ALBERT

Description: The goal of this study was to propose a functional approach to test four metacommunity models. We evaluated the importance of traits related to dispersal and growth strategies and species abundances in the regional pool, for the assembly of woody saplings in the forest–grassland mosaic of the southern Brazilian plateau. Mass effects were the model that best explained the process. Abstract Questions What is the relative influence of: (a) traits related to dispersal (seed size) and growth (specific leaf area [SLA]), and; (b) regional species abundances, on the structure of local sapling communities? Which metacommunity models best explain the process of forest nucleation in grasslands? Location Woody vegetation patches originating from nucleation and the continuous forest that neighbors native grasslands, in the southern Brazilian plateau (29°28′58.00″ S, 50°12′26.00″ W). Methods We calculated community‐weighted trait means (CWM) and functional diversity for woody saplings, considering species occurrences (FD) and abundances (FDis) separately for both seed size and SLA, in woody vegetation patches of different sizes and in the continuous forest. We tested whether mean SLA and mean seed size differed between habitat types, and whether values of FD and FDis were lower than expected by chance. We also evaluated whether the relative abundances of species in woody vegetation patches were determined by their abundances in the continuous forest, using CATS models (Community Assembly by Trait Selection). The results were then compared to the expectations from each metacommunity model. Results Functional diversity of SLA and seed size did not differ from what was expected by chance in most sites. Species in communities of different habitat types did not differ in their mean SLA values, but we observed differences in mean seed size, with large‐seeded species being typical of areas with more shade. Species abundances in the regional species pool were more important factors than species traits for determining species abundances in the local communities. Conclusions The analytical framework employed in this study enables a straightforward test of different metacommunity structuring models. Mass effects were the metacommunity model that best explained the process of forest nucleation in the southern Brazilian plateau, which highlights the importance of dispersal for the structure of local communities.