wenzel, 1997). Also, the Baltic cores of Ruhnu and Ohesaare, which were drilled into open shelf deposits, exhibit N13 C values of the lower Sheinwoodian A-period of 4x to 5x that are higher than synchronous deposits of a deeper facies drilled in Ventspiels and Priekule (3x to 4x) (Kaljo et al., 1998). Similarly, upper Llandovery sections of the shallow-marine sediments exhibit N13 C values 1x to 2x higher than those of the deeper sediments (Kaljo et al., 1998). º ved-Ramb Upper Ludfordian A-period : The O safisa Beds, southern Sweden, are composed mainly of marine siliciclastics and intercalated carbonates. In the late Ludfordian, the 25 m thick oncolitic Bja«rsjo«lagafird Limestone (Scania, South Sweden) was deposited which exhibits a N13 C excursion of 11.2x (Wigforss-Lange, 1999). In the nearby somewhat deeper-marine sections close to Klinte, the excursion does not exceed 7.6x (Wigforss-Lange, 1999). b Hirnantian : A decrease in N13 C amplitudes is reported for the transition from shallow to deep facies in the Hirnantian of Nevada (Finney et al., 1999). The carbonates of the Monitor Range Section, which are deposited in an embayed platform margin, show an excursion of 6x (from 2x to 8x), while the more distal sediments of the Vinini Creek Section exhibit an excursion of only 3x (31x to +2x). On Gotland a similar facies dependence of the N13 C gradients has not been observed, because

even the deepest sediments have been deposited on a carbonate ramp in water depths less than 100 m (Samtleben et al., 2000). The depth dependence of the N13 C values is readily explained by our model: During the Palaeozoic the carbonate production was mostly benthic, as the calcareous plankton had scarcely been developed. Planktonic micro-organisms with calcareous tests were present in the Silurian (Munnecke et al., 1999, 2000, 2001), but they were of little signi¢cance compared to planktonic foraminifera and coccolithophorids since the Mesozoic. Most of the carbonate mud is produced by the decay of calcareous skeletons of benthic organisms. Thus, the N13 Ccarb values reveal the isotopic composition of the bottom water of that time and not that of the corresponding surface waters. With increasing distance from the carbonate platforms the relative amount of allochthonous platform-derived carbonate mud decreases. A strong fractionation of isotope values in the water column, as postulated in our model (Fig. 7), predicts lower N13 Ccarb values and weaker amplitudes in sediments of deeper shelf areas than in upper shelf environments, caused by the increasing admixture of 12 C enriched bottom water with water depth both in A- and in H-periods. 6.2. Transition from H- to A-periods, and possible causes for extinction events The observation that the N13 C values after A-periods returned to similar levels as before the excursions points to a climate system oscillating between two basic states. Hence, unidirectional processes such as vertical or horizontal tectonics are unlikely, and up to now no indications for bolide impacts have been found (Orth et al., 1986; Wilde et al., 1986; Schmitz et al., 1994). The transitions from H- to A-periods, with durations of 104 to 105 years, o¡er possibilities to examine the causes of the climatic changes by a detailed investigation of the time relations between extinction events, facies shifts, and developments of isotope values. The extinction events connected with the carbon isotope excursions are di¡erent from other bioevents in Earth history because a permanent

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

transformation of the marine ecosystem did not occur. Instead, after A-periods, speci¢c environments have been re-settled by similar faunal assemblages (Droser et al., 2000). We conclude that the ecological system might have oscillated between two modes, too. The results of the present investigation indicate a close relationship between the development of facies, isotope values, and palaeontological events. Especially, sediment facies and isotope values seem to be closely correlated. In the Ireviken Event, both show a rapid shift at the base of the Phaulactis layer followed by a further gradual increase of N13 C values in the early Sheinwoodian (Upper Visby Beds). Likewise, the facies shallowed and became more and more carbonaterich, culminating in the Ho«gklint Beds (Samtleben et al., 1996). Interestingly, the ¢rst extinction events are observed nearly 5 m below the Phaulactis layer (Figs. 2 and 5) clearly before the ascent of the N13 C values. Carbon and oxygen isotope values show only small-scale variations. There are no isotope data available from the interval from 36 m to 33 m. However, N13 C values of wholerock samples from bore holes of the Baltic area show that there are no large scale variations in the uppermost Llandovery (Kaljo et al., 1998). This observation implies that the processes which caused the rise of the isotope values did not induce the early extinction events. Obviously, organisms were severely a¡ected long before oceanographic and climatic processes were re£ected in isotope signals. Also the other transitions from H- to A-periods later in the Silurian, in the Hirnantian, and in the late Cambrian show early extinction events prior to the increase of isotope values (Brenchley et al., 1995; Saltzman et al., 2000; Samtleben et al., 2000). This observation excludes many possible causes for the extinctions, i.e., global temperature changes (Barron, 1995) and consequent disturbances of ecological conditions. Such temperature changes should have been connected to changes in N18 O, but these did not occur. A hint for the causes of extinctions is given by the ecological a⁄nities of the extinct organisms. Pelagic planktonic graptolites, hemipelagic nek-

119

tonic conodontophorids as well as pelagic/hemipelagic benthic trilobites were strongly a¡ected in these early events, but reef organisms only to a lesser extent (Brood, 1985; Brenchley, 1989; Chatterton and Speyer, 1989; Melchin, 1994; Copper, 1997, 2002; Jeppsson, 1997c, 1998). This indicates that the mechanism for the extinctions ¢rst acted in the deeper ocean realm and then extended into the shallower marginal seas. In many cases, the shallow marine benthos shows highly diverse assemblages during A-periods (Copper, 2001, 2002; Samtleben et al., 2000). The fact that reef organisms were only a¡ected to a minor extent by extinction events might be explained by their relatively unspecialised mode of life compared to modern reef builders and, thus, their low dependence on reef habitats. While many reef ecosystems were destroyed in H-periods, reef-forming benthos species survived. One likely scenario could be that the extinctions were caused by anoxic events linked to the shift of the oceanic circulation from estuarine to anti-estuarine mode. Such a shift passed inevitably through a state with a standstill of the oceanic circulation. In that situation the fronts between water masses would break down and consequently euxinic deep-water masses could spread in peripheral marine areas. A shallowing of the oxygen minimum zone caused by a standstill of the estuarine circulation (H-period) shortly before the turn-over into an anti-estuarine mode (A-period), could have a¡ected not only the benthic organisms of deep shelves, but also the deep migrating zooplankton, whereas phytoplankton (acritarchs) is less a¡ected. Such a mechanism would explain the step-like extinction of organisms of di¡erent habitats. Later, with the collapse of the climatic system and the reversal of the oceanic circulation from an estuarine to an increasing anti-estuarine mode, the nutrient input to the ocean diminished. Possibly, the abundant occurrence of the solitary rugose coral Phaulactis on Gotland at the transition from the upper Telychian H-period to the lower Sheinwoodian A-period indicates the short-term occupation of an ecological niche until the ecological system came into equilibrium with the altered environmental

PALAEO 3075 9-5-03

120

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

conditions. The sessile shallow-water organisms (brachiopods, bryozoans, corals, and stromatopores) did not su¡er from the ecological changes. On the contrary, these ¢lter feeders which could not search and select their food were not feeding specialists dependent on a speci¢c prey, and obviously had advantages in the new situation. Except for the Hirnantian, the causes for the hiatuses observed at the beginning of many A-periods within the increase of isotope values are still unexplained. Calner and Sa«ll (1999) calculated for the beginning of the upper Homerian A-period a short-term regression of about 16 m, based on ¢eld observations. As a cause for such short-term sea-level changes, the storing of water in continental ice shields seems most likely. However, as discussed above, for most of the A-periods there is no evidence of glacial sediments. As another explanation for the hiatuses, an increased submarine erosion due to stronger surge might be discussed. As indicated by the glacial deposits in the Hirnantian A-period, the polar regions might have been generally colder in A-periods than during H-periods. Even if they were not cold enough for the formation of large polar ice shields, the consequently higher temperature gradients between low and high latitudes would result in intensi¢ed seasonal temperature changes and stronger winds. This would lead to increased wave amplitudes, a lowering of the wave base, and to submarine erosion in shallow platform environments. The several million years long and to some extent varying time intervals between the A-periods demand an explanation di¡erent from astronomical forced climatic cycles. More likely is an internal feed-back mechanism which, at the end of each H-period, by the break of a climatic threshold led to a shift into an A-period mode. A likely mechanism discussed in this respect is the carbon cycle of the coupled ocean-atmosphere system, because all known A-periods occur during times characterised by high atmospheric CO2 levels and nearly without continental biomass. However, such an explanation remains speculative, until the climatic processes and their thresholds are better understood.

Acknowledgements The authors are grateful to Monika Segl and her team for carrying out the isotope measurements in Bremen, and to Hildegard Westphal, James Nebelsick, and Annette Schmid-Ro«hl for helpful comments. The paper bene¢ted from constructive reviews by Dimitri Kaljo (Tallinn, Estonia) and Peter M. Sheehan (Milwaukee, WI, USA). This study was funded by the Deutsche Forschungsgemeinschaft (Bi 657/1, Sa 124/8, We 992/20).

References Adrain, J.M., Westrop, S.R., Chatterton, B.D.E., Ramsko«ld, L., 2000. Silurian trilobite alpha diversity and the end-Ordovician mass extinction. Paleobiology 26, 625^646. Ainsaar, L., Martma, T., Meidla, T., Rubel, M., Sidaraviciene¤, N., 1999. Quantitative stratigraphy of sedimentary sequences: a case study of the Middle Ordovician Event. In: Har¡, J., Lembke, W., Stattegger, K. (Eds.), Computerized Modeling of Sedimentary Systems. Springer, New York, pp. 275^ 287. Ainsaar, L., Meidla, T., Martma, T., 1999b. Evidence for a widespread carbon isotopic event associated with late Middle Ordovician sedimentological and faunal changes in Estonia. Geol. Mag. 136, 49^62. Aldridge, R.J., Jeppsson, L., Dorning, K.J., 1993. Early Silurian oceanic episodes and events. J. Geol. Soc. Lond. 150, 501^513. Andrew, A.S., Hamilton, P.J., Mawson, R., Talent, J.A., Whitford, D.J., 1994. Isotopic correlation tools in the Middle Palaeozoic and their relation to extinction event. APEA J. 34, 268^277. Azmy, K., Veizer, J., Bassett, M.G., Copper, P., 1998. Oxygen and carbon isotopic composition of Silurian brachiopods; implications for coeval seawater and glaciations. Geol. Soc. Am. Bull. 110, 1499^1512. Barron, E.J., 1995. Tropical climate stability and implications for the distribution of life. In: Stanley, S.M. (Ed.), E¡ects of Past Global Change on Life. National Academy Press, Washington, DC, pp. 94^107. Bassett, M.G., Cocks, L.R.M., 1974. A review of Silurian brachiopods from Gotland. Fossils Strata 3, 1^56. Bassett, M.G., Edwards, D., 1991. The Murchison Symposium ^ Proceedings of an International Conference on the Silurian System. Special Papers in Palaeontology 44, 397 pp. Berger, W.H., Vincent, E., 1986. Deep-sea carbonates: reading the carbon-isotope signal. Geol. Rundsch. 75, 249^269. Berner, R.A., 1990. Atmospheric carbon dioxide levels over Phanerozoic time. Science 249, 1382^1384.

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124 Berry, W.B.N., Quinby-Hunt, M.S., Wilde, P., 1995. Impact of Late Ordovician glaciation-deglaciation on marine life. In: Stanley, S.M. (Ed.), E¡ects of Past Global Change on Life. National Academy Press, Washington, DC, pp. 34^46. Berry, W.B.N., 1998. Silurian oceanic episodes: the evidence from central Nevada. N.Y. State Mus. Bull. 491, 259^264. Bickert, T., Pa«tzold, J., Samtleben, C., Munnecke, A., 1997. Paleoenvironmental changes in the Silurian indicated by stable isotopes in brachiopod shells from Gotland, Sweden. Geochim. Cosmochim. Acta 61, 2717^2730. Boucot, A.J., 1991. Developments in Silurian Studies since 1839. In: Bassett, M.G., Edwards, D. (Eds.), The Murchison Symposium ^ Proceedings of an International Conference on the Silurian System. Special Papers in Palaeontology 44, 91^107. Bourque, P.-A., Savard, M., Chi, G., Danserau, P., 2001. Diagenesis and porosity evolution of the Upper Silurian-lowermost Devonian West Point reef limestone, eastern Gaspe¤ Belt, Que¤bec Appalachians. Bull. Cdn. Petrol. Geol. 49, 299^326. Brenchley, P.J., 1989. The late Ordovician extinction. In: Donovan, S.K. (Ed.), Mass Extinctions: Processes and Evidence. Belhaven Press, London, pp. 104^132. Brenchley, P.J., Marshall, J.D., Carden, G.A.F., Robertson, D.B.R., Long, D.G.F., Meidla, T., Hints, L., Anderson, T.F., 1994. Bathymetric and isotopic evidence for a shortlived Late Ordovician glaciation in a greenhouse period. Geology 22, 295^298. Brenchley, P.J., Carden, G.A.F., Marshall, J.D., 1995. Environmental changes associated with the ‘¢rst strike’ of the late Ordovician mass extinction. Modern Geol. 20, 69^82. Brenchley, P.J., Marshall, J.D., Hints, L., No‹lvak, J., 1997. New isotopic data solving an old biostratigraphic problem: the age of the upper Ordovician brachiopod Holorhynchus giganteus. J. Geol. Soc. London 154, 335^342. Brood, K., 1985. Bryozoans from the Ro«nnklint section, Gotland. GFF 107, 71^75. Calner, M., 2002. A lowstand epikarstic intertidal £at from the middle Silurian of Gotland, Sweden. Sediment. Geol. 148, 389^403. Calner, M., Sa«ll, E., 1999. Transgressive oolites onlapping a Silurian rocky shoreline unconformity, Gotland, Sweden. GFF 121, 91^100. Chatterton, B.D.E., Speyer, S.E., 1989. Larval ecology, life history strategies, and patterns of extinction and survivorship among Ordovician trilobites. Paleobiology 15, 118^132. Chen, X., Mitchell, C.E., Melchin, M.J., Jun-Xuan, F., 2002. Patterns and processes of latest Ordovician graptolite extinction and recovery based on data from South China. IPC2002 Geol. Soc. Austr. Abstr. 86, 32^33. Cocks, L.R.M., Scotese, C.R., 1991. The global biogeography of the Silurian period. In: Bassett, M.G., Edwards, D. (Eds.), The Murchison Symposium ^ Proceedings of an International Conference on the Silurian System. Special Papers in Palaeontology 44, 109^122. Copper, P., 1997. Reefs and carbonate productivity: Cambrian through Devonian. Proceedings of the 8th International

121

Coral Reef Symposium, Panama City, Panama 2, 1623^ 1630. Copper, P., 2001. Reefs during the multiple crisis towards the Ordovician-Silurian boundary: Anticosti Island, eastern Canada, and world-wide. Can. J. Earth Sci. 38, 153^171. Copper, P., 2002. Silurian and Devonian reefs: 80 million years of global greenhouse between two ice ages. In: Kiessling, W., Flu«gel, E., Golonka, J. (Eds.), Phanerozoic Reef Patterns. SEPM Special Volume 72, 181^238. Copper, P., Brunton, F., 1991. A global review of Silurian reefs. In: Bassett, M.G., Edwards, D. (Eds.), The Murchison Symposium ^ Proceedings of an International Conference on the Silurian System. Special Papers in Palaeontology 44, 225^259. Cor¢eld, R.M., Siveter, D.J., Cartlidge, J.E., McKerrow, W.S., 1992. Carbon isotope excursion near the Wenlock-Ludlow (Silurian) boundary in the Anglo-Welsh area. Geology 20, 371^374. D|¤az-Mart|¤nez, E., 1997. Latest Ordovician-Early Silurian glaciation and carbonate deposition in the Bolivian Central Andes. Libro de Resu¤menes y Excursiones V International Meeting IGCP Project 351, A Corun‹a, 51^53. D|¤az-Mart|¤nez, E., 1998. Silurian of Peru and Bolivia: recent advances and future research. In: Gutie¤rrez-Marco, J.C., Ra¤bano, I. (Eds.), Proceedings 6th International Graptolite Conference and 1998 Field Meeting of the IUGS Subcommission on Silurian Stratigraphy. Temas Geolo¤gico-Mineros ITGE, 23, 69^75. D|¤az-Mart|¤nez, E., Acosta, H., Cardenas, J., Carlotto, V., Rodr|¤guez, R., 2001. Paleozoic diamictites in the Peruvian Altiplano: evidence and tectonic implications. J. S. Am. Earth Sci. 14, 587^592. Dorning, K.J., 1999. Wenlock cyclicity, palynology, and stratigraphy in the Buildwas, Coalbrookdale, and Much Wenlock Limestone formations, Shropshire, England. Boll. Soc. Paleontol. Ital. 38, 155^166. Droser, M.L., Bottjer, D.J., Sheehan, P.M., McGhee, G.R., Jr., 2000. Decoupling of taxonomic and ecologic severity of Phanerozoic marine mass extinctions. Geology 28, 675^ 678. Finney, S.C., Berry, W.B.N., Cooper, J.D., Ripperdan, R.L., Sweet, W.C., Jacobson, S.R., Sou¢ane, A., Achab, A., Noble, P.J., 1999. Late Ordovician mass extinction: A new perspective from stratigraphic sections in central Nevada. Geology 27, 215^218. Fischer, A.G., 1983. Long-term climatic oscillations recorded in stratigraphy. In: Berger, W.H. (Ed.), Climate in Earth History. National Academy of Sciences, pp. 97^104. Flu«gel, E., Kiessling, W., 2002. Patterns of Phanerozoic reef crisis. In: Kiessling, W., Flu«gel, E., Golonka, J. (Eds.), Phanerozoic Reef Patterns. SEPM Special Volume 72, 691^733. Fry, B., Jannasch, H.W., Molyneaux, S.J., Wirsen, C.O., Muramato, J.A., King, S., 1991. Stable isotopes of the carbon, nitrogen and sulfur cycles in the Black Sea and the Cariaco Trench. Deep-Sea Res. 38 (Suppl.), S1003^S1019. Gagnier, P.Y., Blieck, A., Emig, C.C., Sempere, T., Vachard,

PALAEO 3075 9-5-03

122

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

D., Vanguestaine, M., 1996. New paleontological and geological data on the Ordovician and Silurian of Bolivia. J. S. Am. Earth Sci. 9, 329^347. Gelsthorpe, D., 2002. Microplankton changes across a mass extinction: the Early Silurian Ireviken Event. IPC2002 Geol. Soc. Austr. Abstr. 86, 64. Glumac, B., Walker, K.R., 1998. A Late Cambrian positive carbon-isotope excursion in the Southern Appalachians; relation to biostratigraphy, sequence stratigraphy, environments of deposition, and diagenesis. J. Sed. Res. 68, 1212^ 1222. Glumac, B., Walker, K.R., 2000. Carbonate deposition and sequence stratigraphy of the terminal grand cycle in the Southern Appalachians. J. Sed. Res. 70, 952^963. Grahn, Y., Caputo, M.V., 1992. Early Silurian glaciations in Brazil. Palaeogeogr. Palaeoclimatol. Palaeoecol. 99, 9^15. Hallam, A., Wignall, P.B., 1999. Mass extinctions and sea-level changes. Earth Sci. Rev. 48, 217^250. Heath, R.J., Brenchley, P.J., Marshall, J.D., 1998. Early Silurian carbon and oxygen stable-isotope stratigraphy of Estonia: implications for climate change. In: Landing, E., Johnson, M.E. (Eds.), Silurian Cycles: Linkages of Dynamic Stratigraphy with Atmospheric, Oceanic, and Tectonic Changes (James Hall Centennial Volume). New York State Museum Bulletin 491, 313^323. Hede, J.E., 1925. Beskrivning av Gotlands silurlager. In: Munthe, H., Hede, H.E., von Post, L., Gotlands Geologi, º versikt. Sveriges Geol. Underso«kn. Ser. C 331, 13^30. en O Helfrich, C.T., 1980. Late Llandovery-early Wenlock conodonts from the upper part of the Rose Hill and the basal part of the Mi¥intown formations, Virginia, West Virginia, and Maryland. J. Paleontol. 54, 557^569. Hints, L., Oraspo‹ld, A., Kaljo, D., 2000. Stratotype of the Porkuni Stage with comments on the Ro«a Member (uppermost Ordovician, Estonia). Proc. Estonian Acad. Sci. Geol. 49, 177^199. Jenkyns, H.C., Clayton, C.J., 1997. Lower Jurassic epicontinental carbonates and mudstones from England and Wales: chemostratigraphic signals and the early Toarcian anoxic event. Sedimentology 44, 687^706. Jeppsson, L., 1984. Sudden appearances of Silurian conodont lineages; provincialism or special biofacies? In: Clark, D.L. (Ed.), Conodont Biofacies and Provincialism. Geol. Soc. Am. Spec. Pap. 196, 103^112. Jeppsson, L., 1987. Lithological and conodont distributional evidence for episodes of anomalous oceanic conditions during the Silurian. In: Aldridge, R.J. (Ed.), Palaeobiology of Conodonts. Ellis Horwood Ltd., Chichester, pp. 129^145. Jeppsson, L., 1990. An oceanic model for lithological and faunal changes tested on the Silurian record. J. Geol. Soc. London 147, 663^674. Jeppsson, L., 1997a. A new latest Telychian, Sheinwoodian and Early Homerian (Early Silurian) Standard Conodont Zonation. Trans. R. Soc. Edinb. Earth Sci. 88, 91^114. Jeppsson, L., 1997b. Recognition of a probable secundo-primo event in the Early Silurian. Lethaia 29, 311^315. Jeppsson, L., 1997. The anatomy of the Mid-Early Silurian

Ireviken Event and a scenario for P-S events. In: Brett, C.E., Baird, G.C. (Eds.), Paleontological Events: Stratigraphic, Ecological, and Evolutionary Implications. Columbia University Press, New York, pp. 451^492. Jeppsson, L., 1998. Silurian oceanic events: summary of general characteristics. In: Landing, E., Johnson, M.E. (Eds.), Silurian Cycles: Linkages of Dynamic Stratigraphy with Atmospheric, Oceanic, and Tectonic Changes (James Hall Centennial Volume). New York State Museum Bulletin 491, 239^257. Jeppsson, L., Ma«nnik, P., 1993. High-resolution correlations between Gotland and Estonia near the base of the Wenlock. Terra Nova 5, 348^358. Jeppsson, L., Calner, M. (2003). The Silurian Mulde Event and a scenario for secundo-secundo events. Trans. R. Soc. Edinb. Earth Sci., submitted. Jeppsson, L., Viira, V., Ma«nnik, P., 1994. Silurian conodontbased correlations between Gotland (Sweden) and Saaremaa (Estonia). Geol. Mag. 131, 201^218. Jeppsson, L., Aldridge, R.J., Dorning, K.J., 1995. Wenlock (Silurian) oceanic episodes and events. J. Geol. Soc. London 152, 487^498. Jeppsson, L., Mawson, R., Simpson, A.J., Talent, J.A., Andrew, A., Whitford, D.J., Corradini, C., 2002. The Lau global extinction event (Late Silurian): lithologic, conodont and isotope data and implications. IPC2002 Geol. Soc. Austr. Abstr. 86, 222^223. Johnson, M.E., 1996. Stable cratonic sequences and a standard for Silurian eustasy. In: Witzke, B.J., Ludvigson, G.A., Day, J. (Eds.), Paleozoic Sequence Stratigraphy: Views from the North American Craton. Geol. Soc. Am. Spec. Pap. 306, 203^211. Johnson, M.E., Kaljo, D., Rong, J.-Y., 1991. Silurian eustasy. In: Bassett M.G., Edwards, D. (Eds.), The Murchison Symposium ^ Proceedings of an International Conference on the Silurian System. Special Papers in Palaeontology 44, 145^ 163. Jones, B.G., Chenhall, B.E., Wright, A.J., Pemberton, J.W., Campbell, C., 1987. Silurian evaporitic strata from New South Wales, Australia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 59, 215^225. Kaljo, D., Boucot, A.J., Cor¢eld, R.M., Le He¤risse¤, A., Koren, T.N., Kriz, J., Ma«nnik, P., Ma«rss, T., Nestor, V., Shaver, R.H., Siveter, D.J., Viira, V., 1995. Silurian Bio-Events. In: Walliser, O.H. (Ed.), Global Events and Event Stratigraphy in the Phanerozoic. Springer, Berlin, pp. 173^223. Kaljo, D., Kiipli, T., Martma, T., 1997. Carbon isotope event markers through the Wenlock-Pridoli sequence at Ohesaare (Estonia) and Priekule (Latvia). Palaeogeogr. Palaeoclimatol. Palaeoecol. 132, 211^223. Kaljo, D., Kiipli, T., Martma, T., 1998. Correlation of carbon isotope events and environmental cyclicity in the East Baltic Silurian. In: Landing L., Johnson, M.E. (Eds.), Silurian Cycles ^ Linkages of Dynamic Stratigraphy with Atmospheric, Oceanic, and Tectonic Changes. New York State Museum Bulletin, pp. 297^312. Kaljo, D., Hints, L., Martma, T., No‹lvak, J., 2001. Carbon

PALAEO 3075 9-5-03

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124 isotope stratigraphy in the latest Ordovician of Estonia. Chem. Geol. 175, 49^59. Kiessling, W., 2002. Secular variations in the Phanerozoic reef system. In: Kiessling, W., Flu«gel, E., Golonka, J. (Eds.), Phanerozoic Reef Patterns. SEPM Special Volume 72, 650^690. Kump, L.R., Arthur, M.A., 1999. Interpreting carbon-isotope excursions: carbonates and organic matter. Chem. Geol. 161, 181^198. Kump, L.R., Arthur, M.A., Patzkowsky, M.E., Gibbs, M.T., Pinkus, M.T., Sheehan, P.M., 1999. A weathering hypothesis for glaciation at high atmospheric pCO2 during the Late Ordovician. Palaeogeogr. Palaeoclimatol. Palaeoecol. 152, 173^187. Laufeld, S., 1974. Silurian Chitinozoa from Gotland. Fossils Strata 5, 1^130. Le He¤risse¤, A., 1989. Acritarches et kystes d’algues Prasinophyce¤es du Silurien de Gotland. Sue'de. Palaeontogr. It. 76, 57^302. Le He¤risse¤, A., 2000. Characteristics of the acritarch recovery in the Early Silurian of Saudi Arabia. In: Al-Hajri S., Owens, B. (Eds.), Stratigraphic Palynology of the Palaeozoic of Saudi Arabia. Gulf PetroLink, Bahrain, pp. 57^81. Leggett, J.K., McKerrow, W.S., Cocks, L.R.M., Rickards, R.B., 1981. Periodicity in the early Palaeozoic marine realm. J. Geol. Soc. London 138, 167^176. Lenz, A.C., 1982. Ordovician to Devonian sea-level changes in western and northern Canada. Can. J. Earth Sci. 19, 1919^ 1932. Long, D.G.F., 1993. Oxygen and carbon isotopes and event stratigraphy near the Ordovician-Silurian boundary, Anticosti Island Quebec. Palaeogeogr. Palaeoclimatol. Palaeoecol. 104, 49^59. Manten, A.A., 1971. Silurian reefs of Gotland. Dev. Sedimentol. 13, 1^539. Marshall, J.D., Brenchley, P.J., Mason, P., Wol¡, G.A., Astini, R.A., Hints, L., Meidla, T., 1997. Global carbon isotopic events associated with mass extinction and glaciation in the late Ordovician. Palaeogeogr. Palaeoclimatol. Palaeoecol. 132, 195^210. Meidla, T., Ainsaar, L., Hints, L., Hints, O., Martma, T., No‹lvak, J., 1999. The mid-Caradocian biotic and isotopic event in the Ordovician of the east Baltic. Acta Univ. Carolinae Geol. 43, 503^506. Melchin, M.J., 1994. Graptolite extinction at the LlandoveryWenlock boundary. Lethaia 27, 285^290. Melezhik, V.A., Fallik, A.E., Medvedev, P.V., Makarikhin, V.V., 1999. Extreme 13 Ccarb enrichment in ca. 2.0 Ga magnesite-stromatolite-dolomite-‘red beds’ association in a global context a case for the world-wide signal enhanced by a local environment. Earth Sci. Rev. 48, 71^120. Mesolella, K.J., Robinson, J.D., McCormick, L.M., Ormiston, A.R., 1974. Cyclic deposition of Silurian carbonates and evaporites in Michigan Basin. AAPG Bull. 5b, 34^62. Mikulic, D.G., Kluessendorf, J., 1999. Stasis and extinction of Silurian (Llandovery-Wenlock) trilobite associations related to oceanic cyclicity. J. Paleontol. 73, 320^325.

123

Muir, L., 2002. The lundgreni (Wenlock, Silurian) graptoloid extinction event in the Welsh Borderland, UK. IPC2002 Geol. Soc. Austr. Abstr. 86, 246. Munnecke, A., 1997. Bildung mikritischer Kalke im Silur auf Gotland. Cour. Forschungsinst. Senckenb. 198, 1^71. Munnecke, A., Samtleben, C., Servais, T., Vachard, D., 1999. SEM-observation of calcareous micro- and nannofossils incertae sedis from the Silurian of Gotland, Sweden: Preliminary results. Geobios 32, 307^314. Munnecke, A., Servais, T., Vachard, D., 2000. A new family of calcareous microfossils from the Silurian of Gotland, Sweden. Palaeontology 43, 1153^1172. Munnecke, A., Servais, T., Vachard, D., 2001. New ¢ndings and stratigraphical distribution of the Ovummuridae (Palaeozoic calcareous microfossils). C.R. Acad. Sci. Paris Sci. Terre plane't. 333, 179^185. Nestor, V., 1998. Ireviken Event in the succession of chitinozoans (abstract). CIMP Newslett. 55, 29. Orth, J.J., Gilmore, J.S., Quintana, L.R., Sheehan, P.M., 1986. Terminal Ordovician extinction: geochemical analysis of the Ordovician-Silurian boundary, Anticosti Island, Quebec. Geology 14, 433^436. Paris, F., Bourahrouh, A., Le He¤risse¤, A., 2000. The e¡ects of the ¢nal stages of the Late Ordovician glaciation on marine palynomorphs (chitinozoans, acritarchs, leiospheres) in well N1^2 (NE Algerian Sahara). Rev. Palaeobot. Palynol. 113, 87^104. Patzkowsky, M.E., Slupic, L.M., Arthur, M.A., Pancost, R.D., Freeman, K.H., 1997. Late Middle Ordovician environmental change and extinction: Harbinger of the late Ordovician or continuation of Cambrian patterns? Geology 25, 911^914. Raup, D.M., Sepkoski, J.J., Jr., 1982. Mass extinctions in the marine fossil record. Science 215, 1501^1503. Riding, R., Watts, N., 1981. Silurian algal reef crest in Gotland. Naturwissenschaften 68, 91^92. Rong, J.-Y., Harper, D.A.T., 1999. Brachiopod survival and recovery from the latest Ordovician mass extinctions in South China. Geol. J. 34, 321^348. Saltzman, M.R., 2001. Silurian N13 C stratigraphy: A view from North America. Geology 29, 671^674. Saltzman, M.R., Runnegar, B., Lohmann, K.C., 1998. Carbon isotope stratigraphy of Upper Cambrian (Steptoean Stage) sequences of the eastern Great Basin; record of a global oceanographic event. Geol. Soc. Am. Bull. 110, 285^297. Saltzman, M.R., Ripperdan, R.L., Brasier, M.D., Lohmann, K.C., Robison, R.A., Chang, W.T., Peng, S., Ergaliev, E.K., Runnegar, B., 2000. A global carbon isotope excursion (SPICE) during the Late Cambrian: relation to trilobite extinctions, organic-matter burial and sea level. Palaeogeogr. Palaeoclimatol. Palaeoecol. 162, 211^223. Samtleben, C., Munnecke, A., 1999. Reefmounds im unteren Wenlock auf Gotland: Beispiele fru«her Korallenri¡e. Meyniana 51, 77^94. Samtleben, C., Munnecke, A., Bickert, T., Pa«tzold, J., 1996. The Silurian of Gotland (Sweden): Facies interpretation based on stable isotopes in brachiopod shells. Geol. Rundsch. 85, 278^292.

PALAEO 3075 9-5-03

124

A. Munnecke et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 99^124

Samtleben, C., Munnecke, A., Bickert, T., 2000. Development of facies and C/O-isotopes in transects through the Ludlovian of Gotland: Evidence for global and local in£uences on a shallow-marine environment. Facies 43, 1^38. Samtleben, C., Munnecke, A., Bickert, T., Pa«tzold, J., 2001. Shell succession, assemblage and species dependent e¡ects on the C/O-isotopic composition of brachiopods ^ examples from the Silurian of Gotland. Chem. Geol. 175, 61^107. Schmitz, B., Jeppsson, L., Ekvall, J., 1994. A search for shocked quartz grains and impact ejecta in early Silurian sediments on Gotland, Sweden. Geol. Mag. 131, 361^367. Scotese, C.R., McKerrow, W.S., 1990. Revised world maps and introduction. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geological Society of London Memoirs 12, 1^21. Sheehan, P.M., 1996. A new look at Ecologic Evolutionary Units (EEUs). Palaeogeogr. Palaeoclimatol. Palaeoecol. 127, 21^32. Sheehan, P.M., 2001. The Late Ordovician mass extinction. Annu. Rev. Earth Planet. Sci. 29, 331^364. Sou¢ane, A., Achab, A., 2000. Chitinozoan zonation of the Late Ordovician and the Early Silurian of the island of Anticosti, Que¤bec, Canada. Rev. Palaeobot. Palynol. 109, 85^111. Sutcli¡e, O.E., Dowdeswell, J.A., Whittington, R.J., Theron, J.N., Craig, J., 2000. Calibrating the Late Ordovician glaciation and mass extinction by the eccentricity cycles of Earth’s orbit. Geology 28, 967^970. Talent, J.A., Mawson, R., Andrew, A.S., Hamilton, P.J., Whitford, D.J., 1993. Middle Palaeozoic extinction events: Faunal and isotopic data. Palaeogeogr. Palaeoclimatol. Palaeoecol. 104, 139^152. Tuuling, I., Flode¤n, T., 2000. Late Ordovician carbonate buildups and erosional features northeast of Gotland, northern Baltic Sea. GFF 122, 237^249. Underwood, C.J., Crowley, S.F., Marshall, J.D., Brenchley, P.J., 1997. High-Resolution carbon isotope stratigraphy of the basal Silurian Stratotype (Dob’s Linn, Scotland) and its global correlation. J. Geol. Soc. Lond. 154, 709^718.

Ł lvaro, J.J., Villas, E., 1998. High-latitude pelmaVennin, E., A tozoan-bryozoan mud-mounds from the late Ordovician northern Gondwana platform. Geol. J. 33, 121^140. Wang, K., Orth, C.J., Attrep, M., Jr., Chatterton, B.D.E., Wang, X., Li, J., 1993. The great latest Ordovician extinction on the South China Plate: chemostratigraphic studies of the Ordovician boundary interval on the Yangtze Platform. Palaeogeogr. Palaeoclimatol. Palaeoecol. 104, 61^97. Webby, B.D., 2002. Patterns of Ordovician reef development. In: Kiessling, W., Flu«gel, E., Golonka, J. (Eds.), Phanerozoic Reef Patterns. SEPM Special Volume 72, 129^179. Weissert, H., Mohr, H., 1996. Late Jurassic climate and its impact on carbon cycling. Palaeogeogr. Palaeoclimatol. Palaeoecol. 122, 27^43. Wenzel, B., 1997. Isotopenstratigraphische Untersuchungen an silurischen Abfolgen und deren pala«ozeanographische Interpretation. Erlanger Geol. Abh. 129, 1^117. Wenzel, B., Joachimski, M.M., 1996. Carbon and oxygen isotopic composition of Silurian brachiopods (Gotland/Sweden): paleoceanographic implications. Palaeogeogr. Palaeoclimatol. Palaeoecol. 122, 143^166. Wigforss-Lange, J., 1999. Carbon isotope 13 C enrichment in Upper Silurian (Whitcli⁄an) marine calcareous rocks in Scania, Sweden. GFF 121, 273^279. Wilde, P., Berry, W.B.N., Quinby-Hunt, M.S., Orth, C.J., Quintana, L.R., Gilmore, J.S., 1986. Iridium abundances across the Ordovician-Silurian stratotype. Science 233, 339^341. Wilde, P., Berry, W.B.N., Quinby-Hunt, M.S., 1991. Silurian oceanic and atmospheric circulation and chemistry. Spec. Pap. Palaeontol. 44, 123^143. Williams, G.E., 1991. Milankovitch-band cyclicity in bedded halite deposits contemporaneous with Late-OrdovicianEarly Silurian glaciations, Canning Basin, Western Australia. Earth Planet. Sci. Lett. 103, 143^155. Zhang, T., Kershaw, S., Wan, Y., Lan, G., 2000. Geochemical and facies evidence for palaeoenvironmental change during the Late Ordovician Hirnantian glaciation in South Sichuan Province, China. Global Planet. Change 24, 133^152.

PALAEO 3075 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148 www.elsevier.com/locate/palaeo

The Armorica ‘microplate’: fact or ¢ction? Critical review of the concept and contradictory palaeobiogeographical data Michel Robardet  Ge¤osciences-Rennes, UMR CNRS 6118, Universite¤ de Rennes 1, Campus de Beaulieu, 35042 Rennes Cedex, France Received 27 May 2002; received in revised form 16 September 2002; accepted 15 January 2003

Abstract The problem of an ‘Armorica microplate’, detached from Gondwana and having had during part of the Palaeozoic an independent latitudinal evolution, is reconsidered in terms of a critical review of the palaeomagnetic data that were the very roots of this concept and through an alternative approach based on palaeoclimatic and palaeobiogeographical data. Palaeomagnetic data for the Silurian and the Devonian of the south European regions supposedly constituting the Armorica microplate remain rare and ambiguous. Those from Gondwana are more numerous but contradictory enough to give rise to diverging models regarding the latitudinal evolution of this continent. Consequently, the reality of an Armorica microplate cannot be considered as established. On the contrary, lithological indicators of palaeoclimate and palaeobiogeographical data are in total harmony and indicate that, in actual fact, the southern European regions remained permanently closely connected with Gondwana, of which they composed the northern margin. Although repeatedly maintained for more than 20 years, the concept of an Armorica microplate can thus be considered a fiction. This conclusion should lead to the dismissal of geodynamical models proposed for the Variscan Belt to which this concept was integrated and are contradicted by inescapable palaeobiogeographical constraints. 5 2003 Elsevier Science B.V. All rights reserved. Keywords: Palaeozoic; palaeogeography; Armorica microplate; palaeomagnetism; palaeobiogeography; Variscan belt

1. Introduction For more than 30 years it has been established that during the Early Palaeozoic and especially the Ordovician Period, the palaeogeography of the present-day peri-Atlantic regions (Fig. 1) comprised three major continents, Laurentia, Baltica

* Tel.: +33-2-23-23-61-05; Fax: +33-2-23-61-00. E-mail address: [email protected] (M. Robardet).

and Gondwana, situated in di¡erent latitudes and separated by the Iapetus and Rheic oceans. Subsequently, a more complex palaeogeographical con¢guration has been proposed, with intervening microplates whose evolution was supposedly independent of that of the major continents. The Armorica microplate concept is especially important because, quite apart from pure palaeogeographical problems, reconstruction and geodynamical evolution proposed for the Variscan ( = Hercynian) Belt of SW Europe vary greatly, depending on whether it is accepted or not.

0031-0182 / 03 / $ ^ see front matter 5 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00305-5

PALAEO 3076 9-5-03

126

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

The aim of the present paper is: (1) to propose a critical review of the palaeomagnetic data from which the concept of an Armorica microplate originated and persisted, (2) to restate the sedimentary and palaeontological data, frequently ignored or underestimated, that contrast sharply with this concept, and (3) to study the implications of this concept for the understanding of the Variscan Belt in SW Europe.

2. The concept of the Armorica microplate The existence, in Ordovician palaeogeography, of three major continents, Laurentia, Baltica and Gondwana (Fig. 1), was originally based on distinct faunal provinces, characterised by proper benthic faunas (especially trilobites and brachiopods) inhabiting the epicontinental marine shelves (Spjeldnaes, 1961; Whittington and Hughes, 1972; Williams, 1973, a.o.). The proper characteristics of these benthic faunas, the di¡erences in lithologies and biodiversity, allowed a rough estimation of their respective latitudes (Spjeldnaes, 1961; Webby, 1984; Scotese and Barrett, 1990; Witzke, 1990), which were subsequently corrobo-

Fig. 1. Ordovician palaeogeography of the present-day periAtlantic regions (after Cocks and McKerrow, 1993, simpli¢ed).

rated by palaeomagnetic data (see references in Van der Voo, 1993). Some years before the formal de¢nition of the Armorica microplate (Van der Voo, 1979, 1982), several authors had already suggested the possible existence of microplates in the pre-Variscan palaeogeography of SW Europe. These various models shared the view that all or part of southern Europe constituted an autonomous palaeogeographical unit, sandwiched between Gondwana and northern Europe (Baltica and later Laurussia), from which it was separated respectively by a ‘Proto-Mediterranean’ or ‘Tethys’ Ocean to the south, and a ‘Saxo-Thuringian’ or ‘mid-European’ Ocean to the north (Laurent, 1972; Johnson, 1973; Riding, 1974; Badham and Halls, 1975; Lorenz, 1976; Badham, 1982). However, this hypothesis did not become very popular until it was revived by palaeomagnetic arguments which apparently made it much more credible for a large number of geoscientists. The concept of a microplate or microcontinent Armorica, including southwestern England and Wales, the di¡erent regions of Variscan Europe and probably also the Avalon peninsula and New England, was based originally on palaeomagnetic data (Van der Voo, 1979, 1982; Jones et al., 1979). For the Late Proterozoic and the Cambrian, the palaeomagnetic data from the regions included in the microplate Armorica were similar to those from Gondwana, which showed that both were closely associated. Similar results were later obtained for the Ordovician (see references in Perroud et al., 1984a). Conversely, the palaeomagnetic data for the Late Devonian indicated that Armorica and Gondwana were in different latitudes and that Armorica was at that time juxtaposed with or very close to the southern border of Laurussia ( = Euramerica) that included northern Europe. It thus appeared that, between the Ordovician and the Late Devonian, Armorica had moved independently in latitude, which justi¢ed the status of microplate. The palaeogeographical organisation of both the earliest and the latest Palaeozoic did not pose important problems. On the other hand, the palaeogeographical evolution during the Late Ordovician, the Silurian and the Devonian

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

remained much more enigmatic. The scarcity of palaeomagnetic data for these periods and the ambiguity of some of them allowed various interpretations and scenarios. Some authors considered that Armorica had detached from Gondwana and had drifted northwards on its own; in this interpretation Gondwana had later followed a similar evolution, to join Armorica during the Carboniferous, after closure of the ocean that had separated them. The time schedule of these events remained uncertain, the collision of Armorica and Laurentia being placed either in the Late Ordovician (Van der Voo, 1979) or in the Devonian (Van der Voo et al., 1984). For some others, Gondwana and Armorica had ¢rst drifted northwards as a whole and joined the southern border of Laurussia (Laurentia+Baltica). They had then separated, Armorica holding its position whereas Gondwana retreated southwards, this second phase being accompanied by the opening of a ‘proto-Tethys’ ocean between Armorica and Gondwana. Finally, Gondwana had moved northwards again to join the palaeogeographical units already assembled to form the Late Palaeozoic Pangea. In the type of illustration favoured by the palaeomagnetists, this complex evolution corresponds to the ‘loops’ that appear in the apparent polar wander path (APWP) of Gondwana. However, all the proposed scenarios supposed that Armorica and Gondwana had been separated by a wide ocean at some period of the Palaeozoic. The geographical extent of the Armorica microplate was subsequently reduced, owing to the distinction of Avalonia (Fig. 1), a microplate de¢ned ¢rst on palaeontological arguments (Cocks and Fortey, 1982, 1990) and later on palaeomagnetic data (see references in Van der Voo, 1993; Bachtadse et al., 1995). The regions constituting Avalonia, i.e. northern Germany, Belgium, England, Wales and southern Ireland ( = Eastern Avalonia), together with eastern Newfoundland, Nova Scotia, New Brunswick and eastern New England ( = Western Avalonia) were originally included in the Armorica microplate (Van der Voo, 1979), although some had considered they were a southwestern extension of Baltica (Hughes et al., 1975). According to Cocks and Fortey (1982, 1990),

127

Avalonia was part of Gondwana until the early Ordovician; it detached from this major continent during the Ordovician, with opening of the Rheic Ocean, and moved northwards to join Baltica, with closure of the Tornquist Sea ( = Tornquist’s Ocean), either at the end of this period or in the Early Silurian (Cocks et al., 1997). This implied that the Armorica microplate sensu stricto should be restricted to the Variscan regions of southern and central Europe (Van der Voo, 1988).

3. Critical review of basic palaeomagnetic data A well-argued discussion of the concept of the Armorica microplate necessitates a review of the middle Palaeozoic palaeomagnetic data, both from Gondwana and from southern European Variscan regions, from which the autonomy of Armorica was inferred. First, it must be noted that the palaeogeographical conclusions derived from palaeomagnetic data are generally adopted without any reservation by those who make use of them to support geodynamical models. This is all the more surprising because most of the publications at the root of such palaeogeographical interpretations indicate clearly that ambiguities and problems exist and that the proposed conclusions are the result of a choice or a ‘preference’ between di¡erent hypotheses (see a.o. Scotese and McKerrow, 1990; Van der Voo, 1993). 3.1. Gondwana : Silurian and Devonian palaeomagnetic data 3.1.1. Silurian The basic and crucial data for the Silurian come from igneous rocks of the A|«r region that are part of a long series of igneous complexes extending through Niger and Nigeria. Some of these rocks, dated at ca 435 Ma, contain a complex magnetisation with a component, considered ‘primary’ and of Early Silurian age, that gives a South Pole position in southern Chile and places North Africa in tropical latitudes (Hargraves et al., 1987; Van Houten and Hargraves, 1987; AIR in Fig. 2). This implies that, in the ‘middle’

PALAEO 3076 9-5-03

128

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

cian and the Early Silurian or an erroneous interpretation of the age of the magnetisation observed in the A|«r. Moreover, Moreau et al. (1994) have recently published new geochronological data on the A|«r igneous complexes that give a radiometric age of ca 410 Ma, corresponding to the Silurian^Devonian boundary. With this revised age, the A|«r palaeopole position in southern Chile is rather close to the palaeopole obtained by Schmidt et al. (1987; see below) from the Lower Devonian Snowy River Volcanics of SE Australia (pole SV in Fig. 2).

Fig. 2. Late Ordovician through Late Devonian palaeopole positions deduced from African (dots) and Australian (triangles) palaeomagnetic data (see references in the text). Gondwana reassembly after Bachtadse and Briden (1990). Symbols: AIR, A|«r; BC, Brewer Conglomerate; BO, Bokkeveld Group; BZ, Ben Zireg; CB1 and CB2, Canning Basin; CV, Comerong Volcanics; GB and GC, Gilif Hills; GN, Gneiguira; HG, Hervey Group; MS, Msissi norite; PC, Pakhuis^ Cedarberg formations; SA, Sabakola; SV, Snowy River Volcanics; WP, Worange Point Formation.

Palaeozoic, North Africa was in latitudes similar to those of the southern border of Baltica. This model has been widely reproduced in subsequent palaeogeographical syntheses, although it supposes an extraordinarily high plate velocity (25^ 40 cm/yr, according to various authors) totally unknown in the recent periods, and despite the similar position of Africa in the Carboniferous^ Permian (Bachtadse and Briden, 1991). It can also be noted that Bachtadse et al. (1987) have studied sedimentary rocks slightly older, i.e. of latest Ordovician age, in the western Cape Fold Belt of South Africa. The study of the Pakhuis and Cedarberg Shale formations suggests a pole position in westernmost North Africa (PC in Fig. 2), at a large distance from that deduced from the A|«r (Bachtadse et al., 1987, ¢gure 13). This would imply either an enormous and very fast latitudinal change of Gondwana between the latest Ordovi-

3.1.2. Devonian For more than 20 years, the Late Devonian latitudinal position of Gondwana has been based on the palaeomagnetic data obtained from the Msissi norite in Morocco (Hailwood, 1974). This fundamental reference (pole MS in Fig. 2) was ruled out when it was established that the age of the rock was in fact close to 140 Ma and that no component of its magnetisation could thus correspond to the Devonian (Salmon et al., 1986). However, a localisation of the Devonian South Pole in Central Africa was maintained on the basis of palaeomagnetic data obtained mainly from Australia, and also from Africa. With regard to the Australian data, several problems exist, related (1) to possibly con£icting results obtained in di¡erent parts of the Lachlan Fold Belt of SE Australia, and (2) to their representativeness for Gondwana as a whole after reassembly of the southern continents. 3.1.2.1. SE Australia. The Early Devonian Snowy River Volcanics contain a multicomponent magnetisation with a pre-folding component regarded as acquired during the initial cooling of these rocks. The corresponding Early Devonian palaeopole (SV in Fig. 2) lies SSW of the southern extremity of South America (Schmidt et al., 1987). The Comerong Volcanics of late Middle to early Late Devonian age (ca 370 Ma) were studied by Schmidt et al. (1986). In spite of the di⁄cult study of the complex magnetisation and of the somewhat di¡erent behaviour of samples during demagnetisation processes, Schmidt et al. (1986)

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

considered that the palaeopole corresponding to the pre-folding component was situated west of southern Argentina (CV in Fig. 2) and was a key pole position for Gondwana as a whole in the Middle^Late Devonian. Famennian red sandstones of the Hervey Group have yielded a magnetisation component that indicates a pole position either at the western limit of Central Africa (Li et al., 1988, ¢gure 8) or in westernmost Brazil (Bachtadse and Briden, 1990, ¢gure 1), depending on the model used for reassembly of the southern continents. This pole position (HG in Fig. 2) is clearly distinct from that proposed by Schmidt et al. (1986) for a slightly older period (see above). As no fold test was possible in the Hervey Group, the age of the magnetisation, according to Li et al. (1988), could possibly be Early Carboniferous. The magnetisation of the red sandstones of the Late Devonian Worange Point Formation is dominated by a component that post-dates the mid-Carboniferous folding. However, about onethird of the samples studied by Thrupp et al. (1991) also contain a component that seems to be pre-folding although a fold test is not clearly conclusive. The corresponding pole position (WP in Fig. 2) is rather close to that obtained by Li et al. (1988) from the Famennian rocks of the Hervey Group (HG in Fig. 2). 3.1.2.2. Cratonic Australia. Some results have been also obtained from Devonian rocks of the stable craton of NW Australia, speci¢cally from the Frasnian and Famennian reefal limestones of the Canning Basin (Hurley and Van der Voo, 1987). These results indicate a palaeopole position in central Africa (CB1 in Fig. 2). However, in the absence of a fold test, the age of the characteristic magnetisation could be chemical and Early Carboniferous rather than depositional and Late Devonian (see Schmidt et al., 1987, pp. 146^147; 1990, p. 94). Chen et al. (1995) have reinvestigated various lithofacies of these limestones. Their conclusions are that the characteristic component of magnetisation is of primary origin and that the resulting pole position (CB2 in Fig. 2) is rather close to the Late Devonian palaeopole positions obtained from SE Australia and overlaps that ob-

129

tained originally by Hurley and Van der Voo (1987). Later on, the upper part (Undandita Member) of the Brewer Conglomerate from central Australia was studied by Chen et al. (1993). These rocks are dated as late Famennian, very close to the Devonian^Carboniferous boundary (ca 355 Ma), by spore assemblages. The samples collected from cores previously drilled in the northeast of the Amadeus Basin contain a multicomponent magnetisation. One component is interpreted as drilling-induced; the second one, found only in four samples from a single core, could be a mid to Late Carboniferous overprint; the third one, considered primary and of latest Devonian^earliest Carboniferous age, gives a pole position (BC in Fig. 2) in Africa, close to the pole CB1 obtained by Hurley and Van der Voo (1987) from the Canning Basin. 3.1.2.3. Africa. African data come from South Africa, Sudan, Algeria and Mauritania. In the Western Cape Fold Belt of South Africa, Bachtadse et al. (1987) studied brown and red sandstones from the Bokkeveld Group, considered to be Early to Middle Devonian, probably Middle Devonian as the overlying beds of the Witterberg Group have yielded Givetian fossils. Only a small percentage of the samples studied (13 out of 119) from a single site (out of 17), gave a magnetisation component di¡erent from the present-day ¢eld direction. Although the corresponding pole, situated in Central Africa (BO in Fig. 2), is not far from that deduced from the Canning Basin, the Bokkeveld data are of rather poor quality (Bachtadse and Briden, 1991, p. 637) and the deduced pole should not reasonably be considered as truly conclusive. In the Gilif Hills of the Bayuda Desert (northern Sudan), Bachtadse and Briden (1991) have studied acidic intrusive and e¡usive rocks of mid-Devonian age (377 Ma, Rb/Sr). One of the three components of magnetisation observed is considered to be a Carboniferous overprint that gives a pole position to the south of South America (GB in Fig. 2). Another component was supposedly acquired during or shortly after the closure of the Rb/Sr system and would be represen-

PALAEO 3076 9-5-03

130

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

tative for the Middle Devonian. This component, recognised in 11/18 sites and 39/114 samples, indicates a pole position in southwestern Libya (GC in Fig. 2). However, it must be noted that these results are not constrained by fold tests and come from a region where at least ¢ve pronounced episodes of volcanic activity occurred between 574 and 157 Ma, as mentioned by Bachtadse and Briden (1991, p. 638). More recently, Chen et al. (1993) have suggested a reverse interpretation for these results and considered that (1) magnetisation GC originally interpreted as primary would rather be an Early Carboniferous overprint, and conversely that (2) the magnetisation interpreted as a Carboniferous overprint could be primary, the corresponding pole GB being situated between the Early Devonian Snowy River Volcanics pole (SV) and the early Late Devonian Comerong Volcanics pole (CV). Famennian ‘griotte’ limestones from Ben Zireg in the northern part of the Be¤char Basin (Algeria) have been studied by A|«fa et al. (1990). In addition to a syn- or post-folding component of magnetisation, these authors have observed a prefolding component corresponding to a pole position in southern Africa that places North Africa at a latitude of ca 35‡S. This pole (BZ in Fig. 2) di¡ers strongly from the pole position in Central Africa proposed by the above-mentioned studies. Curiously, this pole BZ is mentioned very rarely in more recent publications, although it is rated as high-quality (‘quality factor’ Q = 6 in Van der Voo, 1993, table A4). Kent et al. (1984) have investigated Early^Middle Devonian reddish sandstones of the Gneiguira Supergroup from the westernmost part of the Taoudeni Basin (Mauritania). The magnetisation component considered ‘characteristic’ corresponds to a pole position to the east of South Africa (GN in Fig. 2) and places the northern part of this continent in low latitudes (ca 20‡S). However, Kent et al. (1984) considered that this component could be a Carboniferous remagnetisation. This was also noted by Kent and Keppie (1988, p. 474), although they favoured the other interpretation. Van der Voo (1988, p. 318) considered that this pole should not be used for the Devonian, an opinion apparently shared by Kent and Van der

Voo (1990), who did not maintain this reference in their list of selected poles for Gondwana. So¡el et al. (1990) published a complete revision of the palaeomagnetic data obtained from the Sabakola ring complex of northern Sudan. Several isotopic studies (K/Ar and Rb/Sr) of different rocks of the complex have given various ages (see references in So¡el et al., 1990, p. 413), but it was considered that they point to an Early to Middle Devonian age. The component of magnetisation considered ‘primary’ indicates a pole position (SA in Fig. 2) to the southeast of South Africa, i.e. not far from the Gneiguira pole of Kent et al. (1984), but also conspicuously close to the Permo^Carboniferous South Pole. Moreover, So¡el et al. (1990) did not exclude the possibility that this component of magnetisation could be Early to Middle Carboniferous, and this pole was excluded for the Devonian by Van der Voo (1993, table A4). 3.1.3. Checkup of the palaeomagnetic data for Gondwana When an attempt is made at compiling a synthesis of the Devonian palaeomagnetic data set for Gondwana, it seems virtually impossible for a non-specialist to make a selection of the most reliable data, even after careful reading of the original publications. Almost every study includes some uncertainties, most frequently concerning the complex structure of the magnetisation and the precise age of the ‘characteristic’, ‘primary’, component that is at the basis of the pole position proposed. It can also be noted that this complexity and these uncertainties, although generally mentioned, are not very precisely documented in synthetic and review papers. However, it has been repeatedly noted by most palaeomagnetists (e.g. Van der Voo, 1993; Bachtadse et al., 1995; Tait et al., 2000) that the APWP for Gondwana during mid-Palaeozoic times, i.e. the drift history of this huge continent, is insu⁄ciently resolved and still remains a controversial matter of debate. All models agree that the South Pole was within, or close to, NW Africa in the Ordovician, and in Antarctica by the Late Carboniferous^Early Permian. The problems concern the intervening mid-Palaeozoic times, and, essentially, two con-

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

trasting models are proposed. The ¢rst one considers a regular and steady northward drift of Gondwana from the Late Ordovician up to the Carboniferous, which translates into a simple APWP for Gondwana (‘Path X’ ; see Fig. 3). The alternative model is much more complex and implies a northward movement (Ordovician to Late Silurian^Early Devonian), followed by a southward ‘retreat’ (Early Devonian to Late Devonian^Early Carboniferous), and, ¢nally, a northward drift (during the Carboniferous). This model translates into a complex APWP for Gondwana (‘Path Y’; see Fig. 3) characterised by two hair-pin loops. In the 1990s, the more complex model was favoured by most palaeomagnetists (e.g. Bachtadse and Briden, 1990; Kent and Van der Voo, 1990; Van der Voo, 1993), although it required very high drift rates and two drastic changes in the movement of the Gondwana plate. More recently, the hypothesis of a steady northward drift of

Fig. 3. Late Ordovician through Late Devonian palaeomagnetically derived APWPs for Gondwana. Black curve: ‘path Y’ after Bachtadse and Briden (1990). Double-lined curve: ‘path X’ after Bachtadse and Briden (1991), Tait et al. (2000) (see references in the text). Abbreviations: O, Ordovician; S, Silurian; D, Devonian; C, Carboniferous; P, Permian; l, lower; m, middle; u, upper.

