The genus Erica has undergone an extreme radiation in the Cape and exhibits a diversity of pollination syndromes and floral traits. This makes Erica well-suited to study the evolution of floral traits and how they impact speciation. The first chapter explored the role of ultraviolet colouration by recording its prevalence across Erica pollination syndromes. Ultraviolet was rare in wind-, rodent and small insect-pollinated species, but it was common in bird-pollinated species and ubiquitous in long-proboscid fly (LPF)-pollinated species. Testing their preference revealed that sunbirds can see ultraviolet, but they have no innate preference. LPFs on the other hand were not attracted to flowers where ultraviolet reflectance was removed, thus displaying a strong preference. Chapter 2 focused on the role of stickiness for nectar robbers. I experimentally added stickiness to Erica flowers of one species and further compared stickiness to nectar robbing across several communities. Stickiness appears to reduce damage due to nectar robbing within and between species. Further, I found that stickiness is strongly correlated with pollination by birds and LPFs which might be due to their large nectar rewards. Chapter 3 investigated how the sister species Erica shannonea and Erica ampullacea co-occur despite sharing a pollination syndrome. Pollination experiments and observations showed that they are pollinated by LPFs from two families. The horizontal flowers of E. shannonea are pollinated by a tabanid which has a fixed forward-pointing proboscis, while the vertical flowers of E. ampullacea are pollinated by a nemestrinid which can swivel its proboscis downwards. The nemestrinid in turn has a shorter proboscis which prevents it from accessing nectar in the long-tubed E. shannonea. Due to their different biomechanics, each fly can only access the flower it pollinates resulting in effective reproductive isolation between these species. Chapter 4 compared flower orientation in relation to the two LPF families across all LPF-pollinated species. Using a phylogenetically corrected analysis, I found that flowers pollinated by Tabanidae tend to be horizontal, while nemestrinid flowers are more variable in orientation and more often vertical. This confirms the importance of pollinator biomechanics for the evolution of floral traits. The last chapter investigated how pollen transfer efficiency differs between Erica pollination syndromes. I found that LPF- and bird-pollinated species have higher pollen transfer efficiency in comparison to bee-pollinated species which might have facilitated the shifts from ancestral bee pollination.

No description available.

Cette thèse aborde l’écologie et l’évolution des relations qu’entretiennent les orchidées du genre Ophrys avec leurs pollinisateurs. L’approche comparative et la combinaison (i) d’analyses chimiques de phéromones sexuelles et de parfums floraux, (ii) d’analyses génétiques avec des outils moléculaires, et (ii) de tests de comportement réalisés sur les insectes in situ nous ont permis d’éclairer certains aspects méconnus de ces interactions inter-spécifiques. La pollinisation des orchidées du genre Ophrys est assurée par des mâles d'abeilles ou de guêpes solitaires qui opèrent une tentative d’accouplement (pseudocopulation) sur le labelle des fleurs. L'attraction des pollinisateurs est généralement hautement spécifique, régie par un mimétisme des signaux (chimiques, visuels, tactiles) des femelles des espèces d'insectes concernés. Malgré cette spécificité, des hybrides se forment occasionnellement en conditions naturelles, témoignant de la perméabilité partielle des barrières d'isolement reproductif entre espèces. Au cours de ce programme de recherche, nous avons entrepris l’étude des interactions Ophryspollinisateurs en mettant l’accent sur trois aspects spécifiques, à savoir (i) la sélection des caractères floraux par les pollinisateurs, (ii) l'isolement reproductif entre espèces d'Ophrys sympatriques, et enfin (iii) l'évolution des caractères floraux au sein d'un complexe d'espèces-soeurs d'Ophrys associées à différents pollinisateurs. Les principaux résultats de ce travail sont repris ci-dessous, ponctués de références aux articles qui rassemblent l’intégralité des études réalisées.

