1. Key takeaways

• Janine Kievits's article brings together two studies: one shows that the waggle dance involves an element of social learning, the other that bees rely heavily on ventral optic flow to control their flight altitude.
• In Dong et al. (2023), young dancers deprived of role models produce more disordered dances, with more directional and distance errors.
• After 20 days of experience, directional errors decrease, but the distance-coding error persists in the same bees.
• In Serres et al. (2022), reducing or eliminating the optic flow beneath the bee causes loss of altitude, up to collision with the floor mirror.
• For the beekeeper, the value lies mainly in interpretation: better understanding orientation flights, the dance and visual cues. The only really concrete practical implication concerns above all the siting of an apiary or a watering point: avoid large, calm, smooth, featureless water surfaces in the immediate vicinity.

2. What the study shows

Fig. In a flight tunnel, the bee flies at equal distance from both sides; but if one side of the tunnel moves towards the bee, it veers towards the other side, because the speed of the moving pattern creates the illusion that the moving wall is closer than the other. (After Srinivasan et al. (1999)).

The article combines two independent experiments to explore the cognition of the forager.

Question. Is the honey bee's waggle dance, which encodes the distance and direction of a food source, an entirely instinctive behaviour, or does it involve an element of social learning? And how do foragers regulate their flight altitude, particularly over surfaces that deprive their vision of ventral cues? These two questions are linked by the central role of vision in foraging and communication.

Method. For the dance, Dong et al. (2023) set up five experimental colonies from cohorts of one-day-old bees, in which the first dancers had never had the opportunity to follow an experienced dancer. These colonies were compared with five control colonies of mixed ages. The bees, individually marked, were trained to a 55% sucrose feeder placed at 150 m. The first dances were recorded at about 9 days in the experimental colonies and about 10 days in the control colonies, and the same dancers were re-observed 20 days later.

For flight, Serres et al. (2022) flew bees in an outdoor tunnel 2.20 m long, 71 cm high and 25 cm wide, motivated by a feeder placed at the end of the tunnel. After training under control conditions, several optical configurations were tested: ceiling mirror, floor mirror, double mirror at ceiling and floor, and partial floor mirror. The aim was to alter visual information from above and from below independently.

Results. Naive dancers produced significantly more disordered dances, with greater angular dispersion and longer waggle duration—thus signalling an erroneous distance—than dancers from control colonies. Twenty days later, the same bees danced with angular precision comparable to controls but retained their distance-coding error.

On the flight side, the trajectory remained normal as long as optic flow beneath the bee was available: impoverishing visual information at the ceiling did not significantly modify the trajectory in this set-up. By contrast, as soon as the floor was reflective, the bee progressively lost altitude until it struck the surface, even when the non-mirrored ceiling was reflected in the floor mirror.

Interpretation. Social learning seems particularly important for calibrating the odometer, that is, for translating the optic flow perceived during flight into the distance signalled in the dance. In this set-up, once established in a young forager, the distance-coding error did not correct itself over the 20-day follow-up. The authors speak of a coding fixed very early, even "for life", but this wording must be read in light of the actual duration of the experimental follow-up.

These results align with work on variations in dance calibration between species, populations or environments, sometimes described as dance "dialects" (Kohl et al., 2020; Schürch et al., 2019). They suggest a socially transmitted component in the calibration of the signal, without erasing the instinctive and sensory basis of the dance. For flight, the bee relies dominantly on ventral optic flow to maintain altitude: above a calm water surface, this flow may become insufficient, the trajectory tilts downward, and drowning may follow, as already observed by Heran and Lindauer (1963) over a lake.

3. Critical assessment

The results are stimulating for biological understanding, but their practical scope remains indirect.

Strengths. The Dong et al. (2023) design is robust: five experimental colonies and five control colonies, individually marked bees, longitudinal follow-up of the same dancers over time, and a clear separation between directional error, which corrects itself, and distance error, which persists over 20 days. Serres et al. (2022), for their part, finely manipulate the availability of optic flow with several mirror configurations and a control tunnel. The convergence of the conditions with a floor mirror toward loss of altitude strongly supports the ventral interpretation.

