Nous nous sommes proposé d’étudier la constitution d’un épuisement professionnel (burnout) chez une population de 105 thérapeutes algériens (psychiatres et psychologues) étant intervenus auprès des victimes des événements tragiques qui ont frappé l’Algérie dans les années 1990, 2001 et 2003, à savoir une vague d’attentats terroristes, l’inondation de Bab el Oued et le séisme de Boumerdès. Notre hypothèse de départ était étayée par le fait que thérapeutes et patients, relevant du même contexte psychosocial, avaient été les uns et les autres psychologiquement perturbés par ces événements et que, pour les thérapeutes, l’écoute empathique des récits d’horreur racontés par leurs patients pouvait exercer un effet de traumatisation vicariante, source de trauma secondaire et d’installation plus rapide d’un épuisement professionnel. Pour explorer l’expérience vécue de ces thérapeutes, nous leur avons fait passer un questionnaire de 28 items, précisé et complété d’une façon anonyme; et nous avons recherché chez eux d’une part l’existence d’un état de stress post-traumatique en nous référant aux critères cliniques du DSM-IV-TR, et d’autre part l’existence d’un burnout par application du Maslach Burnout Inventory (MBI). Nous avons croisé nos résultats avec le sexe, l’âge, la profession et l’ancienneté dans la profession ; et aussi en opposant les deux sous-populations de thérapeutes : ceux qui avaient été directement exposés à un événement potentiellement traumatisant, et ceux qui ne l’avaient pas été. Les résultats ont montré que la presque totalité de ces thérapeutes intervenants remplit les critères A et B du DSM-IV-TR (PTSD).Cela laisse sous entendre que vivre - directement ou indirectement - un traumatisme en Algérie, quels que soient sa violence et les vécus d’horreur et d’impuissance qu’il engendre, est susceptible de faire naître ensuite des symptômes d’intrusion. Par contre, les attitudes d’évitement et les manifestations neurovégétatives (critères C et D du DSM-IV-TR) n’étaient présentes que pour certains, représentant les conséquences de leur détresse, sans constituer pour autant la totalité de cette détresse et des autres aspects éprouvants du vécu. Ces tableaux sémiologiques laissent transparaître un accomplissement personnel plus faible du thérapeute à l’égard de son travail, voire le développement en cours d’un état d’épuisement professionnel. Les réponses au questionnaire montrent aussi que l’auto insatisfaction, le désir de donner plus et la crainte de « mal faire » en temps de crise ont été au centre des préoccupations des thérapeutes algériens. Aucun des critères de sexe, de profession, d’âge et d’expérience professionnelle n’est significativement corrélé avec la vulnérabilité du thérapeute confronté à des visions d’horreur ou des images tragiques sur le terrain ; ni corrélé avec les résultats obtenus à l’inventaire de burnout de Maslach. Il en est autrement pour les éventuelles décharges émotives que les thérapeutes ont pu manifester lors de leurs interventions auprès des traumatisés ; ces décharges émotives sont plus le fait des femmes que des hommes. Par ailleurs, la vulnérabilité à l’impact du trauma trouve ses assises dans la personnalité de chacun, et dans les expériences antérieures de chaque personne. Avoir vécu directement le terrorisme, les catastrophes naturelles ou d’autres événements potentiellement traumatisants ainsi que les autres traumatismes collectifs, ne favorise pas l’épuisement émotionnel, ni l’autoévaluation négative à travers la réduction de l’accomplissement personnel dans l’intervention du thérapeute. Par ailleurs, les thérapeutes qui n’ont personnellement vécu que les catastrophes naturelles semblent avoir moins tendance à réagir par le cynisme dans leur relation au patient, ou par la déshumanisation : le fait de partager le même contexte psychosocial avec leurs semblables les inciterait à faire preuve de plus de compassion. Il semble évident que le thérapeute algérien a