A fundamental question in biology is how animals efficiently locate and use diverse resources. Pollinators foraging on flowers are one of our most thoroughly studied examples of generalist foraging behavior and cognition. Individual pollinators typically specialize on a subset of flowering species available to them. Specialization by nectar-foraging pollinators is often the consequence of learned or innate preferences for floral display traits such as color, pattern, and scent. Pollinators must also typically learn to extract nectar from each floral type. By specializing, pollinators reduce costs associated with learning and forgetting nectar extraction routines. Specialization also benefits the plant by enhancing conspecific pollen transfer. Yet nectar is not the only floral reward. The pollen of hundreds of thousands of plant species is collected by pollinators such as bees, beetles, and flies. In fact, solitary and social bees must collect both pollen and nectar to survive. However, much of the vast literature on bee foraging behavior concerns the collection of nectar. This research investigated mechanisms by which generalist bumblebees (Bombus impatiens) specialize on diverse floral resources. Most foragers in a colony were reward generalists over their lifetime, but specialized daily on either pollen or nectar collection. Lifetime patterns of pollen collection were associated with interindividual differences in sensory morphology. Pollen-foraging bumblebees had weak innate preferences, but learned strong preferences for pollen-only plant species, with preferences mediated primarily by anther properties. The anthers provided indirect cues of concealed pollen, and bees learned to prefer properties of the anthers to select potentially rewarding flowers. While learning was involved in the formation of floral preferences by pollen foragers, pollen extraction behavior relied little on learning. Specifically, floral sonication, which is used by bees to extract concealed pollen, was modified only modestly with experience. Furthermore, bees foraged efficiently for pollen from diverse floral resources without relying on instrumental (associative) learning. Efficient foraging involved switching between two distinct motor routines: floral sonication and scrabbling. Switching was regulated by two ubiquitous floral cues: chemical anther cues eliciting sonication and mechanical pollen cues suppressing it (and eliciting scrabbling). I discuss how mechanisms of floral specialization by generalist pollen-foraging bees could drive floral trait evolution.

Nesse estudo, explorei aspectos da ecologia evolutiva da interação planta-polinizador, ao avaliar o fenótipo floral sob pressões seletivas geograficamente divergentes. Para isso utilizei a interação entre abelhas coletoras de óleo e Krameria (Krameriaceae) como sistema modelo. A tese abrange a história natural do sistema e manipulações experimentais in situ que deram suporte à investigação sobre integração floral no contexto multipopulacional. No primeiro capítulo, estudei a biologia da polinização de Krameria tomentosa, listando as espécies de visitantes florais associadas a suas flores. Assim, mostrei que essa espécie depende de seus polinizadores para o sucesso na polinização e que a maioria das abelhas coletoras de óleo associada a suas flores pertence ao gênero Centris (Centridini). No entanto, observei uma alta frequência de abelhas coletoras de pólen nessa espécie que parecem contribuir com a manutenção da polinização na população. No segundo capítulo, mostrei que há uma variação geográfica na assembleia de polinizadores de K. grandiflora que inclui diferentes comportamentos e ajustes à morfologia floral da espécie. Desse modo, testei a significância funcional das três estruturas especializadas das flores de Krameria (cálice conspícuo, pétalas petaloides e elaióforos) em dois cenários contrastantes de ajuste entre a arquitetura floral e o polinizador. No entanto, diferenças não foram detectadas. O ajuste mecânico que os elaióforos provêm parece essencial para o sucesso na polinização no grupo. No terceiro capítulo, demonstrei que populações similares no padrão de correlação de seus caracteres florais, apresentam composição e diversidade funcional de polinizadores semelhantes, sugerindo que diferenças na morfologia combinadas à abundância de polinizadores influenciam o padrão de integração fenotípica. Portanto, esses resultados mostram uma situação em que a variação geográfica na assembleia de polinizadores parece promover a diferenciação da estrutura correlacional do fenótipo floral / In this