Fish/tetrapod Communities Examensarbete vid Institutionen för geovetenskaper in the Upper Devonian ISSN 1650-6553 Nr 301

Maxime Delgehier Fish/tetrapod Communities

Vertebrate communities including tetrapods and fishes are known from a in the Upper Devonian limited number of Late Devonian localities from several areas worldwide. These localities encompass a wide variety of environments, from true marine conditions of the near shore neritic province, to fluvial or lacustrine conditions. These localities form the foundation for a number of data matrices from which three different sets of cluster analyses were made. The first set practices a strait forward taxonomical framework using present/absent data on species and genus level to test similarity between the various localities. The second set of analyses builds on the Maxime Delgehier first one with the integration of artificial hierarchies to compensate taxonomical biases and instead infer relationship. The third also builds on the previous ones, but integrates morphological data as indicators of relationships between taxa. From this, a critical review was made for each method which comes to the conclusion that the first analysis and the first artificial level of the second analysis provide the distinctions between Frasnian and Frasnian/Famennian locality whereas the second artificial level of the second analysis and the third analysis need to be improved.

Uppsala universitet, Institutionen för geovetenskaper Examensarbete D/E1/E2/E, Geologi/Hydrologi/Naturgeografi /Paleobiologi, 15/30/45 hpISSN 1650-6553 Nr 301 Tryckt hos Institutionen för geovetenskaper, Geotryckeriet, Uppsala universitet, Uppsala, 2014. Examensarbete vid Institutionen för geovetenskaper ISSN 1650-6553 Nr 301

Fish/tetrapod Communities in the Upper Devonian

Maxime Delgehier

Supervisor: Henning Blom

Copyright Maxime Delgehier and the Department of the Earth Sciences Uppsala University Published at Department of Earth Sciences, Geotryckeriet Uppsala University, Uppsala, 2014

Abstract Vertebrate communities including tetrapods and fishes are known from a limited number of Late Devonian localities from several areas worldwide. These localities encompass a wide variety of environments, from true marine conditions of the near shore neritic province, to fluvial or lacustrine conditions. These localities form the foundation for a number of data matrices from which three different sets of cluster analyses were made. The first set practices a strait forward taxonomical framework using present/absent data on species and genus level to test similarity between the various localities. The second set of analyses builds on the first one with the integration of artificial hierarchies to compensate taxonomical biases and instead infer relationship. The third also builds on the previous ones, but integrates morphological data as indicators of relationships between taxa. From this, a critical review was made for each method which comes to the conclusion that the first analysis and the first artificial level of the second analysis provide the distinctions between Frasnian and Frasnian/Famennian locality whereas the second artificial level of the second analysis and the third analysis need to be improved.

Keywords: Palaeobiogeography, vertebrates, Tetrapods, Frasnian, Famennian

Populärvetenskaplig sammanfattning Samhällen som innehåller ryggradsdjur, vilken innefattar såväl fiskar som landryggradsdjur (så kallade tetrapoder), är kända från flera lokaler av sen devonsk ålder runt om i världen. Dessa representerar också flera olika miljöer, vilket gör dessa till en bra grund för vår förståelse över hur ekologin påverkade de tidigaste tetrapoderna. Sammansättningen av ryggradsdjur i dessa lokaler ligger till grund för en jämförande analys där tre olika klusteranalyser har använts. Den första analysen använder närvaro/frånvaro-data av traditionellt etablerade arter och släkten för att på så sätt etablera likheten mellan de olika faunorna. Den andra analysen använder samma grundmetod, men dör art och släkten har ersatts av artificiella hierarkier för att på så sätt kompensera för taxonomisk bias och samtidigt ta hänsyn till andra antaganden om släktskapsförhållanden. Den tredje analysen ersätter namn och hierarkier med morfologiska karaktärer för att på så sätt få en annan och taxonomiskt oberoende test av likheterna mellan faunorna hos de olika lokalerna. Dessa tre analyser kunde sedan jämföras, med resultatet och slutsatsen att traditionell taxonomisk jämförelse och till viss del även artificiella hierarkier, ger bättre och mer tolkningsvänliga resultat där studien visar på en tydlig uppdelning i tid mellan lokalerna, medan den tredje metoden bör utvecklas ytterligare.

Vertebrate communities including tetrapods and fishes are known from Late Devonian localities from several areas worldwide. These localities encompass a wide variety of environments. Their understanding is a key to understand ecologic pressure on these first tetrapods. The vertebrate communities from these localities are the base of data matrices from which three different cluster analyses were made. The first analysis is on forward taxonomical framework using present/absent data on the species and genus to test similarity between the various localities. The second builds on the previous one with the integration of artificial hierarchies to compensate taxonomical biases and instead infer relationship. The third analysis also builds on the previous ones, but integrates morphological data as indicators of relationships between taxa. From this, a critical review was made for each method which comes to the conclusion that the first analysis and the first artificial level of the second analysis provide the distinctions between Frasnian and Frasnian/Famennian localities whereas the second artificial level of the second analysis and the third analysis need to be improved.

Keywords: Palaeobiogeography, vertebrates, Tetrapods, Frasnian, Famennian

Table of Contents 1. Introduction ...... 1 2. Aim ...... 2 3. Background ...... 2 4. Methodology ...... 12 5. Results ...... 15 5.1 First analysis ...... 15 5.1.1 Species level ...... 15 5.1.2 Genus level ...... 17 5.2 Second analysis ...... 21 5.2.1 First artificial level ...... 21 5.2.2 Second artificial level ...... 24 5.3 Third analysis ...... 27 6. Discussion ...... 30 7. Conclusion ...... 35 8. Acknowledgements ...... 36 9. References ...... 36 Appendix ...... 47

1. Introduction

Vertebrate communities including tettrapods and fishes are known from Late Devonian localities from several areas worldwide. These localities, sometimes referred to as tetrapod localities, provide vertebrate assemblages constituted by major groups of vertebrates as agnathans, placoderms, acanthodians, chondrichthyans, actinopterygians and sarcopterygians. These sites show thus a wide diversity of vertebrates in various palaeoenviironmental conditions (see Background). The study of the vertebrate faunas is also essential to understaannd what could be the selection pressures on these first tetrapodss, which represent a key event in the vertebrate evolution. Our understanding of this event, which popularly known as the ‘fish-tetrapod transition’ has progressed greatly since the discovery of Ichthyostega in Greenland by Säve-Söderberggh in 1932 with recent discoveries. There are at present only a few Upper Devonian tetrapod localities known (Figure 1): three Frasnian localities (in Scotland, China and Latvia), one Frasnian-Famennian locality (Australia) and seven Famennnnian localities (in Greenland, USA, Belgium, Latvia and Russia). Today there are twelve tetrapod genera and fourteen species described and even more if includingg the yet undescribed taxa that have been found (Lebedev & Clack, 1993; Ahlberg et al., 1994; Shubin et al., 2004; Daeschler et al., 2009). In order to understand tetrapod evolution it is also important to include the near-tetrapods or elpistostegids. This is why the elpistostegid localities (Lode quarry in Latvia, Ellesmere Island in Cannada and Miguasha in Canada) (Figure 1) have been added to the list of tetrapod localities use in order to study the interrelaationship between localities and the evolution of faunas.

Figure 1: Map showing the near-tetrapod and tetrapod localities world-wide. Green star rreepresents Gogo station locality, yellow stars represent the elpistostegid localities; orange stars represent Frasnian tetrapod localities; red stars represent Famennian localities, brown stars represent Frasnian/Famennian localities.

1 Some studies (Young et al., 1992; Talent et al., 2000; Young et al., 2000; Lebedev & Zakharenko, 2010; Lebedev et al., 2010) were done to understand the global and or provincial relationships between vertebrates during the Devonian as their migrations, their ecology and more particularly on tetrapods (Lelièvre, 2002; Clack, 2006; Blieck et al., 2007). These studies suggest that there are two migration hypotheses for tetrapods: one northern and one southern. The northern one goes through central and southern Asian continental blocks and the southern one along the northern margin of Gondwana (Blieck et al., 2007), neither of which can be confirmed. 2. Aim

The aim of my study is to explore the faunal relationships between different Late Devonian tetrapod yielding localities worldwide. This will be done using several numerical methods using similarity indices, including cluster analysis. Two new methods, based on artificial hierarchies and on morphological data, will be tested. 3. Background

Since the discovery of the first specimen of the well-known tetrapod, Ichthyostega Save- Söderbergh, 1932 in Eastern Greenland (Gauss Halvø and Ymer Ø sites (Figure 2)) more and more tetrapod localities were found world-wide. East Greenland locality have provided a wide diversity of Famennian vertebrates (Blom et al., 2007), including placoderms, chondrichthyans, acanthodians, an actinopterygian, lungfishes, porolepiforms, osteolepiforms and four tetrapod species. The most complete tetrapods are Acanthostega gunnari Jarvik, 1952, Ichthyostega stensioei Säve-Söderbergh, 1932 Ichthyostega watsoni Säve-Söderbergh, 1932 and Ichthyostega eigili Säve-Söderbergh, 1932, while Ymeria denticulata Clack et al., 2012 is recognized based on jaw and tooth morphology only. All the tetrapods come from Famennian sediments of the Celsius Bjerg Group interpreted as being primarily of fluvial origin (Olsen & Larsen, 1993).

2

Figure 2: Map of East Greenland showing the Devonian outcrops (Blom et al., 2007).

Tetrapods are also found in other the localities around the world, including Belgium, USA, Latvia, Russia, Australia, China and Scotland.

The Strud locality in Namur Province, Belgium (Figure 3) provided a parrtt of a lower jaw identified as Ichthyostega-like (Clément et al., 2004) belonging to the Evieux formation. With this discovery the original quarry, which was abanndoned for one century, is now being studied again.

Figure 3: Map showing the localization of Strud locality in Belgium (Gueriau et al., 2014).

The Langlier quarry fossils, near Durnal (Figure 4) were added to the taxa list of Strud because these fossils were found in the same formation (Evieux formation) with sediments which suggest a mixed

3 alluvio-lagoonal environment with estuary or tidal delta influences and Strud sediimments as a deltaic environment (Goemaere, 1995; Thorez et al., 2006). These localities provide remains of placoderms (bothriolepids, phyllolepids, groenlandaspids), holoptychiid porolepiforms, rhynchhodipterid lungfishes (Clément & Boisvert, 2006), tetrapodomorph fishes in Strud and acanthodians (Deerycke & Clément, 2013) and tetrapodomorph fishes (Clément, 2002; Clément et al., 2009).

Figure 4: Map showing the localization of Durnal locality in Belgium (Clémentt et al., 2009).

Several tetrapods have also been fouunnd in Red Hill, Pennsylvania (Figuree 5) including Densignathus rowei Daeshler, 2000 based on a lower jaw, Hynerpeton bassetti Daeschler et al., 1994 based on a left cleithrum and a posterior part of a right lower jaw and a possible watcherid (Shubin et al., 2004; Daeschler et al., 2009). In association with these tetrapods are also vertebrate remains belonging to placoderms, chondrichthyans, accanthodians, actinopterygians, lungfiishes and osteolepiforms (Cressler III et al., 2010). These fossils belong to the Catskill Formation which is interpreted as the result of a deltaic environment during the Famennian.

4

Figure 5: Map showing the location of the Red Hill site, Clinton County, Pennsylvania, USA. (Cressler III et al., 2010)

Andreyevka in Tula region, Russia (Figure 6) provide abundant Famennian vertebrate remains of bones, scales and teeth belonging to sarcopterygians, acanthodians, actinopterygians, antiarch placoderms (Alekseev et al., 1994). The skeleton of the tetrapod Tulerpeton curtum Lebedev, 1984 and a possible second undetermined tetrapod (Lebedev & Clack, 1993) were found in limestones which are interpreted as a marine environment.

Figure 6: Map showing the localization of Andreyevka (A) Location of the Tula region in the European part of Russiia; (B) Location of Andreyevka locality in the Tula region; (C) Andreyevka localiity, the black circles representn the position of outcrop. (Alekseev et al., 1994)

5 There are two tetrapod localities in Latvia dating to the Famennian: Pavari and Ketleri (Figure 7). These sites belong to the Ketleri Formation constituted by sandstones interpretted as marine deposits (Luksevics & Zupins, 2004). Both sites are characterized by rich vertebrate assemblages which belong to the same fauna and includes placoderms, acanthodians, porolepiforms, lungfishes, osteolepiforms and tetrapods: Ventastega curonica Ahlberg et al., 1994 on the basis of mandibular fragment and a possible undetermined tetrapod (Ahlberg et al., 1994).

Figure 7: Maps showing the localization of the Latvian localities (a) Map of Latvia; thee area of light shading indicates the extent of Famennian and later deposits, the square of dark shading is the area of the next map. (b) Detail map of the Skrunda area showing the localities of Pavari and Ketleri. The house indicate Pavari and Ketleri hamlets, the stylized fishes mark the fossil localities. Crosses, Zagare Formation; no shading, Ketleri Formation; light shading, Skervelis Formation. The fossil locality at Ketleri lies just below the top of the Ketleri Formation. Scale bar 3 km. (Ahlberg et al., 1994)

6 In the Oryol region (Russia) near the town of Livny, there is the Gornastayevka quarry (Figure 8). This quarry provided the tetrapodd Jakubsonia livnensis Lebedev, 20044 and vertebrate remains, such as placoderms, acanthodians, chondrichthyans, porolepiforms, lungfishes, osteolepids and megalichtyds. These fossils were found in Famennian sedimentary deposits constituted by sandstone and clay interpreted as a deltaic to near-shore environment.

Figure 8: Map showing the localization of thhe locality in Russia, an asterisk shows the location of the Gornostayevka quarry in the Oryol region, a bllack diamond marked the location of Annddreyevka. (Lebedev, 2004)

A poorly preserved right lower jaw of a tetrapod have been found in Jemalong site (Figure 9) and described as belonging to Meetaxygnathus denticulus Campbell & Bell, 1977 from the Cloghnan Shale. The fish fauna of placodermr s and lungfishes is known from this locality and have suggested a fluvial environment, late Frasnian or early Frasnian in age (Campbell & Bell, 1977). This locality is rather close to the famous Canowindra site, which present no tetrapod but show a fauna constituted by placoderms, lungfish, rhizodonts and tristichopteryds. The vertebrates from Canowindra were found near the base of the Hervey Group with evidence of a marine incursion dating to Frasnian. The tetrapod-fish assemblage was assigned to these two fauna called ‘Jemalong-Canowindra’ by Young (1993) which could be with the previous consideration as a Frasnian or Famennian age. There are

7 probably more tetrapods from Australia and especially in Victoria where trackways have been discovered (Clack, 1997).

Figure 9: Map of Australia showing the localization of Jemalong and Canowindra localitty, fossil site indicated by triangles (from Johanson et al., 2003 modified).

Near Elgin in Scotland (Figure 10), the Scat Craig locality provided bonees of Elginerpeton pancheni Ahlberg et al., 1995 with a fauna constituted by porolepiforms, lungfishes, placoderms and heterostracans. The dating of this fauna due to the presence of heterostracans is Frasnian (Gross, 1950). This vertebrate assemblage was found in sandstones, these deposits were interpreted as coming from a fluvial channel.

Figure 10: Map showing the location of Elgin in Scotland (Ahlberg, 1998).

8 Velna-Ala (Figure 11) in Latvia is the locality where Obruchevichthys grraacilis Vorobyeva 1977 was found in Ogre Formation. This formation is composed of a siliciclastic sequence of sandstone, siltstone, dolomitic marl, clay and gypsum, which are interpreted to bee formed in the shallow waters of a narrow gulf of the Baltic palaeobasin by Sorokin (1978). Thiss is a relatively poor in vertebrate remains and the fauna provides some specimens of heterostracans, pllacoderms, acanthodians, actinistians, porolepiforms, lungfishes and tristichopteryds. The presence of heterostracans has dated the Ogre Formation to Frasnian in age. In western Russia the same tetrapod species was found but the locality is not precisely known and could be in the Novvggorod district.

Figure 11: A, B Localization of Langsede site in Europe and in Latvia; C Star indicate location of Langsede site near Sabile (Luksevics et al., 2011).

The Frasnian sediments of Ningxia Hui autonomous region in northwestern China (Figure 12) have provided only one poorly preserved tetrrapod lower jaw named Sinostega pani Zhu et al., 2002 with a diverse fauna of vertebrates. The precise age of the Zhongning Formation, where these fossils were found, is still problematic to date (Blieck et al., 2007) and needs further biostratigraphic research (Jia et al., 2010). The depositional environment of this locality is interpreted as non-marine.

9

Figure 12: (A) Location of Ningxia Hui Autonomous Region; (B) Devonian outcrops in Ningxia; (C) Correlation of the Zhongning Formation in four selected Devonian sections mainly based on Pan ett al., 1987 (Jia et al., 2010).

In order to understand the interrelationships between these tetrapod localiities, it is important and relevant to include to the localities where elpistostegid (also called near-tetrapods) have been found. Elpistostegid localities are composed by Lode (Latvia), Ellesmere (Canada) and Miguasha (Canadaa). All of these sites are Frasnian.

