<p>The evolution of selfing from outcrossing is one of the most frequent mating system transitions in angiosperms. Plants that are highly selfing typically exhibit a suite of morphological traits termed a "selfing syndrome," including reduced corollas and reproductive structures, loss of corolla pigmentation, little anther-stigma separation, and a low pollen/ovule ratio. The overall consensus among scientist is that the morphological changes that accompany the transition to selfing are adaptive and thus a product of natural selection. Few attempts, however, have been made to determine whether traits of the selfing syndrome are truly an operation of natural selection or if genetic drift could be the acting force. My dissertation examines the roles that natural selection and genetic drift played in the evolution of the selfing syndrome in Ipomoea lacunosa. With the use of field observations, crossing data, and molecular analyses, I show that I. lacunosa has evolved increased selfing ability, decreased anther-stigma distance and smaller, white flowers, compared to its closest relative I. cordatotriloba. Furthermore, using a standard QST - FST comparison, I evaluated the relative importance of selection and drift in the evolution of the selfing syndrome in I. lacunosa. I also identified the genetic basis of flower color divergence between I. lacunosa (white) and I. cordatotriloba (purple) and examined patterns of variation to determine if selection or genetic drift caused the divergence. Analyses revealed that the traits of I. lacunosa characteristic of the selfing syndrome have evolved as a product of natural selection, not genetic drift.</p> / Dissertation

Introdução: apesar de depressão e dor serem altamente prevalentes em pacientes com Doença de Parkinson (DP), há poucos estudos sobre a relação entre esses fatores, apesar da já bem descrita potencial modulação da dor por estados emocionais. Objetivo: avaliar a percepção termoalgésica de calor e dor em método quantitativo e correlacioná-la com sintomas psiquiátricos e da Doença de Parkinson. Método: realizamos um estudo transversal avaliando características clínicas e dados psicofísicos em 31 pacientes com DP sob efeito da medicação dopaminérgica (estado “on”). Verificamos as características da DP utilizando a escala Hoehn and Yahr, realizamos uma avaliação psiquiátrica usando as escalas Inventário de Depressão de Beck (Inventário de Depressão de Beck - IDB), Mini International Neuropsychiatric Interview (MINI) de acordo com os critérios do DSM IV, e Mini Mental State Evaluation (MMSE), avaliamos queixas de dor nos últimos 90 dias usando uma escala visual analógica para dor (EVA – Escala Visual Analógica) e medimos a percepção termoalgésica através do Teste Quantitativo Sensitivo (TQS) para percepção de calor e de dor. Resultados: 31 pacientes foram avaliados. Surpreendentemente, não houve associação entre a percepção termoalgésica e as queixas de dor ou sintomas da DP. Entretanto, houve uma correlação moderada mas significativa entre sintomas depressivos medidos pela BDI e os limiares de calor e de dor (r=0.54 para calor p<0.05 e r=0.47 p<0.05 para dor). Pacientes com sintomas depressivos significativos tiveram limiares de calor e de dor maiores comparados aos sem sintomas depressivos. Esse achado se manteve após correção estatística para severidade dos sintomas da DP. Conclusão: processamento termoalgésico em pacientes com DP é mais influenciado por depressão do que pela severidade da Doença de Parkinson ou pelo nível da dor em si. Essa informação tem implicações importantes para o diagnóstico e abordagem terapêutica para pacientes com DP e dor e/ou depressão. Por exemplo, a depressão poderia ser mais sistematicamente rastreada e tratada em pacientes com DP com processamento de dor alterado. / Introduction: Although depression and pain are highly prevalent in Parkinson’s Disease (PD) patients, there is a lack of studies in their relationship, even though it is well-known that pain is potentially modulated by emotional state. Aims: To assess warm and heat pain perception in a quantitative method and correlate it with psychiatric and parkinsonian symptoms. Methods: We carried out a transversal study assessing clinical and psychophysical data in 31 patients with PD during the effect of dopaminergic medication (on state). We assessed the clinical characteristics of Parkinson's using Hoehn and Yahr (HY), performed a psychiatric evaluation using Beck Depression Inventory (BDI), Mini International Neuropsychiatric Interview (MINI) according to the DSM IV criteria and Mini Mental State Evaluation (MMSE), evaluated pain complaints in the last 90 days using a visual analogue scale for pain (VAS) and measured pain perception by means of quantitative sensory testing (QST) for warm and heat pain perception. Results: 31 patients were evaluated. Surprisingly, there was no association between thermoalgesic perception with pain complaints or parkinsonian symptoms. However, there was a moderate but significant correlation between depressive symptoms measured by BDI and warm sensation and heat pain thresholds(r=0.54 for warm p<0.05 and r=0.47 p<0.05 for heat pain). Patients with significant depressive symptoms had higher warm and heat pain thresholds compared to those without depression. This finding was maintained after statistical correction for the PD symptoms severity. Conclusion: Thermoalgesic processing in PD patients is more influenced by depression than by PD severity or level of pain itself. This information has important implications for diagnostic and therapeutic approaches for patients with PD and pain and/or depression. For instance, depression might be more systematically screened and treated in PD patients with altered pain processing.

