There has been quite a bit of research on monitoring bee flight patterns, as this could help us learn more about potential negative impacts on bee activity, such as traffic or pesticide use. It might also help us learn more about how a bee is able to navigate, fly, and land upside down on a moving target, all with a brain the size of a pinhead.
One option for tracking bees is to glue an RFID chip on the back of each individual bee and then use a network of sensors to collect data. However, this approach requires a large network of sensors, so it is difficult to scale, as bees can easily cover an area of up to 80 square miles around a hive. So scientists looked for cheaper alternatives to monitor bee flights.
So instead of trying to track bees in flight, a team of researchers decided to try to learn about where the bees had been based on the communication between bees when they returned to the hive. When scout bees return to the hive after a successful mission in search of food, they share information about the location and quality of food with their forager colleagues. A forager bee’s lifespan is measured in flight miles, so it is important that they accurately balance flying time with food quality. Both the availability and quality of food will constantly change as different plants flower at different times, so the scout bees return to the hive to share the latest information. It sounds pretty straightforward, but this communication can’t rely on visual cues, as it is very dark inside the hive, and sound isn’t easy either, as there will be 50,000 other bees buzzing at the same time.
The main communication method involves a very specific “waggle dance”, where the bees do a sort of figure of 8, with a ‘waggle’ down the middle. The waggle bit is the most important, and involves gripping the honeycomb floor and shaking vigorously so the other bees can feel the vibrations through their feet. The direction of the waggle shows the direction of the food relative to the sun, and the duration of the waggle indicates the distance. The bees then loop round in a figure of 8, doing the same waggle in the middle each time.
Nearby bees detect the vibration and then join in and follow the moves, and only the ones that follow the moves closely are then able to find the food source. The dancing bee also shares food samples to allow its colleagues to evaluate the quality. It’s pretty amazing that this is even possible inside a hive of thousands of bees, but it’s even more impressive considering that this is all part of a complex voting system, as multiple scouts will be ‘dancing’ at the same time.
The dance-offs only take place in specific areas of the honeycomb, which are identified in advance with scent. Watching forager bees can decide to vote ‘for’ the dancer, by joining in and following the moves, or vote ‘against’ her. For example, if a forager has visited the food source and found it to be dangerous or to have run out of food, she might vote against any dancers advertising this food source by headbutting them mid-dance. It sounds a bit harsh, but it definitely gets the message across.
So the next question is, how can researchers make use of this dance?
In order to better understand the waggle dance, researchers put video cameras inside a bee hive and then used image and pattern recognition to track and classify different bee movements and to isolate the waggle dance. But just tracking the dance isn’t much use, as they need to ‘interpret’ the dance in order to work out what it means.
So first, they set up an artificially rich source of food in order to bias the bees in favour of this location. They then rely on the bees finding it and coming back to the hive to dance about it. They were then able to calibrate their analysis from the videos of the waggle dances by correlating them with the known food location. They used this to come up with a ‘waggle dance translator’, which could look at a dance and then translate that into the location being described.
The scientists also noticed that some bees seemed to be more accurate than others when describing the location. According to the researchers doing the study, it isn’t the model that is inaccurate, it is the bees themselves (if in doubt, blame the bees).
They proposed two main reasons for this :
- Some bees ignored the rich target food and went off to find other sources of food. This makes sense as a backup in case the food source runs out.
- Apparently its quite hard for bees to waggle with high precision, while clinging to honeycomb in a crowded hive. They tested this by getting the bees to dance on a piece of flat cloth instead, and they found the dances more accurately matched their model.
Models like these can be used to help gain a better understanding of the excellent natural navigation and flight control systems of bees, which could help with the design of next-generation autonomous drones. There are already some very cool tech companies, like Opteran, that are using a bio-mimicry approach to recreate insect brains using computers. Rather than following the more mainstream approach of using massive computing power to try to create artificial intelligence, they hope to mimic the simplicity of insect brains to deliver a more natural form of intelligence, which is a pretty cool idea.