Lode quarry (Figure 13) provide a fauna composed by agnathans, placoderms, acanthodians, actinopterygians, actinistians, lungfishes, osteolepiforms (Upeniece, 2001) and Panderichthys rhombolepis Gross, 1941 from the top part of Lode Formation. Also the precise dating still problemmatic (Figure 13). The sediments of the quarry were interpreted as deltaic environment.

10

Figure 13: Map showing the localization of Lode quarry in Latvia and the stratigraphic column of the Lode Formation (there is uncertainty about the position of the Givetian-Frasnian boundary) ((Forey et al., 2000).

The Ellesmere Island is in the Canadian Arctic (Figure 14). Tiktaalik roseae Daeschler et al., 2006 was found here in one locality with other fish remains such as antiarch placoderms, holoptychiid porolepitforms (Downs et al., 2011), lungfishes and osteolepiforms. The come froom the Fram Formation which is characterized by alternating resistant sandstones and recessive siltstones, and interpreted as the deposits of meandering stream systems (Embry & Klovan, 1976; Embry, 1991).

Figure 14: Location and stratigraphic position of the site (NV2K17) on southern Ellesmere Island, Nunavut Territory, Canada. (Daeschler et al., 2006)

11 Miguasha (Figure 15), in Québec Province (Canada), represents one of the best known Late Devonian faunas in the world due to an exceptional conservation (Janvier & Arsenault, 2009). This fauna was found in the Escumiac Formation and includes anaspids, osteostracans,, placoderms, acanthodians, actinopterygians, actinistians, porolepiforms, lungfishes, osteolepiffoorms (Cloutier et al., 2011) and Elpistostege watsoni Westoll 1938. The depositional environment of this site was interpreted as estuarine (Cloutier et al., 2011).

Figure 15: a. Localization of Miguasha in eastern Québec, b. Geological map of the Miguasha area (Miguasha Group includes the Fleurant and Escumiac Formations), c. General view of the east sidee of the syncline of the Escumiac Formation (see in b) (Cloutier et al., 2011).

For the present analyses, Gauss Halvø and Ymer ø localities are referred to as the Greenland locality, Strud as the Belgium locality, Ketleri and Pavari as Pavari, Velna-Ala as Ogre and Ningxia Hui as China. Other names are used unchanged. The complete list of taxa present in each locality is in the Appendix. 4. Methodology

The methodological core of this study is to use several sets of analyses inn order to explore the faunal relationships and similarity between the different tetrapod localities worldwide. For this I have done cluster analysis on several data: the first based on taxonomical framework, the second on

12 artificial hierarchies of the previous taxa and the third on characters from taxa. Cluster analyses were selected, because they are simple to use and provide results which are rather strait forward to analysis. Cluster analyses using the Jaccard similarity index have been used for each database to explore the faunal relationships between localities. Other indices were explored but they don’t provide conclusive results. This is why this index was preferred because it emphasizes datasets which have only few common data (Hammer et al. 2001). PAST version 1.56b (Hammer et al. 2001) was used to perform the cluster analyses.

To do these analyses, the first step was to identify relevant Late Devonian tetrapod localities. In that case, there are ten localities (see Background part), but in order to expand the list of localities and include a broader definition of tetrapods, near-tetrapod localities were added. Thus three localities were added: Ellesmere, Lode and Miguasha. Canowindra (Australia) was also added because of the proximity of this locality with Jemalong, these sites are sometimes considered as having the same fauna (Young, 1993). In that respect, each analysis was done with these two localities separated and grouped called JeCa. The whole of localities is considered as the ingroup. Finally the reference locality without tetrapods or elpistostegids is the Gogo station (Western Australia) where an exceptional preservation (Long & Trinajstic, 2010) provided a great number of species of placoderms, actinopterygians, lungfishes and sarcopterygians.

The basis for all analyses is to use presence or absence of taxa to compare the localities. The first step is the listing of taxa for each and one of the localities. Some localities as Miguasha or Gogo provide lot of Frasnian fossils which give a good picture of the vertebrate fauna during the Late Devonian, but others such as Ellesmere record only some vertebrates or contain still undescribed material. These taxa are then placed in the matrix and subdivided into the different vertebrate groups: Anaspida, Osteostraci, Heterostraci, Galeaspida, Placodermi, Chondrichthyes, Acanthodii, Actinopterygii, Actinistia, Porolepiformes, Dipnoi, Osteolepiforms, Elpistostegalia and Tetrapoda (taxa list in Appendix part). Finally the matrix is coded: absent taxa as 0 and present taxa as 1. Then the matrix is run with the software PAST version 1.56b (Hammer et al. 2001) in order to get dendrograms which will be described and interpreted to understand the faunal relationships.

The second analysis is based on artificial hierarchies of taxa, which is a new method based on assumed interrelationships from available phylogenies rather than traditional taxonomical hierarchies. There are two levels of hierarchies: the first level could be assimilated to a classical family level, the second level which group some of first level taxa could be assimilated to a sub-order or super-family level. Also each localities of the previous analysis are used for this analysis. Taxa for which there are not phylogenetic analyses or which were not included in phylogenetic analyses, the current family level is used. A taxon not described or an indeterminate taxon cannot be grouped with other taxa. In that respect these taxa get a level for them. In the case of the indeterminate taxon Acanthodii indet. A, the artificial level 1 is Acanthodii 7 and the artificial level 2 is Acanthodii C (see the list of taxa). The

13 class Placodermi was divided as in several part, in that respect each group get is own grouping: Antiarcha grouping was created from phylogenetic analysis of Moloshnikov, 2008; Arthrodira grouping was created from phylogenetic analysis of Carr & Hlavin, 2010 and Dupret & Zhu, 2007 and Phlyctaenii grouping was created from phyloogenetic analysis of Dupret & Zhu, 2007. The other class or sub-class were divided as previously: taxa belonging to Acanthodii were groupped according to phylogenetic analysis from Hanke & Davis, 2012; actinopterygian taxa were grouped according to phylogenetic analysis from Choo, 2011; Porolepiformes and Dipnoi were grouped according to phylogenetic analysis from Ahlberg et al., 2006 and Clément & Long, 2010; Osteeolepiformes (Osteolepididae, Megalichthydae, Tristichopteryidae) were grouped according to pphylogenetic analysis from Snitting, 2008; elpistostegids and tetrapods were grouped according to phylogenetic analysis from Clack et al., 2012 (the list of grouping is available in the Appendix part). So the second step, the coding of the matrix with the same method as the previous analysis: a present taxon is coded 1 and an absent taxon is coded 0. The second level reuses the same principle as thhe first level but the taxa of the first level are grouped. So the method used can be resumed as the example of the Figure 16.

Figure 16: tree showing the grouping of taxa in order to get the taxa of the second analyses.

The third analysis is also a new method and uses a morphological comparison between taxa of each clade to compare localities. In this case morphological characters of clades replace taxa. The first step to prepare this analysis is the collect of characters for each clade (character sets used are in the Appendix part). These characters can belong to a phylogenetic analysis as it is the case for arthrodires placoderms (Carr & Hlavin, 2010), actinopterrygians (Choo 2011), lungfishes (Ahhlberg et al., 2006), osteolepiforms (Snitting, 2008), elpistostegids and tetrapods (Clack et al., 2012).When it is not possible, characters, which can belong to diagnostic traits present in the description of species, can be used as it is the case for antiarch placoderms, phyllolepids, groenlandaspids. Characters from phylogenetic analyses should be in the same state for taxa from a same locality which provide several taxa (e.g. state 1 of a character X for all tetrapod taxa from the Greenland locality)). This step is the most important because some characters from phylogenetic analyses or diagnostic traits could be convergences between taxa difficult to distinguish. Moreover, there are taxa not present in some localities so these localities are coded as 0. Caution should be taken with not-welll supported phylogenetic analysis. In the particular case of this analysis, Gogo is not the reference locality but this

14 is Ellesmere which have the greater number of characters in plesiomorphic state. This is explained by the few number of taxa present and whose which are present (Asterolepis spp., Holoptychius, Laccognathus, Glyptolepis and Tiktaalik) show characters in plesiomorphic states. 5. Results 5.1 First analysis 5.1.1 Species level In this first cluster analysis (Figure 17), in which Jemalong and Canowindra are treated as separate localities, presence/absence data for 284 species resulted in very low similarity indices between the various localities. This is most likely due to the low number of common taxa. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Pavari, Ogre, Oryol and Andreyevka. These localities from this cluster come from Latvia and Western Russia. So this cluster can be called the Baltica cluster. Ogre is the only locality of the cluster which is Frasnian.

This cluster is also grouping with another distinct cluster, including Greenland, Jemalong, RedHill and Belgium. Jemalong is present in this cluster because some taxa as Bothriolepis and Phyllolepis species from this locality are only present in Famennian localities of the Northern Hemisphere whereas Jemalong is Late Frasnian to Early Famennian in age. Except this locality, Greenland, RedHill and Belgium belong to Laurussia this is why this cluster can be considered as a Laurussia cluster.

There is one additional cluster showing some similarity between Ellesmere and Lode which are two near-tetrapod localities. Only one porolepiform is common to these localities. This cluster shows no similarity with any of the other localities, including the Miguasha locality.

Other localities (Miguasha, China, ScatCraig and Canowindra) show such low similarity index, considered null. This could be explained by the endemism and the exceptional conservation in Miguasha and by a low number of taxa from China, ScatCraig and Canowindra localities.

15 Canowindra Ellesmere Lode Andreyvka Pavari Oryol Ogre Greenland Jemalong RedHill Belgium Miguasha Gogo China ScatCraig 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4 16

Figure 17: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the species level when Canowindra and Jemalong localities are considered as separated.

In the cluster analysis (Figure 18), in which Jemalong and Canowindra are treated as the same localities called JeCa, presence/absence data for 284 species also resulted in very low similarity indices between the various localities. This is again most likely due to the low number of common taxa. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Pavari, Ogre, Oryol and Andreyevka. This cluster can be identified as the Baltica cluster because these localities coming from Latvia and Western Russia. All localities are Famennian except Ogre

This cluster is also grouping with another distinct cluster, including RedHill, Belgium, Greenland and JeCa. JeCa are Late Frasnian to Early Famennian in age and present some taxa as Bothriolepis and Phyllolepis species. These taxa are only present in Famennian localities of the Northern Hemisphere whereas JeCa is Late Frasnian to Early Famennian in age. This cluster can be

16 considered as a Laurussia cluster because the three other localities come from the Northern Hemisphere.

An additional cluster, showing some faunal similarity between Ellesmere and Lode, is present. These near-tetrapod localities show only one porolepiform species common taxa. This cluster shows no similarity with any of the other localities, including the Miguasha locality.

Other localities (Miguasha, China and ScatCraig) show such low similarity index, considered null. This could be explained by endemism and the exceptional conservation in Miguasha and by a low number of taxa from China and ScatCraig localities. Ellesmere Lode Greenland RedHill Belgium JeCa Pavari Oryol Ogre Andreyvka Miguasha Gogo China ScatCraig 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4

Figure 18: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the species level when Canowindra and Jemalong localities are considered as grouped.

5.1.2 Genus level In this first cluster analysis (Figure 19), in which Jemalong and Canowindra are treated as separate localities, presence/absence data for 175 genera resulted high similarity indices between the

17 various localities. This is most likely due to the increase number of common taxa which is explained by the genus level: a higher level gives by default a higher similarity index. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Pavari, Ogre, Oryol, Andreyevka and ScatCraig. Except ScatCraig, these localities from this cluster come from Latvia and Western Russia. So this cluster can be called the Baltica cluster. There are two Frasnian localities which are Ogre and ScatCraig. This latter occupy a basal position in the cluster. This position could be explained by the number of taxa from this site and only few taxa (Holoptychius, Concochus) are common with other localities of the Baltica cluster.

This cluster is also grouping with another distinct cluster, including Greenland, Belgium Jemalong, Canowindra and RedHill. Jemalong and Canowindra are the closest localities with nearly the totality of taxa in common certainly due to the geographical proximity (see Background part). Jemalong and Canowindra are present in this cluster because some taxa as Bothriolepis species from these locality are only present in Famennian localities of the Northern Hemisphere whereas there two localities are Late Frasnian to Early Famennian in age. Except this locality, Greenland, RedHill and Belgium belong to Laurussia this is why this cluster can be considered as a Laurussia cluster. These three Laurussian localities show the presence of Groenlandaspis, Phyllolepis, Soederberghia and Eusthernodon as common taxa. Greenland and Belgium are close and the common taxa Ichthyostega which is the only tetrapod which is not considered restricted to a single locality.

There is on additional cluster showing some similarity between Ellesmere, Lode and China. Common taxa are represented by Asterolepis, Laccognahus and Glyptolepis. The China locality falls to the base of this cluster because of endemic taxa as antiarchs (Jingxilepis, Ninjxialepis and Sinolepis) and other taxa. This cluster shows no similarity with any of the other localities, including the Miguasha locality. The last localities Miguasha show such low similarity index, considered null.

18 Miguasha China Ellesmere Lode ScatCraig Ogre Andreyvka Pavari Oryol Greenland Belgium Canowindra Jemalong RedHill Gogo 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4 16

Figure 19: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the genus level when Canowindra and Jemalong localities are considered as separated.

In this second cluster analysis (Figure 20), in which Jemalong and Canowindra are treated as grouped localities, presence/absence data for 175 genera resulted higher similarity indices between the various localities. This is most likely due to the increase number of common taxa which is explained by the genus level: a higher level gives by default a higher similarity index. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Pavari, Ogre, Oryol and Andreyevka. All these localities come from this cluster come from Latvia and Western Russia. So this cluster can be called the Baltica cluster. Ogre is the only locality which is Frasnian. This cluster show common taxa as Devononchus, Holoptychius, Glyptolepis, Cryptolepis, and Glyptopomus.

This cluster is also grouping with another distinct cluster, including Greenland, Belgium JeCa and RedHill. JeCa are present in this cluster because some taxa as Bothriolepis and Phyllolepis species

19 from these locality are only present in Famennian localities of the Northern Hemisphere whereas there two localities are Late Frasnian to Early Famennian in age. Except this locality, Greenland, RedHill and Belgium belong to Laurussia this is why this cluster can be considered as a Laurussia cluster. These three Laurussian localities show the presence of Groenlandaspis, Phyllolepis, Soederberghia and Eusthernodon as common taxa. Greenland and Belgium are close and the common taxa Ichthyostega which is the only tetrapod which is not considered restricted to a single locality. Then, there is the ScatCraig locality which is not grouped with another locality.

There is on additional cluster showing some similarity between Ellesmere, Lode and China. Common taxa are represented by Asterolepis, Laccognahus and Glyptolepis. The China locality falls to the base of this cluster because of endemic taxa as antiarchs (Jingxilepis, Ninjxialepis and Sinolepis) and other taxa. This cluster shows no similarity with any of the other localities, including the Miguasha locality. The last locality, Miguasha, is in the basal position. Miguasha China Ellesmere Lode ScatCraig Greenland Belgium JeCa RedHill Ogre Andreyvka Pavari Oryol Gogo 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4

20 Figure 20: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the genus level when Canowindra and Jemalong localities are considered as grouped. 5.2 Second analysis 5.2.1 First artificial level In this first cluster analysis (Figure 21), in which Jemalong and Canowindra are treated as separate localities, presence/absence data for 126 artificial grouping (first level) resulted higher similarity indices between the various localities. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Pavari, Greenland, Oryol and Andreyevka. All the localities are Famennian. This first cluster shows several common taxa as Acanthodii 1, Chondrichthyes 5, Porolepiformes 1, Dipnoi 2 and 3, Onychodontida 2, Osteolepididae 1 and Tristichopterydae 1. Two sister-groups are presents: an Oryol and Pavari grouping and an Andreyevka and Greenland grouping.

This cluster is also grouping with another distinct cluster, including Canowindra, Jemalong, ScatCraig, RedHill and Belgium. There are ScatCraig and a grouping Canowindra and Jemalong. These two localities are the closest localities with nearly the totality of taxa in common. The rest of the cluster is composed by RedHill and Belgium localities are still together. This cluster is due to some common taxa as Phlyctaenii 1, Phyllolepida, Dipnoi 3, 11, Osteolepididae 1 Megalichthydae 1 Tristichopterydae 1.

There is on additional cluster showing some similarity between Ellesmere, Lode, China, Ogre and Miguasha. The Miguasha locality falls at the base of this cluster certainly because of the presence of endemic taxa from this locality. Then there two grouping with Antiarcha 1 as common taxa: Ellesmere and Lode group and Ogre and China.

21 Miguasha Ellesmere Lode Ogre China Greenland Andreyvka Pavari Oryol RedHill Belgium ScatCraig Canowindra Jemalong Gogo 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4 16

Figure 21: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the first artificial level when Canowindra and Jemalong localities are considered as separated.

In this second cluster analysis (Figure 22), in which Jemalong and Canowindra are treated as grouped localities, presence/absence data for 126 artificial grouping (first level) resulted higher similarity indices between the various localities. Of course, a higher level gives by default a higher similarity index. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Pavari, Oryol and Andreyevka. All the localities are Famennian and come from Western Russia and Latvia. This is why this cluster can be called a Baltica cluster. This cluster shows several common taxa as Antiarcha 2 Acanthodii 1, Porolepiformes 2 and Osteolipiformes 1. Andreyevka fall at the base of this group but this group shows always the strong faunal connection between its members.