Methodological limits. For the dance, the experimental follow-up covers 20 days: speaking of an error "for life", as the authors do, therefore extends the observation beyond what the measurement directly shows. It would also be useful to verify whether the effect is found in different bee populations and at foraging distances other than the 150 m used here. For flight, the tunnel measures only 2.20 m: it demonstrates a mechanism but does not reproduce the conditions of an actual flight over a lake or pond, where wind, flight altitude, shoreline relief, vegetation and trajectory choice all come into play. In both cases, the experimental conditions are deliberately impoverished to isolate one variable, which is a methodological strength but a limit on transposition.

Possible biases and confounds. The experiments were conducted in highly controlled experimental contexts, in China for Dong et al. and in France for Serres et al. Transposition to the European bees commonly kept in Switzerland remains plausible, but it has not been directly tested. Kievits's article is, moreover, a popular-science piece: it mentions a follow-up of "10 days" after the first dances, whereas the original Dong et al. paper specifies 20 days. This nuance matters when citing the source precisely.

What cannot be concluded. These works do not show that a colony at the apiary will dance better or worse depending on the average age of the bees at a given moment. Nor do they say how often foragers actually drown at the apiary, nor which precise characteristics of a body of water—size, wind, vegetation, depth, floating supports—modify this risk. Finally, the social learning demonstrated does not, as it stands, allow any recommendation on the age composition of colonies or on queen management.

4. What related studies show

Fig. Figure-eight dance, the wavy lines symbolise the waggle (movement following the arrows) (Wikipedia)

A few independent works help situate these two studies within the existing literature.

An observational forerunner to Dong et al. Before the 2023 experimental study, Ai et al. (2017) had individually tracked young bees in normal colonies and found that they typically begin to follow dances about a week before producing their own, focusing on only a few sources. That work indicated that dance-following normally precedes the first dance, but could not demonstrate that the absence of a model degraded the signal. Dong et al. (2023) provide the missing experimental piece, by suppressing this prior exposure and measuring the consequences.

Distance coding normally readjusts with experience. Chatterjee et al. (2019) moved a feeder from one distance to another and observed that dancers gradually adjust their waggle duration, simultaneously using a recent and an earlier memory of the distance. This result shows that, under ordinary conditions, the translation of distance into waggle duration is not simply fixed. The deficit observed by Dong et al. (2023) in naive dancers therefore appears unusual and reinforces the idea of an early calibration linked to social experience.

Among normal populations, calibrations broadly converge. Schürch et al. (2019) compared the distance/waggle-duration curves of several Apis mellifera populations and found that they are close enough to allow a common dance calibration. This suggests that the lasting deficits observed by Dong et al. (2023) stem from an extreme experimental design, not from ordinary variability at the apiary.

Dance dialects remain a useful nuance. The work of Kohl et al. (2020) shows that the functions linking distance and waggle duration can vary across species and populations of bees, and that these variations may have an adaptive dimension. This literature supports the idea of a non-universal calibration of the signal, but it does not by itself prove a stable cultural transmission comparable to a human dialect.

Heran and Lindauer: the historical observation over a lake. As early as 1963, Heran and Lindauer had trained bees to fly 247 m over open water. The bees crossed the surface when it was rippled by the wind or marked by a floating bridge providing contrast, but progressively lost altitude and drowned when the water was perfectly calm. Serres et al. (2022) reproduce and quantify this observation in the laboratory, showing that the underlying mechanism is the disappearance or strong reduction of ventral optic flow.

A parsimonious mechanism confirmed by bio-inspired robotics. Franceschini et al. (2007) built a micro-helicopter piloted by a ventral optic-flow regulator. Placed in conditions analogous to a headwind or a flight over smooth water, the robot qualitatively reproduces the same kind of descending trajectory described in bees. This result supports the interpretation that a relatively simple visual mechanism is sufficient to explain what bees do under these conditions, without assuming complex reasoning about the nature of the surface.

5. What does this mean for the apiary?

At the apiary, these two studies above all invite us to better understand certain observed behaviours, without changing basic management.