study, I investigate some aspects of the evolutionary ecology of a plant-pollinator interaction, by evaluating the floral phenotype under geographically divergent selective forces. For this, I used the interaction between oil-collecting bees and Krameria (Krameriaceae) as a system model. This thesis comprises the natural history of the studied system and manipulative experiments in situ to support the investigation on the phenotypic integration in a multipopulation context. In the first chapter, I studied the pollination biology of Krameria tomentosa, listing the flower visitor species associated to its flowers. Thus, I showed that this species depends on their pollinator for pollination success and that most oil-collecting bees recorded belongs to Centris genus (Centridini). However, I observed a high frequency of pollen-collecting bees on this species which might contribute to pollination maintenance in the population. In the second chapter, I showed that there is a geographic variation in the pollinator assemblage of K. grandiflora which includes different behavior and matching to Krameria floral architecture. Thus, I tested the functional significance of the three specialized structures of Krameria flowers (showy calyx, petaloid petals and elaiophores) in two contrasting pollinator-matching scenarios; however, no differences were detected. The mechanical fit provided by the elaiophores might be a keystone for the pollination success in Krameria. O ajuste mecânico que os elaióforos provêm parece essencial para o sucesso na polinização no grupo. In the third chapter, I demonstrated that populations with similar patterns of floral trait correlation comprise similar pollinator composition and functional diversity, suggesting that differences in pollinator morphology combined to abundance influences the pattern of phenotypic integration. Therefore, these results show a condition in which geographic variation in pollinator assemblage might provide divergence in the phenotypic correlation structure

Myrtaceae possui cerca de 4.600-5.800 espécies, distribuídas em áreas tropicais e subtropicais. Myrteae é uma das tribos mais ricas da família, com ca. de 2.500 espécies. Estudos sobre desenvolvimento e evolução de flor em Myrteae têm mostrado a condição homoplástica de caracteres anteriormente considerados diagnósticos, culminando em um melhor entendimento taxonômico. O hiper-diverso gênero Eugenia, com ca. de 1.050 espécies, está incluído nessa tribo e tem se mostrado morfologicamente homogêneo. Contudo, gêneros previamente segregados de Eugenia pelo grau de fusão do cálice são fontes de controvérsia taxonômica. A filogenia molecular baseada em cinco regiões de DNA reconstruiu que o cálice fundido evoluiu independentemente diversas vezes em Eugenia. A investigação morfológica da fusão do cálice revelou cinco padrões de desenvolvimento, além da condição dos lobos livres. Embora os padrões de desenvolvimento sejam homoplásticos, o sinal filogenético indicou que eles podem ser utilizados para complementar caracteres e sustentar clados. O padrão membranissépalo é uma exceção, sendo exclusivo e diagnóstico de um clado Amazônico. Espécies tradicionalmente reconhecidas com seis pétalas são interpretadas como padrão petaloide, isto é, quatro pétalas são precedidas por duas sépalas internas e petaloides, e duas sépalas externas e fundidas entre si. Padrão longohipanto é reconhecido como uma condição extremamente rara em Eugenia, em que o hipanto se alonga estende, bem como os verticilos estaminais, resultando em estames curvos no botão em vez de eretos. Mudanças nomenclaturais e uma recircunscrição e revisão taxonômica de Eugenia sect. Schizocalomyrtus derivam destes resultados. Dez espécies são descritas e comentadas em suas distribuições, sua conservação e taxonomia. A inflorescência de Eugenia também é analisada sob uma abordagem integrativa, que revela arranjos previamente ignorados como relevantes para a sistemática e taxonomia. Sete padrões foram descritos, e cinco deles são considerados para uma análise evolutiva. O padrão auxotélico foi recuperado em linhagens que primeiro divergiram. Isso sugere que o seu ancestral hipotético é similar ao padrão auxotélico. O padrão fasciculiforme é amplamente distribuído na seção de maior riqueza, Eugenia sect. Umbellatae, indicando que altas taxas de diversificação podem estar relacionadas com a aquisição desse padrão. Este estudo também fornece novas hipóteses para a evolução dos arranjos da inflorescência, indicando que flexibilidade no padrão racemoso de ramificação é, provavelmente, a inovação chave que promoveu