Inferring reciprocal evolutionary histories in associated species of plants and insects in two european pollination systems

Coevolution is defined as reciprocal evolutionary changes that might arise at any spatiotemporal scale. Despite every organism on Earth undergoes coevolutionary interactions, cases of one-to-one specific relationships are generally rare. However, because of the reduced number of interacting species they concern, these species-specific associations are interesting to evolutionary biologists because they allow testing hypotheses in simple frameworks. Despite the history and evolution of coevolutionary interactions have been studied in several cases in the last decade, this topic remains difficult to fully circumscribe because of the multiplicity of factors that affect one or the other species concerned. Moreover, a lot is known about coevolution at a small scale, but little has been done at larger and more integrative scales spanning wider spatiotemporal ranges. Phylogeography is a young area of biology that allows understanding the distribution of lineages in space and time. Despite that the idea of parallely studying the history of species involved in specific interactions appears simple, this has rarely been done until now probably because of the technical efforts this would represent. From a theoretical point of view, we could propose that in specific interactions, because of the dependence between the partners involved, we should observe some phylogeographic pattern associated to the type of interaction studied. In this way, while partners of mutualistic interactions should present similar postglacial histories, this should not be true for those associated by antagonistic relationships. In this thesis, we exploit different techniques and approaches to test this general hypothesis. The final aim of this study is thus to understand if it is possible to identify a pattern of comparative phylogeography in relation to the type of interaction, using as case-studies two specific and obligate European interactions: the antagonistic relationship established between <i>Arum maculatum</i> L. (Araceae) and its Psychodid (Diptera) pollinating flies, and the nursery pollination mutualism involving <i>Trollius europaeus</i> L. (Ranunculaceae) and the <i>Chiastocheta</i> (Diptera: Anthomyiidae) species complex. Before testing our comparative phylogeographic hypotheses and because studying the phylogeography of interactions requires a wide knowledge of the environmental, taxonomic and historical frameworks in which these ecological relationships arose, it was first needed to clearly delimitate the identity of species, their distribution and the environmental factors influencing their survival to finally understand their comparative history. We thus took advantage of the potentialities that interdisciplinary approaches provide, applying molecular taxonomy, biological and evolutionary methods, biogeographic inferences, ecological niche models and hindcasting techniques, as well as classical and recently-developed phylogeographic analyses. Our results indicate that the phylogeographic patterns of these specific and obligate antagonistic and mutualistic relationships appear to be related to the type of interaction. Antagonistic partners presented incongruent phylogeographic patterns, what can be notably explained by differences in their life-history traits. Species involved in mutualistic interactions partly showed congruent phylogeographic patterns (particularly in the cases of <i>T. europaeus</i> and <i>C. dentifera</i>). Flies interacting with <i>T. europaeus</i> appear moreover to present different histories, regardless of their important ecological similarities. These results demonstrate that the systems studied appear to be far more complex than initially supposed, with crossed effects of environmental and historical features on the dynamics of the interaction. Because of the high complexity and interdependency of factors affecting one or the other partner, performing investigations in an interdisciplinary framework appears indispensable to disentangle the dynamics of interactions.

Thèse de doctorat : Université de Neuchâtel, 2010