Table S1 List of the 19 microsatellite loci used in our study. The amplification was performed in four PCR (A, B, C and D) and screening was realized in two set. The symbol (*) indicated the primers sequences redesigned from Rikalainen et  al. (2008). The table indicated for each marker the number of alleles detected and range size.

Fig. S1 Spatial autocorrelation of genetic relatedness for each site. Relatedness coefficients were estimated for intervals of upper bounds equal to 100, 200, 300, 400, 500, 1000 and 2000 m. Standard deviations were estimated using the Jacknife procedure implemented in spagedi.

Fig. S2 Large scale genetic clustering including the whole dataset. (A) Log posterior probabilities (L(K)) for 100 runs for K values ranging between 1 and 10. (B) Absolute value of the change of likelihood distribution (L”(K)) for K values ranging between 1 and 10. (C) Distribution of Delta(K), i.e. L”(K) divided by the standard deviation of L(K), for K values ranging between 1 and 10. According to Evanno et  al. (2005), the modal value of Delta(K) corresponds to the best-fit number of genetic clusters (K). For K = 4, L(K) values exhibited a bimodal distribution, which reflected the two different modes evidenced by the clumpp analyses. Higher values corresponded to the genetic clustering displayed in Fig.  4b, while lower values corresponded to the genetic clustering illustrated in Fig.  4c.

Fig. S3 Genetic clustering using the spatial model implemented in geneland with the correlated allelic frequencies model (right panels). For each analysis, we represented (a) the number of genetic clusters along the MCMC after burnin, and (b) the map of population membership to detected clusters. On the map, each colour corresponds to one genetic cluster and black point indicated the localisation of each sampled site.

Fig. S4 Isolation by distance. The graph represents the relationship between pairwise values of FST/(1 − FST) and logarithm of geographical distance between sites. Filled squares symbolize values corresponding to forest sites, while empty circles correspond to values involving hedge networks sites. The linear curve represents the significant isolation by distance pattern observed for forest sites (P < 0.05). There is no significant pattern of isolation by distance when considering hedge networks sites.

Fig. S5 Genetic isolation of bank vole populations as estimated from local FST (mean ± 95% confidence interval). Black filled circles correspond to forest sites and open circles to hedge networks habitats.

MEC_5199_sm_figureS1.pdf10KSupporting info item
MEC_5199_sm_figureS2.pdf39KSupporting info item
MEC_5199_sm_figureS3.pdf32KSupporting info item
MEC_5199_sm_FigureS4.pdf6KSupporting info item
MEC_5199_sm_FigureS5.pdf21KSupporting info item
MEC_5199_sm_Supplementarymaterialslegends.doc28KSupporting info item
MEC_5199_sm_TableS1.doc99KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.