Pesticide exposure has been implicated as a contributor to insect pollinator declines. In social bees, which are crucial pollination service providers, the effect of low-level chronic exposure is typically non-lethal leading researchers to consider whether exposure induces sublethal effects on behaviour and whether such impairment can affect colony development. Studies under laboratory conditions can control levels of pesticide exposure and elucidate causative effects, but are often criticized for being unrealistic. In contrast, field studies can monitor bee responses under a more realistic pesticide exposure landscape; yet typically such findings are limited to correlative results and can lack true controls or sufficient replication. We attempt to bridge this gap by exposing bumblebees to known amounts of pesticides when colonies are placed in the field. Using 20 bumblebee colonies, we assess the consequences of exposure to the neonicotinoid clothianidin, provided in sucrose at a concentration of five parts per billion, over 5 weeks. We monitored foraging patterns and pollen collecting performance from 3282 bouts using either a non-invasive photographic assessment, or by extracting the pollen from returning foragers. We also conducted a full colony census at the beginning and end of the experiment. In contrast to studies on other neonicotinoids, showing clear impairment to foraging behaviours, we detected only subtle changes to patterns of foraging activity and pollen foraging during the course of the experiment. However, our colony census measures showed a more pronounced effect of exposure, with fewer adult workers and sexuals in treated colonies after 5 weeks. Synthesis and applications. Pesticide-induced impairments on colony development and foraging could impact on the pollination service that bees provide. Therefore, our findings, that bees show subtle changes in foraging behaviour and reductions in colony size after exposure to a common pesticide, have important implications and help to inform the debate over whether the benefits of systemic pesticide application to flowering crops outweigh the costs. We propose that our methodology is an important advance to previous semi-field methods and should be considered when considering improvements to current ecotoxicological guidelines for pesticide risk assessment.

Arce.et.al_censusCensus data from 20 bumblebee colonies before and after a 5 week semi field experiment.Arce.et.al_forage_freqTotal observed returning foragers from a 5 week semi field experiment. Freq = total number of foragers, success = individuals carrying pollen, fail = individuals not carrying pollenArce.et.al_observer1Observations of foraging behaviour from a five week semi-sield experiment. Data comes from Observer 1 who counted the number of bees returning to each colony and removed the pollen from one corbicula from any bee observed carrying pollen.Arce.et.al_observer2Observations of foraging behaviour from a five week semi-sield experiment. Data comes from Observer 2 who photographed each bee returning to each colony, also included is the 2 dimensional size of the pollen load.JAE_script_sumbissionR script for the analysis of the data from Arce et al. "Combining realism with control: impact of controlled neonicotinoid exposure on bumblebees in a realistic field setting"