Résumé
In ecology, the neutral theory has been brought forth to explain the biological diversity observed in communities. Assuming that all species are ecologically equivalent, this theory explains diversity patterns as the outcome of stochasticity, migration and speciation. The neutral theory constitutes a falsifiable « null model » to test more complex hypotheses on community assembly rules, with two caveats though: (i) neutral theory predicts species diversity in communities with a very good fit, so that only considering species diversity is hardly sufficient to test the likelihood of more complex hypotheses; (ii) the few alternative hypotheses that are designed to be compared to neutral models lack the biological and evolutionary background underlying the postulated niches. This thesis brings forth new approaches linking ecology and evolution aimed at overcoming the two aforementioned caveats. First, the genetic diversity of studied species is integrated in the neutral theory to allow testing for niche dimensions based on new patterns, such as speciesgenetic diversity correlations (SGDCs). The development of a neutral model unifying both genetic and species levels allows for more precise predictions on these correlations. A more general statistical approach of these correlations is also proposed, with a more descriptive statistical decomposition of SGDCs along environmental gradients to better grasp the underlying forces, based on the case of a freshwater snail metacommunity from the Lesser Antilles. Second, alternative hypotheses (i.e. “quasi-neutral” models) that allow for individual variation along a few niche axes are developed and explored them in two different ways. The first approach explores the evolutionary processes underlying ecological differences among species and is based on the analysis of the evolution of dispersal in a metacommunity and its consequences for diversity patterns. The second approach is a descriptive one which seeks to identify the more differentiated species in ecological datasets, which I illustrate through a study of the response of Garrigue plant communities to thyme. Overall, this work puts forward the potential of a theory “presuming neutrality” of community ecology that integrates genes, traits and species as well as the evolutionary constraints tying them together.