Phylogeography and molecular phylogenetics have proven remarkably useful for understanding the patterns and processes influencing historical diversification of biotic lineages at and below the species level, as well as delimiting morphologically cryptic species. In this dissertation, I used an integrative approach coupling comparative phylogeography and coalescent-based species delimitation to improve our understanding of the biogeography and species limits of Central American freshwater fishes. In Chapter 1, I conducted a literature review of the contributions of phylogeography to understanding the origins and maintenance of lower Central American biodiversity, in light of the geological and ecological setting. I highlighted emerging phylogeographic patterns, along with the need for improving regional historical biogeographical inference and conservation efforts through statistical and comparative phylogeographic studies. In Chapter 2, I compared mitochondrial phylogeographic patterns among three species of livebearing fishes (Poeciliidae) codistributed in the lower Nicaraguan depression and proximate uplands. I found evidence for mixed spatial and temporal divergences, indicating phylogeographic “pseudocongruence” suggesting that multiple evolutionary responses to historical processes have shaped population structuring of regional freshwater biota, possibly linked to recent community assembly and/or the effects of ecological differences among species on their responses to late Cenozoic environmental events. In Chapter 3, I used coalescent-based species tree and species delimitation analyses of a multilocus dataset to delimit species and infer their evolutionary relationships in the Poecilia sphenops species complex (Poeciliidae), a widespread but morphologically conserved group of fishes. Results indicated that diversity is underestimated and overestimated in different clades by c. ±15% (including candidate species); that lineages diversified since the Miocene; and that some evidence exists for a more probable role of hybridization, rather than incomplete lineage sorting, in shaping observed gene tree discordances. Last, in Chapter 4, I used a comparative phylogeographical analysis of eight codistributed species/genera of freshwater fishes to test for shared evolutionary responses predicted by four drainage-based hypotheses of Neotropical fish diversification. Integrating phylogeographic analyses with paleodistribution modeling revealed incongruent genetic structuring among lineages despite overlapping ancestral Pleistocene distributions, suggesting multiple routes to community assembly. Hypotheses tests using the latest approximate Bayesian computation model averaging methods also supported one pulse of diversification in two lineages diverged in the San Carlos River, but multiple divergences of three lineages across the Sixaola River basin, Costa Rica, correlated to Neogene sea level events and continental shelf width. Results supported complex biogeographical patterns illustrating how species responses to historical drainage-controlling processes have influenced Neotropical fish diversification.

Les chênes sessiles et pédonculés sont deux espèces sympatriques interfertiles occupant des places distinctes dans la succession écologique. Elles constituent pour cela un bon modèle pour l’étude de la spéciation écologique. Malgré leur écologie contrastée, les deux espèces de chênes coexistent naturellement dans de nombreuses forêts, rendant possible l’étude de leur système de reproduction intra- et interspécifique. Des travaux précédents ont suggéré que l’hybridation entre ces deux espèces serait fréquence-dépendante. Elle dépendrait de la proportion de pollen allospécifique (c'est-à-dire de pollen de l’autre espèce) que reçoit l’arbre mère étudié. Ce phénomène d’hybridation fréquence-dépendante est appelé effet Hubbs, du nom d’un ichtyologue qui découvrit ce mécanisme en 1955. Mon travail a consisté à étudier, dans une parcelle mixte de chênes, les barrières à l’hybridation qui permettent la coexistence de ces deux espèces. Pour cela, j’ai effectué une étude de paternité de grande ampleur (près de 3500 individus typés sur 12 marqueurs microsatellites). Tout d’abord, je me suis intéressée à la délimitation des deux espèces en appliquant pour la première fois in situ le critère d’interfertilité. Je me suis ensuite concentrée sur les facteurs qui influencent l’hybridation au travers d’une modélisation des croisements à l’échelle de l’individu. Les résultats montrent que le maintien de ces deux espèces est régi par deux composantes environnementales importantes : la fréquence de chaque espèce et leur