22 This cluster is also grouping with another distinct cluster, including JeCa, Greenland, ScatCraig, RedHill and Belgium. The ScatCraig is in a basal position due to Heterostraci taxa. The rest of the cluster is defined by taxa as Antiarcha 2, Phyllolepida, Dipnoi 1, 3 and Tristichopterydae 1. Then, there are two grouping: the Greenland and Jeca localities are together as RedHill and Belgium.

There is on additional cluster showing some similarity between Ellesmere, Lode, China, Ogre and Miguasha. The Miguasha locality falls at the base of this cluster certainly because of the presence of endemic taxa from this locality. Then there two grouping with Antiarcha 1 as common taxa: Ellesmere and Lode group and Ogre and China. Miguasha Ellesmere Lode Ogre China Greenland JeCa RedHill Belgium ScatCraig Andreyvka Pavari Oryol Gogo 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4

Figure 22: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the first artificial level when Canowindra and Jemalong localities are considered as grouped.

23 5.2.2 Second artificial level In this first cluster analysis (Figure 23), in which Jemalong and Canowindra are treated as separate localities, presence/absence data for 90 artificial grouping (second level) resulted higher similarity indices between the various localities. Of course, a higher level gives by default a higher similarity index. With this level of grouping, Gogo fall in the ingroup. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Canowindra, Jemalong and ScatCraig. This locality falls at the base of the cluster. The rest of the cluster is composed by the Canowindra and Jemalong localities which are the closest localities with nearly the totality of taxa in common.

This cluster is also grouping with another distinct cluster, including Greenland, Andreyevka, RedHill, Belgium and Ogre. This Ogre locality fall at the base because of the presence of taxa Heterostraci C, Antiarcha C, Dipnoi D and Tetrapoda K. The rest of the cluster is composed by two grouping: the Greenland and Andreyevka localities are together as the RedHill and Belgium sites. This last part of the cluster is due to common taxa: Acanthodii B, Chondrichthyes D, Actinopterygii A and Dipnoi A. Then, there is the Pavari locality.

A cluster which shows similarity between Gogo and Oryol is also present. There are two additional clusters showing some similarity between Ellesmere, Lode, China and Miguasha. This cluster is due to some taxa as Antiarcha A and B, Acanthodii B, Actinopterygii D, Porolepiformes A, Tristichopterydae B. Then, there are two grouping: Ellesmere and Lode group and Ogre and China. Moreover these localities should be considered as the outgroup for this analysis.

24 China Ellesmere Miguasha Lode Oryol Gogo Ogre Greenland Andreyvka RedHill Belgium ScatCraig Canowindra Jemalong Pavari 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4 16

Figure 23: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the second artificial level when Canowindra and Jemalong localities are considered as separated.

In this second cluster analysis (Figure 24), in which Jemalong and Canowindra are treated as grouped localities, presence/absence data for 90 artificial grouping (second level) resulted higher similarity indices between the various localities. With this level of grouping, Gogo fall in the ingroup. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Greenland, Andreyevka, JeCa, RedHill and Belgium. This cluster is supported by Antiarcha A and B, Phlyctaenii, Phyllolepida Acanthodii A B, Chondrichthyes F, Actinopterygii A, Dipnoi A, B and P and Tristichopterydae A. It is possible to distinguish two grouping: the RedHill and Belgium locality are together whereas the other grouping is composed by Andreyevka, Greenland and JeCa. This last one falls at the base of the grouping.

25 This cluster is also grouping with another distinct cluster, including Greenland, Andreyevka, RedHill, Belgium and Ogre. This Ogre locality fall at the base because of the presence of taxa Heterostraci C, Antiarcha C, Dipnoi D and Tetrapoda K. The rest of the cluster is composed by two grouping: the Greenland and Andreyevka localities are together as the RedHill and Belgium sites. This last part of the cluster is due to common taxa: Acanthodii B, Chondrichthyes D, Actinopterygii A and Dipnoi A. Then, there are the Pavari, ScatCraig and Ogre localities which are not grouped with other localities.

The next cluster has as common taxa Arthrodira B, Ptyctodontida, Dipnoi E and G and Osteolepididae A. So, this cluster shows similarity between Gogo and Oryol. There is on additional cluster showing some similarity between Lode and Miguasha. This cluster shows some taxa in common as Actinopterygii D, Actinistia and Tristichopterydae B which could explain the position of these localities in the basal position.

Then, there is a cluster which shows similarity between the Ellesmere and China localities. Moreover these localities should be considered as the outgroup for this analysis.

26 Miguasha Lode Oryol Gogo Ogre Greenland Andreyvka JeCa RedHill Belgium ScatCraig Pavari China Ellesmere 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4

Figure 24: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the second artificial level when Canowindra and Jemalong localities are considered as grouped. 5.3 Third analysis Ellesmere becomes the outgroup locality in this analysis because of the great number of plesiomorphic characters (Figure 25 and 26).

In this first cluster analysis (Figure 25), in which Jemalong and Canowindra are treated as separate localities, characters from phylogenies are used. With this analysis, Ellesmere becomes the outgroup locality due to the number of characters coded as 0. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Belgium, RedHill, Andreyevka, Greenland and Oryol supported by characters 1, 2, 5 and 6 from Antiarcha, 30, 33, 35, 36, 42 and 73 from Dipnoi, 8, 18, 22, 46, 47, 60 and 61 from Tristichopterydae and 50, 74, 80, 81, 92

27 and 96 from Elpistostegalia/Tetrapoda.. At the base of this cluster, the Oryol locality is present. The rest of the cluster is supported by some common characters as 44, 70and 72 from Actinopterygii, 70 from Dipnoi, 6, 9, 11, 34, 50, 51 and 52 from Tristichopterydae and 36, 50, 68, 73, 111 and 112 from Elpistostegalia/Tetrapoda. Then, there are Greenland and Andreyevka which are not grouped with other localities. The grouping composed by Belgium and RedHill is supported by the character 2 from Phlyctaenii which is not present in the other group. This locality falls at the base of the cluster. The rest of the cluster is composed by the Canowindra and Jemalong localities which are the closest localities with nearly the totality of taxa in common.

This cluster is also grouping with another distinct cluster, including Lode, Miguasha, China, Canowindra and Ogre. A cluster with Lode and Miguasha is present and supported by the characters 9, 23, 31, 58, 70, 72 and 76 from Actinopterygii .The characters 1, 2, 5 and 6 from Antiarcha, 6, 9, 11, 22, 46, 47 50, 51 ,52, 60 and 61 from Tristichopterydae support the rest of this cluster is composed by China, Canowindra and Ogre. The China locality fall at the base of this group whereas the grouping Canowindra-Ogre is supported by characters 2 from Antiarcha, 30, 33, 35, 36, 42, 73 from Dipnoi and 8, 18, 34 from Tristichopterydae.

A cluster which shows similarity between Gogo and Oryol is also present. There are two additional clusters showing some similarity between Ellesmere, Lode, China and Miguasha. This cluster is due to some taxa as Antiarcha A and B, Acanthodii B, Actinopterygii D, Porolepiformes A, Tristichopterydae B.

The ScatCraig locality is in the basal position of the cluster composed by two other localities: Jemalong and Pavari. This grouping is supported by characters 1 to 3 and 4 to 5 from Antiarcha, 30, 33, 35, 36 and 73 from Dipnoi, 50, 74, 77, 81 and 112 from Elpistostegalia/Tetrapoda.

Finally, the Gogo locality is in the basal position because of all characters are absent except characters coming from Arthrodira, Actinopterygii and Dipnoi and three characters from Osteolepididae.

28 Ellesmere Gogo ScatCraig Pavari Jemalong Greenland Andreyvka RedHill Belgium Oryol Ogre Canowindra China Miguasha Lode 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4 16

Figure 25: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the morphological level when Canowindra and Jemalong localities are considered as separated.

In this second cluster analysis (Figure 26), in which Jemalong and Canowindra are treated as grouped localities, characters from phylogenies are used. With this analysis, Ellesmere becomes the outgroup locality due to the number of characters coded as 0. The y axis represents the similarity index and the x axis represents the number of localities.

The most distinct cluster, with rather high levels of similarity, includes Belgium, RedHill, Andreyevka, Greenland which is supported by characters 44, 70, 72 from Actinopterygii, 70 from Dipnoi, 6, 9, 11, 34, 50, 51, 52 from Tristichopterydae and 36, 50, 68, 73, 111, 112 from Elpistostegalia/Tetrapoda. At the base of this cluster, the Greenland locality is present. Then, the Andreyevka site is present. The rest part of the cluster is composed by Belgium and RedHill which are together. This grouping is supported by the character 2 from Phlyctaenii.

29 Then, there is a cluster showing similarity between JeCa and Ogre. Oryol and Pavari are in the next cluster which is only supported by characters 10 from Dipnoi and 47 Osteolepididae.

A cluster composed by Lode and Miguasha is present. The characters 9 and 31 from Actinopterygii support this cluster. The China locality is not grouping with another locality.

Finally, there is a cluster in basal position showing similarity between Gogo and Scat Craig This is supported by Arthrodira characters. Ellesmere China Miguasha Lode Oryol Pavari Ogre JeCa Greenland Andreyvka RedHill Belgium ScatCraig Gogo 1

0,9

0,8

0,7

0,6

0,5 Similarity

0,4

0,3

0,2

0,1

0 0 1, 6 3,2 4,8 6,4 8 9,6 11, 2 12 , 8 14 , 4

Figure 26: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the morphological level when Canowindra and Jemalong localities are considered as grouped.

6. Discussion Tetrapods are found in sedimentary rocks interpreted as freshwater, brackish or marine (see Background part) and had a worldwide distribution during the Frasnian and Famennian. During the Upper Devonian (Figure 27), the Earth showed a nearly Pangaean palaeogeographical distribution

30 (Laurussia, Gondwana, North China) (Averbuch et al., 2005). The original environment of the early tetrapods have been addressed through comparisons between faunas of fossiliferous localities by means of cluster analysis, an through analysis of co-occurring of aquatic forms with tetrapods by Clack, 2002. A cluster analysis of 39 Famennian localities was done by Lelièvre, 2002. In this study, different methods have been used, including the classical hierarchical cluster analysis using on the genus and family level, as well as neighbour-joining on a localities/taxa matrix. This latter method can use distances with metric and additive properties of the data in order to provide a dendrogram of minimal length. It gives a single dendrogram in which there is no cluster of localities being freshwater. This is why this study concluded that freshwater environments in the Upper Old Red Sandstone could not be confirmed by vertebrates. Other methods, such as sedimentological and sequence- stratigraphical analyses (references in Friend & Williams 2000) have also been used to determinate if Late Devonian tetrapods were living in freshwater or nearshore marine environments but biological and sedimentological data do not fully agree with each other.

In a palaeogeographic point of view, the area where the tetrapods evolved first is unknown. Although, we don’t know how they did disperse to be present from Pennsylvania in the west to Australia in the east. All tetrapods are endemic and restricted to the locality or the region where they have been collected except, for Ichthyostega which is known both in Greenland and Belgium (Clément et al., 2004). So the occurrence of these taxa could not be used to establish biogeographical relationships among the relevant continents. Moreover, there is no consensus regarding tetrapod interrelationships which could be used to infer their palaeocontinental context (see discussions and references in Schultze 1997; Clack 2002; Ruta & Coates 2003; Ruta et al. 2003, Blieck et al., 2007). On current data, Late Devonian tetrapods may have originated in a generalized area including Laurussia, North China and easternmost Gondwana but there is a high probability that this occurred in the western Laurussia because the elpistostegids, the sister-group of tetrapods, occurred in this area (Blieck et al., 2007). According to Friedman & Brazeau (2011), the relative poverty of finds from outside USA and Europe, including Greenland, reflects differential sampling, rather than a highly provincialized palaeobiogeographical pattern. The emergence and the spreading of tetrapods are contemporaneous with a global change context. Tectono-metamorphic data in the world show that the Upper Devonian is characterized by intense tectonic activity with the incipient collision of major continental blocks (Laurussia, Gondwana, Kazakhstan and Siberia) (Averbuch et al., 2005). Moreover, this event is correlated by palaeontological data: the biotic dispersal of placoderms (Young et al., 2000), land plants (Meyer-Berthaud et al., 1997) and their miospores (Streel et al., 2000) which show a very narrow residual oceanic domain between Laurussia and Gondwana. The result of this collisional process was the closure of oceanic domains and deformation and uplift of wide continental areas (Appalachian belt, European Variscides, Northern African Variscides, Arctic Ellesmerian– Svalbardian belt, Central Asian belt, South Uralian belt) and contemporaneous Frasnian–Famennian

31 oceanic subduction led to the accretion of sediments (western American Antler belt, South American Bolivianides, eastern Australian Lachlan fold belt) (Averbuch et al., 2005). This pre-Pangean context seem to have contributed to a global cooling showing both by the δ18O signature of marine carbonates (Joachimski & Buggisch, 2002) and by the miospore distribution on continental areas (Streel et al., 2000). These events have certainly modified strongly several marine environments (see references in

Averbuch et al., 2005).

The results from the present analyses show some mixed signals and globally the first analysis and the second analysis on the first level hierarchy provide distinguished cluster whereas it is not the case in the second level hierarchy and the third analysis. In all analyses, the grouping of Canowindra and Jemalong shows a better similarity index and should be grouped for future analyses.

Although the signals are quite weak due to endemism of species from localities (only seven species exist in more than one locality), there are some cluster signals in the analysis using species level (Figure 17 and 18). It provides a distinction between a cluster of Frasnian localities and a cluster of Frasnian/Famennian localities in which there are both Laurussia (plus Jemalong or JeCa) and Baltica localities. The presence of Jemalong (Figure 17) or JeCa (Figure 18) in the cluster with Laurussian localities is due to placoderms such as Bothriolepis. These placoderms are only recorded from Famennian in the Northern Hemisphere whereas Jemalong or JeCa is in Australia and is clearly older (see reviews in Young 2003, 2005a–b). This is why the position of this locality in the cluster could show a migration of some taxa as placoderms or other vertebrates along the Northern margin of Gondwana to the Laurussia.

It is rather evident that the genus level analysis show stronger results than the species level with respect to a higher similarity index (Figure 19 and 20). Naturally, this is due to the higher level correlation which by default gives a higher similarity index. Globally there are three clusters: a cluster of Frasnian localities which belong to freshwater or brackish environment and a split between a cluster where Laurussian localities plus the Australian locality (Jemalong or JeCa) showing freshwater or brackish sediments and a cluster with Baltican localities considered as marine. Moreover, these two groups are emphasis with palaeogeographical reconstruction (Averbuch et al., 2005) in which Laurussian localities (Belgium, RedHill and Greenland) are located further inland than Baltican localities (Andreyevka, Oryol, Pavari and Ogre) (Figure 27). As in previous studies (Schultze & Cloutier 1996; Schultze 1997; Lelièvre, 2002), there are no clusters of localities supposed to be exclusively freshwater.

Some taxa are present in nearly all localities such as the placoderms Remigolepis, Groenlandaspis and Phyllolepis. There is only one acanthodian presents in more than one locality: Devononchus which is present in Pavari, Ogre Andreyevka, Oryol and Belgium localities. The porolepiforms, Holoptychius is present in all localities excepted in Jemalong, Canowindra and Gogo localities, whereas Glyptolepis is present in Greenland, Andreyevka, Oryol, Lode, Ellesmere localities.

32 Among the dipnoans, Soederberghia is presennt in Jemalong, Canowindra, RedHilll, Greenland and Belgium, and Jarvikia from Greenland, Oryol and Belgiumu . Onychodus is presentt in Gogo and Lode localities. The Tristichopterydae Eustenodon has been found in Ogre, Greenland, Andreyevka, RedHill and Belgium localities. As it was previously said Ichthyostega is present iin Belgium and Greenland.

In order to do future analyses based on taxonomical framework, two possibilities could be done: usse only described taxa and focus analyses on the genus level to get a better similarity index. Some taxa used in the present analysis are not yet described. They represent informmations on the fauna of localities but they are a weakness because some of these undescribed taxa could belong to existing taxa. Some localities are difficult to access as Ellesmere in the Arctic Canada this is why only few materials come from these localities. The difference of taphonomy between localiities is another weakness in the sense that the localities with an exceptional preservation, as Miguasha, provide a greater number of fossils than others. Unfortunately, the taphonomic bias cannot be removed.

Figure 27: Late Devonian tetrapod and near-tetrapod localities plotted on palaeogeographical reconstruction from Averbuch et al., 2005, showing the active orogenic systems at the Frasnian/Famennian boundary. The nummbbers design localities: 1–Catskill, 2– Greenland, 3–ScatCraig, 4–Belgium, 5–6–Pavari and Ogre, 7– Andreyevka-2, 8– Oryol, 9–China, 10–Jemalong and Canowindra. The letters design localities A-Miguasha, B- Lode, C-Ellesmere. Abbreviations: C.A.b., Central Asian belt; NC, North China; SC, Soouth China; T, Tarim (modified from Blieck et al., 2007).

