In the initial setup of Bateson’s experiment, a group of honeybees was trained to associate two simple odor mixtures with two different foods. One mixture, which consisted of one part hexanol to nine parts octanone, was repeatedly paired with sucrose, which bees find rewarding. The other odor mixture consisted of the same two chemicals in opposite proportions (nine parts hexanol to one part octanone) paired with quinine, a compound that most of us find bitter and bees will actively avoid after tasting. By using this technique, the researchers hoped to overcome the bees’ intrinsic responses to sucrose and quinine and test only their judgment of the new smells. After learning these odor-food associations, the bees responded as expected, uncoiling and extending their mouthparts in anticipation of food when the first odor mixture was presented and retracting them at offers of the second concoction.
This training allowed the scientists to study the bees’ decision making by then testing their mouthing responses to a series of ambiguous odor mixtures. First, half the bees got a trip to the “vortexer.” The experience was probably as unpleasant for them as it sounds to us. In a procedure meant to simulate a badger attack on a hive, the bees were shaken for one minute in a machine typically used to vigorously mix chemicals. If bees can indeed be made to feel cranky, surely this device would do the trick.
Next, both shaken and unshaken bees were tested on five mixtures of hexanol and octanone at different concentrations. Sure enough, both groups preferred extending their mouth to octanone-heavy mixtures, which predicted sugar, rather than hexanol-heavy mixtures, the scent of which predicted quinine. Interestingly, the shaken bees were less likely to advance toward any of the mixtures than their unperturbed counterparts.
In an analogue of the classic scenario of the half-empty glass versus the half-full glass, the bees were also presented with an equal mixture of hexanol and octanone. Bees that were spared the trip to the vortexer gave the concoction the benefit of the doubt, moving their mouth toward the food on close to 60 percent of the trials. Shaken bees, on the other hand, ignored or recoiled from these same ambiguous stimuli more than half the time. The stress of shaking had turned them into pessimists that interpreted the ambiguous odor as half threatening rather than half appetizing.
Both Shaken and Stirred
In addition to these behavioral measures, the scientists also tested for changes in the bees’ neurotransmitter levels after shaking. The quantities of certain chemicals with known roles in insect learning (octopamine), aversive conditioning (dopamine) and aggression (serotonin) were all reduced by the procedure, suggesting that as with their mammalian counterparts, duress in bees causes sustained, system-wide changes in brain state—a possible analogue of mood. Together these behavioral and neurochemical tests reveal an unexpected dimension of bee cognition. Formally, we can say that when agitated, bees can take on a negative disposition, a state that alters both their thinking and their neurochemistry.
For now, however, we cannot conclude anything more sweeping about the emotional life of a bee. Bateson and her co-authors leave us with an intriguing plea for consistency, however, one that nudges us to think clearly about how we regard the minds and emotions of all creatures. Last year researchers tested dogs that appeared to suffer from separation anxiety for a pessimistic bias. When they encountered an uncertain food reward, the perturbed dogs also appeared less inclined to try the ambiguous treat, which the researchers interpreted as evidence that dogs indeed feel anxious when left alone. “It is logically inconsistent,” Bateson and her colleagues say, to deduce that dogs and other similar animals express emotions “but to deny the same conclusion in the case of honeybees.”