Variation of functional traits in trees from a biogeographycally complex Mexican cloud forest

Abstract

Several studies have proposed a group of morpho-functional traits as determinants of the ecological strategy of species. Among these, four morpho-functional traits are considered to be relevant in determining a plant's ecological strategy: specific leaf area (SLA), height at maturity (Hmax), wood density (WD), and seed mass (SM). We examined the variation of these traits and attempted to identify functional groups among 33 tree species with different biogeographical affinities from a montane cloud forest. Covariation among the four traits was examined using Principal Component Analysis (PCA) and species clustering. Bivariate trait relationships were evaluated through two methods: cross-species correlations, and evolutionary divergence correlations using phylogenetically independent contrasts (PICs). Correlations between attributes were overall weak, the most obvious ones being between Hmax and SM, and between Hmax and WD; this latter trait pair was also correlated in PICs. In both analyses SLA was unrelated to all other traits. In the PCA ordination the first two axes explained 66.9% of the between-species variation. Despite a largely continuous between-species variation, species clustering allowed differentiation of two main groups. Observed trait correlations were consistent with those reported for other floras, with the important exception of the independent behaviour of SLA. This study indicates a variety of comparable successful life history strategies among the studied species. The effect of phylogeny in trait covariation was unimportant, in fact, a mixture of clades was represented in several groups among the species they contained, suggesting among-lineage convergence.