Increasing efficiency and reducing bias in the detection of seed-dispersal interactions based on mist-netted birds [Dataset]

The dataset analysed in the study with one row for each ‘interaction event’. As defined in the ‘Data analysis’ section of the article, interaction events are interactions between a seed species i dispersed by an individual bird captured b. The dataset includes the capture of 108 frugivorous bird individuals from five different species, which dispersed seeds from 17 fleshy-fruited plant species (Table S1). These captures yielded a total of 138 interaction events (i.e. rows in the dataset) because some captured birds dispersed multiple seed species. Variables: • plant_sp: plant (seed) species. • bird_sp: bird species. • bird_ring: ring code of captured birds. • date: date of the capture. • nseeds_mesh: number of seeds (of the corresponding plant species) collected on the mesh below captured bird. • nseeds_bag: number of seeds (of the corresponding plant species) collected inside the cloth bag where the captured bird was kept to obtain droppings. • nseeds_total: nseeds_mesh + nseeds_bag • prop_seeds_mesh: proportion of seeds (of the corresponding plant species) on the mesh band. • prop_seeds_bag: proportion of seeds (of the corresponding plant species) inside the cloth bag. • occ_only_mesh: seed occurrence only on the mesh band. • occ_only_bag: seed occurrence only inside the cloth bag. • occ_both: seed occurrence both, on the mesh band and also inside the cloth bag. • occ_bag: seed occurrence inside the cloth bag irrespective of seed occurrence on the mesh band. • bird_weight: body weight (g) of the captured bird. • mean_seed_dry_weight: mean seed dry weight (mg) of the seed species dispersed. We used our own data except for five species (Crataegus monogyna, Ficus carica, Lonicera sp.*, Prunus avium and Rhamnus alaternus) for which the information was gathered from the literature (Torroba Balmori et al. 2013). *Note: For Lonicera sp. (sampled in two interaction events) we used the average value between Lonicera implexa and Lonicera periclymenum mean values

Efficient and unbiased sampling of ecological interactions is essential to our understanding of the functions they mediate. Seed dispersal by frugivorous birds is a key mutualism for plant regeneration and community dynamics. Mist-netting is one of the most widely used methods to sample avian seed dispersal through the identification of seeds in droppings of captured birds kept inside cloth bags. However, birds may drop seeds on the ground before being extracted from the net, leading to a fraction of missing information due to ineffective sampling. Worryingly, this fraction could be unevenly distributed across bird and plant species, leading to sampling biases. Here, we assess the effectiveness of using a 1- m wide mesh below mist nets to sample seeds dropped by entangled birds. We used data from birds mist-netted during one-year-round. We sampled nearly 50% of interaction events and 75% of dispersed seeds on the mesh band below the mist nets (i.e. lost information without this optimization). The proportion of seeds sampled on the mesh bands was not evenly distributed among bird species but strongly related to bird size, ranging from 57-63% in warblers to 84-94% in thrushes. Moreover, the proportion of seeds sampled on the mesh was negatively related to seed size, although this relationship was weaker. We also evaluated accumulation curves of species and pairwise interactions with increasing sampling effort, both with and without using the mesh bands. The number of seed species sampled increased by 21% when using the mesh bands and the number of pairwise interactions by 36%. Our findings provide strong evidence on how inefficient and biased traditional mist-netting can be for sampling community-wide seed-dispersal interactions. We thus urge the use of mesh bands in future studies to increase sampling effectiveness and avoid biases, which will ultimately improve our understanding of the seed dispersal function.

Fundación BBVA, Award: ClaveSER 044-2019

Regional Government of Asturias, Award: IDI/2018/000151

Peer reviewed