At first glance, the idea that bees could be capable of doing math might seem fairly ridiculous. But according to a team of Australian-based researchers, these flying insects are actually smart enough to perform basic mathematics.

Through the Maze

So how did this group, employed by Melbourne’s RMIT University, come to the conclusion that bees can solve math problems? The team started their research process by shepherding honey bees inside a Y-shaped maze. Upon entering the maze, the bees were presented with between one and five shapes.

Blue shapes meant that the bees had to perform addition by following a certain pathway. In contrast, yellow shapes were required the bees to perform subtraction, and again required them to follow a particular path. One branch of the Y-shaped maze represented the correct choice, while the opposite branch represented the wrong decision. After properly navigating this maze, the bees were provided with a reward of sugary water.  Conversely, the bees were given a bitter-tasting quinine solution if they chose the incorrect path.

As you might expect, the idea of adding and subtracting numbers took a while for the bees to grasp. When the study began, the bees simply made random choices. Through much trial and error (involving over 100 learning trials spanning several hundred hours) the flying test subjects finally realized that blue meant +1, whereas seeing a yellow shape required the bees to subtract by the amount of one.

To further determine if the bees had learned basic addition and subtraction, the authors flipped the right and wrong sides of the maze midway through the study. Even with the unexpected switch-up, the bees still made the correct choice.

Learning from Bees?

In light of the report’s conclusions, senior author Adrian Dyer stated that “our bees also used their short-term memories to solve arithmetic problems, as they learned to recognize plus or minus as abstract concepts rather than being given visual aids. Our findings suggest that advanced numerical cognition may be found much more widely in nature among non-human animals than previously suspected.”

Given these findings, Dyer further noted that his team’s work could help with the further development of artificial intelligence. “If math doesn’t require a massive brain, there might also be new ways for us to incorporate interactions of both long-term rules and working memory into designs to improve rapid AI learning of new problems.”