Animals develop food preferences based on taste, nutritional quality and to avoid environmental toxins. Yet, measuring preferences in an experimental setting can be challenging since ecologically realistic assays can be time consuming, while simplified assays may not capture natural sampling behavior. Field realism is a particular challenge when studying behavioral responses to environmental toxins in lab-based assays, given that toxins can themselves impact sampling behavior, masking our ability to detect preferences. We address these challenges by comparing different experimental methods for measuring sucrose concentration preference in bumble bees (Bombus impatiens), evaluating the utility of two preference chamber-based methods (ad libitum versus a novel restricted-sampling assay) in replicating bees’ preferences when they fly freely between artificial flowers in a foraging arena. We find that the restricted-sampling method matched a free-flying scenario more closely than the ad libitum protocol, and we advocate for expanded use of this approach, given its ease of implementation. We then performed a second experiment using the new protocol to ask whether consuming the neonicotinoid pesticide imidacloprid, known to suppress feeding motivation, interfered with the expression of sucrose preferences. After consuming imidacloprid, bees were less likely to choose the higher-quality sucrose even as they gained experience with both options. Thus, we provide evidence that pesticides interfere with bees’ ability to discriminate between floral rewards that differ in value. This work highlights a simple protocol for assessing realistic foraging preferences in bees and provides an efficient way for researchers to measure the impacts of anthropogenic factors on preference expression.

动物根据口味、营养质量和避免环境毒素来形成对食物的偏好。然而，在实验环境中测量偏好可能具有挑战性，因为生态现实化验可能非常耗时，而简化化验可能无法捕捉自然采样行为。在基于实验室的分析中研究对环境毒素的行为反应时，现场真实性是一个特别的挑战，因为毒素本身会影响采样行为，掩盖我们检测偏好的能力。我们通过比较测量熊蜂（Bombus impatiens ）蔗糖浓度偏好的不同实验方法来解决这些挑战)，评估了两种基于偏好室的方法（随意与新型限制采样分析）在复制蜜蜂在觅食场所人造花之间自由飞行时的偏好的效用。我们发现受限采样方法比随意协议更接近于自由飞行场景，我们提倡扩大使用这种方法，因为它易于实施。然后，我们使用新方案进行了第二个实验，以询问食用新烟碱类杀虫剂吡虫啉（已知会抑制进食动机）是否会干扰蔗糖偏好的表达。食用吡虫啉后，蜜蜂不太可能选择更高质量的蔗糖，即使它们获得了两种选择的经验。因此，我们提供的证据表明，杀虫剂会干扰蜜蜂区分价值不同的花卉奖励的能力。这项工作强调了一种用于评估蜜蜂真实觅食偏好的简单协议，并为研究人员测量人为因素对偏好表达的影响提供了一种有效的方法。