distribution, qui influencent la quantité de pollen reçue. Grâce à cette étude empirique et à l’approche de modélisation de ces croisements, nous avons désormais une bien meilleure vision de l’effet de l’environnement sur l’hybridation. J’ai par la suite comparé les caractéristiques du système de reproduction de chaque espèce (dispersion du pollen et fécondité mâle) en cherchant si un lien existait avec leur stratégie écologique. Les résultats suggèrent que les différences de dynamique écologique pourraient être à l’origine de la spéciation du fait de l’existence de compromis différents en termes d’allocation de ressources mais qu’à plus court terme la stabilité de l’environnement est essentielle au maintien des espèces. / Quercus petraea and Q. robur are two interfertile sympatric species. They occupy distinct stages during forest succession and constitute therefore good models for ecological speciation studies. Despite their differences, they often grow together in mixed stands, allowing the study of their intra- and interspecific reproductive system. Hybridization between these two oak species has been suggested to be frequency-dependent. The effect of the relative species abundance on hybridization is known as the “Hubbs’ effect”, from the name of an ichthyologist who described this mechanism in 1955. My work was to study the processes that limit hybridization between these two species, thereby allowing their coexistence. I conducted an extensive paternity analysis (almost 3500 individuals genotyped at 12 microsatellite markers). First, I tried to delimitate the two oak species by applying for the first time in situ the interfertility criterion. Then, I focused on interspecific crosses by studying those factors influencing hybridization. Results showed that the maintenance of these two species depends on the frequency of each species and their distribution, as both factors influence the quantity of pollen received by female flowers. Thanks to this empirical study and to this modeling approach, we have now a much better view of environmental effects on hybridization. I then compared the characteristics of the reproductive system of each species (pollen dispersal and male fertility) in relation with their ecological strategies. The results suggest that differences in species ecological dynamics are at the origin of the speciation process but that at a finer scale the stability of the environment is crucial for species’ maintenance.

Section Restricti is one of the last sections of the genus Aspergillus that hasn't been revised in the modern taxonomic era. All species are osmophilic, have simple morphology and they are mainly important because of the food, feed and seed spoilage, some of them also have negative influence to the indoor air. The section consists of seven species according to the last taxonomic revision from 2008 based only on molecular data. It currently consists of six asexual and one homothalic species. 126 isolates from the section coming from four continents were studied in this thesis, including ex-type strains. Revision of the section was carried out combining molecular phylogenetic analysis and conventional taxonomic methods. Modern methods of species delimitation based on multispecies coalescent model were used for the phylogenetic reconstruction. From the conventional methods analysis of morphology (macro- and micromorphology including scanning electron microscopy) and physiology (ability of growing in osmotic gradient and several different temperatures) was performed. Apart from the seven known species, eight new undescribed species were discovered. Majority of the new species belongs to the Aspergillus penicillioides species complex. Key words: Aspergillus restrictus, osmophilic fungi, species...

This study is part of the Swedish Taxonomy Initiative (STI) - one of the most ambitious all taxa biodiversity inventories in the world. One of the pillars in STI is to support taxonomic research on the most neglected taxonomic groups with the aim to lift the level of knowledge of biodiversity in the country. There is still a lot to be discovered, especially in the microscopic world, and this includes mites. Many aspects of mite biology and diversity are poorly known, such as species richness, abundance, distribution, lifestyle and behavior of species. Mites inhabits all sorts of aquatic, terrestrial, arboreal and parasitic habitats, nevertheless even in well-studied systems mites are often overlooked. Despite being among the smallest of arthropods, they are of medical and economical importance and may be very abundant in the ecosystems they inhabit. This thesis focuses on Parasitengona (Acariformes: Prostigmata), one of the most diverse taxa among the arachnids. It includes the aquatic Hydrachnidia (water mites) and the terrestrial Trombidia (e.g. velvet mites, chiggers). A unifying characteristic of Parasitengona is their complex life cycle of active and inactive stages, parasitic larvae and predatory deutonymphs and adults. They typically parasitize and prey on arthropods, except the chiggers which have vertebrates as hosts. The aim of this thesis is to shed light on the phylogeny and taxonomy of Parasitengona with emphasis on the Swedish fauna. To achieve this, mites were collected from different localities throughout the country between the years 2007-2016. Water mites were sampled with a hand net. Larvae of terrestrial Parasitengona were collected with sweeping nets and sorted out from malaise trap samples from the Swedish Malaise Trap Project. To collect the adults Berlese-Tullgren extractor and pitfall traps were used as well as hand collecting and sifting with litter reducer. The material collected abroad was kindly provided through collaboration.  Methods used in the papers included morphometrics, multivariate analyses, experimental rearing, DNA extraction and sequencing, Bayesian phylogenetic analyses and molecular species delimitation. In paper I and II, we combine molecular species delimitation models and morphological data to resolve taxonomical issues. This integrative taxonomic approach of combining data resulted in Piona dispersa Sokolow, 1926 as a valid species and redescriptions, new synonyms and neotypes provided for Erythraeus phalangoides (De Geer, 1778),  E. cinereus (Dugès, 1834) and E. regalis (C.L. Koch, 1837). Based on the new inventories we produce an updated and annotated checklist of 105 terrestrial Parasitengona species for Fennoscandia in paper III, and use metadata to increase the knowledge on distribution, habitat preferences, life stages and abundance. Out of these, 20 species are new findings for the region and five are potential new species for science. In paper IV, we provide a molecular phylogeny of Parasitengona based on the genes 18S, 28S and COI for 80 taxa with a sampling focus on the terrestrial lineages. Based on the results we offer a revised higher-level classification of the group. In particular the analyses supported Tanaupodoidea Thor, 1935 as a separate superfamily, but Trombiculoidea Ewing, 1929 was not monophyletic and was synonymized, along with Chyzerioidea Womersley, 1954, with Trombidioidea Leach, 1815. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 3: Manuscript. Paper 4: Manuscript.</p><p> </p>

A Amazônia inclui uma grande proporção da biodiversidade encontrada atualmente na Terra. Apesar disso, nosso conhecimento sobre a biodiversidade Amazônica ainda é limitado, dificultando nosso entendimento dos padrões de diversidade nesta região. Entender os processos que levaram à diversidade encontrada na Amazônia representa um grande desafio para a biologia evolutiva. Este estudo foca em Pachyptera (Bignonieae, Bignoniaceae), um pequeno gênero de lianas neotropicais, centrado na Amazônia. Pachyptera tem uma história taxonômica complicada, incluindo problemas na circunscrição genérica e específica. Este estudo visa: (i) reconstruir o parentesco filogenético entre espécies do gênero, (ii) produzir uma revisão taxonômica, incluindo nova circunscrição genérica e específica, (iii) entender a história biogeográfica do grupo e, (iv) desenvolver marcadores microssatélites (SSRs) para futuros estudos filogeográficos. Em primeiro lugar, reconstruímos a filogenia do gênero usando uma ampla amostragem de taxa e uma combinação de marcadores de cpDNA (ndhF and rpl32-trnL) e nDNA (PepC). Em segundo lugar, analisamos a filogenia de Pachyptera utilizando análises de coalescência (GMYC e *BEAST) e morfologia para esclarecer limites específicos dentro do complexo P. kerere. Em terceiro lugar, produzimos uma filogenia datada de Pachyptera, a qual foi utilizada como base para reconstruir a história biogeográfica do gênero utilizando BSSVS e RASP. Por fim, desenvolvemos SSRs utilizando sequenciamento de próxima geração, os quais serão utilizados para guiar estudos filogeográficos futuros com o grupo. Nosso estudo indica que P. ventricosa é mais proximamente relacionada à Mansoa do que Pachyptera, levando ao reestabelecimento de M. ventricosa. Além disso, nossos estudos moleculares e morfológicos sustentam o reconhecimento de P. kerere var. incarnata como uma espécie separada e a descrição de uma espécie nova (P. linearis). Desta forma, reconhecemos um gênero com cinco espécies: (i) P. aromatica, (ii) P. erythraea, (iii) P. incarnata, (iv) P. kerere, e (v) P. linearis. Estas espécies são tratadas em uma revisão taxonômica do gênero. As análises biogeográficas indicam que Pachyptera surgiu durante o Eoceno Tardio, e diversificou durante o Mioceno, um período de intensas perturbações provocadas na América do Sul (i.e., soerguimento dos Andes, eventos de incursões marinhas, e formação de sistemas florestais secos e úmidos). Vinte-e-um SSRs foram desenvolvidos para Pachptera e servirão como base para estudos filogeográficos futuros com este grupo. Esta dissertação faz parte de um projeto multi-disciplinar que visa compreender a evolução da biota amazônica e seu ambiente (FAPESP 2012/50260-6) / The Amazon houses a large proportion of the overall biodiversity currently available on Earth. Despite that, our knowledge of Amazonian biodiversity is still limited, complicating our understanding of diversity patterns within this region. Understanding the drivers of Amazonian biodiversity represents a major challenge in evolutionary biology. This study focuses on Pachyptera (Bignonieae, Bignoniaceae), a small genus of neotropical lianas centered in the Amazon. Pachyptera has a complicated taxonomic history, including problematic generic and species circumscriptions. This study aims to: (i) reconstruct phylogenetic relationships among species of the genus (ii) produce a taxonomic revision, including clear generic and species circumscriptions, (iii) understand the biogeographic history of the group, and (iv) develop microsatellite markers (SSRs) for future phylogeographic and population genetic studies. First, we inferred phylogenetic relationships within a broad sampling of taxa and a combination of cpDNA (ndhF and rpl32-trnL) and nuclear (PepC) markers. Second, we analyzed the phylogeny of Pachyptera using coalescent approaches (GMYC and *BEAST) and morphology to clarify species limits within the P. kerere species complex. Third, we produced a time-calibrated phylogeny of Pachyptera that was used as basis to reconstruct the biogeographical history of the genus using BSSVS and RASP. Lastly, we developed SSRs using next-generation sequencing (NGS) that will be used to guide future phylogeographic studies within this group. Our study indicates that P. ventricosa is more closely related to Mansoa than Pachyptera, leading to the reestablishment of Mansoa vetricosa. Furthermore, our molecular and morphological analyses support the recognition of P. kerere var. incarnata as a separate species, and the description of a new taxon (P. linearis). As such, we here recognize a genus with five species: (i) P. aromatica, (ii) P. erythraea, (iii) P. incarnata, (iv) P. kerere, and (v) P. linearis. These species are treated in a taxonomic revision of Pachyptera. The biogeographical analyses indicate that Pachyptera originated during the Late Eocene, and diversified during the Miocene, a time of intense perturbations in South America (e.g., uplift of the Andes, marine incursions, and formation of dry and wetland systems). Twenty-one SSRs were developed for P. kerere and will serve as basis for future phylogeographic studies. This dissertation is part of a multidisciplinary project that aims to understand the evolution of the Amazonian biota and its environment (FAPESP 2012/50260-6)

Genomes provide windows into the evolutionary histories of species. The recent accessibility of genome-scale data in non-model organisms and the proliferation of powerful statistical models are now providing unprecedented opportunities to uncover evolutionary relationships and to test hypotheses about the processes that generate and maintain biodiversity. This dissertation work reveals shallow-scale species boundaries and population genetic structure in two imperiled groups of salamanders and demonstrates that the number and information content of genomic regions used in species delimitation exert strong effects on the resulting inferences. Genome scans are employed to test hypotheses about the mechanisms of genetic sex determination in cryptobranchid salamanders, suggesting a conserved system of female heterogamety in this group. At much deeper scales, phylogenetic analyses of hundreds of protein-coding genes across all major amphibian lineages are employed to reveal the backbone topology and evolutionary timescales of the amphibian tree of life, suggesting a new set of hypotheses for relationships among extant amphibians. Yet, genomic data on their own are no panacea for the thorniest questions in evolutionary biology, and this work also demonstrates the power of a model testing framework to dissect support for different phylogenetic and population genetic hypotheses across different regions of the genome.