131

Gondwana and a simpler model (similar to ‘Path X’) have regained more advocates (Bachtadse and Briden, 1991; Bachtadse et al., 1995; and especially Tait et al., 2000), who take account of the probable Early Devonian age of the A|«r palaeopole and consider that the data set from SE Australia should not be used to de¢ne the APWP for Gondwana as a whole. However, a pole position in Central Africa is maintained for the whole Devonian period (Tait et al., 2000, ¢gure 3) on the basis of the palaeomagnetic results obtained from cratonic Australia and from Sudan. 3.2. Armorica: Silurian and Devonian palaeomagnetic data 3.2.1. Silurian For the Silurian period, the only palaeomagnetic data for SW Europe come from Spain in the Iberian Peninsula. Basaltic lavas interstrati¢ed within the Llandovery graptolitic black shales of the Almade¤n syncline of the southern Central Iberian Zone have yielded a pre-folding magnetisation considered to be of Early Silurian age. It indicates that the Iberian Peninsula and thus Armorica were in tropical latitudes at that time (Perroud et al., 1991), which ¢ts well with the A|«r pole. However, as already noted by Perroud et al. (1991), the problem of plate velocity arises, as the Armorica microplate should have moved from relatively high latitudes in the latest Ordovician (Hirnantian glaciomarine deposits) into the tropical latitudes suggested for the Llandovery. It has been shown later that this magnetisation was in fact a prefolding remagnetisation of Late Devonian or Early Carboniferous age (Pare's and Van der Voo, 1992). In the Cantabrian Zone, red beds of the San Pedro Formation (close to the Silurian^Devonian boundary) have also yielded a pre-folding component of magnetisation that corresponds to a palaeolatitude of ca 20‡S (Perroud and Bonhommet, 1984). However, despite a positive fold test, the age of this component is loosely controlled and could correspond to any time between the Silurian^Devonian boundary and the middle Carbon-

PALAEO 3076 9-5-03

132

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

iferous (see discussion in Van der Voo, 1993; Van der Voo et al., 1997). Tait et al. (1994) have studied Late Silurian (late Wenlock to late Pridoli) rocks from the Barrandian Basin of central Bohemia (Czech Republic). In addition to a present-day overprint (component A), they isolated two distinct components of magnetisation. The ¢rst one (component B) is post-folding and corresponds to the Late Carboniferous pole position for stable Europe. The other one (component C) passes a fold test and is interpreted as representative of the Late Silurian latitudinal position (23‡S) of Bohemia. Their conclusion is that this part of the Armorica microplate should thus be adjacent to the southern border of Laurussia during the late Wenlock^late Pridoli time interval. However, Tait et al. (1994) noted that these results would imply large-scale (140‡) anticlockwise rotation of the Bohemian Massif before acquisition of the post-folding B component. In addition, it can be noted also that (1) only just half of the studied sites (eight out of 21) and samples (48 out of 106 for B, 56 out of 111 for C) could be used to obtain the mean palaeopoles, and (2) apparently there was no signi¢cant di¡erence in the site means with respect to the age of the rocks studied when the late Wenlock^late Pridoli time interval was long enough to allow recording of latitudinal evolution. 3.2.2. Devonian Tait (1999) reported palaeomagnetic results, from well dated Lower Devonian limestones (L’Armorique Formation, Lochkovian^lower Praguian) of the westernmost Armorican Massif, which she considered indicative of a very low latitude (19‡S) for this part of the Armorica microplate. However, in the absence of any fold test, the only control of the age of the magnetisation being a contact test with a dyke whose precise age is unknown, Tait’s interpretation can hardly be accepted at face value, and it is most probable that these data correspond to a Carboniferous remagnetisation (see more complete discussion in Robardet et al., 2001). Tait and Bachtadse (2000) have recently studied the palaeomagnetism of well dated latest Silurian to Early Devonian limestones from two localities

of the Eastern Pyrenees (EP) and the Catalonian Coastal Ranges (CCR) in NE Spain (Rueda Formation, La Creu and Olorda formations, respectively). These limestones contain a multicomponent magnetisation. The ¢rst component (A) is a present-day overprint. The second component (B), identi¢ed only in the CCR, is post-folding but its age remains uncertain and could be either Late Carboniferous or Early Triassic. The third component (C) is present both in the EP and CCR. A fold test is not totally signi¢cant but suggests a pre-folding age, which is also supported by a positive ‘regional’ fold test when the data from the two regions are combined. This component C corresponds to a pole position that places the CCR and EP at a latitude of about 30‡S in the latest Silurian^Early Devonian. This is considered by Tait and Bachtadse (2000) as similar to the results obtained from the Lower Devonian of the Armorican Massif (Tait, 1999) and the Upper Silurian of Bohemia (Tait et al., 1994). However, it must be noted, as already mentioned by Tait and Bachtadse (2000, p. 23 600), that this component C could alternatively be a Cretaceous or more recent overprint. It is not surprising that such di⁄culties of interpretation are encountered in those regions of the Variscan Belt where a more recent Meso^ Cenozoic tectonic development occurred, and it seems rather curious to have selected these regions for testing the palaeogeographical position of the Iberian Peninsula in Late Silurian^Early Devonian times. The palaeomagnetic results obtained in the Montmartin syncline (Normandy, Armorican Massif) from the red beds of the Hyenville Formation (Jones et al., 1979) were at the very root of the Armorica microplate concept. These red beds, at that time considered Upper Devonian, had yielded a nearly univectorial magnetisation corresponding to very low latitudes (about 8‡S). This magnetisation was similar to the post-folding component obtained from Cambrian and lower Ordovician rocks in the same region (Jones et al., 1979; Perroud et al., 1982). However, due to misappreciation of the geological evolution of the region, especially of the Carboniferous age of the folding, and despite the absence of a fold test, this magnetisation was considered representative for

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

the Armorican Massif in the Late Devonian. The corresponding low latitude was similar to that of northern Europe and very di¡erent from that of Gondwana, which gave rise to the concept of the Armorica microplate. It was later demonstrated, by a negative fold test, that this magnetisation was younger than the Carboniferous folding of the Montmartin syncline, and, additionally, that there were no data proving a Late Devonian age for the Hyenville Formation, which should be considered most probably Early Ordovician in age (Perroud et al., 1984b). The magnetisation observed in metamorphosed volcanic rocks (dated Late Devonian : 379^355 Ma) of southern Limousin in the French Massif Central (Edel, 1987) corresponds to a rather high latitude, but these data from a region where metamorphism was important cannot be interpreted easily (see discussion in Bachtadse et al., 1995). Non-metamorphic and very low-grade metamorphic sedimentary and volcanic rocks of Middle to Late Devonian and Early Carboniferous ages have been studied in the Harz Mountains, the Franconian Forest and the southern Vosges (Bachtadse et al., 1983, 1995). These rocks are situated within three distinct ‘zones’ of the Variscan Belt, that are the Rheno-Hercynian, SaxoThuringian and Moldanubian zones, respectively. The pre-folding palaeomagnetic data indicate similar low latitudes of about 13‡S for the three regions in the Middle^Late Devonian. Bachtadse et al. (1983, 1995) concluded that, at that time, the drift history of Avalonia and Armorica had come to an end and that they were already assembled with Laurussia. 3.2.3. Checkup of palaeomagnetic data for Armorica This critical review shows that uncontroversial palaeomagnetic data for the Silurian and the Devonian of the Variscan regions of southern and central Europe, that constituted the so-called Armorica microplate, are very rare, not to say missing. Armorica is now considered generally as a complex collage of terranes (the ‘Armorican Terrane Assemblage’ ; see references in Franke et al., 2000), this assumption being supposedly sup-

133

ported by the palaeomagnetic results obtained in Late Silurian and Early Devonian rocks from Bohemia, NE Spain and the Armorican Massif. In fact, neither the composite character of Armorica, nor its breako¡ from Gondwana and its independent latitudinal evolution, can be considered convincingly demonstrated by the existing palaeomagnetic results. This is because (1) the age of the magnetisations observed in NE Spain and in the Armorican Massif remains ambiguous, which makes any comparison with those from Bohemia fruitless, and (2) the drift history of Gondwana during the Silurian and the Devonian still remains obscure.

4. Palaeoclimatically based geography and palaeobiogeography An alternative approach to palaeogeographical evolution can be based on sedimentary and faunal characteristics which allow identi¢cation and delimitation of distinct palaeogeographical units and estimation of their palaeolatitudes. Some lithofacies are indicators of palaeoclimate and thus of palaeolatitude (e.g. Webby, 1984; Scotese and Barrett, 1990; Witzke, 1990) and can be used to estimate the latitudinal evolution of a region or a larger unit. Benthic faunas allow evaluation of the respective relations and proximity of palaeogeographical units, which is not possible with palaeomagnetical data that do not provide any constraint on longitude. The benthic faunas that lived on the epicontinental marine shelves had limited dispersive abilities (only through their pelagic larvae) and generally could not cross wide and deep oceanic areas. They are therefore especially important in characterising the di¡erent marine shallow-water shelves, in evaluating their proximity or distance, and in evidencing the ‘faunal barriers’ that controlled the geographical distribution of the palaeobenthos (e.g. Cocks and Fortey, 1982, 1990; Paris and Robardet, 1990). Due to their important dispersive abilities, pelagic and planktonic organisms are mainly used for precise biostratigraphical control, but they can also be helpful for palaeogeography because

PALAEO 3076 9-5-03

134

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

their geographical distribution was controlled by sea-water temperature and latitude (e.g. Cocks and Fortey, 1982; Paris, 1993). On the basis of sedimentary and faunal data, it is therefore possible to reconstitute the respective pre-Carboniferous latitudinal evolution of Gondwana and Armorica, and to clarify the problem of their relationship. 4.1. Lithofacies indicators of palaeolatitudes 4.1.1. Ordovician On the North African part of Gondwana, the almost total absence of carbonates during the Ordovician (except bryozoan biostromes of Ashgill age in Libya and Morocco), the overall terrigenous composition of the sediments, and the glaciomarine deposits of the latest Ordovician (Hirnantian) related to the continental glaciation developed in Africa, indicate that the North Gondwanan regions were situated in high latitudes during the whole Ordovician period. The same conclusions can be drawn for southern European Variscan regions, where the Ordovician successions are of the same type and include in their uppermost part glaciomarine deposits strictly equivalent to those of North Africa (Robardet and Dore¤, 1988). In these Variscan regions, the only episode of carbonate sedimentation occurred in the early^middle Ashgill, with bryozoan and pelmatozoan limestones considered to be temperate- or cold-water carbonates (Prasada Rao and Jayawardane, 1994; Vennin et al., 1998). 4.1.2. Silurian During the Silurian, the anoxic sediments that characterise both North Africa and southern Europe are not by themselves latitude indicators. However, considering the rather short duration of this period and the high latitudes in the Late Ordovician, it seems highly improbable that any part of these regions might have reached low latitudes during the Silurian. This is corroborated by the absence or scarcity of carbonate deposits in North Africa and southern Europe when, at the same time, these lithofacies developed (Bassett, 1989; Bassett et al., 1989) in Avalonia and Baltica

(Wales, Gotland, Estonia, Podolia). In southern Europe, the ¢rst carbonates appeared locally in the Wenlock and their importance increased in the Ludlow and the Pridoli in the areas (Pyrenees, Catalonia, Montagne Noire) situated in the northern distal part of the shelf (see Robardet et al., 1994; Gutie¤rrez-Marco et al., 1998). At that time they were still missing in the Ibero^Armorican regions situated in the inner shelf and in North Africa (Hollard and Willefert, 1985; Legrand, 1985). It can also be added that brachiopod taxa representative of the cold-water ‘Clarkeia fauna’ occur in the Pridoli of the Armorican Massif (Babin et al., 1979). All these data contrast sharply with the models that place, in the Silurian, North Africa and/or Armorica in low latitudes similar to those of the southern border of northern Europe. 4.1.3. Devonian In the southern European regions, the Lower Devonian successions comprise well developed carbonates, with an increasing importance in the middle and upper parts of the System. This is the case in the Armorican Massif, the Iberian Peninsula, Pyrenees and Montagne Noire (see references in Robardet et al., 1994), and also in the Barrandian area of Bohemia and the Carnic Alps, where the carbonates are more important and reefs more frequent (Chlupac, 1988; Scho«nlaub, 1993). These data do not match the models where Armorica is placed in very low latitudes in the earliest Devonian (Tait, 1999). Such models would suppose that Armorica had crossed tropical latitudes during the Silurian, an hypothesis which ¢nds no support in the sedimentary successions of this age. Armorica did not reach relatively low latitudes, favourable to a well developed carbonate sedimentation, before the Devonian. In North Africa (Moroccan Anti-Atlas, Algerian Sahara, Libya) carbonates remain of minor importance in the Lower Devonian and are not actually well developed before the Middle and Late Devonian (Wendt, 1988; Bitam et al., 1996; Gourvennec et al., 1997; Boumendjel et al., 1997a). The possible occurrence of glacial or peri-gla-

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

cial rocks in the Upper Devonian or the Lower Carboniferous of Central Africa has sometimes been considered as supporting a polar position of this region at these times. This assumption was based on Caputo and Crowell’s (1985) publication on the migration of glacial centres across Gondwana during the Palaeozoic. Glacial deposits were mentioned in the mid-Famennian of northern South America (Brazil) and it was suggested that similar rocks possibly occurred in Africa, especially in the Agades region of Niger (Caputo and Crowell, 1985, with previous references therein). A recent study rather favours an Early Carboniferous (early Vise¤an) age for the rocks of the Agades region and suggests that they could correspond to small local icecaps at a latitude of ca 50‡S (Lang et al., 1991). The expansion of carbonate and evaporite sedimentation in the Late Devonian and Early Carboniferous of North Africa is consistent with the steady northward drift of Gondwana and is in con£ict with the hypothesis that places the central part of Africa near the South Pole at this time (Witzke and Heckel, 1988). 4.1.4. Palaeoclimatically derived APWP for Gondwana Scotese and Barrett (1990) have established a statistical technique, based on the geographical distribution of lithological indicators of climate (carbonates, evaporites, coals and tillites), to estimate the past position of the geographic pole. This technique was used to obtain a palaeoclimatically derived APWP of the South Pole across Gondwana during the Palaeozoic, established on eight pole positions corresponding to eight time intervals between the Early Cambrian and the Early Permian (Scotese and Barrett, 1990, ¢gure 13). This APWP (Fig. 4) agrees much better with the palaeomagnetic data that place the Devonian South Pole in Argentina than with the results that place it in Central Africa. In addition, the parallel sedimentary evolution observed in North Africa and in the Variscan regions of southern and central Europe gives no support to the proposal that these regions have been separated at any time between the Late Ordovician and the Late Devonian.

135

Fig. 4. Ordovician through Permian palaeoclimatically derived APWP for Gondwana (after Scotese and Barrett, 1990, simpli¢ed), compared with palaeomagnetically derived path Y (dashed line). Ca, Cambrian; O and o, Ordovician; S, Silurian; D and d, Devonian; C and c, Carboniferous; P and p, Permian; E and e, early; M and m, middle; L and l, late.

4.2. Faunas and £oras 4.2.1. Silurian In the Silurian successions of both southern^ central Europe and North Africa, pelagic organisms dominate the fossil record and do not provide precise palaeobiogeographical data. This is so for the acritarchs, whose patterns of distribution are controlled by latitude, but also by environmental conditions and other factors (Le He¤risse¤ and Gourvennec, 1995; Le He¤risse¤ et al., 1995). However, several Early Silurian (Telychian) graptolite species, such as Metaclimacograptus asejradi, Metaclimacograptus £amandi, and Parapetalograptus meridionalis, are good biogeographical markers for the shallowest areas of the Silurian marine shelf. These species occur in the Central Iberian Zone, in the Iberian Cordillera, as well as in North Africa (Algeria, Libya) and suggest close links between these regions within the inshore areas of the North Gondwanan shelf (Gutie¤rrez-Marco et al., 1998). 4.2.2. Devonian During the Devonian, sediments deposited in oxygenated environments and their corresponding benthic marine faunas were abundant and diver-

PALAEO 3076 9-5-03

136

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

si¢ed, which allows comparisons between various regions of North Africa and Europe and evaluation of their a⁄nities. 4.2.2.1. Armorica^Gondwana relations. The very high values of the coe⁄cients of similarity (CS = 0.90) for Early Devonian (Lochkovian) chitinozoan assemblages preclude any important difference in latitude and also the existence of a wide ocean between the Armorican Massif and the Algerian Sahara. Conversely, these coe⁄cients are much weaker (CS = 0.37^0.48) between these regions and northern Europe (Podolia and Poland) situated at the southern margin of Laurussia (Paris, 1993). This suggests that there was a noticeable di¡erence of latitude for North Gondwana and northern Europe. It had already been noted that invertebrate benthic faunas (and also vertebrates) from Variscan Europe and North Africa show so many af¢nities that it is not possible to imagine that they were separated by a wide ocean or any other faunal barrier during the Devonian (Young, 1987, 1990; Morzadec et al., 1988; Paris and Robardet, 1990; Robardet et al., 1990, 1991, 1993; Racheboeuf, 1990). However, these data have generally been ignored or underestimated, and, despite some slight doubts (Van der Voo, 1993, p. 248; Torsvik et al., 1990, p. 38), the concept of an Armorica microplate separated from Gondwana has been maintained by most palaeomagnetists. More recently, palaeontological investigations have been carried out in the southern £ank of the Tindouf Basin (Bitam et al., 1996; Gourvennec et al., 1997) and in the Ougarta Ranges (Boumendjel et al., 1997b; Plusquellec et al., 1997) of the Algerian Sahara, which were unambiguously part of ‘stable’ Gondwana. They have entirely corroborated and more fully illustrated the strong faunal a⁄nities that existed between Gondwana and southern Europe during the Devonian (Figs. 5 and 6). The Early Devonian vertebrate record (jawless and jawed ‘¢shes’ ; Blieck, 1982; Young, 1987, 1990) shows that: (1) faunas from Laurussia were dominated by actinolepids, pteraspids and cephalaspids; and (2) these groups were absent or very poorly represented in the southern Euro-

pean faunas (Spain, Portugal, Armorican Massif) which are closely similar, in taxonomic composition, to those from Morocco, Algeria and Libya, where the ichthyofauna was dominated by other placoderms, sharks and various acanthodians. These data suggest that southern Europe was part of Gondwana and still separated from Laurussia by a faunal barrier during the Early Devonian. In the Middle and Late Devonian, weaker di¡erences indicate closer proximity between southern Europe^Gondwana and Laurussia and therefore contradict the hypothesis of a wide ocean between Laurussia and Gondwana. In the Lochkovian, the Pragian and the Emsian, a number of invertebrate benthic fossils occur both in the Saharan regions and in Variscan Europe, especially in the Iberian Peninsula and the Armorican Massif and also, to a lesser degree, in Bohemia (Figs. 5 and 6). This is established for corals, various families of brachiopods, crinoids, trilobites and ostracods (Lethiers and Raymond, 1993; Boumendjel et al., 1997b; Plusquellec et al., 1997; Le Menn, 1997; Plusquellec, 1998; Plusquellec and Hladil, 2001). It must be emphasised that, in many cases, these a⁄nities are established on the occurrence of the same species, which shows the absence of any ‘reproductive isolation’ within the benthic populations and allows de¢nition of an ‘Ibarmaghian Domain’ (Plusquellec, 1987) comprising the Saharan regions, the Iberian Peninsula and the Armorican Massif. 4.2.2.2. Armorica^Laurussia and Gondwana^Laurussia relations. It has already been noted that during the Early Devonian the faunal a⁄nities between southern Europe and the southern border of Laurussia increased progressively (Paris and Robardet, 1990; Robardet et al., 1990). During the Pridoli and the Lochkovian, southern European benthic faunas were clearly distinct from those found in northern France (Boulonnais, Artois, Ardenne), in the autochthonous units of the Rhenohercynian Zone (Belgium and Germany) and in all the regions that were at the southern border of the Old Red Sandstone Continent ( = Laurussia). At these times there were virtually no species in common and the rare a⁄nities observed concern only a few genera, which

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

137

Fig. 5. Geographical distribution of some Early Devonian ostracod and trilobite species (see references in the text). Symbols: asterisks, Gibba schmidti; circles, Phacops (Prokops) benziregensis benziregensis; squares, Metacanthina lips; triangles, Treveropyge wallacei. Geographical key-symbols: AA, Moroccan Anti-Atlas; Ar, Ardenne; Bh, Bohemia; Bo, Boulonnais; Br, Britanny; BZ, Ben Zireg; CA, Carnic Alps; CIZ, Central Iberian Zone; CZ, Cantabrian Zone; Ei, Eifel; HA, Moroccan High Atlas; Hz, Harz; M, Maider; MM, Moroccan Meseta; MN, Montagne Noire; No, Normandy; OMZ, Ossa Morena Zone; Ou, Ougarta; Py, Pyrenees; Ti, Tindouf Basin.

shows that the two areas were still separated by the faunal barrier of the Rheic Ocean. From the Pragian onwards, there was an increasing number of brachiopod, coral, trilobite and crinoid species common to both domains, which strongly suggests that the width of the Rheic Ocean was decreasing (Figs. 5^7). Through the Middle and Late Devonian, the rugose corals of the ‘Eastern Americas Realm’ show a decreasing endemism that corresponds to faunal migrations between Laurussia and Africa^ southern Europe (Oliver, 1977; Plusquellec et al., 1997). Studies of Devonian chonetacean brachiopods (Racheboeuf, 1990) have shown that: (1) the identity, at the species level, of Early Devonian faunas from southern Europe and northern Sahara pre-

cludes the existence of any faunal barrier between these regions; and (2) during the Middle Devonian, faunal migrations from the ‘Eastern Americas Realm’ into the Saharan regions do not support at all the hypothesis of a wide ocean between Laurussia and Gondwana. Similar conclusions arise from the geographical distribution of Givetian ostracods (Lethiers and Racheboeuf, 1993). The benthic faunas from the Meguma terrane of Nova Scotia can also be mentioned. The Upper Silurian faunas of this area had northern European a⁄nities corresponding to a localisation in the southern part of Laurussia. The presence of typical southern European and northern Gondwanan species of trilobites, crinoids and brachiopods in the Lochkovian to early Emsian faunas of Meguma also implies faunal migrations and rela-

PALAEO 3076 9-5-03

138

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

Fig. 6. Geographical distribution of some Early Devonian brachiopod and coral species (see references in the text). Symbols: black diamonds, Eucharitina oehlerti; open triangles, Ctenochonetes jouannensis; asterisks, Ctenochonetes aremoricensis; black dots, Renaudia mainensis; black squares, Davoustia davousti; open diamonds, Plicanoplia carlsi; open dots, Plicanoplia alani; open squares, Procterodictyum polentinoi. Geographical key-symbols as in Fig. 5.

tionships between these regions from the Early Devonian onwards (Bouyx et al., 1997). Young (1987, p. 292) noted similarities of the Late Devonian shallow marine ¢shes from SE Morocco with those from eastern North America (especially large arthrodires) and Baltica (osteolepiform sarcopterygians). Close proximity of Gondwana and Laurussia during the late Middle and Late Devonian is attested also by plant micro- and macrofossil evidence. The geographical distribution of land plant miospores in the Givetian and Frasnian shows that NW Gondwana (Libya, Brazil) and southern Laurussia (Boulonnais, Ardenne, Rhenish regions) had similar vegetation patterns and similar climatic conditions (Streel et al., 1990 with references therein). Moreover, Meyer-Berthaud et al. (1997) reported, in the lower Famennian of SE

Morocco, the occurrence of a large trunk of Callixylon erianum, a species already known in North America, which provides a supplementary and clear evidence of a⁄nities between Gondwanan and Laurussian £oras in the Late Devonian. It must be noted that these palaeobotanical data both concern the Late Devonian, i.e. precisely the time interval when, according to palaeomagnetic models, the ocean supposedly separating Gondwana and Laurussia would have reached its maximum width. 4.3. Checkup of palaeobiogeographical data All the sedimentological and palaeontological data summarised above (Figs. 4^7) lead to the same conclusions: (1) There is no argument in favour of a sepa-

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

139

Fig. 7. Geographical distribution of some Middle and Late Devonian species (see references in the text). Symbols: open dots, Polyzygia neodevonica; open squares, Salairocrinus kerevensis; asterisks, Mediocrinus squi⁄ecensis; open diamonds, Laudonomphalus regularis; black squares, Eutaxocrinites kergarvanensis; black triangles, Gilbertsocrinus tenuis; open triangles, Centrorhynchus letiensis. Geographical key-symbols as in Fig. 5.

ration between the Variscan regions of southwestern and central Europe ( = the so-called Armorica microplate) and Gondwana at any moment of the Palaeozoic. On the contrary, the geographical distribution and the evolution of lithofacies and, furthermore, the persistent a⁄nities of planktonic organisms, various groups of marine invertebrates, vertebrates and £oras show that all these regions remained closely linked during the whole Palaeozoic. (2) Neither is there any argument that could suggest association or proximity of Armorica and Baltica^Avalonia (or later Laurussia) during the ‘middle’ Palaeozoic (i.e. between the Late Ordovician and the Early Devonian); these two palaeogeographical units remained separated by the Rheic Ocean during the Ordovician, the Silurian and the earliest Devonian and closure of this oce-

anic faunal barrier did not begin before the Pragian. These conclusions are totally at variance with palaeogeographical models, mostly based on palaeomagnetic data, which purport to show that the southern European regions formed a microplate Armorica derived from Gondwana and consider that it had a separate evolution and history, distinct from that of Gondwana, during a large part of the Palaeozoic.

5. Implications for the Variscan Belt of SW Europe Agreeing (or not) with the concept of a microplate Armorica, separated from Gondwana during a large part of the Palaeozoic, has important

PALAEO 3076 9-5-03

140

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

implications for reconstructing the Late Palaeozoic con¢guration of the Variscan Belt of SW Europe (Fig. 8) and understanding its formation. Accepting the existence of a microplate implies indeed that there were two distinct major oceans in the pre-orogenic palaeogeography, the ¢rst one separating Armorica from Baltica^Avalonia (and later from Laurussia), and the second one between Armorica and Gondwana. This means that two distinct oceanic sutures should be identi¢ed within the Variscan Belt. Conversely, when it is considered that Armorica was permanently part of Gondwana, the formation of the belt should have resulted from the closure of a single ocean separating Gondwana from Baltica^Avalonia (and later from Laurussia). This problem has become even more complicated and confused because a number of authors have turned aside from the original signi¢cation of the term ‘Armorica’ and adapted this denomination in their own way. In its original de¢nition, emended after distinction of Avalonia, the Armorica microplate was supposedly composed of the Variscan regions of southern Europe, i.e. the Iberian Peninsula, most of France, part of Germany, Bohemia and southern Poland (Van der Voo, 1988, 1993 p. 249). Considering the recent literature, it appears that ‘Armorica’ may include all these Variscan regions of southern and central Europe, but it is frequently considered to have been an ‘archipelago’ of semi-autonomous terranes separated by minor oceanic basins rather than a coherent microplate ( = Armorican Terrane Assemblage; Tait et al., 1994; Tait, 1999; Tait and Bachtadse, 2000; Crowley et al., 2000). The hypothesis of several blocks situated between Gondwana and Laurussia reaches its maximum extent in the models proposed by Ziegler (1990, with previous references therein), who considers that the Palaeozoic crustal consolidation of western and central Europe resulted from the stepwise accretion to Laurussia of various Gondwana-derived microcratons. Following the proposed occurrence of two distinct oceanic sutures (see below) within the Variscan Belt (Matte, 1986a,b, 1991), several authors have reserved the denomination ‘Armorica’ for

the regions sandwiched between these two sutures (Fig. 8). In this case, despite some variations, it is considered generally that: (1) Armorica comprised the middle and northern regions of the Armorican Massif, part of the Iberian Peninsula, Saxo-Thuringia and Bohemia ; and (2) the other southern European regions (the southern part of the Armorican Massif, southern France, part of the Iberian Peninsula, Corsica^Sardinia) either composed a separate microplate or were included in Gondwana (e.g. Rey et al., 1997; Faure et al., 1997; Matte, 2001). In less frequently quoted models, Armorica is not distinguished clearly from Avalonia and the middle and northern regions of the Armorican Massif are associated with the Ardenne (Neugebauer, 1989; Lefort, 1990; Pique¤, 1991; Pique¤ et al., 1994). In these models, southern France and Iberia are also considered either to be part of the Gondwana margin or a separate additional microplate ‘Iberia’. These examples illustrate the very real confusion that prevails as regards the signi¢cation of the term ‘Armorica’. It is largely beyond the scope of the present paper to discuss in detail the various and con£icting models that have been proposed for the Variscan Belt of SW Europe (for such a discussion see Robardet, 2002). However, it can be noted here that, until now, most of the models proposed have been based on syn- to late-orogenic structural, metamorphic and magmatic data, and have generally paid very little attention to the pre-Variscan palaeogeographical constraints. A large number of these models have uncritically incorporated the concept of the Armorica microplate, and, in£uenced also by present-day geography, have proposed the existence of two distinct oceanic sutures corresponding to two distinct oceans in the pre-Variscan palaeogeography (Fig. 8): ^ the Rheic Ocean, extending between the Variscan regions of southern Europe and those of northern Europe, whose suture would be found, in France, to the north of the Mid^North Armorican Domain, and, ^ the South-Armorican Ocean ( = Ligerian Ocean or Massif Central Ocean), whose suture, partly obliterated by the South Armorican Shear

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

Zone, would be situated between the Mid^North Armorican and the South Armorican regions, and would be prolongated into the Massif Central. In these models, it is also generally assumed that the prolongation of both oceanic sutures can be followed within the Iberian Peninsula : the former between the Ossa Morena and South Portuguese zones (Pulo de Lobo or Beja suture) and the latter running o¡shore west of Galicia and prolongating between the Central Iberian and Ossa Morena zones as a ‘cryptic’ suture (Fig. 8) obliterated by the Badajoz^Co¤rdoba Shear Zone (e.g. Burg et al., 1981).

141

The above-mentioned results of palaeobiogeographical analysis allow such models to be dismissed because they are inconsistent with: (1) the global distribution of faunas and £oras during pre-Carboniferous times; and (2) the smaller-scale palaeogeographic relations between the Iberian Peninsula and the Armorican Massif (Paris and Robardet, 1977; Robardet et al., 1990; Paris, 1998). A model with a single major Variscan ocean, proposed by several authors (e.g. Quesada, 1991; Quesada et al., 1994; Mart|¤nez Cata¤lan et al., 1997) for the Iberian part of the Variscan Belt, also proves suitable for the French part of the belt

Fig. 8. Schematic map of the Variscan Belt in southwestern and central Europe, with tectonostratigraphic units, major shear zones and oceanic sutures proposed by various authors (see references in the text). Abbreviations: Tectonostratigraphic units: CIZ, Central Iberian Zone; CZ, Cantabrian Zone; GTOM, Galicia^Tra¤s os Montes Zone; LD, Ligerian Domain; MNAD, Mid^North Armorican Domain; MZ, Moldanubian Zone; OMZ, Ossa Morena Zone; RHZ, Rheno-Hercynian Zone; SPZ, South Portuguese Zone; STZ, Saxo-Thuringian Zone; WALZ, West Asturian^Leonese Zone. Shear Zones: BCSZ, Badajoz^Co¤rdoba Shear Zone; SASZ, South Armorican Shear Zone, northern branch (N), southern branch (S). Oceanic sutures: BPLS, Beja^Pulo de Lobo suture; CCS, Coimbra^Co¤rdoba suture; GS, Galicia suture; LS, Lizard suture; MCS, Massif Central suture; MTS, Mu«nchberg-Tepla¤ suture; RHS, Rheno-Hercynian suture; SBS, Southern Brittany suture. The stippled area corresponds to the restricted extent of the Armorica microplate in the models with two oceanic sutures (e.g. Matte, 2001; see text).

PALAEO 3076 9-5-03

142

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

and would much more e¡ectively match the palaeobiogeographical constraints (Robardet, 2002).

6. General discussion and conclusion For a non-specialist, it seems very di⁄cult to apply a reliability ¢lter on published palaeomagnetic results and thereby make a selection of the most conclusive data. This is all the more evident since, at present, there is no clear consensus within the palaeomagnetist community itself, which shows that the extant data do not provide unambiguous and decisive evidence on the respective latitudinal evolution of Gondwana and Armorica during the Silurian and the Devonian. As noted above, palaeomagnetists have proposed various and sometimes contrasting models, based on di¡erent data or on di¡ering interpretations of the same data. This de facto situation has been illustrated above by: (1) the reverse interpretation (primary vs overprint) of two components of magnetisation identi¢ed in Mid-Devonian rocks from Sudan (Bachtadse and Briden, 1991; Chen et al., 1993); (2) the Early Silurian vs Early Devonian age of the A|«r palaeopole (Hargraves et al., 1987; Moreau et al., 1994); and (3) the opposite views concerning the use of SE Australia data as representative for Gondwana as a whole (Schmidt et al., 1990; Tait et al., 2000). Moreover, it appears that the quantitative character of palaeomagnetic data can be somewhat illusory for the palaeolatitude where the rocks were formed because, as a general rule for the Palaeozoic, secondary components have frequently been added to the primary magnetisation. Consequently, the main problem that a¡ects the reliability of the palaeomagnetic data concerns the precise age of the supposed primary magnetisation. The ¢eld tests used to constrain this age provide important constraints but no absolute certainty. For instance, a positive fold test only shows that the magnetisation was acquired before folding, but not necessarily when the rock was formed, because this test cannot identify a possible pre-folding remagnetisation (see e.g. the case of the Silurian volcanic rocks at Almade¤n). The advantage of the palaeoclimatic and pa-

laeobiogeographical data is that their age is unequivocal. These data can be reduced to nothing by penetrative deformation and/or strong metamorphism, but, when preserved, their age has never thereby been modi¢ed. Moreover, these data, although rarely quanti¢ed by means of similarity coe⁄cients of faunal assemblages, allow more precise evaluation of the respective relations and proximity of the regions studied than palaeomagnetism, which allows only comparison of their palaeolatitudes. This is most probably one of the reasons why Scotese and McKerrow (1990, p. 14) have considered palaeoclimatic indicators more weighty arguments for positioning the Devonian South Pole. The objective of the present paper was to reconsider the question of the Armorica microplate by: (1) a return to and a re-evaluation of the palaeomagnetic data that were at the very root of this concept ; and (2) an alternative approach based on palaeoclimatic and palaeobiogeographical data. A critical review shows that palaeomagnetic data for the Silurian and the Devonian are: (1) far too rare and still ambiguous for the southern European regions supposedly constituting the Armorica microplate ; and (2) more numerous for Gondwana but contradictory enough to give rise to various and diverging models for the latitudinal evolution of this continent. Consequently, the reality of an Armorica microplate, detached from Gondwana, and having had an independent latitudinal evolution, separate from that of Gondwana during part of the Palaeozoic (Fig. 9), cannot be considered as established. On the contrary, lithological indicators of palaeoclimate and palaeobiogeographical data, concerning both marine vertebrates and invertebrates, and also land plants, are in total harmony and indicate that the southern European regions have, in actual fact, remained permanently in close connection with Gondwana of which they composed the northern margin (Fig. 9). Although repeatedly maintained for more than 20 years, the concept of the Armorica microplate can thus be considered a ¢ction. These conclusions should lead to a re-evaluation of a number of the models proposed for

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

143

Fig. 9. Two di¡erent models of the Early Ordovician through Late Carboniferous palaeogeographical evolution of the presentday peri-Atlantic regions. Abbreviations: L, Laurentia; B, Baltica; G, Gondwana; Ar, Armorica; Av, Avalonia. Left side: based on palaeomagnetic data (after Tait, 1999 and Tait et al., 2000). Right side: based on palaeoclimatic and palaeobiogeographical data (after Paris and Robardet, 1990; Robardet et al., 1990; Paris, 1998).

PALAEO 3076 9-5-03

144

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

the Variscan Belt that had incorporated this ¢ctitious concept and are in contradiction with inescapable palaeobiogeographical constraints.

Acknowledgements Alain Blieck (Lille), Euan Clarkson (Edinburg) and Walter De Vos (Brussel) are thanked for their careful and constructive reviews, as well as Danie'le Bernard (CNRS Ge¤osciences-Rennes) for the illustrations. This work is a contribution to IGCP Projects 410 and 421 (IUGS-UNESCO) and to the Iberia project of EUROPROBE (European Science Foundation).

References A|«fa, T., Feinberg, H., Pozzi, J.P., 1990. Devonian-Carboniferous paleopoles for Africa: consequences for Hercynian geodynamics. Tectonophysics 179, 287^304. Babin, C., Deunf, J., Me¤lou, M., Paris, F., Pelhate, A., Plusquellec, Y., Racheboeuf, P., 1979. La coupe de Porz ar Vouden (Pridoli de la presqu’|“le de Crozon), Massif Armoricain, France, lithologie et biostratigraphie. Palaeontographica A 164, 52^84. Bachtadse, V., Briden, J.C., 1990. Palaeomagnetic constraints on the position of Gondwana during Ordovician to Devonian times. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, pp. 43^48. Bachtadse, V., Briden, J.C., 1991. Palaeomagnetism of Devonian ring complexes from the Bayuda Desert, Sudan - new constraints on the apparent polar wander path for Gondwanaland. Geophys. J. Int. 104, 635^646. Bachtadse, V., Heller, F., Kro«ner, A.M., 1983. Paleomagnetic investigation in the Hercynian mountain belt of central Europe. Tectonophysics 91, 285^299. Bachtadse, V., Van der Voo, R., Ha«lbich, I.W., 1987. Paleomagnetism of the western Cape Fold belt, South Africa, and its bearing on the Paleozoic apparent polar wander path for Gondwana. Earth Planet. Sci. Lett. 84, 487^499. Bachtadse, V., Torsvik, T.H., Tait, J.A., So¡el, H.C., 1995. Paleomagnetic constraints on the Paleogeographic evolution of Europe during the Paleozoic. In: Dallmeyer, R.D., Franke, W., Weber, K. (Eds.), Pre-Permian Geology of Central and Eastern Europe. Springer, Berlin, pp. 567^578. Badham, J.P.N., 1982. Strike-slip orogens - an explanation for the Hercynides. J. Geol. Soc. London 139, 493^504. Badham, J.P.N., Halls, C., 1975. Microplate tectonics, oblique collisions, and evolution of the Hercynian orogenic systems. Geology 3, 373^376.

Bassett, M.G., 1989. The Wenlock Series in the Wenlock area. In: Holland, C.H., Bassett, M.G. (Eds.), A Global Standard for the Silurian System. National Museum of Wales Geological Series 9, pp. 51^73. Bassett, M.G., Kaljo, D., Teller, L., 1989. The Baltic region. In: Holland, C.H., Bassett, M.G. (Eds.), A Global Standard for the Silurian System. National Museum of Wales Geological Series 9, pp. 158^170. Bitam, L., Gourvennec, R., Robardet, M., 1996. Les formations pale¤ozo|«ques ante¤-carbonife'res du sous-bassin de Djebilet (£anc sud du Bassin de Tindouf, Nord-Ouest du Sahara alge¤rien). Me¤m. Serv. Ge¤ol. Alge¤rie 8, 91^111. Blieck, A., 1982. Les grandes lignes de la bioge¤ographie des He¤te¤rostrace¤s du Silurien supe¤rieur-De¤vonien infe¤rieur dans le domaine Nord-Atlantique. Palaeogeogr. Palaeoclimatol. Palaeoecol. 38, 283^316. Boumendjel, K., Morzadec, P., Paris, F., Plusquellec, Y., 1997a. Le De¤vonien de l’Ougarta (Sahara Occidental, Alge¤rie). Ann. Soc. Ge¤ol. Nord 5, 73^87. Boumendjel, K., Brice, D., Copper, P., Gourvennec, R., Jahnke, H., Lardeux, H., Le Menn, J., Me¤lou, M., Morzadec, P., Paris, F., Plusquellec, Y., Racheboeuf, P., 1997b. Les faunes du De¤vonien de l’Ougarta (Sahara Occidental, Alge¤rie). Ann. Soc. Ge¤ol. Nord 5, 89^116. Bouyx, E., Blaise, J., Brice, D., De¤gardin, J.M., Goujet, D., Gourvennec, R., Le Menn, J., Lardeux, H., Morzadec, P., Paris, F., 1997. Biostratigraphie et pale¤obioge¤ographie du Siluro-De¤vonien de la zone de Meguma (Nouvelle-Ecosse, Canada). Can. J. Earth Sci. 34, 1295^1309. Burg, J.P., Iglesias, M., Laurent, Ph., Matte, Ph., Ribeiro, A., 1981. Variscan intracontinental deformation: the CoimbraCordoba Shear zone (SW Iberian Peninsula). Tectonophysics 78, 161^177. Caputo, M.V., Crowell, J.C., 1985. Migration of glacial centers across Gondwana during Paleozoic era. Geol. Soc. Am. Bull. 96, 1020^1036. Chen, Z., Li, Z.X., Powell, C.McA., Balme, B.E., 1993. Palaeomagnetism of the Brewer Conglomerate in central Australia and fast movement of Gondwanaland during the Late Devonian. Geophys. J. Int. 115, 564^574. Chen, Z., Li, Z.X., Powell, C.McA., 1995. Paleomagnetism of the Upper Devonian reef complexes, Canning Basin, Western Australia. Tectonics 14, 154^167. Chlupac, I., 1988. The Devonian of Czechoslovakia and its stratigraphical signi¢cance. In: McMillan, N.J., Embry, A.F., Gass, D.J. (Eds.), Devonian of the World. Can. Soc. Petrol. Geol. Mem. 14, pp. 481^497. Cocks, L.R.M., Fortey, R.A., 1982. Faunal evidence for oceanic separations in the Palaeozoic of Britain. J. Geol. Soc. London 139, 465^478. Cocks, L.R.M., Fortey, R., 1990. Biogeography of Ordovician and Silurian faunas. In: McKerrow W.S., Scotese C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, pp. 97^104. Cocks, L.R.M., McKerrow, W.S., 1993. A reassessment of the early Ordovician ‘Celtic’ brachiopod province. J. Geol. Soc. London 150, 1039^1042.

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148 Cocks, L.R.M., McKerrow, W.S., van Staal, C.R., 1997. The margins of Avalonia. Geol. Mag. 134, 627^636. Crowley, Q.G., Floyd, P.A., Winchester, J.A., Franke, W., Holland, J.G., 2000. Early Palaeozoic rift-related magmatism in Variscan Europe: fragmentation of the Armorican Terrane Assemblage. Terra Nova 12, 171^180. Edel, J.B., 1987. Paleopositions of the western Europe Hercynides during the Late Carboniferous deduced from paleomagnetic data: consequences for ‘stable Europe’. Tectonophysics 139, 31^41. Faure, M., Leloix, C., Roig, J.-Y., 1997. L’e¤volution polycyclique de la cha|“ne hercynienne. Bull. Soc. Ge¤ol. France 168, 695^705. Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), 2000. Orogenic Processes: Quanti¢cation and Modelling in the Variscan Belt. Geol. Soc. London Spec. Publ. 179, 459 pp. Gourvennec, R., Bitam, L., Robardet, M., 1997. Les formations pale¤ozo|«ques ante¤-carbonife'res du sous-bassin de l’Iguidi (Flanc sud du bassin de Tindouf, nord-ouest du Sahara Alge¤rien). Bull. Serv. Ge¤ol. Alge¤rie 8, 135^148. Gutie¤rrez-Marco, J.C., Robardet, M., Pic[arra, J.M., 1998. Silurian stratigraphy and paleogeography of the Iberian Peninsula (Spain and Portugal). Temas Geol.-Min. ITGE Madrid 23, 13^44. Hailwood, E.A., 1974. Paleomagnetism of the Msissi Norite (Morocco) and the Paleozoic reconstruction of Gondwanaland. Earth Planet. Sci. Lett. 23, 376^386. Hargraves, R.B., Dawson, E.M., Van Houten, F.B., 1987. Palaeomagnetism and age of mid-Palaeozoic ring complexes in Niger, West Africa, and tectonic implications. Geophys. J. R. Astron. Soc. 90, 705^729. Hollard, H., Willefert, S., 1985. Lower Paleozoic rocks of Morocco: Silurian. In: Holland, C.H. (Ed.), Lower Paleozoic Rocks of the World, Vol. 4. Wiley, London, pp. 238^321. Hughes, C.P., Ingham, J.K., Addison, R., 1975. The morphology, classi¢cation and evolution of the Trinucleidae (Trilobita). Philos. Trans. R. Soc. London B 272, 537^607. Hurley, N.F., Van der Voo, R., 1987. Paleomagnetism of Upper Devonian reefal limestones, Canning basin, Western Australia. Geol. Soc. Am. Bull. 98, 138^146. Johnson, G.A.L., 1973. Closing of the Carboniferous Sea in Western Europe. In: Tarling, D.H., Runcorn, S.K. (Eds.), Implications of Continental Drift to the Earth Sciences, vol. II. Academic Press, London, pp. 843^850. Jones, M., Van der Voo, R., Bonhommet, N., 1979. Late Devonian to Early Carboniferous palaeomagnetic poles from the Armorican Massif (France). Geophys. J. R. Astron. Soc. 58, 287^308. Kent, D.V., Keppie, J.D., 1988. Silurian-Permian palaeocontinental reconstryctions and circum-Atlantic tectonics. In: Harris, A., Fettes, D.J. (Eds.), The Caledonian-Appalachian Orogen. Geol. Soc. London Spec. Publ. 38, 469^480. Kent, D.V., Van der Voo, R., 1990. Palaeozoic palaeogeography from palaeomagnetism of the Atlantic-bordering continents. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, 49^56.

