Honey bees do not always immediately act on the directional information provided by the waggle dance to locate food sources [1].
This discovery challenges the long-held belief that bee colonies operate as a seamless, immediate relay system. Understanding how these insects process social information provides insight into the complexity of collective decision-making in nature.
The waggle dance is a social communication behavior used by honey bees (Apis mellifera) to convey the distance and direction of nectar sources [1]. While previous reports suggested that worker bees head straight to a source after observing a dance, recent findings indicate that bees almost never immediately jump up to follow these directions [1].
Researchers found that the impact of the dance depends heavily on the audience present on the comb's dance floor [1]. Specifically, the effectiveness of the communication fluctuates based on the size and composition of the listeners [1]. When too many bees are present, or when the audience is composed of inappropriate listeners, the bees are less likely to follow the directions right away [1].
These observations were supported by new visualization techniques. A method for visualizing the dance floor within the hive was published in February 2026 [2]. This allowed researchers to better observe the wiggling and looping motions of the bees, and how those movements change relative to the crowd [2].
Further studies on the precision of the waggle dance were reported in March 2026 [1]. The data suggests that the social environment of the hive acts as a filter for information, meaning a high-quality food source may not be exploited immediately if the social conditions on the dance floor are not optimal [1].
“Bees almost never immediately jump up to follow the directions they just got.”
This research indicates that honey bee foraging is not a simple stimulus-response mechanism. By showing that audience dynamics influence whether a bee acts on information, the study suggests that collective intelligence in hives involves a layer of social filtering that may prevent the colony from overreacting to single data points.