a diversificação de Eugenia na região Neotropical / Myrtaceae has about 4600-5800 species distributed in tropical and subtropical areas. Myrteae is one of the most species rich tribe in Myrtaceae with ca. 2500 species. Studies on Myrteae flower development and evolution have shown the homoplastic nature of characters previously considered diagnostic that are now culminating in much-improved taxonomic understanding. The hyper-diverse Eugenia with ca. 1050 species is nested in this tribe, and have been shown as morphologically homogeneous. However, genera previously segregated from Eugenia by the degree of calyx fusion are source of controversy. Molecular phylogeny based on five DNA regions reconstructed fused calyx as evolved several times independently in Eugenia. Morphological assessment of the calyx fusion revealed five development patterns besides the standard condition of the free lobes. Although the development patterns are homoplastic, phylogenetic signal indicates that they can be used in combination to complementary characters to support clades. Membranisepalous pattern is an exception, recovered as exclusive allowing to be used to diagnose an Amazonian clade. Traditionally six-petal species are newly interpreted as petaloid pattern, i.e. four petals are followed by two internal petal-like sepals, and two external fused sepals. Longohypanthium pattern is recognized as an extremely rare condition in Eugenia which hypanthium extends as the stamens whorl resulting in a display of curved stamens in the bud instead the standard straight. Nomenclatural changings and a taxonomic revision of Eugenia sect. Schizocalomyrtus newly circumscribed follow these results. Ten species are detailed described and commented under distribution, conservation, and taxonomy. The inflorescence of Eugenia was also analysed under an integrative approach that revealed arrangements previously ignored as relevant for systematic and taxonomy. Seven patterns were described but five are regarded to an evolutionary assessment. Auxotelic pattern was recovered in the early lineages. It suggests that a hypothetical ancestral is similar to the auxotelic pattern. Fasciculiform pattern is widely found in the most speciose Eugenia sect. Umbellatae, indicating that high rates of diversification may be related to the acquisition of this pattern. This study also provides insights in the evolution of inflorescence arrangement by indicating that flexibility in the racemose branching pattern is likely the innovative key that promoted the diversification of Eugenia in the Neotropics

Les relations de parentés profondes au sein des Angiospermes ont été longtemps incertaines. A la fin des années 90, les études phylogénétiques à grande échelle ont contribué à l’obtention de l’arbre actuel des Plantes à fleurs, dans lequel Eudicotylédones, Monocotylédones et Magnoliidae forment les trois plus grands clades. Contrairement aux Monocotylédones et aux Eudicotylédones, la monophylie des Magnoliidae (Canellales, Laurales, Magnoliales et Piperales) n’a été soutenue que plus récemment. Les Magnoliidae contiennent actuellement 20 familles et environ 10 000 espèces majoritairement présentes sous les tropiques. Avant cette thèse, de nombreuses parties de ce groupe avaient été étudiées en détail mais l’histoire évolutive du groupe dans son ensemble était encore mal connue. Le premier chapitre est une étude phylogénétique des relations entre les familles et les ordres de Magnoliidae. Pour réaliser cette étude, j’ai échantillonné 199 espèces du groupe et 12 marqueurs moléculaires issus des trois génomes. J’ai ensuite mené des analyses phylogénétiques avec les méthodes de parcimonie, d’inférence bayésienne et de maximum de vraisemblance. Les résultats confirment avec un plus fort soutien la présence de deux clades dans ce groupe : Canellales + Piperales et Laurales + Magnoliales. De plus, les relations entre les 20 familles sont généralement bien soutenues, les Lactoridaceae et les Hydnoraceae étant incluses dans les Aristolochiaceae (Piperales). Dans le second chapitre, j’ai révisé l’âge et la position de 10 fossiles identifiés comme appartenant aux Magnoliidae. Le but de cette étude était de fournir de nouveaux points de calibration fiables afin de conduire de nouvelles analyses de datation moléculaire. Parmi les nombreux fossiles du groupe, nous avons choisi ces espèces car elles avaient été placées phylogénétiquement par des études antérieures. Le schéma de calibration résultant de ce travail inclut six contraintes fiables d’âges minimum. Le troisième chapitre est une étude de datation moléculaire