145

Kent, D.V., Dia, O., Sougy, J.M.A., 1984. Paleomagnetism of Lower-Middle Devonian and Upper Proterozoic-Cambrian (?) rocks from Mejeria (Mauritania, West Africa). In: Van der Voo, R., Scotese, C.R., Bonhommet, N. (Eds.), Plate Reconstruction from Paleozoic Paleomagnetism. Geodynamic Series, Am. Geophys. Union, Washington, DC, vol. 12, pp. 99^115. Lang, J., Yahaya, M., El Hamet, M.O., Besombes, J.C., Cazoulat, M., 1991. De¤po“ts glaciaires du Carbonife're infe¤rieur a' l’Ouest de l’A|«r (Niger). Geol. Rundsch. 80, 611^622. Laurent, R., 1972. The Hercynides of South Europe, a model. In: 24th Int. Geol. Congress, Montreal, Canada, sect. 3, pp. 363^370. Lefort, J.P., 1990. Cine¤matique de la collision Gondwana-Laurentia entre la Bretagne et la Floride d’apre's les donne¤es du socle submerge¤. Schweiz. Mineral. Petrogr. Mitt. 70, 3^ 16. Legrand, Ph., 1985. Lower Palaeozoic rocks of Algeria. In: Holland, C.H. (Ed.), Lower Paleozoic Rocks of the World, Vol. 4. Wiley, London, pp. 5^89. Le He¤risse¤, A., Gourvennec, R., 1995. Biogeography of upper Llandovery and Wenlock acritarchs. Rev. Palaeobot. Palynol. 86, 111^133. Le He¤risse¤, A., Al-Tayyar, H., Van der Eem, H., 1995. Stratigraphic and paleogeographic signi¢cance of Silurian acritarchs from Saudi Arabia. Rev. Palaeobot. Palynol. 89, 49^74. Le Menn, J., 1997. Crino|«des de¤voniens d’Afrique du Nord: re¤vision syste¤matique et a⁄nite¤s pale¤obiogeographiques. Ann. Soc. Ge¤ol. Nord 5, 129^139. Lethiers, F., Raymond, D., 1993. Nouvelles donne¤es stratigraphiques et pale¤oge¤ographiques dans le De¤vonien des Cha|“nes Ibe¤riques orientales (Espagne) par l’e¤tude des ostracodes. Ge¤obios 26, 319^329. Lethiers, F., Racheboeuf, P.R., 1993. Les ostracodes de la Formation de Kersadiou (Give¤tien du Massif Armoricain): biostratigraphie et pale¤oe¤cologie. Ge¤obios 26, 595^607. Li, Z.X., Schmidt, P.W., Embleton, B.J.J., 1988. Paleomagnetism of the Hervey Group, central New South Wales and its tectonic implications. Tectonics 7, 351^367. Lorenz, V., 1976. Formation of Hercynian subplates, possible causes and consequences. Nature 262, 374^377. Mart|¤nez Cata¤lan, J.R., Arenas, R., D|¤az Garc|¤a, F., Abati, J., 1997. Variscan accretionary complex of northwest Iberia: Terrane correlation and succession of tectonothermal events. Geology 25, 1103^1106. Matte, P., 1986a. La cha|“ne varisque parmi les cha|“nes pale¤ozo|«ques pe¤ri-atlantiques, mode'le d’e¤volution et position des grands blocs continentaux au Permo-Carbonife're. Bull. Soc. Ge¤ol. France 8, II, 1, 9^24. Matte, P., 1986b. Tectonics and plate tectonics model for the variscan belt of Europe. Tectonophysics 126, 329^374. Matte, P., 1991. Accretionary history and crustal evolution of the Variscan belt in Western Europe. In: Hatcher, R.D., Zonenshain, L. (Eds.), Accretionary Tectonics and Composite Continents. Tectonophysics 196, 309^337. Matte, P., 2001. The Variscan collage and orogeny (480^290

PALAEO 3076 9-5-03

146

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

Ma) and the tectonic de¢nition of the Armorica microplate: a review. Terra Nova 13, 122^128. Meyer-Berthaud, B., Wendt, J., Galtier, J., 1997. First record of a large Callixylon trunk from the late Devonian of Gondwana. Geol. Mag. 134, 847^853. Moreau, C., Demai¡e, D., Bellion, Y., Boullier, A.-M., 1994. A tectonic model for the location of Palaeozoic ring complexes in A|«r (Niger, West Africa). Tectonophysics 234, 129^ 146. Morzadec, P., Paris, F., Plusquellec, Y., Racheboeuf, P., Weyant, M., 1988. Devonian stratigraphy and palaeogeography of the Armorican Massif (NW France). In: McMillan, N.J., Embry, A.F., Gass, D.J. (Eds.), Devonian of the World. Can. Soc. Petrol. Geol. Mem. 14, pp. 401^420. Neugebauer, J., 1989. The Iapetus model: a plate tectonic concept for the Variscan belt of Europe. Tectonophysics 169, 229^256. Oliver, W.A., 1977. Biogeography of Late Silurian and Devonian rugose corals. Palaeogeogr. Palaeoclimatol. Palaeoecol. 22, 85^135. Pare's, J.M., Van der Voo, R., 1992. Paleozoic paleomagnetism of Almaden, Spain: a cautionary note. J. Geophys. Res. B 97, 9353^9356 and 14425. Paris, F., 1993. Evolution pale¤oge¤ographique de l’Europe au Pale¤ozo|«que infe¤rieur: le test des Chitinozoaires. C.R. Acad. Sci. Paris 316, 273^280. Paris, F., 1998. Early Palaeozoic palaeobiogeography of northern Gondwana regions. Acta Univ. Carolinae - Geol. 42, 473^483. Paris, F., Robardet, M., 1977. Pale¤oge¤ographie et relations ibe¤ro-armoricaines au Pale¤ozo|«que ante¤-Carbonife're. Bull. Soc. Ge¤ol. France 19, 1121^1126. Paris, F., Robardet, M., 1990. Early Palaeozoic palaeobiogeography of the Variscan regions. Tectonophysics 177, 193^213. Perroud, H., Bonhommet, N., 1984. A Devonian palaeomagnetic pole for Armorica. Geophys. J. R. Astron. Soc. 77, 839^845. Perroud, H., Bonhommet, N., Robardet, M., 1982. Comment on ‘A palaeomagnetic study of Cambrian red beds from Carteret, Normandy, France’ by W.A. Morris. Geophys. J. R. Astron. Soc. 69, 573^578. Perroud, H., Van der Voo, R., Bonhommet, N., 1984a. Paleozoic evolution of the Armorica plate on the basis of paleomagnetic data. Geology 12, 579^582. Perroud, H., Robardet, M., Van der Voo, R., Bonhommet, N., Paris, F., 1984b. Revision of the age of magnetization of the Montmartin red beds, Normandy, France. Geophys. J. R. Astron. Soc. 80, 541^549. Perroud, H., Calza, F., Khattach, D., 1991. Paleomagnetism of the Silurian volcanism at Almaden, Southern Spain. J. Geoph. Res. 96, 1949^1962. Pique¤, A., 1991. Les massifs anciens de France, segments de la Cha|“ne Varisque d’Europe occidentale. Sci. Ge¤ol. Strasb. Bull. 44, 371^385. Pique¤, A., Bogdano¡, S., Que¤nardel, J.M., Rolet, J., Santallier D., 1994. The French Paleozoic terranes. In: Keppie, J.D.

(Ed.), Pre-Mesozoic Geology in France and Related Areas. Springer, pp. 483^500. Plusquellec, Y., 1987. Re¤vision de Michelinia transitoria KNOD. Ann. Soc. Ge¤ol. Nord 105, 249^252. Plusquellec, Y., 1998. Pleurodictyidae de l’Emsien supe¤rieur des Monts d’Ougarta (Alge¤rie). Ann. Soc. Ge¤ol. Nord 6, 5^23. Plusquellec, Y., Hladil, J., 2001. Tabulate corals of Ibarmaghian a⁄nities in the Upper Emsian of Bohemia. Geol. Palaeontol. 35, 31^51. Plusquellec, Y., Boumendjel, K., Morzadec, P., Paris, F., 1997. Les faunes de¤voniennes d’Ougarta dans la pale¤oge¤ographie des re¤gions maghre¤bo-europe¤ennes. Ann. Soc. Ge¤ol. Nord 5, 123^128. Prasada Rao, C., Jayawardane, M.P.J., 1994. Major minerals, elemental and isotopic composition in madern temperate shelf carbonates, Eastern Tasmania, Australia: Implications for the occurrence of extensive ancient non-tropical carbonates. Palaeogeogr. Palaeoclimatol. Palaeoecol. 107, 49^63. Quesada, C., 1991. Geological constraints on the Paleozoic tectonic evolution of tectonostratigraphic terranes in the Iberian Massif. Tectonophysics 185, 225^245. Quesada, C., Fonseca, P.E., Munha, J., Oliveira, J.T., Ribeiro, A., 1994. The Beja-Acebuches Ophiolite (Southern Iberia Variscan fold belt): Geological characterization and geodynamic signi¢cance. Bol. Geol. Min. 105, 3^49. Racheboeuf, P., 1990. Pale¤obioge¤ographie de la marge nordgondwanienne au De¤vonien infe¤rieur et moyen: nouvelles donne¤es de¤duites de l’e¤tude des Brachiopodes Chonetace¤s. C.R. Acad. Sci. Paris 310, 1481^1486. Rey, P., Burg, J.P., Casey, M., 1997. The Scandinavian Caledonides and their relationship to the Variscan belt. In: Burg, J.P., Ford, M. (Eds.), Orogeny Through Time. Geol. Soc. London Spec. Publ. 121, 179^200. Riding, R., 1974. Model of the Hercynian foldbelt. Earth Planet. Sci. Lett. 24, 125^135. Robardet, M., 2002. Alternative approach to the Variscan Belt in southwestern Europe: Pre-orogenic paleobiogeographical constraints. In: Mart|¤nez Cata¤lan, J.R., Hatcher, R.D., Arenas, R., D|¤az Garc|¤a, F. (Eds.), Variscan-Appalachian Dynamics: The Building of the Late Paleozoic Basement. Geol. Soc. Am. Spec. Paper 364, pp. 1^15. Robardet, M., Dore¤, F., 1988. The Late Ordovician diamictic formations from southwestern Europe: north Gondwana glaciomarine deposits. Palaeogeogr. Palaeoclimatol. Palaeoecol. 66, 19^31. Robardet, M., Paris, F., Racheboeuf, P.R., 1990. Palaeogeographic evolution of southwestern Europe during Early Palaeozoic times. In: Mc Kerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, 411^419. Robardet, M., Gross-U¡enorde, H., Gandl, J., Racheboeuf, P.R., 1991. Trilobites et Ostracodes du De¤vonien infe¤rieur de la Zone d’Ossa Morena (Espagne). Ge¤obios 24, 333^348. Robardet, M., Blaise, J., Bouyx, E., Gourvennec, R., Lardeux, H., Le He¤risse¤, A., Le Menn, J., Me¤lou, M., Paris, F., Plusquellec, Y., Poncet, J., Re¤gnault, S., Rioult, M., Weyant,

PALAEO 3076 9-5-03

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148 M., 1993. Pale¤oge¤ographie de l’Europe occidentale de l’Ordovicien au De¤vonien. Bull. Soc. Ge¤ol. France 164, 683^695. Robardet, M., Verniers, J., Feist, R., Paris, F., 1994. Le Pale¤ozo|«que ante¤-varisque de France, contexte pale¤oge¤ographique et ge¤odynamique. Ge¤ol. France 1994, 3^31. Robardet, M., Paris, F., Plusquellec, Y., 2001. Comment on ‘New Early Devonian paleomagnetic data from NW France: Paleogeography and implications for the Armorican plate hypothesis’ by J. Tait. J. Geoph. Res. 106, 13,307^13,310. Salmon, E., Montigny, R., Edel, J.B., Pique¤, A., Thuizat, R., Westphal, M., 1986. The Msissi Norite revisited: K/Ar dating, petrography and paleomagnetism. Geophys. Res. Lett. 13, 741^743. Schmidt, P.W., Embleton, B.J.J., Cudahy, T.J., Powell, C.McA., 1986. Prefolding and premegakinking magnetizations from the Devonian Comerong Volcanics, New South Wales, Australia, and their bearing on the Gondwana pole path. Tectonics 5, 135^150. Schmidt, P.W., Embleton, B.J.J., Palmer, H.C., 1987. Pre- and post-folding magnetizations from the early Devonian Snowy River Volcanics and Buchan Caves limestone, Victoria. Geophys. J. R. Astron. Soc. 91, 155^170. Schmidt, P.W., Powell, C.McA., Li, Z.X., Thrupp, G.A., 1990. Reliability of Palaeozoic palaeomagnetic poles and APWP of Gondwana. Tectonophysics 184, 87^100. Scho«nlaub, H.P., 1993. Stratigraphy, Biogeography and climatic relationships of the Alpine Palaeozoic. In: Raumer, J.F. von, Neubauer, F.N. (Eds.), Pre-Mesozoic Geology in the Alps. Springer, Berlin, pp. 65^91. Scotese, C.R., Barrett, S.F., 1990. Gondwana’s movement over the South Pole during the Palaeozoic: evidence from lithological indicators of climate. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Paleogeography and Biogeography. Geol. Soc. London Mem. 12, pp. 75^85. Scotese, C.R., McKerrow, W.S., 1990. Revised world maps and introduction. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, 1^21. So¡el, H.C., Saradeth, S., Briden, J.C., Bachtadse, V., Rolf, C., 1990. The Sabaloka ring complex revisited: palaeomagnetism and rockmagnetism. Geophys. J. Int. 102, 411^420. Spjeldnaes, N., 1961. Ordovician climatic zones. Nor. Geol. Tidsskr. 41, 45^77. Streel, M., Fairon-Demaret, M., Loboziak, S., 1990. GivetianFrasnian phytogeography of Euramerica and western Gondwana based on miospo“re distribution. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, 291^296. Tait, J., 1999. New Early Devonian paleomagnetic data from NW France: Paleogeography and implications for the Armorican microplate hypothesis. J. Geophys. Res. 104, 2831^ 2839. Tait, J., Bachtadse, V., 2000. Paleomagnetism of Siluro-Devonian sequences, NE Spain. J. Geoph. Res. 105, 23595^23603. Tait, J., Bachtadse, V., So¡el, H., 1994. Silurian paleogeography of Armorica: New paleomagnetic data from central Bohemia. J. Geoph. Res. 99, 2897^2907.

147

Tait, J., Scha«tz, M., Bachtadse, V., So¡el, H., 2000. Palaeomagnetism and Palaeozoic palaeogeography of Gondwana and European terranes. In: Franke, W., Haak, V., Oncken, O., Tanner, D. (Eds.), Orogenic Processes: Quanti¢cation and Modelling in the Variscan Belt. Geol. Soc. London Spec. Publ. 179, 21^34. Thrupp, G.A., Kent, D.V., Schmidt, P.W., Powell, C.McA., 1991. Palaeomagnetism of red beds of the Late Devonian Worange Point Formation, SE Australia. Geophys. J. Int. 104, 179^201. Torsvik, T.H., Smethurst, M.A., Briden, J.C., Sturt, B.A., 1990. A review of Palaeozoic palaeomagnetic data from Europe and their palaeogeographical implications. In: McKerrow, W.S., Scotese, C.R. Eds., Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, 25^41. Van der Voo, R., 1979. Palaeozoic assembly of Pangea: a new plate tectonic model for the Taconic, Caledonian and Hercynian orogenies (abstract). EOS Trans. AGU 60, 241. Van der Voo, R., 1982. Pre-Mesozoic paleomagnetism and plate-tectonics. Annu. Rev. Earth Planet. Sci. 10, 191^220. Van der Voo, R., 1988. Paleozoic paleogeography of North America, Gondwana, and intervening displaced terranes: comparisons of paleomagnetism with paleoclimatology and biogeographical patterns. Geol. Soc. Am. Bull. 100, 311^ 324. Van der Voo, R., 1993. Paleomagnetism of the Atlantic, Tethys and Iapetus Oceans. Cambridge University Press, Cambridge, 411 pp. Van der Voo, R., Johnson, R.J., Perroud, H., 1984. A Caledonian to Hercynian history of collisions in the North American Appalachians on the basis of paleomagnetic data. In: 27th Int. Geol. Congress, Moscow, vol. 8, Geophysics, pp. 375^398. Van der Voo, R., Stamakos, J.A., Pare's, J.M., 1997. Kinematic constraints on thrust-belt curvature from syndeformational magnetizations in the Lagos del Valle Syncline in the Cantabrian Arc, Spain. J. Geophys. Res. 102, 10,105^ 10,119. Van Houten, F.B., Hargraves, R.B., 1987. Palaeozoic drift of Gondwana: palaeomagnetic and stratigraphic constraints. Geol. J. 22, 341^359. Vennin, E., Alvaro, J.J., Villas, E., 1998. High-latitude pelmatozoan-bryozoan mud-mounds from yhe late Ordovician northern Gondwana platform. Geol. J. 33, 121^140. Webby, B.D., 1984. Ordovician reefs and climate. In: Bruton, D.L. (Ed.), Aspects of the Ordovician System. Palaeontol. Contr. Univ. Oslo 295, pp. 89^100. Wendt, J., 1988. Facies pattern and paleogeography of the Middle and Late Devonian in the eastern Anti-Atlas (Morocco). In: McMillan N.J., Embry, A.F., Glass, D.J. (Eds.), Devonian of the World. Can. Soc. Petrol. Geol. Mem. 14, pp. 467^480. Whittington, H.B., Hughes, C.P., 1972. Ordovician geography and faunal provinces deduced from trilobite distribution. Philos. Trans. R. Soc. London B 263, 235^278. Williams, A., 1973. Distribution of brachiopod assemblages in relation to Ordovician palaeogeography. In: Hughes, N.F.,

PALAEO 3076 9-5-03

148

M. Robardet / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 125^148

Ed., Organisms and Continents through Time. Spec. Papers in Palaeontology 12, pp. 241^269. Witzke, B.J., 1990. Palaeoclimatic constraints for Palaeozoic Palaeolatitudes of Laurentia and Euramerica. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, pp. 57^73. Witzke, B.J., Heckel, P.H., 1988. Paleoclimatic indicators and inferred Devonian paleolatitudes of Euramerica. In: McMillan, N.J., Embry, A.F., Gass, D.J. (Eds.), Devonian of the World. Can. Soc. Petrol. Geol. Mem. 14, pp. 49^63.

Young, G.C., 1987. Devonian palaeontological data and the Armorica problem. Palaeogeogr. Palaeoclimatol. Palaeoecol. 60, 283^304. Young, G.C., 1990. Devonian vertebrate distribution patterns and cladistic analysis of palaeogeographic hypotheses. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. London Mem. 12, pp. 243^255. Ziegler, P.A., 1990. Geological Atlas of Western and Central Europe. 2nd edn. Shell Int. Petroleum Maatschappij, 238 pp.

PALAEO 3076 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172 www.elsevier.com/locate/palaeo

Ordovician organic-walled microphytoplankton (acritarch) distribution: the global scenario Thomas Servais a; , Jun Li b , Stewart Molyneux c , Elena Raevskaya d a

b

d

Laboratoire de Pale¤ontologie et Pale¤oge¤ographie du Pale¤ozo|«que (LP3), UMR 8014 du CNRS, USTL, SN5, F-59655 Villeneuve-d’Ascq Cedex, France Nanjing Institute of Geology and Palaeontology, Academia Sinica, 39 East Beijing Road, 210008 Nanjing, PR China c British Geological Survey, Keyworth, Nottingham NG12 5GG, UK Geological Faculty, St. Petersburg State University, Universitetskaya embankment 7/9, 199034 St. Petersburg, Russia Received 2 May 2002; received in revised form 27 August 2002; accepted 15 January 2003

Abstract A number of palaeobiogeographical models for Ordovician organic-walled microphytoplankton (acritarchs, prasinophytes, and related groups) have been published during the past 30 years. A modern synthesis of Ordovician acritarch palaeobiogeography, based on previously published acritarch ‘provinces’ and global distribution models, as well as new plots on recently compiled palaeogeographical maps is presented. Review of the literature and new plots indicate that a number of preliminary conclusions can be drawn. Following minor biogeographical differentiation of acritarch assemblages during the Cambrian, ‘provincialism’ started at the Cambrian^Ordovician boundary. In the late Tremadocian a warm-water assemblage, containing the genera Aryballomorpha, Athabascaella and Lua, but no diacrodians, seems to be limited to low-latitude localities such as Laurentia and North China. From the late Tremadocian and throughout most of the Arenig a peri-Gondwana acritarch assemblage with the easily recognisable taxa Arbusculidium filamentosum, Coryphidium, and Striatotheca is present on the southern margin of Gondwana, and its distribution corresponds almost exactly with that of the Calymenacean^Dalmanitacean trilobite fauna. It seems reasonable to consider the acritarchs of Baltica as belonging to a temperate-water ‘province’, which was probably not restricted to the palaeocontinent of Baltica but had a wider distribution at about the same latitude, as some of the elements recorded from Baltica also occur in South China and Argentina. The maximum separation of the continents during the Arenigian, reflected by a pronounced biogeographical differentiation of most Ordovician fossil groups, led to the development of geographically distinct acritarch assemblages. Data from the late Middle Ordovician and the Late Ordovician remain too poor to elucidate global palaeobiogeographical patterns. The biogeographical distribution of Ordovician acritarchs appears similar to that of the resting cysts of modern dinoflagellates, primarily controlled by latitude but also following the continental margins. 8 2003 Elsevier Science B.V. All rights reserved. Keywords: acritarchs; Ordovician; palaeobiogeography

1. Introduction * Corresponding author. Tel.: +33-3-20-33-72-20; Fax: +33-3-20-43-69-00. E-mail address: [email protected] (T. Servais).

Following publication of Paris and Robardet’s (1990) paper, which questioned the existence of

0031-0182 / 03 / $ ^ see front matter 8 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00306-7

PALAEO 3077 9-5-03

150

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

Tornquist’s Sea between Avalonia and Baltica, Fortey and Mellish (1992) asked whether some fossils were better than others for inferring palaeogeography. They considered this question in relation to the Early Ordovician of the North Atlantic regions, and concluded that planktonic fossils such as graptolites, chitinozoans and acritarchs were of low palaeogeographical value while benthic organisms such as trilobites, brachiopods and ostracodes were more useful for biogeographical discrimination. Fortey and Mellish’s (1992) paper has resulted in an ongoing discussion on the relative merits of palaeobiogeographical indicators in the Ordovician. At the heart of this debate is the question of whether the microphytoplankton were so widely distributed as to be virtually cosmopolitan in the Ordovician world, or whether distinct areas of the Earth were characterised by distinct micro£oras, and if so what controlled their distribution. A number of palaeobiogeographical models for Ordovician organic-walled microplankton (acritarchs, prasinophytes, and other algal groups) have been published during the past 30 years. The aim of this paper is to produce a modern synthesis of their biogeography. The synthesis is based on the Ordovician acritarch data compiled by the acritarch clade team of IGCP project 410 ‘The Great Ordovician Biodiversi¢cation Event’ (Servais and Stricanne, 2001).

2. Ordovician palaeogeography, palaeoclimatology, and oceanic currents Progress made during the last 25 years in understanding Ordovician palaeogeography has resulted from the integration of tectonic, stratigraphic, palaeoclimatological, palaeomagnetic and palaeontological evidence. The palaeogeographical maps of Scotese and McKerrow (1990) have been used to plot the distributions of many fossil groups, including acritarchs (e.g., Playford et al., 1995; Tongiorgi et al., 1995; Vecoli, 1999). On the Ordovician maps of Scotese and McKerrow (1990) and similar reconstructions, the main continents are Laurentia (North America) and Siberia, both found in equatorial positions, Baltica,

located at intermediate latitudes in the southern hemisphere, and the supercontinent Gondwana, by far the largest plate, which included South America, Africa, Antarctica, Australia, India and other marginal terranes. This large continent extended from the South Pole to the Equator but remained separated from Laurentia throughout the Ordovician. An alternative palaeogeographical model proposed by geologists working in South and Central America (e.g., Dalla-Salda et al., 1992; Dalziel et al., 1994; Dalziel, 1997) postulated a collision between Laurentia and Gondwana during the Middle Ordovician. Both models are permissible on palaeomagnetic evidence because palaeolongitudes cannot be determined by palaeomagnetic data. Palaeontological data, however, disprove the collision hypothesis. The continents remained separated throughout the Ordovician, while a smaller terrane, which included the Precordillera of Argentina, rifted o¡ from Laurentia in the Early Ordovician to drift across the Iapetus Ocean to dock with Gondwanan Argentina at some time prior to the Late Silurian (e.g., Benedetto, 1998). In the more recently published Ordovician palaeogeographical maps (e.g., Scotese et al., 2001; Cocks, 2001; Li and Powell, 2001), Gondwanan Argentina is depicted facing the western margin of Laurentia, across a western Iapetus Ocean that is much narrower than shown on earlier reconstructions. Models of Ordovician palaeoclimatology and palaeoceanography are rare and remain much more speculative than palaeogeographical reconstructions. Golonka et al. (1994) published palaeotemperature and palaeoclimate maps, constructed using a palaeoclimate-modelling programme developed by Parrish (1982). While there is broad agreement that the end-Ordovician glaciation was due to the formation of an ice cap, following the movement of Gondwana towards the South Pole, there is less agreement on models of oceanic circulation, which vary greatly from one author to another. Bergstro«m (1990, ¢g. 8), for example, presented an oceanic current circulation model for the Early Ordovician, which included a current in the southern hemisphere that originated in South America, drifted eastwards

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

along the margin of European Gondwana (passing through the Tornquist’s Sea), to reach localities such as Saudi Arabia and South China. This model was used to explain the regional distribution of some conodont taxa. At the same time, graptolite workers presented another circulation pattern model for the Arenig with a current £owing in exactly the opposite direction, originating around Arabia, drifting westwards on the northern margin of European Gondwana (passing through the Tornquist’s Sea), to reach Gondwanan Argentina (Finney and Chen, 1990). A more recent model published by Christiansen and Stouge (1999) also concerns the Arenig. These authors suggested that the temperate lowpressure zones were located at 50‡ latitude, and the subtropical high-pressure zones at 25‡ latitude. According to its authors, this model, associated with circulation of discrete water masses, explains the regional distribution, i.e., the provincialism, of graptolites, trilobites, brachiopods and conodonts.

3. Acritarch ‘provinces’ and previous plots of acritarch distribution In this chapter the previously published palaeogeographical models applied to Ordovician acritarchs are summarised. 3.1. The earliest models The ¢rst attempt to model Ordovician acritarch biogeography was that of Cramer and D|¤ez (1972, 1974, 1977). Following publication of a global distribution model of Silurian acritarchs (e.g., Cramer, 1968), in which several large-scale acritarch assemblages were interpreted as being controlled by palaeolatitude, Cramer and D|¤ez distinguished two major provinces in the Ordovician, the ‘cold African’ (named the ‘Coryphidium bohemicum province’ in Cramer and D|¤ez, 1977) and the ‘warm American Palynological Unit’, which they plotted on an available but slightly modi¢ed palaeogeographical map. Cramer and D|¤ez’s Ordovician and Silurian models are now out of date (Colbath, 1990; Tyson, 1995; Le He¤risse¤ and

151

Gourvennec, 1995; Servais and Fatka, 1997), mainly because they were based on palaeogeographical maps that have been superseded. In Cramer and D|¤ez’s models, acritarch provinces were depicted as being parallel to latitude and following climatic belts. At about the same time, Vavrdova¤ (1974) compared European Ordovician acritarch assemblages and came to a similar conclusion. She distinguished between ‘Arenig^Llanvirnian’ assemblages from Bohemia and those of the Baltic area. The Bohemian assemblages were attributed to a ‘Mediterranean province’, which included Belgium, France, Spain, North Africa, southern Germany, central Bohemia and Bulgaria. Coeval assemblages from the northern part of the former Soviet Union, Sweden, Poland, northern Germany and parts of the British Isles were attributed to the Baltic (or Boreal) province, although assemblages from the British Isles (England and Wales) and northern Germany were subsequently reassigned to the ‘Mediterranean’ province. Vavrdova¤’s (1974) di¡erentiation of the two provinces was based on general characteristics (prevalence of diacromorph acritarchs (diacrodians) in the Mediterranean province and of acanthomorph acritarchs in the Baltic province). 3.2. The ‘Mediterranean’ or ‘peri-Gondwana’ province Most papers on Ordovician acritarchs have been produced from the areas that correspond to Vavrdova¤’s Mediterranean province. Martin (1982) accepted a slightly modi¢ed version of Vavrdova¤’s (1974) model, enlarging the Mediterranean province to include the Anglo-Welsh area, eastern Newfoundland and northwestern Argentina. In doing so, she noted the ‘peri-Gondwanan’ distribution of the province. Li (1987) included his material from the Arenigian Meitan Formation of Guizhou Province, southwest China, in Vavrdova¤’s Mediterranean province, and noted that assemblages described from Hungary, Sardinia, Ireland, southern Britain, Spain, Morocco, Libya and Saudi Arabia also belonged to that province. Li (1987) thus established that Vavrdova¤’s Mediterranean province extended from east-

PALAEO 3077 9-5-03

152

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

ern Newfoundland through southern and western Europe, the Mediterranean area and the Middle East to the Upper Yangtze Region of southern China. In a subsequent paper Li (1989) rede¢ned the ‘Mediterranean’ province of Vavrdova¤ (1974), abandoning the prevalence of diacromorph acritarchs as a diagnostic feature. Instead, a⁄nity to the Mediterranean province was determined by the occurrence of the genera Arbusculidium, Coryphidium and Striatotheca. Following the suggestions of Martin (1982) and Martin and Dean (1988), Albani (1989) observed that the term ‘peri-Gondwanian palaeoprovince’ seemed to be more appropriate than the term ‘Mediterranean’ province, and added Acanthodiacrodium and Dasydiacrodium to its list of diagnostic elements. The term ‘peri-Gondwana province’ has subsequently been used widely in the literature, extending to Egypt (Gueinn and Rasul, 1986), Brazil (Padilha de Quadros, 1988), Jordan (Keegan et al., 1990), southern Turkey (Dean and Martin, 1992), Pakistan (e.g., Tongiorgi et al., 1994) and Iran (e.g., Ghavidel-Syooki, 1997), in addition to the areas mentioned above. The province is regarded as a cold-water province that extends from Argentina on the border of northern Gondwana to the Yangtze Platform (Vecoli, 1999, ¢g. 9). More recent work has proposed re¢nements to the basic concept of a peri-Gondwanan province, either by splitting the province into subdivisions or by expanding the list of its diagnostic elements. Thus Playford et al. (1995) accepted that the periGondwana province comprised a broad, latitudinally extensive, circumpolar, cold to cool-temperate, palaeogeographic belt extending along the northern margin of Gondwana from Argentina through eastern Newfoundland, North Africa, central and southern Europe, and southern Turkey to South China, but split the province into three units of ‘subprovincial’ rank. The ‘Mediterranean subprovince’ was restricted to the area originally designated by Vavrdova¤ (1974), and the ‘South America subprovince’ and ‘South China subprovince’ were added. Vavrdova¤ (1990) selected 20 species to be characteristic of the ‘highlatitude Arenigian sea’, and later Vavrdova¤ (1997) proposed the name ‘Coryphidium bohemicum acri-

tarch bioprovince’ of ‘Arenigian^Llanvirnian’ age for this assemblage. In her later paper, she provided a list of 16 taxa that she considered diagnostic of this province, and added ¢ve species that were considered characteristic for this province in the Llanvirn. Additionally, Vavrdova¤ (1997) noted that ‘a barrier divided the cool, high-latitude peri-Gondwanan region from the warmwater Baltoscandinavia and Laurentia’. Most recently, Li and Servais (in press) con¢rmed the existence of the peri-Gondwanan province, which they plotted on the palaeogeographical reconstruction of Li and Powell (2001). Li and Servais showed that the province extended around the southern part of Gondwana, from Argentina to South China, but was absent from the northern part of the supercontinent, i.e. from North China and Australia. In addition, Li and Servais noted that the province should not necessarily be considered typical of cold- or temperate-water environments, as it is present from high latitudes in European Gondwana to low latitudes such as the South China Plate. 3.3. The ‘Baltic province’ Recognition of Vavrdova¤’s (1974) Baltic province has remained problematical due to the lack of a clear de¢nition. Li (1989) noted that it was di⁄cult to recognise the Baltic province as its characteristic elements, i.e. Baltisphaeridium, Peteinosphaeridium, and Goniosphaeridium (acanthomorph acritarchs), were also found in the Mediterranean province. The Baltic province was therefore only distinguishable by the absence of typical peri-Gondwanan taxa. Brocke et al. (1995) noted that a number of easily recognisable Early to Middle Ordovician taxa, namely Arbusculidium ¢lamentosum, Arkonia, Striatotheca, Aureotesta (and Marrocanium), Coryphidium (and Vavrdovella), Dicrodiacrodium and Frankea, which were considered typical of cold-water assemblages from the peri-Gondwana province between the late Tremadocian and the early Llanvirnian, had never been recorded from Baltica. Servais and Fatka (1997) used the presence/absence of these taxa to distinguish between European micro£oras of peri-Gondwanan (including Avalonia and Ar-

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

153

morica) and Baltic a⁄nities, thereby delineating the Trans-European Suture Zone. These authors considered the Baltic province to be an area with mixed assemblages that include both cold-water (diacrodians) and warm-water forms (see below), a view also expressed by Volkova (1997). In addition, Servais and Fatka (1997) noted that the distinction between Baltica and warm-water areas remained di⁄cult because the data recovered from the latter areas were too poor. None of the publications listed above or the papers by Vavrdova¤ (1997), Raevskaya (1999) and Vecoli (1999) identi¢ed taxa that were unique to and diagnostic of the Baltic province. Playford et al. (1995), however, listed some species of Peteinosphaeridium and related genera (Cycloposphaeridium, Liliosphaeridium) as being ‘con¢ned to the Baltic province’ and others as being ‘verylikely Baltic-restricted’. The subspecies Peteinosphaeridium trifurcatum cylindroferum, for example, was considered to be ‘endemic to the Baltic province’. Nevertheless, the concept of a distinguishable Baltic province has so far been followed only by Tongiorgi and his co-workers (Playford et al., 1995; Ribecai and Tongiorgi, 1995, 1999; Tongiorgi et al., 1995, 1998; Yin et al., 1998; Tongiorgi and Di Milia, 1999).

ognised provincial categories’. So far, however, their model, distinguishing North American and Australian provinces, has not been widely adopted, and only Tongiorgi et al. (1995, 1998) have reported a similar distribution pattern. Volkova (1997) provided an alternative de¢nition of a warm-water province based on the genera Corollasphaeridium and Goniomorpha, which she considered to be indicative of the warm-water province at the Cambrian^Ordovician boundary. At a slightly higher stratigraphical level, in the late Tremadocian, she considered the presence of the genera Aryballomorpha, Athabascaella and Lua to be characteristic of the warm-water area, noting that late Tremadocian occurrences of these taxa were limited to northeast China (Martin and Yin, 1988), i.e. on the North China Plate, and to two areas on Laurentia, the ¢rst being in Alberta, Canada (Martin, 1984, 1992), and the second in Texas (Barker and Miller, 1989). In addition, Volkova (1997) noted that diacrodians were completely absent from these assemblages. Aryballomorpha and Athabascaella, however, have also been recorded from Baltica, a continent that Volkova (1997) placed in temperate-water environments, where they occur in mixed assemblages that also include cold-water taxa (diacrodians).

3.4. Warm-water province(s)

3.5. Previous plots of acritarch distribution

Not only did Playford et al. (1995) retain the Baltic province and subdivide the peri-Gondwanan province into three subprovinces, but they also established three new warm-water provinces of North America, North China, and Australia. Playford et al. (1995) identi¢ed species of Peteinosphaeridium and related genera as the distinctive elements of each palaeogeographical area. The Arenigian species Peteinosphaeridium ? furcatum, for example, was considered to be indigenous to Western Australia, while the Upper Caradocian species Peteinosphaeridium indianense and Peteinosphaeridium spiraliculum were regarded as being restricted to North America. Playford and his co-authors summarised their palaeobiogeographical ideas as follows: ‘representatives of the peteinoid genera discussed connote palaeogeographic di¡erentiation in accordance with the rec-

Cramer (1968) represented the ¢rst attempt to understand the global distribution of Silurian microphytoplankton by plotting their distribution on palaeogeographical maps, but comparable plots for the Ordovician remained rare until the 1990s, although several provincial models had been described. Over the last decade or so, there have been many attempts to depict the distribution of acritarchs on published palaeogeographical maps, to compare and contrast assemblages from di¡erent palaeogeographical areas and to test the in£uence of palaeolatitude (i.e. climatic) and other factors on acritarch distribution. Colbath (1990), for example, plotted the Ordovician genus Frankea on the Arenig^Llanvirn and Caradoc palaeogeographical maps of Scotese (1986), and concluded that its distribution could be elegantly explained if Frankea was restricted to

PALAEO 3077 9-5-03

154

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

high latitudes ( s approximately 60‡S). This model received some criticism from Servais (1993), however, who reviewed the biostratigraphical distribution of the genus and noted that its biogeographical distribution was more complicated than indicated by Colbath. Furthermore, Eiserhardt (1992) plotted Multiplicisphaeridium raspum and Palaeohystrichosphaeridium wimanii on maps derived from Scotese (1986) and Colbath (1990) for the Ordovician^Silurian boundary and the Lower^Middle Silurian, and concluded that the distribution of these taxa was not related to palaeolatitude. Most plots have been intended to depict the distribution of various provincial micro£oras. Thus, Vavrdova¤ (1990) plotted the distribution of some 20 selected taxa from the Mediterranean province on a map of unknown origin in order to compare the assemblage recovered from the Klabava Formation of Bohemia with other areas, such as the Montagne Noire (southern France), southern Britain, Thuringia, Belgium, Newfoundland, Morocco, Sardinia, Hungary and southern China. She subsequently (Vavrdova¤, 1997) indicated the distribution of her ‘Coryphidium bohemicum bioprovince’ on the Arenig palaeogeographic reconstruction of Pickering et al. (1989) and on the Late Ordovician map of Erdtmann (1991), and plotted the occurrences of Aremoricanium syringosagis, present in Gondwana, Baltica and Laurentia, on a reconstruction at the Ordovician^Silurian boundary by Pickering et al. (1988) to demonstrate breakdown in provincialism at that time. Servais et al. (1996) plotted occurrences of Dicrodiacrodium ancoriforme on a map ¢rst published by Brocke et al. (1995), demonstrating that the species is typical of the peri-Gondwanan province. The same map was used to plot additional taxa that show a similar distribution : Aureotesta^Marrocanium (Brocke et al., 1998) and Arbusculidium ¢lamentosum var. ¢lamentosum (Fatka and Brocke, 1999). Servais and Fatka (1997) subsequently published a map for the Arenig, depicting the European and North African parts of Gondwana together with Avalonia and Baltica, on which the presence/absence of Dicrodiacrodium, Frankea and A. ¢lamentosum was

used to distinguish between peri-Gondwanan and Baltican assemblages at either side of the Tornquist’s Sea. Servais and Mette (2000) and Vanguestaine and Servais (2002) plotted the occurrence of the messaoudensis^tri¢dum assemblage, considered to be typical of the late Tremadocian^early Arenig in the peri-Gondwanan area, and showed that the full assemblage only occurs in high-latitude areas, including the British Isles, Belgium, Germany, Bohemia, Spain and Turkey. Some elements, but not the whole assemblage are also found in Argentina, southern China and northwestern Russia. Most recently, Li and Servais (2002) plotted occurrences of the warm-water genera selected by Volkova (1997), namely Aryballomorpha, Athabascaella and Lua, and the three signi¢cant taxa of the peri-Gondwana province selected by Li (1989), i.e. Arbusculidium ¢lamentosum, Coryphidium and Striatotheca, on the global palaeogeographical reconstruction for the Arenig (ca. 480 Ma) by Li and Powell (2001). This plot showed a clear distinction between two distributional areas, one around the southern part of Gondwana and the second at low latitudes. Other authors have used maps by Scotese and McKerrow (1990, 1991), including Playford et al. (1995) and Tongiorgi et al. (1995), for the distribution of Lower to Middle Ordovician peteinoid taxa, and Wicander et al. (1999), who discussed in detail the palaeobiogeography of the Late Ordovician acritarchs at a global scale. The latter authors plotted all investigated localities on a palaeogeographical reconstruction based on maps of Scotese and McKerrow (1991) and Wilde (1991). Volkova (1997) plotted some of the localities bearing elements of the three provinces (periGondwana, Baltica and the warm-water province) on the palaeobiogeographical map of Erdtmann (1986), based in turn on graptolite distribution during the earliest Ordovician (early Tremadocian). 3.6. Interpretations of the impact of ocean currents There is little doubt that ocean currents played an important role in the global distribution of Ordovician acritarchs, similar to the way in which

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

modern currents in£uence the distribution of cysts of marine microphytoplankton in present-day oceans. Matthiessen (1995), for example, reported evidence that the transport of dino£agellate cysts and other organic-walled microfossils by currents modi¢es the assemblages found in Recent oceanic surface sediments from the Norwegian^Greenland Sea. Mudie and Harland (1996) noted that the in£ow of warm, saline North Atlantic Drift water in the eastern Arctic was considered responsible for a mixture of dino£agellate assemblages in the modern Arctic Ocean. Models invoking the in£uence of currents have been applied to Ordovician acritarchs to explain the distribution of selected taxa. Based on the palaeogeographical maps of Scotese and McKerrow (1990) and on the palaeoclimatological maps of Wilde (1991), several authors have proposed interpretations of the in£uence of oceanic currents on the distribution of their fossil groups. The ¢rst attempt to explain acritarch distributions in such a way was that of Li (1991) who discussed the presence of ‘cold-water’ acritarchs in lower latitudes in South China, where warm-water carbonates are indicative of a warmer environment. He noted that cold-water conodonts (An, 1987), trilobites (Zhou and Fortey, 1986) and graptolites (Berry and Wilde, 1990) were found on the Yangtze Platform during the Arenig^Llanvirnian, and proposed a model of oceanic surface currents based on the palaeogeographic map of Scotese and McKerrow (1990) to explain the presence of the ‘cold-water’ acritarchs in the low-latitude, warmer environments. Li (1991, p. 33) suggested that the taxa Striatotheca and Coryphidium, considered to be of cold-water a⁄nity by Cramer and D|¤ez (1976) and Albani (1989), may have been introduced into southern China by an ocean current circulation that brought cold water masses from low-latitude Gondwana to southwest China. Subsequently, Tongiorgi et al. (1995) and Yin (1995) published a similar model, proposing that the northward extension of the Mediterranean province to South China could be attributed to ‘a cool peri-Antarctic (peri-Gondwanan) oceanic current £owing along the subpolar margin of GondwanaT’. In a later paper, Tongiorgi and Di

155

Milia (1999) noted that ‘the present reconstructions of oceanic circuits are probably based on overly simplistic assumptions’, and presented a second possible hypothesis to explain ‘changing palaeogeographic a⁄nities’ by ‘sea-level £uctuations related to eustatic cycles’. Eiserhardt (1992) included a cautious discussion of the possible in£uence of ocean currents in his investigation of Late Ordovician assemblages. He noted that there were major problems in trying to isolate the e¡ects of ocean currents on the composition of acritarch assemblages. In the ¢rst place, he argued that the percentage of individual taxa must be known in detail in all assemblages from the di¡erent areas. Furthermore, he noted that it is almost impossible to determine whether a species produced its cysts at only one locality. Eiserhardt (1992) therefore avoided the use of a simplistic model of current circulation to explain the composition of his Late Ordovician assemblages from the southern Baltica area.

4. Palaeobiogeography or palaeoecology? The spatial distribution of acritarch taxa and morphotypes is a function of their ecological requirements as well as biogeography. Any attempt to describe biogeographical patterns of acritarch distribution must therefore take into account likely palaeoecological controls and distinguish them from palaeobiogeographical in£uences. 4.1. The biology of acritarchs Although the acritarchs were de¢ned as a utilitarian catch-all category to include organicwalled microfossils of variable and unknown biological a⁄nity (Evitt, 1963), most Lower Palaeozoic acritarchs are today considered to represent cysts of organic-walled microphytoplankton (e.g., Martin, 1993; Colbath and Grenfell, 1995; Molyneux et al., 1996; Servais et al., 1997 ; the reader is referred to these papers for reviews of the biological a⁄nities of acritarchs). Most Palaeozoic acritarchs probably constituted the organic-walled marine microphytoplankton of the oceans and may be considered as ‘pre-dino£agellates’ (Le

PALAEO 3077 9-5-03

156

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

He¤risse¤, 1989), and can be compared to various algal groups in modern oceans. Assuming that comparison with modern dino£agellates and prasinophytes is valid, Palaeozoic phytoplankton distribution was probably controlled by a series of parameters, of which the climate (temperature) signal, the coastal/oceanic (inshore^o¡shore trend) signal, the salinity signal, and the productivity (coastal upwelling zones) signal are the most important (e.g., Dale, 1996). On a global scale, modern dino£agellate species tend to occur in broad latitudinal bands, forming low-, middle- and high-latitude assemblages (Taylor, 1987). Similarly, Ordovician acritarchs should provide some information about the palaeolatitude of ancient continents, and also about the position of continental ridges. In addition, acritarchs may indicate the in£uence of longitudinal barriers to oceanic circulation. Despite the reservations of Eiserhardt (1992), ocean currents may also have exerted an in£uence on acritarch distribution. 4.2. The palaeoecology of acritarchs: the temperature signal In Mesozoic, Cenozoic and Recent dino£agellate assemblages, only a few taxa appear to be signi¢cant for biostratigraphy, (palaeo)biogeography or (palaeo)ecology. Dale (1996) proposed a model of dino£agellate cyst ecology and discussed the geological implications. Among the di¡erent environmental signals observed in recent cyst assemblages, one of the most important is surface water temperature. Dale (1996) noted that surface water temperatures in the oceans are determined by interactions of physical factors and that they have a direct impact on the constitution of dinocyst assemblages. He indicated that a temperature di¡erence of only a few degrees over a period of time might reasonably be presumed to cause biological di¡erentiation into biogeographical zones. One of the most pronounced temperature-related boundaries in the distribution of recent dino£agellate cysts has been identi¢ed as a main cooler/warmer water boundary in the North Atlantic. On the western side of the Atlantic Ocean, this boundary occurs

on the coast between Cape Cod and Nova Scotia (at about 42^43‡N) and between the English Channel and southwestern Norway on the eastern side (at about 50^65‡N), depending on which species is used (e.g., Dale, 1983; Taylor, 1987). Dale (1996, ¢g. 1) showed the distribution of selected recent and living dino£agellate cyst types compared with standard biogeographical zones (polar, subpolar, temperate, equatorial) for the Atlantic Ocean. According to this scheme, some recent taxa range from the polar region in the northern hemisphere to subpolar regions in the southern hemisphere. The living species that produce these resting cysts may thus be considered to be insensitive to climate. The distribution of other taxa, however, is limited to smaller geographical areas. While some cysts only occur in polar and subpolar areas in the northern Atlantic, others, for example, are only found in temperate to equatorial biogeographical zones. These taxa may therefore be useful for biogeographical studies. Based on this model of recent dino£agellate distribution, Li and Servais (2002) attempted to select some acritarch taxa that may be indicative of possibly temperature-related zones during the Early to Middle Ordovician. According to Li and Servais (2002), some taxa seem to be restricted by palaeolatitude, and therefore their distribution might be controlled by palaeotemperature. It appears that it is possible to distinguish cold-water (high-latitude) and warm-water (low-latitude) forms, as well as ubiquitous taxa that are found at all latitudes. Li and Servais (2002) proposed the following di¡erentiation: Cold-water forms: the Early to Middle Ordovician (Arenigian) taxa Arbusculidium ¢lamentosum, Coryphidium^Vavrdovella and Arkonia^Striatotheca appear to be most common around the South Pole up to palaeolatitudes of 60‡S, and more rarely up to 30‡S. These taxa were possibly distributed only in polar and subpolar zones, and to a lesser extent in temperate zones, assuming that the Ordovician biogeographical zonal scheme was more or less comparable to that of today (see Wilde, 1991; Christiansen and Stouge, 1999). The galeate taxa (sensu Servais and Eiserhardt, 1995) and the diacromorph acritarchs are also more abundant in higher southern hemisphere lat-

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

itudes, but may occur more frequently than the previously listed taxa at localities around 45^20‡S, i.e., these taxa are probably more likely to be recorded from the temperate zone of the southern hemisphere than the ¢rst group. Warm-water forms: the genus Lua has so far only been recorded from low latitudes, and might therefore indicate an ‘equatorial zone’ in the Ordovician. In addition, the genera Aryballomorpha and Athabascaella have been recorded from low and intermediate latitudes, but never from high latitudes. Both these genera may indicate an equatorial to temperate zone in the Ordovician. Some taxa that were ¢rst considered to be of ‘warmwater’ a⁄nity, because they were ¢rst recorded in lower latitudes, may range further to the south. Rhopaliophora was ¢rst described from Laurentia (Tappan and Loeblich, 1971) and Australia (Playford and Martin, 1984), and so its occurrence was believed to be limited to low latitudes. Subsequently it has been found in other localities from higher latitudes, for example from northern England (Cooper and Molyneux, 1990), Argentina (Rubinstein and Toro, 2001), and Iran (Ghavidel-Syooki, 2001). This taxon therefore has a wide distribution. Ubiquitous taxa: many acritarch taxa are recorded from most palaeocontinents, i.e. in most latitudes. Although some may be biostratigraphically signi¢cant or indicate special palaeoecological conditions, these taxa appear to have little application to palaeobiogeographical studies. Among these taxa, Li and Servais (2002) listed the highly variable genera Micrhystridium, Baltisphaeridium, Peteinosphaeridium and Polygonium ( = Goniosphaeridium). Li and Servais (2002) noted that their model was a ¢rst attempt at selecting climate-related morphotypes in the Early Ordovician, and that future investigations would show if some of these taxa had restricted geographic distributions and thus really occupied di¡erent climatic zones. 4.3. The palaeoecology of acritarchs: the coastal/ oceanic signal In his analysis of the ecology of recent dino£agellates, Dale (1996) noted that one of the stron-

157

gest ecological boundaries for phytoplankton in the marine realm is the limit between nutrientrich coastal/neritic waters and the relatively nutrient-poor oceanic waters. This boundary may be re£ected by an inshore^o¡shore trend in dino£agellate communities, which has been observed by many dino£agellate workers (e.g., Head and Wrenn, 1992). Mudie (1992), for example, provided evidence for inshore^o¡shore trends in transects across the temperate, the subarctic and the arctic regions of the eastern Canadian margin (respectively south of Nova Scotia, from northeastern Newfoundland to the Labrador Sea, and from the Labrador Sea to the Saglek Bank). Together with seasonal parameters, latitudinal di¡erences and surface currents, the inshore^o¡shore data set enabled the preparation of a map that depicted the modern geographical distribution of selected dino£agellate cysts in the northwestern Atlantic Ocean (Mudie, 1992, ¢g. 10A^F). Fossil communities were probably distributed in a similar way. Brinkhuis (1994), for example, provided a distribution pattern of the dino£agellate cysts and other palynomorphs, including prasinophytes, spores, pollen and foraminiferal linings, for the continental shelf-slope transect of the Upper Eocene^Lower Oligocene of the Trento Shelf, Italy. While the prasinophytes are only common in nearshore environments, the dino£agellates include taxa that are more common in the inner neritic, outer neritic or oceanic environments. Similar models for Lower Palaeozoic transects remain rare and it is still di⁄cult to attribute particular acritarch taxa to speci¢c environments. The most cited work is that of Dorning (1981), who documented acritarch distribution in the Ludlovian (Silurian) shelf sea of South Wales and the Welsh Borderland. Dorning (1981) analysed the percentages of selected acritarch taxa in his transects and indicated the relative abundance of 17 Silurian genera. He concluded that acritarchs showed a low diversity nearshore and in deeper-water environments, and a much higher diversity over much of the shelf area. Wright and Meyers (1981) indicated a similar trend in the Middle Ordovician of Oklahoma, where acritarch diversity increases from nearshore to open marine

PALAEO 3077 9-5-03

158

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

sites. Many authors came to similar conclusions and today it is widely accepted that assemblages from o¡shore shelf environments contain the most abundant and diverse acritarchs. Nearshore and deep-water acritarch assemblages are of lower diversity, dominated by sphaeromorphs with rare acanthomorphs and polygonomorphs. The inshore^o¡shore model of acritarch distribution is likely to be simplistic, in the sense that acritarch distribution along an inshore^o¡shore gradient is likely to be a¡ected by other parameters, including complex hydrodynamic factors that involve the distribution of water masses of varying physico-chemical type and sea-level £uctuations (e.g., Jacobson, 1979; Colbath, 1980). Nevertheless, Dorning’s (1981) model has been widely adopted to interpret acritarch distribution throughout the Palaeozoic. Various authors have used the inshore^o¡shore model of acritarch distribution for Ordovician acritarch assemblages to interpret conditions of sedimentation. Hill and Molyneux (1988), for example, interpreted Late Ordovician acritarch assemblages of Libya as indicating a relatively shallow, open marine shelf environment. Wicander et al. (1999) interpreted their Late Ordovician assemblage from northeastern Missouri, USA, as being deposited in a lowenergy, o¡shore, normal marine environment, consistent with the available sedimentological data and the information from other fossil groups. Vecoli (2000) also considered the palaeoenvironmental implications of acritarchs from the Cambrian^Ordovician of the northern Sahara Platform, and concluded that some stratigraphically important acritarch species, e.g., Acanthodiacrodium angustum, appeared to be faciessensitive. In some cases, it may be di⁄cult to distinguish the palaeoecological signals of Ordovician acritarchs from palaeogeographical information. It seems that some species are sensitive to speci¢c ecological conditions and to sedimentological facies, while others are sensitive to temperature, i.e. to latitudes. While the ¢rst group of species may have important palaeoecological implications, the second group may be signi¢cant for palaeobiogeography. As palaeoecological data were generally missing from the earliest acritarch investiga-

tions, many ‘palaeobiogeographical’ distributions found in the literature may thus simply re£ect di¡erent sediment facies or di¡erent palaeoecological conditions. Further study, including reinvestigation of previously published material, is required to understand fully the spatial distribution of selected morphotypes at all scales. In addition, it is important to know the total assemblage and not only a selection of taxa. 4.4. The palaeoecology of acritarchs : the salinity signal Not only does the distribution of modern cystforming dino£agellates cover a wide temperature range, from arctic to tropical waters, but it also covers the full range of present-day salinities, from freshwater to hypersaline aquatic environments. The potential for using dino£agellates as salinity indicators is poorly developed, although it appears that dino£agellates follow a similar distribution to that of recent molluscs, i.e. with three main salinity regimes (Wall et al., 1977). Recent experiments on dino£agellates in culture show that the morphology and the size of the cysts may vary considerably for a single biological species under di¡erent salinity conditions. The greatest variation a¡ects the number, distribution, length and structure of the processes. Wall and Dale (1973) interpreted the varying morphology of the resting cysts of Lingulodinium machaerophorum as being a result of low salinity in estuarine environments of the Black Sea. Turon (1984) came to a similar conclusion, correlating reduced process length with lower salinity. Ellegaard (2000) also demonstrated that a number of dino£agellate species displayed shorter processes under conditions of reduced salinity. In addition, Hallett and Lewis (2001) provided evidence of a relationship between cell biochemistry, process length and salinity. These results may have important repercussions for acritarchs. So far, the e¡ects of changing salinity on Lower Palaeozoic acritarchs have not been documented systematically, although Servais et al. (2001) suggested that an increased length of processes on specimens of the galeate acritarch plexus (sensu Servais and Eiserhardt, 1995) in

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

the Cambrian^Ordovician boundary beds of the Algerian Sahara might be related to water depth and possibly to salinity. The morphology of the peteinoid acritarchs (sensu Playford et al., 1995) is possibly also in£uenced by salinity. The classi¢cation of the peteinoids is based essentially on morphological criteria related to the processes, which for dino£agellate cysts are known to vary with environmental changes such as changes in salinity or temperature (see Kokinos and Anderson, 1995; Ellegaard, 2000; Hallett and Lewis, 2001). Bagnoli and Ribecai (2001) described a continuous morphological change in the processes of Liliosphaeridium from two nearby sections in Sweden, which may indicate that a palaeoecological parameter (such as salinity or temperature) may have changed in the sequence analysed. From this, it is evident that detailed investigations of the variability of Peteinosphaeridium and related genera are needed to separate the in£uence of palaeoecology on morphology from that of palaeobiogeography. 4.5. The palaeoecology of acritarchs: the productivity signal Several microfossil groups, including foraminifera, coccolithophorids and radiolarians, are known to indicate palaeoproductivity signals as well as ancient upwelling zones (e.g., Golonka et al., 1994). Few studies indicate so far that recent dino£agellates may also re£ect modern upwelling zones. Wall et al. (1977), for example, distinguished an upwelling signal from Peru and southwestern Africa on the basis of dino£agellate cyst distribution. The identi¢cation of ancient upwelling zones based on acritarch distribution remains extremely poor to date. The data set is too meagre and much work remains to be done before tentative correlations can be made between Ordovician acritarch assemblages and putative upwelling zones such as those indicated by Golonka et al. (1994) on their palaeogeographical and palaeoclimatic maps. So far, only Raevskaya et al. (in press) have discussed the possible in£uence of upwelling zones on the distribution of acritarch assemblages. These authors noted that the di¡er-

159

ences between the Baltic and South Chinese assemblages might be related to di¡erent physiographic situations, with di¡erent nutrient conditions depending either on £uvial or upwelling input.

