How much sampling is enough? Four decades of understorey bird mist‐netting across Amazonia define the minimum effort to uncover species assemblage structure

Mist‐net sampling comprises a key methodological component of assemblage‐wide avifaunal studies, particularly in the understorey of closed‐canopy tropical forests. To investigate mist‐net bird captures and species assemblage structure, we compiled data from 312 sites across the Pan‐Amazon. Using coverage‐based rarefaction and extrapolation curves, we propose a minimum sampling effort, standardizing samples by completeness, to obtain a robust representation of local understorey bird assemblages in Amazonian forests. Furthermore, we aimed to assess the variation in capture rates across sampling days and hours, as well as to characterize species assemblages based on their ecological and morphological attributes. The sample coverage for 312 sites varied from 0.46 to 1.0. Sample size (represented by numbers of individuals) varied according to the standardized sample coverage. For sample coverage of 0.9, the sample size ranged from 30 to 271, with an average of 128.7 individuals. Species‐rich sites (>60 understorey bird species) required a larger sample size than more depauperate sites to achieve a sample coverage of 0.9. Capture success was greatest on the first day of sampling (44.8% of all individuals and 83.7% of species), and in the morning (70.8% and 96.4%). The archetypal bird sampled by understorey mist‐netting in Amazonian forests is a fairly common small‐bodied insectivore that occupies only one habitat type and spends most of its time perched above ground on shrubs or saplings. We suggest that a representative minimum sample of understorey birds should include, on average, 129 individuals per sampling unit. However, this threshold can vary according to local species richness, and we propose different sample sizes for different levels of species richness. For ecological studies where abundance data are essential to derive robust diversity estimates, we recommend operating mist‐nets from dawn to dusk. Our findings can guide future mist‐net surveys of understorey birds, optimizing resource allocation in sampling effort.