Introdução: apesar de depressão e dor serem altamente prevalentes em pacientes com Doença de Parkinson (DP), há poucos estudos sobre a relação entre esses fatores, apesar da já bem descrita potencial modulação da dor por estados emocionais. Objetivo: avaliar a percepção termoalgésica de calor e dor em método quantitativo e correlacioná-la com sintomas psiquiátricos e da Doença de Parkinson. Método: realizamos um estudo transversal avaliando características clínicas e dados psicofísicos em 31 pacientes com DP sob efeito da medicação dopaminérgica (estado “on”). Verificamos as características da DP utilizando a escala Hoehn and Yahr, realizamos uma avaliação psiquiátrica usando as escalas Inventário de Depressão de Beck (Inventário de Depressão de Beck - IDB), Mini International Neuropsychiatric Interview (MINI) de acordo com os critérios do DSM IV, e Mini Mental State Evaluation (MMSE), avaliamos queixas de dor nos últimos 90 dias usando uma escala visual analógica para dor (EVA – Escala Visual Analógica) e medimos a percepção termoalgésica através do Teste Quantitativo Sensitivo (TQS) para percepção de calor e de dor. Resultados: 31 pacientes foram avaliados. Surpreendentemente, não houve associação entre a percepção termoalgésica e as queixas de dor ou sintomas da DP. Entretanto, houve uma correlação moderada mas significativa entre sintomas depressivos medidos pela BDI e os limiares de calor e de dor (r=0.54 para calor p<0.05 e r=0.47 p<0.05 para dor). Pacientes com sintomas depressivos significativos tiveram limiares de calor e de dor maiores comparados aos sem sintomas depressivos. Esse achado se manteve após correção estatística para severidade dos sintomas da DP. Conclusão: processamento termoalgésico em pacientes com DP é mais influenciado por depressão do que pela severidade da Doença de Parkinson ou pelo nível da dor em si. Essa informação tem implicações importantes para o diagnóstico e abordagem terapêutica para pacientes com DP e dor e/ou depressão. Por exemplo, a depressão poderia ser mais sistematicamente rastreada e tratada em pacientes com DP com processamento de dor alterado. / Introduction: Although depression and pain are highly prevalent in Parkinson’s Disease (PD) patients, there is a lack of studies in their relationship, even though it is well-known that pain is potentially modulated by emotional state. Aims: To assess warm and heat pain perception in a quantitative method and correlate it with psychiatric and parkinsonian symptoms. Methods: We carried out a transversal study assessing clinical and psychophysical data in 31 patients with PD during the effect of dopaminergic medication (on state). We assessed the clinical characteristics of Parkinson's using Hoehn and Yahr (HY), performed a psychiatric evaluation using Beck Depression Inventory (BDI), Mini International Neuropsychiatric Interview (MINI) according to the DSM IV criteria and Mini Mental State Evaluation (MMSE), evaluated pain complaints in the last 90 days using a visual analogue scale for pain (VAS) and measured pain perception by means of quantitative sensory testing (QST) for warm and heat pain perception. Results: 31 patients were evaluated. Surprisingly, there was no association between thermoalgesic perception with pain complaints or parkinsonian symptoms. However, there was a moderate but significant correlation between depressive symptoms measured by BDI and warm sensation and heat pain thresholds(r=0.54 for warm p<0.05 and r=0.47 p<0.05 for heat pain). Patients with significant depressive symptoms had higher warm and heat pain thresholds compared to those without depression. This finding was maintained after statistical correction for the PD symptoms severity. Conclusion: Thermoalgesic processing in PD patients is more influenced by depression than by PD severity or level of pain itself. This information has important implications for diagnostic and therapeutic approaches for patients with PD and pain and/or depression. For instance, depression might be more systematically screened and treated in PD patients with altered pain processing.