5. Ordovician acritarch distribution : the global scenario Thorough compilations of Ordovician acritarch literature (Servais, 1998) and species (Servais and Stricanne, 2001) provide the basis for a reassessment of Ordovician acritarch biogeography. Some secure palaeobiogeographical interpretations are possible, while others should be considered as tentative because the available data set remains too sparse. The Ordovician acritarch data available for each palaeocontinent are summarised below. 5.1. Laurentia Laurentia is well de¢ned (e.g., by Scotese and McKerrow, 1990; Cocks, 2001) as a continental mass that was situated across the Equator throughout the Ordovician. Acritarch occurrences have been reported mainly from the Late Ordovician (Wicander in Servais et al., in press), whereas descriptions of Early Ordovician (Tremadocian and Arenig) Laurentian acritarchs are few. Martin (in Dean and Martin, 1982; Martin, 1984, 1992) described and illustrated Tremadocian and Arenig acritarch assemblages from Wilcox Pass, Alberta, Canada, and Barker and Miller (1989) mentioned similar assemblages from a borehole in the Tremadocian of Texas. Assemblages from both areas include taxa that so far have only been recorded from low latitudes, i.e., Aryballomorpha, Athabascaella and Lua. Similarly, the Middle Ordovician acritarchs of Laurentia are not known in great detail. Loeblich and Tappan (1978 and references therein) described assemblages from the Middle Ordovician (approximately Llanvirn^Llandeilo) of Oklahoma, and Martin (1983) reported coeval material from the St. Lawrence Lowland in Ottawa, Montreal, and Que¤bec City areas of Ontario and Que¤bec, Canada. Nevertheless, the data are too sparse

PALAEO 3077 9-5-03

160

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

to assess the palaeobiogeographical relationships of this material. Late Ordovician acritarchs are by far the best studied from Laurentia. Investigations include localities from Indiana, Kansas, Kentucky, Ohio, Oklahoma, Anticosti Island, the St. Lawrence Lowland and Gaspe¤. The palaeobiogeographical signi¢cance of these assemblages was discussed in some detail by Wicander et al. (1999). 5.2. Siberia and Kazakhstan Siberia was one of the major continental masses located at low latitudes during the Ordovician. It was separated from Baltica, but was possibly connected to Kazakhstan (e.g., Scotese and McKerrow, 1990; Torsvik et al., 1995). Very little has been published on Ordovician microphytoplankton from these areas (e.g., Timofeev, 1963; Sheshegova, 1971 ; Sheshegova in Moskalenko, 1984). In the latter paper, Sheshegova illustrated numerous Ordovician acritarchs from three sections in northwest Siberia, located in the Igaro^Norilsk Region and the Moierokan River basin. The oldest assemblages recorded were of latest Arenigian age, from the Kimaiskiy Substage, Kulumbe section, so there is no information on Tremadocian and Early to Middle Arenigian acritarchs to compare with data from other regions. Nevertheless, no peri-Gondwanan marker species were observed among the diverse acritarch assemblages recorded from higher parts of the Ordovician succession. Furthermore, the Upper Ordovician assemblages from the Moierokan section show a high degree of similarity to coeval assemblages from Laurentia. 5.3. Baltica The palaeogeography and palaeobiogeography of Baltica have been investigated in some detail (e.g., Cocks and Fortey, 1998), and a large number of acritarchs papers covering all series from the Lower to the Upper Ordovician have been published, covering sections in Norway, Sweden, Finland, Poland, the Baltic States and northwest Russia (Servais, 1998). Although Baltica is noted for its endemic bra-

chiopod and trilobite faunas (Cocks, 2001), the identi¢cation of endemic acritarch taxa remains problematical. Several attempts have been made to identify taxa that might be diagnostic of a Baltic province, and selected species of Peteinosphaeridium and related genera have been suggested for that role (Tongiorgi and Di Milia, 1999). However, as yet there is insu⁄cient evidence to show that the intraspeci¢c variability of the peteinoid acritarchs re£ects palaeogeographical separation rather than local environmental effects (see also Section 4). Rather than containing a unique set of taxa, Baltic assemblages on the whole seem to comprise intermediates between high- and low-latitude assemblages. So, although lowermost Ordovician (lower Tremadocian) assemblages contain predominantly cold-water taxa (diacrodians), upper Tremadocian and lowermost Arenig assemblages comprise a mixture of taxa from high- and lowlatitude realms (Volkova, 1997), while middle and upper Arenigian assemblages include a high number of widespread species that are not considered to be typical elements of any province. Widespread acanthomorph acritarchs (Baltisphaeridium, Goniosphaeridium, Peteinosphaeridium, etc.) appear to be the dominant taxa in the latter assemblages, while the less common genera Ampullula (and Stelomorpha), Pachysphaeridium, Rhopaliophora, Sacculidium (Ribecai et al., 2002) and Tongzia are also frequently recorded. However, none of these taxa is limited to Baltica, and most of these acritarchs have also been recorded from coeval sediments from South China and the Precordillera of Argentina, where they are mainly associated with distinctive peri-Gondwanan taxa. For these reasons, it seems reasonable to consider the acritarchs of Baltica as belonging to a temperate-water ‘province’, which was probably not restricted to the palaeocontinent of Baltica but had a wider distribution at about the same latitude. Perhaps because of oceanic water mass circulation, the ‘province’ might have extended to the margins of South China and Argentina, at least during the Arenig. Lack of information on Late Ordovician acritarchs from both Argentina and South China inhibits comparison with the Late Ordovician material from Baltica.

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

5.4. China: three plates Ordovician China can be divided into three major plates, the Sino-Korean (North China) Plate, the Tarim Plate, and the Yangtze Platform, with several additional minor terranes (e.g., Scotese and McKerrow, 1990; Li, 1998). During the Early Palaeozoic, these plates were separated from each other and were located in di¡erent latitudes. During most of the Ordovician, the North China Block was dominated by warm-water carbonates and was located near the Equator and/or in low latitudes in the northern hemisphere. The Yangtze Platform and Tarim Plate were located further south, with the Yangtze Platform occupying a position close to the Equator or in low latitudes in the southern hemisphere and Tarim being located further south. Palaeontological data indicate that the Yangtze Platform was probably located near Australia during most of the Palaeozoic (Scotese and McKerrow, 1990). Few papers dealing with Ordovician acritarchs from North China have been published to date (Li et al., 2001). Most papers (e.g., Martin and Yin, 1988) focus on the acritarchs from the Cambrian^Ordovician boundary in Jinlin Province, northeast China. Volkova (1997) used the occurrence of Aryballomorpha, Athabascaella and Lua to assign the Early Ordovician of the North China Plate to her ‘warm-water province’, along with Laurentia. Ordovician acritarchs from the Tarim Plate are also poorly known, and data from the Early Ordovician are lacking. Middle to Late Ordovician acritarchs have been described from the Tarim Basin of Xinjiang, North China, in a series of papers (for references see Li et al., 2002b). The assemblages show some similarities with assemblages from the type Caradoc of England and with coeval material from the mid-continent of North America (Li, 1995). However, it would be premature to draw de¢nitive palaeobiogeographical conclusions at present. Early to Middle Ordovician acritarchs from localities on the Yangtze Platform, South China, have been described in a number of papers (for references see Li et al., 2002b). Li (1987) included the Yangtze Platform in the ‘Mediterranean prov-

161

ince’ of Vavrdova¤ (1974), before proposing a water mass circulation model that brought ‘coldwater’ acritarchs into the area (Li, 1991). Li and Servais (2002) argued that southern China should be included in the ‘peri-Gondwana province’, adding that this palaeogeographical area should not be considered as being necessarily restricted to a ‘cold-water’ setting (see also below). Patterns of acritarch abundance and diversity within the Yangtze Platform probably result from environmental and facies changes, as documented in several papers. Li et al. (2002a), for example, indicated that acritarch diversity and abundance was probably related to an inshore^ o¡shore gradient that could be observed from the western (nearshore) to the eastern (open marine) part of the Platform. Brocke et al. (2000) considered that it is too early to con¢rm the model of Tongiorgi et al. (1998), who indicated that ‘palaeobiogeographical a⁄nities’ of the acritarch assemblages throughout the early to later Arenigian parts of the Dawan Formation could be attributed to a modi¢cation in the pattern of ocean currents or to sea-level £uctuations. Raevskaya et al. (in press) noted that di¡erent conditions of nutrient input might have a¡ected the composition of the South Chinese assemblages. 5.5. Australia : northern Gondwana The supercontinent of Gondwana, also referred to as Gondwanaland, occupied a major part of the southern hemisphere during the Ordovician. According to the most recent reconstructions (e.g., Cocks, 2001; Li and Powell, 2001), Australia occupied a northerly position in Gondwana, at low latitudes, very probably close to the Equator. Southeast Australia faced Antarctica, an area from which Ordovician acritarchs are so far unknown. To the south, Gondwana included the present-day Indian subcontinent, from which acritarch data also remain absent. Ordovician acritarch data from Australia are too sparse to assess their palaeobiogeographical a⁄nities, with only a few assemblages described from sections of the ‘Arenig^Llanvirn’ (Playford and Martin, 1984; Playford and Wicander, 1988). These authors did describe the total assemblages

PALAEO 3077 9-5-03

162

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

from two areas, but in contrast to areas such as the Yangtze Platform, the complete picture of Australian Ordovician acritarch assemblages, with full descriptions of assemblages from a wider range of palaeoenvironments and all stratigraphical intervals, is not yet known. Playford et al. (1995) indicated that the species Peteinosphaeridium? furcatum was only known from Western Australia. Future studies are needed to clarify if the ‘endemic’ taxa of the ‘Australia province’ (Playford et al., 1995; Tongiorgi et al., 1995) are really limited to the Australian part of the Gondwana continent. Investigations of Arenig acritarchs of Laurentia and Siberia, that remain absent so far, could indicate the presence of ‘Australian’ acritarchs in coeval sequences from other areas at low latitudes. 5.6. South America : western Gondwana Rubinstein and Toro (2001) reviewed papers published on South American acritarchs. Most concern the Early to Middle Ordovician of the Eastern Cordillera, northwestern Argentina, for which Playford et al. (1995) created the ‘South America’ subprovince. Other areas of South America have also been investigated, including Brazil (Padilha de Quadros, 1988), Colombia (The¤ry et al., 1986), and Bolivia (Gagnier et al., 1996). Following an initial investigation by Bultynck and Martin (1982) on the Ordovician acritarchs of the Eastern Cordillera, Ottone et al. (1992, 1995) con¢rmed the presence of distinctive elements of the peri-Gondwanan province. Rubinstein and Toro (2001) noted that the acritarch assemblages of the Argentinian Precordillera appear to indicate a temperate palaeolatitudinal location, near the boundary of the cold-water peri-Gondwana realm, while graptolite a⁄nities are with faunas from intermediate latitudes. To some extent, the assemblages can be compared with the southern Chinese material. They are clearly peri-Gondwanan because the taxa Arbusculidium ¢lamentosum, Coryphidium, and Striatotheca are present. However, elements from lower latitudes, indicating possibly warmer water, such as Rhopaliophora, may also be found in these areas. Future

investigations should aim to reach a better understanding between assemblages from Gondwanan Argentina and the Precordillera microterrane, which is thought to have rifted away from Laurentia in the Early Ordovician (see Benedetto, 1998). 5.7. Avalonia, ‘Armorica’, ‘Perunica’ and southern peri-Gondwana Most Ordovician acritarch studies have been based on material from western and southern Europe, North Africa and the Middle East. More than 50 papers concern the British Isles (which apart from Scotland and northern Ireland comprised part of Avalonia), more than 40 articles deal with the French Ordovician, followed by more than 30 articles on Bohemia (the microcontinent ‘Perunica’), and more than 20 articles of Belgium and Germany respectively. The investigations of French oil companies since the late 1950s in North Africa mean that this latter area is also fairly well investigated with numerous papers from Morocco, Algeria, Tunisia, and Lybia. Additional data have been published from Egypt, Jordan, Saudi Arabia, Iran and Pakistan (Servais, 1998). All these areas yield typical peri-Gondwanan assemblages throughout most of the Arenig and have been included in the ‘Mediterranean province’ since the early 1970s. Nevertheless, local differences in the composition of the assemblages are common. Such di¡erences probably relate to different local environmental conditions, and it appears di⁄cult to separate the Arenig assemblages of the di¡erent microterranes in order to distinguish between assemblages from Avalonia, ‘Perunica’ or ‘Armorica’.

6. New plots 6.1. The ‘Tremadocian^Arenig’ boundary Acritarchs from the Tremadocian^Arenig boundary of peri-Gondwana comprise the distinctive messaoudensis^tri¢dum assemblage, which has a wide distribution (Vanguestaine and Servais,

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

2002) and enables biostratigraphical correlation between localities in England, Wales, southern Ireland, Belgium, Germany, Spain, Bohemia and Turkey (Martin, 1996; Servais and Mette, 2000). Some of the elements of this distinctive assemblage are also found in southern Baltica (Raevskaya, 1999), China (Brocke, 1997) and Argentina (Rubinstein et al., 1999). Plotting the messaoudensis^tri¢dum assemblage on a recent palaeogeographical reconstruction for the late Tremadocian^early Arenig (Fig. 1) indicates that the assemblage is geographically restricted to the border of peri-Gondwana in high latitudes. Most localities bearing the assemblage are located around the South Pole ( s 60‡S), but the distribution also extends slightly northwestwards on the periGondwanan margin to latitudes between 60 and 30‡S. Coeval low-latitude assemblages contain Aryballomorpha, Athabascaella and Lua (Volkova, 1997). Li and Servais (2002) plotted the distribution of these genera on the reconstruction by Li and Powell (2001), and showed that they were

163

restricted to areas of Laurentia (Texas, USA, and Alberta, Canada), North and South China (not visible on the reconstruction of Fig. 1) and Baltica. According to the reconstruction presented in Fig. 1, all of these localities were located at low latitudes, but also extended to higher latitudes up to around 60‡S. 6.2. The ‘Arenigian’ Localities bearing the distinctive elements of the Arenig peri-Gondwanan assemblage, namely Arbusculidium ¢lamentosum, Coryphidium and Striatotheca, are plotted (Fig. 2) on a palaeogeographical reconstruction based on the recently published map of Cocks (2001). The peri-Gondwanan acritarch assemblage on this map shows almost exactly the same distribution as that of the Calymenacean^Dalmanitacean trilobite fauna of Cocks (2001). Both the trilobite fauna and the peri-Gondwanan acritarchs are distributed around the southern margin of the Gondwana continent. The geographical range of the peri-

Fig. 1. Tilted Early Ordovician (Tremadocian^Arenig) palaeogeographical reconstruction of Popov (in Bassett et al., in press) illustrating the distribution of the messaoudensis^tri¢dum acritarch assemblage (black dots). Micro£oral data available from Vanguestaine and Servais (in press). L, Laurentia; S, Siberia; B, Baltica; G, Gondwana; Av, Avalonia; A, Armorica; P, Perunica.

PALAEO 3077 9-5-03

164

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

Fig. 2. Slightly tilted Early to Middle Ordovician (Arenig) palaeogeographical reconstruction of Cocks (2001) illustrating the distribution of the peri-Gondwana acritarch assemblage de¢ned on the occurrence of the taxa Arbusculidium ¢lamentosum, Coryphidium and Striatotheca (black dots). Micro£oral data available from Servais (1997), Fatka and Brocke (1999) and Li and Servais (2002). L, Laurentia; S, Siberia; B, Baltica; G, Gondwana; Av, Avalonia.

Gondwanan acritarch assemblage starts in eastern Gondwana (Argentina, Brazil), extends to Avalonia and areas located around the South Pole, including North Africa and southern Europe (i.e.

on the border of Gondwana and the related terranes of Iberia, Armorica, Perunica), and then to Turkey, Saudi Arabia, Jordan, Iran and Pakistan. It reaches intermediate latitudes in southern Chi-

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

na. Thus, the distinctive peri-Gondwanan elements, A. ¢lamentosum, Coryphidium and Striatotheca, cross lines of latitude to range from high southern latitudes to lower latitudes. The Arenig reconstruction in Fig. 2 also shows the palaeogeographical distribution of the continents of Baltica, at intermediate latitudes, and Laurentia and Siberia continents in equatorial positions. As noted above, the distinctive elements of peri-Gondwana have never been recorded from these areas. 6.3. The ‘Arenigian^Llanvirnian’ boundary The reconstruction in Fig. 3 illustrates the position of the Ordovician palaeocontinents at the Arenig^Llanvirn boundary. The Avalonian microcontinent was located in the Tornquist’s Sea between Gondwana and Baltica, while the terranes of Armorica and Perunica were considered to be still on the periphery of Gondwana. The acritarch genus Frankea ¢rst appeared in the late Arenig and had its widest distribution during the Llanvirn (Servais, 1993). When plotted (Fig. 3), the occurrence of this genus is limited to localities

165

on the peri-Gondwanan border and to terranes that are considered to have rifted away from Gondwana, i.e., Avalonia, ‘Armorica’ and ‘Perunica’. Its distribution is restricted to localities at high and intermediate latitudes ( s 30‡S), but the genus has never been reported from Baltica, although this plate was located at similar latitudes (between 30 and 60‡S) and many sections from this continent have been investigated in great detail. So far the genus has not been recorded from southern China. The distribution of Frankea depicted here (Fig. 3) extends that shown by Colbath (1990), in which known occurrences of Frankea were restricted to areas at approximately 60‡S palaeolatitude or higher.

7. Discussion 7.1. The evolution of acritarch palaeobiogeography through the Palaeozoic Although this paper is focused on Ordovician acritarch palaeobiogeography, it is useful to compare the distributional models presented for dif-

Fig. 3. Tilted Middle Ordovician (Llanvirn) palaeogeographical reconstruction of Popov (unpublished) illustrating the distribution of the acritarch genus Frankea (black dots). Micro£oral data available from Servais (1993) and Servais (unpublished). L, Laurentia; S, Siberia; B, Baltica; G, Gondwana; Av, Avalonia; A, Armorica; P, Perunica; SC, South China.

PALAEO 3077 9-5-03

166

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

ferent times in the Ordovician with those in other Palaeozoic systems. Palaeobiogeographical models for Cambrian acritarchs are almost absent. Nevertheless, Modczydlowska (1998) indicated that taxonomically comparable acritarch assemblages of Cambrian age are found in shelf basins of the Gondwanan margin ranging from low latitudes in the northern hemisphere (Australia, South China) to southern European and North African areas, located at intermediate to high latitudes. The same assemblages were also recorded from the southern margin of Siberia and the northern margin of Baltica, facing each other, as well as around the Laurentian continental margin. Vanguestaine (1991) plotted localities from which Early and Middle Cambrian have been described on a map that was based on the reconstruction of Erdtmann (1982). Vanguestaine (1991) noted that all these localities provided assemblages that show similarities with the material he described from Belgium. Hence there is no pronounced ‘provincialism’ reported for Cambrian assemblages. Volkova (1995, 1997) considered that ‘provincialism’ only arose at the Cambrian^Ordovician boundary. As discussed here, the Arenig^Llanvirn was a time when di¡erences between acritarch assemblages from di¡erent geographical areas were most pronounced. The same distinction has also been observed for many other fossil groups. This pronounced provincialism was probably due to a maximum separation of the continents at that time, as shown in the most recent palaeogeographical reconstructions (Cocks, 2001; Li and Powell, 2001; Scotese et al., 2001). As noted above, the data set remains too poor to comment de¢nitively on palaeobiographical di¡erences in the Late Ordovician (see Wicander et al., 1999). Later in the Palaeozoic, distinct, geographically restricted assemblages have been reported from both the Silurian and the Devonian. Le He¤risse¤ and Gourvennec (1995) reviewed the data published on the Silurian, and argued that the acritarch Dactylofusa maranhensis was limited to intermediate and high latitudes on the Gondwanan border during the late Llandovery to Wenlock, whereas the Estiastra^Hoegklintia^Pulvinosphaeridium association was restricted to low latitudes,

ranging from about 30‡S to about 20‡N. Di¡erentiation of palaeobiogeographical areas in the mid-Silurian was thus similar to that in the Arenigian. On the other hand, Le He¤risse¤ and Gourvennec (1995) noted that the distribution of Deun⁄a and Domasia, both considered to be ‘pelagic’ species, was apparently controlled by parameters that were not directly related to palaeolatitude. Le He¤risse¤ et al. (1997) plotted Devonian assemblages on recent palaeogeographical reconstructions, and concluded that a geographical restriction of several acritarch genera and species could be observed for the Early Devonian. In particular, they reported pronounced di¡erences between North Gondwanan and eastern North American micro£oras. Later in the Devonian, the acritarch micro£oras became more similar, indicating that the provinciality probably decreased (Le He¤risse¤ et al., 1997, 2000). 7.2. Comparison with the Ordovician distribution of the chitinozoans Like acritarchs, the chitinozoans are considered to be a group of planktonic organic-walled microfossils. Achab (1988) summarised palaeobiogeographical information, and indicated that Ordovician chitinozoans from Que¤bec, Canada, belonged to an assemblage that was restricted to low palaeolatitudes. She also plotted the distribution of selected chitinozoan species on palaeogeographical reconstructions for the Early, Middle and Late Ordovician, and showed that some taxa were limited to a given palaeolatitude while others were more widely distributed across latitudes. In a subsequent paper, Achab (1991) noted that a biogeographical di¡erentiation could be made between high-latitude chitinozoan assemblages, present in North Africa, southwest Europe, Great Britain and Bohemia, and low-latitude assemblages recovered from eastern Canada, Australia, Spitsbergen and the United States. According to Achab (1991) the Baltic region provided faunas that appear to have occupied an intermediate position. Paris (1991) presented a similar model and also discussed the possible in£uence of water mass currents. The geographical distribution pattern of

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

Ordovician chitinozoans, also discussed by Oulebsir and Paris (1995), thus parallels the distribution of Ordovician acritarchs. Achab et al. (1992) and Paris et al. (1995) noted that a similar distribution pattern was also observed for the Silurian. While some chitinozoan taxa were cosmopolitan, and thus of limited palaeogeographical importance, other taxa were limited to the Gondwanan border, and others to the Baltica continent, on either side of the Rheic Ocean which separated these two continents.

167

North China and Baltica. The data set remains too poor to make statements bout the palaeobiogeography of the marine organic-walled microphytoplankton in the late Middle Ordovician and the Late Ordovician. The biogeographical distribution of Ordovician acritarchs appears similar to that of the resting cysts of modern dino£agellates, controlled by latitude but also following the continental margins.

Acknowledgements 8. Conclusions Investigation of Ordovician acritarch biogeography is still at an early stage, compared to study of the biogeography of other Ordovician fossil groups. In particular, the present review shows gaps in knowledge for the Ordovician of Laurentia and Australia, and indicates that there is almost no information on Ordovician acritarchs from Siberia and Kazakhstan. Study of acritarch assemblages from these areas is needed in order to document and understand the palaeobiogeographical distribution patterns of Ordovician acritarchs. Nevertheless, a number of preliminary conclusions can be drawn. There was apparently little biogeographical di¡erentiation of acritarch assemblages during the Cambrian, but the situation changed at about the Cambrian^Ordovician boundary. The maximum separation of the continents during the Arenig, re£ected by a pronounced ‘provincialism’ of most Ordovician fossil groups, was probably also responsible for the development of geographically distinct acritarch assemblages. A peri-Gondwana acritarch assemblage with the easily recognisable acritarch taxa Arbusculidium ¢lamentosum, Coryphidium and Striatotheca is present on the southern margin of Gondwana, and its distribution corresponds almost exactly with that of the Calymenacean^ Dalmanitacean trilobite fauna (Cocks, 2001). A warm-water assemblage with the acritarch genera Aryballomorpha, Athabascaella and Lua, but without diacrodians seems to be limited to localities at low to intermediate latitudes, including Laurentia,

We are grateful to a number of colleagues for valuable discussion: A. Blieck (Lille), R. Brocke (Frankfurt/Main), Chen Xu (Nanjing), O. Fatka (Prague), A. Le He¤risse¤ (Brest), F. Paris (Rennes), and Zhou Zhiyi (Nanjing). We thank L. Popov (Cardi¡) for allowing us to use his palaeogeographical reconstructions. C. Cro“nier (Lille) and Yan Kui (Nanjing) provided technical assistance for the drawing of the ¢gures. We are particularly grateful to R. Wicander (Mt. Pleasant, Michigan) and M. Vanguestaine (Lie'ge) for reviewing the manuscript and for valuable comments that improved the paper. This paper is a contribution to the French^Chinese CNRS-Academia Sinica PICS project and bene¢ted from the ¢nancial support of the National Natural Science Foundation of China (NSFC project no. 49972007), the Major State Basic Research Development of China (MSTC-G2000077700), and the Chinese Academy of Sciences (CAS-KZCX2-116 and CAS-KZCX2SW-130). J.L. acknowledges the University of Sciences and Technologies of Lille (USTL) for a position as invited professor. E.R. bene¢ted from a post-doctoral position at the USTL. S.M. publishes by permission of the Director of the BGS (NERC). This is a contribution to the IGCP project no. 410 ‘The Great Ordovician Biodiversi¢cation Event’.

References Achab, A., 1988. Mise en e¤vidence d’un provincialisme chez les chitinozoaires ordoviciens. Can. J. Earth Sci. 25, 635^ 638.

PALAEO 3077 9-5-03

168

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

Achab, A., 1991. Biogeography of Ordovician chitinozoa. In: Barnes, C.R., Williams, S.H. (Eds.), Advances in Ordovician Geology. Geol. Surv. Canada, Paper 90 (9), pp. 135^142. Achab, A., Betrand, R., Van Grootel, G., 1992. Chitinozoan contribution to the Ordovician and Lower Silurian Paleobiogeography. J. Geol. 100, 621^629. Albani, R., 1989. Ordovician (Arenigian) acritarchs from the Solanas Sandstone Formation, Central Sardinia, Italy. Boll. Soc. Paleontol. Ital. 28, 3^37. An, T.X., 1987. The early Paleozoic conodonts of South China. Beijing University Press, Beijing, 238 pp. (in Chinese). Bagnoli, G., Ribecai, C., 2001. On the biostratigraphic signi¢cance of the Ordovician acritarch genus Liliosphaeridium on º land, Sweden. Rev. Palaeobot. Palynol. 117, 195^215. O Barker, G.W., Miller, M.A., 1989. Tremadocian (Lower Ordovician) acritarchs from the subsurface of West Texas. Palynology 14, 209. Bassett, M.G., Popov, L.E., Holmer, L.E., in press. Brachiopods: Cambrian^Tremadoc precursors to Ordovician radiation events. In: Crame, J.A., Owen, A.W. (Eds.), Palaeobiogeography and Diversity Change: the Ordovician and Mesozoic-Cenozoic Radiations. Geol. Soc. London, Spec. Publ. 194, 13^23. Benedetto, J.L., 1998. Early Palaeozoic brachiopods and associated shelly faunas from western Gondwana: their bearing on the geodynamic history of the pre-Andean margins. In: Pankhurst, R.J., Rapela, C.W. (Eds.), The Proto-Andean Margins of Gondwana. Geol. Soc. London, Spec. Publ. 142, pp. 57^83. Bergstro«m, 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. Mem. Geol. Soc. London 12, pp. 105^122. Berry, W.B.N., Wilde, P., 1990. Graptolite biogeography: implications for palaeogeography and palaeoceanography. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Mem. Geol. Soc. London 12, pp. 129^138. Brinkhuis, H., 1994. Late Eocene to Early Oligocene dino£agellate cysts from the Priabonina type-area (northeast Italy): biostratigraphy and palaeoenvironmental interpretation. Palaeogeogr. Palaeoclimatol. Palaeoecol. 107, 131^163. Brocke, R., 1997. First results of Tremadoc to Lower Arenig acritarchs from the Yangtze Platform, southwest China. In: Fatka, O., Servais, T. (Eds.), Acritarcha in Praha. Acta Univ. Carolinae Sect. Geol. 40, pp. 337^355. Brocke, R., Fatka, O., Molyneux, S.G., Servais, T., 1995. First appearance of selected Early Ordovician acritarch taxa from peri-Gondwana. In: Cooper, J.D., Droser, M.L., Finney, S.C. (Eds.), Ordovician Odyssey: Short Papers for the Seventh International Symposium on the Ordovician System. The Paci¢c Section for Sedimentary Geology (SEPM) 77, Fullerton, CA, pp. 473^476. Brocke, R., Fatka, O., Servais, T., 1998. A review of the Ordovician acritarchs Aureotesta and Marrocanium. Ann. Soc. Ge¤ol. Belgique 120, 1^22.

Brocke, R., Li, J., Wang, Y., 2000. Upper Arenigian to Lower Llanvirnian acritarch assemblages from South China: a preliminary evaluation. Rev. Palaeobot. Palynol. 113, 27^40. Bultynck, P., Martin, F., 1982. Conodontes et acritarches de l’Ordovicien Infe¤rieur de la partie septentrional de la Cordille're Argentine. Bull. Inst. R. Soc. Nat. Belgique Sci. Terre 53, 1^21. Christiansen, J.L., Stouge, S., 1999. Oceanic circulation as an element in palaeogeographical reconstructions: the Arenig (early Ordovician) as an example. Terra Nova 11, 73^78. Cocks, L.R.M., 2001. Ordovician and Silurian global geography. J. Geol. Soc. London 158, 197^210. Cocks, L.R.M., Fortey, R., 1998. The Lower Palaeozoic margins of Baltica. Geol. Fo«ren. Fo«rh. (GFF) 120, 173^179. Colbath, G.K., 1980. Abundance £uctuations in Upper Ordovician organic-walled microplankton from Indiana. Micropaleontology 26, 97^102. Colbath, G.K., 1990. Palaeobiogeography of Middle Palaeozoic organic-walled phytoplankton. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Mem. Geol. Soc. London 12, pp. 207^213. Colbath, G.K., Grenfell, H.R., 1995. Review of biological af¢nities of Paleozoic acid-resistant, organic-walled eukaryotic algal microfossils (including ‘acritarchs’). Rev. Palaeobot. Palynol. 86, 287^314. Cooper, A.H., Molyneux, S.G., 1990. The age and correlation of Skiddaw Group (early Ordovician) sediments in the Cross Fell inlier (northern England). Geol. Mag. 127, 147^157. Cramer, F.H., 1968. Silurian palynologic microfossils and paleolatitudes. N. Jahrb. Geol. Pala«ont. Monogr. 10, 591^597. Cramer, F.H., D|¤ez, M., 1972. Lower Palaeozoic palynomorph provinces and paleoclimate. In: Published Abstracts of the S.E.P.M.-A.A.P.G. Meeting, Denver, CO, April 1972, p. 611. Cramer, F.H., D|¤ez, M., 1974. Early Paleozoic palynomorph provinces and paleoclimate. In: Ross, C.A. (Ed.), Paleogeographic Provinces and Provinciality. S.E.P.M. Spec. Publ. 21, pp. 177^188. Cramer, F.H., D|¤ez, M., 1976. Seven new late Arenigian species of the Acritarch Genus Coryphidium. Pala«ontol. Z. 50, 201^208. Cramer, F.H., D|¤ez, M., 1977. Lower Paleozoic phytoplancton from North Africa and adjacent regions. General survey. Ann. Mines Ge¤ol. 28, 21^34. Dale, B., 1983. Dino£agellate resting cysts: benthic plankton. In: Fryxell, G.A. (Ed.), Survival Strategies of the Algae. Cambridge University Press, Cambridge, pp. 69^136. Dale, B., 1996. Dino£agellate cyst ecology: modeling and geological applications. In: Jansonius, J., McGregor, D.C. (Eds.), Palynology: Principles and Applications. American Association of Stratigraphic Palynologists Foundation vol. 3, pp. 1249^1275. Dalla-Salda, L.H., Cingolani, C., Varela, R., 1992. Early Paleozoic orogenic belt of the Andes in southwestern South America. Result of Laurentia^Gondwana collision? Geology 20, 617^620. Dalziel, I.W.D., 1997. Overview: Neoproterozoic^Paleozoic

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172 geography and tectonics: review, hypothesis, environmental speculation. Geol. Soc. Am. Bull. 109, 16^42. Dalziel, I.W.D., Dalla-Salda, L.H., Gahagan, L.M., 1994. Paleozoic Laurentia^Gondwana interaction and the origin of the Appalachian-mountain system. Geol. Soc. Am. Bull. 106, 243^252. Dean, W.T., Martin, F., 1982. The sequence of trilobite faunas and acritarch micro£oras at the Cambrian^Ordovician boundary, Wilcox Pass, Alberta, Canada. In: Bassett, M.G., Dean, W.T. (Eds.), The Cambrian^Ordovician Boundary: Sections, Fossil Distributions, and Correlations. Geological Series 3. National Museum of Wales, Cardi¡, pp. 131^140. Dean, W.T., Martin, F., 1992. Ordovician biostratigraphic correlation in southern Turkey. In: Webby, B.D., Laurie, J.R. (Eds.), Global Perspectives on Ordovician Geology. Balkema, Rotterdam, pp. 195^203. Dorning, K.J., 1981. Silurian acritarch distribution in the Ludlowian shelf sea of South Wales and the Welsh Borderland. In: Neale, J.W., Brasier, M.D. (Eds.), Microfossils from Recent and Fossil Shelf Seas. Ellis Horwood, Chichester, pp. 31^36. Eiserhardt, K.H., 1992. Die Acritarcha des Oejlemyr£intes. Palaeontogr. Abt. B 226, 1^132. Ellegaard, M., 2000. Variations in dino£agellate cyst morphology under conditions of changing salinity during the last 2000 years in the Limfjord, Denmark. Rev. Palaeobot. Palynol. 109, 65^81. Erdtmann, B.D., 1982. Palaeobiogeography and environments of planktonic dictyonemid graptolites during the earliest Ordovician. In: Bassett, M.G., Dean, W.T. (Eds.), The Cambrian^Ordovician Boundary: Sections, Fossil Distributions, and Correlations. Geological Series 3. National Museum of Wales, Cardi¡, pp. 9^27. Erdtmann, B.D., 1986. Early Ordovician eustatic cycles and their bearing on punctuations in early nematophorid (planctic) graptolite evolution. In: Walliser, O.H. (Ed.), Global Bio-Events. Lecture Notes in Earth Sciences 8, pp. 139^152. Erdtmann, B.D., 1991. The postcadomian Early Palaeozoic tectonostratigraphy of Germany (Attempt at an analytical review). Ann. Soc. Ge¤ol. Belgique 114, 19^43. Evitt, W.R., 1963. A discussion and proposals concerning fossil dino£agellates, hystrichospheres and acritarchs. II. Proc. Natl. Acad. Sci. USA 49, 298^302. Fatka, O., Brocke, R., 1999. Morphologic variability in two populations of Arbusculidium ¢lamentosum (Vavrdova¤ 1965) Vavrdova¤ 1972. Palynology 23, 153^180. Finney, S.C., Xu, C., 1990. The relationship of Ordovician graptolite provincialism to palaeogeography. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Mem. Geol. Soc. London 12, pp. 123^128. Fortey, R.A., Mellish, C.J.T., 1992. Are some fossils better than others for inferring palaeogeography? The early Ordovician of the North Atlantic region as an example. Terra Nova 4, 210^216. Gagnier, P.Y., Blieck, A., Emig, C.C., Sempere, T., Vachard,

169

D., Vanguestaine, M., 1996. New paleontological and geological data on the Ordovician and Silurian of Bolivia. J. South Am. Earth Sci. 9, 329^347. Ghavidel-Syooki, M., 1997. Acritarch biostratigraphy of the Palaeozoic rock units in the Zagros Basin, Southern Iran. In: Fatka, O., Servais, T. (Eds.), Acritarcha in Praha. Acta Univ. Carolinae Sect. Geol. 40, pp. 385^411. Ghavidel-Syooki, M., 2001. Palynostratigraphy and paleobiogeography of the Lower Paleozoic sequence in the northeastern Alborz Range (Koret-Dagh Region) of Iran. In: Goodman, D.K., Clarke, R.T. (Eds.), Proceedings of the IX International Palynological Congress, Houston, TX, 1996. American Association of Stratigraphic Palynologists Foundation, pp. 17^35. Golonka, J., Ross, M.I., Scotese, C.R., 1994. Phanerozoic paleogeographic and paleoclimatic modeling maps. Pangea Global Environments and Resources. Can. Soc. Pet. Geol. Mem. 17, 1^47. Gueinn, K.J., Rasul, S.M., 1986. A contribution to the biostratigraphy of the Palaeozoic of the Western Desert, utilising new palynological data from the subsurface. In: Abstracts 8th E.G.P.C. Petroleum Conference, Cairo. Hallett, R., Lewis, J., 2001. Salinity, dino£agellate cyst growth and cell biochemistry. In: Abstracts 34th Annual Meeting of the American Association of Stratigraphic Palynologists. Head, M.J., Wrenn, J.H. (Eds.), 1992. Neogene and Quaternary dino£agellate cysts and acritarchs. American Association of Stratigraphic Palynologists Foundation, 438 pp. Hill, P.J., Molyneux, S.G., 1988. Biostratigraphy, palynofacies and provincialism of Late Ordovician^Early Silurian acritarchs from northeast Libya. In: El-Arnauti, A., Owens, B., Thusu, B. (Eds.), Subsurface Palynostratigraphy of Northeast Libya, pp. 27^43. Jacobson, S.R., 1979. Acritarchs as paleoenvironmental indicators in Middle and Upper Ordovican rocks from Kentucky, Ohio and New York. J. Paleontol. 53, 1197^1212. Keegan, J.B., Rasul, S.M., Shaheen, Y., 1990. Palynostratigraphy of the Lower Palaeozoic, Cambrian to Silurian, sediments of the Hashemite Kingdom of Jordan. Rev. Palaeobot. Palynol. 66, 167^180. Kokinos, J.P., Anderson, D.M., 1995. Morphological development of resting cysts in cultures of the marine dino£agellate Lingulodinium polyedrum ( = L. machaerophorum). Palynology 19, 143^166. Le He¤risse¤, A., 1989. Acritarches et kystes d’algues Prasinophyce¤es du Silurien de Gotland. Sue'de. Palaeontogr. Ital. 76, 57^302. Le He¤risse¤, A., Gourvennec, R., 1995. Biogeography of upper Llandovery and Wenlock acritarchs. Rev. Palaeobot. Palynol. 86, 111^133. Le He¤risse¤, A., Gourvennec, R., Wicander, R., 1997. Biogeography of Late Silurian and Devonian acritarchs and prasinophytes. Rev. Palaeobot. Palynol. 98, 105^124. Le He¤risse¤, A., Servais, T., Wicander, R., 2000. Devonian acritarchs and related forms. Cour. Forsch.-Inst. Senckenb. 220, 195^205. Li, J., 1987. Ordovician acritarchs from the Meitan Formation

PALAEO 3077 9-5-03

170

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

of Guizhou Province, south-west China. Palaeontology 30, 613^634. Li, J., 1989. Early Ordovician Mediterranean province acritarchs from Upper Yangtze Region, China. In: Sun, S. (Ed.), Developments in Geoscience: Contribution to the 28th Geological Congress 1989, Washington, DC. Science Press, Beijing, pp. 231^234. Li, J., 1991. The Early Ordovician acritarchs from Southwest China. Ph.D. Thesis, Nanjing Institute of Geology and Palaeontology, Academia Sinica, Nanjing (in Chinese, with English Abstract). Li, J., 1995. Ordovician (Caradoc) acritarchs from Qilang Formaton of Kalpin, Xinjiang, China (in Chinese, with English Abstract). Acta Palaeontol. Sin. 34, 454^467. Li, J., Servais, T., 2002. Ordovician acritarchs of China and their utility for global palaeobiogeography. Bull. Soc. Ge¤ol. France 173, 399^406. Li, J., Wang, Q.F., Song, Q.Y., Gao, J.Z., 2001. New data on the Ordovician acritarchs from North China. J. Stratigr. 25, 377^382. Li, J., Brocke, R., Servais, T., 2002a. The acritarchs of the South Chinese Azygograptus suecicus graptolite Biozone and their bearing on the de¢nition of the Lower^Middle Ordovician boundary. C. R. Palevol. 1, 75^81. Li, J., Servais, T., Brocke, R., 2002b. Chinese Palaeozoic acritarch research: review and perspectives. Rev. Palaeobot. Palynol. 118, 181^193. Li, Z.X., 1998. Tectonic history of the Major East Asian lithospheric blocks since the Mid-Proterozoic ^ A Synthesis. Mantle dynamics and plate interactions in East Asia. Geodyn. Ser. 27, 221^243. Li, Z.X., Powell, C.McA., 2001. An outline of the paleogeographic evolution of the Australasian region since the beginning of the Neoproterozoic. Earth Sci. Rev. 53, 237^277. Loeblich, A.R., Tappan, H., 1978. Some middle and late Ordovician microphytoplankton from Central North America. J. Paleontol. 52, 1233^1278. Martin, F., 1982. Some aspects of late Cambrian and early Ordovician acritarchs. In: Bassett, M.G., Dean, W.T. (Eds.), The Cambrian^Ordovician Boundary: Sections, Fossil Distributions, and Correlations. Geological Series 3. National Museum of Wales, Cardi¡, pp. 29^40. Martin, F., 1983. Chitinozoaires et acritarches ordoviciens de la plate-forme du Saint-Laurent (Que¤bec et Sud-Est de l’Ontario). Geol. Surv. Canada Bull. 310, 1^59. Martin, F., 1984. New Ordovician (Tremadoc) acritarch taxa from the middle member of the Survey Peak Formation at Wilcox Pass, Southern Canadian Rocky Mountains, Alberta. Current Research, Part A, Geol. Surv. Canada, Paper 841A, pp. 441^448. Martin, F., 1992. Uppermost Cambrian and Lower Ordovician acritarchs and Lower Ordovician chitinozoans from Wilcock Pass, Alberta. Geol. Surv. Canada Bull. 420, 1^57. Martin, F., 1993. Acritarchs: A review. Biol. Rev. 68, 475^538. Martin, F., 1996. Recognition of the acritarch-based ‘tri¢dum £ora’ (Ordovician) in the absence of the eponymous species. Bull. Inst. R. Sci. Nat. Belgique Sci. Terre 66, 5^13.

Martin, F., Dean, W.T., 1988. Middle and Upper Cambrian acritarch and trilobite zonation at Manuels River and Random Island, eastern Newfoundland. Geol. Surv. Canada Bull. 381, 1^99. Martin, F., Yin, L.M., 1988. Early Ordovician acritarchs from Southern Jilin Province, North-East China. Palaeontology 31, 109^127. Matthiessen, J., 1995. Distribution patterns of dino£agellate cysts and other organic-walled microfossils in recent Norwegian^Greenland Sea sediments. Mar. Micropaleontol. 24, 307^334. Modczydlowska, M., 1998. Cambrian acritarchs from Upper Silesia, Poland ^ biochronology and tectonic implications. Fossils Strata 46, 1^121. Molyneux, S.G., Le He¤risse¤, A., Wicander, R., 1996. Paleozoic phytoplankton. In: Jansonius, J., McGregor, D.C. (Eds.), Palynology: Principles and Applications. American Association of Stratigraphic Palynologists Foundation vol. 2, pp. 493^530. Moskalenko, T.A., 1984. Ordovik Sibirskai platformy. Paleontologicheskii Atlas. Akademiya Nauk SSSR. Trudy IGIG, Vyp. 590. Nauka, Novosibirsk, 240 pp. (in Russian). Mudie, P.J., 1992. Circum-arctic Quaternary and Neogene marine palyno£oras: paleoecology and statistical analyses. In: Head, M.J., Wrenn, J.H. (Eds.), Neogene and Quaternary Dino£agellate Cysts and Acritarchs. American Association of Stratigraphic Palynologists Foundation, pp. 347^390. Mudie, P.J., Harland, R., 1996. Aquatic Quaternary. In: Jansonius, J., McGregor, D.C. (Eds.), Palynology: Principles and Applications. American Association of Stratigraphic Palynologists Foundation vol. 2, pp. 843^877. Ottone, E.G., Toro, B.A., Waisfeld, B.G., 1992. Lower Ordovician palynomorphs from the Acoite Formation, northwestern Argentina. Palynology 16, 93^116. Ottone, E.G., Waisfeld, B.G., Astini, R.A., 1995. Acritarcos del Ordovicico temprano de la quebrada de Chalala, noroeste de Argentina. Ameghiniana 32, 237^242. Oulebsir, L., Paris, F., 1995. Chitinozoaires ordoviciens du Sahara alge¤rien: biostratigraphie et a⁄nite¤s pale¤oge¤ographiques. Rev. Palaeobot. Palynol. 86, 49^68. Padilha de Quadros, L., 1988. Zoneamento bioestratigra¢co do Paleozoico inferior e medio (Secao Marinha) da Bacia do Solimoes. Bol. Geoci. Petrobras 2, 95^109. Paris, F., 1991. Conceptions actuelles sur la pale¤oge¤ographie ordovicienne. Ge¤ochronique 40, 20. Paris, F., Robardet, M., 1990. Early Palaeozoic palaeogeography of the Variscan regions. Tectonophysics 177, 193^213. Paris, F., Verniers, J., Al-Hajri, S., Al-Tayyar, H., 1995. Biostratigraphy and palaeogeographic a⁄nities of Early Silurian chitinozoans from central Saudi Arabia. Rev. Palaeobot. Palynol. 89, 75^90. Parrish, J.T., 1982. Upwelling and petroleum source beds, with reference to Paleozoic. The A.A.P.G. Bull. 66, 750^774. Pickering, K.T., Bassett, M.G., Siveter, D.J., 1988. Late Ordovician^early Silurian destruction of the Iapetus Ocean: Newfoundland, British Isles and Scandinavia ^ a discussion. Trans. R. Soc. Edinburgh Earth Sci. 79, 361^382.

PALAEO 3077 9-5-03

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172 Pickering, K.T., Bassett, M.G., Siveter, D.J., 1989. Erratum: Late Ordivician^early Silurian destruction of the Iapetus Ocean: Newfoundland, British Isles and Scandinavia ^ a discussion. Trans. R. Soc. Edinburgh Earth Sci. 80, 69. Playford, G., Martin, F., 1984. Ordovician acritarchs from the Canning Basin, Western Australia. Alcheringa 8, 187^223. Playford, G., Wicander, R., 1988. Acritarch palyno£ora of the Coolibah Formation (Lower Ordovician), Georgina Basin, Queensland. Mem. Assoc. Austral. Palaeontol. 5, 5^40. Playford, G., Ribecai, C., Tongiorgi, M., 1995. Ordovician acritarch genera Peteinosphaeridium, Liliosphaeridium, and Cycloposphaeridium: morphology, taxonomy, biostratigraphy, and palaeogeographic signi¢cance. Boll. Soc. Paleontol. Ital. 34, 3^54. Raevskaya, E.G., 1999. Early Arenig acritarchs from the Leetse Formation (St. Petersburg region, northwest Russia) and their palaeogeographic signi¢cance. Boll. Soc. Paleontol. Ital. 38, 247^256. Raevskaya, E.G., Vecoli, M., Bednardczyk, W., Tongiorgi, M., in press. Early Arenig acritarchs from the East-European Platform (St. Petersburg Region, northwestern Russia, and Leba area, northwestern Poland) and their palaeobiogeographic signi¢cance. Lethaia. Ribecai, C., Tongiorgi, M., 1995. Arenigian acritarchs from º land, Sweden): a preliminary report. Rev. Horns Udde (O Palaeobot. Palynol. 86, 1^11. Ribecai, C., Tongiorgi, M., 1999. The Ordovician acritarch genus Pachysphaeridium Burmann 1970: new, revised, and reassigned species. Palaeontogr. Ital. 86, 113^153. Ribecai, C., Raevskaya, E.G., Tongiorgi, M., 2002. Sacculidium: a new Ordovician genus with a distinctive pseudo-pylome. Rev. Palaeobot. Palynol. 121, 163^203. Rubinstein, C.V., Toro, B.A., 2001. Review of acritarch biostratigraphy in the Arenig of the Eastern Cordillera, Northwestern Argentina: new data and calibration with the graptolite zonation. In: Weiss, R.H. (Ed.), Contributions to Geology and Palaeontology of Gondwana in Honour of Helmut Wopfner. Geological Institute of the University, Cologne, pp. 421^439. Rubinstein, C.V., Toro, B.A., Waisfeld, B., 1999. Acritarch biostratigraphy of the upper Tremadoc-Arenig of the Eastern Cordillera, northwestern Argentina: relationships with graptolite and trilobite faunas. Boll. Soc. Paleontol. Ital. 38, 267^286. Scotese, C.R., 1986. Phanerozoic reconstructions: a new look at the assembly of Asia. Univ. Texas, Inst. Geophys., Tech. Rept. 66, 54 pp. Scotese, C.R., McKerrow, W.S., 1990. Revised world maps and introduction. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Mem. Geol. Soc. London 12, pp. 1^21. Scotese, C.R., McKerrow, W.S., 1991. Ordovician plate tectonic reconstructions In: Barnes, C.R., Williams, S.H. (Eds.), Advances in Ordovician Geology. Geol. Surv. Canada, Paper 90 (9), pp. 271^282. Scotese, C.R., Boucot, A.J., Chen, X., Ruan, Y.P., Peng, S.C., 2001. Reconstructions of the Cambrian to Permian geogra-

171

phy and climatic belts. In: Chen, X., Ruan, Y.P., Boucot, A.J. (Eds.), Paleozoic Climatic Evolution of China. Science Press, Beijing, pp. 127^143 (in Chinese). Servais, T., 1993. The Ordovician acritarch Frankea. Spec. Pap. Palaeontol. 48, 79^95. Servais, T., 1997. The Ordovician Arkonia^Striatotheca acritarch plexus. Rev. Palaeobot. Palynol. 98, 47^79. Servais, T., 1998. An annotated bibliographical review of Ordovician acritarchs. Ann. Soc. Ge¤ol. Belgique 120, 23^72. Servais, T., Eiserhardt, K.H., 1995. A discussion and proposals concerning the Lower Paleozoic ‘galeate’ acritarch plexus. Palynology 19, 191^210. Servais, T., Fatka, O., 1997. Recognition of the Trans-European Suture Zone (TESZ) by the palaeobiogeographical distribution pattern of early to middle Ordovician acritarchs. Geol. Mag. 134, 617^625. Servais, T., Mette, W., 2000. The messaoudensis^tri¢dum acritarch assemblage (Ordovician: late Tremadoc^early Arenig) of the Barriga Shale Formation, Sierra Morena (SW-Spain). Rev. Palaeobot. Palynol. 113, 145^163. Servais, T., Stricanne, L., 2001. Ordovician phytoplankton (acritarch) diversity. In: IGCP 410 Abstracts, Paleobios, 21 (2) 2nd Supplement, p. 11. Servais, T., Brocke, R., Fatka, O., 1996. Variability in the Ordovician acritarch Dicrodiacrodium. Palaeontology 39, 389^405. Servais, T., Brocke, R., Fatka, O., Le He¤risse¤, A., Molyneux, S.G., 1997. Value and meaning of the term acritarch. In: Fatka, O., Servais, T. (Eds.), Acritarcha in Praha. Acta Univ. Carolinae Sect. Geol. 40, pp. 631^643. Servais, T., Montenari, M., Stricanne, L., 2001. Acritarchs at the Cambro^Ordovician boundary: Biostratigraphy or ecophenotypism? In: Harper, D.A.T. (Ed.), Abstracts 45th Annual Meeting of the Palaeontological Association, p. 25. Servais, T., Li, J., Stricanne, L., Vecoli, M., Wicander, R., in press. Acritarchs. In: Webby, B.D., Droser, M.L., Paris, F., Percival, I. (Eds.), The Great Ordovician Biodiversi¢cation Event. Columbia University Press, New York. Sheshegova, L.I., 1971. Akritarkhi paleozoya. In: Vozzhennikova, T.F. (Ed.), Vodozosli Paleozoya i Mezozoya Sibiri. K III Mezhdunarodnai Palinologicheskai Konferentsii Novosibirsk. Izdatelstvo Nauka, Moskva, pp. 9^35 (in Russian). Tappan, H., Loeblich, A.R., 1971. Surface sculpture of the wall in Lower Paleozoic acritarchs. Micropaleontology 17, 385^410. Taylor, F.J.R., 1987. Ecology of dino£agellates. In: Taylor, F.J.R. (Ed.), The Biology of Dino£agellates. Botanical Monographs 21, pp. 399^501. Timofeev, B.V., 1963. Ordovician and Silurian phytoplankton of the Siberian Platform. Akad. Nauk SSSR (Dokl. Earth Sci. Sect.) 149, 399^402. Published English translation dated 1965 in Dokl. Earth Sci. Sect. Am. Geol. Inst. 149, 45^48 (in Russian). The¤ry, J.M., Pe¤niguel, G., Haye, G., 1986. Descubrimiento de Acritarcos del Arenigiano cerca a Araracuara (Caqueta¤, Colombia). Ensayo de reinterpretatio¤n de esta regio¤n de la Saliente del Vaupe¤s. Geol. Norandina 9, 3^17.