The second analysis (Figure 21 and 24) uses artificial groupings building on the species/genus data set from the previous analysis as described in the Background part. For the fiirst artificial level (Figure 21 and 22), the undescribed material becomes problematic: some of undescribed taxa could belong to a taxon present and create a bias in the analysis. Moreover, some phylogenetic analyses are not well supported as actinopterygian or dipnoan phylogeny and caution must be ttaken. In any case the analysis which is done on the first artificial level give nearly the same picture as the previous analysis:

33 a cluster of Frasnian localities showing brackish to freshwater environment, the split is again present between clusters of Laurussian and Baltican localities. As in the genus level, the first cluster shows localities from the Laurussian in brackish or freshwater deposits whereas the second cluster presents localities only marine called Baltican cluster.

These clusters are represented by the similarity of particular taxa, such as the placoderms Antiarcha 3 is present in nearly all localities except Gogo, Ellesmere, Lode localities. In Jemalong, Belgium, RedHill and Greenland, Phyllolepida is present. Phlyctaenii is present in the same localities as Phyllolepida and also in ScatCraig. There is only one acanthodian taxa present in several localities Acanthodii 1 which is present in Pavari, Ogre, Andreyevka, Oryol, Belgium and Miguasha. Among actinopterygians, there are two taxa which are present in several localities: Actinopterygii 1 from RedHill and Greenland localities and Actinopterygii 2 from Andreyevka and Gogo. Actinistia 1 is present in Ogre, Miguasha and Lode localities. One lungfish taxon is present in more than one locality: Dipnoi 3 from Jemalong, Greenland, RedHill, Oryol, ScatCraig, Belgium and Canowindra. Then Struniiformes 2 is present in Gogo, Andreyevka, Oryol, Lode sites. Among osteolepiforms, Tristichopterydae 1 is in Ogre, Greenland Andreyevka RedHill, Belgium, Canowindra, and Osteolepididae 1 has been found Pavari, RedHill, Oryol and Belgium. There is only one tetrapod taxon, Tetrapoda 3, in Greenland and Belgium as previously.

Some problems appear in the second artificial level (Figure 23 and 24). The Gogo locality is in the ingroup of localities whereas it should not be the case. This level doesn’t show some well grouping which could be explained by environmental influences, neither by geographical influences. This second artificial level could be improved with some changes in the grouping of Dipnoi and with a change of outgroup locality which should be older than Frasnian and with agnathans taxa.

The third analysis used morphological sets of characters instead of taxa (Figure 25 and 26) shows some problems which need to be resolved to improve it because there are not well clusters. Globally, there is one main point to improve: some of characters which are present could be convergences between taxa difficult to identify. These convergent characters could be problematic because they lead to cluster with groupings of locality which are wrong and don’t show the real faunal relationships between localities.

The phylogenetic analyses used for the placoderms dataset don’t include all the groups. These analyses include only Arthrodira (Carr& Hlavin, 2010) for example but there are not phylogenies on Antiarcha, Phyllolepida and Phlyctaenii. In that case, the diagnostic traits are useful but are problematic: in the Antiarcha characters some localities present Asterolepis, Remigolepis and Bothriolepis, the two first taxa present the most ancestral characters and Bothriolepis represents characters in the state 1. So when Bothriolepis taxa are present in a locality, this locality is coded as 1. These Antiarcha characters could be change to take in consideration species if future phylogenies are done. It is the same things with Phyllolepida and Phlyctaenii characters, if these characters came from

34 phylogenetic analyses, they could improve the third analysis. For Arthrodira characters, another selection of character could be a better choice.

Some convergences are present in Dipnoi as the long-snout, to improve the choice of characters maybe only characters coming from tooth-plate should be included. Cautions should be taken with taxa in which tooth-plates were not found. Moreover phylogeny from Dipnoi shows a poor resolution.

Actinopterygian phylogeny is relatively well supported with the present taxa and provides a good signal with the set of characters chosen for the third analysis. Unfortunately, nearly complete describe taxa are not present in all localities but if new taxa are found, the Actinopterygii characters could provide a better grouping of localities.

The Osteolepiformes phylogeny provided well supported tree in which Osteolepididae, Megalichthydae and Tristichopterydae are clearly distinguished. This separation provides the better signal for the third analysis. Characters from this phylogeny provide the better grouping of the third analysis.

The Elpistostegalia and Tetrapoda characters should be only on lower jaws to improve the third analysis but some of Tetrapods are not described on their lower jaws as the possible watcheriid from RedHill, Jakubsonia livnensis from Oryol and undetermined tetrapods from Andreyevka and from Pavari. So the characters which are not present from these two taxa should be coded as zero by default.

7. Conclusion The first tetrapods found from body fossil have been dated from Frasnian to Famennian. By that time they were already widespread throughout the world and present in various environmental conditions from true marine conditions of the near-shore neritic province, to fluvial or lacustrine conditions. These tetrapod localities show faunal relationships which are an important for the study of tetrapods evolution and environmental context.

In order to explore these relationships, these localities form the foundation for a number of data matrices from which three different cluster analyses were.

In the analysis of the taxonomical framework, Frasnian faunas are spread around the world and don’t show any provincialism whereas a differentiation between Laurussian faunas from brackish to freshwater environments and Baltican faunas from marine environment during the Famennian is present.

The new method of artificial level by grouping previous taxa of the first analysis provides some good results. The first artificial level of the second analysis shows nearly the same result as

35 previously with Frasnian faunas spread around the world and the differentiation between Laurussian faunas with brackish to freshwater environmental conditions and Baltican but with marine environmental conditions. Also even if this level provides good results, the analysis need to be improve as some changes in grouping of taxa with new phylogenetic analyses. The second level of the second analysis needs to be more improved than the previous level in order to provide well results. This improvement could be done with some changes in grouping of taxa based on present phylogenetic analyses and also with new phylogenies.

The third analysis need to be improved also because it is a new method. In order to do this improvement, some changes in the character sets should be done in characters from antiarch placoderms and dipnoan to remove convergences. Characters from Phlyctaenii and Phyllolepida should be removed. If this is not enough, it should include only characters from well supported phylogeny as those from osteolepiforms.

Further works on these three analyses should be done with the improvements thus that with others methods as Ward’s method and UPGMA which could not be explored because of lack of time. 8. Acknowledgements I thank my supervisor Henning Blom for this subject, the first data base and datasets with unpublished material, for his idea of the second analysis and his help and discussions and Nicolas Campioné about the idea of the third analysis. I thank my faculty reviewer Benjamin Kear for his comments on the Master Thesis. I thank Sébastien Clausen who permits me with the double Master’s degree to be in Sweden and get the opportunity to do my Master Thesis in Uppsala. I thanks to my opponent Pierre Wavelet for his comments and questions. I thank all people I forgot. I thank my parents and family for their support during all these years and for this Master Thesis even if they don’t speak and understand English. À tous ceux qui ne sont plus là. 9. References Ahlberg, P.E., 1992. A new holoptychiid porolepiforms fish from the Upper Frasnian of Elgin, Scotland. Palaeontology, Vol 35, Part 4, 1992, pp. 813-828.

Ahlberg, P.E., 1995. Elginerpeton pancheni and the earliest tetrapod clade. Nature, Vol 373, pp. 420- 425.

Ahlberg, P.E., 1998. Poscranial stem tetrapod remains from the Devonian of Scat Craig, Porayshire, Scotland. in: D.B. Norman, A.R. Milner, A.C. Milner (Eds.), A study of fossil vertebrates, Zoological Journal of the Linnean Society, London., 122, 1998, pp. 99–141.

Ahlberg, P.E. & Johanson, Z., 1997. Second tristichopterid (Sarcopterygii, Osteolepiformes) from the

36 Upper Devonian of Canowindra, New South Wales, Australia, and phylogeny of the Tristichopteridae. Journal of Vertebrate Paleontology, 17:4, 653-673

Ahlberg, P.E., Johanson, Z. & Daeschler, E.B., 2001. The late Devonian lungfish Soederberghia

(Sarcopterygii, Dipnoi) from Australia and North America, and its biogeographical implications. Journal of Vertebrate Paleontology, 21:1, 1-12

Ahlberg, P.E., Luksevics, E. & Lebedev, O., 1994. The First Tetrapods Finds from the Devonian (Upper Famennian) of Latvia. Philosophical Transactions of the Royal Society, London, B343, 303– 328.

Ahlberg, P.E., Smith, M.M. & Johanson, Z., 2006. Developmental plasticity and disparity in early dipnoan (lungfish) Dentitions. EVOLUTION & DEVELOPMENT 8:4, 331–349.

Alekseev, A.S., Lebedev, O.A., Barskova, M.I., Kononova, L.I. & Chizhova, V.A., 1994. On the stratigraphic position of the Famennian and Tournaisian fossil vertebrate beds in Andreyevka, Tula Region, Central Russia. Proceeding of the Geologic Association, 105, 41-52.

Arratia, G. & Cloutier, R., 1996. Reassessment of the morphology of Cheirolepis canadensis (Actinopterygii). In: Schultze, H.P., Cloutier, R. (Eds.), Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, München, pp. 165-197.

Arsenault, M., 1982. Eusthenopteron foordi, a predator on Homalacanthus concinnus from the Escuminac Formation, Miguasha, Quebec. Canadian Journal of Earth Sciences, 1982, Vol. 19, N° 11 : pages 2214-2217

Averbuch, O., Tribovillard, N., Devleeschouwer, X., Riquier, L., Mistiaen, B. & Van Vliet-Lanoë, B. 2005. Mountain building-enhanced continental weathering and organic carbon burial as major causes for climatic cooling at the Frasnian– Famennian boundary (c. 376 Ma)? Terra Nova, 17, 33–42.

Béland, P. & Arsenault, M., 1985. Scauménellisation de l'Acanthodii Triazeugacanthus affinis (Whiteaves) de la formation d'Escuminac (Dévonien supérieur de Miguasha, Québec) : révision du Scaumenella mesacanthi Graham-Smith. Canadian Journal of Earth Sciences, 1985, Vol. 22, N° 4 : pages 514-524

Blieck, A., Clément, G., Blom, H., Lelièvre, H., Luksevics, E., Streel, M., Thorez, J. & Young, G.C., 2007. The biostratigraphical and palaeogeographical framework of the earliest diversification of tetrapods (Late Devonian). From: BECKER, R. T. & KIRCHGASSER, W. T. (eds) Devonian Events and Correlations. Geological Society, London, Special Publications, 278, 219–235.

Blom, H., Clack, J.A., Ahlberg, P.E. & Friedman, M., 2007. Devonian vertebrates from East Greenland: a review of faunal composition and distribution. Geodiversitas 29 (1): 119-141.

37 Burrow, C.J., Trinajstic, K. & Long, J.A., 2012. First acanthodian from the Upper Devonian (Frasnian) Gogo Formation, Western Australia. Historical Biology Vol. 24, No. 4, August 2012, 349– 357

Carr, R.K. & Hlavin, W.J., 2010. Two new species of Dunkleosteus Lehman, 1956, from the Ohio Shale Formation (USA, Famennian) and the Kettle Point Formation (Canada, Upper Devonian), and a cladistic analysis of the Eubrachythoraci (Placodermi, Arthrodira). Zoological Journal of the Linnean Society, 2010, 159, 195–222.

Choo, B., 2011. Revision of the actinopterygian genus Mimipiscis (= Mimia) from the Upper Devonian Gogo Formation of Western Australia and the interrelationships of the early Actinopterygii. Earth and Environmental Science Transactions of the Royal Society of Edinburgh, 102, 77–104.

Choo, B., Long, J.A. & Trinajstic, K., 2009. A new genus and species of basal actinopterygian fish from the Upper Devonian Gogo Formation of Western Australia. Acta Zoologica (Stockholm) 90 (Suppl. 1): 194–210

Clack, J.A., 2002. Gaining Ground: The Origin and Evolution of Tetrapods. Indiana University Press, Bloomington & Indianapolis. Clack, J.A., 2006. The emergence of early tetrapods. Palaeogeography, Palaeoclimatology, Palaeoecology, 232, 167–189.

Clack, J.A., Ahlberg, P.E., Blom, H. & Finney, S.M., 2012. A new genus of Devonian Tetrapod from North-East Greenland, with new information on the lower jaw of Ichthyostega. Palaeontology, Vol 55, Part 1, 2012, pp. 73-86.

Clément, A.M., 2012. A new species of long-snouted lungfish from the Late Devonian of Australia, and its functional and biogeographical implications. Palaeontology, Vol. 55, Part 1, 2012, pp. 51–71.

Clément, A.M. & Long, J.A., 2010. Xeradipterus hatcheri, a new dipnoan from the Late Devonian (Frasnian) Gogo Formation, Western Australia, and other new holodontid material. Journal of Vertebrate Paleontology, 30:3, 681-695

Clément, G., 2002. Large Tristichopteridae (Sarcopterygii, Tetrapodomorpha) from the Late Famennian Evieux Formation of Belgium. Palaeontology, 45(3), 577-593.

Clément, G., Ahlberg, P. E., Blieck, A., Poty, E., Thorez, J., Clack, J. A., Blom, H. and Janvier, P. 2004. Palaeogeography: Devonian tetrapod from Western Europe. Nature, 427, 412–413.

Clément, G. & Boisvert, C. A., 2006. Lohest’s true and false “Devonian amphibians”: evidence for a rhynchodipterid lungfish in the Late Famennian of Belgium. Journal of Vertebrate Paleontology, 26(2), 276-283.

38 Clément, G., Snitting, D. & Ahlberg, P. E., 2009. A new tristichopterid (Sarcopterygii, Tetrapodomorpha) from the Upper Famennian Evieux Formation (Upper Devonian) of Belgium. Palaeontology, 52(4), 823-836.

Cloutier, R., 1996. The primitive actinistian Miguasha bureau,Schultze (Sarcopterygii). In: Schultze, H.P., Cloutier, R. (Eds.), Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, München, pp. 227-247.

Cloutier, R., 2010. Great Canadian Lagerstätten 4. The Devonian Miguasha Biota (Québec): UNESCO World Heritage Site and a Time Capsule in the Early History of Vertebrates. Geoscience Canada, Journal of the geological association of Canada, Volume 40, Number 2.

Cloutier, R., Proust, J-N. & Tessier, B., 2011. The Miguasha Fossil-FishLagerstätte: a consequence of the Devonian land-sea interactions. Palaeobio Palaeoenv (2011) 91:293-323.

Cressler, W.L. III, Daeschler, E.B., Slingerland, R. & Peterson, D.A., 2010. Terretrialization in the Late Devonian: a palaeoecological overview of the Red Hill site, Pennsylvania, USA. Geological Society, London, Special Publication 2010, v.339; p111-128.

Daeschler, E.B., 2000. Early tetrapod jaws from the Late Devonian of Pennsylvania, USA, Journal of Paleontology, Vol. 74, No. 2, pp. 301–308.

Daeschler, E.B., Clack, J.A. & Shubin, N.H., 2009. Late Devonian tetrapod remains from Red Hill, Pennsylvania, USA: how much diversity? Acta Zoologica (Stockholm) 90 (Suppl. 1): 306–317.

Daeschler, E.B., Shubin, N.H. & Jenkins, F.A., Jr, 2006. A Devonian tetrapod-like fish and the evolution of the tetrapod body plan. Nature 440 (2006) 757–763.

Daeschler, E.B., Shubin, N.H., Thomson, K.S. & Amaral, W.W., 2000. A Devonian tetrapod from North America, Science 265 (1994) 639–642.

Davis, M.C., Shubin, N. & Daeschler, E.B., 2004. A new specimen of Sauripterus taylori (Sarcopterygii, Osteichthyes) from the Famennian Catskill Formation of North America, Journal of Vertebrate Paleontology, 24:1, 26-40.

Derycke, C. & Clément, G., 2013. First assemblage of Acanthodian scales and spines from the Famennian (Upper Devonian) of Durnal (Belgium), palaeobiogeographical and palaeoenvironmental implications. GEOLOGICA BELGICA (2013) 16/1-2: 18-26.

Downs, J.P. & Daeschler, E.B., 2001. Variation within a large sample of Ageleodus pectinatus teeth

(Chondrichthyes) from the Late devonian of Pennsylvania, U.S.A., Journal of Vertebrate Paleontology, 21:4, 811-814.

39 Downs, J.P., Daeschler, E.B., Jenkins, F.A. Jr & Shubin, N.H., 2011. A new species of Laccognathus (Sarcopterygii, Porolepiformes) from the Late Devonian of Ellesmere Island, Nunavut, Canada. Journal of Vertebrate Paleontology, 31:5, 981-996.

Downs, J.P., Daeschler, E.B., Jenkins, F.A. Jr & Shubin, N.H., 2013. Holoptychius bergmanni sp. nov. (Sarcopterygii, Porolepiformes) from the Upper Devonian of Nunavut, Canada, and a Review of Holoptychius Taxonomy. Proceedings of the Academy of Natural Sciences of Philadelphia, 162():47- 59.