INTRODUÇÃO: A diabetes mellitus tipo 2 (DM) é uma doença metabólica caracterizada pela presença de hiperglicemia crônica. Estudos prévios demonstraram que pacientes com pré-diabetes (PDM) têm uma história natural de progressão para DM. A neuropatia diabética é a complicação mais comum da DM e avanços recentes na neurofisiologia clínica trouxeram um refinamento das técnicas de avaliação. Entre estas estão à resposta cutânea simpática (SSR) e o teste sensorial quantitativo (QST). Biomarcadores como Enolase Neurônio-Específica (NSE) e a Proteína S100-Beta (S100B) vem sendo descritos por muitos autores como associados a danos em células do sistema nervoso. OBJETIVO: O objetivo deste estudo é avaliar parâmetros neurofisiológicos e compará-los com achados clínicos e bioquímicos (S100B e NSE) em pacientes com DM, PDM e controles saudáveis. MÉTODOS: Pacientes dos ambulatórios de Neurologia e Endocrinologia foram randomizados em um estudo transversal. Os participantes foram submetidos a uma bateria de testes clínicos e neurofisiológicos que englobaram condução nervosa, Onda-F, SSR e QST. Níveis séricos de NSE e S100B foram quantificados através de ensaio ELISA (Enzyme-linked immunosorbent assay). RESULTADOS: A avaliação clínica e os estudos de condução nervosa e Onda-F foram similares nos grupos estudados. Já os limiares QST calor (QSTc) e QST dor (QSTd) foram significativamente elevados nos pacientes PDM e DM com relação aos controles (P<0.05 para todas as comparações). No entanto, estes parâmetros não foram capazes de distinguir pacientes DM vs. PDM (P >0.1 para todas as comparações). O SSR foi capaz de diferenciar o grupo DM do controle (P <0,01) embora não tenha mostrado diferença entre os grupos PDM e controle (P = 0,6). Não houve diferença entre os níveis de S100B (P = 0.6) e NSE (P = 0.2) entre os grupos DM, PDM e controles. CONCLUSÃO: O QST e SSR são testes úteis para a avaliação de pacientes com diferentes graus de tolerância a glicose. Este estudo não encontrou diferenças entre os biomarcadores NSE e S100B em indivíduos com DM e PDM. / BACKGROUND: Type 2 diabetes mellitus (DM) is a metabolic disease characterized by the presence of chronic hyperglycemia. Previous studies demonstrated that patients with prediabetes states (PDM) have a natural history of progression to DM. Neuropathy is the most common and disabling complication of diabetes and recent advances in neurophysiology have enabled a refinement of neurophysiological diagnostic techniques such as sympathetic skin response (SSR) and quantitative sensory testing (QST). Biomarkers like Neuron-specific Enolase (NSE) and S100- Beta Protein (S100B) has been described for many authors as associated with damage at nervous system cells and are related with severity of injury as well as clinical outcomes. OBJECTIVE: The aim of this study is to evaluate neurophysiological findings and compare them with clinical and biochemical findings (S100B and NSE) in patients with DM, PDM and healthy controls. METHODS: Patients at the outpatient Neurology and Endocrinology service were randomized in a cross-sectional study. Participants underwent a battery of clinical and neurophysiological tests that encompassed nerve conduction studies, F-wave, SSR and QST. ELISA (enzyme-linked immunosorbent assay) were perform to quantify serum levels of NSE and S100B. RESULTS: There were no difference regarding clinical evaluation, nerve conduction studies and F-wave were between groups. The QST thresholds of warm (QSTw) and QST pain (QSTp) were significantly elevated in patients with PDM and DM compared to controls (P <0.05 for all comparisons). However, these parameters were not able to distinguish among DM and PDM (P > 0.1 for all comparisons). The SSR was able to differentiate DM from control group (P <0.01) but did not show difference between PDM and control groups (P = 0.6). There was no difference on levels of S100B (P = 0.6) and NSE (P = 0.2) between the DM, PDM and control groups. CONCLUSION: The QST and SSR are useful tests to evaluating patients with different degrees of glucose tolerance. This study found no differences between biomarkers NSE and S100B in subjects with DM and PDM.

