Behavioral Ecology of Pollinator Assemblages

Abstract

Pollinator assemblages -- the multi-species communities of bees, hoverflies, butterflies, and other flower-visiting insects
that provide crop and wildflower pollination services -- are declining globally due to habitat loss, pesticide exposure,
pathogen spread, and climate change, yet their behavioural ecology -- the foraging decisions, flower fidelity, temporal
partitioning, and inter-specific interactions that determine pollination effectiveness -- remains poorly characterised at the
community level across agricultural-natural habitat gradients. This study quantified pollinator assemblage composition,
flower-visiting behaviour, and inter-specific temporal partitioning across 48 sites spanning an agricultural intensity
gradient in France, Switzerland, and Sweden, using standardised transect surveys (n = 18,420 flower visits recorded;
284 plant-pollinator interaction pairs) combined with individual-level RFID tracking of bumblebees (n = 2,842 individually
tagged Bombus terrestris and B. lapidarius workers) and pan-trap malaise trap community sampling. Total pollinator
species richness was positively correlated with floral diversity (r(S) = +0.84, p < 0.001) and negatively correlated with
agricultural intensity index (AII; r(S) = -0.72, p < 0.001). Bumblebee flower fidelity (proportion of visits to most-visited plant
species per foraging bout) was significantly higher at low-intensity sites (0.74 +- 0.08) than at high-intensity sites (0.48 +-
0.10; t = 8.42, p < 0.001), indicating that floral diversity promotes specialist foraging behaviour. Temporal partitioning of
flower visits among bee, hoverfly, and butterfly guilds was significant (PERMANOVA F = 12.4, p < 0.001), with bees
dominating morning visits, hoverflies peaking at midday, and butterflies predominantly visiting in the afternoon. Network
analysis of plant-pollinator interactions showed connectance declining with agricultural intensity (connectance 0.28 +-
0.04 at low-intensity vs. 0.14 +- 0.04 at high-intensity sites; t = 12.4, p < 0.001), indicating increasing network
specialisation and vulnerability to pollinator loss at high agricultural intensities. These results identify floral diversity
restoration as the priority management action for maintaining pollinator assemblage diversity and interaction network
robustness in agricultural landscapes.