Dupret, V. & Zhu, M., 2007.The earliest phyllolepid (Placodermi, Arthrodira) from the LateLochkovian (Early Devonian) of Yunnan (South China). Geol. Mag. 145 (2), 2008, pp. 257–278.

Elliott, D.K., Johnson, H.G., Cloutier, R., Carr, R.K. & Daeschler, E.B., 2000. Middle and Late Devonian vertebrates of the western Old Red Sandstone Continent. In: Blieck, A., Turner, S. (Eds.), Palaeozoic Vertebrate Biochronology and Global Marine-Non-Marine Correlation. Final Report of IGCP 328 (1991-1996). Courier Forschungsinstitut Senckenberg 223, pp. 291-308.

Embry, A. F., 1991. in Geology of the Innuitian Orogen and Arctic Platform of Canada and Greenland (ed. Trettin, H. P.) 263–-279 (Geological Survey of Canada, Calgary, 1991).

Embry, A. & Klovan, J.E., 1976. The Middle-Upper Devonian clastic wedge of the Franklinian Geosyncline. Bull. Can. Petrol. Geol. 24, 485–-639.

Forey, P.L., Ahlberg, P.E., Luksevics, E. & Zupins, I., 2000. A new coelacanth from the Middle Devonian of Latvia. Journal of Vertebrate Paleontology, 20:2; 243-252.

Friedman, M. & Brazeau, M.D., 2011. Sequences, stratigraphy and scenarios: what can we say about the fossil record of the earliest tetrapods? Proc. R. Soc. B (2011) 278, 432–439.

Friedman, M. & Blom, H., 2006. A new actinopterygian from the Famennian of East Greenland and the interrelationship on Devonian ray-finned fishes. Journal of Paleontology 20, 1186-1204.

Friedman, M. & Daeschler, E.B., 2006. Late Devonian (Famennian) Lungfishes from the Catskill Formation of Pennsylvania, USA. [Palaeontology, Vol. 49, Part 6, pp. 1167–1183.

Gagnier, P.-Y., 1996. Acanthodii. In: Schultze, H.P., Cloutier, R. (Eds.), Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, München, pp. 149-164.

Gardiner, B. G. & Bartram, A.W.H., 1977. The homologies of ventral cranial fissures in osteichthyans, p. 227–245. In S. M. Andrews, R. S. Miles, and A. D. Walker (eds.), Problems in Vertebrate Evolution. Academic Press, London.

Goemaere, E. 1995. A model of a regressive depositional system around the Old Red Continent in the Upper Famennian ‘Psammites du Condroz’ in Belgium. 88–104. In BLIECK, A. (ed.). Guidebook for

40 IGCP 328 ⁄ SDS joint field trip Boulonnais (France) – Ardenne (Belgium). Universite´ des Sciences et Technologies de Lille-URA 1365 du CNRS, Lille, 119 pp.

Gross, W., 1941. Über den Unterkiefer einiger devonischer Crossopterygier. Abh. preuss. Akad. Wiss. Math. Naturw. Kl. 1941, 1-51.

Gross, W., 1950. Die paläontologische Beduntung des Wirbeltierfaunen des Old Reds und der marine altpaläozoischen Schichten. Abhandlungen des deutches Akademie des Wissenschaften, Berlin, Mathematisch-Naturwissenschaftliche Klasse 1949, 1-130.

Gueriau, P., Charbonnier, S. & Clément, G., 2014. First decapod crustaceans in a Late Devonian continental ecosystem. Palaeontology,2014 Early View (Online Version of Record published before inclusion in an issue).

Hammer, Ø., Harper, D.A.T. & Ryan P.D., 2001. PAST: Paleontological statistics software package for education and data analysis (version 1.56b). Palaeontol Electron 4. Available at: http://palaeo- electronica.org/2001_2001/past/issue2001_2001.htm

Hanke, G.F. & Davis, S.P., 2012. A re-examination of Lupopsyrus pygmaeus Bernacsek &Dineley, 1977 (Pisces, Acanthodii). Geodiversitas, 34(3):469-487.

Schultze, H.-P. & Arsenault, M., 1987. Quebecius quebecensis (Whiteaves), a porolepiform crossopterygian (Pisces) from the Late Devonian of Quebec, Canada. . Canadian Journal of Earth Sciences, 1987, 24(12): 2351-2361

Janvier, P., 1996. The Miguasha “Anaspida”. In: Schultze, H.P., Cloutier, R. (Eds.), Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, München, pp. 134-140.

Janvier, P. & Arsenault, M., 1996. Osteostraci. In: Schultze, H.P., Cloutier, R. (Eds.), Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, München, pp. 123-133.Janvier, P. & Arsenault, M., 2009. La conservation exceptionnelle des fossiles du parc national de Miguasha. Nat Can 133:78–83.

Jarvik E. 1952. On the fish-like tail in the ichthyostegid stegocephalians. Meddelelser om Gronland 114: 1-90.

Jia, L-T, Zhu, M. & Zhao, W-J., 2010. A new antiarch fish from the Upper Devonian Zhongning Formation of Ningxia, China. Palaeoworld 19 (2010) 136–145.

Joachimski, M. & Buggisch, W., 2002. Conodont apatite δ18O signatures indicate climatic cooling as a trigger of the Late Devonian mass extinction. Geology, 30, 711–714.

Johanson, Z., 1997. New Remigolepis (Placodermi; Antiarchi) from Canowindra, New South Wales, Australia. Geol. Mag. 134 (6), 1997, pp. 813–846.

41 Johanson, Z. & Ahlberg, P.E., 1997. A new tristichopterid (Osteolepiformes: Sarcopterygii) from the Mandagery Sandstone (Late Devonian, Famennian) near Canowindra, NSW, Australia. Transactions of the Royal Society of Edinburgh: Earth Sciences, 88, 39-68.

Johanson, Z. & Ahlberg, P.E., 1998. A complete primitive rhizodont from Australia. Nature, Vol. 394, pp. 669-673.

Krupina, N.I., 2000. Dipnoans from the Upper Devonian Locality Rybnitsa in Oryol Region. Paleontology Journal, Vol. 34, N°5, 2000, pp. 527-533. Translated from Paleontologicheskii Zhurnal, N°5, 2000, pp. 55-61.

Lane, J.A. & Cuffey, R.J., 2005. Phyllolepis rossimontina sp. nov. (Placodermi) from the Uppermost Devonian at Red Hill, North-Central Pennsylvania. Revista Brasileira de Paleontologia 8(2):117-126.

Lebedev, O.A., 1995. Middle Famennian (Upper Devonian) Chondrichthyans and Sarcopterygians from Oryol region, Central Russia.. GEOBIOS, M.S. n ° 19 : 361-368.

Lebedev, O.A. 2004. A new tetrapod Jakubsonia livnensis from the Early Famennian (Devonian) of Russia tetrapod habitats. In: Luksevics, E. & Stinkulis, G. (eds) The Second Gross Symposium on Advances of Palaeoichthyology (Riga, 2003). Acta Universitatis Latviensis, 679, 79–98.

Lebedev, O.A. & Clack, J.A., 1993. Upper Devonian tetrapods from Andreyevka, Tula region, Russia. Palaeontology, 36(3), 721–734.

Lebedev, O.A., Luksevics, E. & Zakharenko, G.V., 2010. Palaeozoogrographical connections of the Devonian vertebrate communities of the Baltica Province. Part II. Late Devonian. Palaeoworld 19 (2010) 108–128.

Lebedev, O.A. & Zakharenko, 2010. Global vertebrate-based palaeozoogeographical subdivision for the Givetian-Famennian (Middle-Late Devonian): Endemism-cosmopoltitanism spectrum as an indicator of interprovincial faunal exchanges. Palaeoworld 19 (2010) 186–205.

Lelièvre, H., 2002. Phylogénie des Brachythoraci (Vertebrata, Placodermi) et ajustement de la phylogénie à la stratigraphie. Les sites du Dévonien terminal, la caractérisation de leur milieu de dépôt par analyse de similitude de leur ichthyofaune. H. D. R. Sciences Naturelles, U.S.T.L., Villeneuve d’Ascq (12 décembre 2002).

Leriche, M., 1931. Les Poissons Famenniens de la Belgique. Mem. Cl. Sci. Acad. Roy. Belgique, 4(10): 1-72.

Lohest, M., 1888a. Recherches sur les poissons des terrains paléozoïques de Belgique. Poissons des Psammites du Condroz, Famennien supérieur. Annales de la Société Géologique de Belgique, mémoires 15: 112-203.

Lohest, M., 1888b. Découverte du plus ancien amphibien connu et de quelques fossiles remarquables dans le Famennien supérieur de Modave. Annales de la Société Géologique de Belgique, 129-137.

42 Long, J. A. 1985. A new osteolepidid fish from the Upper Devonian Gogo Formation, Western Australia. – Records of the Western Australian Museum 12: 361–377.

Long, J.A. & Daeschler, E.B., 2013. First Articulated Phyllolepid Placoderm from North America, with Comments on Phyllolepid Systematics. Proceedings of the Academy of Natural Sciences of Philadelphia, 162():33-46.

Long, J.A. & Trinajstic, K., 2010. The Late Devonian Gogo Formation Lägerstatte of Western Australia: Exceptional Early Vertebrate Preservation and Diversity. Annu. Rev. Earth Planet. Sci. 2010. 38:255–79.

Lukesecics, E. & Zupins, I., 2004. Sedimentology, fauna, and taphonomy of the Pavari site, Late Devonian of Latvia, in: E. Luksevics, G. Stinkulis (Eds.), Advances of Palaeoichthyology, Second Gross Symposium, 2003, Acta Universitatis Latviensis, 679, Riga, 2004, pp. 99–119.

Luksevics, E., Ahlberg, P.E., Stinkulis, G., Vasilkova, J. & Zupins, I. 2011: Frasnian vertebrate taphonomy and sedimentology of macrofossil concentrations from the Langsede Cliff, Latvia. Lethaia, Vol. 45, pp. 356–370.

Meyer-Berthaud, B., Wendt, J. & Galtier, J., 1997. First record of a large Callixylon trunk from the Late Devonian of Gondwana. Geological Magazine, 134, 847–853.

Moloshnikov, S.V., 2001. Unikal’noye mestonakhozhdeniye pozvonochnykh verkhnego devona Orlovskoy oblasti [A unique vertebrate locality in the Upper Devonian of the Oryol Region]. Izvestiya VUZov, Geologiya i Razvedka 3, 29-33 (in Russian).

Moloshnikov, S.V. 2003. A new species of bothriolepidid antiarchs (Pisces, Placodermi) from the Zadonskian Regional Stage (Upper Devonian) of the Central Devonian field. Paleontologicheskiy Zhurnal, 2: 69-72.

Moloshnikov, S.V., 2008. Devonian Antiarchs (Pisces, Antiarchi) from Central and Southern European Russia Paleontological Journal, 2008, Vol. 42, No. 7, pp. 691–773.

Newman, M.J., 2005. A systematic review of the Placoderm genus Cosmacanthus and a description of Acanthodian remains from the Upper Devonian of Scotland. Palaeontology, Vol. 48, Part 5, 2005, pp. 1111–1116

Olsen, H. & Larsen P.-H., 1993. — Lithostratigraphy of the continental Devonian sediments in North- East Greenland. Bulletin of the Gronlands Geologiske Undersogelse 165: 1-108.

Pan, J., Huo, F.C., Cao, J.X., Gu, Q.C., Liu, S.Y.,Wang, J.Q., Gao, L.D., Liu, C., 1987. Continental Devonian System of Ningxia and its Biotas. Geological Publishing House, Beijing, 237 pp. (in Chinese and English).

Pan, J., Wang, S.-T., Liu, S.-Y., Gu, Q.-C. & Hang, J. 1980. Discovery of Devonian Bothriolepis and Remigolepis in Ningxia. Acta Geologica Sinica 54, 176–85.

43 Säve-Söderbergh G. 1932a. Preliminary note on Devonian stegocephalians from East Greenland. Meddelelser om Gronland 94 (7): 1-211.

Ruta, M. & Coates, M.I., 2003. Bones, molecules, and crown-tetrapod origins. In: DONOGHUE, P.C.J. & SMITH, M.P. (eds) Telling the Evolutionary Time Molecular Clocks and the Fossil Record. CRC Press, Boca Raton/Systematic Association, Special Volume Series/Palaeontological Association, 224–262.

Ruta, M., Coates, M.I. & Quicke, D.L.J., 2003. Early tetrapod relationships revisited. Biological Reviews, 78, 251–345.

Schultze, H.-P., 1996. The vertebrates of the Escumiac Formation. In: Schultze, H.P., Cloutier, R. (Eds.), Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, München, pp. 120-122.

Schultze, H.-P. 1997. Umweltbedingungen beim Übergang von Fisch zu Tetrapode [Paleoenvironment at the transition from fish to tetrapod.] Sitzungsberichte der Gesellschaft Naturforschender Freunde zu Berlin, N. F., 36, 59–77.

Schultze, H.-P. & Cloutier, R., 1996. Comparison of the Escuminac Formation ichthyofauna with other late Givetian/early Frasnian ichthyofaunas. In: SCHULTZE, H.-P. & CLOUTIER, R. (eds) Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, Mu¨nchen, 348–368.

Snitting, D., 2008. A redescription of the anatomy of the Late Devoniana Spodichthys buetleri Jarvik, 1985 (Sarcopterygii, Tetrapodomorpha) from Greenaland. Journal of Vertebrate Paleontology; 28:3, 637-655.

Sorokin, V.S., 1978. Etapy razvitiya severo-zapada Russkoy platformy vo Franskom veke [Stages of development of the north-western part of the Russian platform in the Frasnian]. Zinatne, Riga. [In Russian].

Streel, M., Caputo, M.V., Loboziak, S. & Melo, J.H.G., 2000. Late Frasnian–Famennian climates based on palynomorph analyses and the question of the Late Devonian glaciations. Earth-Science Reviews, 52, 121–173.

Talent, J.A.,Mawson, R., Aitchison, J.C., Becker, R.T., Bell, K.N., Bradshaw, M.A., Burrow, C. J., Cook, A.G., Dargan, G.M., Douglas, J.G., Edgecombe, G.D., Feist, M., Jones, P.J., Long, J.A., Phillips-Ross, J.R. Pickett, J.W., Playford, G., Rickards, R.B., Webby, B.D., Winchester-Seeto, T., Wright, A.J., Young, G.C. & Zhen, Y-Y., 2000. 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. Association of Australasian Palaeontologists, Memoirs, 23, 167–257.

44 Thomson, K.S., 1973. Observations on a new rhipidistian fish from the Upper Devonian of Australia. Palaeontographica 143A: 209-220.

Thomson, K.S. & Thomas, B., 2001. On the status of species of Bothriolepis (Placodermi, Antiarchi) in North America. Journal of Vertebrate Paleontology, 21:4, 679-686.

Thorez, J., Dreesen, R. & Streel, M., 2006. Famennian. Geologica Belgica (2006) 9/1-2: 27-45.

Upeniece, I., 2001. The unique fossil assemblage from the Lode Quarry (Upper Devonian, Latvia). Mitt. Mus. Nat.kd. Berl., Geowiss Reihe 4 (2001) 101-119.

Upeniece, I. & Upenieks, J. 1992. Young Upper Devonian antiarch (Asterolepis) individuals form the Lode quarry, Latvia. In Fossil Fish as Living Animals (ed. E. Mark-Kurik), pp. 167–74. Tallinn: Academy of Sciences, Estonia (Palaeozoic Fishes, 2nd Colloquium).

Vézina, D., 1996. Placodermi (Antiarcha and Arthrodira). In: Schultze, H.P., Cloutier, R. (Eds.), Devonian Fishes and Plants of Miguasha, Quebec, Canada. Verlag Dr. Friedrich Pfeil, München, pp. 149-164.

Vorobyeva, E.I. 1977: Morfologiya i osobennosti evolyutsii kisteperikh ryb. Trudy Paleontologicheskogo Instituta 163, 1–240.

Westoll, T.S., 1938. Ancestry of the tetrapods. Nature, 141, 127.

Young, G.C., 1993. Middle Palaeozoic macrovertebrate biostratigraphy of eastern Gondwana. In: LONG, J. A. (ed.) Palaeozoic Vertebrate Biostratigraphy and Biogeography. Belhaven Press, London, 208–251.

Young, G.C., 2003. North Gondwanan mid-Palaeozoic connections with Euramerica and Asia; Devonian vertebrate evidence. Courier Forschungsinstitut Senckenberg, 242, 169–185.

Young, G.C., 2005a. An articulated phyllolepid fish (Placodermi) from the Devonian of central Australia: implications for non-marine connections with the Old Red Sandstone continent. Geological Magazine, 142, 173–186.

Young, G.C., 2005b. New phyllolepids (placoderm fishes) from the Middle–Late Devonian of southeastern Australia. Journal of Vertebrate Paleontology, 25, 261–273.

Young, G.C., Long, J.A. & 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. Association of Australasian Palaeontologists, Memoirs, 23, 209– 219 and 250.