INTRODUÇÃO: A diabetes mellitus tipo 2 (DM) é uma doença metabólica caracterizada pela presença de hiperglicemia crônica. Estudos prévios demonstraram que pacientes com pré-diabetes (PDM) têm uma história natural de progressão para DM. A neuropatia diabética é a complicação mais comum da DM e avanços recentes na neurofisiologia clínica trouxeram um refinamento das técnicas de avaliação. Entre estas estão à resposta cutânea simpática (SSR) e o teste sensorial quantitativo (QST). Biomarcadores como Enolase Neurônio-Específica (NSE) e a Proteína S100-Beta (S100B) vem sendo descritos por muitos autores como associados a danos em células do sistema nervoso. OBJETIVO: O objetivo deste estudo é avaliar parâmetros neurofisiológicos e compará-los com achados clínicos e bioquímicos (S100B e NSE) em pacientes com DM, PDM e controles saudáveis. MÉTODOS: Pacientes dos ambulatórios de Neurologia e Endocrinologia foram randomizados em um estudo transversal. Os participantes foram submetidos a uma bateria de testes clínicos e neurofisiológicos que englobaram condução nervosa, Onda-F, SSR e QST. Níveis séricos de NSE e S100B foram quantificados através de ensaio ELISA (Enzyme-linked immunosorbent assay). RESULTADOS: A avaliação clínica e os estudos de condução nervosa e Onda-F foram similares nos grupos estudados. Já os limiares QST calor (QSTc) e QST dor (QSTd) foram significativamente elevados nos pacientes PDM e DM com relação aos controles (P<0.05 para todas as comparações). No entanto, estes parâmetros não foram capazes de distinguir pacientes DM vs. PDM (P >0.1 para todas as comparações). O SSR foi capaz de diferenciar o grupo DM do controle (P <0,01) embora não tenha mostrado diferença entre os grupos PDM e controle (P = 0,6). Não houve diferença entre os níveis de S100B (P = 0.6) e NSE (P = 0.2) entre os grupos DM, PDM e controles. CONCLUSÃO: O QST e SSR são testes úteis para a avaliação de pacientes com diferentes graus de tolerância a glicose. Este estudo não encontrou diferenças entre os biomarcadores NSE e S100B em indivíduos com DM e PDM. / BACKGROUND: Type 2 diabetes mellitus (DM) is a metabolic disease characterized by the presence of chronic hyperglycemia. Previous studies demonstrated that patients with prediabetes states (PDM) have a natural history of progression to DM. Neuropathy is the most common and disabling complication of diabetes and recent advances in neurophysiology have enabled a refinement of neurophysiological diagnostic techniques such as sympathetic skin response (SSR) and quantitative sensory testing (QST). Biomarkers like Neuron-specific Enolase (NSE) and S100- Beta Protein (S100B) has been described for many authors as associated with damage at nervous system cells and are related with severity of injury as well as clinical outcomes. OBJECTIVE: The aim of this study is to evaluate neurophysiological findings and compare them with clinical and biochemical findings (S100B and NSE) in patients with DM, PDM and healthy controls. METHODS: Patients at the outpatient Neurology and Endocrinology service were randomized in a cross-sectional study. Participants underwent a battery of clinical and neurophysiological tests that encompassed nerve conduction studies, F-wave, SSR and QST. ELISA (enzyme-linked immunosorbent assay) were perform to quantify serum levels of NSE and S100B. RESULTS: There were no difference regarding clinical evaluation, nerve conduction studies and F-wave were between groups. The QST thresholds of warm (QSTw) and QST pain (QSTp) were significantly elevated in patients with PDM and DM compared to controls (P <0.05 for all comparisons). However, these parameters were not able to distinguish among DM and PDM (P > 0.1 for all comparisons). The SSR was able to differentiate DM from control group (P <0.01) but did not show difference between PDM and control groups (P = 0.6). There was no difference on levels of S100B (P = 0.6) and NSE (P = 0.2) between the DM, PDM and control groups. CONCLUSION: The QST and SSR are useful tests to evaluating patients with different degrees of glucose tolerance. This study found no differences between biomarkers NSE and S100B in subjects with DM and PDM.