fait de son mieux pour offrir le meilleur de lui, au détriment de son propre vécu. Il ne s’est pas permis de « lâcher », compte tenu de son statut de soignant. Il avait aussi conscience de contribuer à la reconstruction de toute une génération, à travers l’écoute empathique et le bon geste offerts aux victimes. / We decided to study the development of burnout in a group of 105 Algerian therapists (psychiatrists and psychologists) who treated victims of various tragic events that happened in Algeria during the 1990s, in 2001 and in 2003, namely several terrorist attacks, the Boumerdès earthquake, and the flood of Bab-el-Oued. Our initial hypothesis was supported by the fact that therapists and patients coming from the same psycho-social context were both affected psychologically by those events, and that for the therapists, listening empathetically to the horror stories told by their patients could create a vicarious trauma, which is a source of secondary trauma and a factor favorable to a faster occurrence of burnout.In order to investigate the experiences lived by those therapists, we presented them with a 28-question survey to be answered anonymously. We looked for the presence of post-traumatic stress disorder (PTSD) using clinical criteria referenced in the DSM-IV-TR and the presence of a burnout according to the Maslach Burnout Inventory (MBI). We cross-examined our results with the sex, age, profession and seniority level and by comparing two sub-groups of therapists: those who had been directly exposed to a potentially traumatic event and those who had not. The results demonstrated that almost all the therapists met criteria A and B of the DSM-IV-TR (PTSD). This implies that experiencing – directly or indirectly – a trauma in Algeria, irrelevantly of the violence, the horror and the feelings of helplessness experienced, is likely to lead to symptoms of intrusion. On the other hand, attitudes of avoidance and neuro-vegetative manifestations (criteria C and D of the DSM-IV-TR) were only present for some, representing the result of their distress without however constituting the totality of this distress and the other trying aspects of their experience. These tables of symptoms show a weaker self-fulfillment regarding work and possibly the initial stage of burnout. The survey answers also show that self-dissatisfaction, the desire to give more, and the fear of failure in times of crisis were the main source of anxiety for the Algerian therapists.Criteria such as sex, profession, age and seniority do not correlate significantly with the vulnerability of the therapists exposed to horrific or tragic scenes in the field, nor do they correlate with the results obtained from the MBI. We cannot say the same of the possible emotional releases that the therapists may have expressed while intervening with their traumatized patients: those emotional releases are more present for women than men. In other respects, the vulnerability caused by the trauma is rooted in one’s personality and past experiences. Having directly experienced terrorism, natural catastrophes or other potentially traumatizing events as well as the other mass traumas, does not favor the burnout or the negative self-evaluation through diminished self-accomplishment during the therapists’ interventions. In other respects, therapists who have personally experienced natural catastrophes only seem to be less likely to react by dehumanizing the patient-therapist relationship; sharing the same psycho-social context with their fellow citizens would invite them to show more compassion. It seems obvious that the Algerian therapists did all they could to offer the best of themselves, at the expense of their own experience. They did not allow themselves to give up because of their practitioners’ status. They also found important to contribute to the rebuilding of an entire generation through their empathetic listening and good deeds offered to the victims.