utilisant ce schéma de calibration et le même jeu de données moléculaires que le chapitre 1. Les résultats tendent à repousser l’âge des Magnoliidae (127.1-198.9 Ma), et des quatre ordres Canellales (126.3-141.0 Ma), Piperales (88.2-157.7 Ma), Laurales (111.8-165.6 Ma) et Magnoliales (115.0-164.2 Ma). Dans ce même chapitre, j’ai également étudié le mode de diversification du groupe. Les variations importantes du nombre d’espèces entre les différentes parties de l’arbre s’expliquent le mieux par des modèles de diversification incluant 6 à 14 transition du taux net de diversification. Enfin, dans le dernier chapitre de la thèse, j’ai retracé l’histoire évolutive de 26 caractères floraux pour reconstruire les fleurs ancestrales de nœuds-clés des Magnoliidae. Pour ce faire, j’ai tiré parti de la phylogénie du premier chapitre et utilisé les mêmes espèces dans ma matrice morphologique. Les résultats montrent que l’ancêtre commun le plus récent des Magnoliidae présentait des fleurs bisexuées et actinomorphes avec un périanthe différencié de deux cycles trimères à tépales libres et probablement trois étamines libres. Ce travail de thèse apporte des résultats importants sur l’évolution des Magnoliidae et soulève de nombreuses questions telles que l’impact des crises géologiques sur la diversification du groupe ou l’influence des pollinisateurs et de l’environnement sur l’évolution de la morphologie florale. / Deep phylogenetic relationships in the angiosperms had long been uncertain. However, by the end of the 1990s, large-scale studies contributed to the current well resolved picture of the tree of flowering plants, in which eudicots, monocots, and magnoliids are the three largest clades. Whereas monocots and eudicots have been recognized since the very first phylogenetic analyses, the monophyly of magnoliids (Canellales, Laurales, Magnoliales, and Piperales) is a more recent result. Magnoliidae, as now circumscribed, consist of 20 families and ca. 10,000 species mostly distributed in the tropics. Before the present thesis, several parts of the magnoliid tree had been well studied, but little was known about the evolutionary history of Magnoliidae as a whole. The first chapter of this thesis is a phylogenetic study conducted to clarify the relationships among families and orders of Magnoliidae. To do so, I sampled 199 species of Magnoliidae and 12 molecular markers from the three genomes and conducted phylogenetic analyses using parsimony, maximum likelihood, and Bayesian methods. The results confirm, with a greater level of support, two clades in Magnoliidae: Canellale + Piperales, and Laurales + Magnoliales. In addition, the relationships among the 20 families are generally well supported, and Lactoridaceae and Hydnoraceae are nested within Aristolochiaceae (Piperales). In the second chapter, the ages and phylogenetic positions of 10 fossils attributed to Magnoliidae were reviewed in detail. The goal of this study was to provide new reliable calibration points in order to conduct molecular dating analyses. These fossils were selected from the rich fossil record of the group because of their previous inclusion in phylogenetic analyses with extant taxa. The resulting calibration scheme provides six solid, internal minimum age constraints. The third chapter includes molecular dating analyses using the present calibration scheme and the same molecular dataset of Chapter 1. This study tends to push back in time the ages of the crown nodes of Magnoliidae (127.1-198.9 Ma), and of the four orders, Canellales (126.3-141.0 Ma), Piperales (88.2-157.7 Ma), Laurales (111.8-165.6 Ma), and Magnoliales (115.0-164.2 Ma). In the same chapter, I investigated the mode of diversification in the group. The strongly imbalanced distribution of species appears to be best explained by models of diversification with 6 to 14 diversification rate shifts. Finally, in the last chapter, I traced the evolution of 26 floral characters to reconstruct the ancestral flowers in key nodes of Magnoliidae. I used the phylogeny of Chapter 1 and an exemplar approach. Our results show that the most recent common ancestor of all Magnoliidae was a tree bearing actinomorphic, bisexual flowers with a differentiated perianth of two alternate, trimerous whorls of free perianth parts (outer and inner tepals) and probably three free stamens. This work provides key results on the evolution of Magnoliidae and raises several new questions such as the impact of geological crises on diversification of the group or the influence of pollinators and the environment on the evolution of floral morphology.

