For years, cognitive scientist Lars Chittka felt a bit eclipsed by his colleagues at Queen Mary University of London. Their studies of apes, crows, and parrots were constantly revealing how smart these animals were. He worked on bees, and at the time, almost everyone assumed that the insects acted on instinct, not intelligence. "So there was a challenge for me: Could we get our small-brained bees to solve tasks that would impress a bird cognition researcher?" he recalls. Now, it seems he has succeeded at last.
Chittka’s team has shown that bumble bees can not only learn to pull a string to retrieve a reward, but they can also learn this trick from other bees, even though they have no experience with such a task in nature. The study “successfully challenges the notion that 'big brains' are necessary" for new skills to spread, says Christian Rutz, an evolutionary ecologist who studies bird cognition at the University of St. Andrews in the United Kingdom.
Many researchers have used string pulling to assess the smarts of animals, particularly birds and apes. So Chittka and his colleagues set up a low clear plastic table barely tall enough to lay three flat artificial blue flowers underneath. Each flower contained a well of sugar water in the center and had a string attached that extended beyond the table's boundaries. The only way the bumble bee could get the sugar water was to pull the flower out from under the table by tugging on the string.
The team put 110 bumble bees, one at a time, next to the table to see what they would do. Some tugged at the strings and gave up, but two actually kept at it until they retrieved the sugar water. In another series of experiments, the researchers trained the bees by first placing the flower next to the bee and then moving it ever farther under the table. More than half of the 40 bees tested learned what to do, Chittka and his colleagues report this week in PLOS Biology.
Next, the researchers placed untrained bees behind a clear plastic wall so they could see the other bees retrieving the sugar water. More than 60% of the insects that watched knew to pull the string when it was their turn. In another experiment, scientists put bees that knew how to pull the string back into their colony and a majority of the colony's workers picked up string pulling by watching one trained bee do it when it left the colony in search of food. The bees usually learned this trick after watching the trained bee five times, and sometimes even after one observation. Even after the trained bee died, string pulling continued to spread among the colony’s younger workers.
But pulling a string does not quite qualify as tool use, because it would have to be an independent object that wasn’t attached to the flower in the first place. And other invertebrates have shown they can use tools: Digger wasps pick up small stones and use them to pack down their burrow entrances, for example. But that two bees figured out how to pull the string with no help and further, that other bees could pick up on that ability was “most impressive,” says Ivo Jacobs, a cognitive zoologist at Lund University in Sweden who was not involved with the work. “The fact that bumble bees could learn to do so shows their unexpected behavioral flexibility.”
The findings could also hint at a rudimentary form of culture in bees, Jacobs says. With their ability to learn where others are, find out what they are doing, and experimenting on their own, the insects demonstrated that they can pass on knowledge—a key requirement of culture, normally considered to be a more complex phenomena, he explains. “It's interesting to see that the bees have this capacity."
Rutz is impressed, too, because the work involved almost 300 bees and clearly documented how string pulling spread from bee to bee in multiple colonies. Cognitive studies of vertebrates like birds and monkeys typically involve about an order of magnitude fewer individuals, he notes.
With additional experiments, Chittka hopes to figure out the neural basis of these “smarts” in the bumble bees. He cautions that the insects might not be all that intelligent, but that instead, “these results may mean that culturelike phenomena might actually be based on relatively simple mechanisms.”