Many organisms are sexually dimorphic for ecologically and socially important traits. One of the major foci of biology is to understand the evolution of such sexually dimorphic traits. Here I present my work on the evolution of a dimorphic trait, female-limited Batesian mimicry, in Papilio swallowtail butterflies. I begin by developing a character state path network to study the diversity of mimicry types and directionality of trait change during the evolution of female-limited mimicry. My phylogenetic analysis showed that female-limited mimicry has evolved independently in several groups of swallowtails, mainly via single-step character changes from monomorphic non-mimetic ancestors to female-limited mimetic descendents. Mimetic polymorphism has evolved in tandem with female-limited mimicry, the two being tightly correlated among mimetic species. Most traditional explanations of female-limited mimicry and mimetic polymorphism invoke sexual selection. In reviewing these hypotheses, I show that their key assumptions and predictions remain untested, and that sexual selection cannot maintain female polymorphism under some conditions. Sexual selection hypotheses are also unable to explain community ecological aspects of mimicry rings. Hence, I developed a novel model of female-limited mimicry based on sex-specific, frequency- and density-dependent advantages of mimicry. This model shows that both-sex mimicry, female-limited mimicry and mimetic polymorphism are favored along a gradient of relative mimic frequency. My ecological data from south Indian mimicry rings support a key prediction of this model. Finally, I employ the patterns of female-limited mimicry among swallowtail butterflies to highlight the contrast between Darwin’s sexual selection model and Wallace’s natural selection model of sexual dimorphism. I show that most of the sexual dimorphism in swallowtail wing color patterns is a product of natural selection for protective female coloration, predominantly in the form of female-limited mimicry. Thus, swallowtails support Wallace’s model of sexual dimorphism, underlining the importance of natural selection. / text

Orchids provide a model system for addressing evolutionary and ecological questions both because of their species diversity, and because the packaging of their pollen into pollinia facilitates the estimation of male and female pollination success. This thesis focuses on the ecology and evolution of pollination systems in the African orchid genus Eulophia, with an emphasis on deceptive pollination, mechanisms promoting cross-pollination, and pollinatordriven speciation. Pollination in the deceptive species E. zeyheriana is shown to depend on flower colour and proximity to the rewarding model species, Wahlenbergia cuspidata (Campanulacae). This study demonstrates the functional importance of colour matching between model and mimic in a floral Batesian mimicry system, as well as the importance of facilitation by the rewarding model [chapter 2]. The pollinaria of the vast majority of Eulophia and Acrolophia species undergo reconfiguration following removal by pollinators, similar to the phenomena first described by Darwin in some European orchids and which he hypothesised to be adaptations to limit pollinator mediated self-pollination. In chapter 3, a less common mechanism – anther cap retention – is described for E. foliosa. Observations of reconfiguration times were compared to the respective visit times by pollinators in a number of orchids (including Eulophia and Acrolophia) and asclepiads. In 18 of 19 species, pollinarium reconfiguration times exceed the average visit times, providing empirical support for Darwin’s cross-pollination hypothesis [chapter 4]. All of the 25 species of Eulophia examined are deceptive, but two of the three species in the small, closely related Cape genus Acrolophia examined in chapter 5 are rewarding. This translates into very high levels of pollen transfer efficiency in the rewarding A. cochlearis relative to the deceptive A. capensis and species of Eulophia. In addition, A. cochlearis exhibits high rates of pollinator-mediated self-pollination, as quantified using a novel method based on levels of inbreeding depression during embryo development. In chapter 6 the evolutionary divergence of long- and short-spurred forms of E. parviflora in response to different pollinators is investigated. This shows that divergence has occurred in floral morphology, scent chemistry and flowering phenology and that this can be attributed to adaptations to the respective bee and beetle pollinators of each form. This thesis also includes case histories of bee pollination in an additional five Eulophia species, and beetle-pollination in two other species of Eulophia with dense inflorescences and slow pollinarium reconfiguration [chapter 7]. In addition, four taxa were found to undergo auto-pollination [chapter 8]. The main conclusions of this thesis are that pollination of food-deceptive species can be enhanced by spatial proximity to, and floral colour matching with, sympatric rewarding species; that selection strongly favours traits that promote cross-pollination; that pollinatorshifts can drive speciation; and that floral adaptations for bee-, beetle-, and auto-pollination are found in South African representatives of Eulophia. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2009.

Many Hymenoptera, with their painful stings and noxious chemical defenses, exhibit bright aposematic warning color patterns and are the most frequently mimicked group of organisms. Such aposematic color patterns are found in parasitic wasps of the Neotropical Compsobracon group (Braconidae). Many members of this group exhibit color patterns similar to several thousand other species of Braconidae, Ichneumonidae, sawflies, assassin bugs, flies, moths, and beetles. One hypothesis to explain this observation is that the members of the complex and their colors are generated by multiple cospeciation events resulting in the constituent genera having isomorphic phylogenetic trees. An alternative hypothesis is that the organisms have colonized existing color pattern niches independently and do not have topologically similar phylogenetic histories. In order to test the hypothesis that these patterns are the result of cospeciation events they will be described and mapped onto a phylogenetic tree. If clades are found to have isomorphic topologies; evidence will suggest cospeciation. However, if clades are not found to have similar topologies, evidence will suggest independent colonization of color pattern niches. / Graduation date: 2006

Ce doctorat trouve son origine dans la compréhension des prises de décision et des comportements collectifs des animaux. Comment ces derniers parviennent-ils à effectuer des choix collectivement ?Comment les membres d’un groupe procèdent-ils pour synchroniser leurs comportements spatialement et temporellement ?Mon principal objectif a été de dégager, lors des déplacements collectifs et du fur rubbing chez le capucin moine, les évènements décisionnels dépendants de processus anonymes de ceux dépendants de processus liés à l’identité des individus et à leur réseau de relations sociales au sein du groupe. Dans les prises de décision collective relatives aux déplacements, les membres du groupe sont influencés dans leurs choix par leur identité sociale mais aussi par des mécanismes anonymes, de type mimétique. Le fur rubbing est également un comportement collectif dont les mécanismes sous-jacents incluent une dépendance interindividuelle de type mimétique. Des mécanismes similaires mettant en jeu des interactions entre individus basées sur des règles comportementales simples se retrouvent dans chacun des phénomènes collectifs étudiés. Ces résultats sont les premiers à démontrer l’émergence de prises de décision collective à partir de telles interactions anonymes dans un groupe de primates non humains. Ils permettent de faire le lien entre choix individuels et comportement collectif et de mieux concevoir comment un groupe de primates peut se coordonner, maintenir sa cohésion spatiale et synchroniser ses activités./How do animals reach collective consensus? How do group members spatially and temporally synchronise their behaviour? My main purpose was to demonstrate the respective roles of anonymous processes (contagion, mimetism) and individual-dependent processes (hierarchical rank, age, sex, kin, social relationships) in collective decision-making. During decision-making relating to collective movements, group members’ decisions depend on their social identity (individual-dependent mechanism) as well as anonymous processes. Fur rubbing is also a collective behaviour involving interindividual dependence with mimetic underlying mechanisms. We found similar mechanisms, involving interindividual interactions according to simple behavioural rules, in both collective phenomenon studied. These results are the first to demonstrate the emergence of collective decision-making based on anonymous interactions in a group of non human primates. They help to understand the link between individual choices and collective behaviour and to appreciate how a social group of primates maintain its spatial cohesion and synchronize its activities. / Doctorat en sciences, Spécialisation biologie animale / info:eu-repo/semantics/nonPublished