Plant‐centred sampling estimates higher beta diversity of interactions than pollinator‐based sampling across habitats

Abstract

Summary When describing plant–animal interaction networks, sampling can be performed using plant‐ or animal‐centred approaches. Despite known effects of sampling on network structure, how samplings affect the estimates of interaction β‐diversity across networks is still unresolved. We investigated how the sampling method affects the assessment of β‐diversity of interactions, turnover and rewiring. We contrasted plant‐ and animal‐centred sampling methods applied to pollination networks across habitats in a heterogeneous tropical landscape, the Pantanal Wetlands. We also asked whether plant traits influence the difference in interaction specialization according to sampling. Plant‐centred networks resulted in higher β‐diversity of interactions in space than animal‐centred networks. Turnover explained most of the β‐diversity in both methods, but rewiring was proportionately more important when using the animal‐centred method. While the plant‐centred method indicated lower network modularity and specialization, floral traits modulated the effects of the sampling method on species‐level network metrics. Combining animal‐ and plant‐centred approaches returned intermediate values for β‐diversity of interactions and network metrics. Distinct methods may also be better suited for answering questions at different scales. Our results point out that the method choice, or combination of methods, should always reflect the appropriate scale of the factors determining the interactions being investigated.

Behavioural and morphological traits influence sex‐specific floral resource use by hummingbirds

“…The mean number of individuals sampled for pollen loads was 12 females and 10 males from each species (ranging from 10 to 74 individuals per species) with a minimum of five individuals from each sex. We considered that a minimum sample size of 10 individuals is representative of the interactions for the species included in our study since pollen gives robust, spatially broad, cumulative link evidence and also detects interactions between rare species (Bosch et al, 2009; Ramírez‐Burbano et al, 2017; Souza et al, 2021). Thus, pollen loads can be a suitable estimator of an individual's foraging patterns (Courtney et al, 1981).…”

Integrating nocturnal and diurnal interactions in a Neotropical pollination network

“…Second, we only considered the presence-absence of interactions and not their intensity. Quantitative interaction data were hardly comparable since we compiled information from different sources that used distinct methodologies known to affect the quantitative assessment of interactions (Souza et al, 2021); and some of the original studies only reported presence/absence of interactions. Therefore, we employed a qualitative combination, a conservative approach most suitable for the data we had (Quintero et al, 2021).…”

Plant–hummingbird pollination networks exhibit minimal rewiring after experimental removal of a locally abundant plant species

“…Comprehensively sampling plant–pollinator interactions at the community level is challenging (Jordano, 2016), and we acknowledge that our methodological approach has some limitations. For pollen networks, we cannot exclude the possibility that hummingbirds were transporting pollen from broad geographic extents, therefore introducing unwanted variability (Jordano, 2016; Ramírez-Burbano et al ., 2017; Souza et al ., 2021). Additionally, pollen networks relied on capturing enough hummingbirds (and sampling enough interactions) to create meaningful networks; low capture numbers in some sites limited our overall sample size for pollen network analysis, and estimated sampling completeness for pollen networks was low and highly variable (mean ± SD: 50 ± 19%).…”

“…Our networks also allowed us to compare two parallel methods for sampling plant– pollinator interactions: pollen loads sampled from captured hummingbirds versus direct observations of hummingbirds visiting focal plants. We found higher specialization in the camera networks, consistent with previous comparisons of these network sampling approaches (avian pollinators: Ramírez-Burbano et al ., 2017; Zanata et al ., 2017; insect pollinators: Bosch et al ., 2009; but see Zhao et al ., 2019; Souza et al ., 2021). In our study, differences between sampling methods could arise because camera networks included visits where hummingbirds bypassed the plant’s anthers (robbing was observed in 11% of hummingbird visits).…”