• Do not regard the dance as a strict automatism. It becomes more precise through experience and through observing experienced foragers. This mainly reminds us that the normal functioning of a colony rests on interactions between bees of different ages and experiences, without turning this into a management directive.
• Take large water surfaces into account when siting an apiary. The risk described by Kievits concerns above all open, calm, smooth water surfaces poor in visual cues. For a new apiary, one should therefore avoid placing it directly on the edge of a large bare pond or a very smooth body of water. For watering points, this also suggests favouring shallow water, floating supports, stones, moss, wood or vegetated edges over a smooth, reflective surface.
• Recognise orientation flights. The circular flights of young bees in front of the hive are not swarming. Confusing them with swarming may lead to unnecessary interventions.
• Do not extrapolate the deficits in Dong et al. to an ordinary colony. In normal colonies, without experimental deprivation, dance calibrations broadly converge across Apis mellifera populations. The lasting deficits observed in the study stem from an extreme design, not from a common fragility at the apiary.
• Keep a practical but cautious reading. These results impose no change in beekeeping practice in French-speaking Switzerland; their main contribution is to help interpret what is observed at the apiary.

 

Read the original study

Article read: Kievits, J. (2023). Danser, ça s'apprend! La Santé de l'Abeille, no. 316, July–August 2023, pp. 53–62.

Main scientific studies discussed: Dong et al. (2023), on the social learning of the waggle dance, and Serres et al. (2022), on the role of ventral optic flow in altitude control during flight.

 

Further reading:

• The bee dance
• Behaviour and cognition: what a mini-brain teaches us
• Mini-brain, mega performances
• The senses housed in the bee's antennae
• What water for our bees?

 

References

Ai, H., Kobayashi, Y., Matake, T., Takahashi, S., Hashimoto, K., Maeda, S., & Tsuruta, N. (2017). Development of honeybee waggle dance and its differences between recruits and scouts. bioRxiv. https://doi.org/10.1101/179408

Biesmeijer, J. C., & Seeley, T. D. (2005). The use of waggle dance information by honey bees throughout their foraging careers. Behavioral Ecology and Sociobiology, 59(1), 133–142. https://doi.org/10.1007/s00265-005-0019-6

Chatterjee, A., George, E. A., Prabhudev, M. V., Basu, P., & Brockmann, A. (2019). Honey bees flexibly use two navigational memories when updating dance distance information. Journal of Experimental Biology, 222(11), jeb195099. https://doi.org/10.1242/jeb.195099

Dong, S., Lin, T., Nieh, J. C., & Tan, K. (2023). Social signal learning of the waggle dance in honey bees. Science, 379(6636), 1015–1018. https://doi.org/10.1126/science.ade1702

Franceschini, N., Ruffier, F., & Serres, J. (2007). A bio-inspired flying robot sheds light on insect piloting abilities. Current Biology, 17(4), 329–335. https://doi.org/10.1016/j.cub.2006.12.032

Heran, H., & Lindauer, M. (1963). Windkompensation und Seitenwindkorrektur der Bienen beim Flug über Wasser. Zeitschrift für vergleichende Physiologie, 47, 39–55. https://doi.org/10.1007/BF00342890

Kievits, J. (2023). Danser, ça s'apprend! La Santé de l'Abeille, no. 316, July–August 2023, pp. 53–62.

Kohl, P. L., Thulasi, N., Rutschmann, B., George, E. A., Steffan-Dewenter, I., & Brockmann, A. (2020). Adaptive evolution of honeybee dance dialects. Proceedings of the Royal Society B, 287, 20200190. https://doi.org/10.1098/rspb.2020.0190

Schürch, R., Zwirner, K., Yambrick, B. J., Pirault, T., Wilson, J. M., & Couvillon, M. J. (2019). Dismantling Babel: Creation of a universal calibration for honey bee waggle dance decoding. Animal Behaviour, 150, 139–145. https://doi.org/10.1016/j.anbehav.2019.01.016

Serres, J. R., Morice, A. H. P., Blary, C., Miot, R., Montagne, G., & Ruffier, F. (2022). Floor and ceiling mirror configurations to study altitude control in honeybees. Biology Letters, 18(3), 20210534. https://doi.org/10.1098/rsbl.2021.0534