Myrmecophytic plant species form associations where the ant colony inhabits structures in the plant and offers protection against herbivory in exchange for food and shelter. Widely distributed across the tropics, myrmecophytic mutualisms are particularly diverse in the Neotropics, a region characterized by the rapid and recent uplift of the Andean mountain range. It has been suggested that the abrupt change in terrain triggered the emergence of new niches, new barriers to gene ow and speciation. Studying ant-plant associations in the Neotropics not only provides insight into how associations evolve in time but also the impact that external factors, such as geographic changes, have in the evolution of mutualisms. Because of its wide distribution on both sides of the Andes, The Tococa guianensis- Azteca system is useful to explore the effects the Andean uplift had on the evolution of mutualisms. This thesis aims to 1. Identify the ants associating with T. guianensis and the lineages of ants and plants involved in the mutualisms in different populations on both sides of the Andes, 2. generate genomic data for both ants and plants to increase sampling of loci, and 3. estimate and calibrate the species trees to compare patterns of phylogenetics and temporal congruence between ants, plants and the Andean uplift. Most ant-plant studies focus on only one partner or study both partners by using already collected data for one of them. This project is the first study inferring the evolutionary history of both partners associated at that point in time and across a large area. This thesis identifies two main Azteca lineages associated with T. guianensis, each one distributed on different sides of the Andes. It addresses the monophyly of T. guianensis (and related species) and why such monophyly cannot be confirmed. Results show how both plants and ants were geographically structured congruent with timing of a split of populations coinciding with the Andean uplift. Moreover, four plants and fifteen ant genomes were assembled and used to estimate gene and species trees. For Tococa, candidate markers were selected for future resolution of the plant's phylogeny. Different histories but similar divergence times between ants and plants suggest that the mutualism has evolved in response to geographic changes rather than through codiversication, but that the mutualism persists thanks to the availability of the host. The information generated during this study provides the basis to understand the evolution of mutualisms, the genomic features of ants and plants and opens the possibility for Tococa and Azteca to become a model system.

One of the original goals of phylogeography was to use genetic data to identify historical events that might contribute to breaks among biotic communities. In this study, we examine the phylogeography of a common livebearing fish (Poecilia gillii) from Costa Rica. Our goal was to see if phylogeographic breaks in this species were congruent with previously-defined boundaries among four fish community provinces. We hypothesized that if abiotic factors influence both community boundaries and genetic structuring in P. gillii then we would find four monophyletic clades within our focal species that were geographically separated along community boundary lines. Similarly, we expected to find most of the genetic variation in P. gillii partitioned among these four geographic regions. We generated DNA sequence data (mitochondrial cyt b and nuclear S7 small ribosomal subunit) for 260 individuals from 42 populations distributed across Costa Rica. We analyzed these data using phylogenetic (parsimony and likelihood) and coalescent approaches to estimate phylogenetic relationships among haplotypes, patterns of gene flow, and effective population size. Contrary to our expectations, we did not find four monophyletic groups that mapped cleanly to our geographic community provinces. However, one of our clades was restricted to a single province, suggesting that common earth history events could be responsible for both genetic structuring in P. gillii and fish community composition in this area. However, our results show a complex pattern of gene flow throughout other regions in Costa Rica where genetic structuring is not governed by community province boundaries.

Ecosystems transition quickly in the Anthropocene, whereas biodiversity adapts more slowly. Here we simulated a shifting woodland ecosystem on the Colorado Plateau of western North America by using as its proxy over space and time the fundamental niche of the Arizona black rattlesnake (Crotalus cerberus). We found an expansive (= end-of-Pleistocene) range that contracted sharply (= present), but is blocked topographically by Grand Canyon/Colorado River as it shifts predictably northwestward under moderate climate change (= 2080). Vulnerability to contemporary wildfire was quantified from available records, with forested area reduced more than 27% over 13 years. Both 'ecosystem metrics' underscore how climate and wildfire are rapidly converting the Plateau ecosystem into novel habitat. To gauge potential effects on C. cerberus, we derived a series of relevant 'conservation metrics' (i.e. genetic variability, dispersal capacity, effective population size) by sequencing 118 individuals across 846 bp of mitochondrial (mt)DNA-ATPase8/6. We identified five significantly different clades (net sequence divergence = 2.2%) isolated by drainage/topography, with low dispersal (F-ST = 0.82) and small sizes (2N(ef) = 5.2). Our compiled metrics (i.e. small-populations, topographic-isolation, low-dispersal versus conserved-niche, vulnerable-ecosystem, dispersal barriers) underscore the susceptibility of this woodland specialist to a climate and wildfire tandem. We offer adaptive management scenarios that may counterbalance these metrics and avoid the extirpation of this and other highly specialized, relictual woodland clades.