PALAEO 3077 9-5-03

172

T. Servais et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 149^172

Tongiorgi, M., Di Milia, A., 1999. Di¡erentiation and spread of the Baltic Acritarch Province (Arenig-Llanvirn). Boll. Soc. Paleontol. Ital. 38, 297^312. Tongiorgi, M., Di Milia, A., Lefort, P., Gaetani, M., 1994. Palynological dating (Arenig) of the sedimentary sequence overlying the Ishkarwaz Granite (upper Yarkhun valley, Chitral, Pakistan). Terra Nova 6, 595^607. Tongiorgi, M., Yin, L.M., Di Milia, A., 1995. Arenigian acritarchs from the Daping section (Yangtze Gorges area, Hubei Province, Southern China) and their palaeogeographic signi¢cance. Rev. Palaeobot. Palynol. 86, 13^48. Tongiorgi, M., Yin, L.M., Di Milia, A., Ribecai, C., 1998. Changing paleogeographical a⁄nities of the acritarch assemblages throughout the Dawan Formation (Arenig, Yichang Area, South China). Palynology 22, 181^196. Torsvik, T.H., Tait, J., Moralev, V.M., McKerrow, W.S., Sturt, B.A., Roberts, D., 1995. Ordovician palaeogeography of Siberia and adjacent continents. J. Geol. Soc. London 152, 279^287. Turon, J.L., 1984. Le palynoplankton dans l’environnement actuel de l’Atlantic nord-oriental. Evolution climatique et hydrologique depuis le dernier maximum glaciaire. Me¤m. Inst. Ge¤ol. Bassin Aquitaine 17, 1^313. Tyson, R.V., 1995. Sedimentary Organic Matter. Organic Facies and Palynofacies. Chapman and Hall, London. Vanguestaine, M., 1991. Datation par acritarches des couches cambro-tre¤madociennes les plus profondes du sondage de Lessines (bord me¤ridional du Massif du Brabant, Belgique). Ann. Soc. Ge¤ol. Belgique 114, 213^231. Vanguestaine, M., Servais, T., 2002. Early Ordovician acritarchs of the Lierneux Member (Stavelot Inlier, Belgium): stratigraphy and palaeobiogeography. Bull. Soc. Ge¤ol. France 173, 561^568. Vavrdova¤, M., 1974. Geographical di¡erentiation of Ordovician acritarch assemblages in Europe. Rev. Palaeobot. Palynol. 18, 171^175. Vavrdova¤, M., 1990. Early Ordovician acritarchs from the locality Myto near Rokycany (late Arenig, Czechoslovakia). Cas. Pro Mineral. Geol. 35, 239^250. Vavrdova¤, M., 1997. Early Ordovician provincialism in acritarch distribution. Rev. Palaeobot. Palynol. 98, 33^40. Vecoli, M., 1999. Cambro-Ordovician palynostratigraphy (ac-

ritarchs and prasinophytes) of the Hassi-R’Mel area and northern Rhadames Basin, North Africa. Palaeontogr. Ital. 86, 1^112. Vecoli, M., 2000. Palaeoenvironmental interpretation of microphytoplankton diversity trends in the Cambrian^Ordovician of the northern Sahara Platform. Palaeogeogr. Palaeoclimatol. Palaeoecol. 160, 329^346. Volkova, N.A., 1995. Acritarchs of the Cambrian^Ordovician boundary deposits of the Baltic Phytoplankton Province. Stratigr. Geol. Correl. 3, 31^43. Volkova, N.A., 1997. Paleogeography of phytoplankton at the Cambrian^Ordovician Boundary. Paleontol. J. 31, 135^140. Wall, D., Dale, B., 1973. Paleosalinity relationships of dino£agellates in the Late Quaternary of the Black Sea ^ a summary. Geosci. Man 7, 95^102. Wall, D., Dale, B., Lohmann, G.P., Smith, W.K., 1977. The environmental and climatic distribution of dino£agellate cysts in modern marine sediments from regions in the North and South Atlantic Oceans and adjacent seas. Mar. Micropaleontol. 2, 121^200. Wicander, R., Playford, G., Robertson, E.B., 1999. Stratigraphic and paleogeographic signi¢cance of an upper Ordovician acritarch £ora from the Maquoketa Shale, northeastern Missouri, USA. Paleontol. Soc. Mem. 51 (J. Paleontol. 73 (Suppl. 6)), 1^38. Wilde, P., 1991. Oceanography in the Ordovician. In: Barnes, C.R., Williams, S.H. (Eds.), Advances in Ordovician Geology. Geol. Surv. Canada, Paper 90 (9), pp. 283^298. Wright, R.P., Meyers, W.C., 1981. Organic walled microplankton in the subsurface Ordovician of northeastern Kansas. Kansas Geol. Surv. Subsurf. Geol. Ser. 4, 1^53. Yin, L.M., 1995. Early Ordovician acritarchs from Hunjiang region, Jilin, and Yichang region, Hubei, China (in Chinese, with English Abstract). Palaeontol. Sin. New Ser. A 185, 1^ 175. Yin, L.M., Di Milia, A., Tongiorgi, M., 1998. New and emended acritarch taxa from the lower Dawan Formation (lower Arenig, Huanghuachang Section, South China). Rev. Palaeobot. Palynol. 102, 223^248. Zhou, Z.Y., Fortey, R.A., 1986. Ordovician trilobites from North and Northeast China. Palaeontogr. Abt. A 192, 157^210.

PALAEO 3077 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191 www.elsevier.com/locate/palaeo

Cosmopolitan arthropod zooplankton in the Ordovician seas Jean Vannier a; , Patrick R. Racheboeuf a , Edsel D. Brussa b , Mark Williams c , Adrian W.A. Rushton d , Thomas Servais e , David J. Siveter f a

Universite¤ Claude Bernard Lyon I, UFR Sciences de la Terre, UMR 5125 ‘Pale¤oenvironnements et Pale¤obiosphe're’ du CNRS, 43, Boulevard du 11 novembre 1918, 69622 Villeurbanne cedex, France b CONICET, Universidad Nacional de La Pampa, Ca¤tedra de Paleontolog|¤a I, Facultad de Ciencias Exactas y Naturales, Uruguay 151, 6300 Santa Rosa, Argentina c British Geological Survey, Keyworth, Nottingham NG12 5GG, UK d Department of Palaeontology, The Natural History Museum, London SW7 5BD, UK e Universite¤ des Sciences et Techniques de Lille, Sciences de la Terre, UMR 8014 du CNRS, Cite¤ Scienti¢que SN5, 59655 Villeneuve d’Ascq cedex, France f University of Leicester, Department of Geology, University Road, Leicester LE1 7RH, UK Received 13 May 2002; received in revised form 2 September 2002; accepted 15 January 2003

Abstract Evidence is presented here for a zooplanktonic component in Ordovician marine ecosystems, namely the caryocaridid arthropods, that add to other well-documented midwater organisms such as graptolites, cyclopygid and telephinid trilobites, orthoconic cephalopods and the microphytoplankton (e.g. acritarchs). Although the soft anatomy of caryocaridids is largely hypothetical, their carapace design and ultrastructure, and their phyllocarid-like abdominal morphology (flattened furcal rami, telescopic segments) indicate a swimming lifestyle in midwater niches. Both functional and ecological interpretations are supported by their palaeogeographical and facies distributions and by analogies with modern pelagic ostracods. Caryocaridids occur at numerous localities on the palaeo-plates of Laurentia, Baltica, Avalonia, Perunica, Gondwana and South China and are recurrent faunal components of graptolitic black shales (mainly Tremadoc to Llanvirn). Typical faunal associates are the didymograptid and isograptid graptolites, pelagic cyclopygid and deep-sea benthic atheloptic trilobites. Their depositional environments suggest that the caryocaridids and their pelagic associates (graptolites) most probably thrived in waters above the distal shelf margins, where upwelling-controlled primary productivity possibly reached its maximum. Their exact bathymetrical range within the water column cannot be inferred from fossil evidence. However, their feeding strategies may have led them to exploit food resources across the mesopelagic^epipelagic boundaries as do numerous midwater crustaceans in present-day ecosystems. Caryocaridids represent a significant step in the post-Cambrian colonisation of midwater niches by arthropods and in the construction of complex modern foodwebs. 6 2003 Elsevier Science B.V. All rights reserved. Keywords: zooplankton; arthropods; crustaceans; Caryocarididae; Ordovician; palaeobiogeography

* Corresponding author. Tel.: +33-04-7244-8144.

E-mail address: [email protected] (J. Vannier).

0031-0182 / 03 / $ ^ see front matter 6 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0031-0182(03)00307-9

PALAEO 3078 9-5-03

174

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

1.2. Midwater marine arthropods in the Early Palaeozoic: previous work

1. Introduction Zooplanktonic organisms play a crucial role in modern marine food webs both in terms of biomass and energy £uxes. By exploiting and recycling microscopic phytoplankton, they produce massive quantities of nutrient-rich particles that constitute a permanent and exploitable resource for benthic communities (Butter¢eld, 2002a,b). Furthermore, as prey for larger midwater secondary consumers, many of them also contribute directly to the food chain. Thus, the colonisation of midwater niches by zooplanktonic organisms was a very important step in the construction of modern oceanic foodwebs. Pioneer colonisations are likely to have begun in the Late Proterozoic or Early Cambrian (Williams et al., 1996; Butter¢eld, 1997; Vannier and Chen, 2000) and increased throughout the early Palaeozoic by way of assumed pulses of invasion (Vannier et al., 1998). We are testing this model in the light of fossil evidence from caryocaridid arthropods, which form a recurrent component of graptolitic shales during the Tremadoc to early Caradoc. 1.1. Midwater arthropods of present-day ecosystems Present-day midwater niches are occupied by an extraordinary diversity of organisms ranging from picoplankton (below 2 Wm) to large ¢shes (e.g. sharks) and mammals (e.g. whales). These are either passive drifters, buoyancy regulators or swimmers, some of which engage in daily or seasonal vertical migration. Midwater organisms also comprise the larval stages of numerous benthic animals. Of the arthropods, crustaceans have colonised almost all levels of the water column (Fig. 1). Some are con¢ned to the close vicinity of the sea bottom or thrive in surface waters (epipelagic zone), others have bathymetrical preferences in relation to food supplies and perform vertical migrations through the water column. The most proli¢c midwater crustaceans in modern marine ecosystems are the copepods, the halocyprid ostracods and the euphausiaceans (shrimp-like krill).

The animal invasion of midwater niches started early in the Phanerozoic history of numerous phyla, including arthropods (Rigby and Milsom, 1996; Rigby, 1997). Arthropods are likely to have inhabited part of the water column ecospace already by the Early Cambrian. Whereas the majority of early Cambrian arthropods were probably epibenthic or nektobenthic animals (e.g. phyllocarid-like bivalved forms such as waptiids; Chen and Zhou, 1997; Taylor, 2002) living on the substratum, others such as Isoxys (Williams et al., 1996 : Vannier and Chen, 2000) and possibly other smaller bivalved arthropods (e.g. some bradoriids; Siveter and Williams, 1997) were exploiting true midwater niches. Other fossil evidence for mesozooplanktonic arthropods in the early Cambrian comes from excellently preserved ¢lter-feeding apparatuses (Mount Cap; Butter¢eld, 1994, 1997) that closely resemble those of free-swimming branchiopod crustaceans. Planktonic larval stages, such as nauplius larvae almost identical to Recent forms are present in the Upper Cambrian ‘Orsten’ fauna from Sweden (e.g. Walossek and Mu«ller, 1989) and may have been present earlier. In addition, Fortey (1985) and McCormick and Fortey (1998) have given convincing evidence on the basis of functional morphology, eye design and palaeolatitudinal patterns that some Early Palaeozoic trilobites were midwater dwellers. These are typically the epipelagic Carolinites and the mesopelagic Pricyclopyge, both of which have hypertrophied eyes. These trilobites and their relatives (Cyclopygidae, Telephinidae) are con¢ned to Ordovician strata. Fossil evidence for the earlier occurrence of pelagic trilobites is hypothetical (e.g. Irvingella and Centropleura from the Upper and Middle Cambrian respectively ; Fortey, 1985). The colonisation of midwater niches by ostracods is supposed to have taken place much later, by mid-Silurian times (Siveter et al., 1991; Vannier and Abe, 1995), and is supported by comparing the morphology of fossil and Recent myodocopes, by their facies and palaeogeographical distribution and by the nature of the associated faunas. Caryocaris has long been interpreted as a plank-

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

175

Fig. 1. Diversity of midwater crustaceans in Recent marine environments (typical representatives given in brackets). Larval forms not considered. (A^H) Adaptations to midwater habitats exempli¢ed by free-swimming copepod (A), parasite copepod (B), daphnid (C), anostracan (D), euphausiacean (krill; E) hyperiid amphipod (F), halocyprid ostracod (G), and phyllocarid (H). Legend: 1, anostracan branchiopods (Branchinecta; free-swimmer, ephemeral ponds); 2, gelyelloida copepods (Gelyella; free-swimmer, underground freshwater); 3, haplopod branchiopods (Leptodora; free-swimmer, lakes); 4, anomopod branchiopods (Daphnia, free-swimmer, continental freshwater); 5, cyclopoid copepods (Acanthocyclops; free-swimmer, lakes); 6, remipeds (Speleonectes; free-swimmer in marine caves connected to the sea); 7, mictacean peracarids (Mictocaris; free-swimmer, marine caves); 8, portunid decapods (Portunus (swimming crab), swimming abilities, shallow marine and estuarine habitats); 9, phyllocarids (Nebalia; nektobenthic, coastal environments); 10, cypridinid ostracods (Vargula; nektobenthic active swimmer); 11, harpacticoid copepods (Euterpina; pelagic, very shallow waters); 12, branchiurans (Argulus; free-swimmer and ectoparasite of ¢sh); 13, siphonostomatoid copepods (Trebius; ectoparasite of horn sharks (passive transportation)); 14, calanoid copepods (Calanus; epipelagic, vertical migrations); 15, sergestid decapods (Sergestes; pelagic, vertical migrations); 16, hyperid amphipods hosting on zooplankters (salps, medusae, jelly¢sh, siphonophores), extensive diurnal vertical migrations (over 1000 m vertically in 24 h); 17, mormonilloid copepods (Mormonilla; mesopelagic); 18, halocypridid ostracods (Conchoecia; epipelagic to deep mesopelagic, vertical migrations); 19, euphausiaceans (Euphausia; pelagic gregarious free-swimmers, particularly common in the open ocean (krill) to a depth of 5000 m, major source of food for larger nektonic animals (baleen whales, squid, ¢sh), extensive daily vertical migrations of up to 400 m each way); 20, penaeid decapod (Bentheogennema; bathypelagic); 21, mysid peracarids; 22, amphionidacean eucarid (Amphionides; pelagic to a depth of 1700 m); 23, cypridinid ostracods (Macrocypridina; deep mesopelagic (adults) to about 100 m (juveniles)); 24, phyllocarids (Nebaliopsis typica; bathypelagic); 25, cypridinid ostracods (Gigantocypris muelleri; deep mesopelagic/bathypelagic, vertical migrations); 26, lophogastrid peracarids (Gnathophausia ingens; gigantic (up to 35 cm), bathypelagic oceanic predator on zooplankton); 27, phyllocarids (Dahlella; nektobenthic, deep-sea hydrothermal vents); 28, cypridinid ostracods (Gigantocypris dracontovalis; abyssopelagic and benthopelagic, vertical migrations).

tonic animal on the basis of its unusually wide distribution and its frequent occurrence in black graptolitic shale facies (e.g. Ruedemann, 1934; StMrmer, 1937; Chlupa¤c, 1970; Vannier et al., 1998 and complete references in Racheboeuf et al., 2000) but detailed studies to assess its impor-

tance in Ordovician ecosystems are still lacking. From a palaeoecological point of view, LeGrand and Hannibal (2000) suggested that the occurrences of Caryocaris in the Ordovician of South America may be analogous to mass occurrences of Nebalia in the present-day submarine canyons

PALAEO 3078 9-5-03

176

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

of California (see also Vetter, 1994, 1995, 1996 for studies on productivity)

all characterised by a small, bivalved, elongate carapace that, by contrast with other extinct or extant representatives of Phyllocarida such as Nebalia (Vannier et al., 1997) lacks rostral and dorsal plates. Many caryocaridid species have dorsal, posterior and ventromarginal spines. The £imsy exoskeleton (possibly a head-shield) is in many cases strengthened along the margin by a ridge, and a doublure underlines the ventral margin of the lateral folds (‘valves’ of authors). The general absence of dorsal disruption suggests that the carapace was not hinged as it is in, for example, calci¢ed ostracods but was merely folded dorsally. No adductor muscle scars are known. The ana-

2. Caryocaridid morphology and inferred lifestyles 2.1. General morphology The caryocaridids have long been placed within the Phyllocarida (see Racheboeuf et al., 2000). The family consists of a single genus, namely Caryocaris Salter, 1863, represented by about ten species, and numerous indeterminate forms (Fig. 2). These small arthropods, less than 5 cm long, are

t

ts sp

ts

A sr

db

t

fr

fr

B

C

s

D

Fig. 2. General morphology, carapace ultrastructure and inferred life attitude of caryocaridid arthropods. (A,B) Lateral and dorsal view of an idealised animal showing the bivalved carapace and the protruding posterior trunk. (C) Complete tail-piece of Caryocaris sp. 1 from the early Arenig of Argentina (Figs. 6 and 7: 24). (D) Right furcal ramus of Caryocaris delicatus from the Llanvirn of Argentina (Figs. 6 and 7: 27) showing articulated spinules and setae. (E) Section through compressed carapace of Caryocaris curvilata from the late Arenig/early Llanvirn of Alaska (fragment from USNM 147444). (F) Polished section through the carapace of Caryocaris wrightii from the Llanvirn of Bohemia (fragment from cat. No. 23937, Kloucek coll., National Museum, Prague) showing distribution of phosphorus. (G) Curled incomplete carapace of Caryocaris curvilata from the late Arenig/ early Llanvirn of Alaska (USNM 147444). Abbreviations: c, carapace; db, doublure; fr, furcal rami; t, telson; s, setae; sp, spinules; sr, strengthening ridge; ?ta, thoracic appendages (hypothetical); ts, last trunk segments; va, valve. Scale bars = 10 Wm in (E,F); 500 Wm in (G); 1 mm in (C,D). (E) is a scanning electron micrograph.

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

177

Fig. 3. Slabs of black shales showing caryocaridids associated with graptolites. (A,B) Caryocaris sp. 3 from Argentina (Mun‹ayoc, West Abra Pampa, Jujuy Province, Puna area; Racheboeuf et al. unpublished), mid-Arenig; general view of slab and line-drawing showing about 15 compressed caryocaridid carapaces. (C) Caryocaris stewarti from South Australia (Lance¢eld, 60 km north of Melbourne; specimen No. P 54292; see Jell, 1980), mid-Tremadoc, slab with £attened carapaces of Caryocaris and graptolite remains (Dictyonema). Abbreviations: Gr, graptolites; Ca, Caryocaris.

tomical features of the caryocaridid animal are virtually unknown apart from the abdominal termination of the body (telson attached to limbless telescopic segments) that is often preserved together with the carapace. Typically, a conical telson is £anked by a pair of leaf-like furcal rami often fringed with spine-like outgrowths, or more rarely with articulated spinules and setae (Fig. 2A,D). The bauplan of modern phyllocarids corresponds to a 5^8^7 segmentation pattern (with 5 head, 8 foliaceous thoracopods and 7 abdominal segments + telson). Although the carapace design and abdominal termination of caryocaridids indeed closely resembles those of Recent phyllocarids no evidence of caryocaridid head and complete trunk and ventral anatomy is available to con¢rm or falsify the current assignment of the group to Phyllocarida and, by extension, its crustacean status.

(e.g. Tremadoc of Australia; Fig. 3C) carapaces are preserved as white deposits (iron sulphate) produced by the oxidation of pyrite (Jell, 1980) that replaced the exoskeleton. Typically caryocaridids are laterally compressed with one valve pressed on top of the other (Fig. 5G). 3-D preserved specimens are known from Bohemia (Chlupa¤c, 1970), North America (Churkin, 1966) and the British Isles (Rushton and Williams, 1996; Fig. 5A,B). Transverse sections through mineralised carapaces of Caryocaris from Bohemia and Alaska (Fig. 2E) reveal that the lateral folds are ca. 5^10 Wm thick ( = e) and thus are extremely thin relative to the carapace length of the animal

2.2. Carapace ultrastructure In black graptolitic shales, caryocaridids typically occur as £attened, wrinkled or folded carapaces and are preserved as extremely thin, commonly shiny black imprints on the surface of bedding planes (Fig. 3A). At several localities

Fig. 4. Carapace ultrastructure of Recent planktonic crustaceans exempli¢ed by Conchoecia (Ostracoda, Halocyprididae; Paci¢c Ocean o¡ Japan). (A) Transverse para⁄n section. (B) Scanning electron micrograph showing chitinous layers. Abbreviations: ca, carapace; ec, epidermal cells; il, internal lamella; li, ligament. Scale bar = 1 Wm in (B).

PALAEO 3078 9-5-03

178

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

(L = 10^35 mm). L/e ratio = ca. 500. The thin carapace consists of three mineralised layers that may represent the typical cuticular structure of crustaceans (the epi-, exo- and endocuticle). Microprobe analysis and element mapping performed on polished sections of these specimens by Wavelength Dispersion Spectrometry and Energy Dispersion Spectrometry shows that the carapaces are preserved in phosphate (apatite; Fig. 2F). Phosphorus is also abundant between the valves (Fig. 2F). Specimens from Alaska, Bohemia and Scotland (Fig. 2G) are often spirally coiled with no major breakage between whorls. Preservational states (enrollment, wrinkles, £attening), ultrastructural observations and chemical analysis all suggest that the carapace was originally chitinous and £exible and was phosphatised in early post-mortem stages (Briggs and Kear, 1993). Phosphate concentration between the valves is probably indicative of phosphatic mineralisation of soft parts. 2.3. Comparisons with modern analogues and inferred lifestyle Caryocaridids display similarities with Recent planktonic crustaceans in several important aspects of their bauplan. In particular, there are close resemblances to planktonic halocyprid ostracods such as Conchoecia (Vannier and Chen, 2000). The latter has an extremely thin and transluscent carapace (Fig. 4 ; L/e ratio = ca. 500^1000) that is non-calci¢ed, light, formed by interconnected layers of chitin and highly £exible. The spinosity and streamlined design of the carapace of caryocaridids is also characteristic of planktonic crustaceans either at larval stages as in the

179

planktotrophic larvae of malacostracans, or adult stages (e.g. halocyprid ostracods; some hyperiidean amphipods; Fig. 1G,H). These two important adaptive features minimise the e¡ect of drag and are crucial in animals that migrate vertically through the water column. The £attened tail-piece of caryocaridids (Figs. 2C,D, 5I^M) is also a strong indication of active locomotion in the water column. This ‘leaf-like’ feature was able to fan out at the articulation between the telson and rami (Fig. 2C), and would thus have been able to maximise resistance to the water during power strokes. Fine scale adaptations such as articulated spinules along the inner side of rami (Fig. 2D), also present in Recent phyllocarids (e.g. Nebalia, Nebaliopsis; Vannier et al., 1997; Vannier and Chen, 2000), may also have contributed to lifting forces. Recent phyllocarids, including the only living pelagic phyllocarid Nebaliopsis typica (Fig. 1H ; Cannon, 1946), use their furcal rami in locomotion and directional control (see Vannier et al., 1997 for Nebalia). Unfortunately no information is available on the assumed swimming appendages of caryocaridids. Whether locomotion was performed by the beating of trunk appendages (Fig. 2A) or chie£y by the strokes of the tail fan (or by both) remains an open question. Despite these uncertainties, the morphological evidence largely supported by modern analogues, seems to validate the hypothesis that caryocaridids were midwater arthropods and active swimmers.

3. Caryocaridid occurrences An updated summary of the occurrences of caryocaridids is presented here and is based on

Fig. 5. Ordovician representatives of Caryocarididae. (A,B,L) Caryocaris wrightii from the Arenig of Britain, carapaces and tailpiece (BGS CS 98; BGS GSM 7508, specimen ¢gured by Jones and Woodward, 1892, pl. 14, ¢g. 12, possibly Upper Arenig; BGS RX 1104, respectively). (C,I,J) Caryocaris sp. a¡. wrightii, from the Llanvirn of Bohemia, carapace and tail-piece (ICH 579, NML 7020 and NML 7012, respectively). (D) Caryocaris subula from the Upper Llanvirn of Bohemia (ICH 616). (E,F) Caryocaris curvilata from the latest Arenig/early Llanvirn of East Central Alaska, general view and close-up of posterior spinosity (USNM 147449b and USNM 147442, respectively). (G,K) Caryocaris stewarti Jell, 1980 from the mid-Tremadoc of South Australia compressed slightly dislocated carapace and tail piece (P 54285, P 54288). (H) Caryocaris sp. 3 from the mid-Arenig of Argentina, carapace (Racheboeuf et al., unpublished). (M) Caryocaris delicatus from the Llanvirn of Argentina, setae along inner margin of left furcal ramus (CORD PZ 13256, specimen ¢gured in Racheboeuf et al., 2000, ¢g. 5I,J). Scale bars = 5 mm in (A^C,E,H); 3 mm in (L); 2 mm in (D,F,G,K); 1 mm in (I,J) and 500 Wm in (M). (I,J,M) are scanning electron micrographs.

PALAEO 3078 9-5-03

180

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

both published sources and unpublished information (e.g. British Isles, Belgium). Numerous new observations of type specimens from Bohemia, North America, Scandinavia, the British Isles, and Germany were made in the framework of the Treatise project (University of Kansas; part Crustacea, F.R. Schram, coord.; revision of Phyllocarida, J. Vannier and P.R. Racheboeuf, coord.). Our data also include key information on the stratigraphical distribution and diversity of caryocaridids in the Lower Ordovician of South America, namely Argentina, Bolivia and Peru (ECOS-Sud Program; P. Racheboeuf, coord., E. Brussa and collaborators, in progress). Discussions on taxonomy and synonymies are outside the scope of this paper and will be presented elsewhere. The material discussed here is in the following repositories: British Geological Survey, Keyworth, Nottingham, UK (BGS); Ca¤tedra de Estratigraf|¤a y Geolog|¤a Histo¤rica, Universidad Nacional de Co¤rdoba, Argentina (CEGH^UNC); Museo de Paleontolog|¤a de Cordo¤ba, Argentina (CORD PZ); Faculty of Science, University of Lyon I, France (FSL); I. Chlupa¤c collections, Geological Survey, Prague, Czech Republic (ICH) ; National Museum, Prague (NML) ; National Museum of Victoria, Melbourne, Australia (P); Instituto Miguel Lillo, Tucuma¤n, Argentina (PIL) ; Universidad Nacional de La Pampa, Argentina (PI^UNLPam) and United States National Museum, Washington DC (USNM). Species numbers in brackets are those used in Figs. 6 and 7. 3.1. British Isles 3.1.1. Lake District, Northern England The type species of Caryocaris, Caryocaris wrightii Salter, 1863 ( = 3) was originally described from the Skiddaw Group (Jones and Woodward, 1892). It occurs at many localities in the Kirkstile Formation (Upper Arenig, gibberulus and hirundo graptolite biozones) and is very abundant in places, even covering bedding planes. The species ranges down through the Loweswater and Hope Beck Formations (Mid- to Lower Arenig, simulans, varicosus and phyllograptoides biozones) and ranges up into the Tarn Moor Formation

(Llanvirn, artus Biozone) (Cooper et al., 1995). Small Caryocaris sp. ( = 7); more doubtfully identi¢ed, occurs in the lowest Skiddaw Group, namely the Watch Hill and Bitter Beck Formations (murrayi Biozone, uppermost Tremadoc ^ basal Arenig?) (Rushton and Williams, 1996). The Skiddaw Group consists mainly of turbidites and mud-turbidites deposited in an ocean-facing, outer-shelf to slope setting. Trace fossils indicate that the bottom waters were not fully anoxic. Associated fossils are mainly graptolites with very few other macrofossils, though there are atheloptic and mesopelagic cyclopygid trilobites, all of Gondwanan a⁄nity (Fortey et al., 1989). 3.1.2. North Wales Caryocaris wrightii and its putative synonym Caryocaris marrii Hicks, 1876 occurs in the Upper Arenig and Lower Llanvirn (gibberulus, hirundo and artus biozones) in the Carnarvon area. They are associated mainly with graptolites and cyclopygid trilobites. 3.1.3. South Wales Caryocaris wrightii and Caryocaris sp. ( = 10) are recorded sparsely from the ‘Tetragraptus Shales’ of Pembrokeshire and Carmarthenshire. They ¢rst appear from the Pontyfenni Formation of late Arenig (Fennian) age. Caryocaris has so far not been recorded from the Llanvirn or from the Early Arenig (Moridunian Stage) of South Wales. In South Wales Caryocaris is associated with graptolites and atheloptic trilobites but not with shallower-water fossils such as Neseuretus. 3.1.4. Shelve area, Welsh Borderland Whittard (1931) recorded Caryocaris sp. ( = 8) from the Mytton Flags Formation, in which it is associated with an early Arenig shallow-water Neseuretus fauna. He also mentioned Caryocaris sp. ( = 11) from the Hope Shale Formation (Llanvirn, artus Biozone), where it occurs in association with graptolites and a raphiophorid trilobite community, suggestive of a deeper shelf setting. 3.1.5. Scotland In the Arenig rocks of Ballantrae, just south

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

of Girvan, a small Caryocaris sp. ( = 9) occurs with graptolite faunas assigned to the Bendigonian and Chewtonian stages of the Australasian succession, in a probable deep-water trench setting (Stone and Rushton, 1983). Caryocaris? is also present as phosphate-rich exuviae in cherty mudstones at Stonehaven, at the NE end of the Highland Boundary Fault. The specimens are small and mostly rolled up in the manner described by Churkin (1966) and Chlupa¤c (1970) (see also Fig. 2G). Associates include sparse lingulate brachiopods, but graptolites are not known. The age of these beds remains rather uncertain. The Caradoc age proposed by Curry et al. (1984) is based on the presence of supposed tubes of Polylopia, but these are actually enrolled phyllocarid carapaces, and are therefore not diagnostic. 3.1.6. Ireland Caryocaris wrightii ( = 3) was described from Kiltrea, SE Ireland (Rushton and Williams, 1996) from green^grey early Arenig (ca. varicosus Biozone) graptolitic beds. 3.2. Belgium Specimens of Caryocaris sp. ( = 22) have recently been collected in the Condroz Ridge (Bande de Sambre-et-Meuse) of southern Belgium. The caryocariodids were recovered from the Huy Formation along the railway track at Sart-Bernard, SE of Namur. The horizons bearing Caryocaris sp. are of early Llanvirn artus graptolite Biozone age (Servais and Maletz, 1992). Fossils associated with these caryocaridids are graptolites, mesopelagic cyclopygids (Pricyclopyge and Girvanopyge) and benthic atheloptic trilobites, mostly Placoparia (Owens et al., 2001), and the ichnogenus Tomaculum Groom, 1902, that was recently interpreted as coproliths produced by epibenthic animals (Eiserhardt et al., 2001a). Caryocaridids occur throughout the Huy Formation and at several localities (e.g. Sart-Bernard, Huy). The assumed phyllocarid specimens described as Lamprocaris micans Maillieux, 1939 by Maillieux (1939) are most probably referable to Caryocaris sp.

181

3.3. Germany Caryocaridids have so far been recovered from two areas in Germany, the Ebbe-Anticline in the Rheinisches Schiefergebirge (Rhenish Massif) and from the subsurface of the island of Ru«gen in the Baltic Sea. These two areas are considered to be part of the early Palaeozoic microcontinent of Eastern Avalonia (Servais and Fatka, 1997). Koch and Brauckmann (1998) summarised the previous literature and redescribed specimens (Caryocaris wrightii ( = 3) and Caryocaris sp. ( = 20) from the Rheinisches Schiefergebirge, which were found in the Kiesbert-Tonschiefer Formation (fasciculatus Biozone) and in the Plettenberger-Ba«nderschiefer Formation (lentus Biozone), both currently attributed to the Llanvirn D. artus Biozone (Eiserhardt et al., 2001b). The £oral and faunal associates of these caryocaridids are acritarchs, chitinozoans, conularids, ostracods, foraminifers, hyolithids, trilobites, brachiopods, graptolites, and the ichnofossils Tomaculum and Chondrites. Caryocaris sp. ( = 21) was found in boreholes in Ru«gen (Jaeger, 1967). The caryocaridids occur in the Nobbin-Grauwacken Formation and the Arkona-Schwarzschiefer Formation (see Servais and Katzung, 1993; Beier et al., 2001; Maletz, 1998 for caryocaridid distribution), in levels of early Llanvirn (lentus Biozone) to early Caradoc (gracilis Biozone) age. 3.4. Scandinavia (Sweden and Norway) The only species of caryocaridid described from Sweden is Caryocaris scanicus ( = 12) which occurs in Scania in the Ceratopyge Shale of Tremadoc age (Moberg and Segerberg, 1906). North of Lund, caryocaridids are also represented by two forms (a smaller one with long dorsal spines ( = 16) and a larger species with reduced spinosity ( = 17), both associated with graptolites correlatable with the artus and murchisoni biozones in the lower part of the Llanvirn (Eckstro«m, 1937 ; P. Ahlberg (pers. commun., 1999). Possible caryocaridids ( = 14, 15) are also reported to occur at Killero«d (Scania) in a marginal isograptid biofacies of Arenig age (Bassett and Berg-Madsen, 1993); this is Protocimex siluricus, originally de-

PALAEO 3078 9-5-03

182

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

scribed by Moberg (1892) as an insect wing. Other fragments occur in coeval subsurface horizons in a core at Fa«gelsa«ng (Hede, 1951). Caryocaris cf. monodon ( = 13) occurs in the Lower Didymograptus Shale (beds 3b^3d) of the Oslo region (StMrmer, 1937), associated with Arenig graptolite assemblages (phyllograptoides to densus biozones). 3.5. Bohemia Caryocaris sp. a¡. wrighti ( = 18) is abundant in the Sarka¤ and Dobrotiva formations of Bohemia (Chlupa¤c, 1970), where it is typically associated with either graptolitic grey shales of the retro£exus Biozone or siliceous concretions with trilobites and associated faunas of the clavulus Biozone of Llanvirn age (sensu Fortey et al., 1995). Older occurrences are reported (Chlupa¤c, 1970) in the deeper-water graptolite-rich facies of the underlying Upper Arenig Klabava Formation (see Havl|¤cek and Fatka, 1992, 1994). The lower Llanvirn, where Caryocaris becomes abundant, coincides with transgressive pulses in the basin and with the onset of black shales associated with new benthic and planktonic biota. Caryocaris subula Chlupa¤c, 1970 ( = 19) is another distinctive species of the Ordovician Series of Bohemia that is abundant in the Dobrotiva Formation (retro£exus Biozone ; Llanvirn), where it is associated with a low-diversity pelagic fauna (e.g. cyclopygid trilobites, conularids and microphytoplankton; Havl|¤cek and Fatka, 1992) and deepening conditions in the central part of the basin. 3.6. North America Caryocaris curvilata Gurley, 1896 ( = 4) is present in the Road River Formation of East Central Alaska and is found in large numbers at the base of this unit. These caryocaridids are typically associated with diverse assemblages of isograptids and didymograptids that are probably equivalent to the hirundo and bi¢dus biozones. Conodonts found just below the Caryocaris-bearing horizons indicate a late Arenig age (Churkin, 1966). Other occurrences of C. curvilata ( = 4) have been noted by Churkin (1966) and previous

authors (e.g. Gurley, 1896) in Nevada (teretiusculus Biozone) and Idaho (extensus Biozone) in, again and typically, graptolite-bearing black or grey shales. Additional but poorly documented Caryocaris species are reported from Lower and Middle Ordovician strata from other North American localities; for example: C. raymondi Ruedemann, 1934 in the Athens Shale, Bristol, Tennessee and near Calera, Alabama (Ruedemann, 1934, pp. 92,93, pl. 8, ¢gs. 8^12 and pl. 23, ¢gs. 1^6); C. monodon Gurley, 1896 in the Deep Kill Shale near Melrose, New York (Gurley, 1896, p. 10 ; Ruedemann, 1934, pl. 22, ¢gs. 10^14) and C. silicula Bassler, 1919 from the Martinsburg Shale, Strasburg, Virginia (Ruedemann, 1934, pl. 23, ¢gs. 7^9). G.R. Ganis (pers. commun., 2002) has collected specimens of caryocarids ( = 5) (associated with Adelograptus tenellus) from an olistolith of Tremadoc (Cressagian) rock embedded in the Mid-Ordovician Shellsville Member of the Dauphin Formation in the western part of the Hamburg Klippe, Pennsylvania (Ganis et al., 2001, p. 115). 3.7. South America 3.7.1. Argentina Recent ¢eldwork (P.R.R., E.D.B.) has shown that caryocaridids form a recurrent component of thick graptolitic shale successions also on the South American subcontinent, exempli¢ed by the Argentine Precordillera. Details on the geological framework of the fossiliferous localities are given in Racheboeuf et al. (2000). In the Argentine Precordillera, caryocaridids occur in the Los Azules Formation (Caryocaris delicatus Racheboeuf et al., 2000 ( = 27); austrodentatus to teretiusculus biozones, early^late Llanvirn), the Gualcamayo Formation in the Precordillera Central de San Juan (unpublished Caryocaris sp. 4 ( = 26); austrodentatus and elegans biozones; Llanvirn), and the Las Plantas Formation (Caryocaris sp. 2 ( = 28); bicornis Biozone; early Caradoc). In the Sierra de Famatina Caryocaris is found in the Suri Formation (unpublished; mid-Arenig horizons) and it was recovered recently in shales attributed to the Tremadoc (Caryocaris bodenbenderi Acen‹olaza and Esteban, 1996 ( = 23), from

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

the Volcancito Formation) ; this is the oldest record of caryocaridids in South America (Acen‹olaza and Esteban, 1996). Caryocaridids also occur in Northern Argentina (Puna area, Eastern Cordillera), in the Acoite Formation (unpublished Caryocaris sp.; T. phyllograptoides to bi¢dus biozones; Arenig), the Parcha Formation (Caryocaris sp. 1 ( = 24); phyllograptoides and akzharensis biozones, early Arenig), as well as in the Mun‹ayoc section, Sierra de Quichagua (Caryocaris sp. 3 ( = 25), Middle Arenig). The only fauna associated with these various caryocaridids are graptolites (Fig. 3A,B), rarely pelagic trilobites and, locally, inarticulate lingulid brachiopods (e.g. northern part of Precordillera ; Gualcamayo Formation). In the Sierra de Famatina, the upper member of the Volcancito Formation (Tremadoc) yields abundant caryocaridids (?Caryocaris) associated with agnostid, olenid, and cyclopygid trilobites (Esteban, 1999). These graptolite-bearing black shales are typical basinal and/or slope deposits. In general, caryocaridids are absent from calcareous horizons in which benthic organisms abound and also from graptolitic calcareous shales. 3.7.2. Bolivia and Peru In the Tarija area of southern Bolivia, caryocaridids are present in the uppermost part of the Pircancha Formation (Caryocaris sp. ( = 29), Hannibal and Feldmann, 1996 ; Azygograptus fauna, mid-Arenig, Maletz et al., 1995). Caryocaris sp. also occurs at Culpina, associated with Tetragraptus quadribrachiatus, in the upper part of the Las Cieneguillas Formation (Arenig^?Llanvirn; Sua¤rez Soruco, 1976). A single species, Caryocaris acuta ( = 30) has been described from Huichiyuni, Peru (Bulman, 1931) in possibly Caradoc graptolite-bearing shales. 3.8. China Caryocaris zhejiangensis Shen, 1986 ( = 6) is the only known representative of the group from the South China Plate. It was recovered from the Arenig graptolitic shales (suecicus/abnormis Biozone to lower part of Cardiograptus/nexus Biozone ; Shen, 1986) of Zhejiang Province, SE China.

183

3.9. Australia and New Zealand Caryocaris stewarti Jell, 1980 ( = 1); occurs in South Australia, associated with graptolites (Dictyonema campanulatum, D. scitulum and Staurograptus di⁄sus; Fig. 3C) indicating a mid-Tremadoc age. C. marrii, a probable synonym of C. wrightii (see Chlupa¤c, 1970) is reported from the Lower Ordovician of Marong, near Daylesford and in erratics derived from similar horizons in Tasmania (see Chapman, 1934). C. marrii ( = 2), C. wrightii ( = 3) and a few other ill-de¢ned allied taxa are recorded (Chapman, 1934) from the probable Arenig of Providence Inlet, New Zealand.

4. Distribution in time and space, ecological range 4.1. Caryocaridids through time It is during the Arenig (50% of total occurrences) and the Llanvirn (34%) that caryocaridids seem to attain their highest diversity and numerical abundance. However, the true biodiversity of the group remains di⁄cult to evaluate due to numerous taxonomic uncertainties at species level (‘Caryocaris sp.’; Fig. 6). The oldest known caryocaridids, C. stewarti from Australia and C. bodenbenderi from Argentina, are from the Tremadoc whilst the group ranges up into the Caradoc (8% of the total occurrences, mostly in South America; Fig. 6) when they appear to have become extinct. In the British Isles Caryocaris is commonest in the Arenig Series, ranges down into the top of the Tremadoc (murrayi Biozone; Lake District) but is rare in the Lower Llanvirn (artus Biozone; e.g. in the Lake District, North Wales and Shelve area). Caryocaridids seem to be more frequent elsewhere during the Llanvirn, for example in Bohemia, Germany, Belgium, Baltoscandia, and both North and South America. The Cambrian ancestry of caryocaridids has not been addressed yet, but the waptiid arthropods, which were also bivalved arthropods with a phyllocarid-like body plan and were well represented in the early and middle Cambrian biota (e.g. Burgess Shale, Chengjiang, Sirius Passet Lagersta«tten; Conway

PALAEO 3078 9-5-03

184

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

Morris et al., 1982; Briggs, 1983; Briggs et al., 1994; Chen and Zhou, 1997; Taylor, 2002) may provide some hints concerning the origin of the group. Their external anatomy suggests a nektobenthic lifestyle, probably comparable to that of modern phyllocarids. We propose as an hypothesis to be tested by future studies that caryocaridids may have arisen from the waptiid stock through a benthic-to-pelagic ecological shift taking place in Late Cambrian times. Siveter et al. (1991) have shown the importance of such ecological events in documenting the (early Silurian) radiation of another important arthropod group, the myodocope lineage of the Ostracoda. 4.2. Caryocaridid palaeogeographical distribution We have plotted the Tremadoc to early Caradoc occurrences of caryocaridids on world maps for the early Ordovician (ca. 480 Ma; Cocks, 2001 and Fig. 7). Caryocaridids are present around the margins of several continental blocks such as Gondwana, Laurentia, Baltica, Avalonia, Perunica, and South China within a broad transequatorial latitudinal range, ca. 75‡S to 20‡N. Caryocaris has virtually a cosmopolitan distribution, although there are no records as yet from Siberia and Kazakhstan. Conspeci¢c occurrences between distant regions are uncertain except for Caryocaris wrightii, which is recognised within a relatively narrow latitudinal belt (ca. 75‡S) in Avalonia (British Isles, Northern Germany) and Perunica (C. a¡. wrightii in Bohemia). Caryocaris species

185

are found in Avalonia^Baltica and Laurentia, on either side of the Iapetus Ocean, a feature that acted as a major oceanic barrier for benthic faunas during the early Ordovician (Fig. 7; Williams et al., in press). The pattern of distribution of caryocaridids di¡ers markedly from that of the shallow-water benthic faunas as exempli¢ed by trilobites (Fig. 7). The geographic di¡erentiation and endemicity of shallow shelf benthic trilobites are strongly constrained by oceanic barriers and a host of environmental factors (e.g. temperature) induced by latitude. Indeed, the more widespread distribution of caryocaridids recalls that of Early Ordovician planktonic organisms such as didymograptid and isograptid graptolites (Cocks and Fortey, 1988; Cooper et al., 1991), suggesting that latitude may not have exerted a primary control on their dispersal. Fortey (1980) showed that olenid trilobites were similarly independant of palaeogeographic distribution. The spatial distribution of caryocaridids is also consistent with the assumed free-swimming lifestyle and midwater habitat of the group that is inferred from functional anatomy (see above; Fig. 2). A more accurate de¢nition of the taxa will certainly allow better elucidation of the actual degree of pandemism and dispersal capabilities of the group. 4.3. Depositional environment of caryocaridids Caryocaridids are overwhelmingly commoner in outer-shelf or slope settings. Shallow water inner-shelf occurrences, like that in the Arenig of

Fig. 6. Stratigraphical distribution of Ordovician caryocaridid arthropods (?Phyllocarida). Legend: 1, Caryocaris stewarti (Australia); 2, Caryocaris marrii (South Australia and New Zealand); 3, Caryocaris wrightii (England; North and South Wales, UK; Ireland; Germany; ?South Australia); 4, Caryocaris curvilata (USA); 5, caryocaridids (Pennsylvania); 6, Caryocaris zhejiangensis (SE China); 7, Caryocaris sp. (Lake District, UK); 8, Caryocaris sp. (Welsh Borderland, UK); 9, Caryocaris sp. (Scotland, UK); 10, Caryocaris sp. (South Wales, UK); 11, Caryocaris sp. (Welsh Borderland, UK); 12, Caryocaris scanicus (Sweden); 13, Caryocaris cf. monodon (South Norway); 14,15, Protocimex siluricus and probable caryocaridid (Sweden); 16,17, Caryocaris sp. (form with long dorsal spines and larger form, both from South Sweden); 18, Caryocaris sp. a¡. wrightii (Bohemia, Czech Republic); 19, Caryocaris subula (Bohemia, Czech Republic); 20, Caryocaris sp. (Rheinisch Massif, Germany); 21, Caryocaris sp. (Ru«gen, Germany); 22, Caryocaris sp. (Belgium); 23, Caryocaris bodenbenderi (Argentina); 24, Caryocaris sp. 1 (Northern Argentina); 25, Caryocaris sp. 3 (Argentina); 26, Caryocaris sp. 4 (Precordillera, Argentina); 27, Caryocaris delicatus (Precordillera, Argentina); 28, Caryocaris sp. 2 (Precordillera, Argentina); 29, Caryocaris sp. (Bolivia); 30, Caryocaris acuta (Peru). Graptolite biozonations, time scale and correlations for Australasia, North America, China, Britain and Baltoscandia after B. Webby, R. Cooper, S. Bergstro«m and F. Paris (unpublished chart; F. Paris, pers. commun., July 2001) and Fortey et al. (2000). Stratigraphical distribution of caryocaridids in Germany, Belgium and South America also based on associated graptolites (zonation not given for these three regions).

PALAEO 3078 9-5-03

186

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

Fig. 7. Caryocaridid occurrence plotted (black and grey dots) on global maps for the early Ordovician (Arenig; ca. 480 Ma), centred on the North Iapetus Ocean (A) and on the Gondwanan supercontinent (B). Trilobite benthic faunas indicated for comparison (after Cocks and Fortey, 1990 and Cocks, 2001). Continental distribution as given by Cocks (2001; modi¢ed from C.R. Scotese PaleoGIS for Arcview package). Legend: 1, Northern Laurentia and Chukotka Arc (Natal’in et al., 1999); 2,3, Notre-Dame Arc and Penobscot Arc (Van Staal et al., 1998); 4, Kipchak Arc (Sengo«r and Natal’in, 1996). Species numbers same as in Fig. 2. Carapace outline of representative species after Racheboeuf et al., 2000.

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

187

Fig. 8. Assumed bathymetrical distribution of caryocaridids (evidence from depositional environments and graptolite/trilobite associates). Graptolite and trilobite biofacies after Finney and Berry (1997). (A) Idealised caryocaridid arthropod. (B,C) The epipelagic trilobites Telephina spinifera and Carolinites genacinaca (after Fortey, 1985, ¢g. 1). (D) Isograptus caduceus. (E) Didymograptus (D.) stabilis (D,E, after Cooper et al., 1991, ¢g. 3).

the Welsh Borderland, are extremely rare. Caryocaridids form a recurrent component of black shales that are often intercalated in thick deepsea fan^turbidite sequences deposited in oceanfacing settings. They are are almost invariably associated with graptolites, which apart from their use in dating and correlating the caryocarididbearing horizons (Fig. 6) also provide valuable indications of the depositional setting and bathymetry. The ‘isograptid biofacies’ (Tremadoc^ Llanvirn) contains typically pandemic taxa con¢ned to ocean-facing continental margin and marginal basin sites (Fig. 8 ; see Fortey and Cocks, 1986; Cooper et al., 1991). By contrast the ‘didymograptid biofacies’ which is recognised by the numerical dominance of broad-stiped graptolites and the absence of deeper water elements, is associated with relatively shallower shelf lithofacies (Fig. 8). Although caryocaridids do occur in association with isograptids (e.g. Alaska, Sweden,

England) they are frequently associated with graptolites referred to both deeper and shallower biofacies (Cooper et al., 1991). Trilobites are also recurrent associates of caryocaridids, mainly giant-eyed cyclopygids, that probably lived at depths of ca. 200^500 m in the water column (Fig. 8B,C ; see Fortey, 1985; McCormick and Fortey, 1998), and 2), but also blind and reduced-eyed ‘atheloptic’ benthic forms (Fortey and Owens, 1987) typical of deep-water or slope facies. The occurrences of caryocaridids with olenid and raphiophorid trilobites are far less frequent. 4.4. Caryocaridid biotope The relatively ‘marginal’ depositional setting of early Ordovician caryocaridids, added to their inferred planktonic lifestyle, poses the question of their inferred ecological niche, either epipelagic

PALAEO 3078 9-5-03

188

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

and/or mesopelagic. Similar questions have long been debated concerning another important component of the Ordovician zooplankton, the graptolites (e.g. Cooper et al., 1991; Goldman et al., 1995). Finney and Berry (1997) proposed a model in which graptolites thrived in waters over the outer shelf and the proximal part of the continental slope. This model is in broad accordance with the actual distribution of plankton, including crustaceans, in recent marine ecosystems (Fig. 1). In modern oceans, the relatively narrow band of waters that fringes the continental margins is characterised by high primary productivity and high density of plankton (on account of food supplies in upwelling currents). Although it is evident that the food requirements and feeding strategies of the drifting colonial graptolites were probably di¡erent from those of actively swimming arthropods like caryocaridids, both groups may have shared the same ecological niche, which had abundant food supplies and seasonal water circulation. The anoxic/dysoxic depositional environments that characterise most early Ordovician graptolite and caryocaridid-bearing black shales are likely to have been a consequence of the presence of an oxygen-minimum zone generated by the overlying highly productive waters. This is an important aspect of Finney and Berry’s (1997) graptolite model that possibly applies to numerous midwater organisms such as caryocaridids. Although it becomes clear that caryocaridids were neither abundant in neritic waters over the inner shelf nor probably in open oceanic waters, there is no direct evidence to infer their preferential bathymetrical level within the water column, whether epipelagic or mesopelagic. Recent crustacean zooplankton such as the halocyprid ostracods (Vannier and Chen, 2000) are cosmopolitan and active swimmers. Numerous ‘mesopelagic’ crustaceans (e.g. halocyprid and cypridinid ostracods, euphausiaceans ; Fig. 1, Nos. 18,19,23) exploit intensively the food resources of the epipelagic zone (at least at its lower levels) and for that reason perform signi¢cant migrations through the water column. For example, halocyprids ostracods (Vannier et al., 1998) are microphagous feeders with diatoms, silico£agellates, coccolithophorids, foraminifers, detrital ag-

gregates and various damaged organisms (e.g. crustaceans; Ikeda and Inamura, 1992) constituting the most frequent components of their gut contents. We consider that the Ordovician caryocaridids may belong to the same category of bathymetrical migrants and had comparable feeding strategies.