Young, G.C., Long, J.A. & Ritchie, A., 1992. Crossopterygian fishes from the Devonian of Antarctica: systematics, relationships and biogeographic significance. Records of the Australian Museum, Supplement 14: 1-77.

45 Zhang, G.-R. & Liu, Y.-G., 1991. A new antiarch from the Upper Devonian of Jinagxi, China. In Early Vertebrates and related problems of Evolutionary Biology (eds M.-M. Chang,Y.-H. Liu and G.- R. Zhang), pp. 195–212. Beijing: Science Press.

Zhu, M., Ahlberg, P.E., Zhao, W. & Jia, L., 2002. First Devonian tetrapod from Asia. Nature, 420, 760–761.

46 Appendix List of taxa used for the analysis 1 and 2 (the first column design the second artificial level, the second column design the first artificial level, the third column design the genus/species level, the fourth column design the locality and the last column design references where each species are presents). Agnatha Anaspida Anaspida A Anaspida 1 Euphanerops longaevus Miguasha Stensiö, 1939; Janvier, 1996 Anaspida A Anaspida 2 Legendrelepis parenti Miguasha Woodward, 1900; Janvier, 1996 Anaspida B Anaspida3 Endeiolepis aneri Miguasha Arsenault & Janvier, 1991; Janvier, 1996 Galeaspida Galeapida Galeapida Galeapida indet. China Pan et al., 1987 Osteostraci Osteostraci A Osteostraci 1 Alaspis rosamundae Miguasha Janvier & Arsenault 1996; Cloutier, 2010 Osteostraci B Osteostraci 2 Escuminaspis laticeps Miguasha Janvier & Arsenault 1996; Cloutier, 2010 Osteostraci C Osteostraci 3 Levesquaspis patteni Miguasha Janvier & Arsenault 1996; Cloutier, 2010 Heterostraci Heterostraci A Heterostraci 1 Psammosteus falcatus Ogre Luksevics et al., 2011 Heterostraci A Heterostraci 1 Psammosteus cf. Scat Craig Ahlberg, falcatus 1998 Heterostraci A Heterostraci 1 Psammosteus tenuis Ogre Luksevics et al., 2011 Heterostraci A Heterostraci 1 Psammosteus sp. Ogre Luksevics et al., 2011 Heterostraci A Heterostraci 2 Traquairosteus Scat Craig Ahlberg, postulatus 1998 Heterostraci B Heterostraci 3 Psammolepis alata Lode Upeniece, 2001 Heterostraci B Heterostraci 3 Psammolepis paradoxa Lode Upeniece, 2001 Heterostraci C Heterostraci 4 Obruchevia heckeri Ogre Luksevics et al., 2011

47 Placodermi Antiarcha Antiarcha A Antiarcha 1 Asterolepis ? amulensis Ogre Luksevics et al., 2011 Antiarcha A Antiarcha 1 Asterolepis cristata Ellesmere Daeschler et al., 2006 Antiarcha A Antiarcha 1 Asterolepis ornata Lode Upeniece & Upenieks, 1992; Upeniece, 2001 Antiarcha A Antiarcha 1 Asterolepis radiata Ellesmere Daeschler et al., 2006 Antiarcha A Antiarcha 1 Asterolepis sinensis China Zhang & Liu, 1991 Antiarcha A Antiarcha 2 Remigolepis acuta Greenland Blom et al., 2007 Antiarcha A Antiarcha 2 Remigolepis armata Andreyevka Alekseev et al., 1994 Antiarcha A Antiarcha 2 Remigolepis cristata Greenland Blom et al., 2007 Antiarcha A Antiarcha 2 Remigolepis emarginata Greenland Blom et al., 2007 Antiarcha A Antiarcha 2 Remigolepis incisa Greenland Blom et al., 2007 Antiarcha A Antiarcha 2 Remigolepis kochi Greenland Blom et al., 2007 Antiarcha A Antiarcha 2 Remigolepis kullingi Greenland Blom et al., 2007 Antiarcha A Antiarcha 2 Remigolepis major China Pan et al., 1987 Antiarcha A Antiarcha 2 Remigolepis China Pan et al., microcephala 1987 Antiarcha A Antiarcha 2 Remigolepis ? Greenland Blom et al., tuberculata 2007 Antiarcha A Antiarcha 2 Remigolepis walkeria Canowindra Johanson, 1997 Antiarcha A Antiarcha 2 Remigolepis China Pan et al., xiangshanensis 1987 Antiarcha A Antiarcha 2 Remigolepis xixiaensis China Pan et al., 1987 Antiarcha A Antiarcha 2 Remigolepis China Pan et al., zhongningensis 1987 Antiarcha A Antiarcha 2 Remigolepis China Pan et al., zhongweiensis 1987 Antiarcha A Antiarcha 2 Remigolepis sp. 1 Andreyevka Alekseev et al., 1994 Antiarcha A Antiarcha 2 Remigolepis sp. 2 Oryol Moloshnikov, 2001, Lebedev, 2004 Antiarcha A Antiarcha 2 Remigolepis sp. 3 Jemalong Young, 1993 Antiarcha B Antiarcha 3 Bothriolepis canadensis Miguasha Vézina,

48 1996; Cloutier, 2010 Antiarcha B Antiarcha 3 Bothriolepis ciecere Pavari Luksevics & Zupins, 2004 Antiarcha B Antiarcha 3 Bothriolepis evaldi Ogre Luksevics et al., 2011 Antiarcha B Antiarcha 3 Bothriolepis Greenland Blom et al., groenlandica 2007 Antiarcha B Antiarcha 3 Bothriolepis jarviki Greenland Blom et al., 2007 Antiarcha B Antiarcha 3 Bothriolepis cf. Oryol Lebedev, leptocheria 2004 Antiarcha B Antiarcha 3 Bothriolepis lohesti Belgium Leriche, 1931 Antiarcha B Antiarcha 3 Bothriolepis maxima Ogre Luksevics et al., 2011 Antiarcha B Antiarcha 3 Bothriolepis nielseni Greenland Blom et al., 2007 Antiarcha B Antiarcha 3 Bothriolepis nitida RedHill Thomson & Thomas, 2001 Antiarcha B Antiarcha 3 Bothriolepis China Pan et al., niushoushanensis 1980 Antiarcha B Antiarcha 3 Bothriolepis paradoxa Scat Craig Ahlberg, 1998 Antiarcha B Antiarcha 3 Bothriolepis sosnensis Oryol Moloshnikov, 2003; Lebedev, 2004 Antiarcha B Antiarcha 3 Bothriolepis yeungae Canowindra Young, 1993 Antiarcha B Antiarcha 3 Bothriolepis zdonica Oryol Lebedev, 2004 Antiarcha B Antiarcha 3 Bothriolepis sp. 1 Belgium Leriche, 1931 Antiarcha B Antiarcha 3 Bothriolepis sp. 2 Andreyevka Alekseev et al., 1994 Antiarcha B Antiarcha 3 Bothriolepis sp. 3 Jemalong Young, 1993 Antiarcha B Antiarcha 3 Grossilepis spinosa Oryol Lebedev, 2004 Antiarcha B Antiarcha 4 Jiangxikepis China Zhang & Liu, longbrachius 1991 Antiarcha B Antiarcha 4 Ninjxialepis spinosa China Jia et al., 2010 Antiarcha B Antiarcha 5 Sinolepis szei China Pan et al., 1987 Antiarcha C Antiarcha 6 Walterilepis speciosa Ogre Luksevics et al., 2011 Arthrodira Arthrodira A Arthrodira 1 Coccosteomorphi indet Scat Craig Unpublished Arthrodira A Arthrodira 1 Kimberia heintzi Gogo Long & Trinajstic, 2010 Arthrodira A Arthrodira 2 Compagopisces Gogo Trinajstic & croucheri Hazelton,

49 2007; Long & Trinajstic, 2010 Arthrodira A Arthrodira 2 Gogopiscis gracilis Gogo Trinajstic & Hazelton, 2007; Long & Trinajstic, 2010 Arthrodira A Arthrodira 2 Harrytoombsia elegans Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 2 Kimberleyichthys Gogo Young, 1993; bispicatus Long & Trinajstic, 2010 Arthrodira A Arthrodira 2 Kimberleyichthys Gogo Young, 1993; whybrowni Long & Trinajstic, 2010 Arthrodira A Arthrodira 2 Macnamaraspis kaprios Gogo Long & Trinajstic, 2010 Arthrodira A Arthrodira 2 Plourdosteus Miguasha Vérizina, canadensis 1996; Cloutier, 2010 Arthrodira A Arthrodira 3 Simosteus tuberculatus Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 3 Torosteus pulchellus Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 3 Torosteus tuberculatus Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 4 Eastmanosteus Gogo Young, 1993; calliaspis Long & Trinajstic, 2010 Arthrodira A Arthrodira 5 Gogosteus sarahae Gogo Long & Trinajstic, 2010 Arthrodira A Arthrodira 5 Incisoscutum ritchiei Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 6 Camuropiscis Gogo Young, 1993; concinnus Long & Trinajstic, 2010

50 Arthrodira A Arthrodira 6 Camuropiscis laidlawi Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 6 Fallacosteus turneriae Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 6 Latocamurus coulthardi Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 6 Rolfosteus canningensis Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 6 Tubonasus lennardensis Gogo Long & Trinajstic, 2010 Arthrodira A Arthrodira 7 Bullerichthys fascidens Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira A Arthrodira 8 Bruntonichthys Gogo Young, 1993; multidens Long & Trinajstic, 2010 Arthrodira B Arthrodira 9 Holonema westolli Gogo Young, 1993; Long & Trinajstic, 2010 Arthrodira B Arthrodira 10 Selenosteidae indet. Greenland Blom et al., 2007 Arthrodira B Arthrodira 11 Kendrickichthys Gogo Young, 1993; cavernosus Long & Trinajstic, 2010 Arthrodira B Arthrodira 12 Pachyosteomorph indet. Oryol Lebedev, 2004 Arthrodira C Arthrodira 13 Pinguosteus thulborni Gogo Young, 1993; Long & Trinajstic, 2010 Petalichthyda Petalichthyda Petalichthyda Quasipetalichthys cf. China Pan et al., haikouensis 1987 Phlyctaenii Phlyctaenii A Phlyctaenii 1 Cosmacanthus Scat Craig Newman malcomsoni 2005 Phlyctaenii A Phlyctaenii 1 Groenlandaspis Greenland Stensiö & mirabilis Säve- Söderbergh 1938; Blom et al., 2007

51 Phlyctaenii A Phlyctaenii 1 Groenlandaspis RedHill Daeschler et pennsylvanica al 2003 Phlyctaenii A Phlyctaenii 1 Groenlandaspis thorezi Belgium Janvier & Clément, 2005 Phlyctaenii A Phlyctaenii 1 Groenlandaspis sp. 1 Jemalong Young, 1993 Phlyctaenii A Phlyctaenii 1 Groenlandaspis sp. 2 Canowindra Young, 1993 Phlyctaenii A Phlyctaenii 1 Turriaspis elector RedHill Daeschler et al., 2003 Phlyctaenii A Phlyctaenii 2 Heterostius Greenland Blom et al., groenlandicus 2007 Phyllolepida Phyllolepida Phyllolepida Phyllolepis nielseni Greenland Blom et al., 2007 Phyllolepida Phyllolepida Phyllolepis orvini Greenland Blom et al., 2007 Phyllolepida Phyllolepida Phyllolepis RedHill Lane & rossimontina Cuffey, 2005 Phyllolepida Phyllolepida Phyllolepis thomsoni RedHill Long & Daeschler, 2013 Phyllolepida Phyllolepida Phyllolepis undulata Belgium Lohest, 1888; Leriche, 1931 Phyllolepida Phyllolepida Phyllolepis sp. Jemalong Young, 1993 Ptyctodontida Ptyctodontida Ptyctodontida 1 Austropyyctodus Gogo Long & gardineri Trinajstic, 2010 Ptyctodontida Ptyctodontida 1 Materpisces Gogo Long & attenboroughi Trinajstic, 2010 Ptyctodontida Ptyctodontida 2 Campbellodus decipiens Gogo Young, 1993; Long & Trinajstic, 2010 Ptyctodontida Ptyctodontida 3 Chelyophorus sp. Oryol Lebedev, 2004 Acanthodii Climatiiformes Acanthodii A Acanthodii 1 Devononchus concinus Andreyevka Alekseev et al., 1994 Acanthodii A Acanthodii 1 Devononchus Pavari Luksevics & ketleriensis Zupins, 2004 Acanthodii A Acanthodii 1 Devononchus laevis Andreyevka, Alekseev et Ogre al., 1994 Acanthodii A Acanthodii 1 Devononchus cf. laevis Oryol Lebedev, 2004 Acanthodii A Acanthodii 1 Devononchus Pavari Gross 1933, tenuispinus Luksevics & Zupins, 2004 Acanthodii A Acanthodii 1 Devononchus cf. Belgium Derycke & tenuispinus Clément,

52 2013 Acanthodii A Acanthodii 1 Diplacanthus ellsi Miguasha Woodward 1892, Gagnier 1996; Cloutier, 2010 Acanthodii A Acanthodii 1 Diplacanthus horridus Miguasha Gagnier 1996 Acanthodii A Acanthodii 2 Gyracanthus cf. RedHill Elliott et al., sherwoodi 2000 Acanthodiformes Acanthodii B Acanthodii 3 Cheiracanthus sp. Andreyevka Alekseev et al., 1994 Acanthodii B Acanthodii 3 Homalacanthus Miguasha Gagnier, concinnus 1996; Cloutier, 2010 Acanthodii B Acanthodii 4 Halimacanthodes Gogo Burrow et al., ahlbergi 2012 Acanthodii B Acanthodii 5 Lodeacanthus gaujicus Lode Upeniece, 1996; Upeniece, 2001 Acanthodii B Acanthodii 5 Teneracanthus cf Belgium Derycke & toombaensis Clément, 2013 Acanthodii B Acanthodii 5 Triazeugacanthus Miguasha Béland & affinis Arsenault, 1985; Gagnier, 1996; Cloutier, 2010 Acanthodii B Acanthodii 6 Acanthodidae indet. Greenland Blom et al., 2007 Acanthodii indet Acanthodii C Acanthodii 7 Acanthodii indet. A ScatCraig Unpublished Acanthodii D Acanthodii 8 Acanthodii indet. B Pavari Luksevics & Zupins, 2004 Acanthodii E Acanthodii 9 Acanthodii indet. C Belgium Unpublished Acanthodii F Acanthodii 10 Acanthodii indet. D ScatCraig Unpublished Chondrichthyes Ctenacantiformes Chondrichthyes A Chondrichthyes 1 Ctenacanthus sp. RedHill Elliott et al., 2000 Chondrichthyes A Chondrichthyes 1 Tuberospina nataliae Oryol Lebedev, 1995 Chondrichthyes B Chondrichthyes 2 Symmorium sp. Oryol Lebedev, 1995 Chondrichthyes C Chondrichthyes 3 Ageleodus pectinatus RedHill Downs & Daeschler, 2001

53 Chondrichthyes D Chondrichthyes 4 Cladodus obtusus Andreyevka Alekseev et al., 1994 Chondrichthyes D Chondrichthyes 4 Cladodus sp. Greenland Blom et al., 2007 Chondrichthyes E Chondrichthyes 5 Chondrichthyes indet. A Andreyevka Alekseev et al., 1994 Chondrichthyes F Chondrichthyes 6 Chondrichthyes indet. B Greenland Blom et al., 2007 Chondrichthyes G Chondrichthyes 7 Chondrichthyes indet C Greenland Blom et al., 2007 Actinopterygii Actinopterygii A Actinopterygii 1 Cuneognathus gardineri Greenland Friedman & Blom, 2007 Actinopterygii A Actinopterygii 1 Limnomis delaney RedHill Daeschler, 2000 Actinopterygii A Actinopterygii 2 Moythomasia Gogo Gardiner & durgaringa Bartram 1977, Young, 1993 Actinopterygii A Actinopterygii 2 Moythomasia nitida Gogo Gardiner & Bartram 1977 Actinopterygii A Actinopterygii 2 Moythomasia sp. Andreyevka Alekseev et al., 1994 Actinopterygii B Actinopterygii 3 Gogosardina coatesi Gogo Choo et al., 2009 Actinopterygii B Actinopterygii 3 Mimipisces bartrami Gogo Choo, 2011 Actinopterygii B Actinopterygii 3 Mimipisces tombsi Gogo Gardiner & Bartram 1977, Young, 1993; Choo, 2011 Actinopterygii C Actinopterygii 4 Cheirolepis canadensis Miguasha Arratia & Cloutier 1996, Cloutier, 2010 Actinopterygii C Actinopterygii 4 Cheirolepis sp. Andreyevka Alekseev et al., 1994 Actinopterygii D Actinopterygii 5 Actinopterygii indet. A Pavari Luksevics & Zupins, 2004 Actinopterygii E Actinopterygii 6 Actinopterygii indet. B Belgium unpublished Actinopterygii F Actinopterygii 7 Actinopterygii indet. C Belgium Derycke & Clément, 2013 Sarcopterygii Actinistia Actinistia Actinistia 1 Miguashaia bureau Miguasha Cloutier, 1996, 2010 Actinistia Actinistia 1 Miguashaia grossi Ogre, Lode Forey et al., 2000; Upeniece, 2001 Actinistia Actinistia 2 Coelacanthidae indet. Lode Upeniece,

