Trait conservatism
Admittedly, basic understanding on trait conservatism through phylogenetic signal detection greatly affects our mechanistic interpretations on assembly process (Losos 2008). Even though prior studies have suggested that lambda is not sensitive to the size of species pool (Wang & Clarke 2014; Blomberg et al. 2003), however, we found both K and lambda metrics significantly varied across datasets. Besides, most of morphological traits were phylogenetically convergent, especially in LSD. Significant but weak signal (K< 1, p< 0.05) probably results from adaptive convergence to particular environment and indicates labile phylogenetic signal (Blomberg et al. 2003). These results in phylogenetic signal detection is not surprising especially in the most diverse mammalian lineage(Fabre et al. 2012). Environment driven convergent adaptation in functional traits can largely promote coexistence among distant species, efficiently reduce habitat resources caused competitive exclusion among relatives and ultimately facilitate interspecific differentiation in rodents (Cavender-Bares et al. 2009; Cavender-Bares et al. 2004). Absolutely, a single comparison between two species pools might be less comprehensive to infer general conclusion. However, to some sense, these results have provided us an opportunity to rethink about ‘what is phylogenetic signal per se ’ and ‘how to understand the degree of PNC by detecting phylogenetic signal’. After all, phylogenetic signal obtained from a certain species pool could not represent the complete evolutionary history for every lineage (Swenson 2013).