5. Conclusions (1) Caryocaridids have the morphological characteristics of midwater active swimmers (carapace design and ultrastructure, abdominal morphology) that are recognised in Recent planktonic crustaceans. (2) Their oldest record is Tremadoc. The group attains its highest diversity and abundance during the Arenig and the Llanvirn and probably became extinct during the Caradoc. (3) The cosmopolitan pattern of caryocaridids is indicated by the occurrences of Caryocaris in several major palaeo-plates of Gondwana (South America, Australia), Laurentia, Avalonia, Perunica, Baltica, South China) within a broad transequatorial latitudinal distribution (75‡S to 20‡N). (4) Caryocaridids are overwhelmingly commoner in outer-shelf or slope settings where they are typically associated with didymograptid and isograptid graptolites, and pelagic cyclopygid and deep-sea benthic atheloptic trilobites. (5) Caryocaridids are likely to have occupied pelagic niches situated at the fringes of continental margins where primary productivity is high and upwelling currents active. They may have been vertical migrants within the water column in relation to their feeding strategies, possibly occupying both the mesopelagic and the lower epipelagic zones, as do numerous midwater crustaceans in present-day ecosystems. (7) The caryocaridid arthropods represent a signi¢cant pulse of invasion of the pelagic realm, that started in the Tremadoc, and provide additional evidence for the diversity of zooplankton (e.g. graptolites, pelagic trilobites, orthoconic nautiloids) and for the existence of complex marine food chains in the early Ordovician.

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

Acknowledgements J.V. is very grateful to numerous colleagues for the loan of specimens, especially from the British Geological Survey (Keyworth, Nottingham), the United States National Museum (Washington, DC), the Australian Museum (Melbourne), the National Museum and the Geological Survey (Prague), and to Drs. C. Franzen (Stockholm), B. Weber (Berlin), P. Ahlberg (Lund), V. BergMadsen (Uppsala), P. Noel (Museum d’Histoire Naturelle, Paris) and Mr. L. Koch (Ennepetal). P.R.R. and E.D.B. are grateful to CONICET, to the Universidad Nacional de La Pampa for grants in the framework of a joint French^Argentinian research project (ECOS-Sud ; Grant A99U03) and to their Argentinian colleagues, especially B. Waisfeld and B. Toro for assistance in ¢eldwork. We thank M.N. Podevigne for his assistance in photographic work and the Centre de Microscopie a' Balayage, Universite¤ Claude Bernard Lyon I for SEM facilities. M. Williams publishes by permission of the Director, British Geological Survey. This paper is a contribution to the team project of UMR 5125 (CNRS) on the structure and functioning of aquatic palaeoecosystems.

References Acen‹olaza, F.S., Esteban, S., 1996. Filoca¤ridos (Crustacea) en el Tremadociano del Sistema de Famatina, Provincia de La Rioja, Argentina. Mem. XII Congr. Geol. de Bolivia, Tarija, pp. 281^288. Bassett, M.G., Berg-Madsen, V., 1993. Protocimex: a phyllocarid crustacean, not an Ordovician insect. J. Pal. 67, 144^ 147. Bassler, R.S., 1919. Cambrian and Ordovician. Md. Geol. Surv. 1919, 1^424. Beier, H., Katzung, G., Maletz, J., Schallreuter, R., 2001. Vorpommern und angrenzende Ostsee (Ru«gen^Kaledoniden). In: Stratigraphie von Deutschland II, Ordovizium, Kambrium, Vendium, Ripha«ikum, Teil III. Courier des Forschungsinstitut Senckenberg 235, pp. 120^134. Briggs, D.E.G., 1983. A⁄nities and early evolution of the Crustacea: the evidence of the Cambrian fossils. In: Schram, F.F. (Ed.), Crustacean Phylogeny. Balkema, Rotterdam, pp. 1^22. Briggs, D.E.G., Kear, A.J., 1993. Fossilization of soft-tissue in the laboratory. Science 259, 1439^1442.

189

Briggs, D.E.G., Erwin, D.H., Collier, F.J., 1994. The Fossils of the Burgess Shale. Smithsonian Institution Press, Washington, DC, 238 pp. Bulman, O.M.B., 1931. South American Graptolites with special references to the Nordenskio«ld collections. Ark. Zool. 22A, 1^111. Butter¢eld, N.J., 1994. Burgess Shale-type fossils from a Lower Cambrian shallow-shelf sequence in northwestern Canada. Nature 369, 477^479. Butter¢eld, N.J., 1997. Plankton ecology and the Proterozoic^ Phanerozoic transition. Paleobiology 23, 247^262. Butter¢eld, N.J., 2002a. Cambrian food webs. In: Briggs, D.E.G., Crowther, P.R. (Eds), Palaeobiology II. Blackwell Science, Oxford, 607 pp. Butter¢eld, N.J., 2002b. Leanchoilia guts and the interpretation of three dimensional structures in Burgess Shale-type fossils. Paleobiology 28, 155^171. Cannon, H.G., 1946. Nebaliopsis typica. Discov. Rep. 33, 213^ 222. Chapman, F., 1934. On some phyllocarids from the Ordovician of Preservation Inlet and Cape Providence, New Zealand. Trans. R. Soc. N. Z. 64, 105^114. Chen, J.-Y., Zhou, G.-Q., 1997. Biology of the Chengjiang fauna. Bull. Nat. Mus. Nat. Sci. 10, 11^105. Chlupa¤c, I., 1970. Phyllocarid crustaceans of the Bohemian Ordovician. Sbor. Geol. Ved. Paleontol. 12, 41^75. Churkin, M., Jr., 1966. Morphology and stratigraphic range of the phyllocarid crustacean Caryocaris from Alaska and the Great Basin. Palaeontology 9, 371^380. Cocks, L.R.M., 2001. Ordovician and Silurian global geography. J. Geol. Soc. Lond. 158, 197^210. Cocks, L.R.M., Fortey, R.A., 1988. Lower Palaeozoic facies and faunas around Gondwana. In: Audley-Charles, M.G., Hallam, A. (Eds.), Gondwana and Tethys. Geol. Soc. Lond. Spec. Publ. 37, pp. 183^200. Cocks, L.R.M., Fortey, R.A., 1990. Biogeography of Ordovician and Silurian faunas. In: McKerrow, W.S., Scotese, C.R. (Eds.), Palaeozoic Palaeogeography and Biogeography. Geol. Soc. Lond. Mem. 12, pp. 94^104. Conway Morris, S., Whittington, H.B., Briggs, D.E.G., Hughes, C.P., Bruton, D.L., 1982. Atlas of the Burgess Shale. Palaeontol. Assoc., London, 234 pp. Cooper, R.A., Fortey, R.A., Lindholm, K., 1991. Latitudinal and depth zonation of Early Ordovician graptolites. Lethaia 24, 199^218. Cooper, A.H., Rushton, A.W.A., Molyneux, S.G., Hughes, R.A., Moore, R.M., Webb, B.C., 1995. The stratigraphy, correlation, provenance and palaeogeography of the Skiddaw Group (Ordovician) in the English Lake District. Geol. Mag. 132, 185^211. Curry, G.B., Bluck, B.J., Burton, C.J., Ingham, J.K., Siveter, D.J., Williams, A., 1984. Age, evolution and tectonic history of the Highland Border Complex, Scotland. Trans. R. Soc. Edinb.: Earth Sci. 75, 113^133. Eckstro«m, G., 1937. Upper Didymograptus Shale of Scania. Sver. Geol. Unders. Ser. C 403, 1^51. Eiserhardt, K.H., Koch, L., Eiserhardt, W.L., 2001a. Revision

PALAEO 3078 9-5-03

190

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191

des Ichnotaxon Tomaculum Groom, 1902. Neues Jahrb. Geol. Pala«ontol. Abh. 221, 325^358. Eiserhardt, K.H., Koch, L., Maletz, J., 2001b. No«rdliches Rheinisches Schiefergebirge. In: Stratigraphie von Deutschland II, Ordovizium, Kambrium, Vendium, Ripha«ikum. Teil III. Courier des Forschungsinstitut Senckenberg 235, pp. 90^108. Esteban, S.B., 1999. Cyclopygid trilobites and associated facies from the Ordovician of the Famatina Basin (Northwestern Argentina): paleogeography and evolutionary implications. Acta Univ. Carolinae Geol. 43, 377^379. Finney, S.C., Berry, W.B.N., 1997. New perspectives on graptolite distributions and their use as indicators of platform margin dynamics. Geology 25, 919^922. Fortey, R.A., 1980. The Ordovician trilobites of Spitsbergen, Part III. Remaining trilobites of the Valhallfonna Formation. Skr. Norsk. Polarinstit. 171, 1^163. Fortey, R.A., 1985. Pelagic trilobites as an example of deducing the life habits of extinct arthropods. Trans. R. Soc. Edinb. 76, 219^230. Fortey, R.A., Cocks, L.R.M., 1986. Marginal faunal belts and their structural implications, with examples from the Lower Palaeozoic. J. Geol. Soc. Lond. 123, 151^160. Fortey, R.A., Owens, R.M., 1987. The Arenig Series in South Wales. Bull. Brit. Mus. Nat. Hist. Geol. 41, 69^307. Fortey, R.A., Owens, R.M., Rushton, A.W.A., 1989. The palaeogeographic position of the Lake District in the Early Ordovician. Geol. Mag. 126, 9^17. Fortey, R.A., Harper, D.A.T., Ingham, J.K., Owen, A.W., Rushton, A.W.A., 1995. A revision of Ordovician series and stages from the historical type area. Geol. Mag. 132, 15^30. Fortey, R.A., Harper, D.A.T., Ingham, J.K., Owen, A.W., Parkes, M.A., Rushton, A.W.A., Woodcock, N.H., 2000. A revised correlation of Ordovician rocks in the British Isles. Geol. Soc. Spec. Rep. 24, 1^83. Ganis, G.R., Williams, S.H., Repetski, J.E., 2001. New biostratigraphic information from the western part of the Hamburg Klippe, Pennsylvania, and its signi¢cance for interpreting the depositional and tectonic history of the klippe. Bull. Geol. Soc. Am. 113, 109^128. Goldman, D., Bergstro«m, S.M., Mitchell, C.E., 1995. Revision of the Zone 13 graptolite biostratigraphy in the Marathon, Texas, standard succession and its bearing on Upper Ordovician graptolite biostratigraphy. Lethaia 28, 115^128. Gurley, R.R., 1896. North American graptolites; new species and vertical range. J. Geol. 4, 291^311. Hannibal, J.T., Feldmann, R.M., 1996. Caryocaris (Crustacea: Phyllocarida) from the Ordovician of the Cordillera oriental of southern Bolivia. Kirtlandia 49, 7^11. Havl|¤cek, V., Fatka, O., 1992. Ordovician of the Prague Basin (Barrandian area, Czechoslovakia). In: Webby, B., Laurie, R. (Eds.), Global Perspectives on Ordovician Geology. Balkema, Rotterdam, pp. 461^471. Havl|¤cek, V., Fatka, O., 1994. Perunica microcontinent in the Ordovician (its position within the Mediterranean Province,

series division, benthic and pelagic associations). Sbor. Geol. Ved. Paleontol. 46, 23^56. Hede, J.E., 1951. Boring through Middle Ordovician^Upper Cambrian strata in the Fafigelsafing district, Scania (Sweden), 1. Succession encountered in the boring. Acta Univ. Lund. N. F. 46, 1^78. Hicks, H., 1876. Fossiliferous Cambrian Shales near Caernarvon. Quart. J. Geol. Soc. Lond. 32, 79^88. Ikeda, T., Inamura, A., 1992. Population structure and life cycle of the mesopelagic ostracod Conchoecia pseudodiscophora in Toyama Bay, Southern Japan Sea. Mar. Biol. 1113, 595^601. Jaeger, H., 1967. Ordoviz auf Ru«gen. Datierung und Vergleich mit anderen Gebieten (vorlau«¢ge Mitteilung). Berl. Dtsch. Ges. geol. Wiss. A12, 165^176. Jell, P.A., 1980. Two arthropods from the Lancefeldian (La1) of central Victoria. Alcheringa 4, 37^46. Jones, T.R., Woodward, H., 1892. A Monograph of the British Palaeozoic Phyllopoda (Phyllocarida Packard), part 2. Monogr. Palaeontogr. Soc. (for 1892), pp. 73^124. Koch, L., Brauckmann, C., 1998. Phyllocariden (Crustacea) aus dem unteren Llanvirn (Ordovizium) des Elbe-Sattels (Rheinishes Schiefergebirge). Neues Jahrb. Geol. Pala«ontol. Mon.heft. 1998, 55^64. LeGrand, S., Hannibal, J., 2000. Rock shrimp. Explorer 41, 11^13. Maillieux, E., 1939. L’Ordovicien de Sart Bernard. Me¤m. Mus. Hist. Nat. Belg. 86, 1^59. Maletz, J., 1998. Die Graptolithen des Ordoviziums von Ru«gen (Norddeutschland, Vorpommern). Pala«ontol. Z. 72, 351^372. Maletz, J., Kley, J., Reinhardt, M., 1995. New data on the Palaeontology and Biostratigraphy of the Ordovician in southern Bolivia. Newslett. Strat. 32, 163^173. McCormick, T., Fortey, R.A., 1998. Independent testing of a paleobiological hypothesis: the optical design of two Ordovician pelagic trilobites reveals their relative bathymetry. Paleobiology 24, 235^253. Moberg, J.C., 1892. Om nafigra nya graptoliter frafin Skafines Undre graptolit-ski¡er. Sver. Geol. Und. Ser. C 125, 1^12. Moberg, J.C., Segerberg, C.O., 1906. Bidrag del ka«nnedomen om Ceratopygeregionen med sa«rskild ha«nsyn till dess utveckling i Fogelsafingstrakten. Medd. Lunds geol. Fa«ltklubb Ser. B 2, 1^116. Natal’in, B.A., Amato, J.M., Toro, J., Wright, J.E., 1999. Palaeozoic rocks of northern Chukotka Peninsula, Russian Far East: implications for the tectonics of the Arctic region. Tectonics 18, 977^1003. Owens, R.M., Servais, T., Koch, L., Fatka, O., 2001. Trilobites from the Llanvirn of the Condroz Ridge, Belgium, and their palaeogeographical signi¢cance. Abstracts Early Palaeozoic Palaeogeographies and Palaeobiogeographies of Western Europe and North Africa, Lille, pp. 24^26. Racheboeuf, P.R., Vannier, J., Ortega, G., 2000. Ordovician phyllocarids (Arthropoda; Crustacea) from Argentina. Pala«ontol. Z. 74, 317^333.

PALAEO 3078 9-5-03

J. Vannier et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 173^191 Rigby, S., 1997. A comparison of the colonization of the planktic realm and the land. Lethaia 30, 11^17. Rigby, S., Milsom, C., 1996. Benthic origins of zooplankton: an environmentally determined macroevolutionary e¡ect. Geology 24, 52^54. Ruedemann, R., 1934. Palaeozoic plankton of North America. Mem. Geol. Soc. Am. 2, 1^133. Rushton, A.W.A., Williams, M., 1996. The tail-piece of the crustacean Caryocaris wrightii from the Arenig rocks of England and Ireland. Ir. J. Earth Sci. 15, 107^111. Salter, J.W., 1863. Note on the Skiddaw Slate fossils. Quart. J. Geol. Soc. Lond. 19, 135^140. Sengo«r, A.M.C., Natal’in, B.A., 1996. Paleotectonics of Asia: fragments of a synthesis. In: Yin, A., Harrison, T.M. (Eds.), The Tectonic Evolution of Asia. Cambridge University Press, Cambridge, pp. 486^640. Servais, T., Fatka, O., 1997. Recognition of the Trans-European Suture Zone (TESZ) by the palaeobiogeographical distribution pattern of Early to Middle Ordovician acritarchs. Geol. Mag. 134, 617^625. Servais, T., Katzung, G., 1993. Acritarch dating of Ordovician sediments of the island of Ru«gen (NE-Germany). Neues Jahrb. Geol. Pala«ontol. Mon.heft. 12, 713^723. Servais, T., Maletz, J., 1992. Lower Llanvirn (Ordovician) graptolites and acritarchs from the ‘Assise de Huy’, Bande de Sambre-et-Meuse, Belgium. Ann. Soc. Ge¤ol. Belg. 115, 265^284. Shen, 1986. Caryocaris from the Lower Ordovician of Jiangshan, Zhejiang. Kexue Tongbao 31, 765^769. Siveter, D.J., Williams, M., 1997. Cambrian bradoriid and phosphatocopid arthropods of North America. Spec. Pap. Palaeontol. 57, 1^69. Siveter, D.J., Vannier, J., Palmer, D., 1991. Silurian myodocopes: pioneer pelagic ostracods and the chronology of an ecological shift. J. Micropaleontol. 10, 151^173. Stone, P., Rushton, A.W.A., 1983. Graptolites from the Ballantrae ophiolite complex and their structural implications. Scott. J. Geol. 19, 297^310. StMrmer, L., 1937. Planktonic crustaceans from the Lower Didymograptus Shale (3b) of Oslo. Norsk Geol. Tidsskr. 16, 267^278. Sua¤rez Soruco, R., 1976. El sistema Ordov|¤cico en Bolivia. Rev. Te¤c. YPFB 5, 111^223.

191

Taylor, R., 2002. A new bivalved arthropod from the Early Cambrian Sirius Passet fauna, North Greenland. Palaeontology 45, 97^123. Van Staal, C.R., Dewey, J.F., McNiocaill, C., McKerrow, W.S., 1998. The Cambrian^Silurian tectonic evolution of the Northern Appalachian and British Caledonides: history of a complex, west and southwest Paci¢c-type segment of Iapetus. In: Blundell, D.J., Scott, A.C. (Eds), Lyell: the Past is the Key of the Present. Geol. Soc. Lond. Spec. Publ. 143, pp. 199^242. Vannier, J., Abe, K., 1995. Size, body plan and respiration in the Ostracoda. Palaeontology 38, 843^873. Vannier, J., Chen, J.-Y., 2000. The Early Cambrian colonization of pelagic niches exempli¢ed by Isoxys (Arthropoda). Lethaia 33, 295^311. Vannier, J., Abe, K., Ikuta, K., 1998. Feeding in myodocopid ostracods: functional morphology and laboratory observations from videos. Mar. Biol. 132, 391^408. Vannier, J., Boissy, Ph., Racheboeuf, P., 1997. Locomotion in Nebalia bipes: a model for Palaeozoic crustaceans? Lethaia 30, 89^104. Vetter, E.W., 1994. Hotspots of benthic production. Science 372, 47. Vetter, E.W., 1995. Detritus-based patches of high secondary production in the nearshore benthos. Mar. Ecol. Prog. Ser. 120, 251^262. Vetter, E.W., 1996. Secondary production of a Southern California Nebalia (Crustacea: Leptostraca). Mar. Ecol. Prog. Ser. 137, 95^101. Walossek, D., Mu«ller, K.J., 1989. A second type A-nauplius from the Upper Cambrian ‘Orsten’ of Sweden. Lethaia 22, 301^306. Whittard, W.F., 1931. The geology of the Ordovician and Valentian rocks of the Shelve country, Shropshire. Proc. Geol. Assoc. 42, 322^339. Williams, M., Siveter, D.J., Peel, J.S., 1996. Isoxys (Arthropoda) from the Early Cambrian Sirius Passet Lagersta«tte, North Greenland. J. Pal. 70, 947^954. Williams, M., Floyd, J.D., Salas, M.J., Stone, P., Siveter, D.J., Vannier, J., in press. Patterns of ostracod migration for the ‘North Atlantic’ region during the Mid and Late Ordovician. Palaeogeogr. Palaeoclimatol. Palaeoecol.

PALAEO 3078 9-5-03

Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228 www.elsevier.com/locate/palaeo

Patterns of ostracod migration for the ‘North Atlantic’ region during the Ordovician Mark Williams a; , James D. Floyd b , Maria Jose Salas c , David J. Siveter d , Philip Stone b , Jean M.C. Vannier e a

British Geological Survey, Keyworth, Nottingham NG12 5GG, UK British Geological Survey, Murchison House, West Mains Road, Edinburgh EH9 3LA, UK Catedra de Estratigra¢a y Geologia Historica, Facultad de Ciencias Exactas, Fisicas y Naturales, Universidad Nacional de Cordoba, Av. Velez Sars¢eld 299, Cordoba 5000, Argentina d Department of Geology, University of Leicester, Leicester LE1 7RH, UK e Universite¤ Claude Bernard-Lyon 1, UFR Sciences de la Terre, UMR PEPS, 43 Boulevard du 11 Novembre 1918, 69622 Villeurbanne, France b

c

Received 13 May 2002; received in revised form 2 September 2002; accepted 15 January 2003

Abstract A review of Ordovician neritic ostracods from the ‘North Atlantic’ region including Europe and North America identifies over 100 genera (including 44 palaeocopes and 31 binodicopes) which show a complex pattern of migration between two or more of the palaeocontinents Gondwana, Ibero-Armorica, Perunica, Avalonia, Baltica and Laurentia. Many dispersals were relatively slow, and the migration of a genus between palaeocontinents often took the duration of one or more graptolite biozones. Over 70 migrations appear to have occurred more rapidly, including those of Pseudulrichia, a genus which dispersed to five palaeocontinents within the duration of three graptolite biozones. Longevity clearly facilitated the chances of migration, as the most widespread genera such as Vannieria, Platybolbina, Medianella and Euprimites, are often the most long-ranging. Low migration rates prior to the Llanvirn are, at least in part, related to low ostracod taxonomic diversity. Greatly increased diversity from the late Llanvirn coincided with a much higher rate of migration. Coupled with the spread of carbonate^mudstone shelf marine facies in Laurentia during the early and mid Caradoc, this resulted in the migration of up to 18 Baltic-origin genera to Laurentia. Relative to overall ostracod diversity, migration rates were generally higher during periods of lower global sea level, suggesting that ostracod dispersal may have been aided by mid-ocean islands or outer-shelf carbonate platforms, which provided more extensive island-hopping routes during periods of low sea level. The palaeogeographical convergence of Avalonia, Perunica and Baltica, and subsequently of Avalonia and Baltica with Laurentia, in low latitudes and warm surface waters, is suggested by increasing ostracod migration between these palaeocontinents from the late Llanvirn onwards. This culminated, during the Ashgill, in numerous species^level links. Baltica may have been the source area for more than 40 migrant genera, reflecting its high-diversity faunas and its intermediate palaeogeographical position between Laurentia and Avalonia. Several ostracod genera used Baltica as a staging-post in migrations between Avalonia and Laurentia. Migrations continued during the late Ashgill Hirnantian Stage (24 migrations), especially between Laurentia, Baltica and Avalonia (up to 19 migrations of genera), suggesting close geographical proximity for these palaeocontinents. Some ostracods, particularly the binodicopes Pseudulrichia,

* Corresponding author.

E-mail address: [email protected] (M. Williams).

0031-0182 / 03 / $ ^ see front matter B 2003 Published by Elsevier Science B.V. doi:10.1016/S0031-0182(03)00308-0

PALAEO 3079 9-5-03

194

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Klimphores, Kinnekullea, Aechmina and Spinigerites, could occupy outer-shelf and cooler-water benthic palaeoenvironments. They were part of a widespread deep-shelf fauna from the mid Caradoc onwards, for which distances or climatic barriers were less of an obstacle for trans-oceanic migration. None of these ostracods were bathyal. B 2003 Published by Elsevier Science B.V. Keywords: Ordovician; biodiversity; ostracods; palaeoenvironments; palaeobiogeography; palaeogeography

1. Introduction Ordovician ostracods had limited trans-oceanic dispersal capabilities. Their carapace morphology, facies distribution and faunal associates indicate that they occupied benthic neritic habitats, reaching their highest diversity in carbonate-depositing environments at mid to low latitudes (see Schallreuter and Siveter, 1985; Vannier et al., 1989). They rapidly disappeared where o¡-shelf and deeper-water (pelagic) facies encroached as a result of deepening in the marine environment (Williams et al., 2001b). There is no evidence for a deep marine, bathyal or abyssal plain Ordovician ostracod fauna (see Schallreuter and Siveter, 1985), nor is there any evidence for a pelagic mode of life in ostracods prior to the Silurian (Siveter et al., 1991). Moreover, many Ordovician benthic ostracods appear to have brooded their young, as indicated by carapaces which possess sexually dimorphic pouches that are assumed to have accommodated juveniles (e.g. see Lundin et al., 1995), and these features are particularly well represented in the dominant Ordovician group, the palaeocopes (see examples in Vannier et al., 1989). Therefore, like their modern benthic counterparts, Ordovician ostracods probably lacked a pelagic larval stage. Because of their mainly inner to mid shelf marine habitat and apparently limited dispersal capability, the patterns of Ordovician ostracod provinciality are important indices for determining the relative geographical position of ancient continents. In a broad sense, provincial faunas have been used to identify geographical isolation, whereas cosmopolitan faunas have been used to signal palaeogeographical proximity. For example, some authors have argued for the continued provinciality of Avalonian, Baltic and Laurentian ostracod faunas into the Silurian (e.g. McKerrow

and Cocks, 1976; Cocks and Fortey, 1982; McKerrow and Soper, 1989; McKerrow et al., 1991; Cocks et al., 1997; Van Staal et al., 1998; Cocks, 2000), using this as one line of evidence to support models which show a wide (1000 km or more) Iapetus Ocean separating Baltica and Avalonia from Laurentia during the Late Ordovician. Others have observed the increasingly cosmopolitan nature of the ostracod faunas from the Caradoc onwards (particularly Schallreuter and Siveter, 1985; Vannier et al., 1989; Williams et al., 2001a,b; but see also Kesling, 1960a,b; Copeland, 1973, 1989), suggesting, in accordance with certain plate tectonic reconstructions (e.g. Pickering and Smith, 1995), that the Iapetus Ocean may have been narrower at some point along its length by Late Ordovician times, thereby allowing ostracods to migrate between continents. This paper focusses on migration patterns between the ostracod faunas of the ‘North Atlantic’ region including Europe (palaeocontinental Baltica, Malopolska, Avalonia, Perunica and IberoArmorica) and North America (Laurentia). Where pertinent, it also refers to material from the Argentine Precordillera and from parts of the Gondwana supercontinent, particularly South America, Saudi Arabia and Australia. Ordovician ostracods are also known from Turkey (Sayar and Schallreuter, 1989), including typical widespread forms such as Klimphores and Piretella. They are also known from Siberia (see references in Olempska, 1994) and Kazakhstan (e.g. Melnikova, 1986). Olempska (1994) has noted rare palaeobiogeographical links with the latter two faunas and those of Europe, and further work should seek to analyse their relationships with the faunas documented herein. The data reviewed here are assembled from various published records. In some cases generic names have been up-dated (see the Appendix),

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

though the Bohemian faunas (e.g. see comments in Schallreuter and Krufita, 1988) and in particular the Laurentian faunas still require much primary taxonomic revision (see Williams, 1990). Not all identi¢ed migrants are included. For example, Karinutatia is recorded from the Baltic region and Bohemia (Vannier et al., 1989; Schallreuter and Krufita, 1988), but its stratigraphical range in the former is uncertain, as is the case for the genus Deefgella, which is also tentatively recorded from North America (Schallreuter and Siveter, 1985). Likewise, Parapyxion is recorded from Bohemia, but only very tentatively from the Arenig of Britain (Siveter, in press). Schallreuter and Siveter (1985, p. 586) also detailed numerous Late Ordovician genera in common between Laurentia and Baltica. Detailed primary analysis of these faunas (Copeland, 1970, 1973) and, for example, those from southern Ireland and northern England which are still unpublished (see Jones, 1987; Siveter, in press) will almost certainly identify further faunal links between Europe and North America. Despite these shortcomings, the dataset reviewed here, consisting of more than 100 migrant genera and over 150 migrations, is considered su⁄ciently robust to identify overall trends and patterns of migration in the ‘North Atlantic’ region. The aims of this paper are: b to discern the patterns of ostracod migrations at the generic level between Ibero-Armorica, Perunica, Avalonia, Baltica (including Malopolska) and Laurentia during the interval of the Arenig to Ashgill (Avalonia, Perunica and Ibero-Armorica were initially part of Gondwana, but rifted o¡ from that supercontinent during the Ordovician; see Cocks, 2000); b to determine the source area of migrating genera and determine their route(s) of migration ; b and to interpret the signi¢cance of these migrations in terms of palaeogeography, global sealevel changes, facies compatibility and climate.

2. The fossil record of Ordovician ostracods The fossil record of Ordovician ostracods is extensive, with several thousand species described.

195

Nevertheless, reconstructing the ranges and migration patterns of Ordovician ostracods is hampered by problems of homeomorphy and by spatial and temporal gaps in the fossil record, particularly regarding faunas representative of palaeocontinental Gondwana. 2.1. Problems of homeomorphy Contemporaneous species from di¡erent palaeocontinents which have been referred to the same genus, but which actually represent two or more di¡erent (sometimes unrelated) genera, can skew data being used to interpret palaeogeography (see Schallreuter, 1988a, p. 1041). Problems of homeomorphy in Ordovician ostracods have mainly arisen because of the over-emphasis of plesiomorphic or loosely de¢ned carapace structures in taxonomy. For example, early workers often emphasised a single character in their generic ‘diagnoses’, uniting many disparate species into ‘bag genera’. Thus, Primitia was based on the possession of a single sulcus; Aparchites was de¢ned on the absence of lobes or sulci; and Tetradella was recognised by possession of quadrilobate valves. The vast majority of these taxa have subsequently been referred to a range of di¡erent genera. Nevertheless, even apparently specialised structures, such as a loculate ventral structure, can occur in relatively unrelated palaeocopes (Schallreuter, 1988a). Since the 1930s Ordovician ostracod workers have more precisely de¢ned the morphological characters used to de¢ne ostracod taxa, emphasising a number of apomorphic carapace features in their diagnoses of genera. This has led to a more stable taxonomic usage of four ostracod orders, the Beyrichiocopa (suborders Palaeocopa, Binodicopa, Leiocopa), Platycopa, Podocopa (suborders Metacopa and Podocopa) and Leperditiocopa. For a summary of the characteristics and systematics of these higher taxa see Vannier et al. (1989) and Whatley et al. (1993). Numerically, palaeocopes and binodicopes are the dominant element of the Ordovician faunas (see Vannier et al., 1989, text-¢gures 20, 27). Morphological characteristics of the carapace used to distinguish palaeocopes include the histial,

PALAEO 3079 9-5-03

196

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Fig. 1.

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Fig. 1 (Continued, caption overleaf).

PALAEO 3079 9-5-03

197

198

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

velar and lobal structures: many of these structures are sexually dimorphic, di¡ering markedly between heteromorphs (females ?) and tecnomorphs (males?). Other features used for taxonomy include the structures associated with valve overlap (see Vannier et al., 1989, text-¢gures 6, 9). Binodicopes are also characterised by distinctive lobal and velar structures, but these are not dimorphic. Some podocopes, with their relatively simple morphology, can be more di⁄cult to characterise and identify; internal morphology, particularly the duplicature, stop-pegs and nature of valve overlap is often more diagnostic (see Hessland and Adamczak, 1974). As a result, some podocopes, described largely from external morphology, have been referred to ‘bag’ genera whose true taxonomic status is uncertain. These include, for example, the several smooth-shelled, ‘bean’shaped Ordovican species that have been referred to Bythocypris Brady (for which, see Keenan, 1951; Kraft, 1962; Copeland, 1970, 1973). All of the Ordovician species referred to this genus are probably homeomorphs, as Bythocypris is primarily based on Recent material (see also discussion of Elliptocyprites below). Many of the genera used in this analysis were employed in the palaeogeographical syntheses presented by Schallreuter and Siveter (1985) and Vannier et al. (1989). Sources used for constructing the ranges of genera depicted in Fig. 1 (see below) are indicated in the Appendix. In particular, Vannier et al. (1989) was used as a prime

source for reconstructing the ranges of Baltic and Ibero-Armorican taxa. Where there is doubt about the assignation of material to a particular genus, a questionable status is used. 2.2. Completeness of the fossil record Gaps in the fossil record of Ordovician ostracods will appear as a lower number of records of genera for any particular time period, and could result in erroneous interpretations of possible migration patterns. Because of this, the completeness of the Ordovician ostracod fossil record for each palaeocontinent is assessed in this section (see also Vannier et al., 1989). 2.2.1. Laurentia This palaeocontinent encompassed much of modern North America with the exception of parts of the eastern seaboard of the USA and Canada (see Fig. 2). In terms of its palaeobiogeographical signature it also included Scotland and Ireland north of the Iapetus Suture (see Pickering and Smith, 1995; Cocks, 2000; Armstrong and Owen, 2001). During the Ordovician Laurentia straddled the equator, its southern and northern margins extending into the sub-tropics (Cocks, 2000, ¢gure 9). In Laurentia there are no con¢rmed Tremadoc ostracod faunas. Instead, the palaeoecological niches of ostracods were occupied by bradoriids (see Siveter and Williams, 1997), a group origi-

Fig. 1. Stratigraphical ranges and palaeogeographical routes of migration for 107 Ordovician ostracod genera. Ranges for each genus are compiled from the reference list, with key sources indicated in the Appendix. Range data are plotted against the British Ordovician series and graptolite biozones (for which, see Fortey et al., 1995, 2000). Correlation of the Ordovician sequences of Britain, North America, the Baltic region, Bohemia (Czech Republic), Ibero-Armorica and parts of palaeocontinental Gondwana are based on Barnes et al. (1981), Ross et al. (1982), Webby and Packham (1982), Vannier et al. (1989, text-¢gure 2), Olempska (1994, ¢gure 1), Leslie and Bergstro«m (1995), Meidla (1996), Webby (1998) and Fortey et al. (2000, ¢gure 34). Some Laurentian ostracods depicted as arising in the early multidens Biozone might actually have arisen a little earlier in the late gracilis Biozone. This is because of problems of correlation between the gracilis Biozone as distinguished in North America and Britain (see Finney and Bergstro«m, 1986). Key to symbols adjacent the ostracod ranges: A, Avalonia; B, Baltica; M, Malopolska; L, Laurentia; P, Perunica; G, Ibero-Armorica; GP, Argentine Precordillera; GQ, Gondwana (Australia); GR, Gondwana (Saudi Arabia). Boxed letters indicate that the stratigraphical range of a taxon is imprecise. A double-box indicates even greater imprecision. A dotted line indicates an uncertain range. Ostracod group symbols are: P, palaeocope; B, binodicope; Pd, podocope; O, other (eridostracan, leiocope or unde¢ned). The position of each generic occurrence within the graptolite biozonal scheme is approximate. Schmidtella is recorded from Malopolska and Russia (see Olempska, 1994), but the a⁄nities of this material are uncertain until the type North American material is redescribed (see, for example, problems highlighted by Lundin et al., 1995). Baltonotella* includes material referred to Brevidorsa.

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

199

Fig. 2. Main migration routes for Ordovician ostracods of the ‘North Atlantic’ region plotted onto a Caradoc reconstruction of global palaeogeography (modi¢ed from Bassett and Popov, unpublished; see also Popov et al., 1999, ¢gure 2). The ‘ostracod icon’ denotes the palaeobiogeographical regions evaluated in this study. Ostracods are also known from Siberia, but are not analysed here. Thick arrows denote migrations of ten or more taxa during the Llanvirn to Ashgill. The numbers adjacent the arrows refer to the following: 1, peaks of migration during the early to mid Caradoc and late Ashgill; 2, peak of migration during the late Ashgill; 3^5, peaks of migration during the late Llanvirn; 6, peaks of migration during the mid Caradoc and mid Ashgill; 7, peak in£ux of taxa from Baltica, Laurentia and Ibero-Armorica during the early Caradoc.

nally thought to be ostracods, but whose taxonomic status as crustaceans is now in doubt (Hou et al., 1996). The earliest con¢rmed Laurentian ostracods are from Arenig and Llanvirn rocks, described in detail only from a few areas in Utah (e.g. Berdan, 1988 and references therein), Oklahoma (e.g. Harris, 1957 ; see Derby et al., 1991 for summary) and Canada (e.g. see Copeland, 1982). These ostracod assemblages are of low to moderate diversity, with between 6 and 15 genera (e.g. Joins, Oil Creek and Mclish formations of Oklahoma ; Kanosh Shale and Lehman formations of Utah; Sunblood Formation of northwest Canada). More richly diverse ostracod faunas are known from Caradoc sequences equivalent in age to the British gracilis and multidens graptolite biozones. The faunas of individual Caradoc formations can comprise up to 50 genera

and over 80 species. In the United States they include the ostracod faunas of Minnesota and Iowa (Kay, 1934, 1940; Swain, 1987), Virginia (Kraft, 1962), Oklahoma (Harris, 1957; Levinson, 1961; Williams and Siveter, 1996), Kentucky (Warshauer and Berdan, 1982), New York and Vermont (e.g. Swain, 1957, 1962), Pennsylvania (e.g. Swain, 1957, 1962) and Michigan (e.g. Kesling, 1960a,b). Laurentian ostracods of early Caradoc age are also known from SW Scotland (Williams and Floyd, 2000; Williams et al., 2001a) and are extensively documented from Canada (Copeland, 1965, 1971, 1977a,b, 1982, 2000). Younger, late Caradoc and early Ashgill Laurentian ostracod faunas are less well documented, but include those of the Maquoketa and Eden Shales of Missouri and Ohio (Spivey, 1939; Keenan, 1951). These latter faunas are moderately diverse

PALAEO 3079 9-5-03

200

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

(typically yielding ca 15 genera and ca 28 species ; see Keenan, 1951), but of lower diversity than the earlier Caradoc faunas. Middle and upper Ashgill faunas are mainly documented from Canada, and include the moderately diverse faunas of the Ellis Bay Formation (30 genera) and Vaureal Formation (23 genera; see Copeland, 1970, 1973) of eastern Canada, and the Ordovician part of the Whittaker Formation of NW Canada (18 genera; see Copeland, 1989). Laurentian Ordovician ostracods are also known from Greenland (e.g. Teichert, 1937a,b). 2.2.2. Baltica This palaeocontinent encompassed most of Scandinavia and parts of northern Europe eastwards to the Ural Mountains (see Pickering and Smith, 1995; Cocks, 2000 ; see Fig. 2). During the Early Ordovician, Baltica lay at latitudes greater than 30‡S, but during the Caradoc and Ashgill it moved northwards into the southern sub-tropics (see Cocks, 2000). Baltic Ordovician ostracods are widely monographed and appear to be amongst the most richly diverse (see Hessland, 1949; Henningsmoen, 1953; Jaanusson, 1957; Sarv, 1959; Schallreuter, 1993c; Meidla, 1996 and references therein), with well over 800 species described (see Schallreuter, 1988c). The Arenig^Ashgill fossil record of Baltic ostracods is extensive, with hardly any gaps (see Vannier et al., 1989, p. 167, text-¢gure 2). Baltic Tremadoc faunas may include the oldest documented ostracod (Tinn and Meidla, in press), and diverse ostracod faunas are already present in strata of mid and late Arenig age (e.g. Hessland, 1949; Jaanusson, 1957; Sarv, 1959). For example, the Komstad Limestone of Denmark (Volkov Stage, about equivalent to the gibberulus graptolite Biozone of the mid-Arenig in Britain) contains a fauna of about 11 ostracod species including Euprimites and Conchoprimitia (Tinn and Meidla, 1999). The middle Arenig assemblages of the Lanna and Holen limestones in Sweden (Volkhov and Kunda stages of the Baltic, equivalent to the gibberulus and artus graptolite biozones of the British sequence) yield a fauna of 25 ostracod species (Tinn and Meidla, 2001). This contrasts with the much lower diversity (or more poorly

known) Arenig ostracod faunas of Gondwana (including, at this time, Ibero-Armorica and Avalonia ; see below). Many Baltic Ordovician ostracods are also known from ‘Geschieben’ (erratic boulders) of Scandinavian origin which have been deposited in north Germany. Schallreuter has documented these faunas in immense detail (e.g. Schallreuter, 1975, 1984, 1993c ; see also Meidla, 1996 and references to Schallreuter therein). Baltic faunas may have reached their peak of diversity during the late Llanvirn and early Caradoc (Vannier et al., 1989, text-¢gure 27c), but diversity remained high during the Caradoc and Ashgill (e.g. Henningsmoen, 1954), only diminishing rapidly as the late Ashgill Hirnantian glaciation took hold (Meidla, 1996, pp. 11, 193). That Ashgill ostracod diversity was higher in Baltica than suggested by Vannier et al. (1989) is indicated by Meidla’s (1996) description of some 116 genera and 215 species from the Ashgill of º jlemyr£int Formation Estonia. Similarly, the O (Ashgill) has yielded more than 150 species (see Schallreuter, 1984). The Malopolska Massif of Poland is regarded as a separate palaeocontinental area situated very close to Baltica, but perhaps separated by a narrow, deep (oceanic?) seaway of some 350 km (Dzik and Pisera, 1994). Its Ordovician ostracod fauna is very similar to that of Baltica, sharing 42 out of 51 genera (Olempska, 1994, ¢gure 2) and suggesting that ostracods could migrate relatively freely between these two areas. Because most genera are common to these areas they are treated as a single entity in Figs. 1 and 3. Faunas from the Malopolska Massif (Mo¤jcza Limestone) include over 80 species from the Arenig, Llanvirn and Caradoc, though Ashgill faunas are more poorly represented (Olempska, 1994). 2.2.3. Avalonia This small palaeocontinent (sometimes referred to as a micro-continent) comprised Britain and Ireland south of the Iapetus Suture, Belgium, north Germany and parts of the eastern seaboard of North America (see Pickering and Smith, 1995; Cocks, 2000 ; see Fig. 2). Until the Arenig, Avalonia lay on the margin of Gondwana at high

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

southern latitudes, rifting o¡ during the late Arenig or early Llanvirn and drifting northwards towards Baltica and Laurentia. Avalonian ostracod faunas are known largely from Wales and the Welsh Borderland (Siveter, 1978, in press ; Jones, 1986, 1987), northern England (Williams et al., 2001b), the Irish Republic (Orr, 1987; see also Siveter, in press) and Belgium (Schallreuter et al., 2000). Tremadoc faunas are composed entirely of bradoriids (Williams and Siveter, 1998), except for a possible Rivillina (binodicope ostracod) in the late Tremadoc (salopiensis Biozone of British usage), which may be one of the oldest recorded ostracods. Arenig faunas comprise only three genera, ‘Parapyxion’, ‘Conchoprimitia’ and ‘Leperditia’. Faunas of early Llanvirn age are also poorly diverse, comprising only two undescribed palaeocope species (Siveter, in press). There was a massive increase in diversity during the late Llanvirn (teretiusculus graptolite Biozone), with about 20 genera and more than 25 species recorded in Wales and the Welsh Borderland (Jones, 1986, 1987; Siveter, in press). Ostracod diversity peaked during the early and mid Caradoc (gracilis and multidens graptolite biozones), with, for example, more than 20 genera and some 35 species occurring at horizons recognised as being of gracilis Biozone age (Siveter, in press). In contrast, younger Caradoc ostracod faunas are lower diversity (about three species ; see Jones, 1986, 1987), though undescribed faunas from the Dufton Shale Formation in northern England (see Rushton in Burgess and Holliday, 1979 ; Rushton, in Arthurton and Wadge, 1981) suggest that overall diversity for this interval is probably somewhat higher than recorded (cf. Vannier et al., 1989, text-¢gure 27b). At least 19 genera and more than 30 species are known from the Cautleyan and Rawtheyan (Ashgill) of northern England (Williams et al., 2001b), whilst the similarly aged Portrane Limestone in southern Ireland contains over 100 species (Orr, 1987; see Siveter, in press). The latter fauna is not published and its genera are not incorporated into Figs. 1 and 3. Abundant, but largely unrevised ostracod faunas are also known from the English Lake District, including species of Eoaquapulex and Euprimites (see Jones, 1987; Siveter, in press). As a

201

result, the abundance and diversity of ostracods from the Ashgill of Avalonia is almost certainly much higher than depicted by Vannier et al. (1989, text-¢gure 27b), and there may be yet more substantial links between Avalonian, Baltic and Laurentian Ashgill faunas than are depicted in Figs. 1 and 3. 2.2.4. Gondwana During the Ordovician, Gondwana was a supercontinent encompassing most of South America, Antarctica, Africa, India, Australia, and much of the Middle and Far East (Fig. 2 ; see Barnes et al., 1995; Pickering and Smith, 1995). The southern polar region was situated over North Africa (Pickering and Smith, 1995), but parts of Gondwana, including Australia, extended into the tropics (Barnes et al., 1995, ¢gures 4^6). For the Early Ordovician, Ibero-Armorica and Perunica were also part of Gondwana, but then rifted o¡ to drift northwards towards Baltica and Avalonia, and were probably approaching these palaeocontinents by the Ashgill (see Cocks, 2000, ¢gure 9c). Gondwanan Ordovician ostracod faunas are known mainly from Ibero-Armorica (Vannier, 1983, 1986a,b, 1987 and references therein), which encompassed north and central France, the Montagne Noire of southern France, most of the Iberian Peninsula and Sardinia (see Cocks, 2000). Ibero-Armorica may have comprised two small micro-continents (Iberia and Armorica), though its Ordovician ostracod faunas are almost entirely uniform (Vannier et al., 1989, p. 167), and herein it is treated as a single area. A few Gondwanan ostracods are also known from the Hanadir Formation (Llanvirn) of Saudi Arabia (Vannier and Vaslet, 1987), whilst more extensive faunas are documented from the Caradoc of Australia, for example in the Cliefden Caves Limestone Group (Gleesons and Wyoming Limestone members) (Schallreuter, 1988b and references therein) and Daylesford Limestone (Schallreuter and Siveter, 1988a,b) of New South Wales. There is one upper Tremadoc species from the lower Emanuel Formation at Prices Creek, northern Western Australia (Schallreuter, 1993a) and two Arenig species from the upper Emanuel Formation (Schallreuter,

PALAEO 3079 9-5-03

202

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

1993b) and from the Horn Valley Siltstone (Jones and Schallreuter, 1990). South American ostracod faunas are poorly known, with a few Ordovician ostracods reported from Bolivia (Prflibyl, 1984), and more extensive faunas documented from Argentina. In Argentina most of the ostracod data comes from the Precordillera, which was an allochthonous terrane (Benedetto, 1993; Astini et al., 1995; Thomas and Astini, 1996; Benedetto et al., 1999) that rifted o¡ from Laurentia in the early Cambrian and drifted, during the Ordovician, towards the western margin of Gondwana. By the Ashgill the fauna was indistinguishable from those of autochthonous Gondwanan basins (Benedetto et al., 1999). The Precordillera Ordovician ostracod fauna had a high degree of endemism and a mixture of Laurentian and Baltic forms, and only later in the Ordovician did it exhibit Gondwanan a⁄nities. Very few species are known from the Arenig (Salas, 2001, 2002a), there is an increase in diversity during the early Llanvirn and it reached a peak during the early Caradoc (Salas, 2001, 2002a,b, in press; Schallreuter, 1996, 1999 and references therein). Ashgill faunas are less well known, only three genera are so far reported (Harpabollia, Ansipe and Spinodiphores ; see Schallreuter, 1995a,b,c). Apart from the Precordillera faunas, three ostracod genera are also known from the early Arenig of the Northwestern basin (Cordillera Oriental), an area which was part of Gondwana proper (Rossi de Garc|¤a and Proserpio, 1976). These taxa need modern revision. As is the case in Laurentia, Avalonia and Baltica, ostracod diversity reached an early peak in Ibero-Armorica during the late Llanvirn and early Caradoc (Vannier et al., 1989, p. 192) and in the Argentine Precordillera during the early Caradoc (Salas, 2001), but there is too little published information to con¢rm these trends in other parts of Gondwana. There is a particular dearth of information regarding faunas of Ashgill age both from Ibero-Armorica and Gondwana as a whole, though some Upper Ordovician Gondwanan faunas may be preserved in glacial dropstones in the Ordovician sequence of Thuringia, Germany (see Schallreuter and Hinz-Schallreuter, 1998). Ostra-

cods of Ashgill age occur in the carbonates of the Rosan Formation in the Armorican Massif (Vannier et al., 1989), but are undescribed. This assemblage may yield additional evidence about the relative position of Ibero-Armorica and Baltica during Late Ordovician times. Ostracods have also been described from the Arenig to Ashgill of Bohemia in the Czech Republic (e.g. Schallreuter et al., 1996; Schallreuter and Krufita, 1984, 1988, 1994, 2000a,b, and references therein). This area is regarded as part of the Ordovician Perunica micro-continent, which rifted o¡ Gondwana and drifted northwards towards Baltica (see Cocks, 2000; see Fig. 2). The Bohemian faunas include many widespread ostracod genera, but particularly taxa that also occur in Baltica. The most recently revised Bohemian faunas are those of the late Llanvirn interval (Dobrotiva¤ Formation; Schallreuter and Krufita, 2000b) yielding forms which typically originated in Baltica, and those of the early Caradoc (Letna¤ Formation; Schallreuter and Krufita, 2000a) which include forms which originated in both Avalonia and Baltica.

3. Habitats and mode of life of Ordovician ostracods The factors which contribute most to the utility of Ordovician ostracods as indicators of palaeogeography are their limited environmental preferences (entirely neritic distribution), particularly their benthic lifestyles and lack of a pelagic larval stage (see Schallreuter and Siveter, 1985). The palaeoenvironmental distribution of Ordovician ostracods, their faunal associates and their mode of life are reviewed here. 3.1. Palaeoenvironmental distribution There is no evidence for a deep-marine benthic ostracod fauna in the Ordovician. Bathyal and abyssal plain benthic ostracods appear to have evolved only much later, possibly during the Tertiary (see Schallreuter and Siveter, 1985). Although their preservation potential in sediments deposited below the Calcium Carbonate Compen-

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

sation Depth in the oceans was probably slight, the complete absence of ostracods from deepwater lithofacies such as the Hartfell Shales Formation (Caradoc) of the Southern Uplands of Scotland (Williams et al., 2001a) appears to con¢rm this. The majority of Ordovician ostracods were neritic. A few taxa specialised in very shallow marine habitats (such as the leperditicopes; see Berdan, 1976, 1984; Vannier et al., 2001). Others preferred deeper shelf habitats (see Copeland, 1982; Williams et al., 2001b). In these shelf marine settings faunas reached their highest diversity in tropical and subtropical warm water environments where mudstones and carbonates were deposited. Diversity is lower in very shallow or deeper shelf settings, or at higher latitudes (cooler water settings; see Vannier et al., 1989). Similar patterns of ostracod distribution are documented for Silurian ostracods (Siveter, 1984). Very shallow marine (peritidal, supratidal) Ordovician ostracod faunas, at least in Laurentia, are of low diversity. For example, in the Corbin Ranch Member of the Bromide Formation, Williams and Siveter (1996, table 1) identi¢ed only nine species (of a total fauna of more than 80 species for the whole formation ; Williams, 1990). Very shallow marine faunas tend to be dominated by only a few species of leperditicope or Leperditella (e.g. see Berdan, 1976, 1984; Williams and Siveter, 1996). By contrast, marine environments from the subtidal to the open shelf yield proli¢c, high-diversity faunas. These include most of the documented Laurentian Ordovician ostracod faunas (e.g. Fig. 4), which occupied warm tropical or sub-tropical seas. These faunas are palaeocope-dominated from the late Llanvirn onwards (e.g. Swain, 1987; Williams, 1990; Williams and Siveter, 1996), and binodicopes, eridostracans, podocopes and leiocopes are also well represented. They are of high diversity, with many individual stratigraphical formations yielding 50^80 species (e.g. see Kay, 1940; Harris, 1957; Kraft, 1962; Swain, 1987; Williams and Siveter, 1996). Typically, these faunas include palaeocopes such as Laccochilina, Hippula, Eurychilina, Bromidella, Eoaquapulex and Hithis, podocopes such as Balticella, Krausella and Monoceratella, leiocopes such as Baltonotella

203

and eridostracans such as Cryptophyllus and Eridoconcha. Some Ordovician ostracods had adapted to a wide range of mid to deep shelf marine settings. For example, the long-ranging (Arenig^Ashgill) leiocope Baltonotella is widespread from Gondwana to Laurentia in clastic and carbonate facies (see Vannier, 1990; Williams and Vannier, 1995). The binodicopes Pseudulrichia, Kinnekullea, Klimphores, Spinigerites and Aechmina have species which occur in shallow shelf and deeper shelf settings (e.g. Floyd et al., 1999; Williams et al., 2001a,b, Meidla, 1996; Copeland, 1989), as does the palaeocope Platybolbina (e.g. Kesling, 1960b; Copeland, 1989; Williams and Floyd, 2000) and the podocope Krausella (see Copeland, 1982). Some of these ostracods appear to have been opportunistic colonisers. For example, the Pen y Garnedd Phosphorite, which was probably deposited on a submarine rise in the Welsh Basin (Cave, 1965), yields species of Klimphores, Spinigerites and another widespread binodicope, Vogdesella (see Jones, 1987, p. 108), forming part of a ‘westward outpost’ of the Welsh Borderland shallow neritic fauna (Cave, 1965, p. 292). Di¡erences in the environmental ranges of these neritic ostracods can be used to distinguish depthzoned biofacies. For example, Copeland (1982) identi¢ed two ostracod biofacies in the early Caradoc Esbataottine Formation of NW Canada. His Biofacies 1 comprises ostracods which range from shallow marine (Bathyurus and Isotelus trilobite biofacies) to deeper shelf settings (Dimeropyge trilobite biofacies). These include Eoaquapulex, Leperditella and Hippula (Oecematobolbina of Copeland’s terminology), which range through the Bathyurus to Calyptaulax^Ceraurinella trilobite biofacies, and Krausella, which also occurs in the Dimeropyge trilobite biofacies. All of these genera are ostracod migrants (Fig. 1), though Copeland’s Biofacies 1 also includes endemic forms. His Biofacies 2 contains forms restricted to mid to deep shelf settings (equivalent to the Calyptaulux^ Ceraurinella and Dimeropyge trilobite biofacies). It includes Winchellatia (? = Collibolbina), Tetradella, Platyrhomboides, Cryptophyllus, Steuslo⁄na and Baltonotella (Aparchites of his terminology), all of which are migrant, or possible migrant gen-

PALAEO 3079 9-5-03

204

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

era (Fig. 1). This deeper shelf biofacies also includes several endemic taxa. Copeland (1982, p. 6) noted a decline in palaeocope species towards the deep shelf, which has also been recognised in Avalonian (e.g. Williams et al., 2001b) and Baltic (e.g. Meidla, 1996) faunas. This decline may relate to cooler water at depth, as binodicopes, more characteristic of high-latitude faunas (see Vannier et al., 1989), are commoner in these deeper shelf assemblages. Some Laurentian Ordovician ostracod faunas occupied very deep shelf settings, in clastic facies. These faunas are very low diversity (ca ¢ve species) and are characterised by binodicopes such as Aechmina and Pseudulrichia and palaeocopes such as Platybolbina (see Copeland, 1977b, 1989). They appear to have been living on the margins of deep water tolerance for Ordovician ostracods. Baltic ostracod faunas have many similarities to Laurentian faunas, at least from the Caradoc onwards (Figs. 3 and 4) and show similar trends in depth zonation (see Meidla, 1996, p. 201). Open marine shelf ostracod faunas are of very high diversity, and are palaeocope-rich (Vannier et al., 1989). Deeper shelf, clastic-dominated sequences have lower diversity and are palaeocope-poor. They are characterised by binodicopes such as Klimphores, Aechmina and Kinnekullea and podocopes such as Rectella. Such faunas are typi¢ed by the late Ordovician Klimphores minimus and Rectella composita associations (Oandu and Porkuni stages of Estonian nomenclature, respectively; Meidla, 1996, pp. 195, 200, ¢gure 47). Podocopes such as Steuslo⁄na cuneata appear to have been widespread in Baltic facies, occurring in reefal limestones and open marine shelf carbonate^mudstone deposits (Meidla, 1996, pp. 196^ 200). Steuslo⁄na cuneata is also present in the deep

205

shelf facies of the Cautley Mudstone Formation of northern England (Williams et al., 2001b). Most Avalonian faunas come from siltstones, sandstones and limestones which formed in a range of open marine shelly facies. They are most abundant in the late Llanvirn to early Caradoc, when the contrast between graptolitic and shelly facies was strongest (see Siveter, in press). Typical of these faunas are the genera Ceratopsis, Duringia, Homeokiesowia and Piretopsis (see Jones, 1986), all of which were migrants (Fig. 1). Clastic shelf sequences that accumulated at higher latitudes in cooler water settings, for example in Ibero-Armorica (Vannier, 1986a,b; Vannier et al., 1989), host a generally shallow marine ostracod-bearing biofacies (Vannier et al., 1989, p. 166). They typically have lower-diversity faunas and are characterised by the abundance of binodicopes such as Vogdesella, Laterophores, Klimphores, Satiellina, Rivillina and Eocytherella (Vannier, 1986a). Overall palaeocope diversity is much lower in these faunas (Vannier et al., 1989, text¢gure 27A), and this also appears to be characteristic for the Ordovician faunas from Argentina (Schallreuter, 1996, 1999; Salas, 2002a,b, in press). These faunas resemble the deeper shelf faunas of Laurentia and Baltica (e.g. Copeland, 1989). Nevertheless, even where shallow, warmwater carbonates were accumulating in Gondwana, palaeocopes still appear to have been rare. For example, the ostracods of the Gleesons and Wyoming limestones of New South Wales, Australia (Webby and Packham, 1982; Schallreuter, 1988b) yield only binodicopes and podocopes. This appears contrary to the model of Vannier et al. (1989), based largely on the Baltic sequence, in which carbonates are dominated by palaeocopes, though in New South Wales the absence

Fig. 3. Rates of migration of ostracods plotted for seven migration routes involving Avalonia, Baltica (including Malopolska), Ibero-Armorica, Perunica and Laurentia. The data are quanti¢ed from those given in Fig. 1. Data are plotted against the British graptolite biozones (for correlation see references given in Fig. 1). The numbers adjacent to the plots are the total migrations per biozone, and include total possible migrations (for which see Fig. 1). In the two right-hand columns, the total number of migrations per graptolite biozone (which includes data in addition to the seven migration routes depicted in this ¢gure) are plotted against a simpli¢ed global sea level curve for the Ordovician (based on Cocks and Fortey, 1982; Fortey, 1984): this total migration plot does not include those dispersals of uncertain horizon indicated by the double-box in Fig. 1. Because of the questionable origin of certain genera, some of the calculations should be considered approximate.

