A tangled web in tropical tree-tops: effects of edaphic variation, neighbourhood phorophyte composition and bark characteristics on epiphytes in a central Amazonian forest

Question: How do soil fertility, neighbourhood host tree composition and bark characteristics influence community attributes of vascular epiphytes in a central Amazonian forest? Location: Adolpho Ducke Forest Reserve, Amazonas, Brazil. Methods: The abundances of all species of vascular epiphytes were recorded from 300 host trees with DBH >30 cm. Phorophytes were distributed among 30 sample plots established in lowland, slope and upland habitats, and bark characteristics were classified into five types (rough, peeling, fissured, smooth, rugose). The epiphytic species were also classified into three functional types (true epiphytes, hemi-epiphytes and nomadic vines) for separate analyses in an effort to highlight differences in ecological constraints in relation to neighbourhood phorophyte composition and edaphic variation. Results: More than 21 000 individuals representing 122 species distributed in 66 genera and 20 families of vascular epiphytes were recorded from 300 phorophytes during the course of the study. Multiple regressions demonstrated that neighbourhood phorophyte composition and edaphic variables, nitrogen (N) and phosphorus (P), together explained 65% of the variation in epiphytic species composition. However, edaphic factors were the only significant predictors of epiphytic species richness, accounting for 70% of this variation. Nearly one-third of the true epiphytes occurred only on phorophytes in soils with high P content. Null models demonstrate that a vast majority of the epiphyte species (76% in positive and 96% in negative associations) failed to exhibit a statistically significant preference for bark texture. Conclusion: Neighbourhood phorophyte composition, in concert with N and P availability, accounts for the vast majority of the variation in vascular epiphyte composition and abundance in central Amazonian forests. However, only N and P were shown to be significant predictors of epiphyte richness. Such results suggest that both biotic and abiotic variables influence epiphyte community structure (composition, abundance and richness) at differing intensities. Weak associations between bark type and epiphyte taxa suggest that epiphyte community variation may also be influenced by other unmeasured physical and environmental factors such as phorophyte architecture or microclimatic differences. Nonetheless, these results corroborate with other studies which report significant changes in epiphyte composition and richness as a result of experimentally manipulated ground soil fertility.