Na presente Tese foram usados métodos de inferência filogenética e de filogeografia buscando identificar os processos históricos de diversificação do gênero Conopophaga na América do Sul, em especial na Mata Atlântica. O monofiletismo do gênero e a estrutura filogeográfica das espécies distribuídas no sudeste da Mata Atlântica (Conopophaga lineata e C. melanops), foram testados utilizando seqüências de DNA mitocondrial. Para a filogenia foram utilizadas duas matrizes, sendo uma de 2270 pb (941 pb da subunidade 2 da NADH desidrogenase (ND2), 343 pb do ND3 e 986 pb do citocromo b) e outra de 878 pb (461 pb do ND2 e 417 pb do cit b). Nas análises de filogeografia de C. lineata e C. melanops foram utilizadas seqüências da região controladora de 472 pb e 439 pb, respectivamente. Os resultados demonstraram que o gênero Conopophaga é monofilético e que provavelmente uma rápida radiação ocorreu nesse gênero depois da especiação de C. melanogaster e de C. melanops. Dessa radiação, foram recuperados dois grupos: (1) Um grupo que se distribui somente na Amazônia e mantém a característica ancestral da coloração negra da mandíbula e (2) um grupo distribuindo-se na Amazônia e também na Mata Atlântica e que possui a mandíbula branca. Nesse último grupo, C. l. cearae não se agrupou com C. lineata, demonstrando que essa espécie não é uma espécie monofilética. A relação entre as espécies que apresentam a mandíbula branca parece indicar a ocorrência de uma conexão entre o leste da Amazônia e a Mata Atlântica no passado. O estudo filogeográfico de C. lineata revelou a existência de possíveis eventos de vicariância: (1) na região compreendida pelo Vale do Rio Paraíba do Sul e (2) à oeste de São Paulo e Paraná, separando as populações mais ao sul. Apesar de as inferências filogenéticas realizadas em C. melanops e C. lineata não serem totalmente concordantes, é possível que exista um padrão de vicariância nessa região. Concluindo, a ocorrência desses eventos vicariantes, tais como eventos geológicos e ciclos de alterações climáticas tenham influenciado na diversificação da família Conopophagidae. Além disso, eventos de dispersão e/ou seleção também podem auxiliar no entendimento da história biogeográfica do grupo, bem como de outros grupos na América do Sul. / In order to identify the historical processes of diversification of the gender textitConopophaga in South America, especially in the Atlantic forest, methods of phylogenetic and phylogeography inference were used in the present thesis. The genus phylogeny and the phylogeographic structure of two species ( textitConopophaga lineata and textitC. melanops) which occurs in the Southeast of the Atlantic forest were tested using sequences of mithocondrial DNA. Two matrixes were used to perform the phylogenetic analyses. The first one comprising 2270 bp (941 bp of ND2, 343 bp of ND3 and 986 bp of cytochrome b) and the second one comprising of 878 bp (461 bp of ND2 and 417 bp of cytochrome b). The phylogeography analyses of textitC. lineata and textitC. melanops were done using sequences from the control region consisting of 472 bp and 439 bp, respectively. The results demonstrated that the genus textitConopophaga is monophyletic and probably after textitC. melanogaster and textitC. melanops speciation, a rapid diversification had occurred in this genus. Following this event two distinct groups were recovered: (1) a group distributed in Amazonian, which maintains the ancestral characteristic of black jaw and (2) a group possessing white jaw occurring in the Amazonian and also in the Atlantic forest. In the last group, the subspecies C. l. cearae did not grouped with textitC. lineata demonstrating that this species is not monophyletic. Moreover, the distribution pattern of species presenting white jaw indicates a plausible a connection between the east of the Amazonian and the Atlantic forest in the past. The phylogeographic study of textitC. lineata revealed the existence of possible vicariant events: (1) in the area of Vale do Rio Paraíba do Sul and (2) in the west of São Paulo and Paraná, separating the southern south populations. Although the phylogeographic structure observed in textitC. melanops and in textitC. lineata are not in total agreement, the occurrence of vicariant events still remains as a possible explanation for the phylogeographic patterns in this region. Finally, the occurrence of these vicariant events like, geological events and climatic oscilations, may have influenced the diversification of the family Conopophagidae. Moreover, dispersion events and/or selection should also be considered for the understanding of biogeographic history of this group and also other ones in South America.

