2.3 Community phylogeny and functional traits of small mammals
We extracted the phylogenies of the species in the dataset from a recently published supertree of small mammals (Faurby et al., 2018). We downloaded 10000 final trees from http://vertlife.org/phylosubsets/. Subsequently, we sampled 5000 pseudoposterior distributions and constructed a maximum clade credibility tree using the mean node heights in TreeAnnotator v1.10.4 software of the BEAST package. The resulting tree was used for all subsequent phylogenetic analyses (Fig. S1).
Eight commonly used small mammal functional traits were selected for analysis: (1) body length (mm), 2) body mass (g), 3) ear length (mm), 4) tail length (mm), 5) hind foot length (mm), 6) diet (herbivore and insectivore), 7) foraging stratum (ground and non-ground), and 8) daily activity (nocturnal, crepuscular and diurnal) (Sun et al., 2020). The morphological trait data of a species was calculated as the average value of all its captured individuals, and other traits were compiled from Elton Traits 1.0 (Appendix; Wilman et al., 2014). To create the functional distance matrix (Gower distance), we used the function “vegdist” in the R package “vegan” (Oksanen et al. 2012). We clustered the functional distance matrix and obtained a functional dendrogram (Fig. S2). The resulting tree was used for all subsequent functional analyses (Fig. S2).
We quantified the phylogenetic signals to identify possible differences in the patterns of functional and phylogenetic structures. If a significant phylogenetic signal is detected for a trait, it means that the trait is more similar among closely related species than expected by chance (Revell et al., 2008). For continuous traits, the phylogenetic signal was quantified using the K statistic. K ≥ 1 suggests the presence of a phylogenetic signal. K close to zero suggests the presence of a weak phylogenetic signal (Blomberg et al., 2003). For binary traits, the D statistic was used to quantify the phylogenetic signal. If D approaches 0, there is conserved trait evolution, whereas D < 0 suggests trait clustering. A value of D ≥ 1 indicates no phylogenetic signal or a trait that is overdispersed on the phylogenetic tree (Fritz & Purvis, 2010). The D statistic and K statistic were calculated by using the “phylo.d” and “multiPhylosignal” functions in the R packages “caper” and “picante”, respectively (Kembel et al., 2010; Orme et al., 2013).