Study: Hummeln solve problems spontaneously – first observation in vertebrates

Researchers from the University of Oulu have managed to prove spontaneous problem solving at Hummeln. According to the authors, this is the first time that this ability has been detected in an invertebrates. The study by Bhambore et al. was published in the journal Science. It is part of a growing range of scientific reports on cognitive abilities in insects. She was led by Olli Loukola, behavioral ecologist at the University of Oulu, who has been researching cognition of Hummeln for over a decade. The experiment was divided into two phases. First, the animals learned two independent things: a polystyrene ball can be moved; a blue ring stands for lining. Both associations were trained separately – the connection between them had to establish the bumblebees themselves. Two associations, a new problem

In the actual test, ball and artificial flower were simultaneously in a Plexiglasarena. The flower hung on the ceiling, too high to fly. The nearby solution – rolling the ball under the flower, climbing on it, cassing food – had never seen the animals before. More than 70 percent of the bumblebees came to it. Whether this is a real problem-solving or random herumping, the researchers tried to clarify with several control conditions. In a variant with visual barriers, the flower was not visible to the bumblebees while moving the ball – they still rolled it targeted. In a further variant with three chambers connected in series, the animals had to maneuver the ball through several openings before they could reach the flower at all. Here too, an appreciable proportion solved the task. The only statistically significant behavioral indicator for success was how often a bumblebee had inspected the flower side of the arena. The authors value this as an indication of targeted action. Recommended editorial content

With your consent, an external YouTube video (Google Ireland Limited). I agree that external content is displayed to me. This allows personal data to third-party platforms (Google Ireland Limited). More in our privacy policy. Methodological core: No experience with the solution

What distinguishes the study from a large number of previous research is the methodological approach. Investigations on problem solving in animals – including chimpanzees, crows and parrots – have so far typically worked with animals who have already had extensive experience with similar tasks. They were confronted with puzzles and tool tasks over longer periods before the actual test situation followed. Whether the resulting behaviour can then be considered to be a real spontaneous problem solving is controversial in research. Loukola and his team tried to clear up this objection from the outset. The Hummeln had no previous experience with the specific task structure. Loukola explained to Science News: “Spontaneous problem solving is something that has not been detected in any invertebrates so far.” And further: “Our study is the first one where we can be 100% sure that these individuals have no experience with problem-solving tasks.”

Whether this can actually be regarded as evidence of cognitive flexibility in the narrower sense, or whether associative learning mechanisms are sufficient to explain, should be discussed in behavioral research. Part of a long series of research

The study is not alone. In recent years, Loukola's group at the University of Oulu has presented a series of findings on the cognition of Hummeln. In 2024, the group at Proceedings of the Royal Society B showed that Hummeln can solve problems cooperatively. Hummel couples learned to move a Lego block together or to press a door at the end of a transparent double tunnel – tasks that were only detachable by coordinated interaction. Behavior was noteworthy in the case of delays: if the partner remained, the animals waited longer with the pushing. In the tunnel experiment they changed their direction depending on the behavior of the partner. The authors interpreted this as an indication of socially influenced, possibly actively coordinated cooperation – but admitted that the question whether Hummeln actually understands the role of their partner needs further research. In addition, there are findings from other groups: bumblebees can therefore feel emotions, recognize faces, feel at the pace and learn from each other. Honey bees also show amazing cognitive performances: A new study on numerical cognition in bees suggests that they can really capture quantities – and not only react to visual patterns. The assumption that complex cognitive performances necessarily require large brains is thus at least no longer universally maintained. A more comprehensive insight into Loukola's research work is offered by a roughly one-hour lecture on YouTube, presenting the results of his team. (vza)