Forest dynamics plots, which now span longitudes, latitudes, and habitat types across the globe, offer unparalleled insights into the ecological and evolutionary processes that determine how species are assembled into communities. Understanding phylogenetic relationships among species in a community has become an important component of assessing assembly processes. However, the application of evolutionary information to questions in community ecology has been limited in large part by the lack of accurate estimates of phylogenetic relationships among individual species found within communities, and is particularly limiting in comparisons between communities. Therefore, streamlining and maximizing the information content of these community phylogenies is a priority. To test the viability and advantage of a multi-community phylogeny, we constructed a multi-plot mega-phylogeny of 1347 species of trees across 15 forest dynamics plots in the ForestGEO network using DNA barcode sequence data (rbcL, matK, and psbA-trnH) and compared community phylogenies for each individual plot with respect to support for topology and branch lengths, which affect evolutionary inference of community processes. The levels of taxonomic differentiation across the phylogeny were examined by quantifying the frequency of resolved nodes throughout. In addition, three phylogenetic distance (PD) metrics that are commonly used to infer assembly processes were estimated for each plot [PD, Mean Phylogenetic Distance (MPD), and Mean Nearest Taxon Distance (MNTD)]. Lastly, we examine the partitioning of phylogenetic diversity among community plots through quantification of inter-community MPD and MNTD. Overall, evolutionary relationships were highly resolved across the DNA barcode-based mega-phylogeny, and phylogenetic resolution for each community plot was improved when estimated within the context of the mega-phylogeny. Likewise, when compared with phylogenies for individual plots, estimates of phylogenetic diversity in the mega-phylogeny were more consistent, thereby removing a potential source of bias at the plot-level, and demonstrating the value of assessing phylogenetic relationships simultaneously within a mega-phylogeny. An unexpected result of the comparisons among plots based on the mega-phylogeny was that the communities in the ForestGEO plots in general appear to be assemblages of more closely related species than expected by chance, and that differentiation among communities is very low, suggesting deep floristic connections among communities and new avenues for future analyses in community ecology.

Seedling dynamics play a crucial role in determining species distributions and coexistence. Exploring causes of variation in seedling dynamics can therefore provide key insights into the factors affecting these phenomena. We examined the relative importance of biotic neighborhood processes and habitat heterogeneity using survival data for 5,827 seedlings in 39 tree and shrub species over 2 years from an old-growth temperate forest in northeastern China. We found significant negative density-dependence effects on survival of tree seedlings, and limited effects of habitat heterogeneity (edaphic and topographic variables) on survival of shrub seedlings. The importance of negative density dependence on young tree seedling survival was replaced by habitat in tree seedlings ≥ 4 years old. As expected, negative density dependence was more apparent in gravity-dispersed species compared to wind-dispersed and animal-dispersed species. Moreover, we found that a community compensatory trend existed for trees. Therefore, although negative density dependence was not as pervasive as in other forest communities, it is an important mechanism for the maintenance of community diversity in this temperate forest. We conclude that both negative density dependence and habitat heterogeneity drive seedling survival, but their relative importance varies with seedling age classes and species traits.

The underlying drivers of β‐diversity along latitudinal gradients have been unclear. Previous studies have focused on β‐diversities calculated at a local scale and shed limited light on regional β‐diversity. We tested the much‐debated effects of range size vs. environmental filtering on the β‐gradient using data from the US Forest Inventory Analysis Program. We showed that the drivers of the β‐gradient were scale dependent. At the local scale species spatial patterns contributed little to the β‐gradient, whereas at the regional scale spatial patterns dominated the gradient and a U‐shape latitudinal relationship for the standardised β‐diversity deviation was revealed. The relationship can be explained by spatial variation in climate and soil texture, thus supporting the environmental filtering hypothesis. But it is inconsistent with Rapoport's rule about the effect of range size on β‐gradient. These results resolve the debate on whether species spatial distributions contribute to β‐gradient and attest the importance of environmental filtering in determining regional β‐diversity.