1 Pour l`obtention du Grade de DOCTEUR DE L`UNIVERSITE DE
Transcription
1 Pour l`obtention du Grade de DOCTEUR DE L`UNIVERSITE DE
THESE Pour l’obtention du Grade de DOCTEUR DE L’UNIVERSITE DE POITIERS (Faculté des Sciences Fondamentales et Appliquées) (Diplôme National - Arrêté du 7 août 2006) Ecole Doctorale : Sciences pour l’Environnement Gay Lussac. Secteur de Recherche : Présentée par : Benoit GANGLOFF ************************ Systematics and Phylogeography in gadfly petrels (Aves: Procellariiformes) and implications for conservation Systématique et phylogeographie chez les ptérodromes (Aves: Procellariiformes) et implications pour la conservation ************************ Directeur de Thèse : Vincent BRETAGNOLLE ************************ Soutenue le 17 décembre 2010 devant la Commission d’Examen ************************ JURY , Président Richard CORDAUX, Université de Poitiers Examinateur Michael de L. BROOKE, Cambridge University Rapporteur Alice CIBOIS, Muséum d’Histoire Naturelle de Genève Examinatrice Pierre-André CROCHET, CEFE, Montpellier Examinateur Vincent BRETAGNOLLE, CNRS Chizé Directeur de thèse 1 RESUME Les Procellariiformes sont des oiseaux marins présentant des caractéristiques les rendant particulièrement attractifs pour la recherche. Ces oiseaux constituent le groupe d’oiseaux marins le plus diversifié et probablement le plus menacé, 44.7% des espèces de Procellariiformes étant classées Vulnérables ou En Danger d’Extinction par l’UICN. Comme d’autres oiseaux marins, les Procellariiformes constituent des modèles fascinants pour l’étude des patrons de différentiation des populations et des espèces puisqu’étant confrontés à l’action de forces évolutives opposées : d’un côté leur extrême mobilité leur permet de disperser très loin, ce qui est supposé permettre et augmenter le flux de gènes entre populations et donc atténuer la différentiation des populations ; de l’autre côté, étant extrêmement philopatriques ils reviennent généralement se reproduire dans leur colonie de naissance, une caractéristique censée promouvoir la différentiation des populations. Ajoutées au fait qu’ils vivent dans l’océan, ces caractéristiques les rendent particulièrement intéressants pour étudier les processus de différentiation en l’absence de barrières physiques aux flux de gènes. Par ailleurs, la taxonomie et les relations phylogénétiques chez les Procellariiformes sont complexes et mal établies, entre autres en raison de l’action opposée des forces évolutives citées précédemment. Les limites d’espèces et la taxonomie de cet ordre ont donc continuellement changé au fil des années. Au cours des 20 dernières années, les développements observés dans le domaine de la biologie moléculaire ont fournis des outils de plus en plus puissants pour élucider certaines incertitudes phylogénétiques, comprendre les patrons phylogéographiques observés et avoir une meilleure compréhension des processus promouvant la différentiation des populations chez ces oiseaux. Ces outils permettent aussi de mieux comprendre la structuration des populations et les relations entre ces dernières ce qui peut grandement aider à la définition des actions de conservation entreprises pour ces organismes et à leur ordre de priorité. Dans cette thèse, à l’aide d’outils de biologie moléculaire, j’ai étudié les relations phylogénétiques et la phylogéographie de plusieurs taxons de la famille des Procellariidae, la plus riche en espèce chez les Procellariiformes. Cette étude à portée sur deux niveaux taxonomiques : premièrement, au niveau générique, cette étude décrit les relations phylogénétiques du genre Pseudobulweria, probablement le genre d’oiseaux marins le plus menacé au monde ; deuxièmement, au niveau du taxon et des populations l’accent a été porté sur les ptérodromes de Macaronésie et sur le pétrel de Gould, en particulier concernant les relations entre ses sous-espèces australiennes et néo-calédoniennes. i L’utilisation d’analyses multilocus pour chacune de ces investigations a permis d’améliorer notre compréhension et connaissances des ces taxons : chez Pseudobulweria la construction d’un « arbre d’espèce » a permis de définir les relations phylogénétiques au sein du genre et de résoudre la question du statut taxonomique du pétrel de Beck ; chez les ptérodromes de Macaronésie, j’ai montré que la séparation des trois taxons est beaucoup plus récente que précédemment estimée et que ces populations doivent être considérées comme des Unités Evolutives Significatives ; chez le pétrel de Gould, cette étude a montré que les deux sous espèces ne sont pas différenciées génétiquement et que ces deux lignées ne sont pas séparées. L’ensemble de ces résultats obtenus par l’utilisation de plusieurs gènes nucléaires en plus de gènes mitochondriaux souligne l’importance de ce type d’approche pour l’étude des patrons phylogénétiques et phylogéographiques pour les comprendre dans toute leur complexité. MOTS CLES Taxonomie, phylogéographie, phylogénie, pétrel, Procellariiforme, conservation, biologie moléculaire, Pterodroma, Pseudobulweria, ii ABSTRACT Procellariiformes are seabirds showing a set of characters rendering them particularly attractive to research. They are the most diverse seabird group, and probably also the most threatened, with 44.7% species classified Vulnerable or worse under IUCN criteria. As many seabirds, they are fascinating models to study patterns and processes of population and species differentiation, being under contradictory evolutionary forces: on the one hand being extremely vagile allow them to disperse very far, which is supposed to enhance gene flow between populations, thus reducing their differentiation and diversification; on the other hand, being extremely philopatric, they often return to breed in their natal colony, a pattern supposed to enhance population differentiation. Living in the ocean, they also constitute good models to investigate differentiation processes in the absence of physical barrier to gene flow. In addition, partly as a result of the opposite evolutionary forces just described, their taxonomy and phylogenetic relationships are complex, have proved very frustrating over the decades and have therefore been in a state of flux over the years. The development of molecular ecology in the last two decades have provided some new powerful tools to elucidate some of the phylogenetic uncertainties, to understand the observed phylogeographic patterns and have a better grasp at the underlying processes promoting diversification in these birds. These tools also allow a better understanding of population structure and relationships and can greatly help to the prioritisation and design of conservation actions directed at conserving these organisms. In this thesis, by means of molecular ecology tools, I investigated the phylogenetic relationships and phylogeography of several taxa belonging to family Procellariidae, the most speciose in order Procellariiformes. I studied these at two levels: first at the genus level by describing phylogenetic relationships in genus Pseudobulweria, probably the most endangered seabird genus in the world; and second at the taxon and population level I focused on the Macaronesian group of gadfly petrels and on the Gould’s petrel complex, in particular regarding the relationships between its Australian and New Caledonian subspecies. Using a multiloci approach for each of these investigations provided a mean to improve our understandings: in Pseudobulweria the species tree approach used allowed inferring the phylogenetic relationships between all the taxa in the genus for the first time and to solve a taxonomic issue regarding the status of Beck’s petrel; in Northeast Atlantic gadfly petrels, I showed that the divergence of the three taxa living in that region is much more recent than previously thought and that the three populations deserve at least the status of Evolutionary Significant Units; iii in Gould’s petrel I showed that the two population currently recognised as subspecies are not structured genetically and the lineages have not diverged. These investigations underline the necessity to incorporate new methods and multiple loci when investigating the phylogenetic and phylogeographic patterns in organisms to fully capture their complexity. KEYWORDS Taxonomy, phylogeography, phylogeny, petrel, Procellariiforme, Pterodroma, Pseudobulweria, conservation biology, molecular ecology iv ACKNOWLEDGEMENTS My thanks obviously go to Vincent Bretagnolle, for allowing me to do this work in Centre d’Etudes Biologiques de Chizé, as the director of this lab and as my supervisor during these three years. Many thanks for accepting to give me the chance to undertake this work under your supervision. There was a part of risk in this choice and I am glad you took it. Many thanks also to Eric Pasquet for accepting to be my co-supervisor knowing that I would spend most of the time in the “chizean” woods rather than in the MNHN. Thanks also for giving me access to the Service de Systématique Moléculaire and to the bird collections. My thoughts then go to Kerry-Jayne Wilson, who quite far back in the past, in 2000, in a lift at Lincoln University in New Zealand asked me if I would be interested in camping a few months on an island to study petrels for my MSc research. Well, Kerry-Jayne, I’m very glad I accepted and very grateful to you for giving me this opportunity to discover the fantastic world of petrels that, 10 years later, allows me to write this manuscript about these marvellous birds. I would also like to acknowledge the members of the jury who accepted to evaluate my work and to come all the way to Chizé: Michael de L. Brooke, whose book on albatrosses and petrels is a brilliant piece of work; Alice Cibois, who was always keen to help with my questions or to provide samples from Polynesia.; Pierre-André Crochet who managed to come despite a very demanding timetable; Richard Cordaux from Université de Poitiers. My thanks go to Annie Tillier and all the staff of the SSM. Thank you so much for your help, conviviality and for helping me discovering the “wonderful” world of molecular biology. I was a complete stranger to this world when I first came in the SSM; in a sense I still am, but thanks to you I guess I start to have a good glimpse at this world. And if I talk of molecular biology, then obviously I want to thank Stéphanie Dano for sharing her lab with me (it musn’t have been easy all the time I guess…) and sharing also all the disappointments or joys linked with PCRs failures and successes. Thank you also for your receptiveness and technical help. Many thanks also to Colette Trouvé for your conviviality and for the dry-ice. v I also want to thank Delphine Bonnet and Evelyse Rouby for all the administrative side of these three years. Obviously, much of these investigations would not have been possible without samples. All my gratitude therefore goes to the many people who provided those, Hadoram Shirihai, Nicholas Carlile, Francis Zino, Ruth Brown, Jacob Gonzales-Solis, Dick Watling, Alice Cibois, Jean-Claude Thibault. Various Museums also kindly accepted to provide samples: American Museum of Natural History, Museum National d’Histoire Naturelle, Te Papa Tongarewa-New Zealand Museum, Sydney Museum, Museum de Tahiti et des iles. I also acknowledge a Collection Study grant from the AMNH that allowed me to collect samples in this Museum with the kind help of Paul Sweet and Margaret Hart. I am also very grateful to Philippe Raust and the Société d’Ornithologie de Polynésie for inviting me to take part to the expedition in the Marquesas (Eiao and Hatuta’a islands) and for organising my stay on Raivavae. I hope I will have the opportunity to work with you again in the future. Many many thanks to all the people who accepted to read all, or parts of this manuscript to improve it. In particular Bertrand Gauffre, Alex Villers and Janos Hennicke. Thanks to Thomas Cornullier also for insightful comments on some parts of the introduction. Thanks to all the students and contractors with whom I shared some time in the field following the harriers and bustards. Without you guys I wouldn’t have managed…: thousands of thank you to Alex, Léo, Olivier, Steve, Victor, Audrey, Jessica and the others whose names I forget now due to lack of sleep. I am grateful to all the students I met here in Chizé and want to thank you all. I apologise for I will not write all your names down but I do thank you all for the great working environment you provided! A special thought for the “ECCB 2009” and “Oxford 2010” teams: these were good times I spent with you guys! Hope you weren’t too scared by my driving on snowy English roads and in Prague’s traffic… And “Boeni” good luck with your new office mates! And with your last year!...It was fun and pleasant to share my office with you. vi A special thought also to all the “cat keepers” when I was away in Paris or further away: Alex, Léo & Camille (a special thank you Camille for your kindness all the time; you’re incredible, don’t change a thing!; and Léo for all the time shared in Kerguelen and Chizé), Mirabelle (I hope you finish your thesis at some stage and that Rackam is fine), Alban & Nadège (good luck in Britanny! And thanks Alban for coping with me in the office… must have been tough sometimes…), Audrey (I hope your time in Spain will be rich in meetings and experience!). I am obviously grateful to my parents who gave me their support all the time, even though you didn’t understand all of what I was doing here. You were there as always and that’s already more that I could ask for. Last but obviously not least, thanks to my ET for simply being here and helping me finish this work and removing some of my impatience to grab my backpack and leave to explore the world. vii TABLE OF CONTENTS RESUME I MOTS CLES II ABSTRACT III KEYWORDS IV ACKNOWLEDGEMENTS V TABLE OF CONTENTS 1 CHAPTER 1: INTRODUCTION 4 PART 1 THEORETICAL AND METHODOLOGICAL CONTEXT I SPECIES CONCEPTS AND MODES OF SPECIATION II MOLECULAR PHYLOGENETICS THESIS PROBLEMATIC PART 2. STUDY ORGANISM I SEABIRDS II PROCELLARIIFORMES III GADFLY PETRELS THESIS OBJECTIVES REFERENCES 7 9 23 36 39 39 40 48 50 53 CHAPTER 2: GENERAL METHODS 67 2.I MITOCHONDRIAL VERSUS NUCLEAR DNA 67 2.II SAMPLES: ORIGINS AND PROCESSING 70 2.II.1 FRESH SAMPLES 70 2.II.2 MUSEUM SAMPLES 70 2.II.3 DNA EXTRACTION 71 2.II.4 MITOCHONDRIAL AND NUCLEAR DNA AMPLIFICATION AND SEQUENCING 71 2.III ESTIMATION OF GENE FLOW, POPULATION DIVERGENCE TIME AND EFFECTIVE POPULATION SIZES 73 REFERENCES 76 CHAPTER 3: THE COMPLETE PHYLOGENY OF PSEUDOBULWERIA THE MOST ENDANGERED SEABIRD GENUS: SYSTEMATICS, SPECIES STATUS AND CONSERVATION IMPLICATIONS 80 ABSTRACT KEYWORDS INTRODUCTION MATERIAL AND METHODS SAMPLES PHYLOGENETIC ANALYSES RESULTS THE GENUS PSEUDOBULWERIA: MONOPHYLY, AND RELATIONSHIPS WITH OTHER PETRELS 80 80 81 83 83 86 88 88 1 GENETIC DISTANCES AND TAXA RELATIONSHIPS WITHIN GENUS PSEUDOBULWERIA DISCUSSION POSITION OF PSEUDOBULWERIA WITHIN PROCELLARIIFORMES SUBSPECIES OF TAHITI PETREL BECK’S PETREL TAXONOMIC STATUS FIJI PETREL TIMING OF DIVERGENCE WHY ARE THESE PETRELS SO RARE? CONSERVATION IMPLICATIONS ACKNOWLEDGMENTS REFERENCES APPENDIX 1 APPENDIX 2 89 94 94 94 95 95 96 97 97 98 98 104 106 CHAPTER 4: TRACKING THE PHYLOGEOGRAPHIC HISTORY IN NORTH EAST ATLANTIC GADFLY PETRELS REVEALS MITOCHONDRIAL AND NUCLEAR DNA INCONGRUENCE AND COMPLEX SCENARIOS 110 ABSTRACT KEYWORDS : INTRODUCTION MATERIAL AND METHODS SAMPLE COLLECTION AND LABORATORY METHODS GENETIC DIVERSITY, GENETIC STRUCTURE AND HISTORICAL DEMOGRAPHY EVOLUTIONARY RELATIONSHIPS ESTIMATION ESTIMATION OF GENE FLOW, POPULATION DIVERGENCE TIME AND EFFECTIVE POPULATION SIZES RESULTS PHYLOGENETIC RELATIONSHIPS GENE FLOW, POPULATION DIVERGENCE TIME AND EFFECTIVE POPULATION SIZES DISCUSSION GENETIC DIVERSITY POPULATION STRUCTURE POPULATION DIVERGENCE AKNOWLEDGEMENTS REFERENCES 110 110 111 115 115 116 116 117 118 121 125 127 127 128 129 131 132 CHAPTER 5: PHYLOGEOGRAPHY OF GOULD’S PETREL (PTERODROMA LEUCOPTERA) AND PRELIMINARY TAXONOMIC INVESTIGATIONS IN GOULD’S (PTERODROMA LEUCOPTERA) AND COLLARED PETREL (PTERODROMA BREVIPES) 141 PART 1: PHYLOGEOGRAPHY OF A THREATENED SEABIRD TAXON, THE GOULD’S PETREL (PTERODROMA LEUCOPTERA) 142 ABSTRACT 142 KEYWORDS 142 INTRODUCTION 143 MATERIAL AND METHODS 146 SAMPLE COLLECTION AND LABORATORY METHODS 146 EVOLUTIONARY RELATIONSHIPS ESTIMATION 147 ESTIMATE OF GENETIC DIVERSITY, GENETIC STRUCTURE AND HISTORICAL DEMOGRAPHY 147 ESTIMATION OF GENE FLOW, POPULATION DIVERGENCE TIME AND EFFECTIVE POPULATION SIZES 148 RESULTS 149 PHYLOGENETIC RELATIONSHIPS 149 GENETIC DIVERSITY AND POPULATION STRUCTURE AND HISTORY 151 GENE FLOW, POPULATION DIVERGENCE TIME AND EFFECTIVE POPULATION SIZES 155 DISCUSSION 156 2 ACKNOWLEDGEMENTS 160 REFERENCES 160 PART 2: PRELIMINARY TAXONOMIC INVESTIGATIONS IN GOULD’S (PTERODROMA LEUCOPTERA) AND COLLARED (PTERODROMA BREVIPES) PETRELS 166 ABSTRACT 166 KEYWORDS 166 INTRODUCTION 167 MATERIAL AND METHODS 170 SAMPLE COLLECTION AND LABORATORY METHODS 170 EVOLUTIONARY RELATIONSHIPS ESTIMATION 171 RESULTS 172 DISCUSSION 177 ACKNOWLEDGEMENTS 179 REFERENCES 179 CHAPTER 6: DISCUSSION AND PERSPECTIVES 183 6.I METHODOLOGICAL ASPECTS 6.II TAXONOMIC ASPECTS 6.III GENETIC DIVERSITY IN PROCELLARIIFORMES 6.IV POPULATION DIFFERENTIATION IN PROCELLARIIFORMES 6.IV.1 TIMING OF DIVERGENCE 6.IV.2 DIFFERENTIATION 6.IV.3 PLEISTOCENE AND OCEANOGRAPHIC CONDITIONS 6.V IMPLICATION OF THIS RESEARCH FOR CONSERVATION 6.VI PERSPECTIVES REFERENCES 183 188 190 193 193 194 196 198 199 203 APPENDIX 1: PAPER NOT DIRECTLY RELATED TO WORK PRESENTED IN THIS MANUSCRIPT PUBLISHED IN WATERBIRDS (2009) 210 3 CHAPTER 1: INTRODUCTION The planet is currently experiencing an intense episode of extinction similar to prehistoric mass extinctions. The main differences between the latter and this 6th extinction event are (i) loss of biodiversity is much quicker now than in the past, with much higher extinction rates than background extinction rates found in fossils and (ii) this mass extinction is caused by anthropogenic actions. This current loss of biodiversity and the global environmental changes that are emerging give an ever more important role to conservation biology in the 21st century (Hedrick, 2001). The objectives of this discipline can be defined as the conservation of biological diversity and of the processes that sustain it (Moritz, 2002). As such, conservation biology is a multidisciplinary field that is made of subjects as diverse as ecology, biology, environmental monitoring, captive breeding, epidemiology, historical biogeography, land management, genetics, anthropology, economics, politics to cite but a few. Biodiversity can be approached at various levels (Féral, 2002): first at the population level, where different populations within a species can exhibit different degrees of genetic diversity and differentiation whose loss contribute to the overall decline of diversity in an area; second, at the specific level where abundance and number of species describe the diversity of a region; third at the communities and ecosystems level in a region. In the last decades, progresses in molecular biology techniques have rendered the use of genetics and molecular phylogenetics feasible at all these three levels (Moritz, 1995, Hedrick, 2001). Hence, use of genetics and molecular phylogenetics has gained a crucial importance for the long term persistence of biodiversity (Ehrlich & Wilson, 1991; Hedrick, 2001), and Daugherty et al. (1990) stated, based on their molecular reappraisal of Tuatara phylogeny and taxonomy, that “taxonomies are not irrelevant abstractions, but the essential foundations of conservation practice”. Within species, at the population level, population viability can be assessed by use of population genetics and molecular phylogenetics tools (Moritz, 1995; Haig, 1998; Feral et al., 2002; Storfer et al., 2009). Indeed, estimating the effective population size (Ne) can be extremely important to evaluate the overall health of a population and prioritise conservation actions among populations and is made possible much more rapidly than traditional demographic studies by molecular ecology techniques. Additionally, associated with Ne, estimations of genetic diversity and gene flow within and amongst populations can help evaluate which one are in most urgent need of conservation action or are the most important target of actions to maintain long-term genetic 4 diversity within a species. For instance, in a context of limited financial, technical or human resources, it might be more important to protect a population that acts as a source of genetic diversity for other populations rather than a peripheral population acting as a sink in terms of diversity thus loosely contributing to overall genetic diversity. Conversely, identifying a population concentrating much of the overall species genetic diversity by receiving individuals from many others less genetically diverse populations can be of crucial importance for the long term conservation of the species. Such considerations of inter-population genetic diversity and relationships led to the definition of different categories of populations relevant to conservation biology: Management Units (MU), i.e. “populations that are genealogically close but demographically autonomous” (Avise, 2000) and can be identified through allelic frequencies divergence whatever the depth of the gene tree considered (Moritz, 1994), and Evolutionary Significant Units (ESU), i.e. “sets of populations with a distinct, long-term evolutionary history mostly separate from other such units” (Ryder, 1986). Use of molecular tools in conservation of populations can be illustrated with two examples. Friesen et al. (2006) investigated relationships between different populations of the critically endangered Galapagos petrel (Pterodroma phaeopygia) with microsatellite and mitochondrial DNA data. They showed that populations on five islands should be considered as distinct MUs for the long-term conservation of genetic diversity within the species and, as a consequence, within the archipelago as a whole. On the other hand, investigating molecular relationships between species of the Wandering albatross complex (Diomedea sp.), Burg & Croxall (2004) concluded that two taxa (Diomedea antipodensis and Diomedea gibsoni), formerly considered subspecies of Diomedea exulans then later elevated to specific rank, were not differentiated enough and should therefore be considered together as a single unit in terms of conservation. Furthermore, to understand and maintain evolutionary processes that sustain biological diversity (Moritz, 2002), understanding processes acting at the genetic level within and among populations is essential. At the specific level, taxonomic definition or reinterpretation can affect prioritisation of conservation efforts. Indeed, although in an evolutionary perspective this can be controversial, resources, being either financial, human or technical, are allocated and prioritised based on taxonomic status of targeted organisms. Thus, various organisations (e.g. IUCN) or governments use specific or subspecific statuses to protect taxa, and therefore, recognition that a species, subspecies or population is different or not from a non-threatened related taxon, will deeply affect resources allocations. It should be noted however that various pieces of legislation, such as the Endangered Species Act in the USA, or the UK Wildlife and Countryside Act now incorporate 5 levels of diversity under the species level to take into account the importance of intra-specific diversity. Given the relative facility with which it is now possible to collect genetic material from organisms and use molecular phylogenetic tools to infer phylogenetic relationships, taxonomic recommendations more and more rely on such genetic information. This explains the importance taken by the field of molecular phylogenetics in terms of conservation resources allocations through the definition of new phylogenetic relationships between taxa or reinterpretation of traditional taxonomies. For instance, the genus Diomedea, that comprises the great albatrosses, was for a long time considered to be made of 12 species by traditional taxonomies based on morphological characters. However, when Nunn et al. (1996) investigated the phylogenetic relationship of these taxa with mitochondrial DNA (cytochrome b) they found the genus to be paraphyletic and split it into four different genera, genus Diomedea remaining with only three species D. epomophora, D. exulans and D. amsterdamensis. Further investigations incorporating genetic, morphological and ecological data by Robertson & Nunn (1998), Nunn & Stanley (1998) and Cuthbert et al. (2003) led to the division of these three species into seven, of which six are recognised (Brooke, 2004). Of these six species, several have small populations and restricted ranges hence a great increase in the number of albatross species considered threatened (Gaston, 2004). At the community level, molecular tools developed notably for phylogeographic studies can be used to identify regions with highest number of ESUs, regions that can then take advantage of greater conservation priority (Avise, 2000). Furthermore, it has been shown that Phylogenetic Diversity, which measures biodiversity taking into account the phylogenetic relationships of taxa (Faith, 1992, 1994), is being lost faster than expected from species extinctions (Purvis et al., 2000; vonEuler, 2001), and that its distribution is not constant over space (Sechrest et al., 2002). Hence the idea that evolutionary history of taxa within different regions need be taken into account by conservation planning (Rodrigues et al, 2005). These latter authors showed that although in most cases species richness measures are adequate to pursue this goal (albeit with a slightly lower efficiency; Purvis et al., 2005), there can be some situations in which this is not true, for instance if old species are endemic to relatively species-poor areas (such as the Tuatara, Sphenodon sp. in New Zealand). In this context where evolutionary issues need be taken into consideration, taxonomic issues solved and biodiversity described before being lost, it is important to start with a brief reminder about the theoretical background surrounding evolution and the concept of species, as well as the methodological background that have arisen from the development of molecular phylogenetics. 6 Part 1 Theoretical and methodological context "Nothing in Biology Makes Sense Except in the Light of Evolution” Theodosius Dobzhansky 1973 Evolution is considered the underlying principle that organises life on earth and the “engine” of biology (Hayden, 2009). Although the importance of evolution had been recognised and formulated in earlier days, notably by French naturalist Lamarck (1809) - the Lamarckian modality of evolution being centered at directed variation that is specifically caused by environmental factors (Koonin & Wolf, 2009a)- this great figure was ahead of his time and its audience was not ready for his views (Conn, 1907), and it was not until the second half of the 19th century that the importance and fundamental aspect of evolution was described when two papers from Wallace (1858) and Darwin (1858) presenting the “hypothesis of descent by means of natural selection” were published in London. These papers contained five major statements (Kutschera & Niklas, 2004): (1) all organisms produce more offspring than their environments can support; (2) intraspecific variability of most characters exists in abundance; (3) competition for limited resources leads to a struggle for life (Darwin) or existence (Wallace); (4) descent with heritable modification occurs; and (5) as a result, new species evolve into being. However, it was in the publication of Origin of Species that Darwin (1859) summarised a solid body of observations, amassed for more than twenty years, to transform a concept into a real scientific theory (Kutschera & Niklas, 2004) that influenced the scientific world of this time and for the years to come until now. In this book, Darwin presented his evolutionary theory by which “one species does change into another through natural selection” and the accumulation of minute changes. This theory implied that all modern living organisms are related and originate in one or a very small number of common ancestors, hence the description by Darwin of a Tree of Life linking all species and representing the evolutionary relationships between species. This Tree of Life (TOL), deriving from older tradition of natural history (Ragan et al., 2009) although not stated explicitly as such by Darwin in his first edition of the Origin of Species was nevertheless the only drawing in this original version, indicating its importance as a metaphor of evolution (Koonin & Wolf, 2009b). This metaphor soon took a central place in biology and evolutionary thinking and is still of great importance despite the actual debate regarding its validity to describe the reality of the 7 biological world seen through genomic analysis (but see Koonin & Wolf, 2009b for a detailed presentation of this topic and of the Forest of Life concept proposed by these and other authors to replace the TOL). With the theory of evolution and the concepts of natural selection and TOL, it became possible for Darwin and his successors to explain the diversity of life on earth and how one species can become many through speciation. Interestingly, although the famous Darwin’s finches would become an iconic example of speciation, there was no understanding of the processes behind speciation until further works by Darwin (1871), Wallace (1865, 1875, 1889), Weisman (1892, in Kutschera & Niklas, 2004). Later, the rediscovery of Mendel’s laws and the beginnings of population genetics, followed by the development of the synthetic theory or Modern Synthesis by Dobzhansky (1937), Mayr (1942), Huxley (1942), Simpson (1944), Rensch (1947) and Stebbins (1950) (Kutschera & Niklas, 2004; Koonin, 2009) allowed to understand those processes a step further. This synthesis was later summarised in two points (Mayr & Provine, 1980 p.1): “1.Gradual evolution can be explained in terms of small genetic changes (“mutations”) and recombination, and the ordering of this genetic variation by natural selection; 2. the observed evolutionary phenomena, particularly macro-evolutionary processes and speciation, can be explained in a manner that is consistent with the known genetic mechanisms”. Embedded in this approach, is the idea for Mayr and the tenants of the Modern Synthesis, as for Poulton (1904) - an entomologist whose work and definition of species seem often overlooked (Mallet, 2004)-, that species are the only truly natural grouping of individual in nature, and as such, one of the fundamental units in biology along with genes, cells or organisms (Mayr, 1982). Such ideas strongly contrast with the views of Darwin (1859) for whom accepting evolution forces “to treat species in the same manner as those naturalists treat genera, who admit that genera are merely artificial combinations made for convenience”, and of Dawkins (1976) for whom natural selection can act at the level of an individual gene as much as at the level of an organism as a whole, and that organisms thus are “just” a means for genes to propagate. Independently of this debate on the relative importance of species and genes, species remain an essential operational unit used in most, if not all, fields of biology (deQueiroz, 2005) such as population biology, evolutionary biology, ecology, conservation biology and hence cannot be ignored. The next section will introduce in more details the various concepts of species and models of speciation. 8 I Species Concepts and Modes of Speciation Species are an essential unit in all subfields of biology (deQueiroz, 2005). Furthermore, in a time where most biodiversity remains unknown (Blaxter 2004 ; Savolainen et al. 2005) but is facing its 6th mass extinction, the description of this biodiversity takes a very significant environmental and economical importance (Cincotta et al. 2000; Myers et al. 2000). Species is the most used unit by international or governmental organisations when prioritising conservation actions (Sites & Crandall 1997; Peterson & Navarro-Siguenza 1999; Sites & Marshall 2004; Stockman & Bond, 2007) and cannot be overlooked. However, despite the central importance of species in biology and biodiversity conservation, for more than 150 years since the publication of the Origin of Species (Darwin, 1859) there has been no consensus on what is a species, or on how should we identify species (Barton, 2001; Hey, 2001). In the last few decades many often contradicting definitions and concepts of species have been put forward by biologists from different sub-fields of biology adding complexity to the problem (deQueiroz, 2007). Before the advent of the theory of evolution, the essentialist view of species, preceding Lamarck, Wallace and Darwin, considered species as separately created and immutable units. Associated with these views were empirical perceptions of similarity and inter-fecondity of individuals from the same species. Such conceptions formed the basis of the first attempts to formalise the concept of species. Thus, for John Ray (1628-1705) a species is “a group of individuals, which, by means of reproduction, give birth to individuals similar to themselves” (in Lorion, 2008). Following the rise of the theory of evolution this concept of species was to be replaced by many others. I. 1 Diversity of Species Concepts Unlike other fundamental biological entities such as genes, cells or organisms, species are difficult to perceive due to their temporal and spatial scales. With the ever increasing specialisation of various sub-fields of biology, biologists have developed different concepts that fitted to their field, thus giving to the species concept the specific properties that they were studying. Two dozens of different species concepts originating in various sub-fields of biology have been listed by Mayden (1997; Table 1.1) and more have been described since then (e.g. the genetic 9 species concept of Baker & Bradley, 2006). It should be noted that some of these concepts or their formulations simply correspond to semantic variations on a common theme. Table 1.1 Species concepts (from Mayden, 1997; see therein for more details) •Reproductive Competition •Agamospecies Concept •Internodal Species Concept •Morphological Species Concept Concept* •Biological Species Concept* •Non-dimensional Species Concept •Successional Species Concept •Cladistic Species Concept •Cohesion Species Concept* •Phenetic Species Concept •Taxonomic Species Concept •Composite Species Concept •Phylogenetic Species Concept •Ecological Species Concept* (Diagnosable Version)* •Phylogenetic Species Concept •Evolutionary Significant Unit* •Evolutionary Species Concept* (Monophyly Version) •Genealogical Concordance •Phylogenetic Species Concept Concept (Diagnosable and Monophyly Version) •Genetic Species Concept* •Genotypic Cluster Concept •Polythetic Species Concept •Hennigian Species Concept* •Recognition Species Concept* * Concepts that make reference to biological processes (e.g. reproduction and competition) that occur among organisms within species (and less so between species) and that contribute to a shared process of evolution within species It is usually accepted that these species concepts can be categorised in four classes: (i) biological concepts based on reproductive isolation, used by population biologists for instance; (ii) phenetic concepts based on overall similarity, usually in morphology or other observable traits and mainly used by palaeontologists (iii) phylogenetic concepts based on common derived characters; (iv) ecological concepts based on common ecological niche or adaptive zone. However, additional classes and sub-classes may be recognised, as summarised in Table 1.2 (from deQueiroz, 2007). Classes in Table 1.2 are somewhat “artificial” and several concepts are almost identical. For instance, the Genotypic cluster concept of Mallet (1995) can be seen as an adaptation of the Phenetic concept to genetic data based on independent markers. In addition, the Evolutionary concept of Simpson (1951), Wiley (1978) and Mayden (1997) is very similar to the Lineage Species Concept presented below in I-2. All these concepts are based on some biological reality and have been formulated in a particular context with a special “species issue” to solve. Due to a common confusion between the definition of a taxonomic category (i.e. the definition of a concept) and the criteria that can be used to identify taxa belonging to that category (i.e. species delimitation) (deQueiroz, 1998; Samadi & Barberousse 2006), they often end up leading to different conclusions in terms of species 10 boundaries and numbers. The case of brown and polar bears is a good example: polar bear (Ursus maritimus) and brown bear (Ursus arctos) are very different, both phenetically (size, fur colour…) and ecologically (they live in different ecosystems and use them in different ways). Nevertheless, they are able to reproduce and give birth to fertile hybrids. Thus, Ursus maritimus and Ursus arctos can be seen as belonging to the same species according to a Biological Species Concept! Furthermore, some populations of Ursus arctos from Alaska do share a direct common ancestor with Ursus maritimus (Talbot & Shields, 1996), while more meridional populations of Ursus arctos belong to another branch of the tree. Thus according to that study, the polar bear is monophyletic, while the brown bear is paraphyletic. Ursus arctos is therefore not a species in the phylogenetic sense since all descendants of the common ancestor do not share the same derived characters. Table 1.2 Diversity of contemporary species concepts and properties upon which they are based; properties usually accepted as necessary are marked with an asterisk (from deQueiroz, 2007) Concept Property(ies) References Interbreeding (natural reproduction resulting in viable and fertile offspring) Wright 1940; Mayr Dobzhansky 1950 Isolation *Intrinsic reproductive isolation (absence of interbreeding between heterospecific organisms based on intrinsic properties, as opposed to extrinsic [geographic] barriers) Mayr (1942); Dobzhansky (1970) Recognition *Shared specific mate recognition or fertilization system (mechanisms by which conspecifics organisms, or their gametes, recognize one another for mating and fertilization) Paterson (1985); Masters et al. (1987); Lambert and Spencer (1995) Ecological *Same niche or adaptive zone (all components of the environment with which conspecific organisms interact) Van Valen (1976); Andersson (1990) Evolutionary Unique evolutionary role, tendencies, and historical fate (some interpretations) Cohesion *Diagnosability (qualitative, fixed difference) Simpson (1951); Wiley (1978); Mayden (1997) Grismer (1999, 2001) Phenotypic cohesion (genetic or demographic exchangeability) Templeton (1989, 1998a) Phylogenetic Hennigian Heterogeneous (see next four entries) Ancestor becomes extinct when lineage splits (see next four entries) Hennig (1966); Ridley (1989); Meier and Willmann (2000) Rosen (1979); Donoghue (1985); Mishler (1985) Biological 1942; Monophyletic *Monophyly (consisting of an ancestor and all of its descendants; commonly inferred from possession of shared derived character states) Genealogical *Exclusive coalescence of alleles (all alleles of a given gene are descended from a common ancestral allele not shared with those of other species) Baum and Shaw (1995); see also Avise and Ball (1990) Diagnosable *Diagnosability (qualitative, fixed difference) Nelson and Platnick (1981); Cracraft (1983); Nixon and Wheeler (1990) Michener (1970); Sokal and Crovello (1970); Sneath and Sokal (1973) Phenetic *Form a phenetic cluster (quantitative difference) Genotypic cluster *Form a genotypic cluster (deficits of genetic intermediates; e.g., heterozygotes) Mallet (1995) 11 I. 2 Towards a unification of species concepts: the Lineage Species Concept According to deQueiroz (1998), all phenomena emphasized by the different species definitions are different aspects of the same kind of entity and recognise that species are parts of separately evolving lineages. Therefore, the so-called “species problem” (Mayden, 1997; Hey, 2001) is not as complex and as insurmountable as it could appear at first. deQueiroz (1998, 2005, 2007) advocates that all the concepts described previously can be considered as delimitation criteria within a larger concept called Lineage Species Concept, derived from Simpson (1951) and Wiley (1978). Following deQueiroz (1998), this idea has also been advocated by Samadi & Barberousse (2006, 2009), who consider an internodal species concept that formalises the definitive divergence of evolutionary lineages. This internodal species concept represents a broader version of the Hennigian phylogenetic concept (Table 1.2) where a species disappears when splitting into new lineages (Fig. 1.1) and differs from the Lineage Species Concept of deQueiroz in that it defines more precise boundaries to species (Samadi & Barberousse, 2006). Figure 1.1 From Samadi & Barberousse, 2006. Species are groups of organisms connected to one another through the global genealogical network, and exist either between two successive speciation events or between a speciation and an extinction. Such concepts are rooted within the theory of evolution: heritable characters of genealogically linked organisms change through random mutations. The sorting of these variable characters of organisms is affected by both a random process (i.e. drift), and a selective process (i.e. natural selection). The influence of these two processes is affected by intrinsic constraints of the organisms (described through characters) and dependent on the spatio-temporal context in which organisms are living. These sorting processes act on groups of organisms or populations and, if they last long enough divergences can appear between populations, accumulate and ultimately give rise to completely divergent lineages. All along the divergence process, individuals from different lineages become more similar to each other than to individual of other lineages. And as the lineages 12 diverge, they become diagnosable through their fixed characters, their reproductive systems change to a point where they become incompatible, they gradually stop recognising each other as potential mates and they start using different ecological niches. Hence there is a transition from polyphyly to monophyly (deQueiroz, 2007). The acquisition of these properties does not occur immediately, and not necessarily simultaneously. And the order in which all these changes appear can vary. It is because the various species concepts use different sets of properties as essential and necessary conditions for species categories that they often become incompatible, hence the existence of the “species problem”. This phenomenon is illustrated by Fig. 1.2 (from deQueiroz, 2007) where one can see a region in the divergence process in which different species concepts can contradict each other. This is due to the different abilities of various species concepts to detect divergence (especially when it is recent) and to identify whether this divergence is definitive or not. However, previously described species concepts are mostly based on properties that can be used as evidence to evaluate lineages divergences. Because in the lineage species concept, species are defined as separately evolving lineages, using these properties can help evaluating the presence of lineages evolving towards species when evidence suggest that one is looking at the grey zone of Fig. 1.2. Such approach has been labelled Integrative Taxonomy (Dayrat, 2005; Padial et al., 2010; Schlick-Steiner et al., 2010). Integrative Taxonomy involves the use of different sets of criteria (molecular, morphological, ecological…) and integrating them while assessing the congruence or lack of congruence between the different sets of characters. Although there probably is no perfect methodology to delineate species boundaries, this integrative and evolutionary approach clearly extend the possibilities to identify and describe biodiversity. 13 Figure 1.2 From deQueiroz (2007). Lineage separation and divergence and species concepts. This represents a single lineage splitting to form two lineages. Lineage divergence through time is illustrated by gradations in shades of gray. Horizontal lines represent times at which lineages acquire different properties (i.e., phenetic differences, reciprocally monophyletic, reproductively incompatible, ecologically distinct, etc.) corresponding to different species criteria in different species concepts. Within the gray zone alternative species concepts come into conflict due to adoption of different species criteria. I. 3 Taxonomic characters and Integrative Taxonomy Integrative taxonomy, as other traditional taxonomic methods, is based on taxonomic characters. Those characters, which can be qualitative or quantitative, can be classified in five general classes: biochemical, molecular, morphological, behavioural and ecological. Each character can possess a unique state within a population that differs from another state within another population, or several states whose frequency distribution vary between the compared populations. Choice of characters used to identify species limit is of paramount importance and must be done carefully. An important point is the role played by the characters used in the speciation process. Thus every species delimitation case is likely to have a character more appropriate than others to assess the lineages divergence pattern. For instance, in amphibians or birds call patterns can be important characters to differentiate species, or in groups in which ecological specialisation plays a role in differentiation, such ecological character can be more important to consider than, say, colourations. 14 Generally, traits that contribute to the reproductive isolation of lineages, such as sexually selected characters or traits that promote pre- or post-mating isolation (e.g. calls in some birds or in insects) are more likely to be indicative of species-specific differences. In all five classes of characters cited above, traditionally the most used ones are morphological and molecular traits. Morphological traits can usually be used with living, dead or fossilised organisms, and can often, through visual inspection, allow differentiating individuals from different species (although there are obviously many exceptions), making them particularly attractive. However, despite their advantages, morphological characters suffer from two major drawbacks, i.e. there is often a strong subjective component in their use and they usually are continuous characters. For instance, the shape of marine organisms’ shells is a continuous character and describing limits within this continuum to differentiate lineages often requires some part of subjectivity. Alternatively, molecular characters such as DNA sequences have been increasingly used in the last twenty years. DNA sequences provide many more characters (nucleotides) than morphology and conclusions drawn from these are usually considered less subjective than morphologically-based conclusions, at least among the proponents of the various phylogenetic species concepts. DNA sequences are most often studied with tree-based methods that allow searching for monophyletic lineages that could represent species. However, caution is needed, as monophyly can encompass groups of organisms above species level. Furthermore, as a tree based on a single locus is not necessarily in agreement with the real species tree (Nichols, 2001; Degnan & Rosenberg, 2006, 2009), conclusions based on only one locus can lead to erroneous species delimitations. Actual methodological developments now allow building trees based on coalescent theory using many loci. Such approaches can detect lineage divergence despite individual gene trees incongruence or lack of monophyly (Carstens & Knowles, 2007). When using the Lineage Species Concept and an Integrative taxonomy approach, usual disagreement remains as to the degree of congruence necessary between different sets of characters to delineate species boundaries within diverging lineages. Some consider congruence between morphological and molecular traits absolutely necessary (e.g. Dayrat, 2005; Cardoso et al., 2009), while other authors advocate avoiding a priori combinations of characters (e.g. deQueiroz, 2007; Padial et al., 2010; Schlick-Steiner et al., 2010). These two different approaches have been called respectively "integration by congruence" and "integration by cumulation" (Fig. 1.3) by Padial et al. (2010). Integration by congruence, by considering that congruence between several taxonomic characters is likely to indicate full lineage divergence, presents the advantage of taxonomic stability 15 (Padial et al., 2010) as only small lack of congruence between characters will lead to conclude to incomplete lineage differentiation hence to the lack of complete separation of lineages into species. However, by doing so, it is possible to underestimate the true number of species. Indeed, character differentiation will not occur for all characters at different stages of speciation (Fig. 1.2). Furthermore, such an approach could miss speciation events such as Darwin’s finches or cichlids fishes in which morphological differences and habitat use characterise diverging lineages that have not yet fully diverged genetically and can exhibit incomplete gene lineage sorting. On the other hand, integration by cumulation considers that divergence in any taxonomic character can allow delineating species. Although acknowledging the importance of congruence of characters in species delimitations, this approach does not consider it absolutely necessary. This gives the advantage of being able to use any taxonomic characters deemed most appropriate to evaluate lineage differentiation. Unlike the congruence approach, this approach is likely to be more effective to detect recently diverged species or lineages that are in the divergence process. However, the drawback is that, by allowing the use of a lesser number of taxonomic traits, this approach can lead to an overestimation of species numbers. For instance, one could consider that using a single mtDNA marker is enough to draw conclusions, whereas as mentioned above gene trees do not necessarily reflect the true species tree and evolutionary history of a lineage (Pamilo & Nei, 1988; Nichols, 2001). Although the formal description of Integrative Taxonomy is relatively recent, the practice of combining different kinds of information, morphological, behavioural, molecular etc. is not new and some authors have been advocating it without putting a name on the methodology (e.g. Forister et al., 2008). And it has been recognised by various authors that the evolution of molecular, morphological, ecological and behavioural characters is driven by multiple forces (Rubinoff, 2006) and that all population level processes need be considered when trying to identify conservation units or species limits (Forister et al., 2008). 16 Figure 1.3. From Padial et al., 2010. Schematic representation of two approaches of integrative taxonomy. Background yellow, red, and blue colors represent the spectrum of character variation, each dot being an independent evolutionary linage that requires identification and delimitation as separate species. Integration by cumulation (left) identifies species limits with divergence in one or more not necessarily overlapping taxonomic characters (e. g. mtDNA or morphology), whereas the integration by congruence (right) identifies species limits with the intersection of evidence from two or more independent taxonomic characters (e. g. mtDNA plus morphology). Both methods of integration have relevant limitations and there is a trade-off between the lack of reliability of the species detected by integration by cumulation, and the lower taxonomic resolving power of the integration by congruence. I. 4 Speciation The origin of species is “(…) the single most important event in evolution” (Mayr 1963, p. 11 in Sobel et al., 2010). Yet, speciation is one of the least understood major features of evolution and fundamental issues such as what causes reproductive barriers to emerge and what barriers play a role in speciation are not yet solved (Schluter, 2001; Sobel et al., 2010). I.4.1 Ecological vs. non-ecological speciation Traditionally, different modes of speciation have been classified with a geographical approach depending on the distribution of populations affected by the process, i.e. allopatric, sympatric and parapatric speciation modes (Schluter, 2001; Coyne & Orr, 2004). Such a 17 classification focuses on the effect of gene flow or absence of gene flow on the development of reproductive isolation. In the last two decades however, some authors suggested a shift towards an approach centred on the underlying mechanisms driving the evolution of reproductive isolation (e.g. Schluter, 1998; Orr & Smith, 1998; Via, 2001) and promoted the distinction between ecological and non-ecological mechanisms to elucidate the role of natural selection in speciation. In such an approach, is considered as non-ecological speciation three different processes: (i) speciation by divergence under uniform selection; (ii) polyploid speciation and (iii) speciation by genetic drift (Schluter, 2001; Table 1.3). Although such a shift in thinking might appear new, it actually corresponds to a long-held view, dating back to the early days of the modern synthesis, that ecology and divergent selection are major factors of speciation (Dobzhansky 1937; Stebbins 1950; Grant 1981; Rundle & Nosil, 2005). It can be argued that Darwin himself considered ecological adaptation as a key process in speciation (Sobel et al., 2010). Even among biologists believing that speciation can be non-ecological, the general agreement is that most of the time adaptation does play a major role (Coyne & Orr, 2004; Templeton, 2008). Probably one of the most convincing examples that divergent natural selection can be a major cause of reproductive isolation is the radiation of Hawaiian honeycreepers that, from a single common ancestor, diversified in different species adapted for feeding on various substrates, such as nectar, fruits, seeds or insects (Price, 2008). Different definitions of ecological speciation have been stated (Schluter, 2001; Rundle & Nosil, 2005; Nosil et al, 2009; Schluter, 2009). For instance, Schluter (2009) defines ecological speciation as “the evolution of reproductive isolation between populations or subsets of a single population by adaptation to different environments or ecological niches”. It has been argued recently that, of the three mechanisms proposed as non-ecological (Table 1.3) and presented below, only genetic drift can be considered purely non-ecological (Sobel et al., 2010). 