Emplectonema gracile (Johnston 1837) is a hoplonemertean of marine intertidal hard-bottom communities and is distributed throughout the Northern Hemisphere. Although possessing a planktonic larval stage in its life history, the range of such cosmopolitan marine invertebrate species is often explained by cryptic speciation and anthropogenic transport. The purpose of this study is to test for possible cryptic species using mtDNA markers (COI and 16S rDNA) and to investigate population structure in E. gracile over a portion of its geographic range using mtDNA markers and ddRADseq nuclear SNP data. The results of both phylogenetic- and tree-based species delimitation revealed that E. gracileis a morphotype containing cryptic species. Three North Atlantic and one Pacific coast population are inferred as one species (E. gracile sensu stricto) and two Pacific coast populations (Akkeshi, Japan and Charleston, Oregon) are inferred as another species (Emplectonemasp 1), strongly confirming an earlier study and extending the range of the latter species to the Pacific coast of Japan. Anthropogenic transport is suggested as the likely mode of transport for E. gracile.Both Fst, PCA and haplotype network analyses suggest a lack of differentiation between E. gracile populations separated by large geographic distances.In contrast corresponding analyses forEmplectonemasp. 1 indicate differentiation between the two populations sampled. Further research will be necessary to reveal if rare anthropogenic transport or natural dispersal (larval transport, rafting) between geographically adjacent yet to be delimitedE. gracile morphotype populations is responsible for its seemingly disjunct distribution.

La grande diversité d’insectes et la quantité de spécimens recueillis lors de l’échantillonnage constituent les plus grands défis de la systématique des insectes. Le tri des échantillons au niveau des espèces est nécessaire avant qu’ils puissent être utilisés pour des enquêtes sur les modèles de biodiversité. En raison de l’obstacle taxonomique, le manque d’expertise taxonomique, de nombreuses études sur la diversité des insectes classe les spécimens en Unités Morphologiques Opérationnelles (MorphOTUs), aussi appelées morpho-espèces, en désignant les groupes définis subjectivement en fonction de caractéristiques morphologiques évidentes. Cependant, il est long et douteux de définir avec précision les limites des espèces en se fondant sur les MorphOTUs, surtout dans les groupes où il y a de minuscules insectes et une grande similarité au niveau des espèces, comme chez les Hyménoptères. Le codage à barres de l’ADN, une approche taxonomique discriminatoire qui utilise des séquences d’ADN, a accéléré la classification taxonomique et peut être une approche alternative aux MorphOTUs. Cependant, il est crucial d’utiliser une stratégie fiable et économique de codage à barres ADN pour traiter un grand nombre d’échantillons. En outre, le codage à barres d’ADN devrait fonctionner avec les espèces problématiques dans l’entomologie moléculaire comme on l'observe parfois avec les hyménoptères. Afin de mettre en œuvre une évaluation rapide de la biodiversité des Hyménoptères, optimiser les de étapes de barcodage d`ADN (extraction d’ADN, amplification par PCR et séquençage) était le premier objectif de ce projet de recherche. On a testé et optimisé une extraction d’ADN arrivant à une méthode coûtante 0,20 dollars par spécimen. On a validé la performance adéquate des mini-codes à barres d’ADN, réduits en taille à 313bp, pour établir une classification d’Unités Taxonomiques Opérationnelles Moléculaires (MOTUs) comparable à celle du codage à barres d’ADN couramment utilisé, de longueur de 658bp. On a adopté ce protocole optimisé pour le codage à barres de 517 spécimens d’Hyménoptères échantillonnés par des pièges aspirateurs situés dans la forêt laurentienne de l’Est du Canada. Avec le séquençage multiplexé à haut débit Illumina, impliquant des amplicons étiquetés, on a obtenu des mini-codes à barres pour 88% des spécimens. Le coût et le temps nécessaires pour générer des données MOTU, grâce à notre approche de codage à barres d’ADN, étaient environ la moitié de celui de la classification morphologique en MorphOTUs. Le deuxième objectif de ma recherche était de comparer l’efficacité du tri morphologique des MorphOTUs avec l’identification moléculaire et la délimitation par MOTUs. On a démontré iii une forte congruence entre l’identification morphologique et moléculaire au niveau taxonomique de la famille dansla base de données Barcode of Life (BOLD) et GenBank (93 %), alors que seulement 18 % des mini-codes à barres ont été attribués à des identifications plus précises (genre ou espèce). La délimitation moléculaire s’est faite avec quatre méthodes de regroupement différentes (basée sur la distance : Découverte automatique de l’écart de codes à barres (ABGD) et Assemblage des espèces par partitionnement automatique (ASAP) ; basée sur un dendrogramme : Coalescente mixte généralisée du yule (GMYC) et Processus bayésien de l’arbre de poisson (bPTP)). En générale, les méthodes moléculaires ont plus que doublé la diversité estimée des MorphOTUs des Hyménoptères. Les MOTUs étaient en grande partie incompatibles avec les MorphOTUs (ratio d’appariement <0,35). Les méthodes basées sur la distance ont donné des résultats plus conformes au tri morphologique que les méthodes basées sur les arbres, en particulier dans la superfamille des Chalcidoidea. Compte tenu de la comparaison entre le coût et le temps des méthodes de classification moléculaire et morphologique, nos résultats suggèrent que le codage à barres mini-ADN pour estimer la diversité des espèces d’Hyménoptères est plus économique que le tri par MorphOTU. Cependant, bien que les méthodes MorphOTU et MOTU aient donné de nombres unités taxonomiques différentes, les analyses de la diversité utilisées actuellement tiennent compte de l’abondance et d’autres paramètres. On n’a pas évalué si les MorphOTUs et les protocoles d’entente donneraient des résultats suffisamment équivalents dans la recherche réelle sur les diversités α et β, c’est-à-dire pour évaluer s’ils pouvaient tout de même tous deux être utiles. / The great insect diversity and the quantity of insect specimens collected during sampling constitute the biggest challenges facing insect systematics. Sorting samples to the species level is necessary before they can be used for investigations of biodiversity patterns. Because of the Taxonomic impediment, the lack of taxonomic expertise, many insect diversity studies sort specimens to Morphological Operational Taxonomic Units (MorphOTUs), also known as morphospecies, classifying subjectively defined groups based on obvious morphological features. However, accurately defining species boundaries based on MorphOTUs is time consuming and questionable, especially in groups with tiny insects and great species-level similarity such as Hymenoptera. DNA barcoding, a taxonomic discriminatory approach that employs DNA sequences, has accelerated taxonomic classification and may be an alternative approach to MorphOTUs. However, it is crucial to use a reliable and economic DNA barcoding strategy to deal with a large number of samples. Additionally, DNA barcoding should work with species problematic in molecular entomology as is sometimes observed with Hymenoptera. In order to implement a rapid biodiversity assessment of Hymenoptera, optimizing the DNA barcoding steps (DNA extraction, PCR amplification, and DNA sequencing) was the first objective of this present research. We tested and optimized a DNA extraction arriving at a method costing 0.20CAD per specimen. We validated the adequate performance of 313bp mini barcodes for establishing Molecular Operational Taxonomic Units (MOTUs) classification, comparable to that of the commonly used full-length DNA barcode of 658bp. We adopted this optimized protocol to barcode 517 Hymenoptera specimens sampled with suction traps located in the Laurentian Forest of eastern Canada. With multiplexed Illumina high throughput sequencing of tagged amplicons, we obtained mini-barcodes for 88% of specimens. The cost and time taken to generate MOTU data through our DNA barcoding approach was approximately twice that of morphological identification for MorphOTU designation. The second objective of my research was to compare the efficacy of morphological sorting of MorphOTUs with the molecular identification and delimitation of MOTUs. We found a high taxonomic congruence between morphological and molecular identification at family level in Barcode of Life (BOLD) and GenBank databases (93%), whereas only 18% of mini-barcode data was assigned to more precise identification (genus or species). Molecular delimitation based on four different clustering methods (distance-based: Automatic Barcode Gap Discovery v (ABGD) and Assemble Species by Automatic Partitioning (ASAP); tree-based: Generalized Mixed Yule Coalescent (GMYC) and Bayesian Poisson Tree Processes (bPTP)) resulted in more than doubling the estimated diversity of Hymenoptera as compared to MorphOTUs. The MOTUs were largely incongruent with MorphOTUs (match ratio <0.35). Distance-based methods gave results more congruent with morphological sorting than tree-based methods, especially within the Chalcidoidea superfamily. Taking into account the comparison between the cost and time of molecular and morphological classification methods, our results suggest that mini-DNA barcoding to estimate a proxy for Hymenoptera species diversity is more economical than MorphOTU sorting. However, although MorphOTU and MOTU methods gave different numbers of species, actual diversity analyses take into account abundance and other parameters. We did not evaluate whether MorphOTUs and MOTUs would yield sufficiently equivalent results in actual α- and β diversity research: that is, they may yet both be fit for purpose

No description available.