Deep phylogenetic relationships in the angiosperms had long been uncertain. However, by the end of the 1990s, large-scale studies contributed to the current well resolved picture of the tree of flowering plants, in which eudicots, monocots, and magnoliids are the three largest clades. Whereas monocots and eudicots have been recognized since the very first phylogenetic analyses, the monophyly of magnoliids (Canellales, Laurales, Magnoliales, and Piperales) is a more recent result. Magnoliidae, as now circumscribed, consist of 20 families and ca. 10,000 species mostly distributed in the tropics. Before the present thesis, several parts of the magnoliid tree had been well studied, but little was known about the evolutionary history of Magnoliidae as a whole. The first chapter of this thesis is a phylogenetic study conducted to clarify the relationships among families and orders of Magnoliidae. To do so, I sampled 199 species of Magnoliidae and 12 molecular markers from the three genomes and conducted phylogenetic analyses using parsimony, maximum likelihood, and Bayesian methods. The results confirm, with a greater level of support, two clades in Magnoliidae: Canellale + Piperales, and Laurales + Magnoliales. In addition, the relationships among the 20 families are generally well supported, and Lactoridaceae and Hydnoraceae are nested within Aristolochiaceae (Piperales). In the second chapter, the ages and phylogenetic positions of 10 fossils attributed to Magnoliidae were reviewed in detail. The goal of this study was to provide new reliable calibration points in order to conduct molecular dating analyses. These fossils were selected from the rich fossil record of the group because of their previous inclusion in phylogenetic analyses with extant taxa. The resulting calibration scheme provides six solid, internal minimum age constraints. The third chapter includes molecular dating analyses using the present calibration scheme and the same molecular dataset of Chapter 1. This study tends to push back in time the ages of the crown nodes of Magnoliidae (127.1-198.9 Ma), and of the four orders, Canellales (126.3-141.0 Ma), Piperales (88.2-157.7 Ma), Laurales (111.8-165.6 Ma), and Magnoliales (115.0-164.2 Ma). In the same chapter, I investigated the mode of diversification in the group. The strongly imbalanced distribution of species appears to be best explained by models of diversification with 6 to 14 diversification rate shifts. Finally, in the last chapter, I traced the evolution of 26 floral characters to reconstruct the ancestral flowers in key nodes of Magnoliidae. I used the phylogeny of Chapter 1 and an exemplar approach. Our results show that the most recent common ancestor of all Magnoliidae was a tree bearing actinomorphic, bisexual flowers with a differentiated perianth of two alternate, trimerous whorls of free perianth parts (outer and inner tepals) and probably three free stamens. This work provides key results on the evolution of Magnoliidae and raises several new questions such as the impact of geological crises on diversification of the group or the influence of pollinators and the environment on the evolution of floral morphology.

Les Ranunculales sont un ordre d’angiospermes d’environ 4500 espèces, incluant des plantes communes comme les boutons d’or et les coquelicots. Leurs fleurs sont très diversifiées et ont piqué l’intérêt des botanistes depuis des décennies. Les sept familles de l’ordre sont faciles à reconnaître sur le terrain, mais paradoxalement, certaines familles n’ont pas de synapomorphies et ont des états ancestraux peu clairs, comme les Ranunculaceae (~ 2500 espèces). Cette diversité florale peut être liée à la pollinisation et des innovations ont pu être guidées par cette interaction. De plus, les plantes de ce groupe produisent une grande variété de composés secondaires, certains dont les propriétés sont connues depuis fort longtemps (comme la morphine et le curare). Ces substances pourraient jouer un rôle dans la défense contre les herbivores. Les caractères floraux et chimiques sont donc de bons candidats pour produire des hypothèses adaptatives. Ici nous utilisons des méthodes analytiques modernes pour comprendre les patrons d’évolution expliquant la distribution actuelle de la diversité et l’évolution des fleurs et des toxines de Ranunculales. Un nouveau cadre phylogénétique synthétique a été produit avec des séquences publiées de 144 espèces. Seize caractères floraux et la présence de certains composés métaboliques ont été reconstruits sur cet arbre avec la méthode de parcimonie et la méthode de « Reversible Jump Monte Carlo Markov Chains ». Un nouveau scénario d’évolution florale a été produit, où les Ranunculales avaient une fleur ancestrale trimère avec trois cycles de tépales et où les ancêtres des familles ont évolué en perdant ou en différenciant des cycles du périanthe. Les patrons d’évolution des toxines montrent de la variabilité, certaines apparaissant et disparaissant