18 Table 1.3 Modes of speciation (Schluter, 2001) Mode of Speciation Mechanism of initial divergence Initial form of reproductive isolation Ecological speciation Divergent natural selection Prezygotic postzygotic or Speciation by divergence under uniform selection Different advantageous mutations occur in separate populations experiencing similar selection pressures Genetic drift Prezygotic postzygotic or Prezygotic postzygotic or Hybridization and polyploidy Postzygotic Speciation by genetic drift Polyploid speciation Proximate basis of reduced hybrid fitness Examples of the roles of natural selection Example roles of sexual selection Ecological selection, genetic incompatibility and sexual incompatibility Genetic incompatibility and sexual incompatibility Initial: Drive divergence in phenotypic traits Final: Reinforcement Amplify divergence of mate preferences initiated by natural selection Reinforcement Initial: Drive fixation of incompatible mutations in different populations Final: Reinforcement Drive fixation of alternative incompatible mutations in different populations Reinforcement Genetic incompatibility and sexual incompatibility Initial: None; or opposes divergence Final: Reinforcement caused by drift Initial: None; or promotes further genetic divergence Final: Reinforcement Amplify differences in mate preferences Reinforcement Genetic incompatibility Reinforcement Divergence under uniform selection Although it can be debated whether selective environments can be truly uniform between populations, different beneficial mutations could theoretically arise under similar selective regime, thus leading to speciation. In this case, allopatric populations fix different mutations that can lead to negative impact when combined in hybrids thus leading to intrinsic postzygotic isolation and to speciation. However, the question remains here as to how often such postzygotic isolation can arise with this mechanism. Additionally, selection usually generates reproductive isolation only as a byproduct and does not promote barriers as such, whether it is divergent or uniform (Schluter, 2009; Sobel et al., 2010). Under uniform selection, phenotypic convergence is promoted and even distantly related taxa often converge phenotypically under similar selective environments (Simpson, 1953). Examples of divergence under uniform selection are rare. One case of colouration convergence between different populations of beach mice (Peromyscus polionotus) along the Atlantic coast was described by Steiner et al. (2009): these populations evolved similar colouration by mutations at 19 different loci despite the recognised general importance of the gene Mc1r in colourations in a broad range of taxa (mammals, birds, lizards) and in beach mice of Florida coasts. Thus, diverse populations of beach mice evolved convergent light colour through natural selection for camouflage in pale sand dunes through mutations involving different loci. But in this example the question remains whether such genetic divergence under uniform selection can be the cause of postzygotic isolation. It could theoretically be the case, if the mutation for the colour also affect in some ways the reproductive success of the individuals possessing it or the fitness or their progeny, but evidence of this needs to be described. Furthermore, if such is the case, then the mechanism of differentiation could not be truly stated as non-ecological because different fixed mutations concerned characters directly linked to the environment, i.e. the colouration of sand dunes. Polyploidy Polyploïdy, i.e. the duplication of whole genome by the production of unreduced gametes, affects many organisms, from fishes to yeast through amphibians and plants. Two kinds of polyploidy exist: autopolyploïds originating from individuals of the same biological species, and allopolyploids originating by hybridisation of individuals belonging to different species. Sobel et al. (2010) argue that polyploidy can be seen both as ecological or non-ecological speciation, depending on whether we consider polyploïds as a new species when they arise, independently of their ability to persist or not. If we do, then it is a case of non-ecological speciation. However, if we recognise polyploïds as a new species only when they establish a population isolated from their progenitors, then speciation is often ecological because polyploidy is likely to produce new genetic and phenotypic variability that can be subject to divergent selection between polyploïds and their progenitors thus allowing the former to establish in a new ecological niche. Drift Theoretically, speciation by genetic drift is perfectly plausible: complete reproductive isolation occurs from fixation of different mutations in allopatric populations through a random process. However, in reality, drift alone can hardly produce significant reproductive isolation (Turelli et al., 2001; Coyne & Orr, 2004; Sobel et al., 2010). Two main reasons can explain this: (i) speciation by drift is much slower than speciation by natural selection, and isolation by drift can be easily overcome by even weak selection (Nei et al., 1983). Furthermore, if lineage isolation is incomplete, any secondary contact between such lineages is likely to erase isolation through gene flow. (ii) most of the characters leading to lineage isolation are very likely to be subject to natural selection, and except in severe bottleneck situations, such adaptive traits can hardly be altered by 20 drift alone. In theory, severe bottlenecks, such as in “founder effect”, where a small number of individuals colonise a new environment such as an island, can allow speciation by genetic drift (Gavrilets & Hastings 1996). However tested examples in nature are very few (Sobel et al., 2010) and founder effect speciation remains controversial and do often include interactions between selection and drift (Coyne & Orr, 2004). The often cited example where drift could have played a role is the snail genus Euhadra in which a single, maternally inherited mutation can cause a change in shell chirality (Ueshima & Asami, 2003). As individuals of opposite chirality cannot mate together, this mutation can by itself give rise to reproductive isolation and thus to speciation. The case of sexual selection Role of sexual selection in lineage differentiation and speciation is widely recognised and was investigated by West-Eberhard (1983) and Ritchie (2007) among others. Although speciation by sexual selection alone may happen (Sobel et al., 2010), because drift or natural selection are involved in all known examples where sexual selection had a role in lineage divergence, proving that sexual selection is involved in a speciation event does not remove the debate about speciation process from the ecological/non-ecological speciation context (Schluter, 2001). Indeed, as put forward by Ritchie (2007) sexual selection mostly facilitates divergence through its interaction with natural selection in an ecological divergence process. Thus, in the well known examples of sexual selection in Drosophila (Templeton, 1979) and African cichlids (Seehausen et al., 1999), without the marked niche divergence experienced by these lineages we can wonder whether sexual selection alone could have promoted speciation. Similarly, in the sticklebacks case (Boughman, 2001), nuptial colourations of males and female preferences are light dependent. Thus in that case sexual selection is promoting reproductive isolation but depends on the ecological context. Similarly, there exist some situations where sexual selection plays a role in speciation alongside genetic drift (Lande, 1981) or promotes unique mutations through sexual conflict (e.g. Gavrilets, 2000). I.4.2 Allopatry, sympatry, parapatry Although traditional models of speciation focused on the geographical aspect of it through the models of allopatry and sympatry, it remains unclear how genetically based differences in geographic distribution can be considered as a form of reproductive isolation (Sobel et al., 2010). Furthermore, such models have recently been labelled as overly simplistic (Grant & Grant 1997; Schluter 2001; Rundle & Schluter 2004). However, since most speciation events start with an 21 allopatric phase (Coyne & Orr, 2004), such models can obviously not be considered obsolete, and rather can be incorporated in the context of ecological speciation described above. Indeed, a classical ecological speciation scenario (Schluter, 2001; Rundle & Nosil, 2005) can be described as follow: through an allopatric phase, populations start to accumulate differences through adaptations to unique characteristics of their environment. Upon secondary contact of diverging lineages, ecological interactions between the populations constitute novel potential sources of divergent selection and, if reproductive isolation was not complete, premating isolation can evolve through reinforcement, i.e. the enhancement of prezygotic isolation by natural selection (Coyne & Orr, 2004) caused by the fact that hybrids are maladaptive. If hybridisation does not involve lesser fitness, then the two lineages merge again and no speciation occurs. This scenario is summarised in Fig. 1.4 (from Rundle & Nosil, 2005). If no secondary contact ever happens, or if complete lineage divergence and reproductive isolation occurred before the secondary contact, then speciation was allopatric. On the other hand, if lineage divergence and reproductive isolation finished evolving after the contact, then the speciation process is labelled as sympatric (Schluter, 2001; Rundle & Nosil, 2005). Parapatric speciation, where gene flow is reduced through isolation by distance but not eliminated can be seen as a form of sympatric speciation (Rundle and Nosil, 2005), or as an intermediate form between allopatric and sympatric processes (Coyne & Orr, 2004). Different isolating barriers play a role in speciation. The difficulty remains to decide what importance to give to these various barriers (Coyne & Orr, 2004). In this context, it has often been argued that speciation mechanisms can only be studied in sympatric populations (Coyne & Orr, 2004; Sobel et al., 2010). Thus, it is often said that allopatric taxa cannot be considered true species until secondary contact brings them in sympatry and “tests” their full divergence (Coyne & Orr, 2004). In this line, Rundle & Nosil (2005) considered that relevant ecological differences between species are those that allow coexistence of lineages in sympatry. However, sympatry is not inevitable, and traits involved in habitat isolation could possibly avoid any secondary contact to occur (Coyne & Orr, 2004; Sobel et al., 2010). This leads some authors to consider that geographic separation could be considered as some form of assortative mating (Kirkpatrick & Ravigné, 2002) or that, because allopatric populations often do not exhibit any gene flow, they can be considered species, whether they have evolved real reproductive barriers or not (Wiens, 2004). 22 Figure 1.4 From Rundle & Nosil (2005). Speciation scenario in different geographic contexts. Reproductive isolation between two populations is absent at the beginning of the speciation process (left) and evolves to completion (right). Populations are initially allopatric, but secondary contact can occur at any time (dashed vertical line), commencing the second stage of the speciation process. Ecological causes of divergent selection by which reproductive isolation may evolve are listed within the panel for each stage In this first part, I have introduced the concepts of species, its fundamental importance in biology and the current state of thinking (although by the time this work is submitted, it is likely to be partly out of date) in the study of speciation processes and concepts. Since the theory of evolution states that species are related to each other by descent, (Darwin, 1859), understanding the relationships between taxa and the evolutionary pathways these taxa have taken most often necessitates the use of phylogenetics, i.e. the reconstruction of evolutionary history (Delsuc et al., 2005). In the last decades, the use of molecular data became probably the most important tool to achieve phylogenetics’ objective, hence the apparition of the field of Molecular Phylogenetics. II Molecular Phylogenetics The reconstruction of evolutionary relationships between taxa, phylogenetics, directly follows the work of Darwin and his theory of evolution. In his “Origin of species” the only illustration shows the relationships between species in the form of a phylogenetic tree indicating that all species share a common history through their common ancestry. Thus, any evolutionary study requires understanding phylogenetic relationships between taxa (Delsuc et al., 2005). Different kinds of data can be used to investigate these relationships. Morphological characters 23 constituted the main source to phylogenetic reconstruction until the 1970’s. However, although such data has proved useful and powerful, e.g. through the comparison of fossils and extant taxa, it also have some limitations. For example, understanding the evolutionary relationships of viruses, which do not leave fossils, is not possible by use of morphological characters. Indeed, in microorganisms, morphologically homologous characters needed to understand phylogenetic relationships are almost non-existent, and even in complex organisms such characters are limited (Delsuc et al., 2005). By dealing directly with the evolving substrate, i.e. the sequence of gene, molecular phylogenetics allowed, from the 1970’s, to have access to a much greater number of homologous characters that could be compared. By the 1980’s, molecular phylogenetics reached maturity (Koonin & Wolf, 2009b) with the introduction of rRNA as phylogenetic marker, opening the gates of a whole new world of investigations and the discovery of the third domain of life, the Archaea, whose existence had never been suspected (Delsuc et al., 2005; Koonin & Wolf, 2009b). II. 1 Phylogenetic reconstruction The usual flow chart of phylogenetic inference process is described is Fig. 1.5 (from Holder & Lewis, 2003). Once data is collected, DNA sequences aligned, -i.e. the process of adding gaps to a matrix of data so that nucleotides in one column are related to each other by descent from a common ancestral residue (Holder & Lewis, 2003)-, and a model of sequence evolution chosen (see below), a phylogenetic tree can be constructed. Although DNA sequence data might be seen as an incontrovertible evidence of evolutionary relationships, uncertainties are actually frequent in molecular phylogenetic inferences. Such uncertainties are due on the one hand to the (very) large number of trees that can describe relationships between a set of taxa, and on the other hand, to differences in the pattern of sequence evolution between genes, organisms and over evolutionary time. Four trees can describe the relationships between three DNA sequences, one rooted, and three unrooted (Fig. 1.6). And when the number of sequences increases to 100, a number easily accessible nowadays, the possible number of trees exceeds the number of particles in the known universe (Page & Holmes, 1998). Hence, it is most of the time computationally impossible to examine every tree in order to find the best one, and various methods have therefore been developed that attempt to find the best tree without having to examine every single one. 24 Figure 1.5 From Holder & Lewis (2003). Usual phylogenetic inference flow chart Figure 1.6. Possible phylogenetic relationships between three sequences A B A B C A C B C B A C 25 Two categories of methods exist: distance methods in which aligned sequences are transformed into a matrix of pairwise distances (the total number of nucleotide substitutions between each pair of sequences) that is used to construct a tree; and discrete methods that consider each aligned nucleotide site independently. Additionally, tree building can follow either a clustering method or search for an optimality criterion. With clustering methods, a tree of three taxa is built first, then each additional taxon is added sequentially in the best available position. On the other hand, optimality criterion methods allows selecting the best tree from all possible trees based on how well that tree fits a particular model of character evolution (Pade & Holmes, 1998). The most commonly used methods of tree reconstruction are the distance-based neighbour-joining algorithm (NJ) and the maximum parsimony (MP), maximum likelihood (ML) and the Bayesian inference (BI) (Table 1.4) that all use discrete data and an optimality criterion. II.1.1 Neighbour-joining The NJ algorithm (Saitou & Nei, 1987; Studier & Keppler, 1988) uses both distance data and clustering to infer phylogenetic trees. It is a very fast method and performs well when sequences exhibit low divergence. However, when converting sequence data into a distance matrix some information is lost because the observed distance between sequences is not equivalent to the real evolutionary distance between them (Holder & Lewis, 2003). Indeed, multiple substitutions at the same site can make sequences artificially close to each other. For distantly related sequences it is therefore necessary to use a correction of the pairwise distance that takes into account multiple substitutions at the same site. There exist many models of sequence evolution, each with a different way of correct pairwise distances and the choice of the correction can be debated (Holder & Lewis, 2003). Thus, NJ is useful for an initial exploration of data, or for recently diverged sequences, but is seldom used as the sole phylogenetic reconstruction method. This method is often used as a starting point for more computationally intensive methods such as MP or ML. II.1.2 Maximum Parsimony Unlike distance-based methods such as NJ, MP and ML attempt to map gene sequence history onto a tree by assigning a score to trees corresponding to the plausibility of the mutation required by a particular tree to explain the data. In MP, the score is simply the minimum number of mutations and the tree with the lowest number is chosen as the best tree. It often happens that more 26 than one tree exhibit the same number of mutations to explain the data, in which case a consensus tree can be constructed from this set of most parsimonious trees to estimate the most consistently recovered clades. One serious drawbacks of MP is that this method does not take into account that the number of changes along branches is unlikely to be equal between branches. If the rate of sequence evolution differs between taxa, some branches will be longer than others. In MP such long branches might be considered more closely related than they truly are, simply because they have acquired the same mutations by chance. Thus MP is susceptible to long-branch attraction (Felsenstein, 1978) causing two long branches to be erroneously inferred to be the closest relatives of each other while they are not. Despite such drawbacks, it has been shown that this method can still perform well, even for divergent trees in which long-branch attraction is likely, given that enough sequences are sampled (Hillis, 1996; Kim, 1996). II.1.3 Maximum Likelihood With this method, the tree that is chosen is the one that, out of all possible trees, is the most likely to have produced the observed data according to a specified model of evolution. This method is particularly computationally expensive as it has to investigate two problems: first, for a given topology which combination of branch length is most likely to have given the data, and second, out of all available topologies which one has the greatest likelihood. The likelihood for each site is determined with the probability of that site having different nucleotides following the given model of sequence evolution. For instances if the model states that transitions, - a mutation between two pyrimidines (T↔C) or two purines (A↔G)-, are more common than transversions,- a mutation between a pyrimidine and a purine (A↔C, A↔T,G↔C or G↔T)-, then the probability of a site having both nucleotides A and G will be higher than the probability of a site having nucleotides A and T or A and C. Despite the computational drawback, this method remains attractive because it allows investigating all mutational scenarios compatible with the observed data by using a well known and reliable function (Holder & Lewis, 2003). However, it must be kept in mind that the likelihood value of a tree is not the likelihood that this tree is the true tree, but the indication that this tree has produced the observed data under the chosen model of substitutions. On the ability of the chosen model to describe the real process of nucleotide evolution depends the conclusion whether or not the tree is the true one. As said earlier, the actual number of possible trees given a set of DNA sequences quickly becomes enormous, rendering it impossible for MP or ML algorithms to explore in a reasonable 27 timeframe every possible tree. These methods sidestep this problem by using heuristic search methods, also called hillclimbing. Heuristic searches start with a particular tree, rearrange it to produce different trees and save any trees that are better than their predecessors. It thus allows investigating large datasets that would be impossible to analyse otherwise. As any method, this one has drawbacks. The main one is that it is possible with this method to keep trapped in a local optimum in the tree space that might not include the globally best tree. Another issue when building a tree lies in the fact that phylogenetic trees are complex structures whose confidence intervals are difficult to evaluate. Thus, after an analysis has run, sometimes for days or more, a reasonable question that arises is “how does the data support relationships seen in that tree?” Although several methods exist to assess confidence in a phylogenetic tree (Goldman et al., 2000), the most popular and most used one is ‘bootstrapping’ (Efron, 1979; Felsenstein, 1985). In boostrapping, pseudoreplicate of the data set is created from the original set of sequences by randomly selecting nucleotide sites and building new sequences of the same length as the originals. Each site is replaced after sampling and can be sampled several times, thus, some sites might appear more than once in the pseudoreplicate, while others might be absent altogether. Many pseudoreplicates are produced (usually between 100 and 1000) on which the tree-building algorithm is performed. Bootstrap supports are then calculated for clades, indicating the proportion of times a clade was recovered from all pseudoreplicates. As a result, a clade with a low bootstrap support is sensitive to the exact combination of sites that were sequenced and might therefore not appear if another data set were to be collected. Despite the popularity of bootstrapping, the interpretation of bootstrap proportions remains unclear and what is the cut-off value indicating a ‘good’ clade is not defined. Some have argued that bootstrap values being conservative, 70% might indicate a strong support for a group (Zharkikh & Li, 1992; Hillis & Bull, 1993). In any case, what bootstrap values tell us is whether the same result would be found given additional data collection, not whether this result is correct. Indeed, even a wrong tree can obtain strong bootstrap values, as in the case of a tree obtained because of long-branch attraction (Holder & Lewis, 2003). In the last decade, an alternative method to MP and ML with bootstrapping has emerged that has become very popular and that I have used for all my phylogenetic tree estimations, Bayesian Inference. 28 II.1.4 Bayesian inference of phylogeny Although Bayesian statistics were described in the 18th century, it was not until 1996 that they were used to estimate phylogenies (Rannala & Yang, 1996; Huelsenbeck et al., 2002). In this framework, estimating a phylogenetic tree consists in evaluating the posterior probability of a tree according to Bayes theorem: Pr[Tree | Data] = Pr[Data | Tree] x Pr[Tree] Pr[Data] Where Pr[Tree] is the prior probability of a phylogeny and Pr[Data | Tree] is the likelihood of the observed sequences given that phylogeny. In this model, all trees are considered equally probable a priori and a given model of nucleotide evolution is used to calculate the posterior probability of a tree producing the observed data. The posterior probability of a tree can be seen as the tree is true given the data set and the specified model of evolution (Huelsenbeck et al 2001). The calculation of posterior probabilities necessitates the summing of probabilities over all trees, and all possible combinations of branch lengths and parameter values for each tree. Thus, such calculation is virtually impossible as such and the estimation of posterior probabilities therefore implies the use of an algorithm such as the Markov Chain Monte Carlo (MCMC) algorithm. MCMC methods calculate optimal trees and their accompanying posterior probabilities with randomly chosen parameter values. A calculated tree is compared with the preceding tree and then either accepted or rejected as the new tree based on whether it improves the posterior probability or not. In this way, the MCMC methods can produce a reliable estimation of the posterior distribution and the accompanied parameters in a very quick way. The Metropolis algorithm (Metropolis et al 1953; Hastings 1970) is the most widely used MCMC method: a starting tree T1 and its posterior probability are calculated with randomly chosen parameter values. Then, a closely related tree T2 with slightly changed parameter values is calculated. The algorithm then computes the ratio of the posterior probabilities of these trees. If the ratio is >1, the new tree T2 is accepted and the next neighbouring tree T3 is calculated. If the ratio is <1, a random number between 0 and 1 is drawn. If the drawn number is less than the ratio, the new tree T2 is accepted and the next tree T3 calculated. Otherwise, T2 is rejected and another neighbouring tree Ti is calculated. In this way, a long chain of trees is calculated that over time converge towards trees with high probabilities. The danger of getting stuck in local optima is lowered by the chance that a calculated tree, although worse than its preceding tree, can be accepted if the ratio is higher than a randomly drawn value. This procedure is repeated millions of time and the proportion of time a tree is visited constitutes a valid 29 approximation of the posterior probability of that tree (Huelsenbeck et al 2001, 2002; Tierney 1994). Both ML and BI rely on a likelihood function and can use the same models of evolution. In ML statistical support is provided by bootstapping, while in BI support can be calculated by creating a consensus tree on which are reported the posterior probabilities of individual clades. In this case, these posterior probabilities indicate the posterior probability that a clade is true given the data set and the model of evolution. BI offers the advantage over ML that the estimation of support values is much faster, especially for large data sets that are becoming the norm. One of the problems of BI is that support values derived from posterior probabilities are too liberal, as opposed to the more conservative bootstrap method. Reasons for this discrepancy are not yet fully understood but it has been argued that it might be due to inaccuracies in the model of evolution used (Wròbel 2008). While bootstrap values of 70% are usually seen as indicative of a good support, BI posterior probabilities values of 95% are usually accepted as good support in a phylogenetic tree. When starting a BI analysis, the MCMC usually starts from a random point in the parameter space and it can take some time before the Markov chain reaches a region with high posterior probabilities in that parameter space. Thus, this initial period (“burn-in”) is discarded from the analysis. Once the chain has converged, i.e. it has reached an appropriate region of the parameter space, it will explore other regions through that space. However it happens that moving from one region to another may take the chain a long time, or it might become trapped in one region containing several close local optimums. One of the difficulties of BI is therefore to estimate an appropriate burn-in period and assessing if it is exploring the character space properly. An indication can be obtained by checking the evolution of log likelihood values. The latter tend to increase during the burn-in period and then reaches a plateau from where it fluctuates randomly up and down. Although this can give a hint at the efficiency of the analysis, the best solution probably remains to run several MCMC analyses: if results from these different runs are similar, then it is an indication that the Markov chains appropriately converge and explore the parameter space (Huelsenbeck et al 2002). As mentioned several times in the previous paragraphs, methods such as NJ, ML or BI use a specified model of nucleotide evolution to reconstruct phylogenetic trees that can match the DNA sequence data. 30 Table 1.4 (derived from Holder & Lewis, 2003) Summary of some advantages and disadvantages of the four methods used in phylogenetic reconstruction. Method Neighbour joining Advantages Disadvantages Fast Information is lost in compressing sequences into distances; reliable estimates of pairwise distances can be hard to obtain for divergent sequences Maximum Parsimony Fast enough for the analysis of hundreds of sequences; robust if branches are short (closely related sequences or dense sampling) Can perform poorly if there is substantial variation in branch lengths Maximum Likelihood The likelihood fully captures what the data tell us about the phylogeny under a given model Can be prohibitively slow (depending on the thoroughness of the search and access to computational resources) Bayesian Inference Has a strong connection to the maximum likelihood method; might be a faster way to assess support for trees than maximum likelihood bootstrapping The prior distributions for parameters must be specified; it can be difficult to determine whether the Markov chain Monte Carlo (MCMC) approximation has run for long enough II.1.5 Models of nucleotide evolution Evolutionary processes produce different patterns of nucleotide changes in DNA (Bos & Posada, 2004). Not taking such patterns into account can lead to estimation of wrong phylogenetic relationships (Bos & Posada, 2004; Sullivan & Joyce, 2005). Different models of nucleotide evolution have been proposed over the years that try to explain molecular evolutionary processes, with different degree of complexity. Most models are related but differ in the number of parameters they integrate. Parameters involved in model differences include nucleotide frequency within sequences, the probability of different types of substitution (transitions, transversions) and the substitution rates between different sites (Pages & Holmes, 1998). The simplest model of nucleotide substitution was introduced by Jukes & Cantor (1969). By assuming equal base frequencies within the sequence and equal substitution probabilities this model (JC69) is mathematically simple but is unlikely to reflect the actual nucleotide evolution and to allow an accurate estimation of phylogenetic relationships. Following this model, Kimura (1980) proposed what is known as the Kimura 2 Parameter model (K2P). In this model, different rates for 31 transitions and transversions are used, giving a higher probability to transitions, to take into account observed differences in substitution type, especially in mitochondrial DNA (Wakeley, 1996). It must be noted that if transition and transversion rates are equal this model becomes the JC69 model. Later, Felsenstein (1981) elaborated a model that addresses the possibility that base frequencies in a set of sequence can differ. In that case, it is reasonable to expect that some types of substitutions might be more common than others. Felsenstein model (F81) allows the frequencies of the four nucleotides to be different, but not to the point where base frequencies are not roughly equal across all the sequences any more. This obviously might not always be true. In order to make models closer to observed data, several authors then built more complex models incorporating multiple aspects of sequence evolution. Amongst such models, the Hasegawa Kishino Yano model (HKY; Hasegawa et al., 1985) allows both different rates for transitions and transversions and different base frequencies. Thus, it incorporates the K2P and F81 models. However, this model considers all transversions and all transitions to have the same probabilities. A more complex model taking into account that transitions rates might be different between purines and between pyrimidines, and that all transversions might not have the same probability was proposed by Rodriguez et al. (1990). This model is now referred to as the General Time Reversible model (GTR). These are the most used and best known models. Others exist that I will not present here. All the models presented above assume among-site substitution rate homogeneity. Although this might be true for non-coding portions of DNA, protein coding regions are very likely to be under functional constraint and any mutation in these regions might change the functional properties of the coded protein thus individuals bearing this mutation could have a reduced fitness, or, if the functional change is lethal, be simply removed from the population. Furthermore, since the genetic code is degenerate, nucleotide substitution will not have the same effect depending on the position of the modified nucleotide. Indeed, approximately half of third-codon positions are “fourfold degenerate”, i.e. any of the four nucleotides at this position will code for the same amino acid. Conversely, a nucleotide change in the second codon position will almost invariably lead to an amino acid substitution. Hence it appears intuitively that the second codon position will be more constrained than the third codon position. This was confirmed by Nei (1987) in the globin family of genes, in which substitutions are about three times more frequent at the third codon position than at the first or second positions. The most successful method to incorporate rate variation among sites in models is the use of a gamma distribution (Γ; Bos & Posada, 2005). This statistical representation of rate variation among sites can be added to any model such as the one I described above. The shape parameter, α, 32 of this distribution is determined by the number of nucleotide sites with various rates of substitutions, with a shape of the distribution skewed to the right when most nucleotides are invariable (Fig. 1.7). Thus, when α<1 a few nucleotides are evolving very rapidly and account for most of the variation in the data. When α>20 most sites have intermediate rates of evolution and few nucleotides evolve very slowly and very rapidly. The greater the value of α, the smaller the difference of rate variation between nucleotides and the lesser it is necessary to take among site rate variation into account (Bos & Posada, 2005). Figure 1.7 From Bos & Posada (2005). Gamma distributions calculated using different shape parameters (α). The shape parameter is determined by the number of nucleotides in a sequence evolving at a particular rate. When most nucleotide sites are invariable and variation concentrated at a few rapidly evolving nucleotide sites, the shape parameter is small (<1). It becomes larger while the proportion of variable nucleotide sites increases. This indicates that more sites evolve at a moderate rate and fewer sites have extremely high or low rates. When using model-based phylogenetic methods such as BI or ML, the choice of model greatly influences the outcome and the performance of the analysis. Although for long sequences with low polymorphism model choice is not so much an issue, for sequences with high levels of polymorphism choosing the wrong model can sometime lead to support a wrong tree topology (Bos & Posada, 2005; Sullivan & Joyce, 2005). Therefore, model choice is of critical importance in molecular phylogenetic estimation. Choosing an overly simplistic model will most likely affect negatively phylogenetic estimation, unless no very short or very long branches are present in the phylogeny that is estimated. A temptation that naturally comes to mind is therefore to always use the most complex model to evaluate phylogenies, as it has been suggested by some authors, e.g. for Bayesian phylogenetic inference (Huelsenbeck & Rannala, 2004). However overparameterisation 33 can lead to a loss of degrees of freedom hence to nonidentifiability of parameters and more complex models with better likelihood scores should not be expected to be better for phylogeny estimation (Sullivan & Joyce, 2005). Therefore the question remains as how to choose the best model given a data set of DNA sequences. Several methods exist. The Likelihood Ratio Test (LRT) is one of the most used model selection method (Sullivan & Joyce, 2005). The LRT statistics calculation necessitates a starting tree to estimate the likelihoods of the different models: it first calculates the likelihood score of a null model L0 and an alternative more complex model L1. The two scores are then compared according to the equation δ = 2(ln L1 − ln L0) The test statistic is then evaluated under an assumption of asymptotic convergence to a χ2 distribution, with the degrees of freedom being the difference in number of free parameters in the two evaluated models (Bos & Posada, 2005; Sullivan & Joyce, 2005). This method suffers from several drawbacks. The most serious being that the model space must be traversed by a series of pairwise comparisons, but no relevant theory exists to explain how to do so (Sullivan & Joyce, 2005). Another weakness of LRT is that this method will tend to favour the more complex model. Other methods include the Aikake Information Criterion (AIC; Aikake, 1973). The AIC selects the best-fit model by computing models’ likelihood while penalysing parameter-rich models in an attempt to avoid overparameterisation. The chosen model is then the one with the smallest AIC value, with AIC = -2 ln Li + 2 Ni, whith Li the likelihood for model i and Ni the number of free parameters in model i. The AIC offers several advantages over the LRT. First, it can compare all models simultaneously instead of conducting pairwise comparisons; second, it avoids overparameterisation, at least partially; third, AIC allows model averaging and model selection uncertainty calculation; four, AIC attempts to find the model that best approximates the true unknown model of evolution given the information provided by the data. Alternatively, the Bayesian Information Criterion (BIC) can be used to select the model that provides the best estimation of the true evolution of nucleotides. BIC was first proposed by Schwartz (1978) and assesses the fit of a model through maximum likelihood scores and a heavier penalisation of overparameterisation than AIC. One of the advantages of BIC is that it avoids the tendancy to select more complex models when sample size increases (Sullivan & Joyce, 2005). 34 II. 2 New perspectives in molecular phylogenetics Phylogenetics tries to describe the history and relationships of species and populations (Edwards, 2009). Until recently the common molecular phylogenetic method implied sequencing one gene in individuals from different species, inferring a tree from the obtained sequences and considering this tree derived from one gene to represent the species relationships (Degnan & Rosenberg, 2009). However, it has been recognised for a long time that species trees are not the same as gene trees, where, following Degnan & Rosenberg (2009), a species tree is “a tree of ancestor–descendant relationships for a set of populations” and a gene tree “a tree of ancestor– descendant relationships for a gene, where the same gene is sampled from several individuals”. In the 1960’s and late 1970’s, Cavalli-Sforza (1964) and Felsenstein (1981) were already using simple models to take this difference into account, and in the 1980’s several authors investigated this issue, as presented in Pamilo & Nei (1988). In the early 1990’s, heterogeneity between gene trees within species had been observed but did not receive much attention in the following years (Edwards, 2009) for several reasons. One of these was the lack of phylogenetic methods able to deal with large dataset and with the complexity required to incorporate this homogeneity into analyses (Brito & Edwards, 2009; Edwards, 2009). The development in the last decade of analytical approaches based on the coalescent theory and Bayesian statistics, and the increasing ease with which it is becoming possible to collect molecular data from a wide range of loci, rendered possible to tackle this issue. Fields such as phylogeography (Avise et al., 1987; Avise, 2000) promoted the development of such approach and models in which the estimation of population parameters above gene level implies to embed gene histories in species history (Edwards, 2009). It has been argued that since phylogenies affect populations, species tree parameters and the ways to evaluate them from genetic data are in the same category than tools and models developed in the fields of phylogeography to evaluate demographic parameters (e.g. genetic diversity, gene flow, population divergence …) within species (Maddison, 1997; Avise, 2000; Edwards, 2009). Thus, all these advances, in particular in coalescent theory, are now slowly closing the gap between these fields (Brito & Edwards, 2009) and it seems that methodological and conceptual differences developed between phylogenetics and phylogeography (Hey and Machado 2003; Brito and Edwards 2009) are likely to disappear, thus answering to Felsenstein statement that “systematicists and evolutionary geneticists do not often talk to each other” (Felsenstein (1988) in Edwards, 2009). 35 THESIS PROBLEMATIC Contrary to Darwin’s view that accepting evolution leads to relegating species to artificial combinations of organisms (Darwin, 1859), it is now accepted that species constitutes a fundamental and essential biological entity (Mayr, 1982). This biological unit is used in all fields of biology. Being universally recognised, despite the debates surrounding the concept described by the word, species is fundamental in Conservation Biology and is used by virtually all organisations involved in the conservation of biological diversity and its processes. Besides, in the actual context of erosion of biodiversity and 6th mass extinction, description of this biodiversity, i.e. the identification of species and their relationships, has a renewed and significant environmental and economical importance. To describe and understand biodiversity, in order to conserve it, it is necessary to: 1) carry out taxonomic and phylogenetic description of biodiversity, i.e. what and how many species are there? Where are they? What are the relationships between them? 2) investigate the patterns of differentiation leading to the formation of species, i.e. how do lineages diversify? What are the effects of gene flow between taxa on their differentiation? What are the effects of distance and dispersal abilities? How quickly do taxa diverge to give rise to species? To cite but a few questions. Although answering such questions might maybe be relatively straightforward sometimes in some terrestrial taxa with limited dispersal abilities and easily identifiable taxonomic limits, such is not the case in many marine organisms, in particular seabirds. Indeed, these organisms exhibit somewhat paradoxical evolutionary traits. Being highly vagile organisms mostly living in an open environment, they face very few physical barriers to dispersal, a characteristic supposed to enhance gene flow between populations therefore limiting genetic differentiation between these populations (Wright, 1931). However, counteracting selective forces consisting in the discrete location of breeding localities and extreme philopatric behaviour should, on the reverse, enhance the effect of genetic drift. As a consequence of these conflicting processes acting on their population genetics, seabirds show cases of population structure and/or differentiation in the absence of obvious barrier to gene flow (e.g. Xanthus’s murrelet Synthliboramphus hypoleucus; Friesen et al., 2007). Explanations to such pattern include the existence of non-visible barriers to gene flow, such as 36 oceanic fronts (e.g. the Almeria-Oran front in the Mediterranean sea that seems to prevent gene flow between Atlantic and Mediterranean populations of Cory’s shearwater Calonectris diomedea; Gomez-Diaz et al., 2006), large expanses of low productivity waters such as in the western and eastern equatorial Pacific ocean or the distribution at sea during the nonbreeding season (Friesen et al., 2007). These characteristics create very complex phylogeographic patterns, which, combined to the insularity of most species and their morphological similarities, render taxonomic delineation and understanding of patterns and processes of differentiation extremely difficult in seabirds, in particular in petrels (order Procellariiformes). Because seabirds are declining at a much faster rate than any other bird group (Birdlife International; Fig. 1.8) it is necessary to undertake thorough investigations to bring answers to points 1) and 2) in order to conserve them. It is in this context that this PhD work is embedded. In this study, within seabirds, the biological study organisms used are members of a particularly threatened and complex group, the petrels. To address the issues mentioned above, we have seen that genetic and molecular phylogenetic tools are particularly pertinent. Through the recent theoretical and methodological developments in these fields, it is now possible, not only to reconstruct phylogenetic relationships but also to investigate the relationships of lineages through the use of multiple loci analyses and to understand the complex, sometimes contradicting, histories of the many genes carried by organisms. It becomes also possible to quantify gene migration between populations and reconstruct patterns of diversification through time and space to better understand those to come in the future. 37 Figure 1.8 Evolution of degree of threat in six groups of birds (Birdlife International, 2009) 38 Part 2. Study Organism I Seabirds The distinction between seabirds and other birds, particularly shorebirds, is not easy to define (Brooke, 2002; Schreiber & Burger, 2002). Traditionally, are considered seabirds, members of three Orders, Sphenisciformes (the penguins), Procellariiformes (albatrosses, petrels, shearwaters, diving-petrels and storm-petrels), Pelecaniformes (tropicbirds, gannets, boobies, cormorants, anhingas, frigatebirds), and some families in order Charadriiformes (Stercorariidae i.e. skuas; Alcidae, the auks; and Laridae/Sternidae, gulls and terns) (Brooke, 2002; Friesen et al., 2007a). Historically, from the fossil record, it has been suggested (Warheit, 2002) that, apart from a bird apparently belonging to genus Puffinus from the early Oligocene (circa 29-34 millions years ago), most seabird’s modern genera did not appear until the early Miocene, approximately 16 to 23 millions years ago (mya), although older members of seabird orders are known, e.g. family Diomedeoididae in Procellariiformes (Mayr, 2009). From the fossil records, Warheit (1992, 2002) also pointed out that there seem to be a correlation between the taxonomic diversification of seabirds in the Miocene and the extensive climatic and oceanic changes that occurred during the middle Miocene transition: around 11-17 mya, sharp oceanic temperature drops (Shevenell et al., 2004) and oceanic current changes (Miller & Fairbanks, 1983) happened that might have offered improved ecological conditions due to increased oceanic productivity resulting from the cooling of ocean surface temperatures (Warheit, 1992) thus promoting seabird lineages diversification. Although such a diverse group obviously present much differences between its various members, commonly shared characteristics (Schreiber & Burger, 2002; Friesen et al., 2007a) include mostly pelagic distribution during the non breeding season, marine foraging during breeding, colonial breeding, either on isolated islands or on cliffs, late sexual maturity (with a large interval of age at first reproduction, from 2 to 13 years; Jouventin & Dobson, 2002), low annual fecundity, biparental care and long lives (e.g. Sagar & Warham, 1993 for a case in albatrosses). Evolutionary speaking, seabirds have provided some of the most convincing cases of sympatric speciation, due to temporal segregation of breeding populations (e.g., band-rumped storm petrel Oceanodroam castro; Monteiro & Furness 1998) or extreme natal philopatry (Friesen et al. 1996, Friesen & Anderson, 1997; Congdon et al. 2000; Steeves et al. 2005). Cases of peripatric and parapatric speciation have also been shown, e.g. in the white-capped albatross (Thalassarche steadi; 39 Abbott & Double, 2003) and nazca boobies (Sula granti; Friesen et al., 2002). Nevertheless, it is now generally agreed that almost every speciation events are associated with some kind of geographic separation (Fitzpatrick et al., 2009; Coyne & Orr, 2004) and allopatric speciation seems to be the dominant mode of speciation in birds, including seabirds (Friesen et al., 2007a), with only a mere 5% of speciation events compatible with sympatric scenario (Phillimore et al., 2008). Yet, in many taxa species diagnostics is not straightforward due to the geographic isolation of their populations (Coyne & Orr, 2004). This arises because phenotypic similarity may be maintained by similar ecological conditions even when gene flow has ceased and DNA sequences have diverged (Price, 2008). II Procellariiformes Thorough reviews on the ecology of Procellariiformes can be found in the two masterpieces of Warham (1990, 1996) and Brooke (2004), and most material presented in this brief introduction to these birds are assembled from these references. Order Procellariiformes comprises albatrosses and petrels (Warham, 1990; Brooke, 2004). The term petrel includes a wide range of species, belonging to three families, from the storm-petrel to the diving-petrels and an assemblage made of gadfly petrels, shearwaters and allies. Albatrosses are sometimes included within the general term petrel (Warham, 1990; Brooke, 2004). The main feature characterising Procellariiformes from other seabirds are their tubular nostrils. Another difference with other seabirds is their digestive track that lacks a crop, contrary to other birds. Procellariiformes are probably the best representation of what is a seabird. Typically, Procellariiformes are pelagic, long-lived species, breeding only once a year or every second year. They are colonial breeding birds and lay only one egg that cannot be replaced if breeding fails, although relaying has been documented in diving-petrels and storm-petrels (Warham, 1990). Both sexes share responsibilities in incubation and chick rearing. All species exhibit high partner fidelity between years, a feature that might be associated with the high degree of cooperation needed to successfully rear a chick, the breeding period being longer than in other similar sized seabirds. These birds are also slow to mature and mostly do not breed until several years of age. This slow generation time renders the Procellariiformes particularly sensitive to adult mortality. All species show colouration made of a combination of black, grey, white and brown. In terms of sizes, the order Procellariiformes has the largest range in birds, from the Least storm-petrel (Halocyptena microsoma) and its 20.5cm wing for 20.5g to the Southern royal albatross, wing length 69.8/66.6 cm (males/females) and 10.3/7.7 kg (Brooke, 2004). 40 II.1 Threats and Conservation As shown by Fig. 1.8, seabirds are probably the fastest declining group of birds in the last 20 years, and this poor record is essentially driven by high declines in Procellariiformes populations. Table 1.5 shows the percentage of species categorised in different conservation statuses for the four families in the order (calculated from Birdlife International data; Birdlife International, 2010). Overall, 45.7% of species are threatened in the order, and only 38.2% of species are considered as Least Concern. Two families do not have any species in this last category, and only Hydrobatidae have a majority of species categorised LC. Family Diomedeidae is the most seriously threatened with a staggering 77.3% of its species being threatened! However, no species of albatross is considered extinct by Birdlife International (2010). Contrary to Diomedeidae, in family Procellariidae two species are recognised extinct by Birdlife International (2010), a Bulweria and a Pseudobulweria that both used to breed on St Helena island in the Atlantic Ocean and that were extirpated circa 1500 (Olson, 1975). Jamaica petrel (Pterodroma carribea), although not officially classified as Extinct is very likely to have suffered the same fate than St Helena birds and the last confirmed record dates from 1879. However, recent rediscovery of the formerly presumed extinct Beck’s petrel (Pseudobulweria becki; Shirihai, 2008) proves that, maybe all hope should not be abandoned for the Jamaica petrel. Additionally, Rando & Alcover (2008) reported the extinction of a shearwater from the Canary Islands, Puffinus olsoni, thought to have disappeared later than 1270, probably soon after the arrival of the first European in the Canary islands. Other shearwaters have been reported as getting extinct following human colonisation of their breeding islands. These include P. spelaeus (Holdaway & Worthy 1994) from New Zealand that probably disappeared due to the introduction of Pacific rat (Rattus exulans) by Maoris, and P. parvus from Bermuda (Olson 2004) probably driven to extinction after human arrival in these islands in the 16th century. Other potential examples of human induced extinction or extirpations can be found in the Pacific Ocean islands (such as Easter Island or Ua Huka in the Marquesas; Steadman, 1995), where almost every island colonisation by human has led to the extirpation or extinction of one Procellariidae (Steadman, 2006). Nowadays, family Procellariidae is the second most threatened family within Procellariiformes (Table 5) but the few examples cited above show that this family has suffered a dramatic decline in the last few thousands of years following the progressive colonisation of their remote breeding islands by humans. 