L'introgression est un processus fréquent et qui a d'importantes conséquences évolutives. L'objectif de ce travail était de tester un modèle neutre d'introgression chez des épicéas du Plateau tibétain et des régions voisines. Le travail a permis de mettre en évidence que la direction de l'introgression pouvait être prédite par la dynamique passée des populations d'arbres, et que l'importance de cette introgression était inversement proportionnelle à l'intensité des échanges génétiques au sein de l'espèce invasive, grâce à la comparaison de la structure génétique basée sur des marqueurs chloroplastiques (à hérédité paternelle) et mitochondriaux (à hérédité maternelle). / Introgression is a widespread phenomenon with potentially profound evolutionaryconsequences. Recently, significant progress in our understanding of introgression hasbeen made with the development of a neutral model. This model predicts that, whenone species invades an area already occupied by a related species, introgression ofneutral genes takes place mainly from the local species towards the invading ones. Inaddition, following a contact between two hybridizing species, the model predicts thatintrogression should be particularly frequent for genome components experiencinglittle gene flow. However, to date, there was no empirical example available, in whichone species expanded into the range of a closely related one and two markers withcontrasted rates of gene flow had been studied for both species. Only in such a casecould the two predictions outlined above be tested simultaneously. In addition, basedon these two predictions, species delimitation should be more efficient when usingmolecular markers experiencing high rates of gene flow. The present thesis was designed to test the hypotheses of this model. The biologicalmodel used was conifers, a group in which introgression and hybridization arecommon because of incomplete reproductive isolation. The species investigatedbelong to the genus Picea (spruce). We focused on two species complexes,represented by monographic clades in a phylogenetic study using the chloroplast genematK. All species studied occur in the Qinghai-Tibetan Plateau (QTP) and adjacenthighlands. The phylogeography of these species complexes was reconstructed usingorganelle markers (mitochondrial DNA, mtDNA and chloroplast DNA, cpDNA). Inconifers, mtDNA and cpDNA have contrasted modes of inheritance. The former ismaternally inherited, transmitted by seeds experiencing little gene flow while thelatter is paternally inherited, transmitted by both pollen and seeds experiencing highlevels of gene flow. Therefore, uniparentally inherited mtDNA and cpDNA markersexperience different rates of gene flow in such a group, providing an ideal model to test the relationship between rates of gene flow, introgression and species delimitation.Two mtDNA fragments (nad1intron b/c; nad5 intron1) and three cpDNA fragments(ndhK-C；trnL-trnF；trnS-trnG) were sequenced for nine species belonging to thePicea asperata and P. likiangensis species complex.(1) Nine mtDNA and nine cpDNA haplotypes were detected in 459 individualsfrom 46 natural populations in five species of P. asperata complex. As found in mostconifer species studied so far, low variation is present in the two mtDNA intronsalong with a high level of differentiation among populations (GST = 0.90). In contrast,higher variation and lower differentiation among populations was found at cpDNAmarkers (GST = 0.56). The cpDNA, although far from being fully diagnostic, is morespecies-specific than mtDNA: four groups of populations were identified usingcpDNA markers, all of them related to species or groups of species, whereas formtDNA, geographical variation prevails over species differentiation. A literaturereview shows that mtDNA variants are often shared among related conifer species,whereas cpDNA variants are more species-specific. Hence, increased intraspecificgene flow appears to decrease differentiation within species but not among species.[...]