aléatoirement (comme les saponines), d’autres étant des synapomorphies bien soutenues de certains clades (comme les diterpènes de type aconitine pour la tribu des Delphinieae). Ces résultats offrent une meilleure compréhension de l’histoire naturelle des Ranunculales. En outre, cet ordre est le groupe-frère de toutes les autres eudicotylédones, et a donc une position clef pour comprendre l’évolution précoce de ce clade. Ces résultats vont fournir une compréhension plus profonde des changements floraux et phytochimiques qui ont eu lieu à la base des eudicotylédones, qui contiennent 70% des angiospermes actuelles. / Ranunculales are an order of angiosperms comprising ca. 4,500 species including common plants like buttercups and poppies. Their flowers are highly diversified and have raised the interest of botanists for decades. Each of the seven families of the order is easily recognizable in the field, but paradoxically some families lack floral synapomorphies and have unclear ancestral states, like Ranunculaceae (ca. 2,500 species). This floral diversity may be linked to pollination and innovations may have been driven by this interaction. In addition, the plants of this group produce a variety of secondary compounds, some of which having been known for ages for their properties, like morphine or curare. Those substances could play a role in defense against herbivores. Both floral and chemical characters are thus good candidates to produce adaptive hypotheses. Here we used modern analytical methods to understand the the evolutionary patterns accounting for the extant distribution of diversity and evolution of Ranunculales flowers and toxins. A new synthetic phylogenetic framework of the Ranunculales based on already available sequences of 144 species was produced. Sixteen characters of the flower, and the presence of selected secondary metabolites were reconstructed on this tree using parsimony, and Reversible Jump Monte Carlo Markov Chains. A new scenario for floral evolution was produced, where Ranunculales had a trimerous perianth with three whorls of tepals ancestor, and the ancestors of the families evolved by losing and/or differentiating perianth whorls. The evolutionary patterns of toxins showed variability, some appearing and disappearing randomly (such as saponins), others being well-supported synapomorphies (such as aconitine-like diterpens for the tribe Delphinieae). These results offer a better understanding of the natural history of Ranunculales. Furthermore, this order is the sister-group of all other eudicots, and thus has a key position to understand the early evolution of this clade. These results will provide a deeper understanding of the floral and chemical changes that took place at the base of eudicots, which contains 70% of living angiosperms.

There is considerable evidence that pollinator shifts can explain many differences in flower morphology between closely related plant species, but the extent to which pollinator shifts can explain the maintenance of among-population variation in floral traits within species is poorly known. In this thesis, I combined comparative and experimental approaches to examine the evolution of floral traits in the moth-pollinated orchid Platanthera bifolia. More specifically, I investigated (1) the relationship between flower and pollinator morphology, (2) phenotypic selection on morphology and phenology in populations in contrasting environments, (3) components of prezygotic reproductive isolation among divergent populations, and (4) the adaptive and functional significance of two correlated floral traits. A study of Scandinavian of P. bifolia populations revealed that spur length was positively correlated with proboscis length of local pollinators, which suggests that variation in spur length reflects adaptive evolution in response to geographically variable pollinator-mediated selection. A phenotypic selection study on Öland, SE Sweden, suggested that disruptive selection on spur length contributes to the maintenance of a bimodal distribution of spur length in mixed habitats, but provided very limited evidence of divergent selection on plant morphology and flowering phenology in grassland and woodland habitats. Field experiments revealed strong reproductive isolation between divergent populations on Öland, due to differences in spatial distribution, flowering phenology, and pollinators, and among-population incompatibility. The results suggest that prezygotic reproductive isolation contributes to the maintenance of population differentiation in floral traits in P. bifolia. A field manipulation experiment demonstrated that spur length but not perianth size affects pollination success and seed production. This suggests that among-population differentiation in perianth size may be the result of a genetic correlation with spur length. Taken together, the results of this thesis suggest that pollinator-mediated selection can shape the evolution of intraspecific floral variation.