41 Table 1.5 Percentage of Procellariiformes species in IUCN conservation categories Diomedeidae EX 0 CR 13.6 EN 27.3 VU 36.4 NT 22.7 LC 0 DD 0 Procellariidae 2.4 9.8 12.2 20.7 13.4 41.5 0 Pelecanoididae 0 0 25 0 75 0 0 Hydrobatidae 8.7 8.7 4.3 4.3 56.5 17.4 0 Overall 1.5 9.9 14.5 19.8 13.0 38.2 3.1 EX: Extinct; CR: Critically Endangered; EN: Endangered; VU: Vulnerable; NT: Near Threatened; LC: Least Concern; DD: Data Deficient Many threats affect seabirds and obviously Procellariiformes are no exception and many of these taxa have significantly declined in the last century (Warham, 1990; Brooke, 2004). Causes of decline of Procellariiformes population can be divided in two categories, causes on land in the breeding/nesting habitat, and causes affecting the birds at sea, i.e. foraging habitat during breeding season and pelagic habitat during non-breeding season. Using very different habitats renders seabird sensitive to any change or threat that can affect one of these, thus all age classes can be affected, from eggs and chicks in the nesting grounds, to immatures and adults, breeding or not, at sea. In general, species become endangered due to habitat loss, over-harvest, invasive species, pollution and diseases (Wilcove et al., 1998), to which for seabirds, we can add global climatic change and fisheries by-catch. Their breeding habitat being confined to small geographic regions, i.e. usually oceanic islands and/or cliffs, any loss or damage to these restricted coastal or island environments can have very negative consequences on Procellariiforme populations. Hence, habitat loss or damage, either through direct anthropogenic destruction or more natural causes, is indeed an important threat to their and seabirds’ survival in general (DeeBoersma et al., 2002). For instance, fire in breeding colonies can have catastrophic effects on breeding success and adult survival, as exemplified by the fire that affected Zino’s petrel (Pterodroma madeira) breeding colony in Madeira in August 2010, killing several adults and 65% of this year chicks (Birdlife International, 2010). Human induced decline or extinction is well documented on land and can be linked to either direct harvesting of birds by people or indirect effect of environmental changes (Brooke, 2004; Steadman, 2006). Direct harvesting of Procellariiformes for food or other use used to be a widespread practice described worldwide (Warham, 1990; Brooke, 2004) and some species were driven close to extinction by such harvesting (e.g. Short-tailed albatross, Phoebastria albatrus, Kuro-o et al., 2010). Habitat 42 degradation due to anthropogenic causes include the introduction on breeding grounds of herbivores like goats, sheep, cattle or rabbits that can have very deleterious effect on petrel populations by degrading the breeding habitat through tramping or by modifying vegetation cover and inducing erosion. Habitat modification or destruction for urbanised areas or agriculture can also be a cause of mortality in Procellariiformes, and mortality of juveniles due to artificial lights could contribute to the demise of species already struggling to survive, such as happening the Mascarene petrel (Pseudobulweria aterrima) in Réunion island or Newell’s shearwater (Puffinus newelli) in Hawaii (LeCorre, 2003; Warham, 1996). However, light-induced deaths mainly affect young birds and survival of long-lived species such as petrels is more affected by adult mortality than juvenile mortality. The impact of light-induced mortality on species survival can therefore be controversial. In addition to habitat degradation, another cause of seabird mortality while on land that have probably had a much greater negative impact on petrel populations consists in predation by introduced species (Brooke, 2004). Indeed, Procellariiformes being mostly adapted to oceanic environment are rather clumsy on land making them easy preys for introduced predators such as cats. Furthermore, because they evolved on isolated oceanic islands without predators, Procellariiformes taxa did not developed escape or defensive behaviour, either as chicks or adults, to respond to predation, rendering them very sensitive to this threat. Predators like cats and rats have had a tremendous negative impact on Procellariiformes populations everywhere they have been introduced. Fortunately, methods and know-how now exist that allow to clear islands from these predators while limiting the impact on local fauna and flora (Donlan, 2008). Other predators that have impacted Procellariiformes populations, as well as other fauna, on islands include pigs, stoats and mongooses (Brooke, 2004). While at sea seabirds in general and Procellariiformes in particular can be confronted to several threats. Impact of fisheries bycatch is still, despite much improvement in the last years, very important and threatens many populations of albatrosses and petrels (Grémillet & Boulinier, 2009). Concomitant with bycatch, is the over-exploitation of fish resources by industrial fisheries that can potentially affect many seabirds populations (Grémillet & Boulinier, 2009). Another threat for seabirds at sea lies in marine pollution, either through chemicals, heavy metals, plastics or oil products although its impacts are difficult to assess (Burger & Gochfeld, 2002). Finally, the most recent threat to affect seabird communities consists in climate change (review in Grémillet & Boulinier, 2009). Climate change can impact Procellariiformes on land by driving habitat changes through a vegetation shift in breeding colonies or a shift in parasite loads or attacks (Robertson 1998). Climate change is also likely to induce a mismatch between Procellariiformes ranges and their preys prompting lower reproductive success as has probably happened in the California 43 current (Veit et al., 1997). Furthermore, sea-level rise potentially accompanying climate change could also destroy low lying atolls hosting colonies of some tropical Procellariiformes (Brooke, 2004). It was also pointed out by Grémillet & Boulinier (2009) that effects of climate change and overfishing can hardly be considered separately and they could impact seabirds synergistically. To such impacts these authors describe three potential responses of seabirds: (i) change their feeding behaviour and preferences to survive within the same range; however all studies to date indicate a negative effect of such change on seabird populations (reviewed in Grémillet & Boulinier, 2009). Furthermore, in Procellariiformes, several studies showed the negative impact of global change on demographic parameters of Procellariiformes (e.g. Grosbois & Thompson, 2005; Jenouvrier et al., 2005); (ii) modify their range. It has been speculated that range shift in the non-breeding season in Black (Oceanodroma melania) and Least storm-petrels was caused by change in oceanic productivity in the central, equatorial Pacific caused by climate change (Ainley & Divoky, 2001; Ainley et al., 2005); (iii) some species might be unable to change their feeding behaviour or feeding/breeding range thus facing a direct risk of extinction linked to global change and overfishing (Grémillet & Boulinier, 2009). As a consequence of these threats, seabird communities in particular Procellariiformes have experienced a regular decline and an increase in the number of threatened species over the last decades. The challenge to conserve and protect Procellariiforme population worldwide and the many threats sometimes acting synergistically renders the use of various tools necessary, from mathematical and modelling tools (e.g. Dumont et al., 2010) to molecular tools. II. 2 Systematics and Phylogenetics The order comprises 131 species recognised by Birdlife International (Birdlife International, 2010). These species are grouped in four families, the Diomedeidae (Albatrosses; 22 species in four genera), Hydobatidae (Storm Petrels; 23 species in eight genera), Procellariidae (the petrels, shearwaters, prions, fulmars, and similar; 82 species in 14 genera), and the Pelecanoididae (the diving petrels; four species in one genus). Family Procellariidae, the most speciose in the order, is made of 14 genera, amongst which genus Pterodroma comprises 32 recognised species (Birdlife International, 2010), 39% of the species in the family and 24.4% of all species in the order. Traditionally, the phylogenetic relationships in this order have been difficult to establish and the systematics of the group have been fluctuating over time. Brooke (2002) suggested that, because most albatrosses and petrels are highly philopatric, i.e. return to breed in the colony where they were born (sometimes just a few metres from the nest site where they hatched, or even in the very 44 same burrow!), and because a significant proportion breed on only one or a few islands, few birds disperse to other colonies, which reduces the selection for plumage divergence or morphological differentiation. As a result traditionally used characters such as colourations or morphology could be poorer guide than usual to infer phylogenies. Alternative explanations to colouration similarities can be linked to selective pressures at sea and might probably be linked to a combination of several factors related to competition and predation (Bretagnolle, 1993). This has led to many cases of phylogenetic uncertainties, including cryptic species with unclear phylogeographic histories (Brooke, 2004) causing constant revision of their taxonomy. With the rise of molecular phylogenetics in the 1990’s, it was hoped that phylogenetic relationships and taxonomy in the order could be clarified. Indeed, several pieces of work brought big changes to traditional phylogenies. First, Kuroda et al. (1990) using protein electrophoresis produced a phylogeny showing the storm-petrels as the sister group of all other Procellariiformes. The same year, in their major review of bird classification Sibley & Ahlquist provided the phylogenetic tree shown in Fig. 1.9 by means of DNA hybridisation. Contrary to widely held views considering Sphenisciformes as the sister Order to Procellariiformes (Ho et al., 1976; Saiff, 1976; Mey et al., 2002; Mayr and Clarke, 2003; Ksepka et al., 2006), this tree placed Gavia as closest relatives to Procellariiformes, and placed the storm-petrels outside a clade containing the three other families, i.e. Diomedeidae, Procellariidae and Pelacanoididae. Although this work spurred some criticism it was an interesting step forward that clarified a number of issues (Warham, 1996). Subsequently, in a major work based on Cytochrome b, Nunn & Stanley (1998) provided one of the most complete phylogeny of the order. It recovered the albatrosses, petrels and diving-petrels as monophyletic as previously shown by Sibley & Ahlquist (1990). However, the fourth family, Hydrobatidae, was found to be paraphyletic, although storm-petrel sub-families were found monophyletic (Nunn & Stanley, 1998). Following this study, Kennedy & Page (2002) attempted to build a supertree based on various phylogenetic trees from several studies. They also built a MP tree based on a supermatrix approach that merged data from several cytb-based trees. Their analyses recovered broadly the topology of Nunn & Stanley (1998), with family Hydrobatidae paraphyletic. However, depending on the approach, placements of both storm-petrel clades were not similar, thus raising questions about this family and its relationship with other Procellariiformes. Furthermore, placement of some genera conflicted with other studies, e.g. Bulweria and Pseudobulweria, (Bretagnolle et al., 1998). Finally, the last major attempt to reconstruct Procellariiformes phylogeny, by use of cytb, by Pennhallurick & wink (2004) recovered the paraphyletic nature of storm-petrels, which were placed along with Diomedeidae outside a clade containing 45 Pelecanoididae and Procellariidae. However, this last study was flawed with shortcomings (Rheindt & Austin, 2005) and its results should be taken with caution. In addition to these various attempts to reconstruct phylogenetic relationships at the order level, several studies based on molecular tools also investigated relationships at the genus, species or population level and prompted taxonomic revisions or the discovery of cryptic species or strong genetic population structuring. Here are some examples. In albatrosses, successive works from Nunn et al. (1996), Abbott & Double (2003), Cuthbert et al. (2003) and Burg & Croxall (2004) have prompted the recognition of four genera and 22 species, from two genera and 14 species traditionally accepted before the study of Nunn et al. (1996). It can be noted that, interestingly, some of the genera (such as Thalassarche and Phoebastria) were “resurrected” (Nunn et al., 1996). In storm-petrels, several studies have investigated phylogenetic and population relationships within the Band-rumped storm petrel complex and revealed cryptic species and population structure (Friesen et al. 2007b, Smith & Friesen 2007, Bolton et al. 2008). In family Procellariidae, phylogenetic uncertainties exist in almost all genera, which were only partly explored. For example, in the shearwaters (genus Puffinus) much debate exists surrounding the phylogenetic relationships within the Audubon/Little shearwater (P. lherminieri/assimilis) complex. Different taxonomic treatments of this complex provided from one to eight species and from seven to 19 subspecies depending on the authors (review in Austin et al., 2004). In this major study, Austin et al. (2004) proposed five species, three “ocean-based” (bailloni, lherminieri and assimilis) and two formerly misclassified (subalaris, myrtae), and 14 “lower level” taxa, mostly corresponding to subspecies. In genus Macronectes, long-standing debates concerning the relationships between the two recognised species, the nonmonophyly of one taxon and the potential importance of hybridisation existed. Recently, based on two types of molecular markers, Techow et al. (2010) found support additional to morphological and behavioural evidence, for the definitive recognition of two species recently diverged and still presenting some degree of gene flow. Finally, Techow et al. (2009) confirmed the specific status of spectacled petrel (Procellaria conspicillata) that was for a long time considered part of the wide ranging White-chinned petrel (Procellaria aequinoctialis), by means of molecular evidence. In genus Pterodroma, Browne et al. (1997) showed that populations of petrels from the Galapagos and Hawaï deserved specific status based on a study using allozyme electrophoresis, thus separating the former taxon in two species, one Critically Endangered (Galapagos petrel, Pterodroma phaeopygia) and one Vulnerable (Hawaiian petrel, Pterodroma sandwichensis) (Birdlife International, 2010). Subsequently Friesen et al. (2006) went further and analysed the population 46 structure of P. phaeopygia using microsatellite data and found a strong genetic structure between population breeding on different islands within the archipelago. Based on these results they argued that each island population should be treated as a management unit (MU) for the long-term conservation of the species and its genetic diversity. Figure 1.9. Phylogenetic tree from Sibley & Ahlquist (1990) showing relationships within Order Procellariiformes (as in Kennedy & Page, 2002) 47 III Gadfly petrels Family Procellariidae, which contains the most species and genera in Procellariiformes, is the only one in the order that comprises taxa listed in IUCN Red List that were driven to extinction by anthropogenic factors (Table 5) and it now presents several Critically Endangered species. Within the family, gadfly petrels, including members of genera Pterodroma, Lugensa, Bulweria and Pseudobulweria, is the largest group with 39 species Brooke (2004), 32 in Pterodroma, 1 in Lugensa, 2 in Bulweria and 4 in Pseudobulweria. Most gadfly petrel colouration are made of grey or black/brown on the upperparts, with some white on the head and underparts, some dark taxa showing grey or black underparts (Warham, 1990; Brooke, 2004). Sub-Antarctic and temperate taxa tend to be nocturnal on land and breed in burrows, while tropical species are more easily seen active above breeding grounds by day, with activity peaks continuing for a while in the dark. While diurnal species more readily breed on the surface than their nocturnal counterparts, they usually nest under bushes, rocks or tree roots (Brooke, 2004, pers. obs.). Because of the nocturnal and fossorial habits of many taxa in the group, it is believed that their communication system (and presumably behavioural pre-isolating mechanism: Bretagnolle, 1996) almost never rely on visual cues, hence promoting cryptic and highly convergent plumages in many taxa. Several studies have investigated the role that vocalisations can play on population differentiation in these nocturnal species and used this character for systematics studies. For instance, Bretagnolle (1990) used the calls of males and females together with other behavioural data to confirm the validity of genus Halobaena (family Procellariidae) and its relationship with Pachyptila. More recently, McKown (2008) investigated vocalisations within sub-genus cookilarias (genus Pterodroma) to infer phylogenetic relationships between these closely related taxa. In terms of taxonomy and phylogenetic relationships, gadfly petrels present uncertainties and issues at all levels (Brooke, 2004), from genus level, e.g. long-standing debates regarding the validity of genus Pseudobulweria or the validity of sub-genus cookilaria and the relationships between its constituting taxa (Warham, 1990; McKown, 2008) to the species/subspecies level. At this level, several complexes of taxa have long puzzled biologists and taxonomists. For instance, the Pterodroma arminjoniana/neglecta/heraldica complex living in three ocean basins (Pacific, Indian and Atlantic) and whose composition in terms of species has long been debated and still not resolved due notably to recent colonisation of new islands in the Indian Ocean and to apparent 48 hybridisation (Brown, 2010). Another example consists in the Pterodroma mollis complex: this complex whose taxa are distributed from sub-antarctic to tropical latitudes in three oceans (Atlantic, Indian and Austral), has over the years been described as made of one species and six or four sub-species or three species with or without subspecies (review in Bretagnolle, 1995). The latter finally considered the complex as made of two species, Pterodroma mollis itself in the southern Atlantic, Indian and Austral oceans, and Pterodroma feae in Eastern Atlantic, breeding in Cape Verde, Madeira and Desertas archipelagos. However, the latter taxon is now proving incredibly complex and its taxonomic status is not yet settled. In addition, Gadfly petrels, present a striking characteristic of endemicity. Over a quarter of recognised species are endemic to one archipelago and most of the time of one island in the archipelago (Brooke, 2004)! Only a handful of species are supposed to have recently colonised new breeding grounds, such as the Trindade petrel (Pterodroma arminjoniana) supposed to have recently colonised Round Island in the Indian Ocean from its native Trindade archipelago off the coast of Brasil (Brown et al., 2010), or the Black-winged petrel (P. nigripennis) that have colonised a few archipelagos in the New Zealand/ Australia region and Indian Ocean (Warham, 1990; Thiebot et al., 2010). Furthermore, Steadman (1995) describes a very rich community of Procellariidae in all archipelagos of the Pacific, of which certain forms have been extirpated or driven to extinction. Thus the actual observed pattern of distribution and diversity of Procellariidae is biased and it must be kept in mind that most likely, number of endemic species was much higher prior to the human colonisation of the Pacific Ocean archipelagos and that the actual repartition of species does not reflect the past distribution of many of them. That birds presenting such dispersal capabilities have, during the course of evolution, produced so many forms endemic to one or a few islands raises questions on the underlying causes, processes and patterns of population differentiation in the group rendering further phylogenetic and phylogeographic investigations of this group necessary. 49 THESIS OBJECTIVES Procellariiformes are at a cross point between (i) extensive conservation needs, (ii) taxonomic uncertainty, and (iii) conflicting processes regarding population differentiation and lineage divergence. Understanding the third point can greatly help tackling the first two. Genetics and molecular phylogenetic offer tools that allow dealing with these issues. The aim of this work is to describe phylogenetic relationships and taxonomy, as well as describe population histories and timing of divergence in threatened gadfly petrels. I used techniques related to population genetics, phylogeography and molecular phylogenetics to reach this objective. New methods developed now allow fields that diverged due to methodological and conceptual differences (Hey & Machado, 2003; Brito & Edwards, 2008) like phylogenetics and phylogeography to become more integrated (Edwards, 2009). Although we have seen through the diverse examples cited that already many studies have been conducted in the fields of taxonomy and phylogeography in Procellariiformes and gadfly petrels, the array of investigations needed in this group is so large that it ensures no redundancy between these studies and the one undertaken in this PhD. Furthermore, the use of new methodologies allows revisiting with a new perspective some issues not yet resolved in some taxa. The investigations I undertook in phylogenetic and phylogeographic relationships within gadfly petrels concerned two taxonomic levels: 1- the genus level: following the rediscovery of the Critically Endangered Beck’s petrel, the opportunity arose to obtain for the first time samples from all extant taxa of genus Pseudobulweria. This genus has for a long time been controversial and the phylogenetic relationships between its taxa had never been investigated. It is one of the least known seabird genus and probably one of the most endangered one, being composed of three Critically Endangered taxa and one Near Threatened species. In addition, two extinct taxa have been described as belonging to Pseudobulweria, one in the Atlantic Ocean and one in the Pacific Ocean. When confronting this pattern of vulnerability and extinction in this genus, the question arises immediately about the underlying reasons that lead taxa in this genus to be so much threatened or become extinct. Understanding these underlying processes requires as the first step the description of the phylogenetic relationships within the genus and between the genus and the other genera in the family/order. From these further investigations can then be undertaken. 50 2- at the species/population level: (i) first I explored the relationships and population histories of three taxa breeding in Eastern Atlantic in three archipelagos, Madeira, Desertas and Cape Verde. These taxa have been the topic of much debate and recent publications regarding their taxonomic statuses, i.e. how many species do they represent, one, two or three? In this study, I aimed at describing the population history of the three taxa and determining if they are currently exchanging genes. I also attempted to propose a phylogeographic scenario describing the history of these taxa in this region. (ii) second I investigated the relationships between two taxa living in Eastern Australia and New Caledonia and considered two subspecies of the Gould’s petrel, Pterodroma leucoptera. This taxon is classified Vulnerable, the Australian subspecies is limited to a few hundreds breeding pairs on two small islands, while the Caledonian subspecies restricted to remote mountain ranges has an unknown number of breeding pairs, but it is supposed to probably be over several thousands. The phylogenetic relationships and potential gene flow between these two taxa remain unstudied but could have important conservation implications. Contrary to the petrels of Eastern Atlantic, the relationships and statuses of the two Gould’s petrel populations seem more settled and based on sound evidences. However, we might see that this not necessarily the case. These taxa are part of a complex including another species (Collared petrel, Pterodroma brevipes) and several potential undescribed taxa spread in various archipelagos in the Pacific Ocean whose phylogenetic relationships are unclear. These two cases thus represent two facets of taxonomy in gadfly petrels - in one case, the taxonomy is debated and unsettled despite many investigations in the last three decades, while in the other case there does not seem to be any taxonomic problem-, and were chosen to illustrate the need of thorough investigations in the group to uncover underlying patterns and processes. The thesis manuscript is based on papers submitted or in preparation for submission in scientific journals and is composed of five chapters in addition to the introduction: - Chapter 2 is a brief introduction to the techniques and tools used during this PhD; - Chapter 3 describes the phylogenetic relationships within genus Pseudobulweria and the position of this genus relative to other Procellariiformes; this manuscript has been submitted to the journal Conservation Genetics; 51 - Chapter 4 investigates phylogenetic and phylogeographic relationships within three populations of gadfly petrel of Eastern Atlantic Ocean; this manuscript is in preparation for submission; - Chapter 5 is divided in two parts. Part 1 describes the population history and phylogeography of Gould’s petrel subspecies in Australia and New Caledonia. Part 2 is a preliminary investigation of phylogenetic relationships in the Gould’s/Collared petrel complex by means of the mitochondrial gene used in birds for the Barcode of Life project, the cytochrome oxidase 1 (CO1) and based on fragments of CO1 obtained from museum specimens; these two parts are made of manuscripts in preparation for submission; - Chapter 6 finally is a summary of the main findings and discussion of the latter in a broader context and presents some perspectives to the work undertaken during this PhD. 52 References Abbott C. and Double M., 2003. Phylogeography of Shy and White-capped Albatrosses inferred from mitochondrial DNA sequences: implications for population history and taxonomy. Molecular Ecology 12: 2747–2758 Ainley D.G. and Divoky, G.J., 2001. Seabirds: effects of climate change. In: Encyclopedia of Ocean Sciences. Steele J., Thorpe S. and Tarekian K. eds. Academic Press, London UK Ainley D.G., Clarke E.D., Arrigo K., Fraser W.R., Kato A., Barton K.J. and Wilson P.R., 2005. Decadal-scale changes in the climate and biota of the Pacific sector of the Southern Ocean, 1950s to the 1990s. Antarctic Science 17: 171–182 Akaike H., 1973. Information theory and an extension of the maximum likelihood principle. In Second International Symposium on Information Theory, ed. PN Petrov, F Csaki. pp. 267– 81. Budapest: Akad. Kiado Austin J.J., Bretagnolle V. and Pasquet E., 2004. Aglobal molecular phylogeny of the small Puffinus shearwaters and implications for systematics of the Little-Audubon’s shearwater complex. The Auk 121(3): 847-864 Avise JC, Arnold J, Ball RM et al (1987) Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecol and Systematics 18:489–522 Avise J.C., 2000. Phylogeography. The history and formation of species. Harvard University Press, Cambridge, Massachusetts USA Baker R. J. & Bradley R.D., 2006. Speciation in mammals and the genetic species concept. Journal of mammalogy 87(4): 643-662 Barton N. H., 2001. Speciation. Trends in Ecology and Evolution 16(7): 325 BirdLife International, 2009. http://www.birdlife.org accessed on 10/08/2009 BirdLife International, 2010. http://www.birdlife.org accessed on 10/10/2010 Blaxter M. L., 2004. The promise of a DNA taxonomy. Philosophical Transactions of the Royal Society B: Biological Sciences 359, 669-679 Bolton M., Smith A.L., Gomez-Diaz E., Friesen V.L., Medeiros R., Bried J., Roscales J.L. and Furness R.W., 2008. Monteiro’s storm petrel Oceanodroma monteiroi: a new species from the Azores. Ibis 150: 717-727 Bos D.H. and Posada D., 2005. Using models of nucleotide evolution to build phylogenetic trees. Developmental and comparative Immunology 29: 211-227 53 Boughman J.W., 2001. Divergent sexual selection enhances reproductive isolation in sticklebacks. Nature 411: 944–948 Bretagnolle V., 1990. Behavioural affinities of the Blue petrel Halobaena caerulea. Ibis 132: 102105 Bretagnolle V., 1993. Adaptive significance of seabird colouration: the case of Procellariiforms. The American Naturalist 142(1): 141-173 Bretagnolle V., 1995. Systematics.of the Soft-plumaged Petrel Pterodroma mollis complex: new insight from vocalizations. Ibis 137: 207–218 Bretagnolle, V., 1996. Acoustic communication in a group of non passerines birds, the petrels. In Ecology and Evolution of Acoustic Communication in Birds, Kroodsma D.E. and Miller E.H. eds. Cornell University Press, Ithaca, New York, USA Bretagnolle V., Attié C. and Pasquet E., 1998. Cytochrome b evidence for validity and phylogenetic relationships of Pseudobulweria and Bulweria. The Auk 115(1): 188-195 Brito P. & Edwards S., 2009. Multilocus phylogeography and phylogenetics using sequence-based markers. Genetica 135: 439-455 Brooke M., 2002. Seabird systematics and distribution: a review of current knowledge. In Biology of Marine Birds. Schreiber E.A. & Burger J. eds., CRC Press, Boca raton, Florida USA Brooke M., 2004. Albatrosses and Petrels Across the World. Oxford University Press, New York Brown R.M. et al., 2010. Range expansion and hybridization in Round Island petrels (Pterodroma spp.): evidence from microsatellite genotypes. Molecular Ecology 19: 3157-3170 Browne R.A., Anderson D.J., Houser J.N., Cruz F., Glasgow K.J., Hodges C.N. and Massey G., 1997. Genetic diversity and divergence of endangered Galapagos and Hawaiian petrel populations. Condor 99:812–815 Burg T. and Croxall J.P., 2004. Global population structure and taxonomy of the Wandering Albatross species complex. Molecular Ecology 13: 2345–2355 Burger J. & Gochfeld M., 2002. Effects of chemical and pollution on seabirds. In Biology of Marine Birds. Schreiber E.A. & Burger J. eds., CRC Press, Boca raton, Florida USA Cardoso A., Serrano A. and Vogler A. P., 2009. Morphological and molecular variation in tiger beetles of the Cicindela hybrida complex: is an 'integrative taxonomy' possible? Molecular Ecology 18: 648-664 Carstens BC, Knowles LL (2007) Estimating species phylogeny from gene tree probabilities despite incomplete lineage sorting: an example from Melanoplus grasshoppers. Systematic Biology 56:400–411 54 Cavalli-Sforza L.L., 1964. Population structure and human evolution. Proceedings of the Royal Society of London B 164: 362–379. Cincotta R. P., Wisnewski J. & Engelman R., 2000. Human population in the biodiversity hotspots. Nature 404: 990-992. Congdon B.C., Piatt J.F., Martin K. and Friesen V.L., 2000. Mechanisms of population differentiation in marbled murrelets: historical versus contemporary processes. Evolution 54(3): 974-986 Conn H.W., 1907. Evolution of Today: a Summary of the Theory of Evolution as Held by Scientists at the Present Time, and an account of the Progress Made by the Discussions and Investigations of a Quarter of a Century. G.P. Putnam’s Sons, New York & London Coyne J.A. and Orr H.A., 2004. Speciation. Sinauer associates, Sunderland, MA, USA Cuthbert R., Phillips R. & Ryan P., 2003. Separating Tristan albatrosses and wandering albatrosses using morphometric measurements. Waterbirds 26: 338–344 Dawkins R., 1976. The selfish gene. Oxford University Press, Oxford, UK Darwin C., 1858. On the tendency of species to form varieties; and on the perpetuation of varieties and species by natural means of selection. I.Extract from an unpublished work on species, II. Abstract of a letter from C. Darwin, esq., to Prof. Asa Gray. Journal of the Proceedings of the Linnean Society London 3: 45–53. Darwin C., 1859. On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. John Murray, London. Darwin C., 1871. The descent of man, and selection in relation to sex. John Murray, London Daugherty C.H., Cree A., Hay J.M. and Thompson M.B., 1990. Neglected taxonomy and continuing extinctions of Tuatara (Sphenodon). Nature 347: 177-179 Dayrat B., 2005. Towards integrative taxonomy. Biological Journal of the Linnean Society 85: 407–15 DeeBoersma P., clark J.A. and Hillgarth N., 2002. Seabird Conservation. In Biology of Marine Birds. Schreiber E.A. & Burger J. eds., CRC Press, Boca raton, Florida USA Degnan J. H. and Rosenberg N. A,. 2006. Discordance of species trees with their most likely gene trees. PLoS Genetics 2: 762-768 Degnan J.H. & Rosenberg N.A., 2009. Gene tree discordance, phylogenetic inference and the multispecies coalescent. Trends in Ecology and Evolution 24(6): 332–340. Delsuc F., Brinkmann H. and Philippe H., 2005. Phylogenomics and the reconstruction of the Tree of Life. Nature Reviews Genetics 6(5):361-375 deQueiroz K., 1998.The general lineage concept of species, species criteria, and the process of speciation: A conceptual unification and terminological recommendations. Pages57–75 in 55 Endless forms: Species and speciation (D.J. Howard, and S.H. Berlocher, eds.). Oxford University Press, NewYork. deQueiroz K., 2005. Ernst Mayr and the modern concept of species. Proceedings of the National Academy of Sciences of the USA 102: 6600–6607 deQueiroz K., 2007. Species concepts and species delimitations. Systematic Biology 56(6): 879-886 Dobzhansky T., 1937. Genetics and the origin of species. Columbia University Press, New York Dobzhansky T., 1973. Nothing in biology makes sense except in the light of evolution. American Biological Teacher 35: 125-129 Donlan C.J., 2008. On the ecology of invasive species, extinction, ecological history, and biodiversity conservation. PhD thesis. Cornell University Dumont Y., Russell J.C., Lecomte V. and LeCorre M., 2010. Conservation of endangered endemic seabirds within a multi-predator context: the Barau’s petrel in Réunion Island. Natural Resource Modeling 23(3): 381-436 Edwards S.V., 2009. Is a new and general theory of molecular systematics emerging? Evolution 63(1): 1-19 Efron B., 1979. Bootstrap methods: another look at the jackknife. Annals of Statistics 7: 1–26 Ehrlich P.R. & Wilson E.O., 1991. Biodiversity studies: science and policy. Science 253: 758-762 Faith D.P., 1992. Conservation evaluation and phylogenetic diversity. Biological Conservation 61: 1–10. Faith D.P., 1994. Phylogenetic pattern and the quantification of organismal biodiversity. Philosophical Transactions of the Royal Society, Series B 345: 45–58. Felsenstein J., 1978. Cases in which parsimony or compatibility methods will be positively misleading. Systematic Zoology 27: 401-410 Felsenstein J., 1981. Evolutionary trees from DNA sequences: a maximum likelihood approach. Journal of Molecular Evolution 17: 368–76 Felsenstein J., 1985. Confidence intervals on phylogenies: an approach using the bootstrap. Evolution 39: 783-791 Felsenstein J., 1988. Phylogenies from molecular sequences: inference and reliability. Annual Review of Genetics 22: 521–565 Féral J-P., 2002. How useful are the genetic markers in attempts to understand and manage marine biodiversity? Journal of Experimental Marine Biology and Ecology 268: 121-145 Fitzpatrick B.M., Fordyce J.A. and Gavrilets S., 2009. Pattern, process and geographic modes of speciation. Journal of Evolutionary Biology 22(11): 2342-2347 56 Forister M.L., Nice C.C., Fordyce J.A., Gompert Z. and Shapiro A.M., 2008.Considering evolutionary processes in the use of single-locus genetic data for conservation , with examples from Lepidoptera. Journal of Insect Conservation 12: 37-51 Friesen V.L. & Anderson D.J., 1997. Phylogeny and evolution of the sulidae (Aves: Pelecaniformes): a test of alternative modes of speciation. Molecular Phylogenetics and Evolution 7: 252–260 Friesen V.L., Anderson D.J., Steeves T.E., Jones H. and Schreiber E.A., 2002. Molecular support for species status of the Nazca booby (Sula granti). The Auk 119(3): 820-826 Friesen V.L., Burg T.M. and McCoy K.D., 2007a. Mechanisms of population differentiation in seabirds. Molecular Ecology 16: 1765-1785 Friesen V.L., Gonzalez J.A. and Cruz-Delgado F., 2006. Population genetic structure and conservation of the Galapagos petrel (Pterodroma phaeopygia). Conservation Genetics 7: 107-115 Friesen V.L., Montevecchi W.A., Baker A.J., Barrett R.T. and Davidson W.S., 1996. Population differentiation and evolution in the common guillemot Uria aalge. Molecular Ecology 5: 793–805 Friesen V.L., Smith A.L., Gomez-Diaz E., Bolton M., Furness R.W., Gonzales-Solis J. and Monteiro L.R., 2007b. Sympatric speciation by allochrony in a seabird. Proceedings of the National Academy of Sciences of the USA 104(47): 18589-18594 Gaston A.J., 2004. Seabirds: A Natural History. Yale University Press, New Haven, Connecticut Gavrilets S., 2000. Rapid evolution of reproductive barriers driven by sexual conflict. Nature 403: 886-889 Gavrilets S. and Hastings A., 1996. Founder effect speciation: a theoretical reassessment. The American Naturalist 147: 466–491 Goldman N., Anderson, J.P. and Rodrigo A.G., 2000. Likelihood-based tests of topologies in phylogenetics. Systematic Biology 49: 652-670 Gomez-Diaz E., Gonzales-Solis J., Peinado M.A. and Page R.D.M., 2006. phylogeography of the Calonectris shearwaters using molecular and morphometric data. Molecular Phylogenetics and Evolution 41: 322-332 Grant V., 1981. Plant Speciation. Columbia University Press, NewYork USA Grémillet D. & Boulinier T., 2009. Spatial ecology and conservation of seabirds facing global climate change: a review. Marine Ecology Progress Series 391: 121-137 Grosbois V. & Thompson P.M., 2005. North Atlantic climate variation influences survival in adult fulmars. Oikos 109: 273-290 Haig S.M., 1998. Molecular contributions to conservation. Ecology 79(2): 413-425 57 Hasegawa M, Kishino H, Yano T., 1985. Dating the human–ape split by a molecular clock of mitochondrial DNA. Journal of Molecular Evolution 22: 160–74. Hastings W.K., 1970. Monte carlo sampling methods using Markov chains and their applications. Biometrika 57: 97-109 Hayden T., 2009. What Darwin didn’t know. Smithsonian Magazine February 2009 Hedrick P.W., 2001. Conservation genetics: where are we now? Trends In Ecology and Evolution 16(11): 629-636 Hey J., 2001. The mind of the species problem. Trends in Ecology and Evolution 16(7): 326-329 Hey J. & Machado C.A., 2003. The study of structured populations—new hope for a difficult and divided science. Nature Reviews Genetics 4: 535–543 Hillis D.M. and Bull J.J, 1993. An empirical test of bootstrapping as a methods for assessing confidence in phylogenetic analysis. Systematic Biology 42: 182–192 Hillis D.M., 1996. Inferring complex phylogenies. Nature 383: 130-131 Ho C.Y.-K., Prager E.M., Wilson A.C., Osuga D.T. and Feeney R.E., 1976. Penguin evolution: protein comparisons demonstrate phylogenetic relationships to flying aquatic birds. Journal of Molecular Evolution 8: 271–282 Holdaway R.N. & Worthy T.H., 1994. A new fossil species of shearwater Puffinus from the Late Quaternary of the South Island, New Zealand, and notes on the Biogeography and Evolution of the Puffinus gavia superspecies. Emu 94: 201–215 Holder M. and Lewis P.O., 2003. Phylogeny estimation: Traditional and Bayesian approaches. Nature Reviews Genetics 4:275-284 Huelsenbeck J.P., Rannala B., 2004. Frequentist properties of Bayesian posterior probabilities of phylogenetic trees under simple and complex substitution models. Systematic Biology 904– 13 Huelsenbeck J.P., Ronquist F., Nielsen R., Bollback J.P., 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294: 2310–2314. Huelsenbeck J.P., Larget B., Miller R.E., Ronquist F., 2002. Potential applications and pitfalls of Bayesian inference of phylogeny. Systematic Biology 51: 673–688. Huxley J. S., 1942. Evolution: the modern synthesis. Allen and Unwin, London Jenouvrier S., Barbraud C. and Weimerskirch H., 2005. Long-term contrasted responses to climate of two Antarctic seabird species. Ecology 86(11): 2889-2903 Jouventin P. & Dobson F.S., 2002. Why breed every other year? The case of albatrosses. Proceedings of the Royal Society of London. Series B, Biological Sciences 269: 1955–1961 Jukes TH, Cantor CR., 1969. Evolution of protein molecules. In: Munro HN, ed. Mammalian protein metabolism. Academic Press, New York, USA p.21–132 58 Kennedy M. & Page R.D.M., 2002. Seabird supertrees: combining partial estimates of Procellariiform phylogeny. The Auk 119(1): 88-108 Kim J.H. General inconsistency conditions for maximum parsimony: effects of branch lengths and increasing numbers of taxa. Systematic Biology 45: 363-374 Kimura M., 1980. A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16: 111–20. Kirkpatrick M. & Ravigné V., 2002. Speciation by natural and sexual selection: models and experiments. The American Naturalist 159: S22–S35 Koonin E.V., 2009. Darwinian evolution in the light of genomics. Nucleic Acids Research 37(4): 1011–1034 Koonin E.V. & Wolf Y.I., 2009a. Is evolution Darwinian and/or Lamarckian? Biology Direct 4: 42 Koonin E.V. & Wolf Y.I., 2009b. The fundamental units, processes and patterns of evolution, and the Tree of Life conundrum. Biology Direct 4: 33 Ksepka D.T., Bertelli S. and Giannini N.P., 2006. The phylogeny of the living and fossil Sphenisciformes. Cladistics 22: 412-441 Kuroda N., Kakizawa R. and Watada M., 1990. Genetic divergence and relationships in fifteen species of procellariiformes. Journal of the Yamashina Institute of Ornithology 22: 114–123 Kuro’o M., Yonekawa H., Saito S, Eda M., Higushi H., Koike H. and Hasegawa H., 2010. Unexpectedly high genetic diversity of mtDNA control region through severe bottleneck in vulnerable albatross Phoebastria albatrus. Conservation Genetics 11: 127-137 Kutschera U. & Niklas K.J., 2004. The modern theory of biological evolution: an expanded synthesis. Naturwissenschaften 91: 255-276 Lamarck J-B., 1809. Philosophie zoologique, ou exposition des considérations relatives à l'histoire naturelle des animaux. Paris: Dentu Lande R., 1981. Models of speciation by sexual selection on polygenic traits. Proceedings of the National Academy of Sciences of the USA 78: 3721–3725 Le Corre M., Ghestemme T., Salamolard M. and Couzi F-X., 2003. Rescue of the Mascarene petrel, a critically endangered seabird of Reunion Island, Indian Ocean. The Condor 105: 387-391 Lorion J., 2008. Diversité et évolution des Mytilidae (Mollusca : Bivalvia) associés aux substrats organiques coulés. Thèse de doctorat de l’Université Pierre & Marie Curie, Paris, France Maddison W.P., 1997. Gene trees in species trees. Systematic Biology 46: 523–536 Mallet J., 1995. A species definition for the modern synthesis. Trends in Ecology and Evolution 10: 294-299 Mallet J., 2004. Poulton, Wallace and Jordan: how discoveries in Papilio butterflies led to a new species concept 100 years ago. Systematics and Biodiversity 1(4): 441-452 59 McKown M.W., 2008. Acoustic communication in colonial seabirds: individual, sexual, and species-specific variations in acoustic signals of Pterodroma petrels. PhD thesis University of North Carolina Metropolis N., Rosenbluth A.W., Teller A.H. and Teller E., 1953. Equation of state calculations by fast computing machines. The Journal of Chemical Physics 21(6): 1087-1091 Miller K.G. & Fairbanks R.G., 1983. Evidence for Oligocene−Middle Miocene abyssal circulation changes in the western North Atlantic. Nature 306(5940): 250–253 Moritz C.C., 1994. Applications of mitochondrial DNA analysis in conservation: a critical review. Molecular Ecology, 3:401-411 Moritz C.C., 1995. Uses of molecular phylogenies for conservation. Philosophical transactions of the Royal Society of London B 349: 113-118 Moritz C., 2002. Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51(2): 238-254 Mayden R.L.,1997. A hierarchy of species concepts: The denouement in the saga of the species problem. Pages 381–424 in Species: The units of biodiversity (M.F. Claridge, H.A. Dawah, and M.R. Wilson, eds.) Chapman and Hall, London, UK Mayr E., 1942. Systematics and the origin of species. Columbia University Press, New York Mayr E., 1963. Animal species and evolution. Harvard University Press, Cambridge, Massachusetts USA Mayr E., 1982. The growth of biological thought: Diversity, evolution, and inheritance. Harvard University Press, Cambridge, Massachusetts, USA. Mayr G., 2009. Notes on the osteology and phylogenetic affinities of the Oligocene Diomedeoididae (Aves, Procellariiformes). Fossil Record 12(2): 133-140 Mayr G., Clarke J., 2003. The deep divergences of neornithine birds: a phylogenetic analysis of morphological characters. Cladistics 19: 527–533 Mayr E. & Provine W., 1980. The evolutionary synthesis: perspectives on the unification of biology. Harvard University Press, Cambridge, Massachusetts, USA Mey E., Chastel O. and Beaucournu J.C., 2002. A ‘penguin’ chewing louse Nesiotinus on a Kergelen diving-petrel (Pelecanoides urinatrix exsul): an indication of a phylogenetic relationship? Journal of Ornithology 143: 472–476 Monteiro L.R. & Furness R.W., 1998. Speciation through temporal segregation of Madeiran Storm Petrel (Oceanodroma castro) populations in the Azores? Philosophical Transactions of the Royal Society London 353: 945–953 Moritz C., 2002. Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51(2): 238-254 60 Myers N., Mittermeier R. A., Mittermeier C. G., da Fonseca G. A. B. & Kent J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853-858 Nei M., 1987. Molecular evolutionary genetics. Columbia University Press, New York., USA Nei M., Maruyama T. and Wu C.I., 1983. Models of evolution of reproductive isolation. Genetics 103: 557–579 Nichols N., 2001. Gene trees and species trees are not the same. Trends in Ecology and Evolution 16: 358-364 Nosil P., Harmon L.J. and Seehausen O., 2009. Ecological explanations for (incomplete) speciation. Trends in Ecology and Evolution 24: 145–156 Nunn G.B., Cooper J., Jouventin P., Robertson C.J.R and Robertson G.G., 1996. Evolutionary relationships among extant albatrosses (Procelleriiformes: Diomedeidae) established from complete cytochrome b gene sequences. The Auk 113(4): 784-801 Nunn G.B. & Stanley S.E., 1998. Body size effects and rates of cytochrome b evolution in tubenosed seabirds. Molecular Biology and Evolution 15: 1360–1371 Olson S.L., 1975. Remarks on the generic characters of Bulweria. Ibis 117: 111-113 Olson S.L., 2004. taxonomic review of the fossil Procellariidae (Aves: Procellariiformes) described from Bermuda by R.W. Shufeldt. Proceedings of the Biological Society of Washington 117(4): 575-581 Orr M.R. & Smith T.B., 1998. Ecology and speciation. Trends in Ecology and Evolution 13: 502– 506 Padial JM, Castroviejo-Fisher S, Köhler J, Vilà C, Chaparro JC, De la Riva I., 2010. Deciphering the products of evolution at the species level: the need for an integrative taxonomy. Zoologica Scripta 38:431-447 Page R.D.M. & Holmes E.C., 1998. Molecular Evolution, A Phylogenetic Approach. Blackwell Science, Oxford, UK Pamilo P. & Nei M., 1988. Relationships between gene trees and species trees. Molecular Biology and Evolution 5(5):568–583 Penhallurick J. & Wink M., 2004. Analysis of the taxonomy and nomenclature of the Procellariiformes based on complete nucleotide sequences of the mitochondrial cytochrome b gene. Emu 104: 125-147 Peterson A.T. & Navarro-Siguenza A.G., 1999 Alternative species concepts as bases for determining priority conservation areas. Conservation Biology 13:427–431 Phillimore A.B., Orme C.D.L., Thomas G.H., Blackburn T.M., Bennett P.M., Gaston K.J. and Owens I.P.F., 2008. Sympatric Speciation in Birds Is Rare: Insights from Range Data and Simulations. The American Naturalist 171(5): 646-657 61 Poulton E.B.,1904.What is a species? Proceedings of the Entomological Society of London 1903, lxxvii–cxvi Price T.D., 2008. Speciation in Birds. Roberts & Company Publishers, Greenwood village, Colorado, USA Purvis A., Agapow P-M., Gittleman J.L. and Mace G.M., 2000. Nonrandom extinction and the loss of evolutionary history. Science 288: 328–330 Purvis A., Cardillo M., Grenyer R. and Collen B., 2005. Correlates of extinction risk: phylogeny, biology, threat and scale. In: Phylogeny and conservation. Purvis A., Gittleman J.L. and Brooks T.M. eds., Cambridge University Press, Cambridge UK Ragan M.A., McInerney J.O. and Lake J.A., 2009. The network of life: genome beginnings and evolution. Philosophical Transactions of the Royal Society B 364: 2169-2175 Rando J.C. & AlcoverJ.A., 2008. Evidence for a second Plearctic seabird extinction during the last millennium: the Lava shearwater Puffinus olsoni. Ibis 150: 188-192 Rannala B. & YangZ., 1996. Probability distribution of molecular evolutionary trees: a new method of phylogenetic inference. Journal of Molecular Evolution 43: 304-311 Rheindt F.E. & Austin J.J., 2005. Major analytical and conceptual shortcomings in a recent taxonomic revision of the Procellariiformes – a reply to Penhallurick and Wink (2004). Emu 105: 181-186 Rensch B., 1947. Neuere Probleme der Abstammungslehre. Die transspezifische Evolution. Enke, Stuttgart, Germany Ritchie M.G., 2007. Sexual selection and speciation. Annual Review of Ecology, Evolution and Systematics.38: 79–102 Robertson C.J.R., 1998. Factors influencing the breeding performance of the Northern Royal Albatross. In Albatross biology and conservation. Robertson G. & gates R. eds. Surrey Beatty & Sons, chipping Norton Robertson C. & Nunn G., 1998. Towards a new taxonomy for albatrosses. In: Albatross Biology and Conservation (eds Robertson G, Gales R), pp. 13–19. Surrey Beatty, Chipping Norton Rodrigues A.S.L., Brooks T.M. and Gaston K.J., 2005. Integrating phylogenetic diversity in the selection of priority areas for conservation: does it make a difference? In Phylogeny and Conservation. Purvis A., Gittleman J.L. and Brooks T.M. eds., Cambridge University Press, Cambridge UK Rodriguez F, Oliver J. F., Marin A., Medina J.R., 1990. The general stochastic model of nucleotide substitution. Journal of Theoretical Biology 142: 485–501 Rubinoff D., 2006. Utility of mitochonodrial DNA barcodes in species conservation. Conservation Biology 20(4): 1026-1033 62 Rundle H.D. & Nosil P., 2005. Ecological speciation. Ecology Letters 8: 336–352 Rundle H.D. & Schluter, D., 2004. Natural selection and ecological speciation in sticklebacks. In: AdaptiveSpeciation, Dieckmann U., Doebeli M., Metz J.A.J. and Tautz D. eds. Cambridge University Press, Cambridge, UK Ryder O.A., 1986. Species conservation and the dilemma of subspecies. Trends in Ecology and Evolution, 1: 09-10 Sagar P.M. & Warham J., 1993. A long-lived southern Buller’s mollymawk Diomedea bulleri bulleri with a small egg. Notornis 40: 303–304. Saiff E.I., 1976. Anatomy of the middle ear region of the avian skull: Sphenisciformes. The Auk 93: 749–759 Saitou N. and Nei M., 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4: 406–425 Samadi S. & Barberousse A., 2006. The tree, the network, and the species. Biological Journal of the Linnean Society 89: 509-521 Samadi S. & Barberousse A., 2009. Species: towards new, well-grounded practices. Biological Journal of the Linnean Society 97: 217-222 Savolainen V., Cowan R. S., Vogler A. P., Roderick G. K. & Lane R., 2005. Towards writing the encyclopaedia of life: an introduction to DNA barcoding. Philosophical Transactions of the Royal Society B: Biological Sciences 360: 1805-1811 Schlick-Steiner BC, Steiner FM, Seifert B, Stauffer C, Christian E, Crozier RH: Integrative Taxonomy: a multisource approach to exploring Biodiversity. Annu Rev Entomol 55:421438 Schluter D., 2001. Ecology and the origin of species. Trends in Ecology and Evolution 16(7): 372380 Schluter D., 1998. Ecological causes of speciation. In Endless Forms: Species and Speciation Howard, D. & Berlocher, S., eds, Oxford University Press, UK Schluter D., 2009. Evidence for ecological speciation and its alternative. Science 323:737–741 Schreiber E.A. & Burger J., 2002. Biology of Marine Birds. CRC Press, Boca Raton, Florida USA Schwarz G., 1978. Estimating the dimensions of a model. The Annals of Statistics 6: 461–64 Sechrest W., Brooks T.M., daFonseca G.A.B et al., 2002. Hotspots and the conservation of evolutionary history. Proceedings of the National Academy of Sciences of the USA 99(4): 2067-2071 Seehausen O., Mayhew P.J. and VanAlphen J.J.M., 1999. Evolution of colour patterns in East African cichlid fish. Journal of Evolutionary Biology 12: 514–534 63 Shevenell A.E., Kennett J.P. and Lea D.W., 2004. Middle Miocene Southern Ocean Cooling and Antarctic Cryosphere Expansion. Science 305: 1766-1770 Shirihai H., 2008. Rediscovery of Beck’s petrel Pseudobulweria becki, and other observations of tubenoses from the Bismarck archipelago, Papua New Guinea. Bulletin of the British Ornithologist Club 128:3–16 Sibley C.G. & Ahlquist J.E., 1990. Phylogeny and classification of birds – a study in molecular evolution. Yale University Press, New Haven USA Simpson G. G., 1944. Tempo and mode in evolution. Columbia University Press, New York Simpson G. G., 1951. The species concept. Evolution 5: 285-298 Simpson G.G., 1953. The major features of evolution. Columbia University Press, NewYork USA Sites J.W. Jr. & Crandall K.A., 1997. Testing species boundaries in biodiversity studies. Conservation Biology 11: 1289–1297 Sites J.W. & Marshall J.C., 2004. Operational criteria for delimiting species. Ann Rev Ecol Evol Syst 35: 199–227 Smith A.L. and Friesen V.L., 2007. Differentiation of sympatric populations of the bandrumped storm petrel in the Galapagos Islands: an examination of genetics, morphology and vocalizations. Molecular Ecology 16: 1593-1603 Sobel J.M., Chen G.F., Watt L.R. and Schemske D.W., 2010. The biology of speciation. Evolution 64(2): 295-315 Steadman D.W., 1995. Prehistoric extinctions of Pacific island birds: biodiversity meets zooarchaeology. Science 267: 1123-1131 Steadman D.W., 2006. Extinction and Biogeography of tropical Pacific birds. University of Chicago Press, London UK Stebbins G. L., 1950. Variation and evolution in plants. Columbia University Press, New York Steeves T.E., Anderson D.J. and Friesen V.L., 2005. A role for non physical barriers to gene flow in the diversification of a very vagile seabird, the masked booby (Sula dactylatra). Molecular Ecology 14: 3877-3887 Steiner C.C., Rompler H., Boettger L.M., Schoneberg T. and Hoekstra H.E., 2009. The genetic basis of phenotypic convergence in beach mice: similar pigment patterns but different genes. Molecular Biology and Evolution 26: 35–45 Stockman A.K. & Bond J.E., 2007. Delimiting cohesion species: extreme population structuring and the role of ecological interchangeability. Molecular Ecology 16: 3374–3392 Storfer A., Eastman J.M. and Spear S.F., 2009. Modern molecular methods for amphibian conservation. Bioscience 59(7): 559-571 64 Studier J.A. and Keppler K.J., 1988. A note on the neighbor-joining algorithm of Saitou and Nei. Molecular Biology and Evolution 5: 729–731 Sullivan J. and Joyce P., 2005. Model selection in phylogenetics. Annu. Rev. Ecol. Evol. Syst. 36: 445–66 Talbot S. L. & Shields G. F., 1996. Phylogeography of brown bears (Ursus arctos) of Alaska and paraphyly within the Ursidae. Molecular Phylogenetics and Evolution 5: 477-494. Techow N.M.S.M., O’Ryan C., Phillips R.A. et al., 2010. Speciation and phylogeography of giant petrels Macronectes. Molecular Phylogenetics and Evolution 54: 472-487 Techow N.M.S.M., Ryan P.G. and O’Ryan C., 2009. Phylogeography and taxonomy of Whitechinned and Spectacled petrels. Molecular Phylogenetics and Evolution 52: 25-33 Templeton A.R., 1979. Once again, why 300 species of Hawaiian Drosophila? Evolution 33: 513– 517 Templeton A.R., 2008. The reality and importance of founder speciation in evolution. Bioessays 30: 470–479 Tierney L., 1994. Markov chains for exploring posterior distributions. The Annals of Statistics 22(4): 1701-1728 Thiébot J-B., Barbraud C., Scofield R.P., Cherel Y. and Bretagnolle V., 2010. New petrel record on Ile Saint-Paul, southern Indian Ocean. Notornis 57: 50-53 Turelli M., Barton N.H. and Coyne J.A., 2001. Theory and speciation. Trends in Ecology and Evolution 16: 330–343 Ueshima R. and Asami T., 2003. Single-gene speciation by left-right reversal: a land-snail species of polyphyletic origin results from chirality constraints on mating. Nature 425: 679 Via S., 2001. Sympatric speciation in animals: the ugly duckling grows up. Trends in Ecology and Evolution 16: 381–390 Veit R.R., McGowan J.A., Ainley D.G., Wahls T.R. and Pyle P., 1997. Apex marine predator declines ninety percent in association with changing oceanic climate. Global Change Biology 3: 23-28 vonEuler F., 2001. Selective extinction and rapid loss of evolutionary history in the bird fauna. Proceedings of the Royal Society of London B 268: 127-130 Wakeley J., 1996. The excess of transitions among nucleotide substitutions: new methods of estimating transition biasunderscore its significance. Trends in Ecology and Evolution 11(4): 158-163 Wallace A.R., 1858. On the tendency of species to form varieties; and on the perpetuation of varieties and species by natural means of selection. III. On the tendency of varieties to 65 depart indefinitely from the original type. Journal of the Proceedings of the Linnean Society London 3: 53–62. Wallace A.R., 1865. On the phenomena of variation and geographical distribution as illustrated by the Papilionidae of the Malayan region. Transactions of the Linnean Society of London 25: 1-71 Wallace A.R., 1875. The Malayan Papilionidae or swallow-tailed butterflies, as illustrative of the theory of natural selection. In Wallace, A.R. (ed.), Contributions to the Theory of Natural Selection, 2nd edn. Macmillan & Co., London, pp. 130–200 Wallace A.R., 1889. Darwinism. An exposition of the theory of natural selection with some of its applications. MacMillan, London Warham J., 1990. The Petrels: their ecology and breeding systems. Academic Press, London Warham J., 1996. The behaviour, population biology and physiology of the petrels. Academic Press, London Warheit K.I., 1992.A review of the fossil seabird from the Tertiary of the North Pacific: plate tectonics, paleoceanography, and faunal changes. Paleobiology 18: 401-424 Warheit K.I., 2002. The seabird fossil record and the role of paleontology in understanding seabird community structure. In Biology of Marine Birds. Schreiber E.A. & Burger J. eds., CRC Press, Boca raton, Florida USA Weisman A., 1892. Das Keim plasma. Eine Theorie der Vererbung. Fischer, Jena, Germany West-Eberhard M.J., 1983. Sexual selection, social competition and speciation. Quarterly Review of Biology 58: 155–183 Wiens J.J., 2004. The role of morphological data in phylogeny reconstruction. Systematic Biology 53(4): 653-661 Wiley E. O., 1978. The evolutionary species concept reconsidered. Systematic Zoology 27: 17-26 Wright S., 1931. Evolution in Mendelian populations. Genetics 16: 97–159 Wrobel B., 2008. Statistical measures of uncertainty for branches in phylogenetic trees inferred from molecular sequences by using model-based methods. Journal of Applied Genetics 49(1): 49-67 Zharkikh A. and Li W.-H., 1992. Statistical properties of bootstrap estimation of phylogenetic variability from nucleotide sequences. I. Four taxa with a molecular clock. Journal of Molecular Evolution 9: 1119–1147 66 CHAPTER 2: GENERAL METHODS 2.I Mitochondrial versus nuclear DNA For historical reasons (reviewed in Avise, 2000), as well as practical reasons, mtDNA has long been the molecule of choice for phylogenetic and phylogeographic studies. mtDNA presents the advantage of being present in high quantity in cells and to be relatively easy to sequence. It is also acknowledged for its non-recombining properties and its high mutation rate (Avise et al. 1987, Zink & Barrowclough 2008). In addition, because of its lower effective population size (Ne) compared to nuclear DNA (approximately 4Ne) lineage sorting will occur faster for mtDNA than for nucDNA thus allowing the detection of more recent vicariance events (Zink and Barrowclough, 2008; Fig. 2.1). Furthermore, until recently nuclear genes sequences were more difficult to obtain than mtDNA thus rendering their use limited to a few studies (Lee & Edwards, 2008). mtDNa has therefore been extensively used in the last 20 years and has allowed improving significantly the knowledge of phylogenetic and phylogeographic patterns in all groups of living organisms. However, in the last few years, several authors have questioned the supremacy of mitochondrial DNA (mtDNA) in vertebrate phylogenetic and phylogeographic studies by using nuclear genes in addition to mtDNA and uncovering conflicting results between the two types of markers (Ferris et al., 1983; Shaw, 2002; Leaché & McGuire, 2006; Robertson et al., 2006; Peters et al., 2007; Spinks & Shaffer, 2007; Good et al., 2008; Lee & Edwards, 2008; Spinks & Shaffer, 2009). It thus appears that inferences based on mtDNA alone should be regarded with caution (Funk & Omland 2003; Chan & Levin 2005). Despite this line of evidence, respective merits of nuclear versus mtDNA markers are disputed in regard to disentangling pattern and process in evolution and speciation (see Zink & Barrowclough 2008, Edwards & Bensh 2008). For some authors (e.g. Zink & Barrowclough, 2008), mtDNA remains the molecule of choice to determine geographical or taxonomic limits of recently evolved groups, while for studies that wish to investigate both patterns and processes of evolutionary and phylogeographic history it is necessary to use both types of markers. The combination of these allows investigating gene flow, population growth or phylogenetic relationships above the species level. However, it has been shown that it is now possible to estimate species phylogenies from gene trees despite incomplete lineage sorting (Carstens & Knowles, 2007). 