High gene flow across a heterogeneous landscape
Seasonally stable temperatures and steep ecological gradients on tropical mountains are predicted to limit dispersal across elevations (Janzen, 1967; Ghalambor et al., 2006). This prediction is supported by several studies that show limited gene flow and significant genetic differentiation across elevations in taxa with continuous elevational distributions (Bertrand et al., 2014; Milá, Warren, Heeb & Thébaud, 2010; Gadek et al., 2018), sometimes with elevational parapatry between sister species (Linck et al., 2019; Raxworthy et al. 2008a; DuBay & Witt, 2014). For example, Gadek et al. (2018) found cryptic population genetic structure differentiating high from low elevation populations in 3 of 4 Peruvian songbird species with continuous elevational distributions.
Our results are not consistent with a hypothesis of restricted gene flow across elevational gradients. We report evidence of high gene flow between MK and MT as well as between low and high elevations on both peaks, indicating that neither the lowland habitat connecting the two peaks, nor the steep elevational gradient across which mountain treeshrews occur on each peak, has significantly limited effective dispersal.
The mountain treeshrew population in KNP is best described as comprising two or three clusters. The primary population division does not correspond to the two peaks; rather, the summit region of MT (≥ 2000 masl) is distinct (Figure 5a). When dividing the population into three clusters, individuals at low elevation MK show mixed ancestry with low elevation MT (Figure 5b). Additionally, the MIGRATE-N analysis supports the division of individuals into three population clusters (high elevation MT, low elevation MT, and all of MK), with high migration rates between all pairs (Figure 2).
The temperature lapse rate in KNP is -0.55°C per 100 meters of elevation gain (Kitayama, 1992) which means that mountain treeshrews experience a 12.65°C average range in temperature between 900 and 3200 masl on MK. This temperature range is higher than the thermal neutral zone for most mammals that weigh < 1 kg (Khaliq, Hof, Prinzinger, Böhning-Gaese, & Pfenninger, 2014). On MT, mountain treeshrews experience an 8.8°C range from 900 to 2500 masl (Camacho-Sanchez et al., 2018). On MK, the species occupies four distinct vegetation zones, while on MT it occupies three (Kitayama, 1992). If gene flow were restricted due to limited elevational dispersal or selection against cross-elevation migrants, we would expect to find greater genetic differentiation on MK because of the broader elevational range and associated diversity of environmental factors on this slope compared to MT. However, the summit of MT was consistently recovered as the most distinct population cluster while individuals caught along the entire elevational gradient on MK form a single cluster (Figures 4 & 5). This suggests that elevation is not the primary variable influencing genetic structure in KNP.
Tropical montane species tend to have narrower ecological tolerance and elevational ranges than those in the temperate zone. McCain (2009) reviewed 170 montane gradients across vertebrate taxa to test this hypothesis and found that most groups showed the predicted pattern of decreasing range size with decreasing latitude. However, in rodents there was either no relationship or range size increased with declining latitude. This finding could be explained by the presence of cryptic species pairs or genetic differentiation separating low elevation populations from high (den Tex, Thorington, Maldonado, & Leonard, 2010; Hinckley, Hawkins, Achmadi, Maldonado, & Leonard, in review). However, McCain (2009) hypothesized that rodents may be able to cope with the lower temperatures associated with increasing altitude through behavioral adaptations including burrowing. Further studies are necessary to determine whether mountain treeshrews show behavioral or phenotypic plasticity, adaptive phenotypic (e.g. Branch et al., 2017) or genetic differentiation despite gene flow, or a generalist phenotype (e.g. Feijó et al., 2019) that allows them to persist across such broad environmental conditions which is rare among small mammals in this landscape (Nor, 2001; Camacho-Sanchez et al., 2019).