LAUSR

The dispersal process is crucial in determining the fate of populations over time, but habitat fragmentation limits or prevents it. Landscape genetic is an effective tool to assess the degree to which dispersal still occurs in fragmented landscapes. The purpose of this study was to investigate the landscape determinants of genetic differentiation in the hazel dormouse (Muscardinus avellanarius), a forest-dependent species of conservation concern. By comparing subpopulations in a continuous (SLR) and a fragmented (VTH) population, we (i) searched for the presence of Isolation-by-Resistance (IBR); (ii) estimated migration rates; (iii) evaluated the degree of inbreeding and genetic drift, and searched for their landscape determinants. We found an IBR effect in VTH, which heavily hindered the dispersal process. The overall number of migrants among VTH subpopulations was very low (1 per generation, compared to 15 in SLR), although a between-patch displacement of about 4 km along a well-structured hedgerow probably occurred. The inbreeding (F > 0.2 in most subpopulations) and the genetic drift (four out five subpopulations showed private alleles on several loci, with relatively high frequencies) are of particular concern in VTH. However, they were found to be limited in large patches or in patches connected by hedgerows with a high number of neighbouring patches. As a conservation strategy in the VTH landscape, characterized by small patches, we suggest that the dispersal process among subpopulations is enhanced to sustain a functional metapopulation. For this purpose, an effective ecological network should be created by enhancing the continuity and the internal features of hedgerows.