67 Furthermore, because phylogenetics tries to describe the history and relationships of species and populations, each of which made of many independently segregating genes whose gene trees necessarily show some kind of differences between each others (Edwards, 2009), I follow this author in his statement that the use of a sole gene mitochondrial gene tree to infer the relationships of the taxa in which that gene tree is embedded should be questioned, and Petit & Excoffier (2009) when they suggest not using single uniparentally inherited genome (such as mtDNA) for species delimitations. Zink & Barrowclough (2008) summarise potential outcomes of comparison of mtDNA and nucDNA gene trees in four main scenarios. (i) both mtDNA and nucDNA retrieve reciprocally monophyletic groups. If the two markers retrieve the same groups, then there is a clear phylogenetic signal and taxa have probably diverged long ago, giving enough time to nucDNA lineage to sort. It can also happen that the two markers conflict in the monophyletic groups found. In that case, if only one nucDNA marker is used, then the lineage history is unresolved; if however, several nuclear merkers are used and yield the same grouping conflicting with mtDNA, then the latter, for some reason to identify, is an erroneous indicator of lineage history. (ii) mtDNA retrieve reciprocally monophyletic groups, but nucDNA do not support this monophyly. In the case of closely related taxa recently diverged, this can be explained by the inability for nucDNA to recover the split due to its longer coalescence time. Alternatively, this can be explained by a male-mediated gene flow resulting in the failure of the mtDNA marker to reflect the overall population structure. (iii) mtDNA do not recover differentiation whereas nucDNA does. Such case can happen if a selective sweep, by which an advantageous haplotype arises in one population and quickly expand, replacing other haplotypes at that locus, occurred in the mtDNA genome. Irwin et al. (2009) report the case of the yellowhammer (Emberiza citronella) and pine bunting (Emberiza leucocephalos) in which mtDNA is very likely to have introgressed between them as a result of a selective sweep. The result is a contradiction between the mitochondrial signal, indicating lack of differentiation between the two taxa, and the nuclear signal, pointing towards a differentiation. In this example, such hypothesis is supported partly by the important phenotypic differences. An 68 alternative scenario, by which the two taxa very recently diverged and reached nuclear and phenotypic differentiation, contrary to mtDNA, but this is less likely. (iv) both mtDNA and nucDNA fail to find reciprocally monophyletic groups. This can happen when all populations exchange migrants thus sharing haplotypes. Alternatively, dispersal has only recently ceased between geographically isolated populations and shared ancestral polymorphisms remain at all markers. As a result of what precedes, in this PhD work, all investigations have involved at least one nuclear marker in addition to one or two mitochondrial genes. In cases when sequences were available for only one mtDNA marker (CO1), caution was suggested relating to the interpretation of results and conclusion drawns. Figure 2.1. from Zink and Barrowclough, 2008. Capacity of different markers to detect vicariance 69 2.II Samples: origins and processing Two types of samples were used in the course of this work: samples that I will qualify as “fresh”, despite being sometimes collected on dead birds, and samples collected on museum specimens. 2.II.1 Fresh samples Samples of this type were collected by various fieldworkers around the world and consisted of three types: blood, feathers, tissue. Blood was collected from the veins on the leg or wing using microcapillaries and stored in 70% ethanol then frozen at -20°C until processing, except during mail transfer from the collector lab to the CEBC. Feathers were plucked from live birds, either on the back, at the basis of the neck or on the chest. Three to four feathers per bird were collected. Feathers were subsequently stored in 70% ethanol and frozen when possible until processing. Finally, tissue samples were collected on dead birds collected opportunistically in or around the breeding colonies, or at sea. Again, these samples were stored in ethanol and frozen when possible. 2.II.2 Museum samples These were collected on museum specimens with extra care to avoid damaging the specimen, and with the authorisation of the curator of the museum. A small piece of skin (about 1mm long) was cut from the web between the toes on one foot of the specimen. When processing these samples, extra care was taken to avoid any potential contamination, i.e. these were not processed along with fresh samples and DNA extraction and PCR preparation were carried out separately, on different days and in separate rooms, following standard procedure used in the Service de Systematique Moléculaire in the National Museum of Natural History in Paris. 70 2.II.3 DNA extraction Total genomic DNA was extracted using the DNeasy Tissue Extraction Kit (Qiagen, Valencia, CA, USA) following manufacturer’s instructions except that we increased the time of proteinase digestion to 8h, overnight. 2.II.4 Mitochondrial and nuclear DNA amplification and sequencing Cytochrome Oxydase 1 gene (CO1) was amplified using primers F1B 5’- AACCGATGACTATTYT-CAACC-3’ and R1B 5’-TACTACRTGYGARATGATTCC-3’, derived from primers F1/R1 (Simon et al., 1994). The PCR consisted of 37 cycles following a hot start at 94°C and a 4 minutes initial denaturation step at 94°C. Cycles, i.e. 94°C for 30s, 51°C for 40s and 72°C for 50s, were completed by a final extension at 72°C for 5 minutes. Cytochrome b gene was amplified using primers L14987 5’-TATTTCTGCTTGATGAAACT-3’ and H16025 5’-CTAGGGCTCCAATGATGGGGA-3’ (Jesus et al., 2009) and 40 PCR cycles consisting of 30s at 94°C, 50s at 58°C, 50s at 72°C. These cycles followed a 4 minutes initial denaturation step at 94°C and were completed by a final extension at 72°C for 5 minutes. Primers FIB-BI7U 5’-GGAGAAAACAGGACAATGACAATTCAC-3’ and FIB-BI7L 5’- TCCCCAGTAGTATCTGC-CATTAGGGTT-3’ (Prychitko & Moore, 1997) were used for Beta Fibrinogene intron 7 (βFibint7) amplification. We ran thirty-nine PCR cycles consisting in 1 minute at 94°C, 40s at 58°C and 50s at 72°C preceded by an initial denaturation step of 4 minutes at 94°C. These cycles were followed by a 5 minutes final extension step at 72°C. PCR procedure for cold shock domain containing E1 (CSDE1) and PAX interacting protein 1 (PAXIP1) introns followed protocols and used primers described in Kimball et al. (2009). In addition, for samples collected on dead or old museum birds, DNA was degraded and fragment sizes for amplification were mostly small (<200bp) and proved difficult to amplify and sequence. I therefore used additional primers designed specifically for the taxa studied. Amplification problems were mostly encountered with samples from genus Pseudobelweria. Primers used are summarised in Table 2.1. For all markers and all analyses, sequencing was conducted under BigDyeTM terminator cycling conditions at the “Genoscope - Centre National de Séquençage”, France. 71 DNA sequences were aligned using CodonCode Aligner 3.0.3 (CodonCode Corporation, 2009) and ClustalW (Thompson et al., 1994) as implemented in Mega version 4 (Tamura et al., 2007) and checked by eye. Table 2.1 Primers used in amplification of CO1, Cytb and βFibint7 genes in Pseudobulweria petrels. All primers were designed for this study PRIMERS CO1 F1B 5’AACCGATGACTATTYTCAACC-3’ R1B 5’TACTACRTGYGARATGATTCC-3’ R141 5’AGCATGGGCGGTGACGATT-3’ F78 5’ACTTATTCGTGCAGAACTTGGTC-3’ R208 5’AGGGGGACTAGTCAGTTTCC-3’ F150 5’CGCCCATGCTTTCGTAATAATTT-3’ R254 5’AGCTTATGTTGTTTATACGTGGGA3’ F207 5’TGGAAACTGACTAGTCCCCCT-3’ R323 5’ACCTGCTCCTGCTTCTACGG-3’ F288 5’ACCTCCGTCCTTCCTCCTAT-3’ R416 5’CCTGCCAGGTGGAGGGAGA-3’ F377 5’ATCTAGCCCATGCCGGAGC-3’ R502 5’AAGGGGGTTTGGTACTGTGA-3’ F453 5’GGCAATCAACTTCATTACAACAGC3’ R581 5’AGCATGGTGATGCCTGCGG-3’ F537 5’ACTCATCACTGCCGTCCTAC-3’ R678 5’TGGGTGGCCGAAGAATCAG-3’ F642 5’TGGCGGAGGAGACCCAGTC-3’ PRIMERS CYTB F13 5'-CTTCGAAAGTCCCACCCCT-3' R142 5'-TGGCTAGTAGGAGGCCGGT-3' F103 5'-GGATCCCTCCTAGGCATCTGT-3' R202 5'-ACATTTCGGCAGGTGTGAGC-3' F167 5'-ACACAGCTGACACAACCTTAGC3' R297 5'-CCGTAGTAGAATCCCCGTCCG-3' F249 5'-ACATGCAAACGGAGCCTCA-3' R369 5'-ACGAAGGCAGTTGCCATAAGA-3' F314 5'ACGGCTCCTACCTATACAAAGAG-3' R457 5'-TGGCCAATGTAGGGGATGGC-3' F418 5'-TCATTCTGAGGTGCGACAGTCA3' R529 5'AGTGTAGGGCGAAGAATCGGGT-3' F473 5'GCCAAACCCTTGTAGAATGAGCC-3' R595 5'-AGCCAGATTCGTGGAGGAAGGT3' F553 5'CTCCTACCTTTTGCAATCACAGGA-3' R674 5'AGCCTAGGATATCTTTTAAGGTGA-3' F630 5'-TGGTGTCGTATCAAACTGCGA-3' R758 5'CGCTGGAGTAAAGTTTTCTGGGT-3' F711 5'-TCTCCCACTAACAGCCCTAGCT3' R836 5'-GGAATTGAGCGTAGGATGGCGT3' F793 5'ACACCTCCCCATATTAAACCAGA-3' R915 5'-TGGCTTTATGGAGGAATGGA-3' F885 5'-AGCTGCCTCAGTATTGATCCTA3' R1012 5'-ACTGGCTGGCTGCCTACTCA-3' PRIMERS ΒFIBINT7 F21 5'-TGACAATTCACAATGGCATGTTCT3' R158 5'-ACCACGACATGCTGTGAAAACT-3' F72 5'-GATGGTACGTACTTGCATTAGACA3' R210 5'-TGCATGGACGTTCAGCTGGT-3' F160 5'-TTTTCACAGCATGTCGTGGT-3' R277 5'-ACTTGGCTGTGGAGCAGCA-3' F244 5'-GCCAAGGGCAGGTAAAACT-3' R380 5'-TGCCACCATCAGTCTCTGACA-3' F347 5'ACAAATCAGCAAATCTGGATGCAA-3' R462 5'-CCTGTCTCTTTCCTCAGGACCCA-3' F411 5'-CCACTGACTTGCTTAAGTAGGAA3' R522 5'-ACAATTGAGCTCCTGTCTTCTG-3' F476 5'-AGAGACAGGTAGCATGTCCTATT3' R638 5'-TGAGAACTGTACATCTTCCCCAA3' F574 5'-ACTATGTGCTATGTCTTTCTCT-3' R722 5'-GTCTACCGATTGTAGTCTAACTT-3' F641 5'GGGAAGATGTACAGTTCTCATTGT-3' R796 5'-GCACTTGGAAGGTGAAGCAGC-3' F756 5'-TCCGAAAGAGATGCAGCTAAA-3' R852 5'-AAATCCTCCCTGAACTTTCTGT-3' F807 5'-TTCCAAGTGCACTGTGTAGCA-3' R938 5'-GAGTGGCAGATGAACTGTAAGCA3' F900 5'-TCAGTACAGGGGCAGGTGTA-3' R1048 5'-GGGTTGGCTGAGTGGCAGC-3' 72 2.III Estimation of gene flow, population divergence time and effective population sizes Two extreme scenarios have traditionally been the basis of evolutionary models investigating genetic comparisons in different population, an equilibrium migration scenario, and isolation scenario. In the former, populations are supposed to have exchanged migrants at a constant rate for an infinite period of time. In the latter, populations are considered as having evolved independently without gene flow since the split from the common ancestral population. The problem with measures of genetic distance such as FST is that they cannot really differentiate between the two scenarios of equilibrium migration or isolation. In the case of FST, a low value can mean either a high gene flow over time between the populations studied, or a recent cessation of gene flow without enough time for drift and/or selection to accumulate differences between the populations. Isolation with Migration (IM) models bring a great improvement by allowing both cases to be considered and distinguished. IM can therefore be of great use for conservation biology by allowing differentiating populations that are still connected by migration and gene flow, from those that are isolated and are potential incipient species (or at least some kind of independent management units). This model includes population-size parameters for an ancestral and two descendant populations, as well as a splitting time term, and two parameters for gene exchange (Fig. 2.2). IM models assume selective neutrality, constant population size and no recombination. Early programs implementing these models, such as IMa (Hey & Nielsen, 2007), allowed investigating relationships between only two populations and had several limitations: there should not be other populations more closely related to the sampled populations that these were from each other, nor should there be unsampled populations exchanging genes with either of the study populations or the ancestral population. However, the recently released IMa2 (Hey, 2010) circumvents these limitations by allowing investigating relationships between two and ten populations (Fig. 2.3). The main difficulty when running programs such as IMa or IMa2 is to estimate if the output from markov chain Monte Carlo simulations describes patterns that can be close to reality. Thus many precautions need be taken, as described in the documentation, including running simulations at least three or four times. The first time serves to assess the relevance of chosen input parameters and to refine them, while other runs confirm or not that independent runs give equivalent results. 73 In the work presented in this manuscript I used the program IMa2 to estimate migration rates, dates of divergence and population sizes (at time of divergence due to the assumption of constant population size of the IM model) between populations of gadfly petrels living in the Eastern Atlantic Ocean and of Gould’s petrel of Australia and New Caledonia. Figure 2.2 From Pinho & Hey, 2010. (A) Isolation model with three population size parameters and a splitting time parameter. Effective population sizes (N), migration rates per gene copy per generation (M), and time since population splitting (T) are all scaled by the mutation rate (u) of the genes being studied. Effective population sizes are assumed to be constant. The ancestral population is assumed to have been present indefinitely back into the past. (B) The isolation-with-migration (IM) model includes two parameters for gene flow between the sampled populations. Gene flow rates are assumed to be constant over the time period since population splitting. A B Descendant populations (from which data are collected) Present M2.u 4N1u 4N2u 4N1u M1.u -1 -1 4N2u Splitting time 4NAu 4NAu Ancestral population Past 74 Figure 2.3 From Hey, 2010. Isolation with Migration model with 3 populations. Sampled populations 1 and 2 are more recently diverged from each other than either are with respect to population 3 75 References Avise J.C., 2000. Phylogeography. The history and formation of species. Harvard University Press, London, England. Avise J.C., Arnold J., Ball R.M., Bermingham E., Lamb T., Nigel J.E., Reeb C.A., Saunders N.C. 1987. Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Systematics 18:489–522 Carstens B.C. & Knowles L.L., 2007. Estimating species phylogenies from gene tree probabilities despite incomplete lineage sorting: an example from Melanoplus grasshoppers. Systematic Biology 56(3): 400-411 Chan K.M.A. & Levin S.A. 2005. Leaky prezygotic isolation and porous genomes: rapid introgression of maternally inherited DNA. Evolution 59:720–729 Edwards S.V., 2009. Natural selection and phylogenetic analysis. Proceedings of the National Academy of Sciences, USA 106(22): 8799-8800 Ferris S.D., Sage R.D., Huang C., Nielsen J.T., Ritte U. and Wilson A.C. 1983. Flow of mitochondrial DNA across a species boundary. Proceedings of the National Academy of Sciences, USA. 80:2290–2294. Funk D.J. & Omland K.E. 2003. Species-level paraphyly and polyphyly: frequency, causes, and consequences, with insights from animal mitochondrial DNA. Annual Review of Ecology and Systematics 34:397–423 Good J.M., Hird S., Reid N., Demboski J., Steppan S.J., Martin-Nims T.R. and Sullivan J. 2008. Ancient hybridization and mitochondrial capture between two species of chipmunks. Molecular Ecology 17: 1313–1327 Hey J. 2010. Isolation with migration model for more than two populations. Molecular Biology and Evolution 27(4): 905-920 Hey J & Nielsen R. 2007. Integration within the Felsenstein equation for improved Markov chain Monte Carlo methods in population genetics. Proceedings of the National Academy of Science, USA 104:2785–2790 Irwin D.E., Rubtsov A.S. and Panov E.N., 2009. Mitochondrial ntrogression and replacement between yellowhammers (Emberiza citrinella) and pine buntings (Emberiza leucocephalos) (Aves:Passeriformes). Biological Journal of the Linnean Society 98: 422-438 Jesus J., Menezes D., Gomes S., Oliveira P., Nogales M. and Brehm A., 2009. Phylogenetic relationships of gadfly petrels Pterodroma spp.from the Northeastern Atlantic Ocean: 76 molecular evidence for specific status of Bugio and CapeVerde petrels and implications for conservation. Bird Conservation International 19: 1-16 Kimball R.T. et al., 2009. A well-tested set of primers to amplify regions spread across the avian genome. Molecular Phylogenetics and Evolution 50: 654-660 Leaché A.D. and McGuire J.A. 2006. Phylogenetic relationships of horned lizards (Phrynosoma) based on nuclear and mitochondrial data: evidence for a misleading mitochondrial gene tree. Molecular Phylogenetics and Evolution 39:628–644 Lee J.Y. & Edwards S.V., 2008. Divergence across Australia’s Carpentarian barrier: statistical phylogeography of the Red-backed Fairy Wren (Malurus melanocephalus). Evolution 62: 3117–3134 Peters J.L., Zhuravlev Y., Fefelov I., Logie A. and Omland K.E. 2007. Nuclear loci and coalescent methods support ancient hybridization as cause of mitochondrial paraphyly between gadwall and falcated duck (Anas spp.). Evolution 61:1992–2006 Petit R.J. & Excoffier, L., 2009. Gene flow and species delimitation. Trends in Ecology and Evolution 24(7): 386-393 Pinho C. & Hey J., 2010. Divergence with gene flow: models and data. Annual Review of Ecology and Systematic 41: 215-230 Prychitko T. M. and Moore W. S., 1997. The utility of DNA sequences of an intron from the βfibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Picidae). Molecular Phylogenetics and Evolution 8: 193–204 Robertson A.V., Ramsden C., Niedzwiecki J., Fu J. and Bogart J.P. 2006. An unexpected recent ancestor of unisexual Ambystoma. Molecular Ecology 15:3339–3351 Shaw K.L., 2002. Conflict between nuclear and mitochondrial phylogenies of a recent species radiation: what mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. Proceedings of the National Academy of Science of the USA 99: 16122- 16127 Simon C., Frati F., Beckenbach A., Crespi B., Liu H. and Flook P., 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Annals of the Entomological Society of America 87(6): 651-701 Spinks P.Q. & Shaffer H.B. 2007. Conservation phylogenetics of the Asian box turtles (Geoemydidae, Cuora): mitochondrial introgression, numts, and inferences from multiple nuclear loci. Conservation Genetetics 8:641–657 Tamura K., Dudley J., Nei M. and Kumar S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24(8): 15961599 77 Thompson J.D., Higgins D.G. and Gibson T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22(22): 4673-4680 Zink R.M. & Barrowclough G.F., 2008. Mitochondrial DNA under siege in avian phylogeography. Molecular Ecology 17: 2107-2121 78 79 CHAPTER 3: THE COMPLETE PHYLOGENY OF PSEUDOBULWERIA THE MOST ENDANGERED SEABIRD GENUS: SYSTEMATICS, SPECIES STATUS AND CONSERVATION IMPLICATIONS Gangloff B, Shirihai H., Watling D., Cruaud C., Couloux A., Tillier A., Pasquet E. and Bretagnolle V., in revision, Conservation Genetics Abstract Pseudobulweria is one of the least known and most endangered of all seabird genera. It comprises five or six taxa, of which two are extinct, three are critically endangered and one is near threatened. Phylogenetic relationships between these taxa and position of the genus in Order Procellariiformes have never been studied, and the taxonomic status of several taxa remains unsettled. Conservation management of Pseudobulweria taxa will be enhanced if these uncertainties are resolved. We used a multilocus gene tree approach with two mitochondrial DNA markers (CytochromeOxidase subunit 1 and Cytochrome b gene) and one nuclear intron (βFibrinogen intron 7) to investigate phylogenetic relationships within the genus using sequences from all taxa. We combined gene trees to estimate a phylogeny of the genus using a multispecies coalescent methodology. We confirmed the link between Pseudobulweria and a clade comprising Puffinus and Bulweria genera. The Fiji petrel’s status within the genus is confirmed, as is the specific status of newly rediscovered Beck’s petrel. Maintenance of the two sub-species of Tahiti petrel is not supported. Discovering the breeding grounds of all taxa is the key for their conservation, which is vital to both the marine and fragile insular tropical ecosystems where Pseudobulweria are apical predators. Keywords Conservation; Procellariiformes; Pseudobulweria; seabird; CO1; Cytochrome b; βFibrinogen 80 Introduction Specific status is the most recognised unit used by conservation organisations and international and/or national government agencies to determine conservation policy and actions (Drummond et al. 2009; Farrier et al. 2007; Posingham et al. 2002). As resources (both human and financial) are limited, all conservation organisations need access to the most accurate systematic and taxonomic analyses of all taxa together with their evolutionary history. Seabirds, and more particularly petrels (Order Procellariiformes), have traditionally been difficult to study because much of their life is spent at sea coupled with the remoteness of their breeding grounds, usually on isolated oceanic islands (Brooke 2004; Friesen et al. 2007). In consequence, even today, detailed knowledge of their taxonomy, phylogeography and conservation status is scant and controversial (Brooke 2004). Though there has been considerable improvement in the past twenty years through the development of molecular biology, many groups in this Order remain poorly known, the systematics of many taxa and even genera is still debated, and their phylogenetic relationships unsettled (Brooke 2004; Mayr 2009). This is particularly disturbing at a time where most Procellariiformes taxa are in decline and to some degree threatened with extinction (IUCN 2010). Genus Pseudobulweria (family Procellaridae) is a good example. Of six taxa that compose the genus, two are extinct, three are critically endangered and one is near-threatened (IUCN 2010), making Pseudobulweria the most threatened seabird genus in the world. In addition, it remains one of the least known and most controversial petrel genera (Bretagnolle et al. 1998; Brooke 2004; Shirihai 2008). Indeed, the mere existence and validity of the genus remains controversial, its relationships with other petrels are contradictory, the number of species is debated, and species relationships have never been investigated. First proposed by Mathews (1936) for the Fiji Petrel (P. macgillivrayi), the genus was later merged within genus Pterodroma (Jouanin & Mougin 1979; Del Hoyo et al. 1992; Warham 1990) before being reinstated (Imber 1985; Sibley & Monroe 1990). Based on molecular data from only two taxa of the genus, Bretagnolle et al. (1998) confirmed its validity. Yet, its position Procellariiformes continues to be debated according to four recent phylogenetic analyses. Bretagnolle et al. (1998) found that Pseudobulweria was closely related to genus Puffinus (and next to Bulweria and Procellaria) rather than Pterodroma, but did not analyse Pachyptila. Nunn & Staley (1998) also found that Bulweria was closest to Procellaria and next to Puffinus, though they did not include any Pseudobulweria in their analysis. Subsequently, Kennedy & Page (2002) using a supertree approach to reconstruct the Procellariiformes phylogeny with partial trees from various 81 studies (including the two former ones) concluded that Pseudobulweria was closely related to Pachyptila and Halobaena, and not to Procellaria, Puffinus and Bulweria, in contradiction with Bretagnolle et al. (1998). Then, Penhallurick & Wink (2004), used Cytb as Nunn & Stanley (1998), to analyse the taxonomy in this Order (note critical views in Rheindt & Austin 2005), but did not include sequences from Bretagnolle et al. (1998) in their phylogenetic analyses. These authors found that Procellaria/Bulweria were a sister clade of the Pachyptila/Halobaena clade. These results are therefore clearly contradictory and call for further analyses. Furthermore, the number of species within the genus is still uncertain. Five Pseudobulweria taxa (four species, of which one has two subspecies) are currently known and a further one still remains to be named (Worthy & Tennyson 2004). The phylogenetic relationships between these are unknown (Brooke 2004; Shirihai et al. 2009) and even the precise number of species within the genus is not settled. The extinct taxon P. rupinarum is known only from St Helena Island in the Atlantic Ocean, where it was numerous, and was probably extirpated after 1502 (Olson 1975). From bone remains used to describe the taxon, it appeared to be slightly larger than Mascarene petrel P. aterrima. Another extinct taxon was breeding in the southern Tuamotu archipelago, Pacific Ocean, being apparently very abundant, and was the size of Beck’s P. becki and Fiji petrels (Worthy & Tennyson 2004). All surviving taxa now live in the Indian and Pacific Oceans (Fig. 3.1) and breeding colonies are unknown for all but Tahiti petrel (P. rostrata). Until recently, three taxa were known by no more than two (Beck’s, and Fiji petrels) and seven (Mascarene petrel) specimens (Attié et al. 1997; Bretagnolle et al. 1998; Shirihai 2008; Shirihai et al. 2009). Hence, given the paucity of data and poor number of specimens held in museums, systematic studies on this genus were so far mostly based on scant morphological data and Tahiti petrel is the only taxon that has been studied in some details (Villard et al. 2006). Furthermore, phylogenetic relationships between taxa have been investigated solely between Tahiti and Mascarene petrels, the validity of the two P. rostrata subspecies still being questioned (Bretagnolle et al. 1998; Villard et al. 2006) and the phylogenetics of Beck’s and Fiji petrels having never been investigated. In addition, the taxonomic status of becki is still uncertain. The recent rediscovery of Beck’s petrel (Shirihai 2008) and the first observation at sea of Fiji petrel (Shirihai et al. 2009) threw new light on the conservation status of this genus. In this paper we use three different genes obtained from all extant taxa in order to investigate species limits and species validity, species relationships, genus monophyly and genus position within the other petrels. In particular we use newly developed species tree inference tools based on Bayesian statistics and multispecies coalescent theory (Heled & Drummond 2010; Liu et al. 2009a, 2009b; O’Meara 2010; Yang & Rannala 2010). As even the most probable gene trees topologies are not necessarily congruent with species trees (Degnan & Rosenberg 2009; Nichols 2001), as seen in pines (Syring et 82 al. 2007), grasshoppers (Carstens & Knowles 2007), finches (Jennings & Edwards 2005) or hominids (Ebersberger et al. 2007), implementation of such multilocus approaches allows inference of phylogenies when there are conflicting branching patterns between different genes (Degnan & Rosenberg 2009; Heled & Drummond 2010). Figure 3.1 Repartition of Pseudobulweria taxa, including extinct St Helena petrel P. rupinarum Material and methods Samples Table 3.1 summarises the origin of Pseudobulweria samples used with all three genes sequenced. Tissue samples were collected on dead birds, apart from birds from New Caledonia that were alive and released after being bled. For the two type specimens of P. becki and for two specimens of P. aterrima, a 1mm piece of skin from the palm was collected without damaging the specimen. Total genomic DNA was extracted using the DNeasy Tissue Extraction Kit (Qiagen, Valencia, CA, USA) following manufacturer’s instructions except that we increased the time of proteinase digestion to 2h. Since samples were mostly collected on dead or old museum birds, DNA 83 was degraded and fragment sizes for amplification were mostly small (<200bp) and proved difficult to amplify and sequence. Thus we failed to obtain useable gene sequences for some samples, and for others we could not obtain whole gene sequences. We sequenced two mitochondrial genes and one nuclear intron. Primers used for sequencing are shown in Table 3.2. Table 3.1 Pseudobulweria taxa, identifying codes, origins, and tissue sources analyzed. Museum samples came from Paris National Museum of Natural History (MNHN), New York American Museum of Natural History (AMNH), Geneva Natural History Museum (MHNG), Fiji Museum (FM). Sample BB73 was collected by HS on a dead juvenile bird later deposited in the British Museum of Natural History (BMNH). Taxon Code Locality Tissu source Internal Id P. aterrima MNHN 1995-165 Reunion Island muscle B01 P. aterrima MNHN 1970-102 Reunion Island Skin (from palm) BB69 P. becki AMNH 235376 Bismarck sea Skin (from palm) BB72 P. becki BMNH 2008-1-1 New Ireland Leaver BB73 P. macgillivrayi NA Gau Island, Fiji Neck tissue BB74 P. macgillivrayi NA Gau Island, Fiji Neck tissue BB75 P. rostrata rostrata GenBank U704821 Gambiers Blood B15 P. rostrata rostrata MHNG P08-30 Marquesas Muscle BI46 P. rostrata rostrata MHNG P08-31 Marquesas Muscle BI47 P. rostrata rostrata MHNG P08-32 Marquesas Feather BI48 P. rostrata rostrata NA Tahiti Blood BB86 P. rostrata rostrata FM 170141 Vatuira, Fiji Muscle BD02 P. rostrata trouessarti GenBank U704931 New Caledonia Blood B02 P. rostrata trouessarti NA New Caledonia Feather BB84 P. rostrata trouessarti NA New Caledonia Feather BB85 P. rostrata trouessarti NA New Caledonia Feather 1 GenBank accession number refers to Cytb sequence from Bretagnolle et al. (1998) BD04 Cytochrome Oxydase 1 gene (CO1) was amplified using PCR consisting of 37 cycles following a hot start at 94°C and a 4 minutes initial denaturation step at 94°C. Cycles, i.e. 94°C for 30s, 55°C for 40s and 72°C for 50s, were completed by a final extension at 72°C for 5 minutes. Cytochrome b gene was amplified with 40 PCR cycles consisting of 30s at 94°C, 50s at 58°C, 50s at 72°C. These cycles followed a 4 minutes initial denaturation step at 94°C and were completed by a final extension at 72°C for 5 minutes. 84 Primers FIB-BI7U 5’-GGAGAAAACAGGACAATGACAATTCAC-3’ and FIB-BI7L 5’TCCCCAGTAGTATCTGC-CATTAGGGTT-3’ (Prychitko & Moore, 1997) were used along with other primers specifically designed (Table 3.2) for Beta Fibrinogene intron 7 (βFibint7) amplification. We ran thirty-nine PCR cycles consisting in 1 minute at 94°C, 40s at 58°C and 50s at 72°C preceded by an initial denaturation step of 4 minutes at 94°C. These cycles were followed by a 5 minutes final extension step at 72°C. For all genes, sequencing was conducted under BigDyeTM terminator cycling conditions at the “Genoscope - Centre National de Séquençage”, France. DNA sequences were aligned using CodonCode Aligner 3.0.3 (CodonCode Corporation 2009) and ClustalW (Thompson et al. 1994) as implemented in Mega version 4 (Tamura et al, 2007) and checked by eye. Table 3.2 Primers used in amplification of CO1, Cytb and βFibint7 genes in Pseudobulweria petrels. All primers were designed for this study, except for L14987/H16025 from Jesus et al. (2009) Primers CO1 F1B 5’-AACCGATGACTATTYTCAACC3’ R1B 5’TACTACRTGYGARATGATTCC-3’ R141 5’-AGCATGGGCGGTGACGATT3’ F78 5’ACTTATTCGTGCAGAACTTGGTC-3’ R208 5’-AGGGGGACTAGTCAGTTTCC3’ F150 5’CGCCCATGCTTTCGTAATAATTT-3’ R254 5’AGCTTATGTTGTTTATACGTGGGA-3’ F207 5’TGGAAACTGACTAGTCCCCCT-3’ R323 5’-ACCTGCTCCTGCTTCTACGG3’ F288 5’-ACCTCCGTCCTTCCTCCTAT-3’ R416 5’-CCTGCCAGGTGGAGGGAGA3’ F377 5’-ATCTAGCCCATGCCGGAGC-3’ R502 5’-AAGGGGGTTTGGTACTGTGA3’ F453 5’GGCAATCAACTTCATTACAACAGC-3’ R581 5’-AGCATGGTGATGCCTGCGG3’ F537 5’-ACTCATCACTGCCGTCCTAC3’ R678 5’-TGGGTGGCCGAAGAATCAG3’ F642 5’-TGGCGGAGGAGACCCAGTC3’ Primers Cytb L14987 5’-TATTTCTGCTTGATGAAACT-3’ H16025 5’-CTAGGGCTCCAATGATGGGGA-3’ F13 5'-CTTCGAAAGTCCCACCCCT-3' R142 5'-TGGCTAGTAGGAGGCCGGT-3' F103 5'-GGATCCCTCCTAGGCATCTGT-3' R202 5'-ACATTTCGGCAGGTGTGAGC-3' F167 5'-ACACAGCTGACACAACCTTAGC-3' R297 5'-CCGTAGTAGAATCCCCGTCCG-3' F249 5'-ACATGCAAACGGAGCCTCA-3' R369 5'-ACGAAGGCAGTTGCCATAAGA-3' F314 5'-ACGGCTCCTACCTATACAAAGAG-3' R457 5'-TGGCCAATGTAGGGGATGGC-3' F418 5'-TCATTCTGAGGTGCGACAGTCA-3' R529 5'-AGTGTAGGGCGAAGAATCGGGT-3' F473 5'-GCCAAACCCTTGTAGAATGAGCC-3' R595 5'-AGCCAGATTCGTGGAGGAAGGT-3' F553 5'-CTCCTACCTTTTGCAATCACAGGA-3' R674 5'-AGCCTAGGATATCTTTTAAGGTGA-3' F630 5'-TGGTGTCGTATCAAACTGCGA-3' R758 5'-CGCTGGAGTAAAGTTTTCTGGGT-3' F711 5'-TCTCCCACTAACAGCCCTAGCT-3' R836 5'-GGAATTGAGCGTAGGATGGCGT-3' F793 5'-ACACCTCCCCATATTAAACCAGA-3' R915 5'-TGGCTTTATGGAGGAATGGA-3' F885 5'-AGCTGCCTCAGTATTGATCCTA-3' R1012 5'-ACTGGCTGGCTGCCTACTCA-3' Primers βFibint7 F21 5'-TGACAATTCACAATGGCATGTTCT-3' R158 5'-ACCACGACATGCTGTGAAAACT-3' F72 5'-GATGGTACGTACTTGCATTAGACA-3' R210 5'-TGCATGGACGTTCAGCTGGT-3' F160 5'-TTTTCACAGCATGTCGTGGT-3' R277 5'-ACTTGGCTGTGGAGCAGCA-3' F244 5'-GCCAAGGGCAGGTAAAACT-3' R380 5'-TGCCACCATCAGTCTCTGACA-3' F347 5'-ACAAATCAGCAAATCTGGATGCAA-3' R462 5'-CCTGTCTCTTTCCTCAGGACCCA-3' F411 5'-CCACTGACTTGCTTAAGTAGGAA-3' R522 5'-ACAATTGAGCTCCTGTCTTCTG-3' F476 5'-AGAGACAGGTAGCATGTCCTATT-3' R638 5'-TGAGAACTGTACATCTTCCCCAA-3' F574 5'-ACTATGTGCTATGTCTTTCTCT-3' R722 5'-GTCTACCGATTGTAGTCTAACTT-3' F641 5'-GGGAAGATGTACAGTTCTCATTGT-3' R796 5'-GCACTTGGAAGGTGAAGCAGC-3' F756 5'-TCCGAAAGAGATGCAGCTAAA-3' R852 5'-AAATCCTCCCTGAACTTTCTGT-3' F807 5'-TTCCAAGTGCACTGTGTAGCA-3' R938 5'-GAGTGGCAGATGAACTGTAAGCA-3' F900 5'-TCAGTACAGGGGCAGGTGTA-3' R1048 5'-GGGTTGGCTGAGTGGCAGC-3' 85 Phylogenetic analyses Genetic distances, Haplotype network and Gene trees inference For the three genes, corrected genetic distances (K2P) between taxa were calculated with Mega v.4 (Tamura et al. 2007) using the pairwise deletion option to allow for comparison of complete gene sequences with sequences in which fragments were missing due to poor DNA quality. Since we were able to obtain additional sequences from P. rostrata trouessarti from New Caledonia with CO1 (for a total Tahiti petrel sample size of 68), we investigated the relationship between the specimens belonging to the two recognised subspecies of Tahiti petrel, P. rostrata rostrata and P. rostrata trouessarti (hereafter termed rostrata and trouessarti respectively). We built a median- joining haplotype network using the software Networks v4.5.1 (Bandlet et al. 1999; Fluxus Technology Ltd, 2009) after identifying haplotypes with DnaSP v5 (Librado & Rozas 2009), using all sequences of Tahiti petrel and two sequences of Beck’s petrel. For each of the three loci, phylogenetic relationships were estimated using Bayesian Markov Chain Monte Carlo (MCMC) phylogenetic analyses with MrBayes v3.1.2 (Huelsenbeck & Ronquist 2001; Ronquist & Huelsenbeck 2003). jModelTest v0.1.1 (Guindon & Gascuel 2003; Posada 2008) was used to evaluate the fit to the data of 24 models of nucleotide substitution implemented in MrBayes. We used Bayesian Information Criterion (BIC) to evaluate which model best fitted our data (Sullivan & Joyce 2005). For all Bayesian analyses, default priors of MrBayes 3.1.2 were used for MCMC parameters. We used three heated chains and one cold chain for all analyses and runs were started with random trees. Two independent MCMC runs were conducted with 4.106 generations each. Trees and parameters were sampled every 100 generations. Standard deviations of split frequencies were used to assess stationarity (cut off value of 1% was used), with the average standard deviation of split frequencies expected to approach zero when the two runs converge onto stationarity distribution (Ronquist et al. 2005). Additionally the potential scale reduction factor should approach one when runs converge. For each run the first 25% were discarded as burn-in. Since order Sphenisciformes is supposed to be the closest relative to Procellariiformes (Brooke 2004) and because we could obtain sequences from all three genes on GenBank for this taxon, we used Humboldt penguin Spheniscus humboldti as an outgroup to root the gene trees. Since the availability of sequences from different taxa was not homogeneous between genes in GenBank (Cytb sequences’ availability far exceeding CO1 or βFibint7 sequences’ availability), we could not gather the same data sets for the three genes. Thus, the data set used for βFibint7 was smaller, with five less genera than for the mitochondrial markers. Used sequences’ accession numbers are shown 86 in Appendix 1, Table 3. For clarity, we built CO1 phylogenetic tree with a limited number of sequences corresponding to individuals sequenced with the two other markers, and not with all 68 sequences at our disposal. Analyses were also conducted with the complete CO1 data set to check for potentially spurious results (data not shown). Species tree inference In “species tree”, we here use “species” following the terminology of Heled & Drummond (2010), i.e. it is not necessarily referring to the taxonomic rank, but designates any group of individuals that have diverged sufficiently to no longer have breeding history with individuals outside that group, and thus it can include taxonomic rank or any “diverging population structure”. We found different placements of genera, and variations in taxa positions within the Pseudobulweria clade in gene trees although incongruent pattern were mostly unsupported (posterior probabilities < 95%; see results). Potential gene trees incongruences with species trees have been known for decades (Pamilo & Nei 1988). A common practice to avoid the problem consists in concatenating data assuming that all data have evolved under a single evolutionary tree. However, this method was recently shown to result potentially in support of incorrect species trees (Kubatko & Degnan 2007). Here we used the *BEAST methodology (Heled & Drummond 2010) as applied in the software BEAST v1.5.4 (Drummond & Rambaut 2007) to reconstruct Pseudobulweria species tree. In order to precisely identify the phylogenetic relationships of Procellariiformes genera, we ran an analysis with all sequences used previously to build phylogenetic gene trees with MrBayes, first with both mitochondrial genes, then with these two markers and βFibint7 on a reduced dataset to confirm results from the mitochondrial analysis. We specified genera (Pseudobulweria, Puffinus, Fulmarus, Pterodroma etc.) as “units” to build the species tree. The analysis was run with unlinked substitution and clock models and unlinked trees. For each gene, we used a relaxed clock model with an uncorrelated lognormal distribution of the substitution rate with a fix mean value of 0.794 ± 0.115% per million year for CO1 (Pereira & Baker 2006) and 1.89 ± 0.35% (obtained for Procellariiformes) for Cytb (Weir & Schlutter 2008). Since βFibint7 lacks a well calibrated substitution rate, we did not enter this parameter. We used HKY model, with a discrete approximation of the gamma-distributed rate of heterogeneity (four rate categories, Yang 1994) for CO1 and βFibint7, and a GTR+G model for Cytb following jModeltest analysis. We specified a Yule process species tree prior under a continuous population size model. The analysis was run for 200.106 generations. The run was sampled every 1000 generation, with the initial 50,000 steps being discarded as burn in. We used the software TRACER v1.5 (Drummond & Rambaut 2007) to 87 visualize the results of the run as well as for checking effective sample size of each parameter. Second, we investigated the inter-taxa species tree within Pseudobulweria with the three genes using only Pseudobulweria sequences. The same specifications than the previous analysis were used. All *Beast analyses were run on the BioHPC compute cluster at the Cornell University Computational Biological Service Unit (http://cbsuapps.tc.cornell.edu) Results Sequencing of Pseudobulweria samples provided 1140bp sequences for Cytb, 736bp CO1 sequences and sequences of 993bp for βFibint7. Two samples failed to provide sequences in all three genes, i.e. one of the type of becki and one museum specimen of aterrima (Table 3.1). Some samples failed to give useable sequences for one or two genes despite numerous PCR attempts and therefore do not appear in corresponding gene trees. Sequences are deposited in GenBank under accession numbers XXXX The genus Pseudobulweria: monophyly, and relationships with other petrels With all three genes, although tree topologies differ somewhat in the placement of genera (Fig. 3.2; and Fig. 3.6, 3.7 Appendix 2), Pseudobulweria appears to be strictly monophyletic. Although we could not sequence rupinarum (only bones are available), plates provided in Olson (1975) clearly suggest that rupinarum also belongs to Pseudobulweria (based on characters on skull and especially bill angle in comparison to skull). With CO1, the gene used for the BarCode of Life in birds, Pseudobulweria appears as a sister clade of a clade containing Puffinus, Calonectris, Bulweria, Fulmarus and Macronectes (Fig. 3.2). Node support between these two clades, however, is not strong (0.88). In Cytb, Pseudobulweria is a sister group of Bulweria, although with posterior probability inferior to 0.8 (Fig. 3.6 Appendix 2). With this gene, different subclades within family Procellariidae are weakly supported, with the exception of the Fulmarus/Macronectes group. In βFibint7, Pseudobulweria is the sister group of Puffinus with strong support (posterior probability of 100%; Fig. 3.7 Appendix 2). However we lack data from genera such as Bulweria with this nuclear intron. 88 With both data sets (i.e. mitochondrial and nuclear+mitochondrial), clades recovered with species tree analyses on genera data recovers mostly weakly supported clades. These analyses indicate the presence of two main clades: one clade made of family Hydrobatidae and one clade made of families Diomedeidae and Procellariidae (Fig 3.3; Fig. 3.8 Appendix 2). In the latter clade, Pseudobulweria is the sister group to a sub-clade composed of Bulweria, Calonectris and Puffinus (Fig. 3.3) or Puffinus alone (Fig. 3.8, Appendix 2) though the separation is very weakly supported (Posterior probability 0.66 and 0.81 in mitochondrial and mitochondrial+nuclear analyses respectively; Fig. 3.3, Fig. 3.8 Appendix 2). In both analyses time to the most recent common ancestor of Puffinus and Pseudobulweria was estimated at about 13 millions years ago (95%HPD: respectively 11-23 and 5.5-21.4 million years ago in mitochondrial and mitochondrial+nuclear species tree analyses). Genetic distances and taxa relationships within genus Pseudobulweria Inter-taxa divergences (K2P corrected distances) in the genus with CO1 ranged from 1.21% between trouessarti and becki to 7.4% between this latter taxon and aterrima. Fiji petrel showed the lowest divergence with Beck’s (6.31%) and the greatest with Mascarene (7.15%) petrels respectively. Using βFibint7, the greatest divergence was between becki and Fiji petrel (2.46%), the latter being closest to Mascarene petrel (7.10-4%). Mean Fiji petrel divergence was the greatest with rostrata (7.99%) and the smallest with trouessarti (4.99%) when using Cytb. For the two forms of rostrata, sequence divergence was 0.12% with CO1, 0% with βFibint7 and 1.73% with Cytb. Within genus Pseudobulweria, with CO1 aterrima and macgillivrayi seem the first to have evolved in the sampled taxa, and becki is slightly differentiated from rostrata (Fig. 3.2). However, contrary to the monophyly of the genus that is well supported, nodes within the genus are poorly supported. The situation is different with Cytb. With this gene, macgillivrayi is embedded in a group also containing becki, rostrata and trouessarti, while aterrima is placed outside this group with strong node support. With the nuclear intron βFibint7, two subclades exist within Pseudobulweria, separating rostrata (all but the Fiji specimen) from the other taxa, although node support for these subclades is poor. Thus the relationship between the two Tahiti petrel taxa (i.e. rostrata and trouessarti) is unclear when using βFibint7, and unresolved with the two mitochondrial markers. 89 Figure 3.2 Bayesian Inference Phylogenetic Tree within Procellariiformes obtained with CO1 gene with MrBayes. Node with Posterior Probabilities superior or equal to 95% are indicated by bold lines and labels indicate posterior probabilities 90 Figure 3.3 Species tree reconstructed for genera within Order Procellariiformes with *BEAST, using two mitochondrial loci (CO1, Cytb). Node labels indicate posterior probabilities. Oceanites 0.82 Oceanodroma 0.99 Hydrobates 1 Phoebastria Pachyptila 0.37 1 Halobaena 0.98 Pterodroma Macronectes 0.28 1 Fulmarus 0.42 Pseudobulweria 0.66 0.4 Bulweria Puffinus 0.98 Calonectris 91 The species tree analysis finds support for the hypothesis that becki and rostrata are each other closest relatives and, within that group, that becki differs from the two Tahiti petrel taxa (Fig. 3.4). This tends to confirm the full separation of the becki lineage from the rostrata. According to this analysis, the dark (aterrima, macgillivrayi) and white vented (becki, rostrata) Pseudobulweria taxa diverged around six-seven millions years ago (95%HPD range: 3.2-10.75). Figure 3.4 Pseudobulweria species tree reconstructed with *BEAST, based on two mitochondrial loci (CO1, Cytb) and one nuclear marker (βFibint7). Node labels indicate posterior probabilities. P_mcgillivrayi 1 P_aterrima 0.64 P_becki 0.99 P_trouessarti 0.99 P_rostrata 1.0 92 Finally, haplotype network reconstruction with CO1 using Tahiti and Beck’s petrel sequences (Fig. 3.5) recovers a clear separation of becki from the two Tahiti petrel subspecies (11 mutations). Within Tahiti petrel, rostrata birds from Gambier, Tahiti and Marquesas are differentiated from New Caledonian birds. Rostrata specimen from Fiji, although having a different haplotype from the main trouessarti haplotype, is related to this later taxon rather than to other rostrata birds from Polynesia (Marquesas, Gambiers, Society). Figure 3.5 Haplotype network built with 68 sequences of Tahiti petrel and two sequences of Beck’s petrel with the mitochondrial gene CO1. Bars across lines connecting different haplotypes indicate the number of mutations between haplotypes. Small dots between some haplotypes indicate putative missing haplotypes. Circles sizes are proportional to number of individuals 93 Discussion Position of Pseudobulweria within Procellariiformes With Cytb, Bretagnolle et al. (1998) found that Pseudobulweria was part of a clade containing both Puffinus and Bulweria, with a more direct relationship to Puffinus. Here, with the same gene Pseudobulweria was more directly related to Bulweria than to Puffinus though with low support. With CO1, the relationship of Pseudobulweria was found in a clade including both Puffinus and Bulweria. Our species tree analyses, based on two and three genes simultaneously, strongly support a link between a Puffinus/Bulweria clade and Pseudobulweria, Halobaena and Pachyptila forming a clade distantly related within Procellariidae. Our results therefore clearly contradict Kennedy & Page (2002) findings that placed Pseudobulweria close to Halobaena/Pachyptila and those of Penhallurick & Wink (2004), who placed Bulweria also close to Halobaena/Pachyptila. Thus, despite the lack of nuclear intron data from Bulweria and Procellaria, we believe that our study, by integrating information from several markers, strongly supports the conclusion that Pseudobulweria affinities are to be found with Puffinus/Bulweria rather than with Pachyptila/Halobaena or Procellaria. Subspecies of Tahiti petrel Evidence from individual genes used in this study is somewhat confusing regarding the degree of divergence of these two lineages. CO1 node supports were low and corrected divergence was 0.12%, well under the average intraspecific divergence found among 260 North American bird taxa by Hebert et al. (2004). In addition, species tree reconstruction separated these two taxa from becki but does not allow concluding on their status and degree of divergence. An important limitation however remains: species tree analysis (such as in *Beast and other softwares) requires that the taxa investigated do not exchange genes. However, in the case of rostrata and trouessarti, gene exchanges cannot be ruled out, certainly not on the basis of our results since we have too few sequenced individuals from taxon rostrata. Thus, the genetic difference between these taxa appears very small, a result also found when using birds from American Samoa (M. Rauzon, S. Olson & R. Fleischer pers. comm.). In addition, inconclusive morphological differences was found (Villard et al. 2006), and the birds from Fiji belonging presumably to subspecies rostrata are actually closer to birds from New Caledonia (spp. trouessarti) in the haplotypic network, thus raising additional issues with regard to subspecies geographic delineation. Therefore, given the lack of specimens and 94 genetic data from Vanuatu and New Hebrides, we advocate the suppression of the two subspecies of Tahiti petrels because they are currently ill-defined geographically, until stronger evidence is provided either by morphological or phylogeographic investigation. Beck’s petrel taxonomic status Originally described by Murphy (1928) as warranting full specific status (though at that time he placed becki within genus Pterodroma) because of its smaller size (about 10-15%, with no overlap) compared to Tahiti petrel, Beck’s petrel taxonomic rank has subsequently been debated and challenged, and the taxon was considered either a subspecies of Tahiti petrel (Imber 1985; Jouanin & Mougin 1979; Warham 1990) or a full species (Collar & Andrew 1988; Sibley & Monroe, 1990). In addition to biometrics, at-sea behavioural differences led Shirihai (2008) to advocate for full specific status, though acknowledging that no single criteria (except size if judged correctly) could allow separating both forms at sea. We found that becki was consistently separated from rostrata in all three loci used. The separation of the two taxa in the species tree analysis was supported by a posterior probability of 0.99, giving credit to the full specific status of Beck’s petrel. Despite being well supported, the genetic distance is however small. Using CO1, a divergence of 1.21% was detected. Hebert et al. (2004) found that among 260 North American taxa, maximum average intraspecific divergence was 1.24% with an average value of 0.27%. Our value is therefore just below this average maximum divergence. Interestingly, Hebert et al. (2004) also found 13 species that showed interspecific distances lower than 1.25%. Thus, genetic divergence found with CO1 (as well as the two other genes) seems to indicate the presence of two distinct species, albeit only recently separated. Thus, we provisionally suggest that the two taxa should be considered as fully distinct species. We expect other important species isolating characters such as calls (Bretagnolle 1995) to confirm this separation when becki breeding colonies are discovered and birds recorded. Such pre-mating isolating trait is likely to be important in those taxa since it is rather likely that Tahiti petrels also breed in close vicinity to becki breeding colonies and even possibly together (HS pers. obs.). Fiji petrel The phylogenetic relationship of macgillivrayi with other members of Pseudobulweria had never been studied before. However, based on skull characters (Olson 1975), and at-sea behaviour 95 and flight (Shirihai et al. 2009), there was little doubt that Fiji petrel was a member of the genus. Our genetic data fully confirm the pattern, both gene trees and species tree analyses placing this taxon within Pseudobulweria. The exact placement within the Genus varied slightly with the different genes. The species tree topology, suggests that becki and rostrata branched apart from aterrima and macgillivrayi around 6-7 Myrs ago, in parallel with the colourations of the four species (rostrata and becki are white vented, while the other two are entirely dark). Timing of divergence Values found in these analyses need to be taken with caution given the overall poor node supports and also because we used only one estimation rate per marker and no fossil calibration. The separation between Mascarene/Fiji and Tahiti/Beck’s petrels estimated around 6-7 millions years ago corresponds to the end of Miocene, a time of marked ecological change (Janis 1993). The Messinian stage of the end of Miocene was characterised by important sea-level regression that were subsequently followed by sea transgression in the early Pliocene (Haq et al. 1987). Such pattern could have increased the available habitat on several islands in the Indo-Pacific region, thus promoting the colonisation of new breeding locations by Pseudobulweria common ancestors. New populations would then be isolated due to sea-level increase, promoting the divergence of these taxa and their colouration change by fixation of alternative alleles of the melanocortin-1 receptor gene (MC1R). This gene is known to affect colourations in several bird lineages through a single nonsynonymous change, as well as a wide range of other organisms, from lizards to mammals (Mundy 2005). Alternatively, changes in oceanic conditions could have driven some birds to modify their foraging habits thus promoting the differentiation of these lineages through different foraging patterns and at-sea behaviour. Indeed, colourations in Procellariiformes were suggested to be linked to feeding strategies and selective pressures such as competition and predation (Bretagnolle 1993). Similarly, the split between Puffinus/Bulweria and Pseudobulweria around 13 millions years ago correspond to the mid-Miocene climate transition, a period that experienced sharp oceanic temperature drops (Shevenell et al. 2004) and oceanic current changes (Miller & Fairbanks 1983). It was suggested that this period might have offered improved ecological conditions due to increased oceanic productivity resulting from the cooling of ocean surface temperatures thus possibly promoting taxonomic diversification of seabirds (Warheit 1992, 2002). This transition however remains one of the least understood of such events in the last 34 millions years (Lewis et al. 2008) and its importance in promoting seabird lineages diversification remains to be investigated. 