Since the mid 19th century, the biogeography of island South-East Asia has been the subject of much study. Early researchers explained many of the species distribution patterns by the rise and fall of sea levels and land. This and the work of other researchers culminated in a theory that emphasized the role of Pleistocene sea level low stands in species evolution. With the advent of newly developed molecular techniques, however, it became clear that many species divergence events had taken place before the Pleistocene and a biogeographical theory focusing on Pleistocene sea level changes was inadequate. In this research, I have developed a new biogeographic model that explains present-day distribution patterns and evolutionary relationships between species. I use this new model to explain 10 ‘mammalian riddles’, i.e. evolutionary or distribution patterns in selected mammal species groups that could not be explained with the existing theories. I developed the new model by analyzing the geological literature for this region, and by mapping palaeogeographical and palaeoenvironmental changes for the last 20 million years. In addition I compiled information on the palaeontological record for the region and on divergence times between taxa using a molecular clock assumption. These phylogenetic data were compared with the palaeomaps to assess whether particular divergence events could be correlated with certain palaeogeographical or palaeoenvironmental changes. The combination of these two information sources has resulted in a much-improved understanding of mammalian evolution in island SE Asia. Using this model it is now possible to relate important palaeoenvironmental events, such as the Late Miocene cooling, an Early–Middle Pliocene highstand, or the emergence and submergence of a land bridge between the Malay Peninsula and Java to evolutionary changes in species. I test the accuracy of the new model by analysing the relationships within several mammal groups using craniometric and molecular analysis. The observed relationships and deduced timing of divergence between taxa could in many cases be explained by the model, which indicates that it is relatively accurate. In addition, with the new model I have been able to find solutions to most mammalian riddles, although these results require further testing. Overall, I therefore believe I have made a significant contribution to the biogeographical understanding of island SE Asia.

To test competing hypotheses about the relative roles of vicariance and dispersal in the freshwater fishes in Galaxiidae, a phylogenetic framework and a time scale for species divergence were estimated using mitochondrial and nuclear DNA sequences. Relaxed clock dating revealed that the Galaxiidae originated in Gondwana in the late Cretaceous and thus vicariance cannot be ruled out for the two basal lineages, Galaxiella and Brachygalaxias. These two lineages are ancient relicts that rafted to their present distributions or were fragmented by the separation of Australia from South America via Antarctica. The opening of the Drake passage between South America and Antarctica initiated the proto-Antarctic Circumpolar Current (pACC) and counter-clockwise circulation in the South Atlantic, on which marine stage ancestors could have dispersed to South Africa and New Zealand via Australia during the late Eocene. Thus dispersal explains the disjunct distribution of the clade comprised of G. platei, G. zebratus and Neochanna spp. in South America, South Africa, and Australasia. The narrowing of the Drake passage and collapse of the pACC from about 24-14 Mya likely prevented further contact between South America and South Africa. Tectonic events around the globe produced an anomalous warming event, which along with the uplift of New Zealand provided empty niches and promoted the radiation of Galaxias. Most of the speciation in the other clades occurred during this time. When the Antarctic Circumpolar Current was reactivated at the end of the Mid-Miocene Climatic Optimum (MMCO) about 15–14 Mya, it provided a marine conveyor belt for the dispersal of the ancestor of Galaxias maculatus from Australia to South America, and later to Australia and New Zealand. The integration of divergence times estimated on the phylogeny with ancestral area reconstruction supports an origin in Gondwana and subsequent oceanic dispersal as the explanation for the distribution of the Galaxiidae across the southern continents.

