Résumé
Background: Phylogenies can help understand the processes governing species assembly in natural communities because they convey information on species shared history and evolution. Here, plant phylogenies based on rbcL and matK DNA barcodes were used to investigate the phylogenetic structure of plant communities at two contrasting scales: ( i )a 50 ha plot in western Cameroon, and ( ii ) the different floristic regions of tropical Africa. Plant distribution data came from ( i ) a systematic census of all trees ( 300 000 stems from 450 species), and ( ii ) a newly assembled large floristic database for Tropical Africa, called RAINBIO ( 600 000 occurrence points from 25 000 plant species). For each dataset, we investigate in particular species turnover and lineage turnover to obtain insights on assembly rules. Results: Within the 50 ha plot, a general pattern of phylogenetic clustering occurred at all the scales investigated, and phylogenetic turnover in space was correlated with topographic habitat differentiation. These results suggest that local environmental conditions tend to favor the establishment of phylogenetically related species, and we did not find evidence of processes favoring the local assembly of less related species that would be more complementary in terms of niche partitioning. At the continental scale, species and lineage turnover between floristic units depends both on geographic distance and on ecological contrast, but the turnover of lineages better correlates with ecological distances than with spatial distances. Species turnover and lineage turnover can therefore provide complementary information to define vegetation units that highlight biogeographic isolation and ecological differentiation, respectively. Significance: Plant phylogenies offer new perspectives to understand the origin of floristic assemblages. To this end, new theoretical work is probably needed to better establish the links between patterns and process in order to identify the best ways to integrate community and phylogenetic data.