96 Why are these petrels so rare? Pseudobulweria taxa apparently exhibit poor resilience to human presence and its accompanying invasive predators (e.g. Olson 1975). However, the recent rediscovery of Beck’s petrel and the survival of Mascarene and Fiji petrels on islands that suffered important human alteration and where many introduced predators now roam freely, show that these taxa can still survive for a while in adverse conditions. We suggest that surviving taxa have probably been saved by their formerly very large populations rather than the difficult access of their breeding sites. In comparison to Pterodroma that often breed in cliffs or top of active volcanoes, Pseudobulweria usually breed on more gentle slopes, at medium to low altitudes and even in some cases on the seashore. In addition, breeding sites of rostrata and aterrima at least are close to inhabited areas (e.g. in Marquesas, or Reunion Island; VB, pers. obs.). Despite this, and probably because these petrels breed on inhabited islands, their future survival is strongly impeded, and urgent action is required to save at least three taxa from a likely extinction in the near future. Conservation implications Given the current biodiversity crisis and financial limitations, investments may be justified in the conservation of distinct taxa, ecosystems or evolutionary units that are likely to produce future biodiversity (Bowen 1999). The conservation of Pseudobulweria is unlikely to promote future biodiversity. Despite this, we believe that in addition to being important systematics conservation target (Bowen 1999), these birds also constitute good conservation objectives in an “ecologist perspective” (Bowen 1999). Their conservation, both on land and at sea, would include highly diverse and fragile insular and marine ecosystems. For instance, investing in Beck’s petrel conservation would mean investing in the protection of whole forest ecosystems in one of the world’s most biodiversity rich region (Papua-New Guinea area), which might be at risk of excessive logging and/or forest clearance for oil-palm plantations, particularly in New Britain (Buchanan et al. 2008; Shirihai 2008). Similarly, the conservation of Fiji petrel on Gau Island, Fiji, by promoting the control of populations of introduced predators such as cats Felis catus, rats Rattus spp or pigs Sus scrofa (Priddel et al. 2008) will benefit both native plant and animal communities in remnant forest patches. In conclusion, conservation of these taxa is important both for their intrinsic evolutionary and taxonomic values, as well as for the wider ecosystems of which they are a part. In addition to current action in face of light-induced mortality (Le Corre et al. 2002, 2003; Priddel et al. 2008; P. Raust pers. comm.), conservation of all Critically Endangered species of the genus, i.e. 97 Mascarene, Beck’s and Fiji petrel, urgently requires the discovery of their breeding colonies for predator control, and possibly translocation operations, as it is unlikely that predator removal will be feasible. Acknowledgments This work was supported by the "Consortium National de Recherche en Génomique", and the "Service de Systématique Moléculaire" of the Muséum National d'Histoire Naturelle (CNRS UMS 2700). It is part of the agreement n°2005/67 between the Genoscope and the Muséum National d'Histoire Naturelle on the project "Macrophylogeny of life" directed by Guillaume Lecointre. We are deeply indebted to Joel Cracraft, Curator, Paul Sweet, Collection Manager, and Margaret Hart at the American Museum of Natural History (AMNH) for giving us access to the collections and providing samples from Pseudobulweria becki type specimens. BG also acknowledges receipt of a Collection Study Grant from the AMNH. We thank Alice Cibois for providing us samples from P. rostrata rostrata from Marquesas. Many thanks also to T. Steeves and P. Pelser for useful comments that improved a previous draft of the manuscript. References Attié C, Stahl J-C, Bretagnolle V (1997) New data on the endangered Mascarene Petrel Pseudobulweria aterrima: a third twentieth century specimen and distribution. Colonial Waterbirds 20:406–412 Bandlet HJ, Forster P and Röhl A (1999) Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16:37–48 Bowen BW (1999) Preserving genes, species or ecosystems? Healing the fractured foundations of conservation policy. Molecular Ecology 8:S5-S10 Bretagnolle V (1993) Adaptive significance of seabird colouration: the case of Procellariiforms. The American Naturalist 142(1):141-173 Bretagnolle V (1995) Systematics.of the Soft-plumaged Petrel Pterodroma mollis complex: new insight from vocalizations. Ibis 137:207–218 98 Bretagnolle V, Attié C, Pasquet E (1998) Cytochrome b evidence for validity and phylogenetic relationships of Pseudobulweria and Bulweria. The Auk 115(1):188-195 Brooke M (2004) Albatrosses and Petrels Across the World. Oxford University Press, New York Buchanan GM, Butchart SHM, Dutson G, Pilgrim, JD, Steininger MK, Bishop KD, Mayaux P (2008) Using remote sensing to inform conservation status assessment: estimates of recent deforestation rates on New Britain and the impacts upon endemic birds. Biological Conservation 141:56–66 Carstens BC, Knowles LL (2007) Estimating species phylogeny from gene tree probabilities despite incomplete lineage sorting: an example from Melanoplus grasshoppers. Systematic Biology 56:400–411 Collar NJ, Andrew P (1988) Birds to watch. ICBP Technical Publication No. 8. Cambridge, UK Dayrat B (2005) Toward integrative taxonomy. Biol J Linn Soc 85:407-415 Degnan JH, Rosenberg NA (2009) Gene tree discordance, phylogenetic inference and the multispecies coalescent. Trends in Ecology and Evolution 24(6):332–340 Del Hoyo J, Elliott A, Sargatal J (1992). Handbook of the birds of the world, vol 1. Lynx Edicions, Barcelona Drummond AJ, Rambaut A (2007) "BEAST: Bayesian evolutionary analysis by sampling trees." BMC Evolutionary Biology 7:214 Drummond SP, Wilson KA, Meijaard E, Watts M, Dennis R, Christy L, Posingham HP (2009) Influence of a Threatened-Species Focus on Conservation Planning. Conservation Biology 24(2):441-449 Ebersberger I, Galgoszy P, Taudien S, Taenzer S, Platzer M, von Haeseler A (2007) mapping human genetic ancestry. Molecular Biology and Evolution 24(10):2266-2276 Farrier D, Whelan R, Mooney C (2007) Threatened species listing as a trigger for conservation action. Environmental Science and Policy 10(3):219-229 Friesen VL, Burg TM, McCoy KD (2007) Mechanisms of population differentiation in seabirds. Molecular Ecology 16:1765-1785 Guindon S, Gascuel O (2003) A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood. Systematic Biology 52:696-704 Haq BU, Hardenbol J, Vail PR (1987) Chronology of fluctuating sea levels since the Triassic. Science 235:1156–1167 Hebert PDN, Stoeckle MY, Zemlak TS, Francis CM (2004) Identification of birds through DNA Barcodes. PLOS Biology 2(10):e312 Heled J, Drummond AJ (2010) Bayesian inference of species trees from multilocus data. Molecular Biology and Evolution 27(3):570-580 99 Huelsenbeck JP, Ronquist F (2001) MRBayes: Bayesian inference of phylogeny. Bioinformatics 17:754–755 IUCN (2010) IUCN Red List of Threatened Species. Version 2010.2. <www.iucnredlist.org>. Downloaded on 26 July 2010 Janis CM (1993) Tertiary mammal evolution in the context of changing climates, vegetation and tectonic events. Annual Review of Ecology and Systematics 24:467–500 Jennings WB, Edwards SV (2005) Speciational history of Australian grassfinches (Poephila) inferred from thirty gene trees. Evolution 59:2033–2047 Jesus J, Menezes D, Gomes S, Oliveira P, Nogales M, Brehm A (2009) Phylogenetic relationships of gadfly petrels Pterodroma spp. from the Northeastern Atlantic Ocean: molecular evidence for specific status of Bugio and Cape Verde petrels and implications for conservation. Bird Conservation International 19:1-16 Jouanin C, Mougin JL (1979). Order Procellariiformes. In Mayr E and Cottrell GW (eds) Checklist of birds of the world, vol 1, 2nd edn. Museum of Comparative Zoology, Cambridge, pp 48121 Kennedy M, Page RDM (2002) Seabird supertrees: combining partial estimates of Procellariiform phylogeny. The Auk 119(1):88-108 Kubatko LS, Degnan JH (2007) Inconsistency of phylogenetic estimates from concatenated data under coalescence. Systematic Biology 56:17–24 LeCorre M, Ollivier A, Ribes S, Jouventin P (2002) Light-induced mortality of petrels: a 4 year study from Reunion Island (Indian Ocean). Biological Conservation 105:93–102 Le Corre M, Ghestemme T, Salamolard M, Couzi F-X (2003) Rescue of the Mascarene petrel, a critically endangered seabird of Reunion Island, Indian Ocean. The Condor 105:387-391 Lewis AR et al. (2008) Mid-Miocene cooling and the extinction of tundra in continental Antarctica. Proceeding of the National Academy of Sciences of the USA 105(31):10676-10680 Librado P and Rozas J (2009) DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25(11):1451-1452 Liu L, Yu L, Kubatko L, Pearl DK, Edwards SV (2009a) Coalescent methods for estimating phylogenetic trees. Molecular Phylogenetic and Evolution 53(1):320–328. Liu L, Yu L, Pearl DK, Edwards SV (2009b) Estimating species phylogenies using coalescence times among sequences. Systematic Biology 58:468–477 Mathews GM (1936) A note on the Black Fiji Petrel. Ibis 6:309 Mayr G (2009) Notes on the osteology and phylogenetic affinities of the Oligocene Diomedeoididae (Aves, Procellariiformes). Fossil Record 12(2):133-140 100 Miller KG, Fairbanks, RG (1983) Evidence for Oligocene−Middle Miocene abyssal circulation changes in the western North Atlantic. Nature 306 (5940): 250–253 Mundy NI (2005) A window on the genetics of evolution: MC1R and plumage colouration in birds. Proceedings of the Royal Society B 272:1633-1640 Murphy RC (1928) Birds collected during the Whitney South Sea Expedition. IV. American Museum Novitates 322 Nichols R (2001) Gene trees and species trees are not the same. Trends in Ecology and Evolution 16(7):358-364 Nunn GB, Stanley SE (1998).Body size effects and rates of cytochrome b evolution in tube-nosed seabirds. Molecular Biology and Evolution 15:1360–1371 Olson SL (1975) Remarks on the generic characters of Bulweria. Ibis 117:111-113 O’Meara BC (2010) New heuristic methods for joint species delimitation and species tree inference. Systematic Biology 59(1)59-73 Padial JM, Miralles A, De la Riva I, Vences M (2010) The integrative future of taxonomy. Frontiers in Zoology 7:16 Pamilo P, Nei M (1988) Relationships between gene trees and species trees. Molecular Biology and Evolution 5(5):568–583 Pereira SL, Baker AJ (2006) A mitogenomic timescale for birds detects variable phylogenetic rates of molecular evolution and refutes the standard molecular clock. Molecular Biology and Evolution 23(9):1731–1740 Penhallurick J, Wink M (2004) Analysis of the taxonomy and nomenclature of the Procellariiformes based on complete nucleotide sequences of the mitochondrial cytochrome b gene. Emu 104: 125-147 Posada D (2008) jModelTest: Phylogenetic Model Averaging. Molecular Biology and Evolution 25(7):1253-1256 Posigham HP, Andelman SJ, Burgman MA, Medellin RA, Master LL, Keith DA (2002) Limits to the use of threatened species lists. Trends in Ecology and Evolution 17(11):503-507 Priddel D, Carlile N, Moce K, Watling D (2008) A review of records and recovery actions for the Critically Endangered Fiji Petrel Pseudobulweria macgillivrayi. Bird Conservation International 18:381–393. Prychitko TM, Moore WS (1997) The utility of DNA sequences of an intron from the β-fibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Picidae). Molecular Phylogenetics and Evolution 8:193–204 101 Rheindt FE, Austin JJ (2005) Major analytical and conceptual shortcomings in a recent taxonomic revision of the Procellariiformes – a reply to Penhallurick and Wink (2004). Emu 105:181186 Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19(12):1572-1574 Ronquist F, Huelsenbeck JP, Van der Mark P (2005) MrBayes 3.1 Manual. http://mrbayes.csit.fsu.edu/. Accessed March 2010 Shevenell AE, Kennett JP, Lea DW (2004) Middle Miocene Southern Ocean Cooling and Antarctic Cryosphere Expansion. Science 305:1766-1770 Shirihai H (2008) Rediscovery of Beck’s petrel Pseudobulweria becki , and other observations of tubenoses from the Bismarck archipelago, Papua New Guinea. Bulletin of the British Ornithologist Club 128:3–16 Shirihai H, Pym T, Kretzschmar J, Moce K, Taukei A, Watling D (2009) First observation of Fiji petrel Pseudobulweria macgillivrayi at sea: off Gau island, Fiji, in May 2009. Bulletin of the British Ornithologist Club 129:129-148 Sibley CG, Monroe BL (1990) Distribution and taxonomy of birds of the world. Yale University Press, New Haven Stockman AK, Bond JE (2007) Delimiting cohesion species: extreme population structuring and the role of ecological interchangeability. Molecular Ecology 16(16):3374-3392 Sullivan J, Joyce P (2005) Model selection in phylogenetics. Annual Review of Ecology Evolution and Systematics 36:445-466 Syring J, Farrell K, Businsky R, Cronn R, Liston A (2007) Widespread genealogical nonmonophyly in species of Pinus subgenus Strobus. Systematic Biology 56:163–181 Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24(8):1596-1599 Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22(22):4673-4680 Villard P, Dano S, Bretagnolle V (2006). First data on the breeding biology of the Tahiti Petrel Pseudobulweria rostrata. Ibis 148:285–291 Warham J (1990) The Petrels: their ecology and breeding systems. Academic Press, London Warheit KI (1992) A review of the fossil seabirds from the Tertiary of the North Pacific: plate tectonics, paleoceanography, and faunal change. Paleobiology 18(4):401-424 102 Warheit KI (2002) The seabird fossil record and the role of paleontology in understanding seabird community structure in Schreiber & Burger (eds) Biology of Marine Birds, CRC Press, Boca Raton, Florida, USA Weir JT, Schluter D (2008) Calibrating the avian molecular clock. Molecular Ecology 17:23212328 Worthy TH, Tennyson AJD (2004). Avifaunal assemblages from the Nenega-Iti and Onemea sites. In Conte E. and Vinton Kirch P. (eds) Archaeological Investigations in the Mangareva Islands (Gambier Archipelago), French Polynesia, Chapter 6, pp. 122-127. Archaeological Research Facility Contributions No. 62. University of California, Berkeley. Yang Z (1994) Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: approximate methods. Journal of Molecular Evolution 39(3):306-314 Yang Z, Rannala B (2010) Bayesian species delimitations using multilocus sequence data. PNAS 107(20): 9264-9269 103 Appendix 1 Table 3.3 Outgroup taxa and sequences GenBank accession numbers, BarCode Id number (marked with an asterisk) or internal code Family Genus Pachyptila Taxon Pachyptila_turtur AF076070 Halobaena Halobaena_caerulea CO1 Cytb BROMB700-07* AF076057 Bulweria Bulweria_bulwerii CO1 Cytb BROMB697-07* AJ004156 U74341 Puffinus Puffinus_tenuirostris CO1 Cytb βFibint7 DQ434025 DQ434027 U74352 AY695220 CO1 Cytb βFibint7 DQ434015 AF076085 DQ881991 CO1 DQ432808 DQ433417 AY139626 U74356 DQ433651 DQ432933 AJ004178 U74348 EF552765 DQ881958 Puffinus_lherminieri PROCELLARIIDAE CO1 Cytb Sequence Code BROMB324-06* Calonectris Calonectris_diomedea Cytb Fulmarus Fulmarus glacialis CO1 Cytb βFibint7 Macronectes Macronectes_giganteu s CO1 Cytb FJ027768 U48941 AF076060 Pterodroma Pterodroma_hasitata CO1 Cytb DQ434001 EU167017 Pterodroma_cahow CO1 Cytb KKBNA075-04* U74331 Pterodroma_cookii CO1 Cytb ROMC325-07* U74345 Pterodroma_ultima CO1 Cytb βFibint7 BE84 BE85 BE84 104 Pterodroma_neglecta CO1 DIOMEDIDAE Cytb Phoebastria Cytb βFibint7 HYDROBATIDAE Oceanites SPHENI SCIDAE Phoebastria_nigripes βFibint7 CO1 Oceanites_oceanicus CO1 Cytb βFibint7 Hydrobates Hydrobates_pelagicus CO1 Cytb βFibint7 Oceanodrom Oceanodroma_leucor a hoa Spheniscus Spheniscus humboldti CO1 Cytb βFibint7 CO1 Cytb βFibint7 BE54 BE61 U74341 GQ328985 BG54 DQ433934 DQ433935 U48950 EU166988 EU739406 EF552760 DQ433048 DQ433049 AF076062 EU739449 AY567885 AF076059 AJ004182 DQ881965 AY666284 DQ434684 DQ434685 AF076064 AY695221 AY567888 DQ137220 DQ881996 105 Appendix 2 Figure 3.6 Bayesian Inference Phylogenetic relationships within Procellariiformes obtained with Cytb gene in MrBayes. Node with Posterior Probabilities superior or equal to 95% are indicated by bold lines and labels indicate posterior probabilities 106 Figure 3.7 Bayesian Inference Phylogenetic relationships within Procellariiformes obtained with BetaFibrinogen intron 7 in MrBayes. Node with Posterior Probabilities superior or equal to 95% are indicated by bold lines and labels indicate posterior probabilities 107 Figure 3.8 Species tree reconstructed for genera within Order Procellariiformes with *BEAST, using two mitochondrial loci (CO1, Cytb) and one nuclear intron (βFibint7). Node labels indicate posterior probabilities Oceanodroma 0.98 Hydrobates 0.36 Oceanites 1 Phoebastria Fulmarus 0.49 0.39 Pterodroma 0.99 Puffinus 0.81 Pseudobulweria 108 Feae petrel © Hadoram Shirihai 109 CHAPTER 4: TRACKING THE PHYLOGEOGRAPHIC HISTORY IN NORTH EAST ATLANTIC GADFLY PETRELS REVEALS MITOCHONDRIAL AND NUCLEAR DNA INCONGRUENCE AND COMPLEX SCENARIOS Gangloff et al., in prep Abstract Northeast Atlantic Macaronesia archipelagos have been hosts of complex patterns of colonisation and vicariance in many groups of organisms. Complex phylogeographic patterns are also found in seabirds including the gadfly petrels (genus Pterodroma). Long time considered sub-species of the widely distributed Soft-plumaged petrel, the taxonomic status of the three gadfly petrel taxa breeding in Macaronesia is not yet settled, some authors advocating the presence of three, two or one species. These birds have already been the subject of genetic studies but only with one mtDNA marker. However, several recent studies in a range of taxa have underlined potential discrepancies between mitochondrial and nuclear DNA when investigating phylogenetic and phylogeographic histories. In this study, using a total of five loci (2 mitochondrial and 3 nuclear introns) we investigated the population and phylogeographic histories in these threatened seabirds. Despite their morphological and behavioural (calls) similarities, these taxa showed complete lineage sorting with mtDNA contrary to nucDNA, and no gene flow was detected between them. In addition, genetic diversity was unexpectedly high for taxa with such low current population sizes (supposed to be less than 1,000 birds). It appears that these birds diverged in the late Pleistocene in the last 110,000 years, i.e. 10 times more recently than previous estimates based solely on one mtDNA gene. Finally, contrary to previous studies, our results suggest that Madeira petrel is ancestral rather than Feae petrel from Cape Verde. This study advocates the use of nuclear loci in addition to mtDNA when inferring complex demographic and phylogeographic histories of organisms Keywords : Pterodroma, Madeira, Cape Verde, Phylogeny, Petrel, Intron, CO1, Cytochrome b, evolutionary history, divergence, seabird, nucDNA, mtDNA 110 Introduction The supremacy of mitochondrial DNA (mtDNA) in vertebrate phylogenetic and phylogeographic studies has been recently challenged by the increasing use of, and the sometimes conflicting results that emerged from, nuclear genes (Edwards et al. 2005, Zink & Barrowclough 2008, Edwards & Bensh 2008). mtDNA is acknowledged for its non-recombining properties and its high mutation rate (Avise et al. 1987, Zink & Barrowclough 2008). In addition, because of its lower effective population size (Ne) compared to nuclear DNA (approximately 4Ne) lineage sorting will occur faster for mtDNA than for nucDNA thus allowing the detection of more recent vicariance events (Zink and Barrowclough, 2008). Moreover, until fairly recently, nuclear genes were not available (Lee & Edwards, 2008). However, with the increasing use of nuclear markers along with mtDNA, a growing body of evidence has indicated that inferences based on mtDNA should be regarded with caution (Funk & Omland 2003; Chan & Levin 2005). Many studies indeed revealed incongruent results between mtDNA and nuclear genes (e.g., Shaw 2002, Spinks & Shaffer 2009; see also Zink & Barrowclough 2008 and Lee & Edwards, 2008 for a review and a recent example in birds). Incongruence has been explained by lower mutation rate of the nuclear genes, incomplete lineage sorting, and possibly more complicated evolutionary scenarios including hybridization or natural selection which may actually act on mtDNA (Avise et al. 1987; Edwards et al. 2005, Bazin et al. 2006, Spinks & Shaffer 2009). Although the respective merits of nuclear versus mtDNA markers are disputed in regard to disentangling pattern and process in evolution and speciation (see Zink & Barrowclough 2008, Edwards & Bensh 2008), consensus has eventually emerged that in front of complex evolutionary or biogeographical histories, mtDNA alone will be insufficient (Chan & Levin 2005). Oceanic islands have long been considered laboratories of evolution (Losos & Ricklefs, 2009) where fundamental evolutionary processes such as speciation can be addressed (review in Emerson, 2002). For more than 50 years, debates have been heated over the main processes of speciation, and it is now generally agreed that almost every speciation events are associated with some kind of geographic separation (Fitzpatrick et al., 2009; Coyne & Orr, 2004). This has been especially well studied in birds, this class of vertebrates being for a long time one of the best studied groups (Price 2008). In this context, it may therefore seem paradoxical that oceanic birds, that mainly breed on islands, have been so poorly studied (but see Friesen, 2007; Friesen et al. 2007): the most recent and major review of speciation in birds provides only seven examples from the seabirds (two from boobies, four from gulls, and one from the petrels), only concerning three genera (Price 2008). Pelagic seabirds breeding on oceanic islands can be seen as challenging models as they are highly vagile and physical barriers to their dispersal are virtually non existent, 111 and therefore this should enhance high gene flow between populations. However, counteracting selective forces consist in the discrete location of breeding localities, and extreme phylopatric behaviour shown by these seabirds that should, on the reverse, enhance genetic drift. Therefore, conflicting processes are acting on the populations genetics of seabirds, with the petrels (order Procellariformes, containing 113 recognised species; Birdlife International, 2010) being the extreme example in this group. These are the most pelagic seabirds and almost all species breed on oceanic islands (review in Brooke 2004). They also show an extreme degree of philopatric behaviour (Austin 1996; Brooke 2004). In addition, because many species show nocturnal habits on the breeding grounds, their communication system (and presumably behavioural pre-isolating mechanism: Bretagnolle 1996) almost never rely on visual cues, and many taxa show cryptic and highly convergent plumages, making their taxonomy poorly understood (Brooke, 2004). This has led to many cases of phylogenetic uncertainties, including cryptic species with unclear phylogeographic histories causing constant revision of their taxonomy. Indeed recent use of phylogenetic information has led to several taxonomic revisions (Abbott & Double 2003, Burg & Croxall 2004, Austin et al. 2004, Bolton et al. 2008), or the discovery of cryptic species and strong genetic population structuring (Friesen et al. 2006, 2007, Smith & Friesen 2007, Bolton et al. 2008). Perhaps one of the best example of Procellariiformes phylogenetic complexity exists in the north-east Atlantic islands, where gadfly petrels (genus Pterodroma) breeding in Macaronesia (including the archipelagos of Azores, Madeira, Canarias and Cape Verde) have led to long-lasting and never-ending taxonomic debates (review in Zino et al. 2008, Shirihai et al. 2010). Three Pterodroma taxa breed in these islands: one on Madeira island itself (madeira), one on Bugio island (deserta) which is just 30km south-east of Madeira, and one in Cape Verde archipelago (feae), on four distinct islands (Fig. 1). First considered subspecies of the widely distributed soft-plumaged petrel, P. mollis, the three forms were subsequently considered as one distinct species with three subspecies (Bretagnolle 1995), two species including one with two subspecies (Bourne 1983, Zino et al., 2008), or three distinct species (Robb, 2008, Jesus et al. 2009). Phylogenetic analyses based on mtDNA have represented a major step in these splits, since these populations hardly differ morphologically, in colouration and vocalisations (Zino & Zino, 1986; Bretagnolle, 1995; Zino et al., 2008, Shirihai et al. 2010). Using a total of 26 individuals, Zino et al. (2008) showed that birds from Madeira and Bugio differ on Cytochrome b (Cytb) sequences by 2.2% nucleotide divergence. They further suggested that the two species separated some 2.5 million years ago (i.e., late Pliocene), though Sangster et al. (2002) suggested a much recent split (840,000 years ago, Early Pleistocene), based on unpubl. data on Cytb also. Similarly, Jesus et al. (2009) compared 35 individuals from Bugio and Cape Verde, and found that the average K2P sequence divergence 112 (Kimura, 1980) was 1.58% (2.4% and 2.3%, respectively, for Madeira and Bugio, and Madeira and Cape Verde divergences). The split between Bugio and Cape Verde was suggested at 1.75 million years ago. In both studies, no haplotypes were shared between the three populations. Despite the evidence presented by Jesus et al. (2009), Birdlife International currently recognises only two species, madeira and feae, the latter with two subspecies, feae and deserta. Though never tested accurately, the acknowledged historical scenario proposes that birds came from Cape Verde and colonised Madeira twice, hence explaining the differences between Madeira and Bugio birds despite the close proximity of these islands (see e.g., Bourne 1983). Because taxonomic uncertainties are often due to complex underlying phylogeographic patterns that are uneasy to apprehend with the use of a single mitochondrial locus and, since the separation of Macaronesian petrels in three species is currently uncorroborated by any other criteria such as morphology (Shirihai et al., 2010) or vocalisations (Bretagnolle 1995; but see Robb 2008), we investigated further the population, phylogeographic and biogeographic histories of these petrels. Methods are now available that provide an analytical framework to investigate the origin of island taxa and their evolution (Emerson, 2002; Liggins et al., 2008), by integrating simultaneously several genes histories (e.g. Heled & Drummond, 2010; Hey, 2010). We therefore used one additional mitochondrial gene and three nuclear introns in a much extended sample of birds from the three taxa to investigate Macaronesian petrel relationships. More precisely, i) we test whether nuclear DNA confirms the reciprocal monophyly of the three taxa; ii) we used phylogeography and population genetics tools to infer the population structure, timing of divergence, and demographic history in these threatened seabirds. 113 Figure 4.1. Macaronesian archipelagos and breeding localities of Macaronesian Pterodroma taxa: Madeira island (P. madeira), Bugio island (P. deserta), Fogo and other Cape Verde islands (P. feae) 114 Material and Methods Because of the taxonomic uncertaintiessurrounding these taxa we consider them without reference to their potential species status in the following. Sample collection and laboratory methods Blood samples were obtained from Bugio Island (Desertas islands, off Madeira), several colonies on Madeira Island and two islands of the Cape Verde, Fogo and Santo Antao Islands (Fig. 4.1). Blood was collected from the veins on the leg or wing using microcapillaries and stored in 70% ethanol then frozen at -20°C until processing. In all analyses, no differentiation was found between different colonies within madeira and feae (data not shown), and therefore we pooled the data from all colonies within each of the three taxa. Total genomic DNA was extracted using the DNeasy Tissue Extraction Kit (Qiagen, Valencia, CA, USA) following manufacturer’s instructions except that we increased the time of proteinase digestion to 2h. Cytochrome Oxydase 1 gene (CO1) was amplified using primers F1B 5’-AACCGATGACTATTYT-CAACC-3’ and R1B 5’TACTACRTGYGARATGATTCC-3’, derived from primers F1/R1 (Simon et al., 1994). The PCR consisted of 37 cycles following a hot start at 94°C and a 4 minutes initial denaturation step at 94°C. Cycles, i.e. 94°C for 30s, 51°C for 40s and 72°C for 50s, were completed by a final extension at 72°C for 5 minutes. Cytochrome b gene was amplified using primers L14987 5’-TATTTCTGCTTGATGAAACT-3’ and H16025 5’-CTAGGGCTCCAATGATGGGGA-3’ (Jesus et al., 2009) and 40 PCR cycles consisting of 30s at 94°C, 50s at 58°C, 50s at 72°C. These cycles followed a 4 minutes initial denaturation step at 94°C and were completed by a final extension at 72°C for 5 minutes. Primers FIB-BI7U 5’-GGAGAAAACAGGACAATGACAATTCAC-3’ and FIB-BI7L 5’- TCCCCAGTAGTATCTGC-CATTAGGGTT-3’ (Prychitko & Moore, 1997) were used for Beta Fibrinogene intron 7 (βFibint7) amplification. We ran thirty-nine PCR cycles consisting in 1 minute at 94°C, 40s at 58°C and 50s at 72°C preceded by an initial denaturation step of 4 minutes at 94°C. These cycles were followed by a 5 minutes final extension step at 72°C. PCR procedure for cold shock domain containing E1 (CSDE1) and PAX interacting protein 1 (PAXIP1) introns followed protocols and used primers described in Kimball et al. (2009). For the five markers, sequencing was conducted under BigDyeTM terminator cycling conditions at the “Genoscope - Centre National de Séquençage”, France. DNA sequences were aligned using CodonCode Aligner 3.0.3 (CodonCode Corporation, 2009) and ClustalW (Thompson et al., 1994) as implemented in Mega version 4 (Tamura et al., 2007) and checked by eye. 115 Genetic diversity, genetic structure and historical demography To estimate haplotypes of nuclear loci, the software PHASE 2.1.1 (Stephens et al. 2001) was used as implemented in DnaSp v5 (Librado & Rozas, 2009). The latter was then used to estimate haplotypes of all five loci for use with Arlequin. Nucleotide diversity (π) and haplotype diversity (h) were calculated with Arlequin v3.5 (Excoffier & Lischer, 2010). With this software, we also conducted an exact test of population differentiation (Raymond & Rousset 1995) and investigated population structure with an AMOVA. Furthermore, to assess whether selection has affected the loci examined and also to test for past population expansion, Fu’s Fs (Fu, 1997) and Tajima’s D (Tajima, 1989) tests of neutrality were performed using 1,000 simulations in Arlequin. Although Fu’s indices have been found to be more powerful than Tajima’s (Ramos-Onsins & Rozas, 2002), it was also shown by these authors to have an irregular behaviour. Thus, we considered FS statistic significant when the P value was below 0.02 (Fu, 1997). Significant negative values of both statistics point towards population growth and/or positive selection, whereas positive Tajima's D values indicate bottlenecks and/or balancing selection. We also investigated historical demography of the three taxa using mismatch distribution under a model of sudden range expansion with Harpending’s raggedness index (Harpending, 1994) and the sum of squared deviation SSD (Schneider et al., 2000) between the observed and expected distributions. A significant SSD value indicates a departure from the model of sudden population expansion (Schneider & Excoffier, 1999) and low values of r usually reflect a population expansion. Evolutionary relationships estimation jModelTest v0.1.1 (Posada, 2008; Guindon & Gascuel, 2003) was used to evaluate the model of nucleotide substitution that best fitted the data. We used Bayesian Information Criterion (BIC), which penalises more heavily than AIC overparamerisation therefore avoiding unnecessary complex models when sample size increases (Sullivan & Joyce, 2005). For each of the five loci, phylogenetic relationships were estimated using Bayesian Markov Chain Monte Carlo (MCMC) phylogenetic analyses with MrBayes v3.1.2 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003). For all Bayesian analyses, default priors of MrBayes 3.1.2 were used for MCMC parameters. We used three heated chains and one cold chain for all analyses and runs were started with random trees. Two independent MCMC runs were conducted with 4.106 generations for each run for all five markers. Trees and parameters were sampled every 100 generation. Standard deviation of split frequency were used to assess stationarity, with the average standard deviation of 116 split frequencies expected to approach zero when the two runs converge onto stationarity distribution (Ronquist et al., 2005). Additionally the potential scale reduction factor should approach one when runs converge. For each run the first 25% were discarded as burn-in. For CO1 data, we added sequences from Pterodroma hasitata retrieved on GenBank (accession number DQ434000 and DQ434001) to root the tree, as this species appears to be one of the closest relative of these three taxa (Nunn & Stanley, 1998; Zino et al., 2008). For Cytb data, sequences of Pterodroma mollis (n=1), P. cahow (n=1) and P. hasitata (n=2) retrieved on GenBank (accession numbers U74334, U74331, EU167017, U74332 respectively) were added. βFibint7 tree was rooted using one sequence of Puffinus pacificus (Gangloff et al., MS submitted). Finally, trees obtained with CSDE1 and PAXIP1 were rooted with sequences of Pseudobulweria becki (Gangloff et al., MS submitted). In addition, for all five markers, phylogenetic relationships between haplotypes were inferred using the median-joining networks method as implemented in NETWORK v4.5.1 (Bandlet et al., 1999; Fluxus Technology Ltd, 2009). Estimation of gene flow, population divergence time and effective population sizes A potential consequence of gene flow between populations could be unresolved phylogenetic gene trees. Hence, using IMa2 (Hey, 2010), we estimated this parameter between macaronesian petrel populations. IMa2 allows testing for gene flow between any number of populations between two and ten, but it requires an a priori topology of the populations under study. Since we cannot be certain of the correct species tree topology in our case, in a first step, we ran IMa2 on pairs of petrel populations rather than on all three populations at once. In a second step we ran an analysis using the topology obtained with mitochondrial genes (this study, see Results; Jesus et al., 2009). Several runs were sampled with a burn-in period of 4.105 steps and sampling period of 2.106 steps, using 10 to 100 chains and a geometric heating scheme. A range of mutation rates can be given as prior to the analysis for scaling parameter estimates in demographic units. We used a mean value of 0.794 ± 0.115% per million year for CO1 (Pereira & Baker, 2006) and 1.89 ± 0.35% for Cytb (value found for Procellariiformes; Weir & Schluter, 2008). For introns, we let IMa2 calculate mutation rates scalars for the other loci (i.e. no mutation rate was specified for the three nuclear introns). In addition, to evaluate the degree of gene flow, divergence time and population sizes in demographic unit, it is possible to enter the generation time of a taxon. Pterodroma petrels are, as all other Procellariiformes long-lived seabirds. No precise estimation of generation time currently exists for Macaronesian petrels. However, in the Atlantic petrel (Pterodroma incerta) this parameter is estimated to be 20 years (Cuthbert, 2004) and in another 117 genus belonging to Family Procellariidae, Scott et al. (2008) found two estimates of generation time in the sooty shearwater (Puffinus griseus) of 14.7 and 21.6 years respectively based on different adult survival assumptions. Furthermore, adult survival rate in a similar species was estimated at 0.93 (Simmons et al. 1984, Brooke et al. 2010), giving a generation time of c. 14 years. We thus used an approximate generation time of 15 years for Macaronesian petrels. Runs were monitored by observing effective sample size (ESS) values and inspecting parameter plots for trends following the manual recommendations. Analyses were run three times to ensure convergence and since all results were similar, only one is presented here. As computations with this software can take a very long time, and because we had unbalanced sample sizes between loci, we ran all analyses using a subset of samples from CO1, Cytb and βFibint7 corresponding to individuals from which we obtained CSDE1 and PAXIP1 sequences, rather than integrating all the sequences at our disposal for the three former markers. Results Genetic diversity Overall, madeira and feae had a higher genetic diversity than deserta, the latter consistently having lower h and π values than its macaronesian counterparts in all loci (Table 4.1 for all five loci). Despite the fact that the sample size of individuals in deserta was 1.5-1.6 times more than those for feae and madeira, the latter two showed twice as many haplotypes (9 and 11) as deserta (4 and 6; Table 4.1) with mitochondrial markers. A similar situation arose with PAXIP1 (4 vs. 8 and 10). In the last two other nuclear loci, deserta also had the lowest number of haplotypes, along with one of the other two taxa (Table 4.1). Population history Haplotype and nucleotide diversity can be used to interpret demographic history of populations (Grant & Bowen, 1998). The pattern observed in madeira and feae mtDNA, i.e. high h (>0.5) and low π (<0.5%) is suggestive of a bottleneck followed by a rapid population expansion. Conversely, deserta shows low h and low π, indicating rather a recent population bottleneck or a founder event by a few mitochondrial lineages (Grant & Bowen, 1998). In addition, in madeira, all neutrality tests have negative values, which are significant in three markers (Table 1), thus suggesting past population growth or positive selection. Similarly, non-significant SSD values and low values of r also indicate a population growth. The situation is a bit more contrasted in feae. For 118 this taxon, half of neutrality tests values are significant supporting the hypothesis of population growth; other values are however negative. Mismatch analysis indices, as in madeira, all indicate a population growth. In deserta, although all negative, a majority of Fu’s Fs and Tajima’s D values are not significant. Furthermore, raggedness index’s high values suggest a stable population rather than an increasing one, while non-significant SSD values do not contradict the hypothesis of population growth. Thus in this taxon, the signal is not clear and this could reflect a period of low population followed only very recently by a population increase. Population structure In both mitochondrial genes, exact test of population differentiation indicated a significant differentiation at the global level and for each pairwise comparisons (all P<0.001). In addition, population structure was detected with pairwise Fst values among each pairs of taxa (all Fst>0.9; all P<0.001) indicating a pattern of geographic differentiation. Indeed, most of the genetic variation was observed between populations rather than within population, the latter accounting for only 8.7% and 3.7% of the total variation in CO1 and Cytb respectively (Table 4.2). These results are fully consistent with previous analyses carried out on Cytb only (Zino et al., 2008, Jesus et al., 2009), though obtained on a much larger sample size (c.4-5 times) and with two, rather than a single gene. Results obtained with nuclear introns contrasted strongly, however, with those obtained with mtDNA. Out of the three introns, only one, βFibint7, showed some geographic structure. However, although significant differentiation was detected using Raymond & Rousset test at the global level (i.e. all populations together; P<0.001), only feae actually differed from the other two forms (both P<0.001). Similarly, pairwise Fst values between feae and the two other two taxa were significant (Fst=0.24 and 0.14 with deserta and madeira respectively; both P<0.001). But in contrast to Cytb and CO1 results, more than 80% of genetic variation was found within populations rather than between taxa in βFibint7 (results of AMOVA, see Table 4.2). Furthermore, neither PAXIP1 nor CSDE1 showed differentiation (exact test of population differentiation: P=0.2 and P=0.49, respectively), and only one pairwise Fst value was found significant in PAXIP1, between deserta and madeira (Fst=0.2; P<0.05), while none were significant in CSDE1 (all Fst=0.0). Similarly to βFibint7, more than 90% of genetic variation is detected within populations by the MANOVA in PAXIP1 and CSDE1 (Table 4.2). 119 βFibint7 PAXIP1 CSDE1 Cytb CO1 Table 4.1 Diversity and neutrality estimates for three Macaronesian Pterodroma taxa. Number of sequences (N), number of polymorphic sites (Np), number of haplotypes (Nh), haplotype diversity (h), nucleotide diversity (π, expressed as percentages, i.e. 0.001=0.1%), average number of nucleotide differences (k). Significant values (P<0.05; P<0.02 for Fu’s Fs) are in bold r SSD Taxon N Np Nh h π% k Fu’s Fs Tajima’s D P. deserta 89 5 4 0.474 0.07 0.534 -0.305 -0.714 0.198 0.022 P. feae 58 10 9 0.687 0.15 1.095 -3.426 -0.886 0.078 0.003 P. madeira 57 11 9 0.778 0.23 1.683 -1.616 -0.475 0.04 0.007 P. deserta 94 5 6 0.219 0.03 0.229 -5.552 -1.473 0.367 0.002 P. feae 59 11 11 0.556 0.11 0.963 -6.606 -1.679 0.052 0.0001 P. madeira 59 12 11 0.535 0.09 0.788 -8.01 -1.929 0.086 0.001 P. deserta 12 0 1 0 0 0 NA 0 NA NA P. feae 10 0 1 0 0 0 NA 0 NA NA P. madeira 12 3 4 0.45 0.1 0.5 -2.124 -1.63 0.009 0.153 P. deserta 10 4 4 0.53 0.3 1.4 -0.175 -0.038 0.382 0.055 P. feae 14 15 10 0.95 0.7 3.56 -3.833 -0.999 0.076 0.016 P. madeira 14 12 8 0.87 0.5 2.58 -2.427 -1.257 0.05 0.008 P. deserta 154 1 3 0.026 0.003 0.026 -4.696 -0.904 0.9 0.0 P. feae 56 15 9 0.67 0.17 1.654 -2.558 -1.486 0.074 0.003 P. madeira 58 2 3 0.07 0.007 0.069 -3.239 -1.448 0.75 0.00003 120 Table 4.2 Genetic variation within and between the three taxa of Macaronesian Pterodroma petrels (AMOVA) PAXIP1 CSDE1 βFibint7 Cytb CO1 d.f 2 Sum of Squares 704.787 Variance components 5.3 Percentage of variation 91.28 201 101.835 0.51 8.72 2 1079.63 7.85 96.39 209 61.407 0.29 3.61 2 3.62 0.022 17.4 263 28.446 0.108 82.96 Among populations 2 0.162 0.00 0 Within populations 31 2.75 0.09 100 Among populations 2 5.324 0.11 7.5 Within populations 35 46.229 1.32 92.5 Among populations Within populations Among populations Within populations Among populations Within populations Phylogenetic relationships The model of nucleotide substitution that best fitted the data (using jModeltest) differed between the different markers: Hasegawa, Kishino and Yano model (Hasegawa et al., 1985) with a gamma-distributed rate variation across sites (HKY+G) for Cytb, CO1, and βFibint7, HKY model for CSDE1, and F81 model for PAXIP1. In both mitochondrial markers the phylogenetic trees (Fig. 4.2) recovered the three taxa in two well supported clades: one containing deserta and feae, the other containing madeira. In CO1 this latter taxon presents a well supported sub-clade that is not present in Cytb. The three taxa are closest to Bermuda petrel (P. cahow) then to Black-capped (P. hasitata) and soft-plumaged (P. mollis) petrels (see also Jesus et al. 2009). The haplotype network analyse on Cytb and CO1 clearly indicates complete lineage sorting between the three taxa (Fig. 4.3). Conversely, as expected from the lack of population structuring and/or genetic diversity for nuclear introns, the BI phylogenetic analyses failed to recover monophyletic taxa with any of the 121 three introns studied (Fig. 4.2). All three taxa are mixed together in a large polytomy, although a well supported sub-clade made of feae and madeira specimens is present in PAXIP1. In the haplotype networks (Fig. 4.3) lineages are not sorted and all share a dominant haplotype. In PAXIP1 madeira and deserta do not share any haplotype other than the dominant one found in all three taxa, while on the other hand, feae shares haplotypes with the two other taxa. 122 Figure 4.2. Phylogenetic tree of Macaronesian petrels with five markers. A CO1; B Cytb; C βFibint7; D PAXIP1; E CSDE1; for clarity, in trees A, B and C not all sequences are shown, thus number of leaves is inferior to number of sequences indicated in Table 4.1; trees D and E show all sequences. Nodes with Posterior Probabilities <95% are collapsed. Posterior Probabilities >95% are shown (with exception of one node in CO1 tree with 94% support) 100 P hasitata 100 P mollis mollis P hasitata P cahow 99 98 99 100 100 100 100 96 94 100 D E 99 100 100 99 95 98 100 98 97 100 100 100 100 C A B feae deserta madeira 123 Figure 4.3. Haplotype networks obtained with CO1, Cytb, PAXIP1, βFibint7 and CSDE1 loci with Macaronesian Pterodroma petrels. Size of circles are proportional to number of individuals possessing this haplotype. Green: haplotypes from Madeira, madeira; Red: haplotypes from Bugio, deserta; Blue: haplotypes from Cape Verde, feae. 124 Gene flow, population divergence time and effective population sizes No gene flow was detected between taxa, i.e. distributions of migration rate peaked at zero, hence no significant population migration rate was detected by the likelihood ratio test (Nielsen & Wakeley 2001) in any pairs of taxa. Divergence times between pairs of population and population sizes estimates are summarised in Table 4.3. All population divergences are estimated as having occurred within the Pleistocene: despite large 95% Highest Posterior Density Intervals (95%HPD), all estimates are still below one million year ago. The oldest estimated divergence time concerns the split feae/madeira (around 290,000 years ago). Using the topology obtained with mitochondrial genes (Fig. 4.2) on the three taxa rather than pairs provided very similar results: no significant migration rate is detected between any taxa, the split feae/deserta is estimated around 37,000 years ago (95%HPD 13,000-99,000), and the differentiation between the ancestral feae/deserta and madeira occurred about 110,000 years ago (95%HPD 46,000-222,000; Fig. 4.4). Populations of the three extant Macaronesian petrel taxa have estimated sizes comprised between two and three thousands individuals (Table 4.4), which is lower than estimates found when analysing pairs of taxa (Table 4.3). Table 4.3 Result from Isolation with Migration analysis with IMa2. 95%HPD intervals are given in brackets. Population estimates are for effective population Ne. “Population 1” refers to the first taxon of the pair (in the first column), “population 2” to the second taxon Divergence time (Million years) 0.29 [90 656 – 648 542] 0.092 Madeira/Deserta [18900 – 324538] 0.058 Feae/Deserta [14406 – 204562] Madeira/Feae Population 1 Population 2 Ancestral population 6363 [2354 – 14 964] 4457 [1528 – 11 000] 5202 [1681 – 13152] 8297 [3261 – 18 887] 3793 [1174 – 9652] 3228 [880 – 8777] 13 000 [4358 – 89 914] 23 000 [10 700 – 74 600] 11 000 [2401 – 30786] Table 4.4 Population size estimates obtained with IMa2 for the three populations of Macaronesian petrels when using the mitochondrial gene trees topology Population size estimation (95% highest posterior density interval) Deserta Feae Madeira Ancestral Feae/Deserta Ancestral population 1949 (573 – 5160) 3096 (1032 - 7224) 2179 (1032 - 5848) 7912 (573 – 206 300) 6307 (1261 – 16627) 125 Figure 4.4 Population divergence estimation under IM model with IMa2 for the three populations of Macaronesian petrels. Histories for all three population pairs are represented as boxes (for sampled and ancestral populations), horizontal lines (for splitting times). Time is represented on the vertical axis in each figure, with the sampled species and subspecies names given at the top of each figure at the most recent time point. 95% highest posterior density intervals are shown with arrows in gray for population sizes (i.e., box widths; horizontal arrows) and splitting times (dotted lines; vertical arrows) 126 Discussion Gadfly petrels from Macaronesia have already been the subject of genetic study (Zino et al. 2008, Jesus et al. 2009), although with reduced sample sizes, and with only one mtDNA marker compared to five in this study (including nuclear markers). Main results are as follow: nucleotide diversity and haplotype diversity are unexpectedly high given the current population sizes of all three taxa (between 80 and 250-500 breeding pairs; see Shirrihai et al., 2010). Nucleotide diversity is much lower in deserta than in the other two taxa, despite the fact that current population size of deserta is supposed to be at least 3-5 times larger than, e.g., madeira. In addition, we found that nuclear genes provided a strikingly different picture compared to mtDNA, the latter two genes under study providing a rather similar pattern (though not identical). In contrast to previous statements, our results suggest that madeira is the ancestral taxon, not feae. Finally, our times of divergence between any pairs of taxa are about 10 times smaller than previous estimates, one being estimated less than 40,000 years ago. Our study therefore contradicts almost all previous statements based only on one mtDNA marker and reduced sample sizes. Below we discuss these findings and their implications. Genetic diversity High haplotypic diversity in various mitochondrial genes has been described in a wide range of Procellariiformes species, despite most are currently declining, from albatrosses (Abbott & Double, 2003; Burg & Croxall, 2001,2004) to giant petrels (Macronectes spp.; Techow et al., 2010) and White-chinned petrels (Procellaria aequinoctialis; Techow et al., 2009). In two recent studies of currently bottlenecked populations of Procellariiformes, Lawrence et al. (2008a, b) in Taiko petrel (Pterodroma magentae), and Kuro’o et al. (2010) in Short-tailed albatross (Phoebastria albatrus) found high genetic diversity, while the reverse was expected given the extreme bottleneck faced by these populations (Spielman, 2004; but see Miller & Waits, 2003). The explanation provided in these two species was that current high genetic diversity (despite very low numbers) was linked to high diversity prior to the bottleneck. In addition, these seabirds are long-lived, and the bottlenecks were of relatively short durations hence did not have time to affect genetic diversity. The three Macaronesian petrels also currently show unexpectedly high genetic diversity given their low population sizes. Since they are also long-lived, and since our data support a population growth in the recent past, we suggest that the level of genetic diversity reflects previous levels, i.e. before the bottleneck they are now going through. However genetic diversity found in this study are 127 moderate, and even low in the case of deserta, which requires some explanation. We suspect that this taxon experienced a serious bottleneck and/or a founder event by a very low number of individuals, as shown by the pattern of haplotypic and nucleotide diversity (see Grant & Bowen, 1998). Since then, either the bottleneck period was long, or not enough time elapsed since the bottleneck event, to recover genetic diversity observed in other Procellariiformes taxa. It should be noted that the islands of Madeira and Cape Verde were only recently colonised (17th century) by man, so the bottleneck experienced by deserta must have been much before human colonisation. In this context, the former presence of subfossils of Pterodroma attributed to feae on several islands of northern Europe (Serjeantson, 2005; see Robb 2008 for a review) may indicate a former breeding range much larger than the actual one, which could explain its genetic diversity. However, given the morphological similarities between madeira, feae and deserta, it is almost impossible to ascertain which of these taxa these fossils or historical records refer to until genetic investigations are carried out with ancient DNA to complete past genetic diversity and phylogeographic history of these populations (Lawrence et al., 2008a, b; Steeves et al, 2010). Population structure Lack of population structure and reciprocal monophyly in nuclear intron, as opposed to strong structuration, lineage sorting and monophyly in mitochondrial markers such as we found, although striking, is not a unique case in birds (Zink & baroowclough, 2008). Lee & Edwards (2009) showed that sister species red-backed (Malurus melanocephalus) and white-winged (Malurus leucopterus) fairy wrens in Australia are still sharing their nuclear genes, but not their mtDNA. These authors argued that such pattern can be expected to be common for bird species that recently diverged or had large ancestral populations. This can be due to the differences in coalescence time for nuclear and mitochondrial loci (Moore, 1995; Palumbi et al., 2001) and their different effective population sizes (Zink & Barrowclough, 2008) leading to incomplete lineage sorting. Our data suggest that ancestral populations were not especially high, but divergence times are extremely recent, thus giving credit to this hypothesis. Alternatively, the discrepancy between mitochondrial and nuclear markers in macaronesian petrels could also be explained by extreme female philopatry associated with a nuclear gene flow through males (Zink & Barrowclough, 2008). However male-biased dispersal has never been documented in petrels and, in birds in general, female dispersal is dominant (Newton, 2003), so this explanation is unlikely. In addition, estimating potential gene flow between the three petrel populations indicated the absence of migration between populations. In consequence, it seems reasonable to explain the lack of reciprocal monophyly and 128 population structure in nuclear introns by an incomplete sorting of nuclear lineages due to very recent divergence of macaronesian petrels populations Population divergence North-Eastern Atlantic petrels apparently diverged in the last 150,000 years. These estimates are about 10-15 times younger than previous published estimates (Zino et al., 2008, Jesus et al. 2009). Therefore the use of nuclear loci, in addition to the more traditional cytb, resulted in un (or less) biased estimates of divergence time, and argue for a necessary integration of nuclear loci in phylogeographic studies (see also Lee & Edwards, 2008, Edwards & Bensch, 2009). This fairly recent divergence may also explain, to some extant, the incredible similarity between the three taxa: contrary to fairy wrens that exhibit marked phenotypic differences despite their recent divergence (Lee & Edwards, 2008), in the three Pterodroma studied here, biometric differences occur on average but with some overlap, and they are not diagnosable at sea (Shirihai et al., 2010; see also Bretagnolle, 1995; Zino et al., 2008 and Jesus et al., 2009 for details on morphology and colourations in these taxa). Such similarity between divergent lineages can be maintained by selection when populations are facing similar ecological conditions (Price, 2008). Furthermore, because these are nocturnal, burrowing species, visual cues such as morphology or colouration probably do not play a role in population differentiation processes in these petrels. However, even in calls these populations do not show significant differences despite the importance of this trait as pre-mating isolation mechanism in burrowing petrels (Bretagnolle, 1995, 1996). Alternative hypotheses to explain population differentiation and absence of gene flow in these populations could possibly involve olfactory capabilities of Procellariiformes, breeding allochrony and non-physical barriers to dispersion. While it has been shown that petrels show odour preferences towards their kin/breeding partner (Bonnadonna & Nevitt, 2004; Mardon & Bonnadonna, 2009), the potential role of olfaction in population differentiation remains to be described. Breeding allochrony on the other hand has been shown to be an important isolating mechanism in seabirds (Monteiro & Furness 1998; Smith & Friesen 2007). The three taxa studied here are allochronic breeders: feae lay in December-January (Bannerman & Bannerman, 1968; Cramp & Simmons, 1977; Jesus et al., 2009), deserta lay in July-August (Bourne, 1957; Zino & Zino, 1986; Zino et al., 2008; Jesus et al., 2009) and madeira lay in May-June (Bourne, 1957; Zino & Zino, 1986; Zino et al., 2008), possibly in relation to adaptation to local oceanographic conditions. Friesen (2007), Steeves et al. (2005) and Gomez-Diaz et al. (2006) identified nonphysical barriers to gene flow as playing a role in population differentiation in seabirds. It should be noted that the last 150,000 years (the time period during which all splits occurred) have been 129 marked by a succession of glacials/interglacials events that were accompanied by important sea level and temperature fluctuations that affected oceanic conditions such as thermohaline circulation (Lambeck et al., 2002; Bintanja et al., 2005). Glaciation events and sea level fluctuations in the Pleistocene were already shown to have a strong impact on population structure and also led to speciation in several high latitude seabirds, such as penguins (Ritchie et al., 2004), skuas (Ritz et al., 2008) and Giant petrels (genus Macronectes; Techow et al., 2010). In the order Procellariiforme, Pleistocene variations in climatic conditions have been repeatedly outlined in population structuring and differentiation (e.g., Austin et al., 2004; Cagnon et al., 2004; GomezDiaz et al., 2006). In our study, population split estimates (i.e. 110,000 and 37,000) fall during two periods of climatic cooling and oceanic changes : about 120,000 years ago the earth experienced a warmer than present interglacial, followed by the last major glaciation whose inception stage started around 115,000 years ago, a period that saw temperatures drop dramatically (Bintanja et al., 2005). Later, other cycles of small glacial/interglacial occurred that weer accompanied by several periods of high ice-cap instability linked with sea water temperature drop and changes in oceanic conditions (Heinrich events). One of these occurred between 35,000 and 40,000 years ago (Lambeck et al., 2002). However, this seemingly coincidence in divergent times and climatic major changes must be relativised by the wide confidence intervals in our estimates. Nevertheless, given the cycles of glacial/interglacial periods and number of Heinrich events in the last 150,000 years (Lambeck et al., 2002), it is likely that taxa divergence occurred concomitantly with one of these climatic instability event. Because fossil data are missing and because seabirds, contrary to terrestrial taxa, can be extirpated and recolonise an island many times depending on environmental conditions, establishing a scenario of colonisation of the Eastern Atlantic archipelagos is difficult. However, our data impose a complete re-evaluation of the generally accepted phylogeographic scenario explaining the presence and distribution of the macaronesian petrel taxa. Previous to 115,000 years ago, a petrel population, ancestral to the three currently occurring taxa, was present in the NorthEastern Atlantic (whether it was on all, only parts, or none of the Macaronesian archipelagos can not be ascertained due to lack of fossil evidence). Following the climatic change with sea level and temperature drops starting around 115,000 years ago, this ancestral population split into two populations, one of these being the current population of Madeira (though again, it could have been breeding elsewhere at that time), the other the ancestral population on Cape Verde. These two populations probably faced different oceanic and climatic conditions, and thus migration between the two populations ceased, allowing their differentiation by adaptation to local environmental conditions and/or drift. Hence, following the cold climate episode, these two populations would have sufficiently diverged to stay separated despite potential secondary contact. Population 130 expansion detected in the genetic history of these birds can then be explained by an expansion during one of the warm climatic events that followed the split. During this range expansion, we propose that Bugio would have been colonised by a very few birds from Cape Verde (hence lower diversity in the former). During the cold climatic event characterised by cold North Atlantic waters and ice-rafted debris in North and South Atlantic around 38,000 years ago (Lambeck et al., 2002), these two populations became isolated and diverged. However, the estimates of Ne by IMa2 (c. 2,000) contradict this view, although the confidence interval is large. An alternative possibility would be that the ancestral feae/deserta population was highly structured geographically before the split, and the observed poor diversity in deserta could be the result of this structure: while feae conserved the diversity of several sub-populations, deserta conserved the diversity of a single population. A last hypothesis could be that deserta experienced a bottleneck after the split with feae, thus reducing the genetic diversity in this population. Our proposed scenario must remain, at this time, highly speculative: it is nonetheless more plausible than the supposed scenario (twin colonisation of Madeira archipelago from Cape Verde), which conflicts with our genetic evidence. In conclusion, we have seen that the use of nuclear intron sequence data in addition to mitochondrial data revealed a population history considerably more complex than what was originally though by using mitochondrial DNA only. The use of nuclear loci and Isolation with Migration model allowed us to re-examine the divergence pattern of these three populations of seabird in a new light and reinforced the call for the role of multilocus statistical phylogeography investigations to describe complex patterns of population differentiation. Aknowledgements We would like to thank Annie Tillier from the Service de Systématique Moléculaire (MNHN, Paris) for her help during the laboratory work. This work was supported by the "Consortium National de Recherche en Génomique", and the "Service de Systématique Moléculaire" of the Muséum National d'Histoire Naturelle (CNRS UMS 2700). It is part of the agreement n°2005/67 between the Genoscope and the Muséum National d'Histoire Naturelle on the project "Macrophylogeny of life" directed by Guillaume Lecointre". 131 References Abbott C. and Double M., 2003. Phylogeography of Shy and White-capped Albatrosses inferred from mitochondrial DNA sequences: implications for population history and taxonomy. Molecular Ecology 12: 2747–2758 Austin J.J., 1996. Molecular phylogenetics of Puffinus shearwaters: preliminary evidence from mitochondrial cytochrome b gene sequences. Molecular Phylogenetics and Evolution 6(1): 77-88 Austin J.J., Bretagnolle V. and Pasquet E., 2004. Aglobal molecular phylogeny of the small Puffinus shearwaters and implications for systematics of the Little-Audubon’s shearwater complex. The Auk 121(3): 847-864 Avise J.C., Arnold J., Ball R.M. et al., 1987. Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Systematics, 18, 489–522 Bandlet H.J., Forster P. and Röhl A., 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16, 37–48 Bannerman D.A. and Bannerman W.M., 1968. History of the Birds of the Cape Verde Islands. Birds of the Atlantic Islands, Vol. IV. Edinburgh: Oliver and Boyd. Bazin E., Glémin S. and Galtier N., 2006. Population size does not influence mitochondrial genetic diversity in animals. Science 312:570–572 Bintanja R., van de Wal R.S.W. and Oerlemans J., 2005. Modelled atmospheric temperatures and global sea levels over the past million years. Nature 437: 125-128 BirdLife International, 2010. http://www.birdlife.org accessed on 10/10/2010 Bolton M., Smith A.L., Gomez-Diaz E., Friesen V.L., Medeiros R., Bried J., Roscales J.L. and Furness R.W., 2008. Monteiro’s storm petrel Oceanodroma monteiroi: a new species from the Azores. Ibis 150: 717-727 Bonadonna F. and Nevitt G.A., 2004. Partner-specific odor recognition in an Antarctic seabird. Science 306: 835 Bourne W.R.P., 1957. Additional notes on the birds of the Cape Verde Islands. Ibis, 99: 182–190 Bourne W., 1983. The Soft-plumaged petrel, the Gongon and the Freira, Pterodroma mollis, P. feae and P. madeira. Bulletin of the British Ornithologist Club 103: 52–58 Bretagnolle V., 1995. Systematics.of the Soft-plumaged Petrel Pterodroma mollis complex: new insight from vocalizations. Ibis 137: 207–218 132 Bretagnolle, V., 1996. Acoustic communication in a group of non passerines birds, the petrels. In Ecology and Evolution of Acoustic Communication in Birds, Kroodsma D.E. & Miller E.H. eds. Cornell University Press, Ithaca, New York, USA Brooke M., 2004. Albatrosses and Petrels across the World. Oxford University Press, Oxford, UK Burg T. and Croxall J.P., 2004. Global population structure and taxonomy of the Wandering Albatross species complex. Molecular Ecology 13: 2345–2355 Cagnon C., Lauga B., Hémery G. and Mouchès C., 2004. Phylogeographic differentiation of storm-petrels (Hydrobates pelagicus) based on cytochrome b mitochondrial DNA variation. Marine Biology 145: 1247-1254 Chan K.M.A. and Levin S.A., 2005. Leaky prezygotic isolation and porous genomes: rapid introgression of maternally inherited DNA. Evolution 59(4): 720-729 Coyne J.A. and Orr H.A., 2004. Speciation. Sinauer associates, Sunderland, MA, USA Cramp S. and Simmons K.E.L. (eds), 1977. The Birds of the Western Palearctic, Vol. 1. Oxford: Oxford University Press, UK Dietzen C., Garcia-Del-Rey E., castro G.D and Wink M., 2008. Phylogenetic differentiation of Sylvia species (Aves: Passeriformes) of the Atlantic islands (Macaronesia) based on mitochondrial DNA sequence data and morphometrics. Biological Journal of the Linnean Society 95: 157–174 Domingues V.S., Santos R.S., Brito A., Alexandrou M. and Almada V.C., 2007. Mitochondrial and nuclear markers reveal isolation by distance and effects of Pleistocene glaciations in the northeastern Atlantic and Mediterranean populations of the white seabream (Diplodus sargus ,L.). Journal of Experimental Marine Biology and Ecology 346: 102–113 Edwards S.V. and Bensch S., 2009. Looking forwards or looking backwards in avian phylogeography? A comment on Zink and Barrowclough 2008. Molecular Ecology 18: 2930-2933 Edwards S.V., Kingan S.B., Calkins J.D., Balakrishnan C.N., Jennings W.B., Swanson W.J. and Sorenson M.D., 2005. Speciation in birds: genes, geography and sexual selection. Proceedings of the National Academy of Science of the USA 102:6550-6557 Emerson B.C., 2002. Evolution on oceanic islands: molecular phylogenetic approaches to understanding pattern and process. Molecular ecology 11: 951-956 Excoffier L. and Lischer H.E.L., 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10(3): 564-567 Fitzpatrick B.M., Fordyce J.A. and Gavrilets S., 2009. Pattern, process and geographic modes of speciation. Journal of Evolutionary Biology 22(11): 2342-2347 133 Friesen V.L., 2007. New roles for molecular genetics in understanding seabird evolution, ecology and conservation. Marine Ornithology 35: 89-96 Friesen V.L., Burg T.M. and McCoy K.D., 2007a. Mechanisms of population differentiation in seabirds. Molecular Ecology 16:1765-1785 Friesen V.L., Gonzalez J.A. and Cruz-Delgado F., 2006. Population genetic structure and conservation of the Galapagos petrel (Pterodroma phaeopygia). Conservation Genetics 7: 107-115 Friesen V.L., Smith A.L., Gómez-Díaz E., Bolton M., Furness R.W., González-Solís J. and Monteiro L.R., 2007b. Sympatric speciation by allochrony in a seabird. Proceedings of the National Academy of Sciences USA 104: 18589-18594 Fu Y.X., 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147: 915–925 Funk D.J. and Omland K.E., 2003. Species-level paraphyly and polyphyly: frequency, causes, and consequences, with insights from animal mitochondrial DNA. Annual Review of Ecology, Evolution, and Systematics 34: 397–423 Gomez-Diaz E., Gonzales-Solis J., Peinado M.A. and Page R.D.M., 2006. phylogeography of the Calonectris shearwaters using molecular and morphometric data. Molecular Phylogenetics and Evolution 41: 322-332 Grant W.S. and Bowen B.W., 1998. Shallow population histories in deep evolutionary lineages of marine fishes: insights from Sardines and Anchovies and lessons for conservation. Journal of Heredity 89(5): 415-426 Guindon S. and Gascuel O., 2003. A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood". Systematic Biology 52: 696-704 Harpending R., 1994. Signature of ancient population growth in a low-resolution mitochondrial mismatch distribution. Human Biology 66: 591–600 Hasegawa M., Kishino H and Yano T., 1985. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. Journal of Molecular Evolution 22: 160-174 Heled J. and Drummond A.J., 2010. Bayesian inference of species trees from multilocus data. Molecular Biology and Evolution 27(3): 570-580 Hey J. 2010. Isolation with migration model for more than two populations. Molecular Biology and Evolution 27(4): 905-920 Huelsenbeck J.P. and Ronquist F., 2001. MRBayes: Bayesian inference of phylogeny. Bioinformatics 17: 754–755. Jennings W.B. and Edwards S.V., 2005. Speciational history of Australian grass finches (Poephila) inferred from thirty gene trees. Evolution 59: 2033–2047 134 Jesus J., Menezes D., Gomes S., Oliveira P., Nogales M. and Brehm A., 2009. Phylogenetic relationships of gadfly petrels Pterodroma spp.from the Northeastern Atlantic Ocean: molecular evidence for specific status of Bugio and CapeVerde petrels and implications for conservation. Bird Conservation International 19: 1-16 Jouventin P., Mouret V. and Bonadonna F., 2007.Wilson’s storm petrels Oceanites oceanicus recognise the olfactory signature of their mate. Ethology 113: 1228–1232 Kimball R.T. et al., 2009. A well-tested set of primers to amplify regions spread across the avian genome. Molecular Phylogenetics and Evolution 50: 654-660 Kuro’o M., Yonekawa H., Saito S, Eda M., Higushi H., Koike H. and Hasegawa H., 2010. Unexpectedly high genetic diversity of mtDNA control region through severe bottleneck in vulnerable albatross Phoebastria albatrus. Conservation Genetics 11: 127-137 Lambeck K., Esat T.M. and Potter E-K., 2002. Links between climate and sea levels forthe last three millions years. Nature 419: 199-206 Lawrence H.A., Scofield R.P., Crockett D.E., Millar C.D. and Lambert D.M., 2008a. Ancient genetic variation in one of the world’s rarest seabird. Heredity 101: 543-547 Lawrence H.A., Taylor G.A., Millar C.D. and Lambert D.M., 2008b. High mitochondrial and nuclear genetic diversity in one of the world’s most endangered seabirds, the Chatham Island Taiko (Pterodroma magentae). Conservation Genetics 9: 1293-1301 Lee J.Y. and Edwards S.V., 2008. Divergence across Australia’s Carpentarian barrier: statistical phylogeography of the Red-backed Fairy Wren (Malurus melanocephalus). Evolution 62: 3117–3134 Librado P. and Rozas J., 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25(11): 1451-1452 Liggins L., Chapple D.G., Daugherty C.H. and Ritchie P.A., 2008. Origin and post colonization evolution of the Chatham Islands skink (Oligosoma nigriplantare nigriplantare). Molecular Ecology 17:3290-3305 Losos J.B. and Ricklefs R.E., 2009. Adaptation and diversification on islands. Nature 457: 830-836 Mardon J. and Bonadonna F., 2009. Atypical homing or self-odour avoidance? Blue petrels (Halobaena caerulea) are attracted to their mate’s odour but avoid their own. Behavioral Ecology and Sociobiology 63(4): 537-542 Miller C.R. and Waits L.P., 2003.The history of effective population size and genetic diversity in the Yellowstone grizzly (Ursus arctos): implications for conservation before genetic factors impact them. Proceedings of the National Academy of Science USA 100: 4334–4339 135 Monteiro L.R., Ramos J.A. and Furness R.W., 1996. Past and present status and conservation of the seabirds breeding in the Azores archipelago. Biological Conservation 78: 319-328 Monteiro L.R. and Furness R.W., 1998. Speciation through temporal segregation of Madeiran storm-petrel (Oceanodroma castro) populations in the Azores? Philosophical Transactions of the Royal Society London 353: 945–953 Moore W.S., 1995. Inferring phylogenies from mtDNA variation: mitochondrial-gene trees versus nuclear-gene trees. Evolution 49: 718–726 Moya O., Contreras-Diaz H.G., Oromi P. and Juan C., 2004. Genetic structure, phylogeography and demography of two ground-beetle species endemic to the Tenerife laurel forest (Canary Islands). Molecular Ecology 13: 3153-3167 Newton I., 2003. The speciation and biogeography of birds. Academic Press, San Diego, CA., USA. Nielsen R. and Wakeley J., 2001. Distinguishing migration from isolation. A Markov chain Monte Carlo approach. Genetics 158: 885–896 Nunn G.B. and Stanley S.E., 1998. Body size effects and rates of cytochrome b evolution in tube-nosed seabirds. Molecular Biology and Evolution 15:1360–1371 O’Dwyer T.W., Ackerman A.L. and Nevitt G.A., 2008. Examining the development of individual recognition in a burrow-nesting procellariiform, the Leach’s storm-petrel. Journal of Experimental Biology 211: 337–340 Palumbi S.R., Cipriano F. and Hare M.P., 2001. Predicting nuclear gene coalescence from mitochondrial data: the three-times rule. Evolution 55: 859–868 Pereira S.L. and Baker A.J., 2006. A mitogenomic timescale for birds detects variable phylogenetic rates of molecular evolution and refutes the standard molecular clock. Molecular Biology and Evolution 23(9):1731–1740 Pieper H., 1985. The fossil land birds of Madeira and Porto Santo. Bocagiana 88: 1-6 Posada D., 2008. jModelTest: Phylogenetic Model Averaging. Molecular Biology and Evolution 25(7): 1253-1256 Price T.D., 2008. Speciation in Birds. Roberts & Company Publishers, Greenwood village, Colorado, USA Prychitko T. M. and Moore W. S., 1997. The utility of DNA sequences of an intron from the βfibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Picidae). Molecular Phylogenetics and Evolution 8: 193–204 Ramos-Onsins S.E. and Rozas J., 2002. Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution 19: 2092–2100 136 Raymond M. and Rousset F., 1995. An exact test for population differentiation. Evolution 49(6): 1280-1283 Ritchie P.A., Millar C.D., GibbG.C., Baroni C. and Lambert D.M., 2004. Ancient DNA enables timing of the Pleistocene origin and Holocene expansion of two Adelie Penguin lineages in Antarctica. Molecular Biology and Evolution 21: 240–248 Ritz M.S., Millar C., Miller G.D., Phillips R.A., Ryan P., Sternkopf V., Liebers-Helbid D. and Peter H-U., 2008. Phylogeography of the southern skua complex – rapid colonization of the southern hemisphere during a glacial period and reticulate evolution. Molecular Phylogenetics and Evolution 49(1): 292-303 Robb M. and Mullarney K., 2008. Petrels Night and Day. Poole Ronquist F. and Huelsenbeck J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19(12): 1572-1574 Ronquist F., Huelsenbeck J.P. and Van der Mark P., 2005. MrBayes 3.1 Manual. http://mrbayes.csit.fsu.edu/. Accessed March 2010 SaPinto A., Branco M., Sayanda D. and Alexandrini P., 2008. Patterns of colonisation, evolution and gene flow in species of the genus Patella in the Macaronesian Islands. Molecular Ecology 17:519-532 Schneider S. and Excoffier L., 1999. Estimation of Past Demographic Parameters From the Distribution of Pairwise Differences When the Mutation Rates Vary Among Sites: Application to Human Mitochondrial DNA. Genetics 151: 1079-1089 Schneider S., Roessli D. and Excoffier L., 2000. In: Genetics and Biometry Lab. Dept. of Anthropology, University of Geneva Scott D., Scofield P., Hunter C. and Fletcher D., 2008. Decline of sooty shearwaters, Puffinus griseus, on the Snares, New Zealand. Papers and Proceedings of the Royal Society of Tasmania 142(1): 185-196 Serjeantson D., 2005. Archaeological records of a gadfly petrel Pterodroma sp. from Scotland in the first millennium AD. in Feathers, grit and symbolism; Birds and humans in the ancient Old and New Worlds. Proceedings of the 5th meeting of the ICAZ bird working group in Munich (26.7.-28.7.2004), Grupe and Peters editors, Verlag Marie Leidorf GmbH, Germany Shaw K.L., 2002. Conflict between nuclear and mitochondrial phylogenies of a recent species radiation: what mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. Proceedings of the National Academy of Science of the USA 99: 16122-16127 Shirihai H., Bretagnolle, V. and Zino F., 2010. Identification of Feae’s, Desertas and Zino’s petrels at sea. Birding World, 23(6) in press 137 Simon C., Frati F., Beckenbach A., Crespi B., Liu H. and Flook P., 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Annals of the Entomological Society of America 87(6): 651-701 Smith A.L. and Friesen V.L., 2007. Differentiation of sympatric populations of the bandrumped storm petrel in the Galapagos Islands: an examination of genetics, morphology and vocalizations. Molecular Ecology 16: 1593-1603 Spinks P.Q. and Shaffer H.G., 2009. Conflicting Mitochondrial and Nuclear Phylogenies for the Widely Disjunct Emys (Testudines:Emydidae) Species Complex, and What They Tell Us about Biogeography and Hybridization. Systematic Biology 58(1): 1-20 Spielman D., Brook B.W. and Frankham R., 2004. Most species are not driven to extinction before genetic factors impact them. Proceedings of the National Academy of Science USA 101: 15261–15264 Stephens M., Smith N. and Donnelly P., 2001. A new statistical method for haplotype reconstruction from population data. American Journal of Human Genetics 68: 978–989 Steeves T.E., Holdaway R.N., Hale M.L., McLay E., McAllan I.A.W., Christian M., Hauber M.E. and Bunce M., 2010. Merging ancient and modern DNA: extinct seabird taxon rediscovered in the north Tasman Sea. Biology Letters 6(1): 94-97 Steeves T.E., Anderson D.J. and Friesen V.L., 2005. A role for non physical barriers to gene flow in the diversification of a very vagile seabird, the masked booby (Sula dactylatra). Molecular Ecology 14: 3877-3887 Sullivan J, Joyce P (2005) Model selection in phylogenetics. Annual Review of Ecology Evolution and Systematics 36:445-466 Tajima F., 1989. The effect of change in population size on DNA polymorphism. Genetics 123: 597–601 Tamura K., Dudley J., Nei M. and Kumar S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24(8): 15961599 Techow N.M.S.M., O’Ryan C., Phillips R.A. et al., 2010. Speciation and phylogeography of giant petrels Macronectes. Molecular Phylogenetics and Evolution 54: 472-487 Techow N.M.S.M., Ryan P.G. and O’Ryan C., 2009. Phylogeography and taxonomy of Whitechinned and Spectacled petrels. Molecular Phylogenetics and Evolution 52: 25-33 Thompson J.D., Higgins D.G. and Gibson T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22(22): 4673-4680 138 Weir J.T. and Schluter D., 2008. Calibrating the avian molecular clock. Molecular Ecology 17: 2321-2328 Zink R.M. and Barrowclough G.F., 2008. Mitochondrial DNA under siege in avian phylogeography. Molecular Ecology 17: 2107-2121 Zino F., Brown R. and Biscoito M., 2008. The separation of Pterodroma madeira (Zino’s petrel) from Pterodroma feae (Feae’s petrel) (Aves: Procellariidae). Ibis 150: 326-334 139 Gould’s petrel ©”Chris” southportpelagics.wildiaries.com 140 CHAPTER 5: PHYLOGEOGRAPHY OF GOULD’S PETREL (Pterodroma leucoptera) and PRELIMINARY TAXONOMIC INVESTIGATIONS IN GOULD’S (Pterodroma leucoptera) AND COLLARED PETREL (Pterodroma brevipes) . 141 Part 1: Phylogeography of a threatened seabird taxon, the Gould’s petrel (Pterodroma leucoptera) Gangloff B et al. in prep Abstract Gould’s petrel (Pterodroma leucoptera) is a small gadfly petrel breeding in Eastern Australia and New Caledonia, whose two populations are considered subspecies. It is classified as Vulnerable and populations are supposed in decline in New Caledonia due to introduced predators. In Australia, the population survives only through intensive conservation management on the two breeding islands. Subspecies were defined on phenotypical criteria (morphology, colouration, breeding range and dates) but their phylogenetic relationships are unknown. Information on genetic relatedness and structure is important for the conservation of both populations. In this study we used five markers (2 mitochondrial and 3 nuclear) to investigate whether these populations are genetically isolated from each other. Lineage sorting, was not found in any marker. Furthremore, estimation of gene flow indicated that these populations exchanged migrants after splitting, although this flux was not significant. Despite lack of lineage sorting, a weak but significant population structure was found in nuclear markers but not in mitochondrial loci. This study shows tht the two Gould’s petrel populations are not genetically independent from ech other. This needs to be taken into account when devising conservation strategies for the two taxa. Keywords mtDNA, nucDNA, Pterodroma, petrel, New caledonia, Australia, CO1, Cytb, intron, phylogeography 142 Introduction The currently occurring 6th mass extinction, that affects all types of organisms, combined with the period of global change that the biosphere is entering into give more and more importance to the field of Conservation Biology in the starting 21st century (Hedrick, 2001). This field aims at protecting biological diversity and the processes that sustain it (Moritz, 2002) and is a multidisciplinary field of which conservation genetics and molecular phylogenetics are important components. In the last 20 years, the rapid progresses of molecular biology techniques have given a crucial importance to these last two fields for the long term persistence of biodiversity (Ehrlich & Wilson, 1991; Hedrick, 2001), leading Daugherty et al. (1990) to argue that “taxonomies are not irrelevant abstractions, but the essential foundations of conservation practice”. As a consequence, genetic tools are more and more used for conservation biology objectives. However, in the case of wild fragmented populations of threatened species, use of these tools is still in an early stage and biodiversity conservation at the species or subspecies level still requires identifying proper conservation and taxonomic units and the relationships between them (Frankham, 2010). When management decisions are critical for the survival of a taxon, in both the short and long terms, information on the genetic differences and demographic history, phylogenetic relationships and potential gene flow between these units is necessary to adopt appropriate management options (Miller et al., 2009). Indeed, in endangered taxa, evaluation of such parameters can help maintaining current levels of diversity (Roques & Negro, 2005) and avoid genetic diversity loss, through inappropriate conservation strategies promoting the merging of diverging genetic lineages, thus helping these taxa survive future environmental fluctuations (Bouzat, 2010). Such information are particularly important for populations that have suffered demographic decline and when interpopulation migrations might be reduced by large geographical separation and/or differences in foraging ecology or natal philopatry and breeding site fidelity (Irwin & Gibbs 2002; Martinez-Cruz et al. 2004; Pearce 2007) such as commonly observed in seabirds, particularly in family Procellariidae. Gadfly petrels of the genus Pterodroma (family Procellariidae) are pelagic seabirds showing extreme dispersal abilities and extreme philopatric behaviour (Brooke, 2004). From these two characteristics, these birds may be expected to exhibit either highly structured populations due to their breeding site fidelity, or unstructured population due to migration promoted by their dispersal abilities. It has appeared from several studies (e.g. Friesen et al., 2006, Rayner et al., 2010) that the presence of population structure might be dominant, even in the absence of physical barriers to gene flow (Friesen et al., 2007). Furthermore, 21 of the 32 recognised Pterodroma species now have an unfavourable conservation status, i.e. Vulnerable, Endangered or Critically Endangered (Birdlife 143 International, 2010) hence rendering conservation actions particularly important for the survival of these taxa and maintenance of their evolutionary history and potential. Sub-genus cookilaria (Bonaparte, 1855; Imber, 1985) comprises six to eight species (Bourne 1983, Imber 1985) breeding in the south Pacific. Only two of these breed on more than one archipelago including Gould’s petrels (Pterodroma leucoptera). This petrel comprises two discrete populations (Fig. 5.1.1) now treated as subspecies (Imber & Jenkins, 1981; de Naurois, 1978): P. leucoptera leucoptera (hereafter called leucoptera) breeding on two small islands in Western Australia, Cabbage Tree and, more recently, Boondelbah (Priddel & Carlile, 1997), the other P. l. caledonica (hereafter called caledonica) breeding in the main island of New Caledonia, in at least two mountain massifs (Bretagnolle, 2001). However, the situation is rendered more complex by the presence in the Australs archipelago of a Pterodroma colony (first discovered by R. & J. Seitre in 1989, identified by V. Bretagnolle) whose taxonomic affiliation is still to be clarified but that could belong to caledonica (V. Bretagnolle pers. obs.) and the capture in the Society archipelago of three birds identified as belonging to either P. leucoptera or to the closely related Collared petrel P. brevipes. Taxonomic identity of these birds, which could come from breeding colonies established in the Society, need to be assessed by genetic means. Gould’s petrel is classified as Vulnerable due to its small breeding range, small number of breeding locations and likely predation by introduced mammals (Birdlife International, 2010). The leucoptera population in Australia was estimated at around 200-500 breeding pairs in the 1970’s (Fullagar, 1976) and about 250 breeding pairs in the 1990’s (Priddel & Carlile 2009). Although the population possibly fluctuated in the first half of the century (Hindwood & Serventy, 1941) it remained low throughout the years. Following intensive conservation actions started in the 1990’s, breeding success was improved and the population slowly started to increase, reaching currently more than 750 breeding pairs (Priddel & Carlile, 2009). Despite this conservation success, the status of the population remains critical and further management is still needed to ensure its survival. In New Caledonia a bird was collected by McMillan in 1931, but it is only when a breeding colony was discovered by deNaurois in 1977 that the morphological and colouration differences with Australian birds were noticed. Imber & Jenkins (1981) finally raised this population to subspecies status, based on birds collected in Northern New Zealand and on the work of DeNaurois. Seven colonies of caledonica are currently known, each with an approximated 100-200 pairs, but sightings at sea suggest greater number of colonies and much greater number of birds (Villard et al., in prep.). Although morphological and colouration differences between the two forms are recognised (listed in Imber & Jenkins 1981), overlap in these two criteria does exist (Bretagnolle, unpubl. data) and the phylogenetic relationships between the two taxa remain unstudied. Given the unfavourable status of the two populations and the threats 144 they face in their respective habitats, investigation of possible gene flow and exchange of birds would bring important information for the management and conservation of the two taxa. It is now widely recognised that single-locus investigations, although having proven much utility and promoted great advances in the description of phylogenetic relationships between taxa, can sometimes lead to evolutionary diagnosis errors (Forister et al., 2008) and need to be improved by use of several loci if one wants to investigate complex evolutionary patterns that are shaping taxa’s history such as gene flow, population divergence or effective population sizes (Edwards, 2009). In this context, by means of five molecular markers, two mitochondrial genes and three nuclear introns, we investigated the molecular ecology of Gould’s petrels from New Caledonia and Australia in order to (i) infer their phylogenetic relationships and degree of divergence, (ii) investigate populations’ divergence and demography history by means of Isolation with Migration models and molecular diversity indices, (ii) check potential gene flow between the population that might prevent the two lineages from differentiating. Figure 5.1.1 Breeding locations of Gould’s and Collared petrels populations caledonica leucoptera 145 Material and Methods Sample collection and laboratory methods Feathers were collected on live adult birds in New Caledonia in a single breeding colony, in breeding season 2005-2006. They were stored in 70% ethanol then frozen at -20°C until processing. Blood samples were collected on Australian birds from 2005 to 2008. Blood was collected from the veins on the leg or wing using microcapillaries and, as for feathers, was stored in 70% ethanol then frozen at -20°C until processing. Samples of skin from birds of Society Archipelago, small pieces of skin were cut from the palm. Total genomic DNA was extracted using the DNeasy Tissue Extraction Kit (Qiagen, Valencia, CA, USA) following manufacturer’s instructions except that we increased the time of proteinase digestion to 8h overnight to ensure complete digestion. Cytochrome Oxydase 1 gene (CO1) was amplified using primers F1B 5’- AACCGATGACTATTYT-CAACC-3’ and R1B 5’-TACTACRTGYGARATGATTCC-3’, derived from primers F1/R1 (Simon et al., 1994). The PCR consisted of 37 cycles following a hot start at 94°C and a 4 minutes initial denaturation step at 94°C. Cycles, i.e. 94°C for 30s, 51°C for 40s and 72°C for 50s, were completed by a final extension at 72°C for 5 minutes. Cytochrome b gene was amplified using primers L14987 5’-TATTTCTGCTTGATGAAACT-3’ and H16025 5’-CTAGGGCTCCAATGATGGGGA-3’ (Jesus et al., 2009) and 40 PCR cycles consisting of 30s at 94°C, 50s at 58°C, 50s at 72°C. These cycles followed a 4 minutes initial denaturation step at 94°C and were completed by a final extension at 72°C for 5 minutes. Primers FIB-BI7U 5’-GGAGAAAACAGGACAATGACAATTCAC-3’ and FIB-BI7L 5’- TCCCCAGTAGTATCTGC-CATTAGGGTT-3’ (Prychitko & Moore, 1997) were used for Beta Fibrinogene intron 7 (βFibint7) amplification. We ran thirty-nine PCR cycles consisting in 1 minute at 94°C, 40s at 58°C and 50s at 72°C preceded by an initial denaturation step of 4 minutes at 94°C. These cycles were followed by a 5 minutes final extension step at 72°C. PCR procedure for cold shock domain containing E1 (CSDE1) and PAX interacting protein 1 (PAXIP1) introns followed protocols and used primers described in Kimball et al. (2009). For the five markers, sequencing was conducted under BigDyeTM terminator cycling conditions at the “Genoscope - Centre National de Séquençage”, France. DNA sequences were aligned using CodonCode Aligner 3.0.3 (CodonCode Corporation, 2009) and ClustalW (Thompson et al., 1994) as implemented in Mega version 4 (Tamura et al., 2007) and checked by eye Sequences are deposited on GeneBanks under accession numbers XXXXXX. 146 Evolutionary relationships estimation Genetic distances in CO1 can be used when investigating species delineation in the Barcode of Life project (Hebert et al., 2003a, 2003b, 2004). We therefore computed corrected mean K2P distances between taxa, and within taxon K2P distances in Mega v.4 using the “Pairwise Deletion” option with CO1 data. Because phylogenetic trees obtained with all markers did not differentiate the two Gould’s petrel taxa (data not shown), we used haplotype network reconstruction to infer relationships between populations of caledonica and leucoptera. These methods, contrary to phylogenetic tree reconstruction, do not impose bifurcating relationships between sequences and are well suited for intraspecific investigations (Posada & Crandall, 2001; Forister et al., 2008). For all five markers, phylogenetic relationships between haplotypes were inferred using the median-joining networks method implemented in NETWORK v4.5.1 (Bandlet et al., 1999; Fluxus Technology Ltd, 2009). Estimate of genetic diversity, genetic structure and historical demography To estimate haplotypes of nuclear loci, the software PHASE 2.1.1 (Stephens et al. 2001) was used as implemented in DnaSp v5 (Librado & Rozas, 2009). The latter was then used to estimate haplotypes of all five loci for use with Arlequin. Nucleotide diversity (π) and haplotype diversity (Hd) were calculated with Arlequin v3.5 (Excoffier & Lischer, 2010). This software was also used to conduct an exact test of population differentiation (Raymond & Rousset 1995) and test for population structure with an AMOVA. Furthermore, to assess whether selection has affected the loci examined and also to test for past population expansion, Fu’s Fs (Fu, 1997), Tajima’s D (Tajima, 1989) and Ramos-Onsins & Rozas R2 (Ramos-Onsins & Rozas, 2002) tests of neutrality were performed using 1,000 simulations in Arlequin and DnaSp. We considered Fs statistic significant when the P value was below 0.02 to take into account the potential irregular behaviour of this indice (Fu, 1997; Ramos-Onsins & Rozas, 2002). Significant negative values of Fs and D statistics point towards population growth and/or positive selection, whereas positive Tajima's D values indicate bottlenecks and/or balancing selection. R2 test was found to be more powerful than the two previous tests in most cases by its authors (Ramos-Onsins & Rozas, 2002) and significant value suggests population growth. We also investigated historical demography of P. leucoptera using mismatch distribution under a model of sudden range expansion with Harpending’s raggedness index (Harpending, 1994) and the sum of squared deviation SSD (Schneider et al., 2000) between the observed and expected distributions. A significant SSD value indicates a 147 departure from the model of sudden population expansion (Schneider & Excoffier, 1999) and low values of r usually reflect a population expansion. Estimation of gene flow, population divergence time and effective population sizes Given the distance between breeding localities and the philopatric behaviour of most petrel taxa, one would expect an absence of migration between Australian and New Caledonian breeding colonies. Potential gene flow between populations may homogenise the gene pool of caledonica and leucoptera and prevent the divergence of these two lineages, a process that could have important conservation consequences. Using IMa2 (Hey, 2010), we estimated this parameter between the two populations. This software allows one to test for gene flow between any number of populations between two and ten and requires the user to provide a topology of the populations it wishes to study. Several runs were sampled with a burn-in period of 4.105 steps and sampling period of 2.106 steps, using 10 to 100 chains and a geometric heating scheme. In order to scale parameter estimates in demographic units, we provided the software with a range of mutation rates for the two mitochondrial loci, using a mean value of 0.794 ± 0.115% per million year for CO1 (Pereira & Baker, 2006) and 1.89 ± 0.35% for Cytb (value found for Procellariiformes; Weir & Schlutter, 2008). We let IMa2 calculate mutation rates scalars for the other loci (i.e. no mutation rate was specified for the three nuclear introns, as these are currently unknown). In addition, to evaluate the degree of gene flow, divergence time and population sizes in demographic unit, it is also possible to enter the generation time of a taxon. Pterodroma petrels are, as all other Procellariiformes, long-lived seabirds. No precise estimation of generation time currently exists for Gould’s petrels. However, in the Atlantic petrel (Pterodroma incerta) this parameter is estimated to be 20 years (Cuthbert, 2004) and in another genus belonging to Family Procellariidae, Scott et al. (2008) found two estimates of generation time in the sooty shearwater (Puffinus griseus) of 14.7 and 21.6 years respectively based on different adult survival assumptions. We consequently decided to use an approximate generation time of 15 years for Pterodroma leucoptera petrels. Runs were monitored by observing effective sample size (ESS) values and inspecting parameter plots for trends following the manual recommendations. 148 Results Phylogenetic relationships Mean intra-taxon K2P distances with gene CO1 were 0.29% and 0.21% in caledonica and leucoptera respectively. The two taxa were virtually undifferentiated when calculating the mean corrected K2P distance between the two groups (0.02%). Calculating the overall mean K2P distance between all sequences of caledonica and leucoptera gave a value of 0.23%. When analysing Gould’s petrels with a haplotype network approach, all three nuclear introns produce a network characterised by the presence of two dominant haplotypes and a star-like repartition of mostly private haplotypes differing from the main ones by one or two mutations (Fig. 5.1.2). Furthermore, in each network, the main haplotype is shared by birds from both populations of Gould’s petrel, and with βFibint7 and CSDE1, it is the case also for the second main haplotype. The βFibint7 network differs from the others by the presence of one Caledonian haplotype separated from the main one by a putative unsampled haplotype and five mutational steps. The observed structure of the nuclear intron networks is typical of populations that expanded recently from small or limited numbers of founders (Avise, 2000), a pattern congruent with what can be deduced from genetic diversity indices. The CO1 haplotype network presents a structure similar to that of nuclear introns in that it is dominated by one haplotype shared by individuals of both populations and a star-like “radiation” of haplotypes. More of these are shared between the two populations than what is found with the nuclear loci. The cytochrome b network is somewhat similar to that obtained with CO1. However, with Cytb only one dominant haplotype is recovered and its prevalence is not as strong as what is found with CO1 or the nuclear introns. Most private haplotypes are differentiated from their closest neighbour by two or more mutations. Overall, all networks are characterised by one or two dominant haplotypes shared between the 2 taxa. Although private haplotypes are present for both Gould’s petrel taxa in all five loci, no lineage sorting appears in any of the loci. All networks are characteristic of populations that experienced a rapid expansion. 149 Figure 5.1.2 Haplotype networks obtained with five loci in P. leucoptera leucoptera and P. leucoptera caledonica. Circle size proportional to number of individual sharing that haplotype. Bars across line joining haplotypes represent mutational differences. 150 Genetic diversity and population structure and history Mitochondrial loci With both loci caledonica and leucoptera populations show high haplotypic diversity (Hd>0.5; Table 5.1.1) and low nucleotide diversity (i.e. π <0.5%; Grant & Bowen, 1998). This type of pattern can be linked to a scenario of population expansion following a period of low effective population size, population growth enhancing the retention of new mutations (Avise et al., 1984; Grant & Bowen, 1998). In leucoptera, all indices indicate a population increase, although Fu’s Fs is not significant with CO1 (Table 5.1.1). Conversely, in caledonica all but SSD tests support a stable population and departure from the population growth scenario. However, in this taxon, non significant negative values of Tajima’s D and Fu’s Fs in CO1 and of Tajima’s D in Cytb suggest a potential selective sweep or population increase. In terms of population structure, with both markers pairwise FST are small and not significant (Table 5.1.2) and the exact test of population differentiation of Raymond and Rousset (1995) is not significant in Cytb, and only weakly significant in CO1. In addition, the genetic variation is mainly due to within population variation (>95%) rather than to interpopulation variation (<5%) as shown by AMOVA (Table 5.1.3), with virtually no interpopulation variation with Cytb data. Nuclear intron With the three nuclear markers, caledonica population exhibits high haplotypic diversity and low nucleotide diversity (Table 5.1.1) thus supporting a hypothesis of population expansion following a population low. Neutrality indices corroborate the scenario of population growth, all markers showing either two or three of the Fu’s Fs, Tajima’s D or Ramos-Onsins and Rozas R2 significant. This is also supported by the mismatch analysis with the three markers. (Table 5.1.1). Thus we observe a contradiction between mitochondrial and nuclear loci for this taxon. In leucoptera significant Fs and R2 with βFibint7 and PAXIP1 both indicate that this taxon experienced a population expansion. Although not significant, negative values of D with the three markers indicate such an increase or a potential selective sweep. High haplotypic and low nucleotidic diversities observed with these two markers also suggest that following a low population size this taxon experienced an expansion. However, the signal given by the various indices with CSDE1 is somewhat confusing relative to the two other introns: with this gene, negative values of Fs and D would suggest a potential population increase, as does SSD value, while non significant R2 and high raggedness index value do not support the population expansion hypothesis. However, under a scenario of population expansion, R2 value is supposed to be low 151 (Ramos-Onsins & Rozas, 2002). Here with CSDE1, R2 value in leucoptera is 0.054, i.e. similar to the significant values found with the other markers in this taxon (Table 5.1.1), thus suggesting that a potential population growth. As found with mitochondrial loci, most of genetic variation (>90%) is found within populations rather than among populations (AMOVA; Table 5.1.4) with the three nuclear introns. Similarly, FST values are low (Table 5.1.3). However, contrary to mitochondrial loci, pairwise FST values are significant indicating population structure between the two taxa. Furthermore, exact test of differentiation are highly significant (P<0.01; Table 5.1.3) contrary to mitochondrial data. Thus we observe contradicting signals between mtDNA and nucDNA: in caledonica the former indicates a stable population, while the latter supports the hypothesis of population increase; in addition, mtDNA contradicts nucDNA in that it detects only a weakly significant population structure in one marker, while all three nucDNA loci indicate significant population structure between the two taxa. 152 Table 5.1.1 Diversity and neutrality estimates and mismatch anlysis indices under a model of population expansion for two Pterodroma leucoptera taxa. Number of sequences (N), number of polymorphic sites (Np), number of haplotypes (Nh), haplotype diversity (Hd), nucleotide diversity (Pi, expressed as percentages, i.e. 0.001=0.1%), average number of nucleotide differences (k). Significant values (P<0.05; P<0.02 for Fu’s Fs) are in bold Np Nh Hd π% k Fu’s Fs R2 Tajima’s D r SSD 30 11 11 0.821 0.346 2.531 -3.065 0.097 -0.733 0.102 0.033 P. leucoptera 43 34 13 0.714 0.379 2.777 -3.276 0.057 -2.101 0.078 0.026 P. caledonica 5 4 0.9 0.457 4.2 0.212 0.164 -0.197 0.41 0.178 P. leucoptera 16 20 11 0.925 0.388 3.442 -4.543 0.074 -1.726 0.054 0.02 P. caledonica 26 18 17 0.917 0.251 2.354 -13.779 0.069 -1.777 0.054 0.003 P. leucoptera 68 8 9 0.605 0.085 0.792 -4.747 0.055 -1.358 0.132 0.012 P. caledonica 20 10 8 0.774 0.333 1.6 -3.024 0.0794 -1.512 0.05 0.058 P. leucoptera 92 6 6 0.399 0.097 0.465 -2.892 0.054 -1.379 0.152 0.002 P. caledonica 12 8 5 0.667 0.308 1.47 -0.952 0.15 -1.778 0.05 0.008 P. leucoptera 72 12 11 0.61 0.245 1.17 -4.866 0.05 -1.47 0.066 0.146 PAXIP1 Cytb CO1 P. caledonica βFibint7 N CSDE1 Taxon 9 153 Table 5.1.2 Pairwise FST values and exact test of population differentiation P values between Gould’s petrel populations from New Caledonia and Australia Pairwise FST CO1 0.026 Exact test of differentiation (P value) 0.045 Cytb 0.01 0.488 βFibint7 0.034 0.0001 CSDE1 0.092 0.0006 PAXIP1 0.068 0.006 Table 5.1.3 Genetic variation within and between the two Gould’s petrel taxa (AMOVA) PAXIP1 CSDE1 βFibint7 Cytb CO1 d.f Among populations 1 Sum of Squares 2.591 Within populations 71 95.026 1.338 97.42 Among populations 1 2.061 0.019 0.93 Within populations 19 36.462 1.919 99.04 Among populations 1 1.42 0.022 3.43 Within populations 92 55.952 0.608 96.57 Among populations 1 1.434 0.034 9.22 110 36.352 0.33 90.78 Among populations 1 1.52 0.044 6.85 Within populations 82 49.611 0.605 93.15 Within populations Variance components 0.035 Percentage of variation 2.58 154 Gene flow, population divergence time and effective population sizes Using a mean exponential migration rate prior value of 0.1 yields an estimation of population split between caledonica and leucoptera approximately 40,000 years ago (Table 5.1.4; Fig. 5.1.3). Population size at time of splitting is estimated very high, ≈85,000 for caledonica and ≈29000 for leucoptera. However, confidence intervals (95% Highest Posterior Density) are extremely large for all these parameters. Migration was found to happen in both directions between populations (Fig. 5.1.3), though being more important from leucoptera to caledonica than the reverse. However these migration rates are not significant (Nielsen and Wakeley test, P>0.05; Nielsen & Wakeley, 2001). Table 5.1.4 Result from Isolation with Migration analysis with IMa2. 95%HPD intervals are given in brackets. Population estimates are for effective population Ne. “Population 1” refers to the first taxon of the pair (first column), “population 2” to the second taxon Divergence time (Million years) Population size caledonica Population size leucoptera Ancestral population 0.041 [0.020844 – 0.23197] 84954 [31164 – 639078] 28720 [13159 – 81550] 19000 [4376 – 19744] Figure 5.1.3 History of caledonica and leucoptera under an Isolation with Migration model, with an exponential migration prior with mean m=0.1. Horizontal lines represent splitting times (dashed lines for 95% HPD) and curved arrows migration per generation. Time is represented on the vertical axis, with the sampled subspecies names given at the top of the figure at the most recent time point. Plain line boxes give population size, and lighter blue boxes and horizontal arrows represent 95% HPD for population estimates 155 Discussion Phylogenetic relationships between the two Gould’s petrel taxa and taxonomic implications With mitochondrial gene CO1, we found that the mean distance between all specimens of caledonica and leucoptera together was 0.23%, exactly the same value than the mean intraspecific distance found by Kerr et al. (2007) in 643 species of North American birds, and slightly lower than the mean intraspecific distance found in Scandinavian birds by Johnsen et al. (2010). In addition, within group mean K2P distance was less than 0.3% in both taxa. Hebert et al. (2004) found a maximum intraspecific difference of 1.24% in 260 species of North American birds. Thus, these CO1 data suggest that caledonica and leucoptera are not differentiated with this gene and, if considered in the framework of the BarCode of Life initiative (Hebert et al., 2003a, 2003b), that they do not deserve specific status. This conclusion is corroborated by the haplotype network analysis: these taxa do not exhibit any lineage sorting in any of the five loci investigated. Although a sample of five loci is small comparatively to the total number of genes of these seabirds, these results nevertheless indicate that these population lineages have not (yet) diverged. This is supported by the values of gene flow identified by the Isolation with Migration analysis: although non significant, contemporary or recent migration was detected between the two populations in both directions. This study therefore brings clear evidence that the two Gould’s petrel taxa constitute a single Evolutionary Significant Unit (Moritz, 2002) and surely do not deserve full species status. Given the apparent lack of genetic differentiation between the two populations, it could also be debated whether the subspecies status is warranted. Furthermore, it appears that morphological and colouration differences although present do show some overlap between the two taxa hence only weakly support their subspecies status. However, it has been argued (Winker & Haig, 2010) that good subspecies, even if described on other characters than molecular ones can have conservation importance and be important evolutionary speaking. Here, given differences in other characters such as breeding ranges or ecological differences, notably in terms of breeding habitat and breeding phenology, we follow Winker & Haig and argue in favour of the retention of the actual subspecies status attributed to the two populations. Population history In terms of population demographic history, leucoptera clearly exhibits signs of a rapid population expansion, possibly consecutive to a founder event or population bottleneck. This scenario is supported by a majority of indices in all five loci, and the structure of the haplotype 156 networks is an expected pattern for species that experienced population growth following a founder event (Avise, 2000). On the other hand, the overall signal given by the five loci in caledonica is somewhat more confused. This taxon could be experiencing a prolonged period of relatively important population size, as seems to be indicated by the mitochondrial data, that would follow a period of population expansion in a more distant past that can be observed in some nuclear loci. The pattern of population size found with the IM analysis somewhat contradicts this scenario: a Ne of ≈29000 does not support the idea of a founder event or bottleneck for leucoptera. One potential explanation would consist in a population crash shortly afterward the split followed by a rapid population expansion that would lead to the recovery of past diversity: because the IM model assumes constant population sizes it is possible that it will not detect a bottleneck occurring after the population divergence. Another explanation could be that, from a very large caledonica population, a small number split and gave rise to leucoptera but due to important migration and gene flow between the two populations and small data set used, in particular in terms of number of loci, the IM model fails to recover the proper population and migration parameter estimates. Indeed, all parameter estimates in the IM analysis show extremely large confidence intervals (Table 5.1.4), suggesting caution is needed in the interpretation of these estimates. Population structure The absence of population structure among mtDNA sequences suggests contemporary gene flow and/or recent evolutionary connections among populations of Gould’s petrel (Slatkin & Maddison, 1989). The IM analysis supports this hypothesis: although detecting migration (i.e., movements between the populations), this was not significant. The pattern observed of migration more important from leucoptera to caledonica can make sense if one supposes that leucoptera at sea join the large populations of caledonica, which are known to forage in the Tasman Sea (Imber & Jenkins, 1981; Thomas & Bretagnolle 1986), and very likely follow the East Australian Current (EAC) and the Tasman Front and their associated eddies. Because the EAC bifurcates from the coast approximately at 32°S, i.e. very close from Boondelbah and Cabbage Tree Islands (Bostock et al., 2006) this could help leucoptera to mix with caledonica birds and promote mixing of populations. However, data is needed on the movements of these birds in the Tasman Sea to confirm this hypothesis. These data are currently beingcollected with GLS loggers. Furthermore, it was shown by Birky et al., (1983) and Slatkin (1987) that an exchange of just one or two females per generation is susceptible to counteract genetic drift in the mitochondrial genome. This could explain the opposite patterns of population structure found between mitochondrial and nuclear loci: it can be suggested that in a small population such as leucoptera, 157 limited amount of female mediated gene flow, through immigration of caledonica females, could affect more the mitochondrial lineage than the nuclear lineage due to the maternal inheritance of mitochondrial DNA and its smaller effective population size compared to nuclear DNA (Ne vs 4 Ne). This would concur with the well known pattern of dominant female mediated gene flow through female dispersal observed in most bird species (Newton, 2003). A hypothetic phylogeographic scenario We may propose that, following the colonisation of New Caledonia by caledonica or its ancestral form, leucoptera is the result of a colonisation in the late Pleistocene of Western Australian islands by New Caledonian birds that disperse widely in the Tasman Sea (Imber and Jenkins, 1981). Given the possible contemporary or recent gene flow between populations and that this divergence is very recent, lineages have not had time yet to diverge sufficiently, as shown by their low genetic differentiation and the presence of shared haplotypes. In a somewhat similar pattern of population expansion signature in genetic diversity and haplotype structure, in their study of Puffinus tenuirostris shearwaters in southern Australia, Austin et al. (1994) observed a low diversity between individuals and that most haplotypes were shared between all the 11 colonies investigated. Furthermore, the mean mtDNA sequence divergence between individuals was 0.25%, i.e. just slightly above the values found in our study with CO1 data. These authors concluded that this species have probably been greatly reduced in numbers in the past, probably through one or more bottlenecks during the last glaciation events in the late Pleistocene, followed by subsequent expansions. It is well recognised that cycles of glacials/interglacials during the Pleistocene are likely to have had a marked effect in seabird avian species, through population bottlenecks and founder events during subsequent population expansion from breeding refugia, variations in habitat availability or likely changes in foraging distribution linked to oceanic conditions variations (Moum et al., 1991; Ovenden et al., 1991; Birt-Friesen et al., 1992; Avise & Walker, 1998; Cagnon et al., 2004; Gomez-Diaz et al., 2006; Techow et al., 2010). Here we find that the population split between caledonica and leucoptera would have happened approximately 40,000 years ago during the last glaciation. The period 35-30,000 thousands years ago corresponds to an interglacial period characterised by warmer periods and higher sea levels (Kershaw et al., 2003) as well as warmer sea surface temperatures in the southern oceans (Barrows et al., 2007). In addition, Kawahata (2002) showed that the Tasman Front moved several times in latitude during the late Pleistocene. We propose that the split between the populations of Gould’s petrel might have been promoted by these changes in oceanic conditions in the Tasman Sea and changes of habitat availability linked to variable sea levels and variations of vegetation in relation to climatic conditions. The present observed differences in breeding habitats, caledonica nests in 158 mountains, along river beds in the forest while leucoptera nests pretty much at sea level, in crevices or rudimentary burrows on cabbage tree island, indicate that this species might be very adaptable in terms of breeding environment, rendering even more complicated to hypothesise on the underlying reasons of the colonisation of West Australian islands. In such a scenario of a large caledonica population, from which emerges the leucoptera population in a recent past, the pattern of nucleotide and haplotype diversity observed for caledonica can seem surprising. However, it must be noted that one of the limitations of this study lies in the biased sampling of caledonica colonies: indeed, not all colonies have been sampled due to their remoteness and difficult access and that not all colonies are known in the different mountain massifs (Villard et al., in prep.). It is likely that with an appropriate sampling of all breeding colonies in New Caledonia, the observed genetic diversity would be higher and haplotype network structure would be more complex for this taxon. Furthermore, sample size difference between the two populations can also be invoked to explain in part the observed pattern: while the sample size used for leucoptera might capture most, if not all, genetic diversity and present haplotypes in this small population, the smaller n used for caledonica, a population supposed to be much larger than leucoptera, might represent only a fraction of the total genetic and haplotypic diversity of New Caledonian birds, especially so if this population is structured between colonies. Conservation implications In terms of conservation, it can be debated whether it is worth investing in management actions of all sub-species or all populations within a species. Intraspecific units of conservation were originally based on taxonomically defined subspecies (Moritz, 2002) but this has sometimes led to incongruent subdivisions between molecular studies and recognised subspecies (Moritz, 2002). However, it has been argued that any independently evolving population is worth conserving, even if not a species (Tobias et al., 2010). In the case of Gould’s petrel subspecies from Australia and New Caledonia, we have shown here that they do not differ significantly genetically in the five loci investigated, which questions the independent evolution of these two lineages and consequently the conservation efforts directed towards one of the subspecies rather than the other. However, ecological differences, notably in terms of breeding environment, and geographically separated breeding grounds argue in favour of the retention of conservation efforts directed at both populations. Indeed, even if they cannot be treated as different ESU (due to lack of mitochondrial DNA monophyly), we argue that these populations need be conserved as different management units for different reasons. First, the species as a whole is considered Vulnerable (Birdlife International, 2010) and conserving geographically separated breeding colonies can greatly reduce its extinction risk. Furthermore, even though these populations have probably exchanged genes 159 until recently and apparently do not exhibit independently evolving genetic lineages, it cannot be ruled out that such independent evolution will eventually occur in the future. Thus investing in the conservation of both populations can help achieve one of the goals of conservation biology, i.e. the conservation of evolutionary potential of lineages. Finally, given the success of management and recovery actions undertaken in Australia, and the relative ease to manage the leucoptera population due to its location and restriction on two small islands, compared to the conservation management difficulties inherent to the location of caledonica breeding colonies in New Caledonian mountains and the many potential threats faced by this taxon (Bretagnolle, 2001) it is well worth preserving what can be most easily be preserved, even if it does not represent the most important part of the Gould’s petrel population. Acknowledgements This work was supported by the "Consortium National de Recherche en Génomique", and the "Service de Systématique Moléculaire" of the Muséum National d'Histoire Naturelle (CNRS UMS 2700). It is part of the agreement n°2005/67 between the Genoscope and the Muséum National d'Histoire Naturelle on the project "Macrophylogeny of life" directed by Guillaume Lecointre. References Austin J.J., White R.W.G. and Ovenden J.R., 1994. Population genetic structure of a philopatric, colonially nesting seabird, the short-tailed shearwater (Puffinus tenuirostris). The Auk 111(1): 70-79 Avise J.C., 2000. Phylogeography. The history and formation of species. Harvard University Press, Cambridge, Massachusetts Avise J.C. & Walker DE, 1998. Pleistocene phylogeographic effects on avian populations and the speciation process. Proceedings of the Royal Society of London B Biological Science 265: 457-463 Avise J.C., Neigel J.E. and Arnold J., 1984. Demographic influences on mitochondrial DNA lineage survivorship in animal populations. Journal of Molecular Evolution 20: 99-105 160 Bandlet H.J., Forster P. and Röhl A., 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16, 37–48 Barrows T.T., Juggins S., De Dekker P., Calvo E. and Pelejero C., 2007. Long-term sea surface temperature and climate change in the Australian – New Zealand region. Paleoceanography 22: 2215 BirdLife International, 2010. http://www.birdlife.org accessed on 10/10/2010 Birky C.W., Maruyama T. and Fuerst P., 1983. Genetic theory for genes in mitochondriaand chloroplasts and some results. Genetics 103: 513-527 Birt-Friesen V.L., Montevecchi W.A., Gaston A.J. and Davidson W.S., 1992. Genetic structure of thick-billed murre (Uria lomvia) populations examined using direct sequence analysis of amplified DNA. Evolution 46: 267–272 Bostock H.C., Opdyke D.N., Gagan M.K., Kiss A.E. and Fifield L.K., 2006. Glacial/interglacial changes in the East Australian current. Climate Dynamics 26: 645-659 Bourne W.R.P., 1983. The appearance and classification of the cookilarias petrels. Sea Swallow 32: 67-71 Bouzat J.L., 2010. Conservation genetics of population bottlenecks: the role of chance, selection and history. Conservation Genetics 11(2): 463-478 Bretagnolle V., 2001. Le Petrel de la Chaine Pterodroma (leucoptera) caledonica: Statut et menaces. Rapport de fin d’étude pour le compte de la Province Sud, Nouvelle Calédonie Brooke M., 2004. Albatrosses and Petrels Across the World. Oxford University Press, New York Cagnon C., Lauga B., Hémery G. and Mouchès C., 2004. Phylogeographic differentiation of storm-petrels (Hydrobates pelagicus) based on cytochrome b mitochondrial DNA variation. Marine Biology 145: 1247-1254 Cuthbert R., 2004. Breeding biology of the Atlantic petrel, Pterodroma incerta, and a population estimate of this and other burrowing petrels on Gough Island, South Atlantic Ocean. Emu 104: 221-228 Daugherty C.H., Cree A., Hay J.M. and Thompson M.B., 1990. Neglected taxonomy and continuing extinctions of Tuatara (Sphenodon). Nature 347: 177-179 deNaurois R., 1978. Procellariidae reproducteurs en Nouvelle Calédonie pendant l’été austral. Comptes Rendus de l’Académie des Sciences, Paris 287 Ser. D 269-271 Edwards S.V., 2009. Is a new and general theory of molecular systematics emerging? Evolution 63(1): 1-19 Ehrlich P.R. and Wilson E.O., 1991. Biodiversity studies: science and policy. Science 253: 758-762 161 Excoffier L. and Lischer H.E.L., 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10(3): 564-567 Forister M.L., Nice C.C., Fordyce J.A., Gompert Z. and Shapiro A.M., 2008. Considering evolutionary processes in the use of single-locus genetic data for conservation, with exemples from the Lepidoptera. Journal of Insect Conservation 12: 37-51 Frankam, R., 2010. Where are we in conservation genetics and where do we need to go? Conservation Genetics 11: 661-663 Friesen V.L., Burg T.M. and McCoy K.D., 2007. Mechanisms of population differentiation in seabirds. Molecular Ecology 16:1765-1785 Friesen V.L., Gonzalez J.A. and Cruz-Delgado F., 2006. Population genetic structure and conservation of the Galapagos petrel (Pterodroma phaeopygia). Conservation Genetics 7: 107-115 Fu Y.X., 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147: 915–925 Fullagar P.J., 1976. Seabird islands No. 35: Cabbage Tree Island, New South Wales. Australian Bird Bander 14: 94-97 Gomez-Diaz E., Gonzales-Solis J., Peinado M.A. and Page R.D.M., 2006. phylogeography of the Calonectris shearwaters using molecular and morphometric data. Molecular Phylogenetics and Evolution 41: 322-332 Grant W.S. and Bowen B.W., 1998. Shallow population histories in deep evolutionary lineages of marine fishes: insights from Sardines and Anchovies and lessons for conservation. Journal of Heredity 89(5): 415-426 Harpending R., 1994. Signature of ancient population growth in a low-resolution mitochondrial mismatch distribution. Human Biology 66: 591–600 Hebert P.D.N., Cywinska A., Ball S.L. and DeWaard J.R., 2003a. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London B Biological Science 270: 313– 321. Hebert P.D.N, Ratnasingham S. and DeWaard J.R., 2003b. Barcoding animal life: Cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society of London B Biological Science 270: S596–S599 Hebert P.D.N., Stoeckle M.Y., Zemlak T.S. and Francis C.M., 2004. Identification of birds through DNA barcodes. PLoS Biology 2: 1657–1663 Hedrick P.W., 2001. Conservation genetics: where are we now? Trends in Ecology and Evolution 16(11): 629-636 162 Hey J., 2010. Isolation with migration model for more than two populations. Molecular Biology and Evolution 27(4): 905-920 Hindwood K.A. & Serventy D.L., 1941. The Gould petrel of Cabbage Tree Island. Emu 41: 1-20 Imber M.J., 1985. Origins, phylogeny and taxonomy of the gadfly petrels Pterodroma spp. Ibis 127:197–229 Imber M.J. & Jenkins J.A.F., 1981. The New Caledonian Petrel. Notornis 28(3): 149-160 Irwin D.E. & Gibbs H.L., 2002. Phylogeographic breaks without geographic barriers to gene flow. Evolution 56: 2383–2394 Jesus J., Menezes D., Gomes S., Oliveira P., Nogales M. and Brehm A., 2009. Phylogenetic relationships of gadfly petrels Pterodroma spp.from the Northeastern Atlantic Ocean: molecular evidence for specific status of Bugio and CapeVerde petrels and implications for conservation. Bird Conservation International 19: 1-16 Johnsen A., Rindal E., Ericson P.G.P., Zuccon D., Kerr K.C.R., Stoeckle M.Y., Lifjeld J.T., 2010. DNA barcoding of Scandinavian birds reveals divergent lineages in trans-Atlantic species. Journal of Ornithology DOI 10.1007/s10336-009-0490-3 Kawahata H., 2002. Shifts in oceanic and atmospheric boundaries in the Tasman Sea (Southwest Pacific) during the Late Pleistocene: evidence from organic carbon and lithogenic fluxes. Palaeogeography, Palaeoclimatology, Palaeoecology 184: 225-249 Kerr K.C.R., Stoeckle M.Y., Dove C.J., Weigt L.A., Francis C.M. and Hebert P.D.N., 2007. Comprehensive DNA barcode coverage of North American birds. Molecular Ecology Notes 7: 535–543 Kershaw P., Moss P. and Van der Kaars S., 2003. Causes and consequences of long term climatic variability on the Australian continent. Freshwater Biology 48: 1274-1283 Kimball R.T. et al., 2009. A well-tested set of primers to amplify regions spread across the avian genome. Molecular Phylogenetics and Evolution 50: 654-660 Lambeck K., Esat T.M. and Potter E-K., 2002. Links between climate and sea levels forthe last three millions years. Nature 419: 199-206 Librado P. and Rozas J., 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25(11): 1451-1452 Martinez-Cruz B., Godoy J.A. and Negro J.J., 2004. Population genetics after fragmentation: the case of the endangered Spanish imperial eagle. Molecular Ecology 13: 2243–2255 Miller K.A., Chapple D.G., Towns D.R., Ritchie P.A. and Nelson N.J., 2009. Assessing genetic diversity for conservation management: a case study of a threatened reptile. Animal Conservation 12: 163-171 163 Moritz C., 2002. Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51(2): 238-254 Moum T., Erikstad K.E. and Bjørklid E., 1991. Restriction fragment analysis of mitochondrial DNA in common murres. Canadian Journal of Zoology 69: 1577–1584. Newton I., 2003. The speciation and biogeography of birds. Academic Press London UK. Nielsen R. & and Wakeley J., 2001. Distinguishing migration from isolation: a Markov Chain Monte Carlo approach. Genetics 158: 885-896 Ovenden J.R., Wustsaucy A., Bywater R., Brothers N. and White R.W.G., 1991. Genetic-evidence for philopatry in a colonially nesting seabird, the Fairy Prion (Pachyptila turtur). The Auk 108: 688–694 Pearce J.M., 2007. Philopatry: a return to origins. The Auk 124: 1085–1087 Pereira S.L. and Baker A.J., 2006. A mitogenomic timescale for birds detects variable phylogenetic rates of molecular evolution and refutes the standard molecular clock. Molecular Biology and Evolution 23(9):1731–1740 Posada D. & Crandall K.A., 2001. Intraspecific genealogies: trees grafting into networks. Trends in Ecology and Evolution 16(1): 37-45 Priddel D. & Carlile N., 1997. Boondelbah Island confirmed as a second breeding locality for Gould’s Petrel Pterodroma leucoptera leucoptera. Emu 97: 245-248 Priddel D. & Carlile N., 2009. Key elements in achieving a successful recovery programme: a discussion illustrated by the Gould’s petrel case study. Ecological Management and Restoration 10: S97-S102 Prychitko T. M. and Moore W. S., 1997. The utility of DNA sequences of an intron from the βfibrinogen gene in phylogenetic analysis of woodpeckers (Aves: Picidae). Molecular Phylogenetics and Evolution 8: 193–204 Ramos-Onsins S.E. and Rozas J., 2002. Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution 19: 2092–2100 Raymond M. and Rousset F., 1995. An exact test for population differentiation. Evolution 49(6): 1280-1283 Rayner M.J., Caraher C.J..F.Clout M.N. and Hauber M.E., 2010. Mitochondrial DNA analysis reveals genetic structure in two New Zealand Cook’s petrel (Pterodroma cooki) populations. Conservation Genetics 11(5): 2073-2077 Roques S. & Negro J.J., 2005. MtDNA genetic diversity and population history of a dwindling raptorial bird, the red kite (Milvus milvus). Biological Conservation 126: 41–50 164 Schneider S. and Excoffier L., 1999. Estimation of Past Demographic Parameters From the Distribution of Pairwise Differences When the Mutation Rates Vary Among Sites: Application to Human Mitochondrial DNA. Genetics 151: 1079-1089 Schneider S., Roessli D. and Excoffier L., 2000. In: Genetics and Biometry Lab. Dept. of Anthropology, University of Geneva Scott D., Scofield P., Hunter C. and Fletcher D., 2008. Decline of sooty shearwaters, Puffinus griseus, on the Snares, New Zealand. Papers and Proceedings of the Royal Society of Tasmania 142(1): 185-196 Simon C., Frati F., Beckenbach A., Crespi B., Liu H. and Flook P., 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Annals of the Entomological Society of America 87(6): 651-701 Slatkin M., 1987. Gene flow and the geographic structure of natural populations. Science 236: 787792 Slatkin M. & Maddison W.P., 1989. A cladistic measure of gene flow inferred from the phylogenies of alleles. Genetics 123: 603-613 Stephens M., Smith N. and Donnelly P., 2001. A new statistical method for haplotype reconstruction from population data. American Journal of Human Genetics 68: 978–989 Tajima F., 1989. The effect of change in population size on DNA polymorphism. Genetics 123: 597–601 Tamura K., Dudley J., Nei M. and Kumar S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24(8): 15961599 Techow N.M.S.M., O’Ryan C., Phillips R.A. et al., 2010. Speciation and phylogeography of giant petrels Macronectes. Molecular Phylogenetics and Evolution 54: 472-487 Thompson J.D., Higgins D.G. and Gibson T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22(22): 4673-4680 Tobias J.A., Seddon N., Spottiswoode C.N., Pilgrim, J.D., Fishpool L.D.C. and Collar N.J., 2010. Quantitative criteria for species delimitation. Ibis 152: 724–746 Weir J.T. and Schluter D., 2008. Calibrating the avian molecular clock. Molecular Ecology 17: 2321-2328 Winker K. & Haig S.M. (eds), 2010.Avian subspecies. Ornithological Monographs 67: 1–200 165 Part 2: Preliminary taxonomic investigations in Gould’s (Pterodroma leucoptera) and Collared (Pterodroma brevipes) petrels Gangloff et al., In Prep Abstract Phylogenetics and taxonomy of small Procellariformes (Aves) of genus Pterodroma present numerous uncertainties leading to frequent revisions that can hamper conservation efforts in the many threatened taxa. Gould’s (Pterodroma leucoptera) and Collared (Pterodroma brevipes) are two closely related species living in the western Pacific whose differentiation based on colourations is very complex. Gould’s petrel comprises two subspecies breeding in New Caledonia and Australia, while Collared petrel is considered monotypic. Taxonomic confusion exists between the two due to the presence in French Polynesia of yet to be identified birds similar in some ways to both species, and within Collared petrel phenotypic variations question the presence of only one taxon. In this study, based on museum samples we carried a preliminary investigation on the phylogenetic relationships between and within the two species with the mitochondrial gene CO1. It appears that the two species are separated based on this locus, supporting the actual taxonomic differentiation. Within Collared petrel this preliminary evidence corroborates the assumption that this species is likely to be made of several differentiated taxa, which could have conservation implications. Additional genetic analyses using complete sequences of CO1 and nuclear genes are necessary to completely solve the phylogenetic relationships within the taxa. Keywords CO1, conservation, Barcode, Pterodroma, petrel, seabird, taxonomy 166 Introduction The concept of DNA barcode, i.e. short sequences of mtDNA that consistently differentiate species, has been advocated for several years now (Hebert et al., 2003a; Hebert et al., 2003b) and studies using the methods have flourished, particularly in birds. In this last group, for which the mitochondrial gene CO1 was selected as a good candidate for the Barcode project, extensive investigations by Hebert et al. (2004), Kerr et al. (2007) or Jonhsen et al. (2010) in North American and Scandinavian birds have proven the potential validity and utility of the method to distinguish species. However much scepticism and critics have arisen over the years (e.g. Moritz & Cicero, 2004; Hickerson et al., 2006; or Will et al., 2005) stating among other impediments that the method could be inappropriate to detect recently diverged, closely related sister species. This latter claim was however refuted by Tavares & Baker (2008). In addition to its potential use in taxonomy, DNA barcoding has already been proved useful in conservation biology, through detection of underestimated species diversity (Lohman et al., 2010), or identification of illegal trade of protected species, as shown in snakes (Dubey et al., 2010) or vertebrates in general (Eaton et al., 2010). Gadfly petrels of the genus Pterodroma (family Procellariidae) are pelagic seabirds whose conservation status is globally unfavourable, 21 of the 32 recognised Pterodroma species being classified as Vulnerable, Endangered or Critically Endangered (Birdlife International, 2010). Conservation actions are therefore particularly important for the survival of these taxa and maintenance of their potential evolutionary history. However, as for many other Procellariidae, the taxonomic statuses of these small gadfly petrels have long been debated and are still in motion (Brooke, 2004), which complicates assessment of conservation needs of taxa. Sub-genus cookilaria (Bonaparte, 1855; see also Imber, 1985) comprises six to eight species (Bourne 1983, Imber 1985; Steeves et al., in prep) breeding in the south Pacific. Of these, only two breed on more than one archipelago, i.e. Gould’s (Pterodroma leucoptera) and Collared (Pterodroma brevipes; hereafter called brevipes) petrels (Fig. 5.2.1), two closely related taxa. Although formerly considered a subspecies of a large Pterodroma leucoptera complex (Hindwood & Serventy, 1941), the Collared petrel, breeding in Fiji and Vanuatu and Rarotonga (Birdlife International, 2010), was considered a full species by Imber & Jenkins (1981) based on distribution, polymorphic plumage and morphological variation, a classification subsequently followed by Sibley & Monroe (1990) based on their DNA hybridisation data. This distinction was followed by Brooke (2004) and accepted by Birdlife International in its IUCN Red List classification of birds (Birdlife International, 2010). Gould’s petrel comprises two populations treated as subspecies 167 (Imber & Jenkins, 1981; de Naurois, 1978), one P. leucoptera leucoptera (hereafter called leucoptera) breeding on two small islands in Western Australia, Cabbage Tree and, more recently, Boondelbah (Priddel & Carlile, 1997), the other P. l. caledonica (hereafter called caledonica) breeding in New Caledonia, in at least two mountain massifs (Bretagnolle, 2001). However, the situation between these three taxa is rendered more complex by the presence in the Australs archipelago of a Pterodroma colony (first discovered by R. & J. Seitre in 1989, identified by V. Bretagnolle) whose taxonomic affiliation is still to be clarified but that could belong to caledonica (V. Bretagnolle) and the capture in the Society archipelago of three birds belonging to one of these taxa but needing to be identified by genetic means. Of these three birds, that could come from breeding colonies established in the Society islands, two are supposedly Collared petrels, while the third one could be a Gould’s petrel based on colourations (VB and HS pers. obs.). Furthermore, the Collared petrel populations in Vanuatu are quite divergent in both plumage and size and might in fact be two different taxa (Bretagnolle & Shirihai, in press). Gould’s petrel is classified as Vulnerable, while Collared petrel is considered Near Threatened (Birdlife International, 2010), classifications that are susceptible to change depending on the taxonomic statuses given to all these populations. The Australian leucoptera population was estimated at around 200-500 breeding pairs in the 1970’s (Fullagar, 1976) and about 250 breeding pairs in the 1990’s (Priddel & Carlile 2009). Following intensive conservation actions started in the 1990’s, breeding success was improved and the population slowly started to increase, reaching more than 750 breeding pairs (Priddel & Carlile, 2009). In New Caledonia seven colonies of caledonica are currently known, each with an approximated 100-200 pairs, but sightings at sea suggest greater number of colonies and much greater number of birds (Bretagnolle, 2001). The exact relationships between Gould’s and Collared petrels and the exact phylogenetic and taxonomic statuses of the different populations of the latter are not known at present and investigations are needed to evaluate the degree of divergence between these taxa and reassess their taxonomic status. Here, using museum samples from brevipes from Fiji, Vanuatu and Rarotonga, as well as samples from two of the three as yet unidentified birds collected in the Society archipelago (Table 5.2.1) and samples from one colony of caledonica and from leucoptera we investigated phylogenetic relationships between these taxa with the mitochondrial gene CO1. This will constitute the basis of a more detailed analysis to be undertaken when samples from all colonies, including that of the Australs, are available and complete sequences of CO1 and other mitochondrial and nuclear genes can be obtained. 168 Figure 5.2.1 Localisation of Gould’s and Collared petrels populations. 3 3 1 2 3 5 4 1 2 3 4 5 caledonica leucoptera brevipes brevipes/caledonica brevipes/caledonica 1 New Caledonia; 2 Cabbage Tree and Boondelbah Islands; 3 Vanuatu, Fiji, Cook Islands archipelagos (from West to East); 4 Australs archipelago; 5 Society archipelago Table 5.2.1 Origin of Museum samples used. AMNH: American Museum of Natural History, New York USA; MTI: Musée de Tahiti et des Iles, Papeete, French Polynesia; Te Papa: Museum of New Zealand TePapa Tongarewa, Wellington, New Zealand Museum AMNH AMNH AMNH AMNH AMNH AMNH AMNH AMNH AMNH AMNH AMNH AMNH AMNH AMNH MTI MTI Te Papa Catalog number Taxon Origin 215400 216919 216920 216921 216923 222193 250893 250899 336700 336703 336705 336707 336709 528336 2004-3-12 2004-3-13 OR.023110 Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma brevipes Pterodroma sp Pterodroma sp Pterodroma brevipes Vanuatu-Mera Lava Island Vanuatu-Mera Lava Island Vanuatu-Mera Lava Island Vanuatu-Mera Lava Island Vanuatu-Mera Lava Island Vanuatu-Mera Lava Island Fiji - Kandavu Island Fiji - Kandavu Island Vanuatu- Tanna Island Vanuatu- Tanna Island Vanuatu- Tanna Island Vanuatu- Tanna Island Vanuatu- Tanna Island Fiji - Viti Levu Island Society - Tahiti Society - Tahiti Cook - Rarotonga 169 Material and Methods Sample collection and laboratory methods Feathers were collected on live adult birds in New Caledonia in a single breeding colony, in breeding season 2005-2006. They were stored in 70% ethanol then frozen at -20°C until processing. Blood samples were collected on Australian birds from 2005 to 2008. Blood was collected from the veins on the leg or wing using microcapillaries and, as for feathers, was stored in 70% ethanol then frozen at -20°C until processing. For birds of Society Archipelago and brevipes specimens, small pieces of skin were cut from the palm. Total genomic DNA was extracted using the DNeasy Tissue Extraction Kit (Qiagen, Valencia, CA, USA) following manufacturer’s instructions except that we increased the time of proteinase digestion to 8h overnight to ensure complete digestion. Cytochrome Oxydase 1 gene (CO1) was amplified using primers F1B 5’- AACCGATGACTATTYT-CAACC-3’ and R1B 5’-TACTACRTGYGARATGATTCC-3’, derived from primers F1/R1 (Simon et al., 1994). The PCR consisted of 37 cycles following a hot start at 94°C and a 4 minutes initial denaturation step at 94°C. Cycles, i.e. 94°C for 30s, 51°C for 40s and 72°C for 50s, were completed by a final extension at 72°C for 5 minutes. DNA of old museums samples from brevipes proved difficult to amplify and sequence. DNA was degraded and fragment sizes for amplification were small (<200bp). Thus we used additional primers (Table 5.2.2) to amplify these small DNA fragments. However, we nevertheless failed to obtain useable gene sequences for some samples, and for old AMNH samples we could not obtain whole sequences. Samples from Society Islands and Rarotonga yielded good PCR results. Complete sequences are deposited on GeneBanks under accession numbers XXXXXX. Sequencing was conducted under BigDyeTM terminator cycling conditions at the “Genoscope - Centre National de Séquençage”, France. DNA sequences were aligned using CodonCode Aligner 3.0.3 (CodonCode Corporation, 2009) and ClustalW (Thompson et al., 1994) as implemented in Mega version 4 (Tamura et al., 2007) and checked by eye 170 Table 5.2.2 Primers used in amplification of CO1 in Pterodroma brevipes. All primers were designed for this study Primers CO1 R141 5’-AGCATGGGCGGTGACGATT-3’ F78 5’-ACTTATTCGTGCAGAACTTGGTC-3’ R208 5’-AGGGGGACTAGTCAGTTTCC-3’ F150 5’-CGCCCATGCTTTCGTAATAATTT-3’ R254 5’-AGCTTATGTTGTTTATACGTGGGA-3’ F207 5’-TGGAAACTGACTAGTCCCCCT-3’ R323 5’-ACCTGCTCCTGCTTCTACGG-3’ F288 5’-ACCTCCGTCCTTCCTCCTAT-3’ R416 5’-CCTGCCAGGTGGAGGGAGA-3’ F377 5’-ATCTAGCCCATGCCGGAGC-3’ R502 5’-AAGGGGGTTTGGTACTGTGA-3’ F453 5’-GGCAATCAACTTCATTACAACAGC-3’ R581 5’-AGCATGGTGATGCCTGCGG-3’ F537 5’-ACTCATCACTGCCGTCCTAC-3’ R678 5’-TGGGTGGCCGAAGAATCAG-3’ F642 5’-TGGCGGAGGAGACCCAGTC-3’ Evolutionary relationships estimation Corrected mean K2P distances between taxa, and within taxon K2P distance were calculated in Mega v.4 (Tamura et al., 2007) using the “Pairwise Deletion” option. Groups were as follow: caledonica, leucoptera, brevipes from Vanuatu, brevipes from Fiji and, because they were each other closest relatives on the trees (see Results), birds from Tahiti and Rarotonga were grouped together. jModelTest v0.1.1 (Posada, 2008; Guindon & Gascuel, 2003) was used to evaluate the model of nucleotide substitution that best fitted the data. We used Bayesian Information Criterion (BIC), which penalises overparamerisation more heavily than AIC therefore avoiding unnecessary complex models (Sullivan & Joyce, 2005). Phylogenetic relationships were estimated using Bayesian Markov Chain Monte Carlo (MCMC) phylogenetic analyses with MrBayes v3.1.2 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003). Default priors of MrBayes 3.1.2 were used for MCMC parameters. We used three heated chains and one cold chain for all analyses and runs were started with random trees. Two independent MCMC runs were conducted with 4.106 generations for each run for all five 171 markers. Trees and parameters were sampled every 100 generation. Standard deviation of split frequency were used to assess stationarity, with the average standard deviation of split frequencies expected to approach zero when the two runs converge onto stationarity distribution (Ronquist et al., 2005). Additionally the potential scale reduction factor should approach one when runs converge. For each run the first 25% were discarded as burn-in. In addition, Mega v.5Beta was used to build a Maximum Likelihood tree using the Nearest Neighbour Interchange heuristic method. Missing data was treated with the “Partial Deletion” option with a site coverage cutoff value of 95% and using all codon positions. Both BI and ML trees were rooted with a sequence from Puffinus puffinus. At the intraspecific level, processes such as recombination can produce reticulate relationships. Such reticulations are not taken into account by molecular phylogenies because these are constructed under the assumption that the data arose through a branching process. However, phylogenetic networks can incorporate reticulations and display them (Huson, 1998; Posada & Crandall, 2001). We therefore used SplitsTree 4.11.3 (Huson & Bryant, 2006) to produce a phylogenetic network with our CO1 data. We used uncorrected ‘p’ genetic distances and the NeighborNet method (Bryant & Moulton 2004). Besides, we performed a bootstrap analysis with one thousand replicates. We also tested for recombination with the pairwise homoplasy index (PHI) statistics (Bruen et al., 2006) implemented in SplitsTree 4.11.3. Results The mean intra-taxon K2P distances recovered are shown in Table 5.2.3. Caledonica and leucoptera are virtually undifferentiated, based on net K2M distance (<0.01%), while they are both differentiated from Polynesian (Society /Rarotonga) birds by distances of respectively 1.6% and 1.5% (Table 5.2.4). Birds from Rarotonga/Society show a striking difference in their net distance with both brevipes groups defined here: while the difference is only 0.8% with Vanuatu birds, it is twice as more important with Fiji birds. 172 Table 5.2.3 Mean within group K2P distances in Gould’s and Collared petrels populations (and standard errors) Rarotonga Society brevipes Vanuatu brevipes Fiji caledonica leucoptera d 0.001 0.009 0.005 0.003 0.002 s.e. 0.001 0.003 0.004 0.001 0.001 Table 5.2.4 Net K2P distances between groups of CO1 sequences from Gould’s and Collared petrels below diagonal and standard errors above diagonal. Rarotonga_tahiti brevipes_Vanuatu brevipes_Fiji caledonica leucoptera Rarotonga_Society 0.004 0.005 0.005 0.005 brevipes_Vanuatu 0.008 0.001 0.006 0.006 brevipes_Fiji 0.016 0.001 0.004 0.004 caledonica 0.016 0.013 0.018 0.000 leucoptera 0.015 0.013 0.017 0.000 Table 5.2.5 shows pairwise K2P distances between all but caledonica and leucoptera birds. This table allows understanding the pattern of intra-group mean distance found in Table 5.2.3: among birds from Fiji, two exhibit much higher genetic distance than other birds in the study, while the third Fiji bird seems very close from specimens from Rarotonga/Society. A somewhat similar pattern is found in birds from Vanuatu, four birds collected near Mera Lava Island showing distance of less than 1% with Rarotonga/Society birds; three collected on Mera Lava and Tanna Island have a K2P distance comprised between 1.2 and 1.5% with Rarotonga/Society birds, and three from Tanna Island are separated by more than 2% from birds from Polynesia. The pattern between birds from Fiji and Vanuatu is more complex (Table 5.2.5). 173 Table 5.2.5 Pairwise K2P distances between CO1 sequences from brevipes birds from Vanuatu, Fiji and Rarotonga below diagonal (standard errors above diagonal). OR.023110 OR.023110_Rarotonga 2004-3-12_ Tahiti 2004-3-13_ Tahiti 216923_Mera Lava 216920_Mera Lava 216921_Mera Lava 222193_Mera Lava 215400_Mera Lava 216919_Mera Lava 336700_Tanna 336703_Tanna 336707_Tanna 336709_Tanna 336705_Tanna 250899_Kandavu 528336_Viti Levu 250893_Kandavu 2004-3-12 0.001 0.001 2004-3-13 216923 216920 216921 222193 215400 216919 336700 336703 336707 336709 336705 250899 528336 250893 0.001 0.004 0.004 0.004 0.005 0.005 0.005 0.007 0.007 0.012 0.011 0.008 0.005 0.008 0.008 0.000 0.004 0.004 0.004 0.005 0.005 0.005 0.007 0.007 0.012 0.011 0.008 0.005 0.008 0.008 0.004 0.004 0.004 0.005 0.005 0.005 0.007 0.007 0.012 0.011 0.008 0.005 0.008 0.008 0.000 0.000 0.004 0.004 0.000 0.000 0.005 0.009 0.007 0.007 0.004 0.000 0.000 0.003 0.005 0.002 0.002 0.000 0.005 0.009 0.006 0.004 0.004 0.000 0.004 0.004 0.004 0.003 0.005 0.006 0.009 0.009 0.006 0.004 0.000 0.005 0.006 0.005 0.007 0.007 0.011 0.009 0.010 0.000 0.008 0.006 0.003 0.000 0.006 0.012 0.006 0.005 0.005 0.008 0.005 0.003 0.006 0.009 0.009 0.005 0.004 0.000 0.005 0.000 0.014 0.000 0.003 0.000 0.000 0.007 0.011 0.006 0.008 0.006 0.007 0.006 0.017 0.013 0.011 0.000 0.009 0.006 0.007 0.013 0.009 0.008 0.011 0.007 0.008 0.005 0.001 0.000 0.004 0.004 0.004 0.007 0.007 0.007 0.000 0.007 0.007 0.007 0.000 0.005 0.008 0.008 0.008 0.004 0.008 0.004 0.012 0.012 0.012 0.004 0.002 0.007 0.011 0.012 0.012 0.012 0.000 0.002 0.005 0.008 0.005 0.014 0.014 0.014 0.000 0.000 0.007 0.007 0.000 0.003 0.015 0.015 0.015 0.008 0.008 0.011 0.015 0.011 0.011 0.000 0.020 0.020 0.020 0.013 0.013 0.013 0.020 0.020 0.013 0.020 0.020 0.020 0.020 0.020 0.007 0.007 0.013 0.013 0.007 0.013 0.000 0.007 0.029 0.021 0.021 0.021 0.010 0.006 0.012 0.020 0.009 0.009 0.003 0.015 0.027 0.006 0.008 0.008 0.008 0.004 0.004 0.004 0.000 0.008 0.004 0.000 0.011 0.020 0.007 0.015 0.008 0.008 0.008 0.000 0.000 0.000 0.008 0.008 0.000 0.000 0.008 0.000 0.013 0.016 0.008 0.042 0.040 0.040 0.000 0.007 0.012 0.011 0.012 0.012 0.017 0.011 0.013 0.013 0.018 0.008 0.000 0.000 174 In both ML and BI trees built with CO1 data, caledonica and leucoptera are differentiated from brevipes and from the birds of Tahiti. The ML analysis sorts the two taxa, but with non significant bootstrap support (<50%), while the BI analysis does not sort these specimens at all (Fig. 5.2.2). In both trees, birds from French Polynesia and Rarotonga are grouped with strong support. No significant recombination was detected by the PHI (P=0.29). Data are therefore appropriate for a phylogenetic network analysis. The analysis recovers two groups with a good support (bootstrap value =73.6) corresponding to the leucoptera/caledonica population on one side and to the brevipes on the other (Fig 5.2.3). One brevipes sample from Viti Levu, Fiji, is found with the leucoptera/caledonica. Within brevipes, the position of Polynesian birds (French Polynesia, Rarotonga) is confirmed as being slightly differentiated from other birds in the taxon. The network show a sorting of birds from Vanuatu: specimens from Mera Lava are grouped together while birds from Tanna in the same archipelago, although related to birds from Mera Lava, seem more distant. Specimens from Fiji exhibit a peculiar distribution in the network, one individual being grouped with Gould’s petrel taxa, while one is found related to birds from Mera Lava. The third specimen from Fiji is more distantly related to the other brevipes and is linked to birds from Tanna. Figure 5.2.3 Haplotype network obtained with CO1 in Pterodroma caledonica, P. leucoptera, and P. brevipes. Labels indicate island of origin and reference number for birds from AMNH. Boostrap support value for the split between caledonica/leucoptera and brevipes is shown leucoptera/caledonica P.spp Tahiti caledonica brevipes-Rarotonga leucoptera brevipes-Fiji brevipes-Vanuatu 175 A B leucoptera caledonica brevipes Vanuatu brevipes Fiji unknown Society brevipes Rarotonga 95 99 99 97 53 85 Tanna 336703 Viti Levu 528336 Tanna 336707 Tanna 336705 96 Tanna 336709 Tanna 336700 Kandavu 250893 56 Mera Lava 221193 96 Kandavu 250899 64 Mera Lava 216923 Mera Lava 216921 Mera Lava 216920 Mera Lava 216219 Mera Lava 215400 100 Mera Lava 216921 Mera Lava 222193 Kandavu 250899 Tanna 336703 Mera Lava 216923 Kandavu 250893 Mera Lava 215400 Mera lava 2116920 Tanna 336700 tanna 336705 Tanna 336709 Mera Lava 216919 Tanna 336707 Viti Levu 528336 100 176 0.01 0.5 Figure 5.2.2 Phylogenetic analysis of Pterodroma caledonica, P. leucoptera, P. brevipes and as yet unidentified Pterodroma based on complete (caledonica, leucoptera, unidentified birds) and partial sequences of CO1 mitochondrial gene. Both trees rooted with a sequence from Puffinus puffinus. Island of origin and museum label are given for brevipes samples from Fiji and Vanuatu A- Molecular Phylogenetic analysis by Maximum Likelihood method, based on the HKY+G model conducted in MEGA v.5Beta. Values above branches show bootstrap support; branches with less than 50% support are collapsed. B- Bayesian Inference phylogenetic analysis using the HKY+G model with MrBayes. Posterior Probabilities superior or equal to 95% are shown above branches. Discussion The mean distance between all specimens of caledonica and leucoptera together was 0.23%, exactly the same value than found by Kerr et al. (2007) in 643 species of North American birds. In addition, within group mean K2P distance was less than 0.3% in all groups but brevipes and the mean distance between each of the Gould’s petrel taxa and the birds from Rarotonga/Society and the brevipes groups was above 1.3%. In Scandinavian birds average intraspecifc distance of 0.24% was found (Johnsen et al., 2010) and Hebert et al. (2004) found a maximum intraspecific difference of 1.24% in 260 species of North American birds. Thus, this CO1 data comfort the view that Gould’s and Collared petrels are different species, but conclusions based on this single marker need to be treated cautiously and additional loci will be needed to draw firm conclusions. In addition, birds from Rarotonga/Society are well differentiated from caledonica and leucoptera and clearly placed within the brevipes group by our data. This is corroborated by both phylogenetic trees, as well as by the haplotype network. These birds from Polynesia are somewhat separated from other brevipes individuals, which call for further investigation. In particular, it would be worth investigating whether these Polynesian birds exchange genes with their Fijian and Vanuatu counterparts or if they constitute a genetically isolated population. The relationships between brevipes birds seem more complex. This might be due to the presence of different taxa, as expected based on morphological differences observed between birds in Vanuatu (Bretagnolle & Shirihai, in press). The haplotype network suggests that, although closely related, brevipes from North (Mera Lava Island) and South (Tanna Island) Vanuatu might be slightly differentiated genetically. In addition, the position of the three birds from Fiji in the haplotype network is intriguing: although finding one bird in each “group” of birds from Vanuatu is not surprising given the slight genetic difference between birds from North and South Vanuatu, the position 177 of the bird from Viti Levu grouped with the caledonica/leucoptera is unexpected. This could be explained by a misidentification of the specimen. However this bird was closely inspected by two of us (VB and HS) and its taxonomic identification did not seem erroneous based on phenotypic inspection. Alternatively, this bird might bear the sign of a hybridisation between Gould’s and Collared petrels, the hybrid exhibiting a brevipes-like phenotype (size, colourations) despite carrying some genetic material from Gould’s petrel maternal lineage. Cases of hybridisation between closely related Pterodroma taxa have been documented, of which probably the most interesting is the hybridisation involving one taxon from the Atlantic Ocean (P. arminjoniana) and one from the Pacific Ocean (P. neglecta) on Round Island, Indian Ocean. From microsatellite data, Brown et al. (2010) showed that there is admixture between the two taxa living on this island and that hybridisation is common. Brown et al. (2010) also documented the case of a hybrid chick that hatched from a mixed morpho-type pair of birds. Genetic analyses confirmed the chick to be a true hybrid. These authors consider this case as an example of secondary contact between two closely related species following long-distance dispersal. Given the closer proximity of Gould’s and Collared petrels breeding grounds and their apparent close genetic relationship, hybridisation between these two taxa seem plausible. However this hypothesis will need confirmation with more genetic data. Furthermore, these results might also be caused by the small fragments of CO1 gene used that could fail to recover the real divergence and relationships between the brevipes birds. Nevertheless, this preliminary work strongly make the case for additional investigations using complete sequences of CO1 gene, as well as multiple nuclear loci to investigate in more details the relationships between brevipes populations and their relationships with Gould’s petrel taxa. This study used DNA extracted from old museum specimens. As could be expected from working with specimens that were not prepared or stored with genetic studies in mind, this DNA was degraded making it difficult to recover fragments of more than 200bp and resulted in low amplification success, as also found by Zimmermann et al. (2008). Using small fragments of the CO1 gene as barcodes has proved successful in several taxa, particularly in insects (Hajibabaei et al. 2006). However these authors showed that these minibarcodes were not as successful as full-length barcodes and were likely to work only in restricted taxonomic groups. Here, we found that our restricted fragments of CO1 gene successfully segregated Gould’s and Collared petrels, two closely related species. Although targeting smaller overlapping regions for amplification or using DNA repair enzymes (Evans, 178 2007; Patel et al., 2010) could have permitted to recover a complete sequence of CO1 from our brevipes samples, this investigation show that even for closely related species limited fragments of the cytochrome oxydase 1 gene can be useful to identify species. This also points out the utility of CO1 barcode or mini-barcode to identify species or populations that deserve deeper investigations, a result also highlighted by Campagna et al. (2009) in their study of the genus Sporophila (Aves Procellariiformes). Acknowledgements This work was supported by the "Consortium National de Recherche en Génomique", and the "Service de Systématique Moléculaire" of the Muséum National d'Histoire Naturelle (CNRS UMS 2700). It is part of the agreement n°2005/67 between the Genoscope and the Muséum National d'Histoire Naturelle on the project "Macrophylogeny of life" directed by Guillaume Lecointre. We are deeply indebted to Joel Cracraft, Curator, Paul Sweet, Collection Manager, and Margaret Hart at the American Museum of Natural History (AMNH) for giving us access to the collections and collecting samples from brevipes. BG also acknowledges receipt of a Collection Study Grant from the AMNH. H. Shirihai also provided some samples collected in the AMNH and A. Cibois collected the samples on the birds in the Musée de Tahiti et des îles, Tahiti. Gilllian Stone from TePapa Museum, New Zealand provided the sample for the bird from Rarotonga. References Bandlet H.J., Forster P. and Röhl A., 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16, 37–48 BirdLife International, 2010. http://www.birdlife.org accessed on 10/10/2010 Bourne W.R.P., 1983. The appearance and classification of the cookilarias petrels. Sea Swallow 32: 67-71 Bretagnolle V., 2001. Le Pétrel de la Chaine Pterodroma (leucoptera) caledonica: Statut et menaces. Rapport de fin d’étude pour le compte de la Province Sud, Nouvelle Calédonie 179 Brooke M., 2004. Albatrosses and Petrels Across the World. Oxford University Press, New York Campagna L., Lijtmaer D., Kerr K.C.R., Barreira A.S., Hebert P.D.N., Lougheed S.C. and Tubaro P.L., 2009. DNA barcodes provide new evidence of a recent radiation in the genus Sporophila (Aves: Passeriformes). Molecular Ecology Resources doi: 10.1111/j.1755-0998.2009.02799.x deNaurois R., 1978. Procellariidae reproducteurs en Nouvelle Calédonie pendant l’été austral. Comptes Rendus de l’Académie des Sciences, Paris 287 Ser. D 269-271 Dubey B., Meganathan P.R. and Haque I., 2010. DNA mini-barcoding: An approach for forensic identification of some endangered Indian snake species. Forensic Science International: Genetics doi:10.1016/j.fsigen.2010.03.001 Eaton M.J., Meyers G.L., Kolokotronis S-O., Leslie M.S., Martin A.P. and Amato G., 2010. Barcoding bushmeat: molecular identification of Central African and South American harvested vertebrates. Conservation Genetics 11: 1389-1404 Evans T., 2007. DNA damage. NEB expressions 2.1, 1-3 Fullagar P.J., 1976. Seabird islands No. 35: Cabbage Tree Island, New South Wales. Australian Bird Bander 14: 94-97 Guindon S. and Gascuel O., 2003. A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood". Systematic Biology 52: 696-704 Hajibabaei M., Smith A.M., Janzen D.H., et al., 2006. A minimalist barcode can identify a specimen whose DNA is degraded. Molecular Ecology Notes 6, 959-964 Hebert P.D.N., Cywinska A., Ball S.L. and DeWaard J.R., 2003a. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London B Biological Science 270: 313–321. Hebert P.D.N, Ratnasingham S. and DeWaard J.R., 2003b. Barcoding animal life: Cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society of London B Biological Science 270: S596–S599 Hebert P.D.N., Stoeckle M.Y., Zemlak T.S. and Francis C.M., 2004. Identification of birds through DNA barcodes. PLoS Biology 2: 1657–1663 Hickerson M.J., Meyer C.P. and Moritz C., 2006. DNA Barcoding will often fail to discover new animal species over broad parameter space. Systematic Biology 55(5): 729-739 Hindwood K.A. & Serventy D.L., 1941. The Gould petrel of Cabbage Tree Island. Emu 41: 120 180 Huelsenbeck J.P. and Ronquist F., 2001. MRBayes: Bayesian inference of phylogeny. Bioinformatics 17: 754–755. Imber M.J., 1985. Origins, phylogeny and taxonomy of the gadfly petrels Pterodroma spp. Ibis 127:197–229 Imber M.J. & Jenkins J.A.F., 1981. The New Caledonian Petrel. Notornis 28(3): 149-160 Johnsen A., Rindal E., Ericson P.G.P., Zuccon D., Kerr K.C.R., Stoeckle M.Y., Lifjeld J.T., 2010. DNA barcoding of Scandinavian birds reveals divergent lineages in transAtlantic species. Journal of Ornithology DOI 10.1007/s10336-009-0490-3 Kerr K.C.R., Stoeckle M.Y., Dove C.J., Weigt L.A., Francis C.M. and Hebert P.D.N., 2007. Comprehensive DNA barcode coverage of North American birds. Molecular Ecology Notes 7: 535–543 Lohman D.J., Ingram K.K., Prawiradilaga D.M. et al., 2010. Cryptic genetic diversity in "widespread" Southeast Asian bird species suggests that Philippine avian endemism is gravely underestimated. Biological Conservation 143: 1885-1890 Moritz C. & Cicero C., 2004. DNA barcoding: promise and pitfalls. PLoS Biology 2: 1529– 1531 Patel S., Waugh J., Millar C.D. and Lambert D.M., 2010. Conserved primers for DNA barcoding historical and modern samples from New Zealand and Antarctic Birds. Molecular EcologyResources 10: 431-438 Posada D., 2008. jModelTest: Phylogenetic Model Averaging. Molecular Biology and Evolution 25(7): 1253-1256 Priddel D. & Carlile N., 2009. Key elements in achieving a successful recovery programme: a discussion illustrated by the Gould’s petrel case study. Ecological Management and Restoration 10: S97-S102 Ronquist F. and Huelsenbeck J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19(12): 1572-1574 Ronquist F., Huelsenbeck J.P. and Van der Mark P., 2005. MrBayes 3.1 Manual. http://mrbayes.csit.fsu.edu/. Accessed March 2010 Sibley C.G. and Monroe B.L., 1990. Distribution and taxonomy of birds of the world. Yale University Press, New Haven, USA Simon C., Frati F., Beckenbach A., Crespi B., Liu H. and Flook P., 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Annals of the Entomological Society of America 87(6): 651-701 181 Sullivan J., Joyce P., 2005; Model selection in phylogenetics. Annual Review of Ecology Evolution and Systematics 36:445-466 Tamura K., Dudley J., Nei M. and Kumar S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24(8): 1596-1599 Tavares E.S. & Baker A.J., 2008. Single mitochondrial gene barcodes reliably identify sisterspecies in diverse clades of birds. BMC Evolutionary Biology 8: 81 Thompson J.D., Higgins D.G. and Gibson T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22(22): 4673-4680 Will K.W., Mishler B. and Wheeler Q. D., 2005. The perils of DNA barcoding and the need for integrative taxonomy. Systematic Biology 54: 844–851 Zimmermann J., Hajibabaei M., Blackburn D., et al., 2008. DNA damage in preserved specimens and tissue samples: a molecular assessment. Frontiers in Zoology 5, 18 182 CHAPTER 6: DISCUSSION AND PERSPECTIVES Evolutionary biology has a major, yet not fully recognised importance in conservation (Hendry et al., 2010). Evolutionary biology can help to describe actual diversity accurately (fields of systematics, deSalle & Amato, 2004), but is also a framework to predict future patterns of biodiversity (fields of phylogeography and population genetics; Avise, 2000; Edwards, 2009). The objective of this PhD work was to unveil the complex evolutionary history and phylogeographic patterns of threatened gadfly petrels in a conservation biology context. This was achieved by investigating taxonomy, phylogenetic relationships, phylogeography and population history in these birds. The conservation genetics tools used in the study were based on recent methodological developments that allow an increased integration of phylogeography, molecular phylogenetics and population genetics to understand and describe the patterns of biodiversity. Indeed, Isolation with Migration models and species tree analyses are relatively new methods that have not yet been extensively used in the study of phylogenetic and phylogeographic patterns in Procellariiformes. 6.I Methodological aspects Because species are made of many genes, each of these recording a particular demographic history, all gene trees are embedded in the species tree depicting the relationships between species. Due to factors such as lineage sorting (or deep coalescence), horizontal gene transfer or gene duplication (Maddison, 1997), one particular gene tree will not necessarily match the species tree in which it is embedded. For instance, in Fig. 6.1, while the tree of gene 2 reflects the topology of the species tree for populations 1, 2 and 3, this is not the case for gene 1. Therefore, in systematic studies aiming at depicting the phylogenetic relationships between species and at defining species limits, integrating information carried by several genes will lead to a better picture of populations/species history than the use of a single marker (Edwards & Bensch, 2009). The use of nuclear DNA in addition to mitochondrial markers is thus crucial to progress in the knowledge of phylogeographic processes in taxa (Knowles & Carstens, 2007; Lee & Edwards, 2008; Edwards & Bensch, 183 2009). In this PhD, the use of mitcochondrial and nuclear DNA sequences with newly developed methods based on multispecies coalescence (species tree reconstruction) and multiloci Immigration with Migration model proved determinant in the analysis of complex phylogenetic and phylogeographic patterns. Figure 6.1 Potential discordance between species tree (in black) and genes trees (red and blue): in this case the tree of gene 1 does not depict the same history than the species tree in which it is embedded. Therefore, basing a phylogenetic study of populations 1, 2 and 3 only on this gene would likely lead to inferring inappropriate population/species limits Species tree reconstruction allowed inferring the validity of Beck’s vs. Tahiti petrel differentiation (Chapter 3) and to reconstruct the phylogenetic tree in genus Pseudobulweria while individual gene trees produced with the three markers used in this analysis did not recover the same relationships scenarios within the genus. In terms of phylogeographic and population differentiation pattern, the work presented in Chapter 4 estimates divergence time between gadfly petrels of Cape Verde, Madeira and Desertas much more recently (c. 37,000 and 110,000 years ago between feae and deserta and 184 between madeira and feae respectively) than what had been suggested before based on a single marker (>1.5 million years ago; Zino et al., 2008; Jesus et al., 2009; Table 6.1). Previous estimates of divergence between these taxa used a single mitochondrial marker (cytochrome b), a divergence rate estimation (Nunn & Stanley, 1998) that has been recently reassessed (Weir & Schluter, 2008) and estimated divergence based on this rate and on corrected genetic distance between taxa with this locus. This type of methodology, which albeit have proved useful over the years, can greatly mislead about the divergence and age of these taxa because it estimates a gene divergence rather than a population divergence. When using coalescent based analyses, such as bayesian estimation of time to most recent common ancestor (TMRCA), estimations differ slightly from those obtained by the previously described methodology, notably due to the uncertainty in the value of divergence rate of cytochrome b. For instance, preliminary analyses with software BEAST (Heled & Drummond, 2010) indicated a TMRCA of approximately 1.5 millions years ago between Madeira and Feae petrels with cytochrome b. This corresponds to the late Pliocene/early Pleistocene period and approximately matches previous estimates placing the split between madeira and feae in the late Pliocene (Zino et al., 2008). However, TMRCA differs from time of population divergence in that it represents the time that lineages shared a common relative, while divergence time corresponds to when populations genetically diverged from one another. Furthermore, TMRCA of different markers will differ, nuclear markers such as intron showing older TMRCA than smaller effective population-sized mitochondrial DNA. Also, TMRCA will usually be older than divergence time (Jennings & Edwards, 2005). However, rapid population structuring linked to high dispersal capacities of some taxa and Pleistocene climate fluctuations can lead to divergence times close to TMRCAs as found by Barton (2010) in striped skunks (Mephitis mephitis) in North America. Interestingly, in Table 6.1 all but Friesen et al.’s (2007a) divergence dates estimated gene divergences, and all indicated divergence times older than the last major galaciation in the late Pleistocene that started around 110-120,000 years ago. These estimates are also older than those of Friesen et al. (2007a) and those in the present thesis. Although these differences might reflect the reality, these estimates indicative of TMRCA rather than of populations divergence potentially overestimate the true population divergence time. Re-estimating these divergence dates with immigration with migration models (or other methodology calculating population split rather than genes’ TMRCA) and additional nuclear loci would possibly provide more recent estimates of differentiation. 185 Furthermore, all the studies presented in Table 6.1 used only one mitochondrial locus to estimate divergence times. These estimates were therefore based only on the maternal lineage histories rather than on the whole population histories. However, I found that the use of nuclear introns in addition to mitochondrial data allowed identifying patterns of lineage divergence more complex than expected in both Macaronesian and Gould’s petrels. The incongruent patterns found between the types of markers in both cases revealed that different genetics lineages within populations/species do not necessarily record the same signal, a crucial point to understand phylogeographic and population differentiation patterns. This limitation was pointed out by Friesen et al. (2007a) and Techow et al. (2009) and it would be worth investigating whether adding nuclear DNA data would lead to different estimates of population divergence and histories in the taxa studied by these authors. Table 6.1 Divergence date estimates between sister taxon in Procellariiformes based on mtDNA markers Taxa Divergence date Study (million years) Pterodroma madeira/feae 2.5 Zino et al., 2008 Pterodroma feae/deserta 1.75 Jesus et al., 2009 Puffinus mauretanicus/yelkouan 1 Genovart et al., 2007 Calonectris diomedea diomedea/C.d. borealis 1 Gomez-Diaz et al., 2006 Procellaria aequinoctialis/conspicillata 0.9 Techow et al., 2009 Macronectes giganteus/halli 0.2 Techow et al., 2010 Hydrobates pelagicus pelagicus/H.p.melitensis 0.55 Cagnon et al., 2004 Oceanodroma castro* 0.18 – 0.075 Friesen et al., 2007a Diomedea exulans/amsterdamensis 0.84 Milot et al., 2007 * sympatric populations in the Azores and Cape Verde Having said this, some limitations remain with respect to the methods used in this PhD. First, the number of loci, although providing an improvement relative to previous studies as mentioned above, is limited and it can be questioned whether more loci could have improved the results and estimates found. In two studies on Australian landbirds, Jennings & Edwards (2005) and Lee & Edwards (2008) used more than 30 nuclear loci to estimate population demographic parameters, timing of divergence, and phylogeography of these taxa in Northern Australia. Both studies found that the variance in most parameters estimates decreased quickly when using more than five loci and stabilised around 10-15 loci. Lee & Edwards (2008) pointed that estimating the number of loci to integrate to obtain parameters estimates 186 with as low a variance as possible is difficult to predict and that it is likely to change from one study to another. Nevertheless, these cases indicate that integrating between five and 10 more nuclear loci in the analyses presented in this thesis could probably allow to obtain more accurate estimates of population history and divergence parameters. The same is true for the species tree reconstruction analysis: adding more loci might allow to obtain better supported species trees or to obtain better resolution of the tree, even at shallow nodes (Edwards, 2009). However, limitation in the number of loci used will disappear altogether in the near future with the advent of new generation DNA sequencing methods delivering huge amount of data quickly and cheaply (Metzker, 2010). Furthermore, one of the limitations of the Isolation with Migration model as implemented in the software IMa2 resides in that several type of demographic history cannot be investigated. Indeed this method, by implying constant population size and constant gene flow (if any) between the populations prevents the estimation of parameters such as population structure, population size changes over time and potential variations of gene flow over time between the studied populations (Pinho & Hey, 2010). These authors suggest two solutions to overcome these issues: first, to add more parameters to the model. However, doing so requires having a great amount of data and computing time available, rendering this option difficult to implement. In the case of this PhD, without adding a very significant number of loci this option would not work: the data available was already very limited to estimate the migration and population parameters and adding more parameters would result in unreliable results; second, to simulate data with known violations of the assumptions of the IM model and then evaluate the performance of the program on this simulated data set. Once again, this requires a lot of computing time. Over the last years, a tremendous progress has been made in the development of statistical and computational tools for reconstructing population sizes and trends through time, estimating divergence times, migration rates and integrating all these parameters. Some promising methods are based on Approximate Bayesian Computation (ABC). These methods allow overcoming two limiting factors of likelihood approaches based on coalescent theory, i.e. the restriction to simple evolutionary models due to the difficulty to calculate likelihood functions, and the use of ever increasing amount of molecular data that is supposed to increase even more in the near future with the high-throughput DNA sequencing. For a general discussion of ABC and its use see Beaumont (2010), Csillery et al. (2010) for a 187 presentation of the method and software implementing it, and Lopes & Boessenkool (2010) for a review of conservation studies using ABC approach to infer population history of endangered populations or Cornuet et al. (2010) for the presentation of a user-friendly software, DIYABC. At this stage however, most ABC softwares still suffer from some of the limitations of the IM model cited above. However, it is likely that those issues will be solved soon.. For instance, DIYABC at the moment allows inferring effective population size, timing of population size changes and timing of divergence but assumes the absence of migration after population diverged. Next developments will aim at allowing estimation of migration among diverged populations (Cornuet et al., 2010). This will then constitutes a very powerful tool to investigate more realistically population history of related taxa. 6.II Taxonomic aspects Species correspond to a fundamental biological reality (Mayr, 1982) and accurate delineation of their limits and thorough taxonomic identification plays an essential role in conservation biology (Brooks & Helgen, 2010). Indeed, because taxonomic identification can sometime fail to capture all phylogenetic diversity, for instance in cases of morphologically criptic lineages or single marker-based investigations (Rissler et al., 2006), lineages or species can be overlooked and disappear along with their evolutionary history and diversity. Many studies (cf. Introduction Part 2 II.2) have tackled the phylogenetic and taxonomic issues in the order Procellariiformes with varying success. For instance, Browne et al. (1997) found clear support for splitting Galapagos petrel populations from the Galapagos and Hawaii in two different species (P. phaeopygia and P. sandwichensis), and Techow et al. (2009) identified what they considered sufficient difference between spectacled petrel and white-chinned petrels (Procellaria conspicillata and P. aequinoctialis respectively) to elevate the former to full species status. On the other hand, Abbott & Double (2003) although finding some arguments in support of species status of both shy and white-capped albatrosses (Thallasarche cauta and T. steadi) reserved their judgement on the taxonomic treatment of these birds due to insufficient evidence. Similarly, Austin et al. (2004) although clarifying the phylogeny of the little/audubon’s shearwater complex (Puffinus assimilis/lherminieri) by mean of mitonchondrial DNA analyses recommended more thorough analyses of the complex with additional markers such as nuclear genes to refine and confirm their findings. 