1. Did Real Honeybees Follow the Robot Bee's Dance?
Researchers from institutions including the Free University of Berlin in Germany conducted the experiment. They placed a robot named "RoboBee" inside a honeybee hive, where it danced the figure-eight dance just like a real bee [1].
The results showed that on 8 out of 13 days of the experiment, bees were observed following the robot’s dance. This happened a total of 80 times [1]. That means following behavior was seen on about 60% of the days.
When looking only at dances that were followed at least three times, the average number of abdominal waggles per dance was 7.19 for the robot and 5.45 for real dancers [1]. Dividing 7.19 by 5.45 gives about 1.3 times. However, this does not mean the robot was "better." The reason for the longer following is not clear from this result alone. This article will look at why researchers use machines and how the robot was built. First, what does the dance actually communicate?
2. What Does a Honeybee's Figure-Eight Dance Tell Other Bees?
Honeybees use dance to tell other bees about food, water, or nest locations. For places farther than 100 meters, the dance has two parts: the "waggle run," where the bee moves forward while shaking its body, and the "return run," where it circles back to the start [1]. Repeating these two parts creates a figure-eight shape.
Distance is communicated by the length of the waggle run. The longer the shake, the farther the location [2]. Direction is communicated on the vertical honeycomb surface. The bee treats straight up as the direction of the sun. The angle of the body from straight up shows the angle from the sun’s direction to the food source [2]. For nearby locations, the dance is said to become circular [2].
This meaning was decoded by Karl von Frisch (1886–1982) from Austria. He showed that the dance contains information about distance and direction. In 1973, he received the Nobel Prize in Physiology or Medicine along with Tinbergen and Lorenz [3].
So far, this is what humans have understood by watching from the outside. But in the darkness inside the hive, what do bees rely on to read the dance?
3. Why Is It Hard to Test One Cue in the Dark Hive?
Inside the hive, it is dark. Bees following the dance (dance followers) cannot see the dance with their eyes. They follow the movements of the dancing bee’s body and sometimes touch it with their antennae [1].
During the waggle, the thorax and wings vibrate. This vibration creates fine air movements, vibrations in the comb, air currents, and changes in electric fields. Other cues include smell, the taste of nectar, and the high body temperature of the dancer [1].
After decades of research, it is still not known which of these cues attract the following bees, motivate them, and show them the way [1]. Because it is difficult for humans to adjust a bee’s body one part at a time to test this, researchers thought of a method: replace the dancer with a machine that performs a fixed dance [1].
4. How Far Did the 1992 Mechanical Honeybee Experiment Go?
The first group to try a mechanical dancer was the research group led by Michelsen. In a 1992 paper, they reported that bee foraging increased near the location shown by the mechanical honeybee [1].
According to an explanatory article, the body was made of copper covered in wax. A small motor moved it in a figure-eight pattern. Electromagnets vibrated the wings to make sound, and a small pump released sugar water [4]. Although the transmission was not as precise as a real bee’s, communication was said to have occurred [4].
However, a 2018 paper states that this earlier study did not report dance-following behavior, and interactions inside the hive were not filmed [1]. "Bees moved," but "how they read the dance" was not visible.
So, how can we make those invisible parts visible?
5. How Was RoboBee Designed to Dance Like a Real Bee, But Quietly?
The base of RoboBee was a modified commercial pen plotter. It moves on a plane and can rotate. At the tip is a life-sized model of a honeybee. The model is soft sponge wrapped in thin film. Instead of wings, multi-layered plastic pieces are vibrated by a speaker. Sugar water is released in small drops from the head [1]. To give it the hive’s smell, the model was placed in the hive’s wire cage for several days before the experiment [1].
The waggle happens 13 times per second. To reduce noise, it is vibrated by a single rotation motor. The model is attached about 20 mm from the rotation axis, linking the angle of swing with the body’s lateral movement [1].
The return run was slowed by 40%, from the real bee’s 20 mm per second to 12 mm per second. This was to reduce the machine’s sound [1]. The drop from 20 to 12 is a decrease of 8; 8 divided by 20 is 0.4.
The dance’s design values were created from records of real dances signaling food about 230 meters away. In the experiment, bees were trained at two artificial food sources (places with sugar water) located 200 meters from the hive [1]. It was not a perfect copy, but a dance made as a compromise with "quietness."
6. Did Bees That Followed the Dance Fly in the Direction It Showed?
The flights of bees that followed the dance were recorded by harmonic radar. This method tracks the flight path by having the bees carry a small transponder (a small component that reacts to radio waves). The flight paths of all tracked flights matched the direction indicated by the robot. The paper interprets this as directional information having been transmitted [1]. However, of the 6 marked dance followers, radar records exist for only 4. The paper itself describes the number as "anecdotal" [1].
One individual flew to a location in the direction shown by the dance, which was different from where it had found food the day before. Another individual flew to the previous location [1]. These are examples of individual bees; it cannot be said that all bees do this.
Researchers noted challenges. Around a real dancer, many bees gather at the same time. With the robot, there were only 2 cases out of 80 where two bees read it simultaneously. That is 2 divided by 80, or 2.5%. Including the fact that bees gathered on only 8 of the 13 days, improvements to attract more bees are listed in the paper as "unresolved research tasks" [1].
In other words, the robot has not reached the point where it can "freely call bees." Still, it was a clear step toward investigating which signals work in the dark hive.
7. Can a Robot Communicate a Treasure's Location by Dance Alone?
The idea of imitating dance is also used in robot-to-robot cooperation. A 2022 study by the Indian Institute of Science and the University of Maryland investigated a method where robots communicate direction to other robots based on the orientation of shapes traced on the floor, and distance based on the time taken to trace [6]. The goal is cooperation in places where communication networks cannot be used [6].
Try a "Dance Treasure Hunt" at home. With a family member or friend, pretend to be honeybees. First, decide together what body orientation and shaking time represent. Then, one person hides a small object somewhere in the room. The hider shows the location by shaking their body without using words. Body orientation shows direction, and the length of shaking time shows distance. The other person looks and searches.
Usually, communication does not work without agreed-upon rules. It is good to decide first "what body orientation represents." Honeybees have this rule by treating straight up on the comb as the direction of the sun [2].
Considering that honeybees use many more signals simultaneously in the dark hive, you might feel the weight of the robot being followed. Next, try thinking about how to read the graphs and figures in the original paper.