54 2001 Porolepiformes Porolepiformes A Porolepiformes 1 Duffichthys mirabilis ScatCraig Ahlberg, 1992 Porolepiformes A Porolepiformes 1 Holoptychius berganni Ellesmere Daeschler et al., 2006 ; Downs et al., 2013 Porolepiformes A Porolepiformes 1 Holoptychius flemingi Belgium Lohest, 1888 ; Leriche, 1931 Porolepiformes A Porolepiformes 1 Holoptychius jarviki Miguasha Cloutier & Schultze, 1966; Cloutier, 2010 Porolepiformes A Porolepiformes 1 Holoptychius Belgium Leriche, 1931 murchisoni Porolepiformes A Porolepiformes 1 Holoptychius cf. Ogre, Pavari Luksevics & nobilissimus Zupins, 2004; Luksevics et al., 2011 Porolepiformes A Porolepiformes 1 Holoptychius sp.1 Andreyevka Alekseev et al., 1994 Porolepiformes A Porolepiformes 1 Holoptychius sp. 2 RedHill Elliott et al., 2000 Porolepiformes A Porolepiformes 1 Holoptychius sp. 3 Greenland Blom et al., 2007 Porolepiformes A Porolepiformes 1 Holoptychius sp. 4 ScatCraig Ahlberg, 1998 Porolepiformes A Porolepiformes 1 Holoptychius sp. 5 China Pan et al., 1987 Porolepiformes A Porolepiformes 1 Holoptychius sp. 6 Miguasha Cloutier & Schultze Porolepiformes A Porolepiformes 1 Laccognathus embryi Ellesmere Downs et al., 2011 Porolepiformes A Porolepiformes 1 Laccognathus panderi Lode Upeniece, 2001 Porolepiformes A Porolepiformes 1 Quebecius quebecensis Miguasha Schultze & Arsenault, 1987, Cloutier & Schultze, 1966; Cloutier, 2010 Porolepiformes A Porolepiformes 2 Glyptolepis baltica Lode Gross, 1930; Upeniece, 2001 Porolepiformes A Porolepiformes 2 Glyptolepis ? dellei Oryol Gross, 1942; Lebedev, 1995 Porolepiformes A Porolepiformes 2 Glyptolepis sp. 1 Andreyevka Unpublished

55 Porolepiformes A Porolepiformes 2 Glyptolepis sp. 2 Ellesmere Daeschler et al., 2006 Porolepiformes A Porolepiformes 2 Glyptolepis sp. 3 Greenland Unpublished Porolepiformes A Porolepiformes 2 Ventalepis ketleriensis Pavari Luksevics & Zupins, 2004 Porolepiformes B Porolepiformes 3 Porolepiformes indet. Greenland Blom et al., 2007 Dipnoi Dipnoi A Dipnoi 1 Fleurentia denticulata Miguasha Cloutier, 1996, Cloutier 2010 Dipnoi A Dipnoi 1 Jarvikia lebedevi Oryol Krupina, 2000 Dipnoi A Dipnoi 1 Jarvikia sp. 1 Belgium Clément & Boisvert, 2006 Dipnoi A Dipnoi 1 Jarvikia sp. 2 Greenland Blom et al., 2007 Dipnoi A Dipnoi 1 Scaumenacia curta Miguasha Cloutier, 1996; Cloutier, 2010 Dipnoi A Dipnoi 2 Andreyevichthys Andreyevka Alekseev et epitomus al., 1994 Dipnoi A Dipnoi 2 Oervigia nordica Greenland Blom et al., 2007 Dipnoi A Dipnoi 2 Rhinopdipterus Gogo Clément, kimberleyensis 2012 Dipnoi B Dipnoi 3 Apatorhynchus RedHill Friedman & opistheretmus Daeschler, 2006 Dipnoi B Dipnoi 3 Conchodus excussus Oryol Krupina, 2000 Dipnoi B Dipnoi 3 Conchodus ostreiformis ScatCraig Ahlberg, 1998 Dipnoi B Dipnoi 3 Soerberghia Greenland, Young, 1993; groenlandica Jemalong, Blom et al., RedHill 2007 Dipnoi B Dipnoi 3 Soerberghia cf. Belgium Clément & groenlandica Boisvert, 2006 Dipnoi B Dipnoi 3 Soerberghia simpsoni Canowindra Ahlberg et al., 2001 Dipnoi B Dipnoi 3 Soerberghia sp. Belgium Clément & Boisvert, 2006 Dipnoi C Dipnoi 4 Orlovichthys cf. Pavari Luksevics & limnatis Zupins, 2004 Dipnoi D Dipnoi 5 Dipterus arcanus Oryol Krupina, 2000 Dipnoi D Dipnoi 5 Dipterus expressus Oryol Krupina, 2000 Dipnoi D Dipnoi 5 Dipterus fourmarieri Belgium Leriche, 1931

56 Dipnoi D Dipnoi 5 Dipterus nelsoni Belgium Leriche, 1931 Dipnoi D Dipnoi 5 Dipterus marginalis Oryol Krupina, 2000 Dipnoi D Dipnoi 5 Dipterus cf. marginalis Ogre Luksevics et al., 2011 Dipnoi D Dipnoi 5 Dipterus pacatus Oryol Krupina, 2000 Dipnoi D Dipnoi 5 Gogodipterus Gogo Long, 1992, paddyensis Young, 1993 Dipnoi E Dipnoi 6 Chirodipterus australis Gogo Young, 1993 Dipnoi E Dipnoi 6 Chirodipterus Oryol Krupina, interstitus 2000 Dipnoi F Dipnoi 7 Grossipterus venustus Oryol Krupina, 2000 Dipnoi G Dipnoi 8 Asthenorhynchus Gogo Long & meemannae Trinajstic, 2010 Dipnoi G Dipnoi 8 Holodipterus gogoensis Gogo Young, 1993; Long & Trinajstic, 2010 Dipnoi G Dipnoi 8 Holodipterus sp. Oryol Lebedev, 2004 Dipnoi G Dipnoi 8 Xeradipterus hatcherii Gogo Clément & Long 2010 Dipnoi G Dipnoi 9 Griphognathus whitei Gogo Young, 1993 Dipnoi G Dipnoi 9 Robinsondipterus longi Gogo Long & Trinajstic, 2010 Dipnoi H Dipnoi 10 Dipnoi indet. A RedHill Friedman & Daeschler, 2006 Dipnoi I Dipnoi 11 Dipnoi indet. B RedHill Friedman & Daeschler, 2006 Dipnoi J Dipnoi 12 Dipnoi indet. C RedHill Friedman & Daeschler, 2006 Dipnoi K Dipnoi 13 Dipnoi indet. D RedHill Friedman & Daeschler, 2006 Dipnoi L Dipnoi 14 Dipnoi indet. E RedHill Friedman & Daeschler, 2006 Dipnoi M Dipnoi 15 Dipnoi indet. F RedHill Friedman & Daeschler, 2006 Dipnoi N Dipnoi 16 Dipnoi indet. G ScatCraig Ahlberg, 1998 Dipnoi O Dipnoi 17 Dipnoi indet. H Andreyevka Unpublished Dipnoi P Dipnoi 18 Dipnoi indet. I Ogre, RedHill, Unpublished Oryol, Belgium Dipnoi Q Dipnoi 19 Dipnoi indet. J Lode Upeniece,

57 2001 Struniformes Struniformes A Struniformes 1 Latvius sp. Lode Upeniece, 2001 Struniformes A Struniformes 2 Onychodus jaekeli Lode Upeniece, 2001 Struniformes A Struniformes 2 Onychodus jandemarrai Gogo Long & Trinajstic, 2010 Struniformes A Struniformes 2 Strunius cf. rolandi Oryol Lebedev, 1995 Struniformes A Struniformes 2 Strunius sp. Andreyevka Alekseev et al., 1994 Struniformes B Struniformes 3 Pycnacanthus fischeri Oryol Lebedev, 1995 Osteolepiformes Osteolepididae Osteolepididae A Osteolepididae 1 Cryptolepis grossi Oryol, Pavari Lebedev, 1995; Luksevics & Zupins, 2004 Osteolepididae A Osteolepididae 1 Glyptopomus ? bystrowi Oryol, Pavari Lebedev, 1995; Luksevics & Zupins, 2004 Osteolepididae A Osteolepididae 1 Osteolepididae cf. Belgium Unpublished Glyptopomus sp. Osteolepididae A Osteolepididae 1 Osteolepididae cf. RedHill Unpublished Litoptychius sp. Osteolepididae A Osteolepididae 2 Gogonasus andrewai Gogo Long 1985; Young, 1993 Osteolepididae B Osteolepididae 3 Osteolepididae g. nov. Andreyevka Unpublished sp. 1 Osteolepididae C Osteolepididae 4 Osteolepididae g. nov. Andreyevka Unpublished sp. 2 Osteolepididae D Osteolepididae 5 Osteolepididae indet. 1 Lode Upenice, 2001 Osteolepididae E Osteolepididae 6 Osteolepididae indet. 2 Belgium Unpublished Osteolepididae F Osteolepididae 7 Canowindrae grossi Canowindra Thomson, 1973 Osteolepididae G Osteolepididae 8 Sarcopterygii indet. Belgium Derycke & Clément, 2013 Megalichthydae Megalichthydae Megalichthydae 1 Megapomus markovskyi Oryol Lebedev, 2004 Megalichthydae Megalichthydae 2 Megalichthydae indet.1 Belgium Unpublished Megalichthydae Megalichthyidae 2 Megalichthydae indet. 2 RedHill Elliott et al., 2000 Tristichopterydae Tristichopterydae Tristichopterydae Cabonnichthys burnsi Canowindra Ahlberg & A 1 Johanson,

58 1997 Tristichopterydae Tristichopterydae Eusthenodon wangsjoi Belgium, Elliott et al., A 1 Greenland, 2000 ; RedHill Clément, 2002, Blom et al., 2007 Tristichopterydae Tristichopterydae Eusthenodon sp. 1 Andreyevka, Alekseev et A 1 RedHill al., 1994, Elliott et al., 2000 Tristichopterydae Tristichopterydae Eusthenodon sp. 2 Ogre Luksevics et A 1 al., 2011 Tristichopterydae Tristichopterydae Hyneria lindae RedHill Elliott et al., A 1 2000 Tristichopterydae Tristichopterydae Langlieria socqueti Belgium Clément et A 1 al., 2009 Tristichopterydae Tristichopterydae Mandageria fairfaxi Canowindra Johanson & A 1 Ahlberg, 1997 Tristichopterydae Tristichopterydae Tristichopterydae sp. Greenland Blom et al., A 1 nov. 2007 Tristichopterydae Tristichopterydae Platycephalichthys Ogre Luksevics et A 2 bishoffi al., 2011 Tristichopterydae Tristichopterydae Eusthenopteron foordi Miguasha Arsenault, B 3 1982 Tristichopterydae Tristichopterydae Eusthenopteron kurshi Lode Upeniece, B 3 2001 Tristichopterydae Tristichopterydae Callistiopterus clappi Miguasha Schultze, B 3 1996 Tristichopterydae Tristichopterydae Tristichopterydae indet. China Unpublished C 4 1 Tristichopterydae Tristichopterydae Tristichopterydae indet. Belgium Unpublished D 5 2 Tristichopterydae Tristichopterydae Tristichopterydae indet. Belgium Unpublished E 6 3 Rhizodontida Rhizodontida A Rhizodontida 1 Gooloogongia loomesi Canowindra Johanson & Ahlberg, 1998 Rhizodontida A Rhizodontida 2 Sauripterus taylori RedHill Davis et al., 2004 Rhizodontida A Rhizodontida 2 Sauripterus indet. RedHill Unpublished Rhizodontida B Rhizodontida 3 Rhizodontidae indet. 1 RedHill Unpublished Rhizodontida B Rhizodontida 4 Rhizodontidae indet. 2 ScatCraig Unpublished Rhizodontida B Rhizodontida 5 Rhizodontidae indet. 3 Belgium Unpublished Elpistostegalia Elpistostegalia A Elpistostegalia 1 Elpistostege watsoni Miguasha Westoll, 1938, Cloutier, 2010 Elpistostegalia B Elpistostegalia 2 Tiktaalik roseae Ellesmere Daeschler et al., 2006

59 Elpistostegalia C Elpistostegalia 3 Panderichthys Lode Gross, 1941; rhombolepis Upeniece, 2001 Tetrapoda Tetrapoda A Tetrapoda 1 Watcheridae indet. RedHill Shubin et al., 2004; Daeschler et al., 2009 Tetrapoda B Tetrapoda 2 Tulerpeton curtum Andreyevka Lebedev, 1984 Tetrapoda B Tetrapoda 2 Tetrapoda indet. 1 Andreyevka Lebedev & Clack, 1993 Tetrapoda C Tetrapoda 3 Ichthyostega eigili Greenland Säve- Söderbergh 1932; Blom et al., 2007; Tetrapoda C Tetrapoda 3 Ichthyostega stensioei Greenland Säve- Söderbergh 1932; Blom et al., 2007 Tetrapoda C Tetrapoda 3 Ichthyostega watsoni Greenland Säve- Söderbergh 1932; Blom et al., 2007 Tetrapoda C Tetrapoda 3 Ichthyostega sp. Belgium Clément et al., 2004 Tetrapoda D Tetrapoda 4 Ymeria denticulara Greenland Clack et al., 2012 Tetrapoda E Tetrapoda 5 Acanthostega gunnari Greenland Jarvik 1952, Blom et al., 2007 Tetrapoda F Tetrapoda 6 Metaxygnathus Jemalong Campbell & denticulatus Bell, 1977 Tetrapoda G Tetrapoda 7 Ventastega curonica Pavari Ahlberg et al., 1994, Luksevics & Zupins, 2004 Tetrapoda G Tetrapoda 7 Tetrapoda indet. 2 Pavari Ahlberg et al., 1994 Tetrapoda H Tetrapoda 8 Densignathus rowei RedHill Daeschler, 2000 Tetrapoda H Tetrapoda 8 Hynerpeton basseti RedHill Daeschler et al., 1994 Tetrapoda I Tetrapoda 9 Jakubsonia livnensis Oryol Lebedev, 2004 Tetrapoda J Tetrapoda 10 Sinostega pani China Zhu et al., 2002 Tetrapoda K Tetrapoda 11 Obruchevichthys Ogre Vorobyeva gracilis 1977 Tetrapoda L Tetrapoda 12 Elginerpeton pancheni ScatCraig Ahlberg et al., 1995 List of characters used in the third analysis, (after the name of taxa in parentheses numbers of characters used in the analysis):

60 For anthiarch placoderms based on description in Moloshnikov, 2008 (1-6): 1 Presence of the dorsal fin: 0 absent, 1 present 2 Elongated pectoral fins: 0 absent, 1 present 3 Dorsal fin: 0 absent, 1 present 4 Anal fins: 0 absent, 1 present 5 Heterocercal caudal fin: 0 absent, 1 present 6 Surface sculpturing of bones: 0 with tubercules, 1 reticular

For arthrodir placoderms, characters from Carr & Hlavin, 2010 (7-16): 5. Rostral plate posterior development: not developed (0); developed (1). 36. Median dorsal plate shape (width/length): short and broad to equidimensional (0); body of plate long and narrow, not including a posteriorly produced spine (1). 45. Spinal plate (Stensiö, 1959): absent (0); present (1). 47. Interolateral plate branchial lamina: absent (0); present (1). 48. Interolateral plate ventral lamina: small or absent (0); developed with either an enlarged contact with the anterior median ventral plate, or an enlarged overlap with the anterior ventrolateral plate (1). 49. Interolateral plate branchial lamina ornamentation: absent (0); present (1). 63. Inferognathal with a developed blade portion (Dennis & Miles, 1979a: modified after character 13); absent (0); present (1). 64. Inferognathal anterior dental field developed anteriorly of the cusp region: absent (0); present (1), indicated by a projection of the inferognathal anterior to the dorsal functional region. 90. Skull roof and thoracic armour articulation (Goujet, 1984): articulation medial (0); articulation laterally displaced (1). 91. Dermal ornamentation: absent (0); present (1).