INTRODUÇÃO: A diabetes mellitus tipo 2 (DM) é uma doença metabólica caracterizada pela presença de hiperglicemia crônica. Estudos prévios demonstraram que pacientes com pré-diabetes (PDM) têm uma história natural de progressão para DM. A neuropatia diabética é a complicação mais comum da DM e avanços recentes na neurofisiologia clínica trouxeram um refinamento das técnicas de avaliação. Entre estas estão à resposta cutânea simpática (SSR) e o teste sensorial quantitativo (QST). Biomarcadores como Enolase Neurônio-Específica (NSE) e a Proteína S100-Beta (S100B) vem sendo descritos por muitos autores como associados a danos em células do sistema nervoso. OBJETIVO: O objetivo deste estudo é avaliar parâmetros neurofisiológicos e compará-los com achados clínicos e bioquímicos (S100B e NSE) em pacientes com DM, PDM e controles saudáveis. MÉTODOS: Pacientes dos ambulatórios de Neurologia e Endocrinologia foram randomizados em um estudo transversal. Os participantes foram submetidos a uma bateria de testes clínicos e neurofisiológicos que englobaram condução nervosa, Onda-F, SSR e QST. Níveis séricos de NSE e S100B foram quantificados através de ensaio ELISA (Enzyme-linked immunosorbent assay). RESULTADOS: A avaliação clínica e os estudos de condução nervosa e Onda-F foram similares nos grupos estudados. Já os limiares QST calor (QSTc) e QST dor (QSTd) foram significativamente elevados nos pacientes PDM e DM com relação aos controles (P<0.05 para todas as comparações). No entanto, estes parâmetros não foram capazes de distinguir pacientes DM vs. PDM (P >0.1 para todas as comparações). O SSR foi capaz de diferenciar o grupo DM do controle (P <0,01) embora não tenha mostrado diferença entre os grupos PDM e controle (P = 0,6). Não houve diferença entre os níveis de S100B (P = 0.6) e NSE (P = 0.2) entre os grupos DM, PDM e controles. CONCLUSÃO: O QST e SSR são testes úteis para a avaliação de pacientes com diferentes graus de tolerância a glicose. Este estudo não encontrou diferenças entre os biomarcadores NSE e S100B em indivíduos com DM e PDM. / BACKGROUND: Type 2 diabetes mellitus (DM) is a metabolic disease characterized by the presence of chronic hyperglycemia. Previous studies demonstrated that patients with prediabetes states (PDM) have a natural history of progression to DM. Neuropathy is the most common and disabling complication of diabetes and recent advances in neurophysiology have enabled a refinement of neurophysiological diagnostic techniques such as sympathetic skin response (SSR) and quantitative sensory testing (QST). Biomarkers like Neuron-specific Enolase (NSE) and S100- Beta Protein (S100B) has been described for many authors as associated with damage at nervous system cells and are related with severity of injury as well as clinical outcomes. OBJECTIVE: The aim of this study is to evaluate neurophysiological findings and compare them with clinical and biochemical findings (S100B and NSE) in patients with DM, PDM and healthy controls. METHODS: Patients at the outpatient Neurology and Endocrinology service were randomized in a cross-sectional study. Participants underwent a battery of clinical and neurophysiological tests that encompassed nerve conduction studies, F-wave, SSR and QST. ELISA (enzyme-linked immunosorbent assay) were perform to quantify serum levels of NSE and S100B. RESULTS: There were no difference regarding clinical evaluation, nerve conduction studies and F-wave were between groups. The QST thresholds of warm (QSTw) and QST pain (QSTp) were significantly elevated in patients with PDM and DM compared to controls (P <0.05 for all comparisons). However, these parameters were not able to distinguish among DM and PDM (P > 0.1 for all comparisons). The SSR was able to differentiate DM from control group (P <0.01) but did not show difference between PDM and control groups (P = 0.6). There was no difference on levels of S100B (P = 0.6) and NSE (P = 0.2) between the DM, PDM and control groups. CONCLUSION: The QST and SSR are useful tests to evaluating patients with different degrees of glucose tolerance. This study found no differences between biomarkers NSE and S100B in subjects with DM and PDM.