To test competing hypotheses about the relative roles of vicariance and dispersal in the freshwater fishes in Galaxiidae, a phylogenetic framework and a time scale for species divergence were estimated using mitochondrial and nuclear DNA sequences. Relaxed clock dating revealed that the Galaxiidae originated in Gondwana in the late Cretaceous and thus vicariance cannot be ruled out for the two basal lineages, Galaxiella and Brachygalaxias. These two lineages are ancient relicts that rafted to their present distributions or were fragmented by the separation of Australia from South America via Antarctica. The opening of the Drake passage between South America and Antarctica initiated the proto-Antarctic Circumpolar Current (pACC) and counter-clockwise circulation in the South Atlantic, on which marine stage ancestors could have dispersed to South Africa and New Zealand via Australia during the late Eocene. Thus dispersal explains the disjunct distribution of the clade comprised of G. platei, G. zebratus and Neochanna spp. in South America, South Africa, and Australasia. The narrowing of the Drake passage and collapse of the pACC from about 24-14 Mya likely prevented further contact between South America and South Africa. Tectonic events around the globe produced an anomalous warming event, which along with the uplift of New Zealand provided empty niches and promoted the radiation of Galaxias. Most of the speciation in the other clades occurred during this time. When the Antarctic Circumpolar Current was reactivated at the end of the Mid-Miocene Climatic Optimum (MMCO) about 15–14 Mya, it provided a marine conveyor belt for the dispersal of the ancestor of Galaxias maculatus from Australia to South America, and later to Australia and New Zealand. The integration of divergence times estimated on the phylogeny with ancestral area reconstruction supports an origin in Gondwana and subsequent oceanic dispersal as the explanation for the distribution of the Galaxiidae across the southern continents.

Na presente Tese foram usados métodos de inferência filogenética e de filogeografia buscando identificar os processos históricos de diversificação do gênero Conopophaga na América do Sul, em especial na Mata Atlântica. O monofiletismo do gênero e a estrutura filogeográfica das espécies distribuídas no sudeste da Mata Atlântica (Conopophaga lineata e C. melanops), foram testados utilizando seqüências de DNA mitocondrial. Para a filogenia foram utilizadas duas matrizes, sendo uma de 2270 pb (941 pb da subunidade 2 da NADH desidrogenase (ND2), 343 pb do ND3 e 986 pb do citocromo b) e outra de 878 pb (461 pb do ND2 e 417 pb do cit b). Nas análises de filogeografia de C. lineata e C. melanops foram utilizadas seqüências da região controladora de 472 pb e 439 pb, respectivamente. Os resultados demonstraram que o gênero Conopophaga é monofilético e que provavelmente uma rápida radiação ocorreu nesse gênero depois da especiação de C. melanogaster e de C. melanops. Dessa radiação, foram recuperados dois grupos: (1) Um grupo que se distribui somente na Amazônia e mantém a característica ancestral da coloração negra da mandíbula e (2) um grupo distribuindo-se na Amazônia e também na Mata Atlântica e que possui a mandíbula branca. Nesse último grupo, C. l. cearae não se agrupou com C. lineata, demonstrando que essa espécie não é uma espécie monofilética. A relação entre as espécies que apresentam a mandíbula branca parece indicar a ocorrência de uma conexão entre o leste da Amazônia e a Mata Atlântica no passado. O estudo filogeográfico de C. lineata revelou a existência de possíveis eventos de vicariância: (1) na região compreendida pelo Vale do Rio Paraíba do Sul e (2) à oeste de São Paulo e Paraná, separando as populações mais ao sul. Apesar de as inferências filogenéticas realizadas em C. melanops e C. lineata não serem totalmente concordantes, é possível que exista um padrão de vicariância nessa região. Concluindo, a ocorrência desses eventos vicariantes, tais como eventos geológicos e ciclos de alterações climáticas tenham influenciado na diversificação da família Conopophagidae. Além disso, eventos de dispersão e/ou seleção também podem auxiliar no entendimento da história biogeográfica do