Researchers Pinpoint Honey Bee 'Dance Floors' Inside Hives with a New Data-Driven Method

For a honey bee returning from a productive food source, communication is a matter of geometry. The waggle dance - a figure-eight movement performed on the vertical surface of a honeycomb - encodes both the direction and the distance of a nectar or pollen source relative to the sun's position. Nestmates observing the dance can then fly directly to the resource, often covering kilometers of unfamiliar terrain. The dance is among the most studied forms of animal communication in biology, and for good reason: it is one of the few known examples of symbolic language outside the human lineage.

Yet for all the attention waggle dances have received, a basic question about where they happen inside the hive has not had a rigorous answer. Forager bees returning to the hive do not dance at random locations. They tend to cluster in specific regions of the comb - what researchers have informally called the "dance floor." These regions are not precisely defined, they vary between colonies, and they shift with season and colony composition. The informal definitions researchers have used to study dance behavior have made it difficult to compare results across studies, colonies, or experimental conditions.

A study published in PLOS ONE, by researchers from Canada and the United States, addresses that problem with a data-driven statistical method that objectively identifies dance floor boundaries from observational data, without requiring researchers to make subjective judgments about where the dance floor ends.

The Problem with Informal Definitions

In practice, researchers studying waggle dances have historically designated dance floor regions based on visual inspection of hive observation frames, prior knowledge of typical dance locations, or rough geometric divisions of the comb. These approaches work well enough for single experiments but create obstacles when trying to compare behavioral patterns across different colonies, different hive designs, different seasons, or studies conducted by different research groups.

Whether a bee dancing near the edge of the conventionally designated dance floor counts as dancing "on" or "off" the floor can affect measures of dance frequency, dance floor occupancy, and the relationship between dance behavior and environmental variables. Inconsistent definitions introduce noise into datasets and can obscure real biological patterns or create apparent patterns that reflect definitional choices rather than bee behavior.

A Statistical Approach to Boundary Detection

The new method treats dance floor identification as a spatial statistics problem. By analyzing the spatial distribution of observed dance events - the locations on the comb where individual waggle dances are performed - the algorithm identifies regions of high dance density and draws objective boundaries separating the dance floor from surrounding comb areas where dances rarely occur.

The approach is data-driven in the sense that the boundaries it identifies are determined by the actual dance location data from a specific colony rather than by a researcher's prior expectations. Applied to different colonies, the method produces colony-specific dance floor maps that reflect each colony's actual behavioral patterns. Applied to the same colony over time, it can track shifts in dance floor location as colony conditions change.

Critically, the method generates dance floor definitions that are reproducible: two researchers analyzing the same dataset will obtain the same result, eliminating the between-researcher variability that plagues approaches based on visual inspection.

Enabling Comparisons Across Colonies and Studies

The most direct application is in studies comparing dance behavior across multiple colonies or across experimental conditions. When dance floor regions are defined consistently by the same method across all colonies in a study, any observed differences in dance floor area, position, or occupancy reflect real variation in bee behavior rather than definitional inconsistency.

The method also supports the growing body of research connecting dance behavior to external environmental factors - local forage availability, landscape composition around the hive, competition from neighboring colonies. These questions require accurate, consistent measurement of where and when dances occur inside the hive. The availability of an objective dance floor identification method removes a methodological constraint that has limited such analyses.

Waggle dance research has increasingly moved toward automated video monitoring systems that track bee position and movement continuously without requiring a human observer to watch the hive. The data-driven dance floor detection method is compatible with this automation: dance location coordinates extracted from video tracking can be fed directly into the algorithm to produce dance floor maps without manual intervention.

Scope and Limitations

The study was conducted by research teams in Canada and the United States, with funding from the Natural Sciences and Engineering Research Council of Canada (Discovery Grant RGPIN-2020-05690) and the U.S. National Science Foundation Division of Mathematical Sciences (NSF 1128954). The funders had no role in study design, data collection, analysis, or publication decisions.

The method was developed and validated for honey bees (Apis mellifera), the primary model species in waggle dance research. Whether the approach applies directly to other eusocial species - other bee genera, stingless bees, or other social insects that use spatial communication - would require testing in those systems. The algorithm also requires sufficient dance event data to reliably estimate the dance distribution; in colonies with low dance activity, the statistical boundaries may be less precisely defined.

The work is published as open-access research in PLOS ONE.