<p>The main aim of this work was to identify local adaptation processes in amphibian populations, thereby improving the general understanding of genetics and mechanisms behind the evolution and maintenance of biological diversity. Phenotypic and genetic variation in life-history traits was studied within and between populations common frog (<i>Rana temporaria</i>) populations along a 1600 km transect from southern Sweden to northern Finland.</p><p>Embryonic and larval development and growth was investigated both under field and laboratory conditions. The results suggest ample genetic diversity in larval life-history traits among Fennoscandian common frog populations. Larval developmental rate along the gradient has evolved a countergradient variation pattern of genotypes and phenotypes as indicated by the positive relationship between developmental rate and latitude under laboratory conditions and the lack of such a relationship in the field. The data suggest that this pattern has evolved because of time constraints due to decreasing length of growth season with latitude. Neither field-caught adults nor laboratory raised larvae displayed a linear latitudinal size cline as expected from the so called Bergmanns rule. Rather, size increased towards the mid-latitude populations and decreased thereafter, indicating that body size is a product of direct environmental induction or a trade-off with other life-history characters. Age and size at hatching showed no consistent latitudinal pattern, indicating that the embryonic stage is not as time constrained as the larval stage.</p><p>A large part of the variation in age and size at metamorphosis among populations was due to additive genetic effects. However, small, but significant maternal effects, mostly due to variation in egg size and non-additive genetic effects also contributed to among population variation. A comparison of divergence in presumably neutral molecular genetic markers (F_ST) and quantitative characters (Q_ST) revealed that although both estimates of divergence were relatively high, estimates of Q_ST was generally higher than those of F_ST, indicating that the genetic variation observed in larval traits is primarily a result of natural selection rather than genetic drift. Hence, our results reinforce the conclusion that intraspecific genetic heterogeneity in the young northern European ecosystems may be more widespread than previously anticipated</p>

<p>In a prospective randomised double-blind study eighty patients with MRI verified lumbar disc herniation and corresponding clinical findings underwent microscopic disc removal. The patients were peroperatively given systemic and local corticosteroids or placebo, and followed for 2 years. The hospital stay and time to return to full-time work was significantly shorter in the treatment group. Pain measured as worst pain during the last week was also lower in the corticosteroid group. The results indicate that peroperative treatment with corticosteroids reduces pain and improves the functional outcome in patients operated for lumbar disc herniations.</p><p>To evaluate whether thermal quantitative sensory testing (QST) is applicable in the study of sensory dysfunction in lumbar disc herniations 66 patients with disc herniations underwent thermal QST. We found that thermal QST reflects sensory dysfunction in patients with lumbar disc herniations. However, thermal QST seems to have a poor predictive value for identifying the anatomic location of a herniated lumbar disc.</p><p>Quantitative sensory testing (QST) was used to detect damage to the myelinated A-delta fibres (cold sense) and the unmyelinated C-fibres (warmth sense). Corticosteroids combined with surgery in lumbar disc surgery improved the normalisation for the warmth disturbance compared to the control group. </p><p>A prospective analysis was performed on the predictive value of preoperatively determined lumbar lordosis and flexion for pain and disability in patients treated by microscopic lumbar disc surgery. Preoperative hyperlordosis correlated to more pain postoperatively (p=0.004). In patients with hypoflexion there was an association between hyperlordosis and moderate or severe pain postoperatively (p<0.001). The same outcomes were found for DRI. The stiff and straight back indicates a good outcome of lumbar disc surgery concerning pain and disability. </p>

Le second chapitre introduit les différents couplages présents dans une analyse mécanique thermique métallurgique. Il rappelle ensuite les notions de métallurgie nécessaires à la compréhension de ce travail. Le troisième chapitre décrit la modélisation adoptée pour l'analyse thermique métallurgique non couplée au problème mécanique. Le modèle choisi est celui de FERNANDES (CP30). Le code développé par ce dernier était limité à l'étude de cylindres infiniment longs, c'est-à-dire à un problème unidimensionnel. L'intégration de ce modèle, dans le code LAGAMINE apporte une grande liberté quant aux formes géométriques traitées. Actuellement, les cas bidimensionnels (état plan ou axisymétrique) peuvent être traités, une extension aux cas tridimensionnels ne pose pas de problème si ce n'est celui des temps de calcul. Du