grupo, bem como de outros grupos na América do Sul. / In order to identify the historical processes of diversification of the gender textitConopophaga in South America, especially in the Atlantic forest, methods of phylogenetic and phylogeography inference were used in the present thesis. The genus phylogeny and the phylogeographic structure of two species ( textitConopophaga lineata and textitC. melanops) which occurs in the Southeast of the Atlantic forest were tested using sequences of mithocondrial DNA. Two matrixes were used to perform the phylogenetic analyses. The first one comprising 2270 bp (941 bp of ND2, 343 bp of ND3 and 986 bp of cytochrome b) and the second one comprising of 878 bp (461 bp of ND2 and 417 bp of cytochrome b). The phylogeography analyses of textitC. lineata and textitC. melanops were done using sequences from the control region consisting of 472 bp and 439 bp, respectively. The results demonstrated that the genus textitConopophaga is monophyletic and probably after textitC. melanogaster and textitC. melanops speciation, a rapid diversification had occurred in this genus. Following this event two distinct groups were recovered: (1) a group distributed in Amazonian, which maintains the ancestral characteristic of black jaw and (2) a group possessing white jaw occurring in the Amazonian and also in the Atlantic forest. In the last group, the subspecies C. l. cearae did not grouped with textitC. lineata demonstrating that this species is not monophyletic. Moreover, the distribution pattern of species presenting white jaw indicates a plausible a connection between the east of the Amazonian and the Atlantic forest in the past. The phylogeographic study of textitC. lineata revealed the existence of possible vicariant events: (1) in the area of Vale do Rio Paraíba do Sul and (2) in the west of São Paulo and Paraná, separating the southern south populations. Although the phylogeographic structure observed in textitC. melanops and in textitC. lineata are not in total agreement, the occurrence of vicariant events still remains as a possible explanation for the phylogeographic patterns in this region. Finally, the occurrence of these vicariant events like, geological events and climatic oscilations, may have influenced the diversification of the family Conopophagidae. Moreover, dispersion events and/or selection should also be considered for the understanding of biogeographic history of this group and also other ones in South America.

The fish order Pleuronectiformes (flatfishes) include over 800 fish species, and are generally considered to be monophyletic. Phylogenetic relationships within the order remain largely unresolved, particularly for the Paralichthyidae, a very poorly defined family. In addition to the lack of consensus on intrarelationships within this order, flatfish diversification patterns are still poorly understood. Here we want to test the hypothesis that (i) a densely sampled phylogenetic tree will help to unravel the phylogenetic relationships within flatfish, and (ii) that the formation of the Isthmus of Panama has played a role in the diversification of flatfish, following their distribution in both oceans, Atlantic and Pacific. Maximum likelihood and Bayesian analyzes were performed on a dataset consisting of nine mitochondrial and nuclear loci on a set of 309 flatfish species. The heterogeneity of substitution rates, known to cause phylogenetic errors in large multigene alignments, can be corrected by data partitioning, or by using heterogeneous site models. Very few studies have compared the effects of using these approaches. We present a robust phylogenetic tree with different heterogeneous and well-supported models, and the intraordinal relationships of flatfish are discussed in light of these results. We present evidence for three new suborders within the flatfishes. The monophyletic status of most families is corroborated. The family Paralichthyidae and the family Poecilopsettidae were found to be polyphyletic. Heterogeneous models have not been able to solve the deeper relationships between flatfish. We also show that flatfish diversification occurred in a few families in at least two stages, related to distinct geological events.