PALAEO 3079 9-5-03

206

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

of palaeocopes might relate to the isolated o¡shore island or carbonate platform setting of the limestones which surmount a thick pile of volcanic rocks.

The Argentine Precordillera faunas have lower diversity than Baltica or Laurentia and are characterised by the scarcity of palaeocopes and the abundance of binodicopes and podocopes. The

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

fauna, in the early Llanvirn, is dominated by binodicopes which occur in a distal carbonate ramp setting. Accordingly, the main control on the distribution of the group appears to have been water depth. In the early Caradoc (Las Aguaditas Formation) the faunas were dominated by podocopes, occurring in a carbonate margin to foreslope environment (carbonate slope-apron) (Astini, 1995). Ordovician ostracods have not previously been recorded from such deep marine environments, and on this basis it is possible to suggest that podocope ostracods displayed the greatest tolerance of deep water during the Ordovician (see also Williams et al., 2001b). 3.2. Associated fauna Ordovician ostracods typically co-occur with a rich benthos of brachiopods, trilobites, echinoderms, bryozoan and other marine invertebrates. Ordovician ostracod biocoenoses are not known from pelagic faunal associations. There are no close associations of ostracods with graptolite colonies or algae, such as are occasionally documented for pseudoplanktonic Ordovician brachiopods (e.g. Botting and Thomas, 1999). The ostracods of the Upper Ordovician (Caradoc) Ardwell Subgroup of the Girvan district, Scotland, are sometimes associated with graptolites, as decalci¢ed mould faunas (Williams et al., 2001a ; Williams and Floyd, unpublished information), but the ostracods appear to be transported benthos from shallower marine settings. In sequences where pelagic biofacies replace benthic biofacies, for example in the Cautley Mudstone Formation of northern England (Williams et al., 2001b), or the Viola Group succeeding the Bro-

207

mide Formation in Oklahoma (see Williams and Siveter, 1996; Finney, 1986), ostracods rapidly disappear. Ordovician ostracods are not reported from deep water associations with olenid trilobites which may have lived in anoxic environments (see Fortey, 1999). 3.3. Ostracod autecology Without evidence from soft parts, particularly knowledge of the appendages used for swimming, interpreting the life strategies of Ordovician ostracods remains speculative. Nevertheless, the carapaces of most palaeocope, binodicope and podocope ostracods suggest a benthic or nektobenthic mode of life. These are ‘thick-shelled’ forms which di¡er radically from the ‘thin-shelled’ pelagic myodocopids of the Silurian (see Siveter et al., 1991). Using morphological criteria detailed by Siveter (1984 and references therein), the podocope and straight-hinged binodicope and palaeocope Ordovician ostracods have carapaces that suggest mainly benthic modes of life. Typical palaeocopes (see Schallreuter and Siveter, 1985; see also Siveter, 1984), such as Platybolbina, possess several of the following carapace features : a mineralised, rigid (relatively thick) carapace; strong lobes, the lobes occupying much of the lateral surface of the carapace; a ‘tetrahedral’ shape and/or a ventrally situated (i.e. ‘low’) centre of gravity; a broad £attened ventral margin, often with well-developed adventral structures (to provide positional stability on the sea £oor); and strongly overlapping valves. Many binodicopes, for example Aechmina, Pseudulrichia, Warthinia, Klimphores and Kinnekullea, possess convex dome-like valves, with a

Fig. 4. Assessment of degree of provinciality of ostracod faunas for selected Laurentian and Avalonian stratigraphical formations of Llanvirn to Ashgill age. For Laurentia, ostracod data are from: Harris (1957; revised by Williams, 1990) for the Oil Creek, Mclish, Tulip Creek and Bromide formations of Oklahoma; Swain (1987) for the Decorah Formation of Minnesota; Berdan (1988) for the Lehman Formation of Utah; Kraft (1962) for the Edinburg and Lincolnshire formations of Virginia; Williams and Floyd (2000 and unpublished data in British Geological Survey collections) for the Craighead Limestone Formation of SW Scotland; Copeland (1970, 1973) for the Vaureal and Ellis Bay formations of Canada (see also Schallreuter and Siveter, 1985); and Floyd et al. (1999) for the Lady Burn Formation of SW Scotland. For Avalonia the data are from Siveter (in press) for the Llanvirn and Caradoc, and from Williams et al. (2001b) for the Cautley Mudstone Formation. Some of the calculations include unpublished revisions of taxa (e.g. Siveter, in press). For some of the Ashgill age formations, ongoing revision of faunas will almost certainly identify greater faunal links than are depicted, especially for the Vaureal and Ellis Bay formations of Canada (see Schallreuter and Siveter, 1985, for further information).

PALAEO 3079 9-5-03

208

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

tapering ventral margin and limited valve overlap. The absence of a broad ventral margin and £ange-like adventral structure suggests that the carapaces of these ostracods might have been relatively unstable on the sea £oor, though this could have been compensated by the valves gaping at 20‡ (a typical ¢gure for gape according to Vannier and Abe, 1992). In the case of Aechmina species, the elongate, hollow anterodorsal spine of both valves might add to this instability, shifting the centre of gravity of the carapace dorsal-wards, at least at the anterior end of the carapace (see Vannier, 1986a, text-¢gure 51; Williams et al., 2001b). This dorsal-wards shift in the centre of gravity could aid positional stability if the ostracod adopted a mainly swimming mode of life, particularly if the main mass of the animal was to the fore of the carapace. The spines might also have provided a stabilising (keel-like) structure whilst the ostracod was swimming, or may have acted as a £otation device, especially if ¢lled with gas (see Siveter, 1984). The dorsally situated lobes of Kinnekullea, Klimphores, Warthinia and Pseudulrichia might have functioned similarly. Although these features suggest that some binodicope carapaces were designed for a mode of life that may have involved swimming, this does not mean that binodicopes were pelagic. Their shells have similar thicknesses to that of podocopes and palaeocopes and are certainly more heavily mineralised than those of pelagic myodocopids or entomozoaceans. This suggests that an active, but benthic mode of life may have been most likely for many of the binodicopes (see Williams et al., 2001b). Vannier et al. (1989, p. 210) have noted the apparent wider dispersal capacities of binodicopes, and, in part, this may be consistent with a more active mode of life. Many Early Palaeozoic podocopes are non-lobate, but their elongate and mineralised rigid carapaces with strong ventral valve overlap, may be adapted for a benthic mode of life. Similarly shaped modern podocopid ostracods are also benthic (see Whatley and Jones, 1999). The common occurrence of this type of carapace design in mud-rich palaeoenvironments suggests that some may have adopted an infaunal, burrowing existence (see Siveter, 1984 and references therein),

or they may have been more tolerant of turbid water. Athough some Ordovician leiocopes (e.g. Baltonotella) are very widely distributed, and may have had good swimming capability (see Vannier, 1990), they invariably occur as part of diverse benthic ostracod faunas and none are known from o¡-shelf or deep marine (i.e. pelagic) biofacies. Eridostracans (e.g. Eridoconcha, Cryptophyllus), characterised by moult retention, had very heavily mineralised carapaces (e.g. see Williams and Jones, 1990), and are unlikely contenders for anything but a benthic mode of life. Carapace morphology also indicates that many Ordovician ostracods were brooding young. This is particularly prevalent amongst the numerically dominant migrating group, the palaeocopes. Some of the most successful palaeocope migrants such as Platybolbina, Hippula, Eoaquapulex and Euprimites, have very well developed brood pouches (the dolonal antrum), suggesting that this behaviour, at least in some ostracods, was not detrimental to wide geographical dispersal.

4. Patterns of migration for Ordovician ostracods The stratigraphical distribution of those Ordovician ostracod genera which show migration between two or more palaeocontinents is given in Fig. 1. This chart shows the major taxonomic group to which each genus is assigned, the earliest known place of origin of each genus (in the areas considered), and the timing of their earliest known appearance in other palaeocontinents. All of the genera listed are considered to be benthic and neritic (see discussion above). 4.1. Main migrating groups For the entire Ordovician, the greatest number of genera that migrate are palaeocopes (44) and binodicopes (31), followed by podocopes (22). Eridostracans and others (genera whose higher taxonomic position is unresolved or contentious) comprise only a small number of migrant genera. The numbers of genera migrating in each group re£ects the overall diversity of Ordovician ostra-

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

cod taxonomic groups. For example, in Baltica and Avalonia the greatest percentage of genera are palaeocopes (more than 50%), followed by binodicopes (see Vannier et al., 1989, text-¢gures 19, 20, 27). Palaeocope genera also represent the dominant element of many Laurentian faunas, particularly during early Caradoc times. Thus, in one of the richest ostracod-bearing sequences of North America, the Simpson Group of Oklahoma, palaeocopes account for more than 50% of the total genera (Williams, 1990). Only in Ibero-Armorica, the Argentine Precordillera and parts of Gondwana were binodicopes and podocopes more dominant overall than palaeocopes, and only in the interval following the late Llanvirn (Vannier et al., 1989, text-¢gure 27A; see also Schallreuter, 1988b, 1996, 1999; Salas, 2002a,b). There is an overall decrease in the abundance of palaeocopes during the late Caradoc and Ashgill in Avalonia and Baltica (Vannier et al., 1989, text-¢gure 27B,C; Meidla, 1996, p. 15). This reduction in palaeocope-diversity is particularly pronounced in Ibero-Armorica, where binodicopes become increasingly common (Vannier et al., 1989). However, binodicopes and podocopes also increase in diversity in Baltica as palaeocope diversity drops (Meidla, 1996, p. 15, ¢gures 3, 4). This is re£ected in the numbers of migrant palaeocope genera, which during the Caradoc and Ashgill account for less than 20 of the newly arising migrating taxa (Fig. 1), compared with more than 30 palaeocope migrants for the Arenig and Llanvirn. By comparison, the overall numbers of binodicope and podocope migrant genera newly arising in the Caradoc and Ashgill may be as many as 31 (see Fig. 1). 4.2. Rates of migration Rates of migration are calibrated with the British graptolite biozones. There are currently ten biozones for the interval of the Llanvirn to Ashgill, a period of time representing some 27 million years (see Gradstein and Ogg, 1996). For the purpose of this study each graptolite biozone is assumed to have lasted between 2 and 3 million years. Nevertheless, some of the Ashgill graptolite

209

biozones may have had a shorter duration than those of the earlier Ordovician. The majority of ostracod migrant genera dispersed to only two or three palaeocontinents (Fig. 1). Some, particularly Vannieria, Klimphores, Euprimites, Duringia, Pseudulrichia, Medianella, Vogdesella and Platybolbina dispersed to four or more palaeocontinents. Dispersal of some of these genera may have been favoured by their widespread deep shelf distribution (see above), but also by the fact that several of these widespread genera are also long-lived. Thus, from the beginning of the Caradoc, very few newly arising (i.e. shorter-ranging) genera migrated to more than two palaeocontinents. Graptolites, particularly mesopelagic forms such as Nemagraptus gracilis, show very rapid dispersal between Avalonia, Baltica and Laurentia, and are vital index species for correlation of Caradoc sequences worldwide (see Bettley et al., 2001). By comparison, the rate of dispersal of ostracods appears to have been much slower, generally encompassing the duration of one or more graptolite biozones (Fig. 1). This slower rate of migration is one further line of evidence to indicate that there were no pelagic ostracods in the Ordovician. Including those migrations that could have taken place between Avalonia, Perunica and IberoArmorica with Gondwana prior to their rifting from that supercontinent, amongst the 107 migrating ostracod genera over 70 migrations occurred within an interval about equivalent to the duration of one graptolite biozone or less (Fig. 1, calculation not including those occurrences shown in double-boxes). For ostracods, all of these may be regarded as ‘rapid’ migrations. The great majority of these ‘rapid’ migrations occurred in the late Llanvirn and Caradoc (about 54 dispersals). At least 14 ‘rapid’ migrations occurred in the late Llanvirn (teretiusculus Biozone, Fig. 1). Many of these involved dispersals between Avalonia, Perunica and Ibero-Armorica (e.g. Parinconchoprimita, Jeanlouisiella, Reuentalina, Duringia, Quadritia, Klimphores, Euprimites, Conchoprimitia), and may represent taxa that were already widespread in Gondwana before these micro-continents rifted o¡. However, several of these late

PALAEO 3079 9-5-03

210

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Llanvirn migrations, particularly those of Piretia, Piretella and Levisulculus (Fig. 1), might suggest that Perunica was approaching Baltica (and Malopolska). Several ‘rapid’ migrations occurred between Laurentia and Baltica. They include early migrants such as Bromidella, but the majority took place during the early and mid Caradoc, including those of the palaeocopes Platybolbina, Hippula, Tetradella and Hesperidella, the binodicopes Spinigerites, Pseudulrichia, Pedomphalella and possibly Easchmidtella, and the podocope Balticella (Fig. 1). There were fewer ‘rapid’ migrations between Avalonia and Baltica. These involved the migration of Homeokiesowia from Avalonia to Baltica in the late Llanvirn, the probably rapid migration of the Laurentian Eoaquapulex from Baltica to Avalonia in the Ashgill, the migration of Gotula from Avalonia to Baltica in the Ashgill, and possibly the migrations of Pseudohippula and Bulbosclerites in the late Ashgill (Fig. 1). Between Ibero-Armorica and Baltica there were few ‘rapid’ migrations, though Vittella and Primitiella might have migrated from Baltica to Ibero-Armorica during the early Llanvirn, when Ibero-Armorica may still have been part of Gondwana (Fig. 1). The most rapidly migrating ostracod of all was the binodicope Pseudulrichia. This arose during the late Llanvirn, probably in Avalonia (see Siveter, in press). Within the duration of three graptolite biozones it had migrated to Perunica, the Argentine Precordillera, Baltica, Ibero-Armorica and Laurentia (Fig. 1). Other genera also dispersed rapidly, but only to one or two palaeocontinents. These include the widespread, early^mid Caradoc Balticella of Laurentia (e.g. Harris, 1957; Kraft, 1962), which occurs contemporaneously in Baltica (Fig. 1). Similarly, Piretia, a longranging Baltic genus, occurs contemporaneously in Perunica during the late Llanvirn (Fig. 1). The ¢rst migrations of many genera, such as Euprimites, Kinnekullea and Eoaquapulex took place after a duration greater than one graptolite biozone, but some of their subsequent migrations were much more rapid (Fig. 1). For example, Euprimites originated in Baltica during the Arenig and migrated to Ibero-Armorica and Perunica

during the late Llanvirn. It subsequently migrated to Avalonia and thence rapidly to Laurentia during the complanatus and anceps biozones of the Ashgill (see Fig. 1). In marked contrast to the ‘rapid’ ostracod migrants, a few ostracod genera only migrated after the duration of ¢ve or more graptolite biozones (at least four out of 107 ¢rst migrations). In these cases, the longevity of some taxa perhaps contributed to their ultimate dispersal. For example, Uhakiella, a typical Baltic genus which arose in the Arenig, migrated to Avalonia only in the mid Ashgill (Fig. 1). ‘Slow’ migrants also include the podocopes Eographiodactylus and Monoceratella, both of which originate in the early Caradoc of Laurentia but dispersed to Baltica only in the late Ashgill (Fig. 1). Cryptophyllus, a widespread eridostracan genus in Laurentian faunas from the late Arenig onwards (see Harris, 1957; Berdan, 1988; Copeland, 1982), migrated to the Argentine Precordillera during the early Caradoc (see Schallreuter, 1981, 1996, 1999), but only appeared in Baltica during the late Ashgill, more than 20 million years after its ¢rst occurrence. 4.3. Quantifying routes of migrations through time Llanvirn to Ashgill migrations of genera originating in Avalonia, Baltica (including Malopolska), Laurentia, Ibero-Armorica and Perunica are quanti¢ed in Fig. 3. Migrations involving parts of Gondwana depicted in Fig. 1 are not plotted because information about the total stratigraphical ranges of ostracod genera in Gondwana remains sparse. Nevertheless, this area may have acted as an intermediary in the dispersal of some ostracods. Prior to the Llanvirn there were relatively few ostracod migrations, though during the Arenig Conchoprimitia appears to have been common to Baltica (see Vannier et al., 1989) and Gondwana (Saudi Arabia ; Vannier and Vaslet, 1987), which included Avalonia at this time (see Siveter, in press). At about the same time Glossomorphites occurs in Perunica (Schallreuter and Krufita, 1988) and Baltica (Vannier et al., 1989). The presence of the binodicope Rivillina in the lower Llanvirn Lehman Formation of Utah, USA (Berdan,

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

1988) represents the earliest migration from Gondwana (Ibero-Armorica; see Vannier, 1983) or Avalonia (see Williams and Siveter, 1998) to Laurentia, whilst Baltonotella occurs in Ibero-Armorica at about the same level as its earliest occurrence in Laurentia (early Llanvirn; see Fig. 1). Subsequently, there were several migrations between Baltica and Ibero-Armorica during the early Llanvirn, involving Primitiella, Vittella and Medianella, whilst the migration of Laccochilina from Baltica to Laurentia also occurred at this time. During the late Arenig^early Llanvirn there were two migrations from Australia to the Argentine Precordillera, involving the binodicopes Eodominina and Pilla. Acanthoscapha and Ningullela are also recorded in the Argentine Precordillera from the late Arenig^early Llanvirn, migrating to Laurentia later, during the early Caradoc. The early peak of Ordovician ostracod migration occurred in the late Llanvirn (teretiusculus Biozone, ca 464 Ma; Fig. 3). Many of these migrations involve Baltic-origin taxa (Figs. 1^3). Over half of these migrants involved dispersals between Baltica, and Avalonia and Perunica. Most ostracod migrants followed a route from Baltica to Avalonia (up to eight migrants) and from Baltica to Perunica (up to six migrants). Migrants from Baltica to Avalonia include the palaeocopes Vittella and Piretopsis, the binodicopes Laterophores and Klimphores, and the podocope Medianella (Fig. 1). Following the opposite route, Homeokiesowia, probably originating in Avalonia, migrated to Baltica at this time. Baltic migrants to Perunica include Klimphores, and the palaeocopes Piretia, Laccochilina and Euprimites. Several genera were also migrating to Ibero-Armorica from Baltica during the late Llanvirn, including the podocope Miehlkella, and the palaeocopes Ogmoopsis, Euprimites and possibly Quadritia (Fig. 1). The number of migrations from Baltica to Avalonia, Perunica and Ibero-Armorica decreased during the early Caradoc, re£ecting an overall slight reduction in the rate of ostracod migration at this time (Fig. 3). However, migrations from Baltica to Laurentia were increasing at this time, leading to a peak during the gracilis to clingani biozones, with up to 18 Baltic-origin genera mi-

211

grating to Laurentia (Figs. 1^3). During this time many characteristic Baltic genera, particularly palaeocopes such as Hithis, Hippula, Platybolbina, Oepikella and Levisulculus, became widespread amongst the Laurentian faunas (see Fig. 4). Migrations of genera originating in Ibero-Armorica to Baltica and Avalonia, and vice versa, continued during the Caradoc but in much reduced numbers (Figs. 1 and 3). These migrations include the dispersal of Avalonian genera such as Ceratopsis and Pseudulrichia. Some of these dispersals might indicate the northwards drift of Ibero-Armorica towards Baltica and Avalonia (see Cocks, 2000), though both Pseudulrichia and Ceratopsis were highly mobile in the Caradoc, migrating to several palaeocontinents in a relatively short time (Fig. 1). Thereafter, only two migrations are recorded for genera arising in Ibero-Armorica, involving the binodicopes Satiellina and possibly Kinnekullea, which appear in Baltica during the Caradoc and Ashgill (see Vannier, 1986a; Vannier et al., 1989): further links may be identi¢ed when the ostracod fauna of the Rosan Formation (Ashgill) of Armorica is described. During the early Caradoc (gracilis Biozone) there were several migrations to the Argentine Precordillera. Ectoprimitioides and Cryptophyllus arrived from Laurentia, Pachydomelloides from Ibero-Armorica, Longiscula from Malopolska and Conchoprimitia, Klimphores, Baltonotella and Medianella from Avalonia, Baltica or Laurentia (Fig. 1). Moreover, Steuslo⁄na occurs at the same time in Laurentia and the Argentine Precordillera. Several genera arising in the Argentine Precordillera during the gracilis Biozone migrated to Baltica and Perunica during the Caradoc and Ashgill (Fig. 1). During the mid to late Ashgill, migrations between Baltica and Avalonia reached a second acme (Figs. 1^3). In contrast to the earlier (teretiusculus Biozone) Baltica to Avalonia migrations, this exchange of genera was more of a two-way process. Many typical Baltic genera migrate to Avalonia during the Cautleyan and Rawtheyan (anceps Biozone). These include the palaeocopes Distobolbina and Hippula, which had already migrated to Laurentia during the Caradoc, and the

PALAEO 3079 9-5-03

212

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

very long-ranging Baltic palaeocope Uhakiella : these migrations also included the podocope Duplicristatia (Fig. 1). The Baltica to Avalonia route was followed by migrations of Avalonian genera to Baltica during the early Hirnantian (Figs. 1 and 3), though the occurrence of some typically Baltic forms (e.g. Bulbosclerites, Pseudohippula) at an earlier period in Avalonia remains tentative (see Fig. 1 and Williams et al., 2001b). As a result of these migrations, Baltic and Avalonian ostracod faunas were cosmopolitan by the mid Ashgill, having many species and genera in common (see Williams et al., 2001b ; Fig. 4). Migrations of Baltic genera to Laurentia had an early peak during the early to mid Caradoc (see above). A second peak, during the late Ashgill, is contemporary with the peak of migrations of Laurentian genera to Baltica (Fig. 1). This twoway migration involved ‘rapid’ dispersals, for example of the palaeocopes Anticostiella and Caprabolbina, probably during less than one graptolite biozone after their ¢rst appearance, and the migration of much longer-ranging Laurentian genera such as the podocopes Eographiodactylus and Monoceratella (Fig. 1). A number of genera were common to Avalonia and Laurentia by the Late Ordovician, but relatively few ostracod migrations occurred directly between these palaeocontinents, possibly only Eridoconcha and Warthinia (Fig. 1). Ceratopsis may also have migrated directly from Avalonia to Laurentia, though it also occurs in Perunica and Ibero-Armorica (Fig. 1). Several genera common to Laurentia and Avalonia appear to have used Baltica as a ‘staging-post’ for migrations from the late Llanvirn onwards, mainly for Laurentian genera migrating to Avalonia (e.g. Steuslo⁄na, Tetradella and Eoaquapulex; see Fig. 1). Otherwise, in many cases Baltica was the origination point for genera which subsequently migrated to Avalonia and Laurentia, such as Distobolbina, Medianella, Platybolbina, Hippula and possibly Easchmidtella (Fig. 1). During the Ashgill there are several species^level links between Avalonia, Baltica and Laurentia, particularly amongst binodicopes such as Kinnekullea comma, Spinigerites unicornis, Aechmina maccormicki and Pseudulrichia simplex (see Floyd

et al., 1999; Williams et al., 2001a,b), but also including palaeocopes such as Eoaquapulex maccoyii (see Siveter, in press). Migration of ostracods continued between Avalonia, Baltica and Laurentia in the late Ashgill (Hirnantian), with up to 20 migrations, involving palaeocopes (e.g. Anticostiella), podocopes (e.g. Monoceratella) and eridostracans (e.g. Cryptophyllus ; see Fig. 1). Overall, a large proportion of the ostracod migrations (over 40%) involved genera which may have originated in Baltica (or Malopolska). This re£ects the higher diversity of ostracod faunas in this area, relative to Avalonia and Ibero-Armorica (see Vannier et al., 1989, text-¢gure 27), and also its intermediate palaeogeographical setting between these palaeocontinents and Laurentia (see Cocks, 2000 and below).

5. Interpreting the patterns of migration In brief, ostracod migrations in the Ordovician are characterised by: few migrations prior to the Llanvirn; a rapid increase in migration during the late Llanvirn, particularly between Ibero-Armorica, Perunica, Baltica and Avalonia; an early peak of migration from Baltica to Laurentia during the early to mid Caradoc; increased migration between Baltica and Avalonia, and between Laurentia and Baltica during the Ashgill; reduced migration to and from Ibero-Armorica after the mid Caradoc. In this section, the complex pattern and rate of migrations are examined in relation to: overall ostracod diversity; availability of compatible marine environments; global sea-level changes; climate ; and palaeogeography. Dispersal of ostracods via intermediate vectors such as other marine biota, mid-ocean islands or o¡shore carbonate platforms is also addressed. 5.1. Links between diversity and the rate of ostracod migration The major Palaeozoic ostracod groups (e.g. beyrichiocopes, podocopes etc.) ¢rst appear during the Arenig as part of the ‘Palaeozoic Fauna’ (Sepkoski, 1981): earlier records of ostracods are

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

scant and the Tremadoc faunas are composed mainly of bradoriids (see Williams and Siveter, 1998), some of which at least, appear not to be ostracods (e.g. Hou et al., 1996). The Rivillina-like forms (binodicopes?) in the Tremadoc of southern Britain (Williams and Siveter, 1998) may be amongst the oldest ostracods. Arenig ostracod faunas of Gondwana (mainly known from Ibero-Armorica), Avalonia, Baltica and Laurentia were mostly low diversity (Vannier et al., 1989, text-¢gure 27). For example, based on the Anglo-Welsh fauna (Siveter, in press), only three Arenig genera of ostracods (‘Leperditia’, ‘Parapyxion’ and ‘Conchoprimita’) are represented in Avalonia. In Laurentia ostracod faunas of the Kanosh Shale of Utah (six genera; see Berdan, 1988), Joins Formation of Oklahoma (six genera; see Harris, 1957; Williams, 1990) and Sunblood Formation of NW Canada (seven genera; see Copeland, 1982, ¢gure 4) are all of relatively low diversity. Diversity amongst Baltic faunas (including Malopolska) was higher (e.g. Hessland, 1949; Olempska, 1994), but lower than for Llanvirn, Caradoc or Ashgill Baltic faunas (see Vannier et al., 1989; Meidla, 1996). The very low diversity is re£ected in the small number of ostracod migrations recorded from this interval (Figs. 1 and 3). Nevertheless, some ostracod genera which later became migrants, particularly Baltic forms such as Laccochilina, Hithis, Ogmoopsis, Euprimites and Uhakiella, already existed in these early ostracod faunas and were widespread in their local palaeogeographical setting (e.g. Cryptophyllus in Laurentia). This indicates that barriers to migration (palaeogeographical, climatic) were probably substantial during the earlier Ordovician, especially between Laurentia, Baltica and Avalonia (see Figs. 3 and 4). The very marked increase in ostracod migration during the late Llanvirn, at the time of the teretiusculus graptolite Biozone (Fig. 3), correlates with the most rapid increase in ostracod generic diversity recorded in Ibero-Armorica, Avalonia and Baltica during the Ordovician (Vannier et al., 1989, text-¢gure 27). The increase in ostracod diversity may partly be linked to the Llanvirn eustatic transgression, which £ooded cratonic areas, providing a larger marine^shelf area for

213

ostracod colonisation and diversi¢cation. However, increased diversity is also probably a product of the major restructuring of marine communities that occurred at this time, a result of which is faunas more typical of the Palaeozoic in general. Similar diversi¢cation is evident amongst other arthropods, for example trilobites (see Barnes et al., 1995; Adrain and Westrop, 2000). Following the major diversi¢cation of ostracods in the Llanvirn, migrations continued at a pace of 13 or more migrations per graptolite biozone until the late Ashgill (Fig. 3). The gradual reduction in ostracod diversity through the mid Caradoc to Ashgill (Vannier et al., 1989, text-¢gure 27) is probably exaggerated by the lack of data from Ibero-Armorica, partial data from Avalonia, particularly from the late Caradoc and Ashgill, and by incomplete data from the Ashgill of the Baltic region. Thus, extensive high-diversity Avalonian faunas are known from southern Ireland, though not formally described (see Siveter, in press), and from northern England (see Williams et al., 2001b). Meidla (1996, p. 11), noted high diversity in Baltic ostracod faunas until the Hirnantian. Laurentian faunas may have peaked in diversity during the early Caradoc, but remained diverse throughout the Caradoc and Ashgill (e.g. Spivey, 1939; Keenan, 1951; Copeland, 1970, 1973, 1989). In the Argentine Precordillera, faunas showed the highest diversity during the early Caradoc (gracilis Biozone) whilst late Caradoc and Ashgill faunas were sparse. Perhaps partly as a result of continuing incomplete data from Ibero-Armorica, the number of migrating genera (and particularly new genera contributing to migration) appears to have decreased rapidly after the mid Caradoc (Fig. 3). There is a similar dearth of knowledge for Ashgill ostracods from Gondwana. Despite an overall recorded reduction in diversity, the number of ostracod migrations remained relatively high during much of the Ashgill : the apparently reduced rate of migration during the latest Caradoc and early Ashgill (linearis Biozone, see Fig. 3) might partly re£ect poorer documentation of ostracods from this interval, particularly in Avalonia and IberoArmorica. Nevertheless, the rate of migration certainly began to increase again during the mid to

PALAEO 3079 9-5-03

214

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

late Ashgill, rising to 24 migrations during the Hirnantian, representing the fourth highest peak of dispersal for the Ordovician (Fig. 3). As overall ostracod diversity was much reduced at this time (e.g. see Meidla, 1996), factors apparently unrelated to diversity were in£uencing widespread ostracod dispersals (see below). 5.2. The in£uence of compatible marine facies, global sea level, palaeogeography and climate In this section, the distribution of Ordovician ostracods is assessed in terms of the availability of compatible marine facies, and the in£uence of global sea level, palaeogeography and climate 5.2.1. Availability of compatible marine facies In general, Ordovician benthic ostracod faunas reached their highest diversity in mid-shelf, warmwater tropical and sub-tropical environments where mudstones and carbonates were deposited. Clastic-dominated sequences, deposited in deeper water settings or at higher latitudes in cooler water settings, had lower diversity. During the Caradoc, the movement of Baltica and Avalonia towards the equator gradually brought shelf environments into latitudes where compatible environments existed on both sides of the Iapetus Ocean (see Fortey, 1984, p. 47; Cocks, 2000). Carbonates were widespread in Baltica, became increasingly common in Avalonia during the late Caradoc and Ashgill (see Jones, 1986, text-¢gure 6) and were also deposited in Ibero-Armorica during the Ashgill as this microcontinent drifted northwards. Although carbonate^mudstone facies existed in Laurentia from Early Ordovician times, and those of Llanvirn age (e.g. Oil Creek Formation of Oklahoma ; Fig. 4) already yielded moderately diverse ostracod faunas, these have no genera that occur contemporaneously in Baltica. However, during the early and mid Caradoc there was a steady in£ux of Baltic genera into Laurentia, and this occurred in tandem with the spread of carbonate^mudstone shelf^marine facies across Laurentia as a result of the early Caradoc transgression (e.g. see Barnes et al., 1995). These facies are represented by formations such as the Deco-

rah Shale (see Kay, 1940; Swain, 1987), Edinburg and Lincolnshire formations (see Kraft, 1962), Bromide Formation (Harris, 1957; Williams and Siveter, 1996) and Esbataottine Formation (Copeland, 1982 ; see also Fig. 4). Approaching 20% of the genera present in these Laurentian formations occur contemporaneously in Baltica by multidens Biozone times. Typical Baltic migrants to Laurentia are Hithis, Levisulculus, Oepikella and Platybolbina. Reduced migration of ostracods from Baltica to Laurentia during clingani Biozone times might, in part, relate to the spread of deeper water facies across the Laurentian palaeocontinent. Thus, in Oklahoma, the richly diverse ostracod-bearing Bromide Formation (80+ species) was replaced upwards by graptolite-bearing sediments of the Viola Group (see Finney, 1986). Nevertheless, in some areas shallower water environments persisted during this time (e.g. Lexington Limestone of Kentucky, see Warshauer and Berdan, 1982; Craighead Limestone of southern Scotland, see Fig. 4) and in some of the ostracod faunas of these formations over half of the genera occur contemporaneously in Baltica (Williams and Floyd, 2000). Although migrations from Baltica to Laurentia reached a peak during the early and mid Caradoc, and appear to have followed the spread of carbonate^mudstone facies across the Laurentian craton, migration of Laurentian genera to Baltica only peaked during the late Ashgill, in tandem with a second peak of migration from Baltica to Laurentia (Figs. 1^3). These migrations involved Monoceratella, Eographiodactylus, Aechmina, Foramenella, Anticostiella and Cryptophyllus, representing a mix of mid to deep shelf forms. At this time both Baltica and Laurentia were situated at low latitudes (see Cocks, 2000) and shared similar carbonate depositing facies (e.g. see Copeland, 1970, 1973 for Laurentian Canada, and Meidla, 1996 for the Baltic region). The availabilty of appropriate environments is hence an important control, but the late Ashgill two-way peak of migration between Laurentia and Baltica also suggests that other factors were at play, possibly associated with low sea level and palaeogeographical convergence. The latter seems to be favoured

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

as the peak of migration between Laurentia and Baltica follows a second (mid Ashgill) peak of migration of ostracods from Baltica to Avalonia (Fig. 3). The availability of compatible facies may also have played a role in migrations between Avalonia and Baltica. For example, the sparse ostracod faunas in the Arenig of southern Britain may be due to the absence of carbonate facies, and are in stark contrast to the much higher diversity, carbonate-hosted Arenig ostracod faunas of Scandinavia. Many of the late Llanvirn occurrences of ostracods in southern Britain, which are associated with an in£ux of Baltic taxa (see Fig. 3), are of secondarily silici¢ed faunas extracted from carbonates (e.g. see Jones, 1986). These contain Baltic-origin elements such as Laterophores, Vittella and Tallinella(?). Nevertheless, carbonates occur in the Caradoc of southern Britain, though migration of ostracods from Baltica was relatively subdued during this time, possibly as a result of higher global sea level (see below). The second peak of ostracod migration from Baltica to Avalonia during the mid Ashgill, occurred at a time when carbonates were widespread in southern Britain and Ireland (e.g. Keisley Limestone, Portrane Limestone etc.; see Jones, 1986, text-¢gure 6; Siveter, in press). Even in facies of Ashgill age which are clastic dominated, for example the Cautley Mudstone Formation of northern England, it is the rarer carbonate-rich levels which bear the ostracod faunas, and these are typically genera and species which originated in Baltica, such as Duplicristatia asymmetrica and Steuslo⁄na cuneata (see Williams et al., 2001b). The apparent lack of migration of ostracods to and from Ibero-Armorica after the mid Caradoc might be due to the lack of compatible facies, but, as the ostracods of the carbonate Rosan Formation (Ashgill age) of Armorica have yet to be described, it may also re£ect a lack of data. During the Caradoc there were migrations from Avalonia (Pseudbollia), Baltica (Trianguloschmidtella) and Laurentia (Elliptocyprites) to Gondwanan latitudes in the tropics (Australia), where carbonates of the Cliefden Caves Group were being deposited (see Webby and Packham, 1982). This indicates that, where compatible ma-

215

rine facies were developed, Ordovician ostracods could disperse between distant palaeogeographical entities during the Caradoc. 5.2.2. Palaeogeography There is no simple relationship between palaeogeography and ostracod migration patterns. Thus, some migrations appear to confound supposed palaeogeographical con¢gurations, including the migration of Elliptocyprites from Laurentia to Gondwana during the Caradoc, or that of Rivillina from Gondwana to Laurentia during the early Llanvirn. In the case of Elliptocyprites, climate and the availability of compatible marine facies appear to have been more in£uential than palaeogeography. Nevertheless, the closer palaeogeographical proximity of the Baltica, Avalonia and Laurentia palaeocontinents in low latitudes from Caradoc times onwards (Pickering and Smith, 1995; Cocks, 2000) may have been the most in£uential factor in determining patterns of mid Caradoc to Ashgill ostracod dispersals between these palaeocontinents. Migrations which may have been facilitated by this convergence include the in£ux of Baltic species into Avalonia. The earliest signi¢cant links between these palaeocontinents involved the migration of several Baltic genera to Avalonia in the late Llanvirn (teretiusculus Biozone), including Tallinnella, Piretopsis, Laterophores, and Medianella (Figs. 1^3). These migrations might be associated with the northwards drift of Avalonia, after its separation from Gondwana in the late Arenig (see Cocks, 2000). However, they are mirrored by a number of migrations from Baltica to Perunica, and from Baltica to Ibero-Armorica at the same time (Figs. 2 and 3), and, as these palaeocontinents may have drifted northwards somewhat later than Avalonia (see Cocks, 2000, ¢gure 9), other factors, perhaps related to lower sea level, may have been more in£uential in these migrations (see below). Nevertheless, during the Ashgill, and particularly from mid Cautleyan times the Baltic and Avalonian ostracod faunas became essentially cosmopolitan at the generic level and shared many species in common (Williams et al., 2001b). This supports the Late Ordovician^Early Silurian amalgamation of these palaeocontinents

PALAEO 3079 9-5-03

216

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

by closure of the Tornquist Sea (see Cocks et al., 1997; Cocks, 2000), a hypothesis underpinned by numerous other palaeontological and geological observations (see Cocks and Fortey, 1982; McKerrow and Cocks, 1986, p. 185). Widespread benthic trilobite species inhabitating former marine shelves are usually taken as evidence of palaeogeographical contiguity. For example, the late Caradoc^?early Ashgill shelf species Remopleurella burmeisteri embraces Scandinavia, Thailand and China and may indicate closure of the Tornquist Sea between Baltica and Gondwana (McCormick and Fortey, 1999). A Late Ordovician approach of an amalgamated Avalonia^Baltica towards Laurentia is also suggested by the many ostracod migrations (Fig. 3), and particularly by the several species which are common to Laurentia and Baltica^Avalonia by the mid to late Ashgill (Rawtheyan Stage). These include the binodicopes Spinigerites unicornis, Pseudulrichia simplex, Aechmina maccormicki and Kinnekullea comma (Williams et al., 2001b), and the palaeocope Eoaquapulex maccoyii (see Siveter, in press). Given the constraints of ostracod lifestyles and their entirely neritic habitat, these migrations may suggest early shelf contacts between Laurentia and a combined Avalonia^Baltica, which is predicted by some reconstructions of palaeogeography to have occurred during the late Ashgill (Pickering et al., 1988; Pickering and Smith, 1995; but cf. Soper and Woodcock, 1990; Soper et al., 1992; Cocks, 2000). These trans-Iapetus Ocean pioneer species might be relative chronometers of a soft collision between these palaeocontinents (see Williams et al., 2000). That several Laurentian and Avalonian genera appear to have secured an indirect route to each other via Baltica (see Fig. 1) conforms to the intermediate palaeogeographical setting of this palaeocontinent (see Pickering and Smith, 1995; Cocks, 2000). These migrants include typically mid to deep shelf forms such as Steuslo⁄na (see Copeland, 1982; Meidla, 1996; Williams et al., 2001b), but also shallower shelf forms such as Eoaquapulex (see Williams and Siveter, 1996; Copeland, 1982). The intermediate position of Baltica is also indicated by the large number of Baltic genera that contributed to the overall os-

tracod migrations: more than 40 migrating genera may have originated in Baltica (Fig. 1). This compares with possibly as many as 23 from Laurentia, or up to 21 from Avalonia. There were a number of migrations between Baltica, Avalonia and Gondwana (including Ibero-Armorica) during the late Llanvirn to mid Caradoc (Fig. 3), but migrations to Ibero-Armorica appear to have declined rapidly during the late Caradoc and Ashgill (see also Vannier et al., 1989), at a time when migrations between Laurentia, Baltica and Avalonia were accelerating (Figs. 1 and 3). Vannier et al. (1989) interpreted this decline as due to an expansion of the Rheic Ocean, between Ibero-Armorica and Avalonia and Baltica. However, as Ibero-Armorica may have begun to drift northwards towards Baltica and Avalonia by the Ashgill (see Cocks, 2000), the lack of migration may, at least in part, re£ect the dearth of information about Ashgill ostracods from Ibero-Armorica. There is a similar paucity of data about Ashgill ostracods from Gondwana, though migrations between Gondwana (e.g. Argentine Precordillera, Australia) were taking place during the Caradoc (see Schallreuter, 1981, 1988b, 1996, 1999; Salas, 2002a,b, in press). During the early Caradoc there was an in£ux of taxa from Laurentia, Avalonia, Perunica, IberoArmorica and Baltica into the Argentine Precordillera. These migrations include Ectoprimitioides, Cryptophyllus, Klimphores, Pseudulrichia, Aechmina(?), Pachydomelloides, Longiscula, Conchoprimitia, Baltonotella and Medianella (Fig. 1). Several genera arising in the Argentine Precordillera during the gracilis Biozone migrated to Baltica and Perunica, including Ordovizona, Velapezoides, Garciana and Kosuriscapha. Given the great disparity in facies between these areas, the intermediate palaeogeographical position of the Argentine Precordillera relative to the other palaeocontinents, and the closer palaeogeographical proximity of Baltica, Avalonia and Laurentia during the Caradoc, might have favoured these migrations. 5.2.3. Climate The in£uence of climate on ostracod migration patterns is closely related to palaeogeography,

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

and the two may have worked in tandem, together with the spread of compatible marine facies, to facilitate migrations between Laurentia, Baltica and Avalonia as these palaeocontinents converged. Early, signi¢cant migrations between Baltica and Avalonia occur in the late Llanvirn, perhaps about 10 million years after the Avalonian microcontinent rifted o¡ Gondwana and began drifting northwards. Notwithstanding the e¡ects of increased generic diversity at this time (see discussion above), many of these migrations involved the dispersal of Baltic taxa (e.g. Tallinnella, Piretopsis, Laterophores) that may have been facilitated by Avalonia’s drift into warmer water latitudes, as well as its palaeogeographical approach to Baltica. For example, Cocks (2000, ¢gure 9b) shows Avalonia at a similar latitude to Baltica by the late Llanvirn. Nevertheless, Perunica and Ibero-Armorica also received an in£ux of Baltic genera at this time (Fig. 3), though they may have been situated at higher latitudes (see Cocks, 2000). This suggests that some ostracod genera had relatively wide tolerance of mid to high latitude cool water environments, and accords with the facies distribution of some of the migrant ostracods to Perunica, which include Klimphores and Euprimites (Fig. 1). By the Ashgill, trans-oceanic ostracod dispersals between Avalonia, Laurentia and Baltica may not have been greatly limited by climatic controls. Even if the Iapetus Ocean was at least 1000 km wide at all points along its length during the Late Ordovician (e.g. McKerrow and Soper, 1989; Cocks, 2000; but cf. Pickering and Smith, 1995), Avalonia was probably situated no more than about 35‡ south of the equator. Baltica was situated at intermediate latitudes (between the equator and ca 30‡S) whilst Laurentia straddled the equator, its northern and southern quarters extending into sub-tropical latitudes (see Pickering and Smith, 1995; Cocks, 2000, ¢gure 9c). Thus, during the Ashgill, a combined Avalonia^Baltica and the Laurentia palaeocontinent would have had marine shelves which occupied areas of both warm tropical and cooler surface waters (e.g. see Wilde, 1991). A number of migrations of genera originating in Laurentia, ¢rst to

217

Baltica and then to Avalonia, suggest the in£uence of warmer climate, coupled to palaeogeographical convergence. These include Eoaquapulex, Tetradella and possibly Steuslo⁄na (Fig. 3). The latter is considered as a possible here, because its occurrence in deep shelf environments (e.g. Copeland, 1982; Williams et al., 2001b) suggests it was tolerant of cooler water. These ostracod genera migrated from Laurentia to Baltica in the mid Caradoc to early Ashgill, and thence to Avalonia during the mid Ashgill, presumably when the latter micro-continent had drifted into a warmer climatic zone. As such, these genera may be a relative proxy for surface water palaeo-temperature and this may also explain the relatively late migrations to Baltica of the longlived Laurentian genera Eographiodactylus and Monoceratella (Fig. 1). Elliptocyprites is a podocope with a morphologically simple external carapace, ¢rst documented from North America (see Swain, 1962, pl. 111, ¢gures 9a^c, 10), but also reported from the Argentine Precordillera (Salas, 2002b), Australia (Schallreuter, 1996) and the Baltic region (Tinn and Meidla, 1999). A key feature of this genus is the incurvature of the mid-ventral part of the carapace viewed in lateral pro¢le (Swain, 1962, p. 742). This feature, although plesiomorphic to many podocope ostracods, is present in the Australian species Elliptocyprites nesowa (see Schallreuter, 1988b, pl. 1, ¢gure 7), and if this can be con¢rmed as belonging to Elliptocyprites, and is not a homeomorph, Elliptocyprites is a migrant which appears to have had very wide distribution in the tropics. In contrast, the much earlier Baltic material referred to Elliptocyprites, from the mid Arenig of Sweden, lacks the mid-ventral incurvature of the carapace (see Tinn and Meidla, 1999, pl. 1, ¢gure 9; Tinn and Meidla, 2001, ¢gure 4AA) and may belong to a di¡erent genus. Elliptocyprites probably appeared in Laurentia during the early Caradoc (see Swain, 1962) and its Australian occurrence, in the Cliefden Caves Limestone Group, is only slightly younger (about the level of the British multidens Biozone). Australia is placed just north of, or straddling, the palaeoequator in most reconstructions of Caradoc palaeogeography (e.g. Barnes et al., 1995, ¢gure 6;

PALAEO 3079 9-5-03

218

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

see Fig. 2), with New South Wales presumably in the sub-tropics. There is no indication from carapace morphology that Elliptocyprites was anything but a benthic ostracod, suggesting that its dispersal to Gondwana may have been fortuitous (e.g. see Eager, 2000), or was secured by some unknown mechanism (island-hopping?, see below). Nevertheless, discounting the Baltic record, the distribution of the genus appears to have been tropical. Together with Elliptocyprites, Pseudbollia and Trianguloschmidtella are also reported from the Cliefden Caves Limestone Group. The latter two genera originated in the early Caradoc, Pseudbollia in Avalonia and Trianguloschmidtella in Baltica. Even if Pseudbollia had an earlier origin, before Avalonia rifted o¡ Gondwana, its occurrence in the Ordovician of Australia involves a temperate to tropical shift in its climatic range. Indeed, many Ordovician ostracod migrants appear to have been less in£uenced by climate, and this is particularly the case for binodicopes. Thus, the ¢rst ostracod migrations across the Iapetus Ocean involved species capable of occupying the more distal parts of the continental shelf which were presumably tolerant of cooler water. For example, Pseudulrichia had colonised Perunica, Baltica, Laurentia and Ibero-Armorica within three graptolite biozones after its ¢rst appearance in the late Llanvirn of Avalonia (Fig. 1), encompassing tropical and temperate latitudes. One of its species, Pseudulricha simplex, was common to Avalonia and Laurentia by the early Ashgill. Aechmina was also one of the most widespread binodicope taxa of the Caradoc and Ashgill, colonising Laurentia, Avalonia and Baltica after its probable ¢rst appearance in the early Caradoc of the Argentine Precordillera (Salas, 2002a). One of its species, A. maccormicki, was an early trans-Iapetus Ocean migrant. Kinnekullea comma and Spinigerites unicornis are additional examples, both known from deeper shelf facies (Williams et al., 2000; Williams et al., 2001b). 5.2.4. Sea level Prior to the Llanvirn there are too few ostracods for changes in migration rate relative to sea level to be quanti¢able. However, from the Llan-

virn until the late Ashgill recorded ostracod diversity was high, at least in low and mid latitudes. The rate of ostracod migration, relative to overall diversity, appears to have slowed during periods of higher global sea level (Fig. 3). This suggests that sea level may have been one of the most profound in£uences on Ordovician migration patterns. For example, during the early Caradoc global transgression, when sea level reached its probable Ordovician peak (see Barnes et al., 1995, p. 154), and there were more than 80 potential ostracod migrant genera (Fig. 3), there were fewer migrations (about 29) than occurred during the late Llanvirn (31 migrations), when sea level and ostracod diversity were much lower. In contrast, the early Caradoc was a time of wide dispersal for trilobites (Barnes et al., 1995). Partly this re£ects the onshore migration of a deep-shelf, widespread trilobite fauna (e.g. Fortey, 1984). Some deep-water trilobite species, for example olenids, already had widespread distribution and were adapted to conditions that could be found at the margins of more than one palaeocontinent (McCormick and Fortey, 1999). Their distribution was limited only by the viability of their supposed pelagic larvae. However, unlike trilobites, there are few deep marine ostracod contenders, so that ostracod migrations related to the Caradoc transgression are probably not the product of shelf-wards migration of an already widespread deep marine (cooler water) ostracod fauna (cf. Fortey, 1984; Barnes et al., 1995, pp. 156, 157). Even Pseudulrichia (see above), did not have a range that extended beyond the deep neritic. The constrast with the trilobites is suggested by: the small reduction in ostracod migration rate during the early Caradoc (Fig. 1); and by the rarity of ostracod migrants from cool water areas fringing Gondwana (see Fig. 2). This contrasts with the appearance of the peri-Gondwanan cyclopygid trilobite biofacies in Laurentia during the Caradoc. Ostracods may have been responding to global sea level in di¡erent ways from trilobites. The initial small reduction in migration during the early Caradoc (Fig. 3), when sea level was at its highest, may have isolated some ostracod migrants by increasing the distances of migration

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

and cutting o¡ some island-hopping routes (see below). This may have severely restricted ostracod migrations from Baltica to Avalonia, Perunica and Ibero-Armorica (see Fig. 3). It followed the early peak of ostracod migration for the Ordovician during the late Llanvirn, when sea level was low. Other peaks in ostracod migration rate also occurred whilst global sea level was low, for example in the early Llanvirn, and the mid to late Ashgill (Fig. 3). The early and late Llanvirn peaks of ostracod migration involved some taxa dispersing between Gondwana, Ibero-Armorica, Perunica and Avalonia (Figs. 1 and 3), which may already have been widespread before the separation of these continental masses, from the late Arenig onwards (see Cocks, 2000). Conversely, they also include many migrations of Baltic taxa which cannot be explained in this way (Fig. 1). These latter migrations may have been associated with the palaeogeographical approach of these palaeocontinents facilitated in tandem with lower sea level. A similar mechanism is also suggested by the mid to late Ashgill migrations of marine shelf ostracods between Laurentia and Baltica (Fig. 1). Lower sea level may have favoured dispersal in two ways. Firstly it narrowed the marine seaways between di¡erent palaeocontinents. Secondly, it favoured dispersal via outer shelf or mid-ocean islands, ‘island-hopping’ routes (see Fortey, 1984, p. 40). These staging-post intermediaries could contribute to increased migration if other factors, such as palaeogeographical narrowing of oceanic barriers increased the chances of transmission of faunas. Migration by this means might explain the increased levels of dispersal at times of low global sea level. This form of dispersal may account for the distribution patterns of some modern shallow marine Paci¢c ostracods, several of which have migrated eastwards (possibly against some ocean currents) from the Indonesian and Australian marine shelves and the Indian Ocean to reach such isolated outposts as Easter Island (Whatley and Jones, 1999; Whatley, 2000). There is evidence for the existence of such islands in the Iapetus Ocean (see Fortey, 1984; McKerrow and Cocks, 1986; Armstrong, 2000; Harper et al., 1996) and further a¢eld (e.g. Webby and Packham, 1982). For example, the Wrae Lime-

219

stone of the Southern Uplands of Scotland (see Owen et al., 1996), which comprises allochthonous fossiliferous limestone blocks (of late Llanvirn or early Caradoc age), occurs in deep water mudstones of late Caradoc or Ashgill age (Owen et al., 1996). The faunas include brachiopods and trilobites and some poorly preserved ostracods including possible Eoaquapulex and Baltonotella (see Williams et al., 2001a). The limestones may have been derived from a now vanished seamount (Leggett, 1980), or more probably from an exhumed carbonate platform from an outer shelf setting (Armstrong, 2000; Armstrong and Owen, 2001). A similar setting is envisaged for the Cliefden Caves Limestone Group of New South Wales, Australia (Webby and Packham, 1982), which surmounts a thick volcanic rock pile, and from which ostracod genera of mixed Baltic, Avalonian and probable Laurentian origin are recorded (see Schallreuter, 1988b). Together with the occurrence of marine shelf ostracods in the o¡shore platform Pen y Garnedd Phosphorite of North Wales (Cave, 1965), these distributions suggest the wide potential of islands or o¡shore platforms as ‘staging posts’ for ostracod dispersal. 5.2.5. Possible dispersal via other marine biota Other potential mechanisms for dispersing ostracods include transportation via marine biota such as trilobites, orthoconic nautiloids or ¢sh (see Schallreuter and Siveter, 1985). These could have carried ostracods across ocean barriers. Ostracods are known to be able to sustain drastic environmental conditions, with little or no food, for several weeks (see Eager, 2000). Thus, the possibility of ostracod dispersal via other marine biota might explain some of the migrations, particularly those which appear to confound overall trends, such as the migration of Rivillina from Gondwana (Ibero-Armorica) or Avalonia to Laurentia during the Llanvirn, or the migration of Elliptocyprites from Laurentia to Gondwana (Australia) during the mid Caradoc (Fig. 1). This form of dispersal might have contributed to a progressively greater number of ostracod migrations during the Caradoc and Ashgill (relative to the Arenig and Llanvirn) if potential hosts (e.g.