For phyllolepid placoderms based on description of taxa (see the taxa list) (17-21): 1 Ornementation with tubercules: 0 absent, 1present 2 Median dorsal plate: flatterd 0, elongated 1 3 Ornementation: weakly developed or absent 0, well developed 1 4 Denticles on the posterior margin of the MD plate: absent 0, present 1 5 The inverted V-shaped pattern of the main lateral-line: absent 0, present 1groove on the posterior dorsolateral plate

For groenlandaspid placoderms based on description of taxa (see the taxa list) (22-25): 1 Ornementation constituted by concentric ridges: 0 absent, 1 present 2 Posterior median ventral plate: present 0, absent 1 3 Posterior dorsolateral plate: present 0, absent 1 4 Preorbital plates meet mesially: yes 0, no 1

For actinopterygians from Choo, 2011 (26-41): 5. Rostral (or equivalent ethmoid bone) with transverse ornamentation: 0=absent; 1=present. 7. Posterior nostril – contribution to margin by premaxillae: 0=absent; 1=present. 9. Pineal foramen: 0=present; 1=absent. 15. Extrascapulae – number of paired examples: 0=one pair; 1=two pairs. 23. Number of infradentaries: 0=two (angular and surangular); 1=one (angular only). (modified) 26. Porous ganoine ornamentation on lower jaw: 0=absent; 1=present. 29. Parasphenoid – anterior tripartite subdivision: 0=absent; 1=present. 31. Scale rows, number of: 0=fewer than 60; 1=more than 60. 35. Scales with well-developed pores/punctae on ganoine surface: 0=absent; 1=present. 38. Pelvic fins: 0=present; 1=absent or greatly reduced. 44. Pectoral fin segmentation: 0=roughly even segmentation to fin-base; 1=proximal segments

61 elongated with terminal segmentation (modified) 58. Quadratojugal as distinct ossification: 0=present; 1= absent. 66. Serrated linear ganoine ornament on skull roof: 0=absent; 1=present. 70. Extrascapulae – medially-directed branch of sensory canals: 0=present; 1=absent. 72. Branchiostegal rays – 1st ray ventral to suboperculum: 0=of similar depth to adjacent branchiostegal ray; 1=deeper than 2nd branchiostegal ray. 76. Scales – anastomosing linear ornament: 0=absent; 1=present.

For lungfishes, characters from Ahlberg et al., 2006 (42-57): 1. Teeth: 0. Present, 1. Absent 2. Addition of large dentine elements at regular intervals to lateral margin of pterygoid/prearticular: 0. Yes, 1. No 6. Addition of marginal blisters to pterygoid/prearticular: 0. No, 1. Yes 8. Addition of inter-row dentine along edge of pterygoid/ prearticular: 0. No 1. Yes 10. Diffuse dentine deposition on surface of palate/lower jaw: 0. Yes, diffusely across whole palate 1. No 14. Distinct vertically growing ‘‘heel’’ on prearticular: 0. No 1. Yes 21. Dental material on parasphenoid: 0. present 1. Absent 22. Position of pterygoid toothplate centre relative to first and last tooth rows: 0. not recessed, tooth rows essentially straight and diverging by 1001 or less 1. Slightly recessed into curving posteromesial margin of tooth plate 30. Premaxilla: 0. Present 1. Absent 33. Autostyly: 0. Absent 1. Present 35. Coronoids: 0. Present 1. Absent 36. Lip fold: 0. Absent 1. Present 42. Curvature of ventral mandibular margin: 0. strongly convex 1. essentially flat 48. Ossified dentary in adult: 0. Present 1. Absent 70. Elongated snout: 0. Absent 1. Present 73. Scale morphology (ordered): 0. semi-rounded, cosmine-covered 1. Rounded, no cosmine

For osteolepiforms (used for osteolepids, megalichtyds, tritischopteryds, these groups were separated for the analysis), characters from Snitting, 2008 (58-77, 78-97, 98-117): (6) Coronoid proportions: Posterior coronoid similar in length to, or shorter than middle coronoid (0); posterior coronoid longer than middle coronoid (1). (7) Marginal denticle band on coronoids: narrower than fang bearing lamina (0); broader than fang- bearing lamina, at least posteriorly (1). (8) Dentary fang pair: absent (0); present (1). (9) Number of fang positions on posterior coronoid: one (0); two (1); none (2). (10) Tooth folding: generalized polyplocodont (0); labyrinthodont (1); dendrodont (2). (11) Number of fang positions on ectopterygoid: one (0); two (1); none (2). (18) Enlarged anterior tooth on premaxilla: absent (0); present (1). (22) Cosmine: present (0); absent (1). (30) Number of nasals: more than two (0); one or two (1). (31) Frontals: absent (0); present (1). (34) Shape of maxilla: distinct posterodorsal process (0); posterodorsal process weak or absent (1). (41) Position of orbits: lateral and widely separated (0); dorsal and close together (1). (46) Contact margin for clavicle on cleithrum: straight or faintly convex (0); strongly concave (1). (47) Anocleithrum: subdermal (0); exposed (1). (49) Dorsal and anal fins: present (0); absent (1). (50) Caudal fin: heterocercal (0); diphycercal (1). (51) Epichordal radials in caudal fin: absent (0); present (1). (52) Well-ossified ribs: absent (0); present (1).

62 (60) Choana: absent (0); present (1). (61) Pineal foramen/eminence surrounded by parietals: no (0); yes (1).

For elpistostegids and tetrapods, characters from Clack et al., 2012 (118-140): 34. Palatine row of smaller teeth: present = 0, absent = 1 36. Premaxillary tooth proportions: all approximately same size = 0, posteriormost teeth at least twice height of anteriormost teeth = 1 46. Coronoid (middle) separated from splenial: yes, by prearticular = 0, no = 1, yes, by postsplenial = 2 47. Coronoid (middle) contacts postsplenial: no = 0, yes = 1 50 Dentary external to angular + surangular, with chamfered ventral edge and no interdigitations: no = 0, yes = 1 51. Dentary ventral edge: smooth continuous line = 0, abruptly tapering or ‘stepped’ margin = 1 52. Dentary suture with splenial + postsplenial marked by deep furrow: no = 0, yes = 1 53. Mandibular sensory canal: present = 0, absent = 1 55. Mandible: oral sulcus/surangular pit line: present = 0, absent = 1 68. Prearticular with longitudinal ridge below coronoids: no = 0, yes = 1 73. Coronoids: at least one has fang pair recognisable because at least twice the height of coronoid teeth: yes = 0, no = 1 74. Coronoids: at least one has fangs recognisable because noticeably mesial to vertical lamina of bone and to all other teeth: yes = 0, no = 1 75. Coronoids: at least one has organised tooth row: yes = 0, no =1 76. Coronoids: at least one carries shagreen: no = 0, yes = 1 77. Coronoids: size of teeth (excluding fangs) on anterior and middle coronoids relative to dentary tooth size: about the same = 0, half height or less = 1 78. Dentary tooth row: homodont = 0, markedly heterodont = 1 79. Dentary with parasymphysial fangs internal to marginal tooth row: yes = 0, no = 1 80. Dentary teeth: same size as maxillary teeth = 0, smaller than maxillary teeth = 1 (modified) 81. Dentary with a row of very small teeth or denticles lateral to tooth row: yes = 0, no = 1 82. Parasymphysial tooth plate: present = 0, absent = 1 92. Nature of ornament: ‘starbursts’ of radiating ornament on at least some bones: no = 0, yes = 1 96. Digits: absent = 0, present = 1 111. Squamation: complete body covering of scales, all similar = 0, ventral armour of gastralia = 1 112. Pectoral process: absent = 0, present = 1.

Data matrix of the third analysis

10 20 30 40 50 60 Jemalong 1110110000 0000001000 1111000000 0000000000 0000000001 1111011000 Pavari 1110110000 0000000000 0000000000 0000000000 0000010001 1111011000 Ogre 1110110000 0000000000 0000000000 0000000000 0000100101 1111001000 Greenland 1110111111 0110111010 1111110001 0000110011 0010000001 1111011000 Andreyvka 1110110000 0000000000 0000000010 1001011111 0000110101 1111111000 RedHill1 110110000 0000001110 1111010000 0000110011 0000000001 1111011000 Oryol 1110110000 0000000000 0000000000 0000000000 0110111101 1111001000 ScatCraig 1110111111 1110111000 1000000000 0000000000 0000000001 1111001000 China 1010110000 0000000000 0000000000 0000000000 0000000000 0000000000 Belgium 1110110000 0000001101 1111000000 0000000000 0000000101 1111011000 Gogo 0000001111 1110110000 0000001011 1101011111 0111111011 1111001000 Miguasha 1110111111 1110110000 0000000101 0010001011 1000010101 1111111000 Ellesmere 0000100000 0000000000 0000000000 0000000000 0000000000 0000000000 Lode 0000100000 0000000000 0000000101 0010001011 1000000001 1111001000

63 Canowindra1100110000 0000001000 1000000000 0000000000 0000000001 1111011000 JeCa 1110110000 0000001000 1111000000 0000000000 0000000001 1111011000

80 90 100 110 120 130 Jemalong 0000000000 0000000000 0000000000 0000000000 0000000000 0000000000 Pavari 0000000000 1000011000 0000000000 0000000000 0000000000 0000000010 Ogre 0000000000 0000000000 0000000000 0000000101 1011100101 1011111000 Greenland 0000000000 0000000000 0000000000 0000000101 1011100101 1011111010 Andreyvka 0000000000 1000011000 0000000000 0000000101 1011100101 1011111010 RedHill1 0000000000 1000011001 0001000000 1000011101 1011100101 1011111010 Oryol 0000000000 1000011001 0001000000 1000011001 0001100001 1000011000 ScatCraig 0000000000 0000011000 0000000000 0000001000 0000000000 0000000000 China 0000000000 0000000000 0000000000 0000000100 1010100001 1011111000 Belgium 0000000000 1000011001 0001000000 1000011101 1011100101 1011111010 Gogo 0000000000 1000011000 0000000000 0000000000 0000000000 0000000000 Miguasha 0000000000 0000000000 0000000000 0000000100 1010100001 1010111000 Ellesmere 0000000000 0000000000 0000000000 0000000000 0000000000 0000000000 Lode 0000000000 1000011000 0000000000 0000000100 1010100001 1010111000 Canowindra0000000000 0000000000 0000000000 0000000101 1011100101 1011111000 JeCa 0000000000 0000000000 0000000000 0000000101 1011100101 1011111000

130 141 Jemalong 1000000100 10011011000 Pavari 0100001010 01000101100 Ogre 0000000000 00000001101 Greenland 0100011110 01001101111 Andreyevka 0100011110 01001101111 RedHill1 0100001110 01001101111 Oryol 0100000010 01001101100 ScatCraig 0001001100 00000001101 China 0000000000 00000000000 Belgium 0100011110 01001101111 Gogo 0000000000 00000000000 Miguasha 0100000000 00000000000 Ellesmere 0000000000 00000100000 Lode 0000000000 00000000000 Canowindra 0000000000 00000000000 JeCa 0100000010 01001101100

Figure and Table captions: Figure 1: Map showing the near-tetrapod and tetrapod localities world-wide. Green star represents Gogo station locality, yellow stars represent the elpistostegid localities; orange stars represent Frasnian tetrapod localities; red stars represent Famennian localities, brown stars represent Frasnian/Famennian localities.

Figure 2: Map of East Greenland showing the Devonian outcrops (Blom et al., 2007).

Figure 3: Map showing the localization of Strud locality in Belgium (Gueriau et al., 2014).

Figure 4: Map showing the localization of Durnal locality in Belgium (Clément et al., 2009).

Figure 5: Map showing the location of the Red Hill site, Clinton County, Pennsylvania, USA. (Cressler III et al., 2010)

64 Figure 6: Map showing the localization of Andreyevka (A) Location of the Tula region in the European part of Russia; (B) Location of Andreyevka locality in the Tula region; (C) Andreyevka locality, the black circles represent the position of outcrop. (Alekseev et al., 1994)

Figure 7: Maps showing the localization of the Latvian localities (a) Map of Latvia; the area of light shading indicates the extent of Famennian and later deposits, the square of dark shading is the area of the next map. (b) Detail map of the Skrunda area showing the localities of Pavari and Ketleri. The house indicate Pavari and Ketleri hamlets, the stylized fishes mark the fossil localities. Crosses, Zagare Formation; no shading, Ketleri Formation; light shading, Skervelis Formation. The fossil locality at Ketleri lies just below the top of the Ketleri Formation. Scale bar 3 km. (Ahlberg et al., 1994)

Figure 8: Map showing the localization of the locality in Russia, an asterisk shows the location of the Gornostayevka quarry in the Oryol region, a black diamond marked the location of Andreyevka. (Lebedev, 2004)

Figure 9: Map of Australia showing the localization of Jemalong and Canowindra locality, fossil site indicated by triangles (from Johanson et al., 2003 modified).

Figure 10: Map showing the location of Elgin in Scotland (Ahlberg, 1998).

Figure 11: A, B Localization of Langsede site in Europe and in Latvia; C Star indicate location of Langsede site near Sabile (Luksevics et al., 2011).

Figure 12: (A) Location of Ningxia Hui Autonomous Region; (B) Devonian outcrops in Ningxia; (C) Correlation of the Zhongning Formation in four selected Devonian sections mainly based on Pan et al., 1987 (Jia et al., 2010).

Figure 13: Map showing the localization of Lode quarry in Latvia and the stratigraphic column of the Lode Formation (there is uncertainty about the position of the Givetian-Frasnian boundary) (Forey et al., 2000).

Figure 14: Location and stratigraphic position of the site (NV2K17) on southern Ellesmere Island, Nunavut Territory, Canada. (Daeschler et al., 2006)

Figure 15: a. Localization of Miguasha in eastern Québec, b. Geological map of the Miguasha area (Miguasha Group includes the Fleurant and Escumiac Formations), c. General view of the east side of the syncline of the Escumiac Formation (see in b) (Cloutier et al., 2011).

Figure 16: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the species level when Canowindra and Jemalong localities are considered as separated.

Figure 17: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the species level when Canowindra and Jemalong localities are considered as grouped.

Figure 18: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the genus level when Canowindra and Jemalong localities are considered as separated.

Figure 19: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the genus level when Canowindra and Jemalong localities are considered as grouped.

65 Figure 20: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the first artificial level when Canowindra and Jemalong localities are considered as separated.

Figure 21: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the first artificial level when Canowindra and Jemalong localities are considered as grouped.

Figure 22: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the second artificial level when Canowindra and Jemalong localities are considered as separated.

Figure 23: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the second artificial level when Canowindra and Jemalong localities are considered as grouped.

Figure 24: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the morphological level when Canowindra and Jemalong localities are considered as separated.

Figure 25: Dendrogram showing faunal relationship between near-tetrapod and tetrapod localities during the Late Devonian for the morphological level when Canowindra and Jemalong localities are considered as grouped.

Figure 26: Late Devonian tetrapod and near-tetrapod localities plotted on palaeogeographical reconstruction from Averbuch et al., 2005, showing the active orogenic systems at the Frasnian/Famennian boundary. The numbers design localities: 1–Catskill, 2– Greenland, 3–ScatCraig, 4–Belgium, 5–6–Pavari and Ogre, 7– Andreyevka-2, 8– Oryol, 9–China, 10–Jemalong and Canowindra. The black point represents the near-tetrapod localities: on the top Ellesmere, near the locality 3 Miguasha, near the locality 5 Lode. Abbreviations: C.A.b., Central Asian belt; NC, North China; SC, South China; T, Tarim (modified from Blieck et al., 2007).

66 Tidigare utgivna publikationer i serien ISSN 1650-6553

Nr 1 Geomorphological mapping and hazard assessment of alpine areas in Vorarlberg, Austria, Marcus Gustavsson

Nr 2 Verification of the Turbulence Index used at SMHI, Stefan Bergman

Nr 3 Forecasting the next day’s maximum and minimum temperature in Vancouver, Canada by using artificial neural network models, Magnus Nilsson

Nr 4 The tectonic history of the Skyttorp-Vattholma fault zone, south-central Sweden, Anna Victoria Engström

Nr 5 Investigation on Surface energy fluxes and their relationship to synoptic weather patterns on Storglaciären, northern Sweden, Yvonne Kramer

Nr 291 Inläckage till spillvattennät i östra Knivsta - Volymbestämning med hjälp av beräkningar, modellering och fältarbete, Johan Hamelius, Juni 2014

Nr 292 Single coccolith weight estimates using two birefringence methods on cultured Gephyrocapsa oceanica, Annalena Antonia Lochte, Juli 2014

Nr 293 Kan den hydrologiska modellen S-HYPE användas för att beräkna grundvattennivåer med tillräcklig noggrannhet?, Hanna Lagergren, September 2014

Nr 294 One- and Three-dimensional P- and S-wave Velocity Models of Central and Southern Sweden Based on SNSN Data , Ne Xun Chan, September 2014

Nr 295 Tooth Replacement of Euhelopus zdanskyi (Dinosauria: Sauropoda) and the Evolution of Titanosaurian Tooth Morphology, Seela Salakka, September 2014

Nr 296 Processing and Modeling of Gravity, Magnetic and Electromagnetic Data in the Falkenberg Area, Sweden, Soroor Mohammadi, September 2014

Nr 297 Calving Front Dynamics: External Forces that Lead to Specific Sized Calving Events, Daniel Wainwright, September 2014

Nr 298 Atmospheric Mercury Deposition in Alpine Snow, Northern British Columbia and Southwestern Yukon, Canada, Pia Karlsson, September 2014

Nr 299 Mobilization and Transport of Different Types of Carbon-based Engineered and Natural Nanoparticles through Saturated Porous Media, Maryeh Hedayati, Oktober 2014

Nr 300 Surface Climatology of the Icefield Ranges, St. Elias Mountains, Yukon Territory, Canada, Justinas Baltrenas, Oktober 2014