Investigating non‐random assemblages emerging in response to environmental gradients is relevant to understand mechanisms and processes affecting biodiversity. Species may be filtered from fractions of environmental gradients that limit dispersal, survival or ontogenetic development, which ultimately leads to biotic complementarities among sites. Non‐random assemblages as a response to environmental filtering have been widely demonstrated in Amazonian forests, but are rarely assessed in non‐forest ecosystems such as macrophyte meadows covering lakes. In this study, we sampled 50 plots (50 m long, 6 m wide) along continuous macrophyte meadows in a lake system in the lower Amazon River. Our main goal is to test the effects of distance from the lake bank, macrophyte height and composition (frequency of morphotype occurrence), air temperature and physicochemical properties of water (pH, dissolved oxygen, depth and temperature) on frog α and β‐diversity estimates, and frequency of species traits occurrence (abundance‐weighted body size, toe pads, foot webbing and tadpole habit). We found 16 species, for which local assemblages quantified by α and β‐diversity estimates were not random, but predicted by macrophyte height, morphotype composition and water depth. We have explicitly shown that species are filtered from fractions of these gradients through ecomorphological relationships, since morphological traits and tadpole habits were also selected by the vertical stratification provided by the vegetation cover and water depth. Overall, we present an investigation of assemblage ecology that is relevant to conservation, because the results suggest biotic complementarities within habitats that are rarely considered as distinct biogeographic units from the surrounding várzea forests.
Abstract in Portuguese is available with online material.