PALAEO 3079 9-5-03

220

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

bathypelagic trilobites) had evolved to provide a more ready medium for transport. However, there is no evidence of association for dispersal of ostracods via orthoconic nautiloids, ¢sh, trilobites or other marine hosts. Some ocean-going animals, for example graptolites, may, in adult life, have been restricted to relatively narrow belts of upwelling oceanic currents (see Finney and Berry, 1997) and are unlikely vectors for transport. Vannier (1986a) has speculated that some Ordovician ostracods might have used their long spines to attach to £oating seaweed, though there is no con¢rmation of this association. Also, this method of dispersal may not be well suited to the dispersal of benthic invertebrates poorly adapted to survive long periods as epiplankton in the open sea (see Schallreuter and Siveter, 1985, p. 591). The absence of ostracods from deep marine Ordovician fossil assemblages suggests that this mode of dispersal did not occur commonly.

6. Conclusions Migration patterns of Ordovician ostracods based on European and American faunas are complex, and were in£uenced by the interplay of several factors: b The main migrating groups were palaeocopes and binodicopes, re£ecting the most diverse Ordovician ostracod higher taxa. Of newly arising migrant taxa, palaeocopes dominated the Llanvirn, but binodicopes and podocopes were more important migrants during the Caradoc and Ashgill. This probably re£ects reduced palaeocope diversity in the Late Ordovician. b Most ostracod genera migrated to two or three palaeocontinents and a few, generally longlived forms (e.g. Euprimites), migrated to four or more palaeocontinents. Migration rates were generally slow, particularly when compared to the rates of migration for pelagic organisms, and each migration to a new palaeocontinental area often took the duration of one or more graptolite biozones. Several ostracods were ‘rapid’ migrants, particularly the widespread genus Pseudulrichia. b The low migration rate prior to the Llanvirn

was partly in£uenced by overall low taxonomic diversity, though future migrant genera such as Laccochilina, Cryptophyllus and Euprimites already existed in Arenig faunas: these may have faced formidable climatic and palaeogeographical barriers to migration. b Higher generic diversity provided a larger pool of ostracods for migration from the Llanvirn onwards. b Relative to overall ostracod diversity, migration rates show an overall negative trend relative to high global sea level, with increased migration occurring during lower global sea level. This pattern may have been faciliated by a narrowing of marine sea-ways during low sea level and by increased availability of island-hopping routes. Possible dispersal by such routes is indicated by the existence of biogeographically mixed island faunas, for example in the Cliefden Caves Limestone Group of New South Wales, Australia. b Migration was facilitated by palaeogeographical convergence, coupled with the availability of similar marine facies and climate. This was probably the primary driving force for migrations between Laurentia, Baltica and Avalonia from Caradoc times onwards. Amalgamation of Avalonia and Baltica during the Ashgill is suggested by the cosmopolitan ostracod fauna. Approach of Avalonia^Baltica to Laurentia is indicated by the migration of numerous ostracod genera and several ostracod species between these areas (e.g. Kinnekullea comma, Aechmina maccormicki, Spinigerites unicornis, Pseudulrichia simplex), by the Ashgill. b Some ostracods, particularly the binodicopes Pseudulrichia, Klimphores, Kinnekullea, Aechmina and Spinigerites, could occupy deep shelf and cooler water benthic environments. They were part of a widespread deeper shelf fauna from the mid Caradoc onwards which may have e¡ectively narrowed the distances, or reduced climatic barriers, for trans-oceanic migration, though none of these ostracods were bathyal. b Dispersal via other marine hosts may have contributed to migration, though there is no evidence of association of ostracods with, for example, bathypelagic trilobites, orthoconic nautiloids or planktonic biota.

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Acknowledgements We thank Deborah Di Naccio for help with micropalaeontological samples, Asih Williams for help preparing the manuscript and Mike Bassett and Leonid Popov (National Museum of Wales) for the palaeogeographical reconstruction used in Figure 2. Stewart Molyneux and Tim McCormick (BGS) suggested improvements to an earlier draft, whilst Alan Owen (Glasgow) and Giles Miller (Natural History Museum, London) provided valuable reviews. Thomas Servais (Lille) is thanked for his editorial comments. M.W. acknowledges Leicester University for a Honorary 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

221

Visiting Fellowship. This paper is a contribution to the team project of UMR 5125 (CNRS) on the structure and functioning of aquatic palaeoecosystems and to IGCP 410, ‘The Great Ordovician Biodiversi¢cation Event’. M.W., J.D.F. and P.S. publish by permission of the Director, British Geological Survey.

Appendix The following references provide a basis for the reconstruction of ostracod genera ranges and migration routes reproduced in Fig. 1.

Eodominina: Salas, 2002a; Schallreuter, 1993b, 1998 Rivillina: Berdan, 1988; Vannier et al., 1989 Gracquina: Vannier et al., 1989; Siveter, in press Conchoprimitia: Vannier et al., 1989; Olempska, 1994, Tinn and Meidla, 1999; Schallreuter and Krufita, 2000b; Siveter, in press Parinconchoprimita: Schallreuter and Siveter, 1985; Vannier and Vaslet, 1987; Vannier et al., 1989 Collibolbina: Vannier et al., 1989; Harris, 1957; Williams and Siveter, 1996 for Winchellatia Pilla: Jones and Schallreuter, 1990; Salas, 2002a; Schallreuter, 1996; Schallreuter and Siveter, 1988a,b Pyxion: Schallreuter and Krufita, 1988; Vannier et al., 1989 Miehlkella: Vannier et al., 1989 Tallinnella: Schallreuter and Krufita, 1988; Vannier et al., 1989; Siveter, in press Piretopsis: Vannier et al., 1989; Schallreuter and Krufita, 2000a; Tinn and Meidla, 2001 Laccochilina: Berdan, 1988; Vannier et al., 1989; Olempska, 1994; Tinn and Meidla, 2001; Schallreuter and Krufita, 2000b Ogmoopsis: Vannier et al., 1989; Tinn and Meidla, 2001; Siveter, in press; Williams and Floyd, 2000 Euprimites: Vannier et al., 1989; Olempska, 1994; Schallreuter et al., 1996; Tinn and Meidla, 1999, 2001; Schallreuter and Krufita, 1988, 2000b Vittella: Vannier et al., 1989; Olempska, 1994; Floyd et al., 1999; Siveter, in press Uhakiella: Schallreuter and Krufita, 1988; Vannier et al., 1989; Olempska, 1994; Siveter, in press Laterophores: Vannier et al., 1989; Tinn and Meidla, 2001; Siveter, in press Leperditella: Harris, 1957; Berdan, 1988; Williams, 1990; Vannier et al., 1989 Primitiella: Vannier et al., 1989; Olempska, 1994; Williams and Siveter, 1996 Hithis: Vannier et al., 1989; Melnikova, 1999; Williams et al., 2001a Medianella: Kraft, 1962; Vannier et al., 1989; Meidla, 1996; Schallreuter, 1996; Salas, 2002b Punctaparchites: Harris, 1957; Vannier et al., 1989; Williams, 1990; Williams and Siveter, 1996 Ectoprimitoides: Berdan, 1988; Williams, 1990; Schallreuter, 1996 Ningulella: Warshauer and Berdan, 1982; Berdan, 1988; Williams, 1990; Schallreuter, 1996; Salas, in press Baltonotella: Vannier et al., 1989; Williams, 1990; Williams and Vannier, 1995; Tinn and Meidla, 2001; Salas, in press Cryptophyllus: Harris, 1957; Berdan, 1988; Vannier et al., 1989; Williams, 1990; Schallreuter, 1981, 1996, 1999 Krausella: Berdan, 1988; Williams, 1990, Meidla, 1996 Klimphores: Schallreuter and Siveter, 1985; Vannier and Vaslet, 1987; Vannier et al., 1989; Olempska, 1994; Schallreuter et al., 1996; Salas, 2002a; Siveter, in press Pachydomelloides: Vannier et al., 1989; Salas, 2002b Conchoprimitiella: Schallreuter and Siveter, 1985; Vannier et al., 1989; Siveter, in press Easchmidtella: Schallreuter and Krufita, 1988; Vannier et al., 1989; Siveter, in press; Williams and Siveter, 1996 Bromidella: Harris, 1957; Vannier et al., 1989; Williams, 1990 Chilobolbina: Copeland, 1973; Vannier et al., 1989 Jeanlouisiella: Vannier et al., 1989; Schallreuter et al., 1996 Reuentalina: Vannier et al., 1989; Schallreuter et al., 1996

PALAEO 3079 9-5-03

222 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89.

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228 Brephocarieis: Schallreuter and Krufita, 1988, 1994; Siveter, in press Schallreuteria: Vannier et al., 1989; Siveter, in press Homeokiesowia: Vannier et al., 1989; Siveter, in press Pedomphalella: Swain et al., 1961; Vannier et al., 1989; Siveter, in press Duringia: Copeland, 1982, Olempska, 1994, Schallreuter et al., 1996, Schallreuter and Krufita, 2001b, Siveter, in press Ceratopsis: Keenan, 1951; Swain, 1987; Vannier, 1987; Schallreuter and Krufita, 1988; Vannier et al., 1989; Siveter, in press Bullaeferum: Vannier et al., 1989; Siveter, in press Piretella: Vannier et al., 1989; Olempska, 1994; Schallreuter and Krufita, 1988, 2001a; Schallreuter et al., 1996 Quadritia: Vannier et al., 1989; Olempska, 1994; Schallreuter et al., 1996; Siveter, in press Piretia: Vannier et al., 1989; Olempska, 1994; Schallreuter et al., 1996 Vannieria: Schallreuter et al., 1996 (Copelandia), Schallreuter, 1999 and references therein, Siveter, in press (Copelandia) Pseudorayella: Olempska, 1994; Williams et al., 2001b Pseudulrichia: Vannier et al., 1989; Schallreuter, 1996, 1999; Schallreuter and Krufita, 2001a; Salas, 2002a; Siveter, in press Longiscula: Schallreuter, 1981; Vannier et al., 1989; Olempska, 1994; Williams et al., 2001b; Salas, 2002b Hesperidella: Vannier et al., 1989; Olempska, 1994 Levisulculus: Vannier et al., 1989; Kesling, 1960a, Williams and Floyd, 2000 Oepikella: Vannier et al., 1989; Copeland, 1965, Warshauer and Berdan, 1982; Williams and Floyd, 2000 Platybolbina: Kesling, 1960b; Schallreuter and Krufita, 1988; Vannier et al., 1989; Olempska, 1994; Williams and Floyd, 2000 Vogdesella: Schallreuter and Siveter, 1985; Vannier et al., 1989; Olempska, 1994, Siveter, in press Hippula: Jones, 1987; Vannier et al., 1989; Williams, 1990; Schallreuter and Krufita, 1988, 2001a Sigmoopsis: Vannier et al., 1989; Williams et al., 2001a; Siveter, in press Acanthoscapha: Swain, 1962; Schallreuter, 1996; Salas, 2001 Bolboscapha: Schallreuter, 1999 and references therein Hastatellina: Schallreuter and Krufita, 1988; Vannier et al., 1989; Siveter, in press Kosuriscapha: Blumenstengel, 1965, Knu«pfer, 1968, Salas, 2002b, Schallreuter, 1996 Garciana: Schallreuter, 1994; Schallreuter and Hinz-Schallreuter, 1998 Ordovizona: Vannier et al., 1989; Salas, in press Aechmina: Kay, 1940; Williams et al., 2001b and references therein, Salas, 2002a Trianguloschmidtella: Schallreuter, 1988b; Vannier et al., 1989 Byrsolopsina: Schallreuter and Siveter, 1985; Vannier et al., 1989 Velapezoides: Vannier et al., 1989; Schallreuter, 1996; Salas, 2002b Dornbuschia: Vannier et al., 1989; Schallreuter, 1996; Salas, 2002b Revisylthere: Vannier et al., 1989; Schallreuter, 1996; Salas, 2002b Oejlemyra: Vannier et al., 1989; Schallreuter, 1996, 1999 Eocytherella: Vannier et al., 1989 Crescentilla: Vannier et al., 1989; Schallreuter and Krufita, 1988, 2001a; Siveter, in press Pseudbollia: Schallreuter, 1988b; Siveter, in press Latebina: Schallreuter and Krufita, 1994; Siveter, in press Steuslo⁄na: Copeland, 1982; Vannier et al., 1989; Olempska, 1994; Williams et al., 2001b; Salas, 2002b Elliptocyprites: Swain, 1962, Schallreuter, 1988b, Tinn and Meidla, 1999, Salas, 2002b Eridoconcha: Williams and Jones, 1990; Siveter, in press Balticella: Schallreuter and Siveter, 1985; Williams and Siveter, 1989; Vannier et al., 1989 Kayina: Vannier et al., 1989; Williams and Vannier, 1993 Eoaquapulex: Kraft, 1962 (Oepikella frequens), Vannier et al., 1989; Williams, 1990; Siveter, in press Platyrhomboides: Kraft, 1962; Copeland, 1982; Vannier et al., 1989; Williams, 1990; Williams and Siveter, 1996 Monoceratella: Henningsmoen, 1954; Kraft, 1962; Williams and Floyd, 2000 Uninodobolba: Schallreuter, 1999 and references therein Orechina: Schallreuter and Siveter, 1985; Schallreuter and Krufita, 1988; Vannier et al., 1989 Hemeaschmidtella: Schallreuter and Siveter, 1985; Vannier et al., 1989 Rectella: Schallreuter and Siveter, 1985; Vannier et al., 1989 Moeckowia: Schallreuter and Siveter, 1985; Vannier et al., 1989 Henningsmoenia: Vannier et al., 1989, Siveter, in press Eographiodactylus: Kraft, 1962; Schallreuter and Siveter, 1985; Vannier et al., 1989 Distobolbina: Vannier et al., 1989, Williams et al., 2001a,b

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

223

90.

Spinigerites: Copeland, 1973 (Primitiella unicornis), Vannier et al., 1989; Swain, 1987 (Primitiella plattevillensis), Schallreuter, 1999; Williams et al., 2001b 91. Ulrichia: Vannier et al., 1989, Siveter, in press 92. Warthinia: Warshauer and Berdan, 1982; Williams et al., 2001b 93. Tetradella: Schallreuter and Krufita, 1988; Vannier et al., 1989; Williams et al., 2001b 94. Satiellina: Vannier et al., 1989 95. Daliella: Copeland, 1973; Vannier et al., 1989 96. Foramenella: Copeland, 1973; Vannier et al., 1989 97. Brevibolbina: Schallreuter and Siveter, 1985; Vannier et al., 1989 98. Kinnekullea: Vannier et al., 1989; Floyd et al., 1999; Williams et al., 2000, 2001b 99. Laevonotella: Meidla, 1996; Williams et al., 2001b 100. Duplicristatia: Vannier et al., 1989; Williams et al., 2001b 101. Gotula: Vannier et al., 1989; Siveter, in press 102. Bulbosclerites: Vannier et al., 1989; Williams et al., 2001b 103. Pseudohippula: Copeland, 1970, 1973 (Monoceratella castorensis), Vannier et al., 1989, Williams et al., 2001b 104. Anticostiella: Copeland, 1973; Vannier et al., 1989 105. Scanipisthia: Schallreuter and Krufita, 1988; Vannier et al., 1989 106. Bollia: Schallreuter and Krufita, 1988; Vannier et al., 1989 107. Caprabolbina: Schallreuter and Siveter, 1985; Vannier et al., 1989 The list is not inclusive of all relevant references given in this paper, but concentrates on the earliest records of taxa from di¡erent palaeocontinents. For Baltica and Ibero-Armorica many ranges are based on data in Vannier et al. (1989, text-¢gures 2, 10^ 17, and references therein). Ranges for Avalonia are based primarily on Siveter, in press (revising Vannier et al., 1989). Ranges for Laurentia are based on comparison of individual records with the range charts published by Barnes et al. (1981) and Ross et al. (1982) with some modi¢cations. In some of the above publications the genus may appear under a di¡erent name. Some genera depicted as migrants by Vannier et al. (1989) and Schallreuter and Siveter (1985) are omitted subject to further taxonomic revision.

References Adrain, J.M., Westrop, S.R., 2000. An empirical assessment of taxic paleobiology. Science 289, 110^112. Armstrong, H.A., 2000. Conodont micropalaeontology of mid-Ordovician aged limestone clasts from Lower Old Red Sandstone conglomerates, Lanark and Strathmore basins, Midland Valley, Scotland. J. Micropalaeontol. 19, 45^60. Armstrong, H.A., Owen, A.W., 2001. Tectonic evolution of the paratectonic Caledonides of Northern Britain. J. Geol. Soc. Lond. 158, 475^486. Astini, R.A., 1995. Sedimentolog|¤a de la Formacio¤n Las Aguaditas (talud carbona¤tico) y evolucio¤n de la cuenca precordillerana durante el Ordov|¤cico medio. Rev. Asoc. Geol. Argent. 50, 143^164. Astini, R.A., Benedetto, J.L., Vaccari, N.E., 1995. The Early Paleozoic evolution of the Argentine Precordillera as a Laurentian rifted, drifted and collided terrane: A geodynamic model. Bull. Geol. Soc. Am. 107, 253^273. Arthurton, R.S., Wadge, A.J., 1981. Geology of the country around Penrith. Mem. Geol. Surv. GB, Sheet 24. HMSO, London, 177 pp. Barnes, C.R., Norford, B.S., Skevington, D., 1981. The Ordovician System in Canada. Correlation chart and explanatory notes. Int. Union Geol. Sci. Publ. No. 8, 27 pp. Barnes, C.R., Fortey, R.A., Williams, S.H., 1995. The pattern of global bio-events during the Ordovician period. In: Walliser, O.H. (Ed.), Global Events and Event Stratigraphy in the Phanerozoic, Springer, pp. 141^172. Benedetto, J.L., 1993. La hipo¤tesis de la alocton|¤a de la Pre-

cordillera Argentina: un test estratigra¤¢co y bioestratigra¤¢co. XII Congreso Geolo¤gico Argentino y II Congreso de Exploracio¤n de Hidrocarburos, Actas III, pp. 375^384. Benedetto, J.L., Sa¤nchez, T.M., Carrera, M.G., Brussa, E.D., Salas, M.J., 1999. Paleontological constraints on successive paleogeographic positions of Precordillera terrane during the early Paleozoic. In: Ramos, V.A., Keppie, J.D. (Eds.), Laurentia-Gondwana connections before Pangea. Geol. Soc. Am. Spec. Pap. 336, pp. 21^42. Berdan, J.M., 1976. Middle Ordovician leperditicopid ostracodes from the Ibex Area, Millard County, Western Utah. Brigham Young Univ. Geol. Stud. 23, 37^65. Berdan, J.M., 1984. Leperditicopid ostracodes from Ordovician rocks of Kentucky and nearby states and characteristic features of the Order Leperditicopida. US Geol. Surv. Prof. Pap. 1066-J, J1^J40. Berdan, J.M., 1988. Middle Ordovician (Whiterockian) palaeocopid and podocopid ostracodes from the Ibex Area, Millard County, Western Utah. N.M. Bureau Mines Miner. Resour. Mem. 44, 273^301. Bettley, R.A., Fortey, R.A., Siveter, D.J., 2001. High-resolution correlation of Anglo-Welsh Middle to Upper Ordovician sequences and its relevance to international chronostratigraphy. J. Geol. Soc. London 158, 937^952. Blumenstengel, H., 1965. Zur Ostrakodenfauna eines Kalkgero«lls aus dem Thu«ringer Lederschiefer (Ordovizium). Freiberger Forschrift (C) 182, 63^78. Botting, J.P., Thomas, A.T., 1999. A pseudoplanktonic inarticulate brachiopod attached to graptolites and algae. Acta Univ. Carol. Geol. 43, 333^335.

PALAEO 3079 9-5-03

224

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Burgess, I.C., Holliday, D.W., 1979. Geology of the country around Brough-under-Stainmore. Mem. Geol. Surv. GB, Sheet 131. HMSO, London, 131 pp. Cave, R., 1965. The Nod Glas sediments of Caradoc age in North Wales. Geol. J. 4, 279^298. Cocks, L.R.M., 2000. The early Palaeozoic geography of Europe. J. Geol. Soc. London 157, 1^10. Cocks, L.R.M., Fortey, R.A., 1982. Faunal evidence for oceanic separations in the Palaeozoic of Britain. J. Geol. Soc. London 139, 465^478. Cocks, L.R.M., McKerrow, W.S., Van Staal, C.R., 1997. The margins of Avalonia. Geol. Mag. 134, 627^636. Copeland, M.J., 1965. Ordovician Ostracoda from Lake Timiskaming, Ontario. Geol. Surv. Can. Bull. 127, 1^52, pls. 1^11. Copeland, M.J., 1970. Ostracoda from the Vaure¤al Formation (upper Ordovician) of Anticosti Island, Quebec. Geol. Surv. Can. Bull. 187, 15^29, pls. 4, 5. Copeland, M.J., 1971. Ostracoda. In: Steele, H.M., Sinclair, G.W., A middle Ordovician fauna from Braeside, Ottawa Valley, Ontario. Geol. Surv. Can. Bull. 211, 42. Copeland, M.J., 1973. Ostracoda from the Ellis Bay Formation (Ordovician) Anticosti Island, Quebec. Geol. Surv. Can. Pap. 72/43, 1^49. Copeland, M.J., 1977a. Ordovician Ostracoda, southeastern District of Franklin. In: Bolton, T.E., Sanford, B.V., Copeland, M.J., Barnes, C.E., Rigby, J.K., Geology of Ordovician Rocks, Melville Peninsula and Region, Southeastern District of Franklin. Geol. Surv. Can. Bull. 269, 77^97. Copeland, M.J., 1977b. Early Palaeozoic Ostracoda from Southwestern District of Mackenzie and Yukon Territory. Geol. Surv. Can. Bull. 275, 1^54, pls. 1^16. Copeland, M.J., 1982. Bathymetry of middle Ordovician (Chazy) ostracodes, Lower Esbaotaottine Formation, District of Mackenzie. Geol. Surv. Can. Bull. 347, 1^39. Copeland, M.J., 1989. Silici¢ed upper Ordovician - lower Silurian ostracodes from the Avalanche Lake area, southwestern District of Mackenzie. Geol. Surv. Can. Bull. 341, 1^ 100, 18 pls. Copeland, M.J., 2000. Ordovician Ostracoda from Southern Ba⁄n Island, Nunavut. In: Geology and paleontology of the southeast Arctic Platform and southern Ba⁄n Island, Nunavut. Geol. Surv. Can. Bull. 557, 217^236. Derby, J.R., Bauer, J.A., Creath, W.B., Dresbach, R.I., Ethington, R.L., Loch, J.D., Stitt, J.H., McHargue, T.R., Miller, J.F., Miller, M.A., Repetski, J.E., Sweet, W.C., Taylor, J.F., Williams, M., 1991. Biostratigraphy of the Timbered Hills, Arbuckle, and Simpson Groups, Cambrian and Ordovician, Oklahoma: a review of correlation tools and techniques available to the explorationist. Okla. Geol. Surv. Circ. 92, 15^41. Dzik, J., Pisera, A., 1994. Sedimentation and fossils of the Mo¤jcza Limestone. In: Dzik, J., Olempska, E., Pisera, A. (Eds.), Ordovician carbonate platform ecosystem of the Holy Cross Mountains Poland. Palaeontol. Pol. 53, 5^ 41. Eager, S.H., 2000. Freshwater ostracoda from Tarawa, Kiri-

bati: their implications for dispersal mechanisms. J. Micropalaeontol. 19, 68. Finney, S.C., 1986. Graptolite biofacies and correlation of eustatic, subsidence, and tectonic events in the middle to upper Ordovician of North America. Palaios 1, 435^461. Finney, S.C., Bergstro«m, S.M., 1986. Biostratigraphy of the Ordovician Nemagraptus gracilis Zone. In: Hughes, C.P., Rickards, R.B. (Eds.), Palaeoecology and biostratigraphy of graptolites. Geol. Soc. London Spec. Publ. 20, 47^59. Finney, S.C., Berry, W.B.N., 1997. New perspectives on graptolite distributions and their use as indicators of platform margin dynamics. Geology 25, 919^922. Floyd, J.D., Williams, M., Rushton, A.W.A., 1999. Late Ordovician (Ashgill) ostracodes from the Drummuck Group, Craighead Inlier, Girvan district, southwest Scotland. Scot. J. Geol. 35, 15^24. Fortey, R.A., 1984. Global earlier Ordovician transgressions and regressions and their biological implications. In: Bruton, D.L. (Ed.), Aspects of the Ordovician System. Contributions from the University of Oslo No. 295. Universitetsforlaget, Oslo, pp. 37^50. Fortey, R.A., 1999. Olenid trilobites as chemoautotrophic symbionts. Acta Univ. Carol. Geol. 43, 355^356. Fortey, R.A., Harper, D.A.T., Ingham, J.K., Owen, A.W., Rushton, A.W.A., 1995. A revision of Ordovician series and stages from the historical type area. Geol. Mag. 132, 15^30. Fortey, R.A., Harper, D.A.T., Ingham, J.K., Owen, A.W., Parkes, M.A., Rushton, A.W.A., Woodcock, N.H., 2000. A revised correlation of Ordovician rocks in the British Isles. Geol. Soc. Spec. Rep. No. 24. Gradstein, F.M., Ogg, J., 1996. A Phanerozoic time scale. Episodes 19, 3^5. Harper, D.A.T., Mac Niocaill, C., Williams, S.H., 1996. The palaeogeography of early Ordovician Iapetus terranes: an integration of faunal and palaeomagnetic constraints. Palaeogeogr. Palaeoclimatol. Palaeoecol. 121, 297^312. Harris, R.W., 1957. Ostracoda of the Simpson Group. OK Geol. Surv. Bull. 75, 1^333, 10 pls. Henningsmoen, G., 1953. The Middle Ordovician of the Oslo Region, Norway. 4. Ostracoda. Nor. Geol. Tidsskr. 32, 35^ 56. Henningsmoen, G., 1954. Upper Ordovician ostracods from the Oslo region, Norway. Nor. Geol. Tidsskr. 33, 69^108. Hessland, I., 1949. Lower Ordovician ostracods of the Siljan District, Sweden. Bull. Geol. Inst. Univ. Uppsala 33, 97^ 408. Hessland, I., Adamczak, F., 1974. On the taxonomic position of Steuslo⁄na Teichert (Ostracoda). Geosci. Man 1, 59^ 64. Hou, X-g., Siveter, D.J., Williams, M., Walossek, D., Bergstro«m, J., 1996. Preserved appendages in the arthropod Kunmingella from the early Cambrian of China: its bearing on the systematic position of the Bradoriida and the fossil record of the Ostracoda. Philos. Trans. R. Soc. London B 351, 1131^1145. Jaanusson, V., 1957. Middle Ordovician ostracodes of central

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228 and southern Sweden. Bull. Geol. Inst. Univ. Uppsala 37, 173^442. Jones, C.R., 1986. Ordovician (Llandeilo and Caradoc) Beyrichiocope Ostracoda from England and Wales. Part 1. Monogr. Palaeontogr. Soc., London, pp. 1^76, pls. 1^22 (Publ. No. 569, part of Vol. 138 for 1984). Jones, C.R., 1987. Ordovician (Llandeilo and Caradoc) Beyrichiocope Ostracoda from England and Wales. Part 2. Monogr. Palaeontogr. Soc., London, pp. 77^114, pls. 23^ 30 (Publ. No. 571, part of Vol. 139 for 1985). Jones, P.J., Schallreuter, R.E.L., 1990. On Pilla latolobata Jones and Schallreuter sp. nov. Stereo-Atlas of Ostracod Shells 17, 93^96. Kay, G.M., 1934. Mohawkian Ostracoda: species common to Trenton faunules from the Hull and Decorah formations. J. Paleontol. 8, 328^343. Kay, G.M., 1940. Ordovician Mohawkian ostracoda: Lower Trenton Decorah fauna. J. Paleontol. 14, 234^269. Keenan, J.E., 1951. Ostracodes from the Maquoketa Shale of Missouri. J. Paleontol. 25, 561^574. Kesling, R.V., 1960a. Middle Ordovician Black River ostracods from Michigan, Part II, Levisulculus and Eurychilina. Contrib. Mus. Paleontol. Univ. Michigan 15, pp. 349^363, 8 pls. Kesling, R.V., 1960b. Middle Ordovician Black River ostracods from Michigan, Part III, Platybolbina. Contrib. Mus. Paleontol. Univ. Michigan 15, pp. 365^385, 8 pls. Knu«pfer, J., 1968. Ostrakoden aus dem Oberen Ordovizium Thu«ringens. Freiberger Forschrift (C) 234, 5^29. Kraft, J.C., 1962. Morphologic and systematic relationships of some Ordovician Ostracoda. Mem. Geol. Soc. Am. 86, 1^ 104. Leggett, J.K., 1980. Palaeogeographic setting of the Wrae Limestone: an Ordovician submarine-slide deposit in Tweedale. Scot. J. Geol. 16, 91^104. Leslie, S.P., Bergstro«m, S.M., 1995. Revision of the North American late middle Ordovician standard stage classi¢cation and timing of the Trenton transgression based on Kbentonite bed correlation. In: Cooper, J.D., Droser, M.L., Finney, S.C. (Eds.), Ordovician Odyssey: short papers from the 7th International Symposium on the Ordovician System. SEPM, Fullerton, CA, pp. 49^54. Levinson, S.A., 1961. New genera and species of Bromide (Middle Ordovician) ostracodes of Oklahoma. Micropaleontology 7, 359^364. Lundin, R.F., Williams, M., Siveter, D.J., 1995. Domatial dimorphism occurs in leperditellid and monotiopleurid ostracodes. J. Paleontol. 69, 886^896. McCormick, T., Fortey, R.A., 1999. The most widely distributed trilobite species: Ordovician Carolinites genacinaca. J. Paleontol. 73, 202^218. McKerrow, W.S., Cocks, L.R.M., 1976. Progressive faunal migration across the Iapetus Ocean. Nature 263, 304^306. McKerrow, W.S., Cocks, L.R.M., 1986. Oceans, islands and olistostromes: the use of fossils in distinguishing sutures, terranes and environments around the Iapetus Ocean. J. Geol. Soc. London 143, 185^191.

225

McKerrow, W.S., Soper, N.J., 1989. The Iapetus suture in the British Isles. Geol. Mag. 126, 1^8. McKerrow, W.S., Dewey, J.F., Scotese, C.R., 1991. The Ordovician and Silurian development of the Iapetus Ocean. Spec. Pap. Palaeontol. 44, 165^178. Meidla, T., 1996. Late Ordovician ostracodes of Estonia. Fossilia Baltica 2. Tartu University Press, 222 pp., 32 pls. Melnikova, L.M., 1986. Ordovician ostracods of Kazakhstan (in Russian). Tr. Paleontol. Inst. Akad. Nauk SSSR 218, 3^ 102. Melnikova, L.M., 1999. Ostracodes from the Billingen Horizon (Lower Ordovician) of the Leningrad Region. Palaeontol. J. 33, 147^152. Olempska, E., 1994. Ostracodes of the Mo¤jcza Limestone. In: Dzik, J., Olempska, E., Pisera, A. (Eds.), Ordovician carbonate platform ecosystem of the Holy Cross Mountains Poland. Palaeontol. Pol. 53, 129^212. Orr, R.J., 1987. Upper Ordovician ostracodes from the Portrane Limestone, eastern Ireland. Thesis, Queen’s Univ., Belfast, 357 pp. Owen, A.W., Harper, D.A.T., Clarkson, E.N.K., 1996. The trilobites and brachiopods of the Wrae Limestone, an Ordovician limestone conglomerate in the Southern Uplands. Scot. J. Geol. 32, 133^149. Pickering, K.T., Smith, A.G., 1995. Arcs and back-arc basins in the early Paleozoic Iapetus Ocean. Island Arc 4, 1^67. Pickering, K.T., Bassett, M.G., Siveter, D.J., 1988. Late Ordovician-early Silurian destruction of the Iapetus Ocean: Newfoundland, British Isles and Scandinavia - a discussion. Trans. R. Soc. Edinb. Earth Sci. 79, 361^382. Popov, L., Bassett, M.G., Holmer, L.E., Gorjansky, V,Yu., 1999. Ordovician patterns of diversi¢cation in craniiformean brachiopods. Acta Univ. Carol. Geol. 43, 321^324. Prflibyl, A., 1984. Ostracodes from the Ordovician and Silurian of Bolivia. Casopis Mineral. Geol. 29, 353^368. Ross, R.J., Adler, F.J., Amsden, T.W., Bergstro«m, D., Bergstro«m, S.M., Carter, C., Churkin, M., Cressman, E.A., Derby, J.R., Dutro Jr., J.T., Ethington, R.L., Finney, S.C., Fisher, D.W., Fisher, J.H., Harris, A.G., Hintze, L.F., Ketner, K.B., Kolata, D.L., Landing, E., Neuman, R.B., Sweet, W.C., Pojeta Jr., J., Potter, A.W., Rader, E.K., Repetski, J.E., Shaver, R.H., Thompson, T.L., Webers, G.F., 1982. The Ordovician System in the United States. Correlation chart and explanatory notes. Int. Union Geol. Sci. Publ. No. 12, 73 pp. Rossi de Garc|¤a, E., Proserpio, C., 1976. Ostra¤codos Ordov|¤cicos de la Sierra de San Bernardo. VI Congr. Geol. Argent. Actas I, 557^563. Salas, M.J., 2001. Taxonom|¤a y paleobiogeograf|¤a de los ostra¤codos ordov|¤cicos de la Precordillera Argentina. Thesis, Univ. Nacional de Co¤rdoba, 171 pp. Salas, M.J., 2002a. Ostra¤codos binodicopas del Ordov|¤cico de la Precordillera de San Juan, Argentina. Ameghiniana 39, 41^58. Salas, M.J., 2002b. Ostra¤codos podocopas del Ordov|¤cico de la Precordillera de San Juan, Argentina. Ameghiniana 39, 129^ 149.

PALAEO 3079 9-5-03

226

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Salas, M.J., in press. Ostra¤codos paleocopas, leiocopas y eridostracas del Ordov|¤cico de la Precordillera de San Juan, Argentina. Ameghiniana. Sarv, L., 1959. Ordovician ostracods in the Estonian S.S.R. (in Russian, English summary). ENSV Teaduste Akad. Geol. Inst. Uurimused 4, 1^207. Sayar, C., Schallreuter, R.E.L., 1989. Ordovician ostracodes from Turkey. N. Jb. Geol. Pala«ontol. Monatsh. 4, 233^242. º jlemyrgeschieben Schallreuter, R.E.L., 1975. Ostrakoden aus O (Ordoviz). II. N. Jb. Geol. Pala«ontol. Abh. 150, 270^293. Schallreuter, R.E.L., 1981. Ordovizische Ostrakoden Argentiniens. Mitt. Geol. Pala«ontol. Inst. Univ. Hamburg 51, 9^21. Schallreuter, R.E.L., 1984. Geschiebe-Ostrakoden I. N. Jb. Geol. Pala«ontol. Abh. 169, 1^40. º jlemyr£int-GeSchallreuter, R.E.L., 1986. Ostrakoden aus O schieben von Sylt. In: von Hacht, U. (Ed.), Fossilien von Sylt II. Verlag und Verlagbuchhandlung Inge-Maria von Hacht, Hamburg. Schallreuter, R.E.L., 1988a. Homeomorphy, phylogeny and natural classi¢cation: case studies involving Palaeozoic ostracods. In: Hanai, T., Ikeya, N., Ishizaki, K. (Eds.), Evolutionary Biology of Ostracoda. Proceedings of the Ninth International Symposium on Ostracoda. Kodansha, Tokyo, pp. 1041^1049. Schallreuter, R.E.L., 1988b. Ordovizische Ostrakoden Australiens. N. Jb. Geol. Pala«ontol. Monatsh. 9, 571^579. Schallreuter, R.E.L., 1988c. Ostrakoden und Geschiebeforschung. Geschiebe aktuell 4, 27^29. Schallreuter, R.E.L., 1993a. On Eopilla ingelorae Schallreuter gen. et sp. nov. Stereo-Atlas of Ostracod Shells 20, 117^120. Schallreuter, R.E.L., 1993b. On Eodominina nuela Schallreuter gen. et sp. nov. Stereo-Atlas of Ostracod Shells 20, 121^124. Schallreuter, R.E.L., 1993c. Ostrakoden aus ordovizischen Geschieben II. Beitra«ge zur Geschiebekunde Westfalens II. Geol. Pala«ontol. Westf. 27, 1^273. Schallreuter, R.E.L., 1994. On Garciana jaanussoni (Rossi de Garc|¤a and Proserpio). Stereo-Atlas of Ostracod Shells 21, 123^126. Schallreuter, R.E.L., 1995a. On Ansipe anseripediculus Schallreuter gen. et sp. nov. Stereo-Atlas of Ostracod Shells 22, 78^81. Schallreuter, R.E.L., 1995b. On Harpabollia argentina Schallreuter sp. nov. Stereo-Atlas of Ostracod Shells 22, 82^85. Schallreuter, R.E.L., 1995c. On Spinodiphores praepletus Schallreuter gen. et sp. nov. Stereo-Atlas of Ostracod Shells 22, 74^77. Schallreuter, R.E.L., 1996. Ordovizische Ostrakoden Argentiniens II. Mitt. Geol. Pala«ontol. Inst. Univ. Hamburg 79, 139^169. Schallreuter, R.E.L., 1998. On Eodominina ema Schallreuter sp. nov. Stereo-Atlas of Ostracod Shells 24, 13^16. Schallreuter, R.E.L., 1999. Eine neue Ostrakodenfauna aus dem Ordoviz Argentiniens. Greifswalder Geowissensch. Beitr. 6, 55^71. Schallreuter, R.E.L., Krufita, M., 1984. The Baltoscandian ostracode genus Piretella in the Ordovician of Bohemia. N. Jb. Geol. Pala«ontol. Monatsh. 11, 684^688.

Schallreuter, R.E.L., Siveter, D.J., 1985. Ostracodes across the Iapetus Ocean. Palaeontology 28, 577^598. Schallreuter, R.E.L., Krufita, M., 1988. Ordovician Ostracodes of Bohemia. Mitt. Geol. Pala«ontol. Inst. Univ. Hamburg 67, 99^199. Schallreuter, R.E.L., Siveter, D.J., 1988a. On Pilla piformis Schallreuter and Siveter gen. et sp. nov. Stereo-Atlas of Ostracod Shells 15, 25^28. Schallreuter, R.E.L., Siveter, D.J., 1988b. On Webbylla reticulata Schallreuter and Siveter sp. nov. Stereo-Atlas of Ostracod Shells 15, 21^24. Schallreuter, R.E.L., Krufita, M., 1994. Bohemian Ordovician ostracodes with relations to Britain. N. Jb. Geol. Pala«ontol. Monatsh 6, 361^367. Schallreuter, R.E.L., Hinz-Schallreuter, I., 1998. Ein Geschiebe aus Armorica im Thuringer Lederschiefer (Ordoviz). Arch. Geschiebekunde 2, 323^360. Schallreuter, R.E.L., Krufita, M., 2000a. Ostracodes from the Letna¤ Formation (Ordovician) of Bly¤skava (Bohemia). Acta Mus. Nat. Pragae Ser. B 56, 85^94. Schallreuter, R.E.L., Krufita, M., 2000b. Ostracodes from the Dobrotiva¤ Formation (Ordovician, Bohemia). Acta Mus. Nat. Pragae Ser. B 56, 95^103. Schallreuter, R.E.L., Krufita, M., Marek, L., 1996. Ordovician (Dobrotiva¤ Formation) ostracodes and trilobites from Ejpovice (Bohemia) and their relations to faunas of northern and western Europe. Pala«ontol. Z. 70, 439^460. Schallreuter, R.E.L., Verniers, J., de Geest, P., 2000. An Ordovician ostracode from Belgium. N. Jb. Geol. Pala«ontol. Monatsh. 9, 570^576. Sepkoski, J.J., 1981. A factor analytic description of the fossil record. Paleobiology 7, 36^53. Siveter, D.J., 1978. The Ordovician. In: Bate, R.H., Robinson, E. (Eds.), A Stratigraphical Index of British Ostracoda. Geol. J. Spec. Issue No. 8. Seel House Press, Liverpool, pp. 41^56. Siveter, D.J., 1984. Habitats and modes of life of Silurian ostracodes. In: Bassett, M.G. (Ed.), The autecology of Silurian organisms. Spec. Pap. Palaeontol. 32, 71^85. Siveter, D.J., in press. The Ordovician. A Stratigraphical Index of British Ostracoda. Br. Micropalaeontol. Soc. Siveter, D.J., Williams, M., 1997. Cambrian bradoriid and phosphatocopid arthropods of North America. Spec. Pap. Palaeontol. 57, 1^69. Siveter, D.J., Vannier, J.M.C., Palmer, D., 1991. Silurian Myodocopes: pioneer pelagic ostracods and the chronology of an ecological shift. J. Micropalaeontol. 10, 151^173. Soper, N.J., Woodcock, N.H., 1990. Silurian collision and sediment dispersal patterns in southern Britain. Geol. Mag. 127, 527^542. Soper, N.J., Strachan, R.A., Holdsworth, R.E., Gayer, R.A., Greiling, R.O., 1992. Sinistral transpression and the Silurian closure of Iapetus. J. Geol. Soc. London 149, 871^880. Spivey, R.C., 1939. Ostracodes from the Maquoketa Shale, Upper Ordovician, of Iowa. J. Paleontol. 13, 163^175. Swain, F.M., 1957. Early middle Ordovician Ostracoda of the eastern United States. Part I. Stratigraphic data and descrip-

PALAEO 3079 9-5-03

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228 tion of Leperditiidae, Aparchitidae and Leperditellidae. J. Paleontol. 31, 528^570. Swain, F.M., 1962. Early middle Ordovician Ostracoda of the eastern United States. Part II. Leperditellacea (Part), Hollinacea, Kloedenellacea, Bairdiacea and Superfamily Uncertain. J. Paleontol. 36, 719^744. Swain, F.M., 1987. Middle and upper Ordovician ostracoda of Minnesota and Iowa. In: Sloan, R.E. (Ed.), Middle and late Ordovician lithostratigraphy and biostratigraphy of the upper Mississippi Valley. Minn. Geol. Surv. Rep. Invest. 35, 99^130. Swain, F.M., Cornell, J.R., Hansen, D.L., 1961. Ostracoda of the families Aparchitidae, Aechminidae, Leperditellidae, Drepanellidae, Eurychilinidae and Punctaparchitidae from the Decorah Shale of Minnesota. J. Paleontol. 35, 345^372. Teichert, C., 1937a. Ordovician and Silurian faunas from Arctic Canada. Report of the Fifth Thule Expedition 1921-24. Gyldendalske Boghandel, Nordisk Forlag, Copenhagen, 169 pp., 24 pls. Teichert, C., 1937b. A new Ordovician fauna from Washington Land, North Greenland. Meddelelser om Gr]nland 119, 65 pp., 7 pls. Thomas, W.A., Astini, R.A., 1996. The Argentine Precordillera: a traveler from the Ouachita embayment of North American Laurentia. Science 273, 752^757. Tinn, O., Meidla, T., 1999. Ordovician ostracodes from the Komstad Limestone. Bull. Geol. Soc. Denmark 46, 25^30. Tinn, O., Meidla, T., 2001. Middle Ordovician ostracods from the Lanna and Holen Limestones, south-central Sweden. GFF 123, 129^136. Tinn, O., Meidla, T., in press. Phylogenetic relationships of early and middle Ordovician ostracodes of Baltoscandia. Palaeontology. Van Staal, C.R., Dewey, J.F., Mac Niocaill, C., McKerrow, W.S., 1998. The Cambrian-Silurian tectonic evolution of the northern Appalachians and British Caledonides: history of a complex, west and southwest Paci¢c-type segment of Iapetus. In: Blundell, D.J., Scott, A.C. (Eds.), Lyell: the past is the key to the present. Geol. Soc. Spec. Publ. 143, 199^242. Vannier, J.M.C., 1983. Rivillina, a new ostracode (Bradoriida?) genus from the Armoricain Massif, France. Alcheringa 7, 1^ 8. Vannier, J.M.C., 1986a. Ostracodes Binodicopa de l’Ordovicien (Arenig-Caradoc) Ibero-Armoricain. Palaeontographica A193, 77^143. Vannier, J.M.C., 1986b. Ostracodes Palaeocopa de l’Ordovicien (Arenig-Caradoc) Ibero-Armoricain. Palaeontographica A193, 145^218. Vannier, J.M.C., 1987. Le genre Ceratopsis (Ostracoda, Palaeocopa) dans l’Ordovicien de l’Europe et de l’Ame¤rique du Nord: phylogene'se, pale¤oe¤cologie et implications pale¤obioge¤ographiques. Geobios 20, 725^755. Vannier, J.M.C., 1990. Functional morphology and mode of life of Palaeozoic leiocope ostracodes. Lethaia 23, 103^ 112. Vannier, J.M.C., Vaslet, D., 1987. Ostracodes from the early Ordovician of central Saudi Arabia. Saudi Arabian Direc-

227

torate General of Mineral Resources, Open File Report BRGM-OF-07-31, 20 pp. Vannier, J.M.C., Abe, K., 1992. Recent and early Palaeozoic Myodocope ostracodes: functional morphology, phylogeny, distribution and lifestyles. Palaeontology 35, 485^517. Vannier, J.M.C., Siveter, D.J., Schallreuter, R.E.L., 1989. The composition and palaeogeographical signi¢cance of the Ordovician ostracode faunas of southern Britain, Baltoscandia, and Ibero-Armorica. Palaeontology 32, 163^222, pls. 24^30. Vannier, J.M.C., Wang, S.Q., Coen, M., 2001. Leperditicopid arthropods (Ordovician-Devonian): functional morphology and ecological range. J. Paleontol. 75, 75^95. Warshauer, S.M., Berdan, J.M., 1982. Palaeocopid and Podocopid Ostracoda from the Lexington Limestone and Clays Ferry Formation (Middle and Upper Ordovician) of central Kentucky. US Geol. Surv. Prof. Pap. 1066-H, H1^80, pls. 1^ 19. Webby, B.D., 1998. Steps towards a global standard for Ordovician stratigraphy. Newsl. Stratigr. 36, 1^33. Webby, B.D., Packham, G.H., 1982. Stratigraphy and regional setting of the Cliefden Caves Limestone Group (Late Ordovician), central-western New South Wales. J. Geol. Soc. Aust. 29, 297^317. Whatley, R., 2000. The zoogeographical signi¢cance of the marine ostracoda of Easter Island. Rev. Esp. Micropaleontol. 32, 147^156. Whatley, R., Jones, R., 1999. The marine podocopid Ostracoda of Easter Island: a paradox in zoogeography and evolution. Mar. Micropalaeontol. 37, 327^343. Whatley, R., Siveter, D.J., Boomer, I., 1993. Arthropoda (Crustacea: Ostracoda). In: Benton, M.J. (Ed.), The Fossil Record 2. Chapman and Hall, London, pp. 343^356. Wilde, P., 1991. Oceanography in the Ordovician. In: Barnes, C.R., Williams, S.H. (Eds.), Advances in Ordovician geology. Geol. Surv. Can. Pap. 90-9, 83^298. Williams, M., 1990. Ostracoda (Arthropoda) of the middle Ordovician Simpson Group, Oklahoma, USA. Thesis, Univ. Leicester, 192 pp. Williams, M., Siveter, D.J., 1989. On Balticella deckeri (Harris). Stereo-Atlas of Ostracod Shells 16, 94^99. Williams, M., Jones, P.J., 1990. On Eridoconcha simpsoni Harris. Stereo-Atlas of Ostracod Shells 17, 13^18. Williams, M., Vannier, J.M.C., 1993. On Kayina hybosa Harris. Stereo-Atlas of Ostracod Shells 20, 41^44. Williams, M., Vannier, J.M.C., 1995. Middle Ordovician Aparchitidae and Schmidtellidae: the signi¢cance of ‘featureless’ ostracodes. J. Micropalaeontol. 14, 7^24. Williams, M., Siveter, D.J., 1996. Lithofacies-in£uenced ostracod associations in the middle Ordovician Bromide Formation, Oklahoma, USA. J. Micropalaeontol. 15, 69^81. Williams, M., Siveter, D.J., 1998. British Cambrian and Tremadoc bradoriid and phosphatocopid arthropods. Monogr. Palaeontogr. Soc. London Publ. No. 607 (part of Vol. 152 for 1998), 49 pp., 6 pls. Williams, M., Floyd, J.D., 2000. Mid-Caradoc (Ordovician) ostracodes from the Craighead Limestone Formation, Girvan district, SW Scotland. Scot. J. Geol. 36, 51^60.

PALAEO 3079 9-5-03

228

M. Williams et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 195 (2003) 193^228

Williams, M., Floyd, J.D., Miller, C.G., Siveter, D.J., Stone, P., 2000. Kinnekullea comma (Jones, 1879), a transIapetus ostracod locum for the late Ordovician Dicellograptus anceps graptolite Biozone. J. Micropalaeontol. 19, 163^ 164. Williams, M., Floyd, J.D., Miller, C.G., Siveter, D. J., 2001a (for 2000). Scottish Ordovician ostracodes: a review of their

palaeoenvironmental, biostratigraphical and palaeobiogeographical signi¢cance. Trans. R. Soc. Edinb. Earth Sci. 91, 499^508. Williams, M., Stone, P., Siveter, D.J., Taylor, P., 2001b. Upper Ordovician ostracods from the Cautley district, northern England: Baltic and Laurentian a⁄nities. Geol. Mag. 138, 589^607.

PALAEO 3079 9-5-03