Sampling completeness in seed dispersal networks: When enough is enough

摘要: Abstract Ecological networks are an increasingly popular tool to explore community assembly rules and frame practical conservation issues. However, most described vary largely in sampling effort, hampering the distinction of true biological patterns from artefacts caused by poor sampling. Identifying entire seeds droppings mist-netted birds is generally considered a preferred method for building unbiased, quantitative seed dispersal networks. We retrieved 936 captured during five days seven sites Portugal estimated completeness as diversity species, disperser links detected with respect those predicted Chao 2 estimator. In one sites, effort was extended 25 evaluate sensitivity ten network structure descriptors increasing effort. After we 93% 97% 79% 2, however resulted detection more seeds, dispersers, than at day 5. Most only began stabilize around 8, except Connectance Weighted that stabilized earlier. Similarly, after 8 significantly departed confidence interval null models exclusively constrained species abundances, thus reflecting independent ecological patterns. Nestedness exception, it never models. suggest may slightly underestimate real our case least eight were needed build sound 67% 88% 71% detected. Our results have important implications interpretation because under-sampled produce biased do not suitably characterize focal communities.