Published June 4 In Science, The Study By Researchers At The University Of Oulu Found That Bumblebees Placed In Front Of An Out-Of-Reach Reward Could Roll A Foam Ball Underneath It, Climb Over The Ball, And Claim The Prize; Replicating Insight Behaviour Previously Documented Only In Great Apes, Elephants, And A Small Number Of Bird Species
According to NPR, EurekAlert, CNN, and a paper published June 4 in Science by Akshaye A. Bhambore, Olli J. Loukola and colleagues at the University of Oulu, Finland, a new study demonstrates that bumblebees are capable of spontaneous, goal-directed problem solving; the same cognitive behaviour that Wolfgang Köhler documented in chimpanzees over a century ago, and that researchers had long assumed required a brain of meaningful size.
The Köhler connection and what makes spontaneous problem solving special. In the early 1900s, the German psychologist Wolfgang Köhler conducted what became one of the most cited experiments in the history of animal cognition. He suspended a banana just out of reach of a chimpanzee and placed a pile of boxes and crates nearby.
Without training, without being shown the solution, and without trial and error, the chimpanzee surveyed the situation, stacked the boxes, climbed them, and grabbed the fruit. Köhler believed this was evidence of insight; an “aha” moment in which the animal assembled its knowledge of the world into a novel solution it had never been taught.
Insight of this kind is not the same as learned behaviour, conditioned response, or trial and error. It is qualitatively different from any of those things. An animal that solves a problem through insight has not been rewarded for the steps that led to the solution. It has assessed a situation it has never encountered, recognised that elements of its existing experience are relevant, and combined them in a new way to achieve a goal.
This is something human beings do so routinely that we do not think of it as remarkable. In the rest of the animal kingdom, it is genuinely rare. Before this study, spontaneous problem solving of this kind had been convincingly documented in only a handful of species: great apes, elephants, and certain bird species, including ravens, rooks, and New Caledonian crows. There is ongoing scientific debate about whether octopuses and some spiders should join that list.
What the researchers did and what the bees did. The experimental setup was elegant in its simplicity. Bumblebees were first trained to associate a blue artificial flower placed on the floor of an arena with a sugary reward. This training established the flower as a known reward signal. In the critical test phase, the researchers moved the flower to the ceiling of the transparent arena just out of reach. A plastic foam ball was placed on the arena floor.
The bees had never been trained on what to do with the ball. They had never been trained on the combination of ball and elevated flower. The elements had been introduced separately. What happened next is what makes the study remarkable: the bees spontaneously rolled the foam ball underneath the elevated flower, climbed on top of the ball, and used the additional height to reach the reward.
Stringent controls ruled out simpler explanations such as trial-and-error or direct visual guidance. The results suggest that spontaneous, goal-directed problem-solving can emerge in animals with brains vastly smaller than those of vertebrates traditionally studied in insight research.
The bumblebee brain contains approximately one million neurons. The human brain contains approximately 86 billion. By any conventional measure of neural architecture, the bumblebee has no business doing what it just did.
Loukola and the long arc of bee intelligence research. Lead researcher Olli Loukola, who now works as a Senior Researcher at the University of Turku, has spent years building the scientific case that bees are far more cognitively capable than their brain size would suggest.
His earlier work demonstrated that bumblebees can learn to use tools, that they learn these tools socially from each other by observation, and that they can even understand the role of a partner in cooperative tasks. Each of those findings was itself contested when it first appeared.
The new study extends that body of work into the most demanding cognitive territory yet. “We are not claiming that bees think like humans,” says Loukola. “But our findings show that miniature brains can generate flexible solutions to novel problems in ways we are only beginning to understand.”
Loukola connects the bees’ spontaneous problem-solving ability directly to the ecological pressures they face in the wild. “Today they might find flowers from here, but tomorrow those flowers are not blooming anymore,” says Loukola. “If the workers can flexibly find new ways to get food for the colony, that’s the skill that they need to have.”
The cognitive flexibility the study documents is not an academic curiosity; it is a foraging survival mechanism operating in an insect whose entire life is structured around finding and retrieving food from a world that changes seasonally, daily, and hourly.
What the experts outside the study say. Cat Hobaiter, a primatologist at the University of St. Andrews who was not involved in the research, said: “We had this underlying assumption that somehow bigger brains means more powerful computations. And so demonstrating this in the bumblebees is really wonderful.”
Hobaiter says the study does a good job replicating similar experiments conducted on animals across the animal kingdom. “Intelligent brains come in really diverse shapes and sizes,” she concluded.
Hobaiter’s observation points to something that the study is, in a quiet way, dismantling. The assumption that large brains are necessary for sophisticated cognition is not merely a folk belief; it has shaped decades of comparative animal cognition research, determining which species were studied, which were considered capable of genuine cognitive complexity, and which were dismissed as operating purely on instinct. The bumblebee study adds to a growing body of evidence suggesting that assumption has been setting the bar in the wrong place.
What comes next. Loukola wants to examine the bees’ body movements, microgestures, and grooming behaviours to see if the insects have a tell preceding their moment of insight. One day, it may even be possible to image the bumblebee brain while it is solving a problem.
These are not incremental refinements to an existing research programme; they are genuinely new questions about what is happening inside a brain the size of a poppy seed at the moment it generates a novel solution to a novel problem.
Loukola reflects on how far the field has moved: “When I started, the cognitive limit was somewhere here,” he says, indicating a low point with his hand. “And now it’s much higher. We have to be smarter to develop or design experimental setups where we can test their real limits.” He’s not sure what those limits are, but he knows he hasn’t reached them yet.
The verdict. A bumblebee with one million neurons just did something that a century of animal cognition research assumed required a vertebrate brain. It did not learn the solution. It was not trained on the combination of elements that produced the solution. It assessed a situation, identified a movable object as relevant, and used it to reach an out-of-reach reward.
The researchers are careful not to claim human-like consciousness or reasoning. But the behaviour on display is the same category of behaviour that Köhler believed demonstrated genuine insight in chimpanzees and it is happening in an animal the size of a thumbnail.
The question this raises is not whether bumblebees are intelligent in the way that chimps are intelligent. The question is what intelligence actually is, whether it requires the architecture we have always assumed it does, and how many other small-brained animals are doing things we have never thought to test them for.
To check out our previous coverage on animal cognition, ecology, and scientific research, read our articles here.
