3.2 Phylogenetic and functional community structures
At the landscape scale, SES.MPD values were less than zero in all forest types, while SES.MFD values were close to (but greater than) zero. At the site scale, SES.MPD values were significantly lower than zero in all forest types (t-test, t =-3.719, df =17, p=0.002; t-test,t =-2.851, df =17, p=0.011; t-test, t =-3.281,df =17, p=0.004), and there was no significant difference in the SES.MPD values among the three forest types (one-way ANOVA, df =2,F =0.178, p=0.838). SES.MFD values were close to zero in forest types (Wilcoxon test, V =111, p=0.284; t-test, t =-0.798,df =17, p=0.436; t-test, t =-0.609, df =17, p=0.551), and there was no significant difference in SES.MFD among forest types (one-way ANOVA, df =2, F =0.589, p=0.559, Fig. 5).
The K values of body length, tail length, hind foot length and ear length were all less than 1, indicating weak phylogenetic signals of these functional traits (i.e. trait convergence). D statistics revealed weak phylogenetic signals for foraging stratum and daily activity (0<Dobs<1; i.e. trait convergence) but strong signals for the other four binary traits (Dobs<0; i.e. trait conservatism). The distributions of foraging stratum and daily activity differed significantly from random distributions along the phylogeny, while only that of daily activity differed from a Brownian distribution (Table S1).