point de vue théorique, nous avons proposé une formulation différentielle de la modélisation des transformations par diffusion. Cette approche, indépendante de celle de SJOSTROM (CP38) a permis de vérifier cette dernière par une voie différente. En outre, la comparaison de cette approche avec celle plus classique de FERNANDES constitue un élément nouveau intéressant. Pour terminer, des tests de validation des développements correspondants dans le code LAGAMINE sont présentés. Le chapitre quatre est consacré aux problèmes numériques à résoudre pour réaliser une simulation de refroidissement d'une pièce métallique. Une bibliographie des schémas d'intégration temporelle des problèmes thermiques linéaires et non linéaires est présentée. A ce niveau, nous ne pouvons parler de contribution originale puisqu'il s'agit de schémas devenus classiques. Cependant, il est assez rare de trouver un résumé de ces notions souvent dispersées dans de nombreux articles. A ce titre, nous considérons ce point comme un apport intéressant de cette thèse. La prise en compte des changements de phase a occasionné un grand nombre de recherches. Après avoir fait le point sur les méthodes proposées par la littérature, nous décrivons les développements introduits dans le code LAGAMINE. Le calcul des flux et de la matrice tangente est détaillé. Ce dernier point constitue un apport personnel original, surtout au niveau de la contribution des transformations de phase à la matrice tangente. Le chapitre cinq est consacré aux interactions existant entre les phénomènes mécaniques, thermiques et métallurgiques. Une présentation approfondie de chaque couplage est réalisée et la notion de plasticité de transformation est introduite. Notre contribution personnelle et originale réside dans :  une loi élasto-plastique thermique couplée pour un matériau monophasique,  une loi élasto-plastique thermique métallurgique pour un matériau multiphasique. L'intégration des lois élasto-plastiques thermiques, métallurgiques ou non, dans le code LA GAMINE constitue le chapitre six. Les points suivants constituent un apport original de cette thèse:  schéma d'intégration temporelle de base pour ces lois couplées,  apport des sous-intervalles d'intégration,  étude de la matrice tangente couplée d'un élément isoparamétrique à 8 nuds,  amélioration du schéma d'intégration temporelle de base pour assurer la stabilité en cas de transformation martensitique,  développement de la notion d'état plan généralisé pour l'étude du refroidissement d'une poutrelle métallique. Le chapitre sept présente le problème du remaillage. Après une description succincte des solutions proposées par la littérature, nous justifions notre choix d'un remaillage avec maillage indépendant. Ce dernier semble en effet le mieux adapté aux problèmes de forgeage et d'estampage que nous désirons résoudre. Le chapitre huit est consacré à la détermination de l'instant opportun pour réaliser un remaillage. Nous distinguerons deux types de critères: a. les critères avec estimation de l'erreur, développés dans le cadre de l'optimisation des maillages d'éléments finis lors d'une analyse linéaire, ou adaptés aux problèmes non linéaires de formage des métaux, b. les critères sans estimation de l'erreur, basés sur des mesures de distorsion des éléments, les valeurs propres, les mesures des interpénétrations de la pièce forgée et de la matrice. Concernant les critères du type a, notre contribution personnelle est la suivante:  étude bibliographique traduisant dans un vocabulaire d'ingénieur, les développements mathématiques souvent présents lors des études d'estimateurs d'erreur,  adaptation du critère proposé par ZIENKIEWICZ aux lois élastoviscoplastiques présentes dans LAGAMINE et essai d'une variante basée sur une idée personnelle. Concernant les critères de type b, nous proposons des mesures de distorsions originales pour un élément isoparamétrique plan à 8 nuds et nous présentons une analyse des valeurs propres de la matrice tangente. Pour terminer, nous présentons une comparaison des critères de type a et b programmés dans LAGAMINE. A notre connaissance, une telle comparaison n'avait encore jamais été réalisée. Elle nous permet de vérifier la concordance des deux approches. Les avantages et inconvénients de chacune de ces méthodes sont établis. Ils permettent de réaliser un choix des critères à conserver dans l'optique d'un remaillage automatique. Le chapitre neuf présente succinctement la création du nouveau maillage. Les exigences relatives à un mailleur automatique adapté au problème de remaillage sont précisées. Une première approche fut réalisée par nos soins grâce à un programme interactif REMDATA. Les développements originaux de ce programme sont détaillés en annexe 3. Le chapitre dix traite le transfert des informations de l'ancien maillage au nouveau maillage. A ce niveau, deux types d'informations sont considérés: a. les informations relatives aux éléments de solide, b. les informations relatives aux éléments modélisant le contact Le transfert des informations de type a est une opération souvent traitée par la littérature. Après une revue bibliographique des méthodes proposées, nous présentons une méthode originale et personnelle