Abstract
This thesis is concerned with the distribution of genetic variation within subdivided populations. Studying the rare aquatic fern Marsilea strigosa Willd. (Marsileaceae, Pteridophyta), I show how the concepts of genetic identity and effective size are crucial to understand which evolutionary forces shape the observed genetic variation. In particular, I developed theoretical models to show the influence of spatial and temporal structure on the distribution of genetic variation. From these models, three methods have been developed to analyse the genetic variation in natural populations. The first method enables to estimate the effective size, which determines the intensity of genetic drift. The second one concerns the estimation of sex-biased dispersal rates. Finally, a method has been developed to identify molecular markers that are potentially responding to selection.