Intraspecific neutral and adaptive variation
The persistence of populations that are increasingly isolated depends largely upon the level of connectivity between them (Hodgson et al. 2011). Thus, the high differentiation observed among some populations within NIDGS might result from loss of suitable habitat and dispersal corridors between populations, currently leading to the isolation (and possible extinction) of populations (Yensen 1999; Sherman & Runge 2002; Barrett 2005). Such patterns are frequent in other small mammals with historically reduced ranges across altitudinal gradients (Waterhouse et al. 2018; Bi et al. 2019). For NIDGS, the neutral dataset identified two main groups (termed here ‘western’ and ‘eastern’, Figure 6A, K = 2) which corroborates previous studies that used microsatellite loci (Garner et al. 2005; Hoisington-Lopez et al. 2012; Zero et al. 2017). Our results also identified more fine-scale structure within the eastern part of the range than within the western part of the range, indicating higher connectivity across the populations sampled from the western region. This result is similar to that of previous microsatellite analyses for the eastern part of the range (although this study did not analyze the exact same populations; Garner et al. 2005), but contrasts somewhat with our hypothesis (a) of IBD and the results of an allozyme study, which found significant IBD within the western region (Gavinet al. 1999) and with unpublished results from Hoisington (2007) which found evidence for additional substructure and restricted gene flow within both the eastern and western groups. Very similar patterns of differentiation have also been found using mitochondrial DNA (Hoisington 2007). For SIDGS, the neutral dataset identified the highest differentiation between populations on different sides of the Weiser River, which corroborates previous studies (Garner et al. 2005; Hoisington 2007; Zero et al. 2017). Interestingly, although Olds Ferry was the most distinct population at the neutral level, it was not substantially differentiated from other populations on the eastern side of the Weiser River at the adaptive level, suggesting that its distinction is mostly due to demographic (i.e. neutral) processes (Garner et al. 2005; Hoisington-Lopez et al. 2012; Zeroet al. 2017). Previous studies sampled a larger number of sites (10-11) and detected additional subgroups both east and west of the Weiser River (Hoisington 2007; Zero et al. 2017). Earlier work on SIDGS, which included a larger number of populations and larger spatial area, found human disturbance (impervious surfaces and agriculture) and small-scale topographic complexity to restrict gene flow while higher at site heat load index, growing season precipitation and frost free period facilitated gene flow (Zero et al. 2017). Translocations among populations east of the Weiser River were performed to supplement small and isolated populations, and might have led to an increased level of genetic homogeneity at the neutral level (Yensen et al. 2010; Weeks et al. 2011; Yensen & Tarifa 2012; Landguth & Balkenhol 2012). Successful translocations would result in low neutral and adaptive differentiation, but previous studies have found low rates of translocation success (Panek 2005; Busscher 2009; Yensen et al.2010; Smith et al. 2019). These translocations were performed mainly into areas near the Weiser River population and might explain the low genetic differentiation observed between this population and all others east of the Weiser River (Figure S11C, Supporting information). Similarly, adaptive differentiation was lowest for WR, suggesting some benefits of translocations (increased genetic diversity and masking of deleterious alleles) outweighing the potential negative effects such as outbreeding depression (Weeks et al. 2011). Still, most of the SIDGS populations showed lower than expected neutral heterozygosity (Table 4), which suggests effects of inbreeding leading to reduced genetic diversity. This agrees with previous studies that have found lower genetic diversity in SIDGS than NIDGS, and point to the need of increased protection of this species (Garner et al. 2005).