188 In the present study, the issues concerning the validity of the genus Pseudobulweria and its close affinity with genera Bulweria and Puffinus within Procellariidae were solved by using both nuclear and mitochondrial markers and species tree reconstruction that integrate phylogenetic histories of several genes. Furthermore, this study showed that the recently rediscovered Beck’s petrel should receive full species status. This ends the debate whether it is a subspecies of Tahiti petrel. However, for Tahiti petrel, as was the case for albatrosses (Abbott & Double, 2003) and shearwaters (Austin et al., 2004), the evidence found here was not sufficient to fully solve the issue concerning the validity of its currently defined subspecies and additional investigations are needed in this respect. Additionally, the multiloci analysis of Gould’s petrel showed that the two currently recognised taxa (subspecies) surely do not deserve full species status and even the validity of the two subspecies is questionable if based strictly on phylogenetic evidence (but see 6.V). Furthermore, the investigation of sole mitochondrial CO1 data underlined the separation of Gould’s petrel and Collared petrel. However, this separation did not appear in nuclear loci between Gould’s taxa and birds collected in French Polynesia indicating a pattern similar to that of Macaronesian gadfly petrels: a marked differentiation of mitochondrial lineages but no differentiation of nuclear lineages. This investigation also pointed towards a complex phylogenetic pattern in the Collared petrel, which reflects the morphological variations found between different populations (Bretagnolle & Shirihai, in press). Hence, further investigations with additional loci are needed. Similarly, despite using several markers, the investigation on Macaronesian petrels did not allow to settle the issue concerning the taxonomic status of gadfly petrels in Northeast Atlantic. Previous work supported the definition of three species (Jesus et al., 2009). Here, as suggested by Abott & Double (2003) for shy and white-capped albatrosses, the results indicate that gadfly petrels of Macaronesia are seen now in the process of speciation (grey zone on Fig 1.2; deQueiroz, 1998, 2007), so that for some species concepts diagnostic levels of divergence (such as in vocalisations, morphology or nuclear DNA lineages) have not yet been reached causing the taxonomic confusion. 189 6.III Genetic diversity in Procellariiformes Table 6.2 summarises haplotype and nucleotide diversity found in several Procellariiformes with the two mitochondrial genes used in the course of my PhD (CO1 and Cytb) and with Control Region and ATPase genes. A striking feature with CO1 is that all taxa (but deserta) investigated in this PhD exhibit (much) higher haplotype (h) and nucleotide (π) diversity than Cook’s petrel (Pterodroma cooki). Population size of the latter in the main breeding colony (Little Barrier Island, New Zealand) was estimated at c. 286,000 breeding pairs (Rayner et al., 2010), a lot more than the estimated Ne found for Gould’s and Macaronesian petrels. In another colony of Cook’s petrel, Rayner et al. (2010) found a single haplotype for all birds sampled, indicating that overall, this taxon has a relatively low genetic diversity despite its population size. Interestingly, values found in Deserta petrel are quite close to those of Cook’s petrel despite the much smaller population size of the former. Although the investigation of Macaronesian petrels presented in Chapter 4 hypothesised a founder event with small number of birds or a bottleneck to explain the pattern observed in deserta, this is unlikely to be the case in P. cooki, a taxon formerly found in all New Zealand, whose present colonies are but the remnants of the former breeding range (Rayner et al., 2010). In this case, low genetic diversity observed could be the result of the fragmentation of a formerly larger breeding range: actual Cook’s petrel breeding colonies correspond to both extremes of a formerly wider range that encompassed all of New Zealand (Rayner et al., 2010). These authors suggested that historically populations of Cook’s petrel were probably connected through a haplotype cline with the populations at the range extremes showing different haplotypes. With the loss of the intermediate populations, these two remaining populations retained only a small portion of the former genetic diversity of the species and different haplotypes. A similar scenario could be proposed to explain the lower genetic diversity of deserta petrel: considering the potential presence of feae petrel in the Canary and Selvagens Islands as well as Cape Verde and the Desertas Islands, these last two populations are located at both extremes of the species range. Habitat fragmentation and disappearance of the intermediate populations leave the two populations separated. The larger Cape Verde population retains a high genetic diversity and multiples colonies on several islands of this archipelago, while the small population on Bugio becomes isolated with low genetic diversity. However, given the genetic diversity pattern and 190 haplotype network structure observed in deserta, the scenario involving a bottleneck or founder event seems more likely. With Cytb, most taxa exhibit high h and low π values that are indicative of population growth following a bottleneck or founder event (Grant & Bowen, 1998). The exceptions are deserta and spectacled petrel (as well as some colonies of white-chinned petrel; Techow et al., 2009) with low h and low π values indicative of a recent bottleneck or founder event (Grant & Bowen, 1998). The Taiko (Lawrence et al., 2008a) and southern giant petrels (Techow et al., 2010) have both high h and π (>0.5) that are the sign of large, stable populations (Grant & Bowen, 1998). As underlined by Lawrence et al. (2008a, 2008b), high genetic diversity in the Taiko is puzzling and probably the result of a very large population that collapsed quickly thus still showing genetic diversity from before the bottleneck it is now experiencing. We saw that the haplotype diversity in madeira is very high for a species with less than 30 breeding pairs (Zino et al., 2008). This is likely due to a rapid decline of a formerly larger population that experienced a Pleistocene population increase, as the balearic shearwater (Puffinus mauretanicus; Genovart et al., 2007). Southern giant petrels have experienced bottlenecks in the past but most likely before the last major ice-age (Techow et al., 2010) and thus enough time have passed for the population to grow and stabilise. Other studies of Procellariiformes using different markers (e.g. Control region of the mitochondrial genome or ATPase) show a pattern of high genetic diversity with h and π values (Table 6.2). Like madeira and the Taiko, the short-tailed albatross (Phoebastria albatrus), is experiencing a severe bottleneck since the 1940’s (it was believed extinct in the 1940’s) and is now listed as Vulnerable (Kuro’o et al., 2010). Like the two former taxa, it exhibits also very high h (0.96) and π (1.3%) despite having been as low as 50-60 individuals in the 1950’s (Kuro’o et al., 2010). All these studies show high levels of genetic diversity of Procellariiformes in general which is corroborated by results obtained in this thesis, although Milot et al. (2007) found low level of diversity in polymorphic loci and heterozygosity in wandering and Amsterdam albatrosses (Diomedea exulans/amsterdamensis respectively) and using amplified fragment length polymorphisms. These levels can be conserved for several generations even when facing extreme bottlenecks, like experienced by short-tailed albatross, Taiko, balearic shearwater or Madeira petrel. The different pattern shown by taxa like deserta or Spectacled petrels could be explained by a prolonged bottleneck that lasted long enough to finally erode a 191 formerly larger genetic diversity. Alternatively such low genetic diversity can be the result of a recent founder event from a small group of specimen through colonisation of a new habitat or fragmentation of a large breeding range as proposed for the Cook’s petrel (see above). Table 6.2 Haplotype (h) and nucleotide (π) diversity found in Procellariiformes with mitochondrial genes CO1, Cytb, Control Region and ATPase. Numbers in brackets indicate sample sizes; Ranges indicated for studies with multiples populations in which no mean values were provided CO1 Pterodroma deserta (94) Pterodrma madeira (59) Pterodroma feae (59) Pterodroma caledonica (5) Pterodroma leucoptera (16) Pterodroma magentae(90) Puffinus mauretanicus (105) Procellaria aequinoctialis (6 to 30) Procellaria conspicillata (23) Macronectes giganteus (74) Macronectes halli (51) Hydrobates pelagicus (65) Pterodroma magentae(90) Oceanodroma castro (5 to 49) Puffinus mauretanicus (105) Thalassarche melanophrys (56) Thalassarche chrysostoma (50) Phoebastria immutabilis (358) Phoebastria albatrus (44) Fulmarus glacialis (9 to 20) ATPase Pterodroma deserta (89) Pterodrma madeira (58) Pterodroma feae (57) Pterodroma caledonica (30) Pterodroma leucoptera (43) Pterodroma cooki (26) Cytb Taxon (n) Control Region Marker Pterodroma phaeopygia (11 to 51) h π (%) Study 0.45 0.78 0.69 0.82 0.71 0.34 0.07 0.23 0.15 0.35 0.38 0.1 Gangloff et al., in prep Gangloff et al., in prep Gangloff et al., in prep Gangloff et al., in prep Gangloff et al., in prep Rayner et al, 2010 0.22 0.54 0.56 0.9 0.93 0.68 0.81 0-0.83 0.49 0.78 0.73 0.7 0.03 0.09 0.11 0.46 0.39 0.58 0.4 0-0.29 0.09 0.5 0.2 0.4 Gangloff et al., in prep Gangloff et al., in prep Gangloff et al., in prep Gangloff et al., in prep Gangloff et al., in prep Lawrence et al., 2008b Genovart et al., 2007 Techow et al., 2009 Techow et al., 2009 Techow et al., 2010 Techow et al., 2010 Cagnon et al., 2004 0.92 0.78-1.0 0.97 0.91 0.97 0.99 0.96 0.76-0.92 2.23 0.5-5.7 3.72 2.21 3.0 4.5 1.3 - Lawrence et al., 2008b Friesen et al., 2007a Genovart et al., 2007 Burg & Croxall, 2001 Burg & Croxall, 2001 Young, 2009 Kuro’o et al., 2010 Burg et al., 2003 - 0-0.05 Friesen et al., 2006 192 6.IV Population differentiation in Procellariiformes 6.IV.1 Timing of divergence The investigations conducted in this thesis found that taxa in Northeast Atlantic gadfly petrels diverged recently, within the last 100,000 years, while the divergence time estimates for Gould’s petrel fall around 40,000 years ago. These divergence time estimates can be compared to investigations in other Procellariiformes keeping in mind the caveats and limitations mentioned in 6.I. Techow et al. (2009, 2010) estimate the timing of divergence of Spectacled and White-chinned petrels and of Southern and Northern giant petrels around 0.9 and 0.8 million years ago respectively, in the middle Pleistocene. In European storm-petrels, Cagnon et al. (2004) estimated that the two Atlantic and Mediterranean lineages coalesced between 0.35 and 0.55 million years ago, depending on the evolutionary rate used. In the same region, Gomez-Diaz et al. (2006) place the separation between the two Cory’s shearwater subspecies (Calonectris diomedea diomedea living in the Mediterranean Sea, and C.d. borealis found in the Atlantic Ocean) at 1 million year, and the separation of Cape Verde shearwater (C. edwardsii) from C.d. borealis around 0.7 million years ago. In Puffinus, Genovart et al. (2007), estimated the divergence between balearic and yelkouan (Puffinus yelkouan) shearwaters to about one million years ago. In the Azores and Cape Verde, Frisen et al. (2007a) estimated the divergence between hot and cool season populations of band-rumped storm-petrel (Oceanodroma castro) between 110,000 and 180,000 years ago in the Azores and around 73,000 years ago in Cape Verde. All divergence events in these taxa occurred during the Pleistocene and, apart from Oceanodroma castro and the divergence estimates in this thesis manuscript, all divergences were dated well before the last major glaciation initiated around 110,000 years ago. These estimates underline the importance of the Pleistocene glacial/interglacials cycles in the differentiation of taxa. The exception concerns the estimates obtained in genus Pseudobulweria (Chapter 3). Within this genus, divergence between the black-vented and white-vented taxa was estimated around 6-7 millions years ago, i.e. at the end of Miocene, during a period of marked ecological changes. However, this estimate is based on very low samples sizes and only three loci and therefore need to be taken with great caution. It would 193 seem however that this genus did not experience as great a number of taxon divergence as observed in other Procellariiformes genera like Pterodroma or Puffinus during the Pleistocene. 6.IV.2 Differentiation Important debates have occurred for the last fifty years about the main processes of population differentiation, and geographic separation has appeared to be associated to most speciation events (Fitzpatrick et al., 2009; Coyne & Orr, 2004), including in birds (Phillimore et al., 2008). In a review on population differentiation in seabirds, Friesen et al. (2007b) showed that land barriers to gene flow played an important role in these organisms: populations separated by contemporary or historical land barriers showing genetic differences and phylogeographic structure (e.g. case of masked boobies Sula dactylatra separation by the isthmus of Panama; Steeves et al., 2005), suggesting that allopatric differentiation is common in this group. However, Friesen et al. (2007b) also showed that, in the absence of physical barriers such as land masses, geographic distance and simple isolation by distance model did not provide conclusive explanations for the observed population genetic structure in seabirds. For instance, within archipelagos, or even within islands, several species show genetic structure between populations, including Procellariiformes, e.g. two storm-petrels (bandrumped storm-petrel in the Galapagos islands and the Azores in Macaronesia, Smith et al., 2007; and Leach’s storm-petrel Oceanodroma leucorhoa on Guadalupe Island, Atkey & Gulavita unpubl. in Friesen et al., 2007b) and one gadfly petrel, the Galapagos petrel (Friesen et al., 2006). With reference to the role of isolation by distance, the present work corroborated the points made by Friesen et al. (2007b): (i) in both Gould’s and Macaronesian petrels geographic separation and isolation by distance do not explain the observed structure, or lack thereof, between populations: in Gould’s petrel, despite an approximate distance of 2,000 km between breeding sites of the two taxa, populations are not genetically differentiated and only weakly structured in nuclear loci, while in the Northeast Atlantic, birds breeding at only 40 km from each other are more differentiated between them than with the third taxon breeding c. 1,000 km away, and; (ii) as is the case for storm-petrels, gadfly petrels show complex genetic and phylogeographic structure in both the Galapagos Islands (Pterodroma phaeopygia) and Macaronesia Islands (P. feae, P. madeira and P. deserta) in the absence of apparent barrier to gene flow. 194 As one potential explanation to population structure and differentiation in the absence of past or contemporary physical barrier to gene flow, Friesen et al. (2007b) suggested that distribution at sea during the nonbreeding season provides a barrier to gene flow in these organisms. Indeed, in their review, all but one species with phylogeographic structure also had population specific nonbreeding distributions. However, although potential interpopulation mixing is necessarily reduced by differences in at-sea distribution, this phenomenon alone can not explain most of the patterns observed. For example, in the Eastern tropical Pacific Ocean, Spear & Ainley (2007) pointed that many of the 26 distinct taxa of Hydrobatidae, though nesting allopatrically, mix strongly at sea in both breeding and/or nonbreeding seasons. In Gould’s and Macaronesian petrels, if the hypothesis of Friesen et al. (2007b) holds, we can expect to find that Gould’s petrel taxa share the same range, while gadfly petrels from Macaronesia do not during the non-breeding season. In Gould’s petrel, caledonica are known to disperse in the Eastern Pacific during the non breeding season and, when breeding, these birds go as far south as southern New Zealand, while leucoptera is mostly observed in the Tasman Sea. Although much still need to be investigated about the distribution at sea of these two taxa, it is likely that the two populations share the same distributions, at least some times of the year. Work is underway with GLS to define more precisely the movements of these birds. Similarly, in the Atlantic, distribution at sea of gadfly petrels from Macaronesia during the nonbreeding season is mostly unknown and analyses are underway from GLS data to identify their movements at sea. In addition to geographic separation and nonbreeding distribution, Friesen et al. (2007b) suggested alternative causes of population differentiation. In particular, the importance of founder events and bottlenecks is underlined by several authors advocating that these could cause drift-induced genetic differentiation, as suggested for the shy and whitecapped albatrosses (Abbott & Double, 2003) and gull (Larus cachinnans-fuscus; Liebers et al., 2001) cases. However, as stated in the Introduction (Part 1 I.4.1), drift alone is unlikely to shape population differentiation and other factors are expected to be involved, such as adaptation to local environmental conditions and selection. The results presented here do not allow inferring conclusions on this point. Although bottlenecks occurred in populations of petrel studied in Northeastern Atlantic and Western 195 Australia/New Caledonia, evidence is lacking to disentangle the role of these population lows and that of selection and local adaptation in the differentiation of taxa. 6.IV.3 Pleistocene and oceanographic conditions This PhD work showed that both Gould’s petrel in Oceania and gadfly petrels of the Northeast Atlantic diverged during the last major ice age of the Pleistocene. The influence of Pleistocene, marked by a succession of intense glacial/interglacial periods, on organisms is well known, and has been well documented in seabirds: e.g. Common Murres Uria aalge Moum et al., 1991; Fairy prion Pachyptila turtur, Ovenden et al., 1991; Thick-billed Murre Uria lomvia, Birt-Friesen et al., 1992; Lesser Black-backed Gulls Larus fuscus, Liebers & Helbig, 2002; European storm-petrel Hydrobates pelagicus, Cagnon et al., 2004; Calonectris shearwaters, Gomez-Diaz et al., 2006; Giant petrels Macronectes sp., Techow et al., 2010. The usual Pleistocene-driven population differentiation scenario (Avise & Walker, 1998) involves range fragmentation during glaciations and survival of population in isolated refugia that evolve separately under different environmental conditions. During interglacial events, these populations can grow and secondary contact can occur, either erasing the differences between populations when these have not diverged enough or reinforcing isolation when divergence is more pronounced. The importance of Pleistocene on bird populations at various latitudes is debated, some authors arguing that in tropical/subtropical regions, where the influence of glaciations is likely to have been less important, one should expect a stronger population differentiation than in temperate/high latitude regions that were only recently recolonised following deglaciation (Friesen et al., 2007b). In the Atlantic Ocean, Macaronesia Archipelagos are located in temperate/subtropical latitudes. Oceanographic patterns around these islands are particularly complex: the Azores Current splitting from the North Atlantic Current flows south and then eastwards. One branch of the current passes north of and around Madeira, then heads south between Canary Islands and Africa, before turning west at the level of Cape Verde and entering the North Equatorial Current. (Jia, 2000; Hernández-Guerra et al., 2001). Cyclical and seasonal variation in current strength, as well as formation of eddies add complexity to the system. In addition, westward flow also occurs north and south of the Azores current (Cromwell et al. 1996; Alves & Verdière 1999; Johnson & Stevens 2000; Zhou et al., 2000; Alves et al. 2002). These current 196 patterns were formed during and following the closure of the Panama isthmus (≈3 million years ago) and are supposed to have stayed relatively stable since then, even during Pleistocene glacial maxima (Pflaumann et al. 2003). Westward flows might have been dominant at some stages in the Pleistocene due to a weaker Gulf Stream (Rind et al. 2001). Furthermore, Madeira Island is located on the sub-tropical front where cold waters from the north meet warm waters from the south and sea surface temperatures are markedly different between north and south coasts of the island (Zhou et al., 2000; Caldeira et al., 2002). Also, localised upwellings do occur near the coast of the islands, in particular between Madeira and the Desertas Islands and several eddy corridors exist, north and south of the Azores front, south of Madeira island, and near Canary Islands (Sangra et al., 2009). Furthermore, both Azores and Canary current are constrained between narrow bands due to sea-level gradients (Zhou et al., 2000). In addition, during the late Pleistocene, ice-cap instability in northern Atlantic have several time led to the discharge of extensive amount of ice in the ocean, thus perturbing the oceanic conditions, in particular sea surface temperatures, in the region (Lambeck et al., 2002). Oceanographic conditions around Madeira are thus particularly complex and resemble the situation described in the eastern Pacific by Spear & Ainley (2007) where these authors proposed that these environmental gradients played a role in the differentiation of the 26 storm-petrel taxa present in that region. On the other hand, oceanic conditions in the Tasman Sea seem simpler. North of New Caledonia, between Vanuatu and the Solomon islands, South Equatorial Current flowing from the East collides with the Queensland plateau which creates the East Australian Current (EAC). The EAC flows south and around 32°S its upperlayer turns east and crosses the Tasman Sea, forming the Tasman Front. The latter forms different meanders and eddies mostly related to the north/south ridges present in the basin (Lord Howe rise, New Caledonian Trough, Norfolk ridge, from West to East). Some deeper layers of the EAC continue south along with several eddies (Bostock et al., 2006). Several studies have shown that during the last glacial maxima, the Tasman Front moved north, up to 26°S (Kawahata, 2002; Bostock et al., 2006), but that some flow continued southward. Around 11,000 years BP the front moved south progressively reaching its current location. In this oceanic and historical context the investigation in this thesis found a strong population differentiation in mitochondrial lineages in gadfly petrels of the Northeast Atlantic despite the recent divergence estimated for these taxa. However that divergence corresponded to Pleistocene climatic fluctuations impacting oceanic conditions (mainly SST). On the contrary, in the simpler oceanic context of the Tasman Sea, no population divergence was 197 detected in Gould’s petrel taxa despite similar estimated divergence times. Although this remains very speculative due to lack of evidence and is not directly based on the findings of this PhD, one can then hypothesise that environmental conditions at sea combined with geographic isolation were the primary factors of differentiation in gadfly petrels of Macaronesia, following the above mentioned Pleistocene population differentiation scenario: Pleistocene glaciation-induced isolation of populations were accompanied by an adaptation to different oceanographic conditions promoting lineages divergence. On the other hand, Gould’s petrel populations, which did not experienced as extreme climatic and oceanographic variations due to their tropical distribution, did not differentiate, despite breeding approximately 2,000 km apart. 6.V Implication of this research for conservation In Pseudobulweria, given the conservation status of Beck’s petrel, as well as the genetic, morphological (albeit based on small sample sizes) and behavioural differences found between this taxon and Tahiti petrel, it seems necessary to suggest full species status for these birds. By rendering conservation needs more visible for this taxon, species rank will help promote conservation actions and measures that will beneficiate not only this taxon, but the whole fragile forest ecosystems where it is supposed to breed in New Ireland. Full species status might also avoid Beck’s petrel to suffer the same fate than the extinct Tuatara subspecies: although two species of Tuatara (Sphenodon spp.) were described in 1875, including one with two sub-species, subsequent taxonomic grouping together of these species led to the demise of several populations due to lack of attention. When Daugherty et al. (1990) reassessed the taxonomy of the group and reinstated the two species, one was surviving only on one island due to extinction of its other populations and one subspecies described in the 19th century had gone extinct. In the case of gadfly petrels from Macaronesia on their way to speciation (if not already full species), this study showed that these taxa are three different Evolutionary Significant Units. Based on this conclusion, a reassessment of the conservation status of Deserta and Feae petrels and further conservation actions are needed. Indeed, the two taxa are still considered subspecies of Feae petrel by Birdlife International (2010) and the species is considered Near Threatened. Taking into account their evolutionary differences and isolation 198 leads necessarily to changing the status of Deserta petrel given its very restricted range on only one island and its small population size. The situation in Gould’s petrel is even more complex. Winker & Haig (2010) argue that good subspecies, even if not characterised with molecular characters, are worth conserving and could be important to preserve the evolutionary history and potential of the considered taxa. Based only on the genetic results presented in Chapter 5, the two taxa are not differentiated. The validity of the subspecies is further questioned by the subtle morphological variations used by Imber & Jenkins (1981) and the large overlap between the populations. However, following Winker & Haig (2010), ecological differences observed between the two taxa (different breeding ranges, different breeding dates) support the validity of the actual taxonomy. Given these ecological differences, managing these populations as different units is therefore needed. However managers should bear in mind the genetic characteristics uncovered in this study, i.e. the lack of lineage sorting between the populations, the lack of population structure in maternal lineages and the weak but significant structure in nuclear lineages. The work undertaken in this PhD confirms the importance of conservation genetics for the protection of petrel taxa and for the conservation of biodiversity in general. Only by identifying phylogenetic lineages and by understanding their history can it be possible to effectively conserve the current biodiversity. Failing to incorporate aspects of phylogenetics or phylogeography as shown in this PhD can lead to incomplete knowledge of the diversity to be protected and to an inaccurate understanding of patterns of taxonomic diversity. 6.VI Perspectives This work shed light on several taxonomic and phylogeographic aspects in gadfly petrels. The results opened up many interesting and urgent questions which deserve further attention to successfully protect the taxa in question. The presented investigations could, for instance, be complemented by work outlined in the following. In relation with the three groups of gadfly petrels studied in this thesis, specific investigations would be needed: (i) in Pseudobulweria: population genetic analyses are 199 needed once the breeding grounds of the three Critically Endangered taxa are discovered to describe their population structure and history and solve the issue surrounding the validity of Tahiti petrels subspecies; (ii) in gadfly petrels from Macaronesia: larger genetic sampling (i.e. more nuclear loci) and taxon sampling (i.e. adding related taxa from the Atlantic Ocean) will help disentangling the complex population and phylogeographic history of Pterodroma in this ocean basin; (iii) in Gould’s and Collared petrel, as for Macaronesian petrels, sampling of more loci and of all populations is needed to elucidate the phylogenetic relationships between these taxa. In addition to these specific lines of research, three directions could be followed on a larger scale: 1- To gain a better understanding of patterns of differentiation in petrel populations, investigating widely distributed taxa such as the species complex Pterodroma arminjoniana/neglecta/heraldica/alba and comparing results with those presented in this thesis would be highly informative. These taxa are found the Pacific, Atlantic and Indian Oceans and phylogenetic relationships between them are particularly difficult to disentangle. Indeed, they exhibit large ranges, lack of mitochondrial lineage sorting (Brown, 2008; Hakoun, 2009; Fig. 6.2) and apparent hybridisation of two or three taxa on Round Island in the Indian Ocean following long distance colonisation by P. neglecta and P. arminjoniana (Brown, 2010). Undertaking detailed investigation of evolution of populations and gene flow in the various populations of these taxa would provide a comparison to the patterns observed in the more restricted taxa studied in this work. 2- Current distribution of gadfly petrels and other Procellariidae is certainly very different to pre-human colonisation of oceanic islands throughout the world (Steadman, 2006). Genetic analyses of bone subfossils have the potential to bring important information about past distribution and genetic diversity of taxa. In seabirds this has already been implemented in petrels (Lawrence et al., 2008a) and boobies (Steeves et al., 2010). Undertaking such analyses on bones from different ocean basins (e.g. Atlantic Ocean; Olson, 1975, Rando & Alcover, 2008; or Pacific Ocean; Steadman, 2006) would help understanding former petrel distributions and diversity. Compared to other such studies on seabird would inform about past and current patterns of differentiation in highly vagile organisms such as seabirds. 200 3- Sets of primers to amplify nuclear introns have now been produced (e.g. Backström et al., 2008; Kimball et al., 2009) and techniques to amplify nuclear anonymous markers described (Jennings & Edwards, 2005; Lee & Edwards, 2008). These cuold be used with petrel DNA obtained during this PhD (more than 1300 DNA samples from 45 different Procellariiformes taxa). Combining such data with published genes sequences with species tree reconstruction techniques such as *Beast (Heled & Drummond, 2010) would make the construction of a complete phylogeny of Procellariiformes resolved at the genus, and maybe species, level at hand. 201 Figure 6.1. Haplotype networks obtained with P. arminjoniana, P. neglecta, P. heraldica and P. alba in two distinct studies, with different mitochondrial markers. A- From Brown, 2008. Cytochrome b statistical parsimony network of haplotypes from Round Island (RI) and Trindade (T). Haplotype name (upper value) and haplotype frequency (lower value) are given within each circle. Slashes diagonal to the lines show the number of inferred base substitutions between haplotypes when >1. Small filled circles represent hypothetical unsampled haplotypes. Shaded areas show the proportion of each haplotype assigned to each population B- From Hakoun, 2009. Median Joining haplotype network obtained with gene ND3 in birds from Pacific, Indian and Atlantic oceans. Red numbers show position and number of mutations between haplotypes. Circles size proportional to number of individual specimens sharing that haplotype. A B 202 References Abbott C. and Double M., 2003. Phylogeography of Shy and White-capped Albatrosses inferred from mitochondrial DNA sequences: implications for population history and taxonomy. Molecular Ecology 12: 2747–2758 Austin J.J., Bretagnolle V. and Pasquet E., 2004. Aglobal molecular phylogeny of the small Puffinus shearwaters and implications for systematics of the Little-Audubon’s shearwater complex. The Auk 121(3): 847-864 Alves M., Gaillard F., Sparrow M., Knoll M. and Giraud S. 2002. Circulation patterns and transport of the Azores Front Current system. Deep Sea Research II 49: 3983–4002 Alves M.L.G.R. and Verdière A.C., 1999. Instability dynamics of a sub-tropical jet and applications to the Azores Front Current System: eddy-driven mean flow. Journal of Physical Oceanography 29: 837–864 Avise J.C., 2000. Phylogeography. The history and formation of species. Harvard University Press, Cambridge, Massachusetts Avise J.C. & Walker D.E., 1998. Pleistocene phylogeographic effects on avian populations and the speciation process. Proceedings of the Royal Society of London B Biological Science 265: 457-463 Backström N., Fagerberg S. and Ellegren H., 2008. Genomics of natural bird populations: a gene-based set of reference markers evenly spread across the avian genome. Molecular Ecology 17: 964–980 Barton H.D., 2010. Disease ecology of rabies in the great plains: synthesizing the effects of viral properties, host attributes, and landscape on disease emergence. PhD thesis, Kansas State University, USA Beaumont M.A., 2010. Approximate Bayesian Computation in Evolution and Ecology. Annual review of Ecology, Evolution and Systematics 41: 379-405 BirdLife International, 2010. http://www.birdlife.org accessed on 10/10/2010 Birt-Friesen V.L., Montevecchi W.A., Gaston A.J. and Davidson W.S., 1992. Genetic structure of thick-billed murre (Uria lomvia) populations examined using direct equence analysis of amplified DNA. Evolution 46: 267–272 203 Bretagnolle V., 2001. Le Petrel de la Chaine Pterodroma (leucoptera) caledonica: Statut et menaces. Rapport de fin d’étude pour le compte de la Province Sud, Nouvelle Calédonie Bolton M., 2007. Playback experiments indicate absence of vocal recognition among temporally and geographically separated populations of Madeiran storm petrels Oceanodroma castro. Ibis 149: 255–263 Bostock H.C., Opdyke D.N., Gagan M.K., Kiss A.E. and Fifield L.K., 2006. Glacial/interglacial changes in the East Australian current. Climate Dynamics 26: 645659 Brooks T.M. & Helgen K.M., 2010. A standard for species. Nature 467: 540-541 Brown R.M., 2008. Molecular ecology of petrels (Pterodroma spp.)from the Indian Ocean and North East Atlantic, and implications for their conservation management. PhD Thesis. Queen Mary University of London and Zoological Society of London Brown R.M. et al., 2010. Range expansion and hybridization in Round Island petrels (Pterodroma spp.): evidence from microsatellite genotypes. Molecular Ecology 19: 3157-3170 Browne R.A., Anderson D.J., Houser J.N., Cruz F., Glasgow K.J., Hodges C.N. and Massey G., 1997. Genetic diversity and divergence of endangered Galapagos and Hawaiian petrel populations. Condor 99:812–815 Burg T. and Croxall J.P., 2001. Global relationships amongst black-browed and grey-headed albatrosses: analysis of population structure using mitochondrial DNA and microsatellites. Molecular Ecology 10: 2647-2660 Burg T. and Croxall J.P., 2004. Global population structure and taxonomy of the Wandering Albatross species complex. Molecular Ecology 13: 2345–2355 Burg T., Lomax J., Almond R., Brooke M. and Amos W., 2003. Unravelling dispersal patterns in an expanding population of a highly mobile seabird, the northern fulmar (Fulmarus glacialis). Proceedings of the Royal Society of London B 270: 979-984 Cagnon C., Lauga B., Hémery G. and Mouchès C., 2004. Phylogeographic differentiation of storm-petrels (Hydrobates pelagicus) based on cytochrome b mitochondrial DNA variation. Marine Biology 145: 1247-1254 Caldera R.D.M., Groom S., Miller P., Pilgrim D. and Nezlin N.P., 2002. Sea-surface signatures of the island mass effect phenomena around Madeira Island, Northeast Atlantic. Remote Sensing of Environment 80: 336-360 204 Cornuet J.M., Ravigné V. and Estoup A., 2010. Inference on population history and model checking using DNA sequence and microsatellite data with the software DIYABC (v1.0). BMC Bioinformatics 11: 401 Coyne J.A. and Orr H.A., 2004. Speciation. Sinauer associates, Sunderland, MA, USA Cromwell D., Challenor P.G. and New A.L., 1996. Persistent westward flow in the Azores Current as seen from altimetry and hydrography. Journal of Geophysical Research 101: 11,923–11,911,933 Csillery K., Blum M.G.B., Gaggiotti O.E. and François O., 2010. Approximate Bayesian Computation (ABC) in practice. Trends in Ecology and Evolution 25(7): 410-418 Daugherty C.H., Cree A., Hay J.M. and Thompson M.B., 1990. Neglected taxonomy and continuing extinctions of Tuatara (Sphenodon). Nature 347: 177-179 deQueiroz K., 1998.The general lineage concept of species, species criteria, and the process of speciation: A conceptual unification and terminological recommendations. Pages57–75 in Endless forms: Species and speciation (D.J. Howard, and S.H. Berlocher, eds.). Oxford University Press, NewYork deQueiroz K., 2007. Species concepts and species delimitations. Systematic Biology 56(6): 879-886 deSalle R. & Amato G., 2004. The expansion of conservation genetics. Nature Reviews Genetics 5: 702-712 Edwards S.V., 2009. Is a new and general theory of molecular systematics emerging? Evolution 63(1): 1-19 Edwards S.V. and Bensch S., 2009. Looking forwards or looking backwards in avian phylogeography? A comment on Zink and Barrowclough 2008. Molecular Ecology 18: 2930-2933 Fitzpatrick B.M., Fordyce J.A. and Gavrilets S., 2009. Pattern, process and geographic modes of speciation. Journal of Evolutionary Biology 22(11): 2342-2347 Friesen V.L., Burg T.M. and McCoy K.D., 2007b. Mechanisms of population differentiation in seabirds. Molecular Ecology 16:1765-1785 Friesen V.L., Gonzalez J.A. and Cruz-Delgado F., 2006. Population genetic structure and conservation of the Galapagos petrel (Pterodroma phaeopygia). Conservation Genetics 7: 107-115 Friesen V.L., Smith A.L., Gomez-Diaz E., Bolton M., Furness R.W., Gonzales-Solis J. and Monteiro L.R., 2007a. Sympatric speciation by allochrony in a seabird. Proceedings of the National Academy of Sciences of the USA 104(47): 18589-18594 205 Genovart M., Oro, D., Juste J. and Bertorelle G., 2007. What genetics tell us about the conservation of the critically endangered Balearic Shearwater? Biological Conservation 137: 283-293 Gomez-Diaz E., Gonzales-Solis J., Peinado M.A. and Page R.D.M., 2006. phylogeography of the Calonectris shearwaters using molecular and morphometric data. Molecular Phylogenetics and Evolution 41: 322-332 Grant W.S. and Bowen B.W., 1998. Shallow population histories in deep evolutionary lineages of marine fishes: insights from Sardines and Anchovies and lessons for conservation. Journal of Heredity 89(5): 415-426 Hakoun C., 2009. Utilisation d’AND mitochondrial pour clarifier la phylogénie d’un groupe de Pterodroma (Procellariidae). MSc. Thesis. Université Pierre et Marie Curie Paris 6 Heled J. and Drummond A.J., 2010. Bayesian inference of species trees from multilocus data. Molecular Biology and Evolution 27(3): 570-580 Hendry A.P., Lohmann L.G., Conti E. et al., 2010. Evolutionary biology in biodiversity science, conservation and policy: a call to action. Evolution 64(5): 1517-1528 Hernández-Guerra A., López-Laatzen F., Machín F., Armas D. and Pelegrí J.L., 2001. Water masses, circulation and transport in the eastern boundary current of the North Atlantic subtropical gyre. Scientia Marina 65: 177–186. Hey J. 2010. Isolation with migration model for more than two populations. Molecular Biology and Evolution 27(4): 905-920 Imber M.J. & Jenkins J.A.F., 1981. The New Caledonian Petrel. Notornis 28(3): 149-160 Jesus J., Menezes D., Gomes S., Oliveira P., Nogales M. and Brehm A., 2009. Phylogenetic relationships of gadfly petrels Pterodroma spp.from the Northeastern Atlantic Ocean: molecular evidence for specific status of Bugio and CapeVerde petrels and implications for conservation. Bird Conservation International 19: 1-16 Jia Y., 2000. Formation of an Azores Current due to Mediterranean overflow in a modelling study of the North Atlantic. Journal of Physical Oceanography 30: 2342–2358 Jennings W.B. and Edwards S.V., 2005. Speciational history of Australian grass finches (Poephila) inferred from thirty gene trees. Evolution 59: 2033–2047 Johnson J. & Stevens I., 2000. A fine resolution model of the eastern North Atlantic between the Azores, the Canary Islands and the Gibraltar Strait. Deep-Sea Research 47: 875– 899 206 Kawahata H., 2002. Shifts in oceanic and atmospheric boundaries in the Tasman Sea (Southwest Pacific) during the Late Pleistocene: evidence from organic carbon and lithogenic fluxes. Palaeogeography, Palaeoclimatology, Palaeoecology 184: 225-249 Kimball R.T. et al., 2009. A well-tested set of primers to amplify regions spread across the avian genome. Molecular Phylogenetics and Evolution 50: 654-660 Knowles L.L. & Carstens B.C., 2007. Estimating a geographically explicit model of population divergence. Evolution doi:10.1111/j.1558-5646.2007.00043.x Kuro’o M., Yonekawa H., Saito S, Eda M., Higushi H., Koike H. and Hasegawa H., 2010. Unexpectedly high genetic diversity of mtDNA control region through severe bottleneck in vulnerable albatross Phoebastria albatrus. Conservation Genetics 11: 127-137 Lambeck K., Esat T.M. and Potter E-K., 2002. Links between climate and sea levels forthe last three millions years. Nature 419: 199-206 Lawrence H.A., Scofield R.P., Crockett D.E., Millar C.D. and Lambert D.M., 2008a. Ancient genetic variation in one of the world’s rarest seabird. Heredity 101: 543-547 Lawrence H.A., Taylor G.A., Millar C.D. and Lambert D.M., 2008b. High mitochondrial and nuclear genetic diversity in one of the world’s most endangered seabirds, the Chatham Island Taiko (Pterodroma magentae). Conservation Genetics 9: 1293-1301 Lee J.Y. and Edwards S.V., 2008. Divergence across Australia’s Carpentarian barrier: statistical phylogeography of the Red-backed Fairy Wren (Malurus melanocephalus). Evolution 62: 3117–3134 Liebers D. Helbig A., 2002. Phylogeography and colonization history of lesser black-backed gulls (Larus fuscus) as revealed by mtDNA sequences. Journal of Evolutionary Biology 15: 1021–1033 Liebers D., Helbig A.J. and de Knijff P., 2001. Genetic differentiation and phylogeography of gulls in the Larus cachinnans-fuscus group (Aves: Charadriiformes). Molecular Ecology 10: 2447–2462 Lopes J.S. and Boessenkool S., 2010. The use of approximate Bayesian computation in conservation genetics and its application in a case study on yellow-eyed penguins. Conservation Genetics 11: 421–433 Mayr E., 1982. The growth of biological thought: Diversity, evolution, and inheritance. Harvard University Press, Cambridge, Massachusetts, USA. Metzker M.L., 2010. Sequencing technologies – the next generation. Nature Reviews Genetics 11: 31-46 207 Milot E., Weimerskirch H., Duchesne P. and Bernatchez L., 2007. Surviving with low genetic diversity: the case of albatrosses. Proceedings of the Royal Society of London B 274: 779-787 Moritz C., 2002. Strategies to protect biological diversity and the evolutionary processes that sustain it. Systematic Biology 51(2): 238-254 Moum T., Erikstad K.E. and Bjørklid E., 1991. Restriction fragment analysis of mitochondrial DNA in common murres. Canadian Journal of Zoology 69: 1577–1584 Nunn G.B. and Stanley S.E., 1998. Body size effects and rates of cytochrome b evolution in tube-nosed seabirds. Molecular Biology and Evolution 15:1360–1371 Olson S.L., 1975. Remarks on the generic characters of Bulweria. Ibis 117: 111-113 Ovenden J.R., Wustsaucy A., Bywater R., Brothers N. and White R.W.G., 1991. Geneticevidence for philopatry in a colonially nesting seabird, the Fairy Prion (Pachyptila turtur). Auk 108: 688–694 Pflaumann U. et al., 2003. Glacial North Atlantic: sea-surface conditions reconstructed by GLAMAP 2000. Paleoceanography 18: 1065 doi: 10.1029/2002PA000774 Phillimore A.B., Orme C.D.L., Thomas G.H., Blackburn T.M., Bennett P.M., Gaston K.J. and Owens I.P.F., 2008. Sympatric Speciation in Birds Is Rare: Insights from Range Data and Simulations. The American Naturalist 171(5): 646-657 Pinho C. & Hey J., 2010. Divergence with gene flow: models and data. Annual Review of Ecology, Evolution, and Systematics 41:215-230 Rando J.C. & Alcover J.A., 2008. Evidence for a second Palearctic seabird extinction during the last Millenium: the Lava shearwater Puffinus olsoni. Ibis 150: 188-192 Rayner M.J., Caraher C.J..F.Clout M.N. and Hauber M.E., 2010. Mitochondrial DNA analysis reveals genetic structure in two New Zealand Cook’s petrel (Pterodroma cooki) populations. Conservation Genetics 11(5): 2073-2077 Rind D. et al., 2001. Effects of glacial meltwater in the GISS coupled atmosphere-ocean model, 1, North Atlantic deep water response. Journal of Geophysical Research 106: 27,335–327,353 Rissler L.J., Hijmans R.J., Graham C.H., Moritz C. and Wake D.B., 2006. Phylogeographic Lineages and Species Comparisons in Conservation Analyses: A Case Study of California Herpetofauna. The American Naturalist 167(5): 655-666 Sangra P. et al., 2009. The Canary Eddy Corridor: A major pathway for long-lived eddies in the subtropical North Atlantic. Deep-sea research I 56: 2100-2114 208 Smith A.L. & Friesen V.L., 2007. Differentiation of sympatric populations of the bandrumped storm-petrel in the Galapagos islands: an examination of genetics, morphology and vocalisations. Molecular Ecology 16: 1593-1603 Smith A.L., Monteiro L., Hasegawa O. and Friesen V.L., 2007. Global phylogeography of the band-rumped storm-petrel (Oceanodroma castro; Procellariiformes: Hydrobatidae). Molecular Phylogenetics and Evolution 43: 755–773 Spear L.B. & Ainley D.G., 2007. Storm-petrels of the eastern Pacific Ocean: species diversity and assembly along marine habitat gradients. Ornithological Monographs 62: 1-77 Steadman D.W., 2006. Extinction and Biogeography of tropical Pacific birds. University of Chicago Press, London UK Steeves T.E., Anderson D.J. and Friesen V.L., 2005. A role for non physical barriers to gene flow in the diversification of a very vagile seabird, the masked booby (Sula dactylatra). Molecular Ecology 14: 3877-3887 Steeves T.E., Holdaway R.N., Hale M.L., McLay E., McAllan I.A.W., Christian M., Hauber M.E. and Bunce M., 2010. Merging ancient and modern DNA: extinct seabird taxon rediscovered in the North Tasman Sea. Biology Letters 6: 94-97 Techow N.M.S.M., O’Ryan C., Phillips R.A. et al., 2010. Speciation and phylogeography of giant petrels Macronectes. Molecular Phylogenetics and Evolution 54: 472-487 Techow N.M.S.M., Ryan P.G. and O’Ryan C., 2009. Phylogeography and taxonomy of White-chinned and Spectacled petrels. Molecular Phylogenetics and Evolution 52: 2533 Weir J.T. and Schluter D., 2008. Calibrating the avian molecular clock. Molecular Ecology 17: 2321-2328 Winker K., 2009. Reuniting Phenotype and Genotype in Biodiversity Research. BioScience 59(8): 657-665 Winker K. & Haig S.M., .2010. Avian subspecies. Ornithological Monographs 67: i–viii, 1– 200 Young L.C., 2009. Foraging ecology, population genetics and risk of fisheries bycatch for the Laysan albatross (Phoebastria immutabilis). PhD thesis, University of Hawaii Zhou M, Paduan J.D. and Niiler P.P., 2000. Surface currents in the Canary basin from drifter observations. Journal of geophysical Research 105(9): 21,893-21,911 Zino F., Brown R. and Biscoito M., 2008. The separation of Pterodroma madeira (Zino’s petrel) from Pterodroma feae (Feae’s petrel) (Aves: Procellariidae). Ibis 150: 326-334 209 Appendix 1: Paper not directly related to work presented in this manuscript published in Waterbirds (2009) 210 iuuuiuuuuuuuuu¥ ¦§¦¨©ª«¥¬©®§©¯¥°±²³µ́¶¶¥·¸¹º¥¥»±³¼¥©®¥¥½¾¿¹±ÀÁ¥ÂÃÀ́¥¦¨Á¥«ÄÄÅ¥¥¦Æ¦¦¥»Ç !"#$ %&'&(&)&*+,&-'& 4 3 3 2 1 ./01234504617789:;<262;;/=5>?3@:A/50BC65>D/3<2.5>63850DE20C/03.=/3540166/8 CFGHIFJK/HLJFM.NOPOQNRLFMJFCSNTU:C0=?>;=8V86:WV8XYENPPNFIMFG.ONM:7IZG[F ?O[NUHUJK1IGNHSOPOQNFJF;OP\GUMNF:.];]@8YV^:V^W88;Z_FFHF:3ZSNHN:;OP\GUMNF7IZG`ZNMF 8;ZI[ 0ZHLIFP=UQNOGZPJFCOIMF:ILFaZbOI6ZcdIOM[SNGN:.;68W:@Y8^65bZ[[NO:7IZG[F 9COIIFM_OGJNGQZLHSOIe/BcZNPfQZGQPOggh[Fd[][GIM]gI jklmnopmqr3SF;SOFGNs;FHIFPtvwxyz{yz|}u}~}NM[PZMMNgNFJZM/GJZGQFIFJZGJNHM_O_LPZHNOGNMJF[PNGNGQOG cOMHOgNHMdIFFJNGQNMPZGJM]<OFFI:HSFFsZ[H_O_LPZHNOGMNTFNMGOH[PFZIP\GOGZGJJZHZZIFMOcFHNcFMOPJ ZGJNGZ[[LIZHF:ML[SZMgOIHSFaZIRLFMZM5I[SN_FPZQO:7IFG[S;OP\GFMNZ]<ZHLHZKZ2MPZGJ:aZIRLFMZM:ZMNMNHFJ gOIHSIFFJZ\MNGaZI[S@YYWHOJFHFIcNGFSFHSFIvu}~}ZMMHNPPdIFFJNGQOGHSNMNMPZGJ]3SF_O_LPZHNOGMNTFZM FMHNcZHFJHOdFcOIFHSZG@YY_ZNIM:[Oc_ZIFJNHSHSFOGFNGJNNJLZPMFFG@Y\FZIMZQO]3SLM:<ZHLHZKZKMNc_OIB HZG[FgOIdNIJ[OGMFIZHNOGNGHSFIFQNOGSZMNG[IFZMFJ] xxx{uu}}yuu}xwx{uu}u inlqr[OGMFIZHNOG:7IFG[S;OP\GFMNZ:<ZHLHZKZ:aZIRLFMZM:;SOFGNs;FHIFP:vwxyz{yz|}u}~}:MFZdNIJ] ZHFIdNIJM8@t8f678B67^:@YYV 8 @ ;SOFGNs;FHIFPMtvwxyz{yz|}u}~}ZIFcFB ZPMOdFPNFFJHOdFJNGJPNGQNGHSFaZIRLFB JNLcBMNTFJQZJ\_FHIFPM:[POMFP\IFPZHFJHO MZM5I[SN_FPZQO]<OFFI:HSF[LIIFGH_O_LB <FIZPJtvuxy}~{}ZGJ3INGJZJFtvu}y| PZHNOGMNTFOgHSFM_F[NFMIFcZNGMLGGOG z}};FHIFPMt2cdFI8V^7e.IOOFZGJ t.IOO[email protected]@YYX: =OF8VVXe.IOOF@YY6:HSZHdIFFJFs[PLB SFG[FHSFNc_OIHZG[FOgcOIFZ[[LIZHFFMHNB MNFP\NGHSF[FGHIZP;Z[N[1[FZG]5MLIgZ[FB cZHFMNGOIJFIHOdFHHFIZMMFMM[OGMFIZHNOG GFMHNGQM_F[NFM:SOMFdIFFJNGQSZMdFFG MHZHLMZGJ[OGMFIFHSFM_F[NFM]NHSHSNM MHLJNFJOGP\d\4ZPPZQSFIt8VXYZGJ ZNc:;=ZGJ.4NMNHFJ<ZHLHZKZ2MPZGJNGHSF ?[SIFNdFIZGJ5MScOPFt8VWY:vu}~}GFMHM aZIRLFMZM5I[SN_FPZQONGaZI[S@YYWHOJFB OGPOZHOPPMZGJcOHLMZGJOGZgFtPZIQFP\ HFIcNGFHSFMHZHLMOgvu}~}_O_LPZHNOGZGJ _IFJZHOIBgIFFOP[ZGN[NMPZGJMt.IOOF HSF_OHFGHNZPINMMgOIHSFMFdNIJMOGHSFNMB @YY6];SOFGNs_FHIFPMFIFgOLGJHOPZ\ PZGJ] FQQMNGZPcOMHZPPcOGHSMNG8V7^B7V:8VX8BX6 ZGJNG8VXWOGHSFNIPZIQFMHdIFFJNGQMNHFOG a/3<1D? NINHNcZHNtCSINMHcZM2MPZGJt4ZPPZQSFI 8VXYe?[SIFNdFIZGJ5MScOPF8VWY]<OFB ?HLJ\5IFZ FI:HSFcZNG_FZOgFQQBPZ\NGQZMgOLGJHO <ZHLHZKZtW]V@¡?86Y]7W¡e7NQ]8NMPO[ZHFJNGHSF dF0OFcdFIBAZGLZI\:NHSZMF[OGJZI\_FZ GcO OIHSFIGQIOL_OgHSFaZIRLFMZMZI[SN_FPZQO]3SNMIFB GOIHSFIGcOMHOgHSF887NMB NG5_INPBALGF]=NGQNGQIF[OFINFMMSOFJ PZGJHFMOOgP[7ZIGFGN[[SNMP;ZGOPJ\GNMFMHNSZF]<Z GOIHSB HSZHHSF_O_LPZHNOG[OGMNMHFJOgHO[PZMMFM FMHOg0LL<NZ:aZIRLFMZMH2LMPHZZGKZJMPN:FNMG7H6SYFZc INJFRLZB OgdNIJMfMLccFIdIFFJFIMZGJNGHFIdIFFJB H@OINc ZPTOGFOgHSF;Z[N[ZGJNMWcPOGQZGJPFMMHSZG JFgOIZHOHZPZIFZOgX]Xc@t7]AZ[R:_FIM] FIM:ZGJHSZHHSFIFZMGONGHFI[SZGQFdFB [Occ]N ] FNMPZGJINMFML_HO6@7cZMPZGJ_IFMFGHMZ HFFGHSFHOQIOL_M]3SFM_F[NFMNM[PZMMNB HZdLPZII3S FPNFg]3SFZdMFG[FOg_FIcZGFGHMHIFZcM FJZM/GJZGQFIFJt2>C0@YYW:NHSHSF cF ZGMHSZHHSFNMPZGJSZMNGMLg[NFGHZHFIHOML__OIH :SN[S:[OcdNGFJNHSHSFJNg[LPH\OgZ[[FMM: _O_LPZHNOGNGHSF;SOFGNs2MPZGJMMSONGQZ LFsG_QPLZPNZGHMFMS IFNMGO_FIcZGFGHSLcZG_IFMFG[F MSZI_JF[PNGFMNG[FHSF8VXYMt;NFI[Fxwu}~ tDF[FI8VW\8HS]F3S LM:<ZHLHZKZNM[LIIFGHP\LGNGSZdNHB @YYX]3SF_O_LPZHNOGMOG3OGQZZGJOG FHOJI:N[ZPHHNScF OLQSNHZMO[[L_NFJd\;OP\GFMNZGMNG_IFSNMB ZIFMLPHOg_ZMHSLcZGZ[HNNHNFM:;Z[N[ HSF3LZcOHL2MPZGJMSZFZ__ZIFGHP\dFFG =ZHMt }wwM]¢u5M IFNGHIOJL[FJZGJZIFGO[OcB FsHFIcNGZHFJd\_IFJZHOIM:ZMFPPZMOGCZGB cOGtDF[FxI£8V~}W8¢]F 3SF\ZIFHSFOGP\cZccZPM_F[NFM HOGZGJaZPJFG2MPZGJMt.IOOF@YY6e _NGIHFIMOFJGLH[OFGJHSOFGNM<Z PZGJ]5[ZHt¤x~¢u}w¢ZMML__OMFJP\ HLHZKZNGHSFPZHF8VWYMt_FIMOGZP .NIJ6NgF2GHFIGZHNOGZP@YYX]3SFM_F[NFMNM 678 ;;;;;;; ¡ ¢£¤¥¦§¨©©ª«¬®¤¦¯ °±²¬¤³´µ¶§³´··¸¹®º 676 89 '%&&W$$'.R&5R, '!%3'%'3' &3$''!'.RED;ABX-S88Y" (88Z3'& /&R&5'0$&' 3#$'$/. &0$ &'''3/'/& $.9''%0,/&3 $''R'' &-P/.(". 9'0&//$0''$/ &9/['-S88("3'9 &8.(#'''/&-3& #''$&#'"-&&, '\&'$/&'3' '"&($/'/&-#, $&3'\3&]$/" $/&&/$&3''.8/ '/&$&0& &/$/$/&.'3'$ '''''$''5/. &#' $&#&'!/&3W3, ''3*8Z&'''%. &3'-..R1 "$#%!'&/&% /$'$'!33 #$&3#/&0/ W&3#'%&//,, /#. 4 3 3 2 1 ijklmnopqorstltsusovwxsyzo{jt|o}|~nyjo}ntmnxo omnnzjykosmnsqomsywntosyzo~jyto~lytox~s tj~ywosmnow|~{yqomsywntoxjyno{jzt|ot~owsxnsm| q !"# $#% &#'&()*7()+*!, - &#'()+)". &!/0'&/&1 #% 223&&#-()*4"0 '%#%3'&5, 3&%&67-()*80()*49 &#'()+)".:/#%&&# <=>?@A;BCDEA;&3';F@G=H@A;I>AH?@G '/. &#'%&'& &#/'&#'(*3, #-3'0,3'03#0##0/, #"'# 3;JKABB@L=BCMNA;>CNEGLEHGO3- &#' ()+)".P&0&'$', !#%&Q!P&R'% ()*(. 2$'!#%&'/()+* - &#'()+)". 6& 8! 2#%##%, '&33/&R&5R' 33'3'&. &' &%$&%0(6(*6&S88*.8& '/0$&&, '. 8'#/R&5R'/ -!'&T$,$&3/UV % %, '())8"3'#/.9S.7 '/(88$/& #/W-P/.("!/335, ^_ 9&3'`!'!%-#, $(7V88(+V48&"0!(Y ]%/#$#!''% %.R&5R'$#! ]/&!&'0$ 3!'%$& &3-./.0$'$ (*8'". &%#/3$ &/0&#!488 &$/&&/&. R&5R'$&//&3 9&##/#!& '%%0#+V88&'\ #/&''. 8!S8$&&5&0 //0()'$&'. $'&#aEbD=LcB=A '&/&$FAGG@d=>A eED@?A&&/<=>?@A BCDEA.8/''%&5 (7&()'01/5' '. [/''3&'/'!#' /&&/&'!'%0 /15'3&$#! -369f:90PgS.40Rh8.(". $!05, BBBB ¡¢£¤¥¦§¢̈©ª«¬©ª®¯¤°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`abcdecfgahigjg_^kljcmbln`abmbopncgqcrsctigbo^uchbpabjncvwxyz{|z{}~~~cvcmljbnc cqbmljbncqagmclp`pll njlocfla`bnlnflakicde AAAAAAAA]l]t] ¡¢¡£¤¥¦§¨©ªª«¬®¯¥§°¡±²³¥´µ¶·¨´µ¸¸¹º¯» 678 9 ^_`ab]cd]eb_f]gb_hijbgbfkh]mno]_fp]j_fqb]rs]uvwxyzxy{|t|}~|]_piakh]]sbg_abh]_fp]]g_abh]_fp] h]]] frk]hbbp]rf]_kik__]ha_fp]e_jibh_h]j ba_qre_j]d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c-c---- hihjklmnokpqikprnstuvwxyynz{|}nn~tvnkpjnn|tn wnhjnmnnhhhn~ 6789 699 9 768 9 4!&)&!;= 6"6#&%>B !""#$%&! #'(!!) 6!6%#Y!#!'#86 &!&&A&6Ɋ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`` 8;";?2-01@0#&#$&#;6!6%#B M888 65!$7&""*N]6%MNN5 &#!=66'D.EFGH.I.//0-GJ1KIKLFE04 (!!)759X5 *++5;9#$9!6!$%!;$&B :6#&96'(6##&!#&6MNN347&6B a%6?5 /.HFbHFQ0-5!;6"&#^";&B ;! $;;#"!;6&"!# &]!"#&6"!7;!6#&#!&##S!"&#%-3"^c!M#*RB!M6R*55(!6!$"&6 ;6&##"!#<=&!#5!! )#!86B (!!)7X55MNN569&!B&"!; $7;6;!#>66! 6&!#9B U!658>!\#:%67 8!)5 5765]5]!#7]55_")%U&6#5 #$!#;6&#;!#6%!#!;B (!*!+)+375X5 ;!7a!8B ";!8#$&#&&##"&57&&A& #$&#&!#&6#$&!##&!9B6&!$"!&68 &6!;!=!!;8!6A! 6&!#! 6(!!"&MC^*3B*+5 #&66&"<"B &#W5!85*++35(;&:! &6&# ;#&#$ O011KPF1FQ@[email protected] (!!)67"X56&5 :#"!"<"&!6$;&# ;!7:%!;! 69&)#! =! ;!;7&666?/.HFbHFQ0-d.H01bKP05 =&#&&A&A#9&&# 9&)*=! RC^MNBRM5 *+N566X&#&#&#"&#X&B !:#=&==&6&##! "!#7 V"e)7&(5#W5 _&66%7#";6!6%#&5\# 96;6;5V5 "&66%!=;#&9%9&!6&#'!#6% ;7V #!W!$&;%7\#:%!&6B R)=&8&%4!#8;";"&7$&#;& !#&5()&6= %5 W5*+R5\#e %B6	!&#!B #7&6!#$8;=&#%&6##:&:6&# V")"7!(5 ## !;8#X&e &6&#'6X&B "5 45(!6!$"&6!#:&!#c^33B35 #:8!;S&#:$&!#"!:! 6e75 *+M5 #$""!;B&6!"B T !665 )-C+^3*B3+5 ;6&&"&!#!#7&&A&8! 6 W&669&$;!#7X5 *+3N5(#!!=;=&6 9&969 "!6% !;=!# 6&"<"8"V5 & # *NM^C+BNM5 &#!6&!#!;69!8:7&6B 7!6%!&)7V55-9]5B5 ;9&65*+C5!#9 !#f !&>6!6%#Y!#&65XY=! ;! $;&=!:&68! 6!9&96%9#< 6XAYY= !#&67! &6676&'Y547 6"&866&O011KPF1FQ@07; g!!6!$7 & * M =&%#!9&!%!;#&!8B =97X5]5*+*5^*B6M!N+65=!6&#;&8&5 8g&6&#]! #&6!X&#&#;8&B #$!;?2-01@0! 6&!##"&&#& & ";^R+3BNR5 ;$;$#"%!";&"!#!#!;B =97X5 5*+C58$#7;%6!$#%&#&>!#!=%! 6&###";6!6%#&5 !#;678 ;$&]a% 6?/.HFbHFQ059*M^*+BMM+5 MNN5-\ !;&#"2\ 5 "!==#"6!=!#!#$!&	 \ `8885"#65!$7&""MMX&%MNNC5 ! 6&!#!#;6&#!"=&" X ;;%^7`5 5&#]5X56##!%5*+M5&$6! !&!#;#&#$&:& #&5 ;$# ?/.HFbHFQ05="&#X = !:& 4 3 3 2 1 T 8U9VWX9 ;:!7&&A&8&=&!969%;V"B !#6A9#:!##=#6!6%#Y&#Z& 'V9 4;&6:& ;!&!#!6&#!#;B 6&#&##:;!"YYA8#;!6!$ *CN^*BR5 !;="&#(&##$! #"65MNN*5; !; ="&#(&#A%W 5 !;="&# (&##$! #"676!#%&!#7&6!#&7 \555 6"75]57597_5T795& 675&&&7X5 ;!#&#W5U&$$5MNN356;!#>6&#"!#B :&!#:%&#&=#!!&!#&B ]^]_`abcd`ef^`egchijklmnncopqrccsiktc̀egccuvwqixycz{lxc]_ycb||}cc]~]]cs 456 789 !"#! CD$%&E0*0.&!! !!!$!%&'&"!"!# 9&%&'&8*+5555857)58:: ()8*&+!,!#"-8*&+! &.&! 9!!"!-8<!A8956680#* /-70!1-234! FGHIJKIJLMN"!"!"O3)30"P2) &0-7!2%8"0$5678 !":7595)596 !1""!!0"$""!-%&'&.&! 90*-<)5666/Q#*!"R" S*Q "559:;:):64 C%!>"-%&'T**E$'&!"!#) -!<,!*"""! !*"*UUV$"V&WXYZ[HM\N#*D!?" ""*=""%!>"!?"% % Q.!0156:85):94 .&&"!%"2@6*0%&'&$$") 73-%7567990!*0"* "!-2*$>-23! 8-203"!S3!"!"8) -/!A8!5666B01&) "&0*!5;45):56 !&"#&1&"11!!" 3 2 2 1 0 A THOUSAND miles from land are we, Tossing about on the roaring sea; From billow to bounding billow cast, Like fleecy snow on the stormy blast: The sails are scatter’d abroad, like weeds, The strong masts shake like quivering reeds, The mighty cables, and iron chains, The hull, which all earthly strength disdains, They strain and they crack, and hearts like stone Their natural hard, proud strength disown. Up and down! Up and down! From the base of the wave to the billow’s crown, And midst the flashing and feathery foam The Petrel finds a home,— A home, if such a place may be, For her who lives on the wide, wide sea, On the craggy ice, in the frozen air, And only seeketh her rocky lair To warm her young, and to teach them spring At once o’er the waves on their stormy wing. O’er the Deep! O’er the Deep! Where the whale, and the shark, and the sword-fish sleep, Outflying the blast and the driving rain, The Petrel telleth her tale—in vain; For the mariner curseth the warning bird Who bringeth him news of the storms unheard! Ah! thus does the prophet, of good or ill, Meet hate from the creatures he serveth still: Yet he ne’er falters:—So, Petrel! spring Once more o’er the waves on thy stormy wing! adapted from Barry Cornwall