de transfert. Ses qualités principales sont sa précision, sa facilité de mise en uvre et sa souplesse d'utilisation. Une comparaison de cette méthode avec deux autres méthodes basées respectivement sur les moindres carrés et sur l'extrapolation nodale des valeurs aux points d'intégration est présentée. Concernant les informations de type b, nous n'avons trouvé aucune référence qui traitait leur transfert d'un maillage à l'autre. Ce point est cependant important, comme le montrent des essais de remaillage réalisés sans interpolation de ces variables. Les deux méthodes de transfert présentées dans cette thèse montrent l'évolution de notre approche de ce problème. Toutes deux constituent des apports personnels et originaux. Le chapitre onze présente cinq exemples d'application des développements réalisés: 1. l'analyse thermique métallurgique d'une poutrelle métallique refroidie par arrosage et refroidissement naturel, 2. l'analyse thermique métallurgique mécanique d'une poutrelle métallique refroidie par arrosage et refroidissement naturel, 3. l'analyse thermique métallurgique mécanique de la trempe d'un cylindre en acier 60NCD11, 4. la simulation de l'estampage d'une aube de turbine (analyse mécanique avec remaillage), 5. la simulation de l'estampage d'un pignon de machine (analyse mécanique avec remaillage). Hormis l'application 3 qui constitue l'application du code à un cas test déjà traité par DENIS et SJOSTROM (CP31), ces exemples constituent des contributions originales. Quelques conclusions sont proposées au chapitre douze. Parmi nos contributions originales, certaines nous semblent d'une ampleur et d'un intérêt plus grands :  la loi élastoplastique thermique métallurgique et son schéma d'intégration prennent en compte les couplages existant entre les différents phénomènes. Le nombre de codes réalisant une telle analyse est assez réduit alors que des calculs réalistes de contraintes résiduelles de trempe, de soudure ne peuvent être obtenus par des analyses découplées. De plus, l'adaptation des modèles développés à d'autres traitements thermiques (revenu n.) constitue de nouvelles perspectives de recherches pour l'équipe de M.S.M.,  une présentation des estimateurs d'erreur disponibles dans le cadre de la méthode des éléments finis est une information précieuse pour les ingénieurs numériciens comme pour les industriels qui ont recours à ces calculs,  les différents avantages et inconvénients des critères de remaillage basés respectivement sur les estimateurs d'eITeur et les mesures des distorsions des éléments constituent un élément important pour ceux qui doivent faire un choix entre ces deux types de critères,  l'importance du transfert des informations relatives au contact a été mis en évidence, alors que ce sujet n'est pas abordé dans la littérature.

The main aim of this work was to identify local adaptation processes in amphibian populations, thereby improving the general understanding of genetics and mechanisms behind the evolution and maintenance of biological diversity. Phenotypic and genetic variation in life-history traits was studied within and between populations common frog (Rana temporaria) populations along a 1600 km transect from southern Sweden to northern Finland. Embryonic and larval development and growth was investigated both under field and laboratory conditions. The results suggest ample genetic diversity in larval life-history traits among Fennoscandian common frog populations. Larval developmental rate along the gradient has evolved a countergradient variation pattern of genotypes and phenotypes as indicated by the positive relationship between developmental rate and latitude under laboratory conditions and the lack of such a relationship in the field. The data suggest that this pattern has evolved because of time constraints due to decreasing length of growth season with latitude. Neither field-caught adults nor laboratory raised larvae displayed a linear latitudinal size cline as expected from the so called Bergmanns rule. Rather, size increased towards the mid-latitude populations and decreased thereafter, indicating that body size is a product of direct environmental induction or a trade-off with other life-history characters. Age and size at hatching showed no consistent latitudinal pattern, indicating that the embryonic stage is not as time constrained as the larval stage. A large part of the variation in age and size at metamorphosis among populations was due to additive genetic effects. However, small, but significant maternal effects, mostly due to variation in egg size and non-additive genetic effects also contributed to among population variation. A comparison of divergence in presumably neutral molecular genetic markers (FST) and quantitative characters (QST) revealed that although both estimates of divergence were relatively high, estimates of QST was generally higher than those of FST, indicating that the genetic variation observed in larval traits is primarily a result of natural selection rather than genetic drift. Hence, our results reinforce the conclusion that intraspecific genetic heterogeneity in the young northern European ecosystems may be more widespread than previously anticipated