Relational Learning In Honeybees (Apis Mellifera): Same-Different Discrimination

Abstract

Previous studies of learning in honeybees (Apis mellifera) have found that the basic learning phenomena are similar to vertebrate learning. Recent research is focused on more complex cognitive phenomena such as concept learning. These experiments explored the honeybees’ ability to learn about same-different relationships using two different discrimination tasks. Individual free-flying bees were trained to visit a laboratory window for reward, fly back to the hive to unload and return to the window. In all the experiments described here, bees were rewarded with sucrose for choosing correctly and were punished with stevia, an unpleasant solution for bees, for choosing incorrectly. Experiment 1 was an oddity discrimination study in which bees were presented with a unique set of three stimuli, two the same and one different, on each trial. Bees were rewarded for choosing the odd stimulus in every trial. The results showed that honeybees learned to choose the odd stimulus. Experiment 2 was a nonoddity discrimination study in which again bees were presented with a unique set of three stimuli, two the same and one different, on each trial. In this case, however, bees were rewarded for choosing either of the same stimuli. The results showed that bees did not learn to choose the nonodd stimuli, possibly because of an unlearned preference for oddity. Experiment 3 was a simple color discrimination study to test a new procedure. Bees were presented with two pairs of colored targets, one yellow and one blue. Behind each pair of targets was a clear block to hold the solutions. Half of the bees were rewarded for choosing the pair of yellow targets and the other half was rewarded for choosing the pair of blue targets. The bees were not able to solve the simple-color discrimination problem with the new training procedure. Experiment 4 was another simple-color discrimination study, which used exactly the same procedure as used in Experiment 3 but with the solutions placed directly on top of the targets instead of on the clear blocks. The bees were able to discriminate the colors easily using this modified procedure. The modified procedure was then used in Experiment 5 in a simultaneous same-different discrimination experiment. Bees were presented with two pairs of stimuli, one pair with the same targets and the other pair with different targets. For bees in the Same Group, choice of the same pair was rewarded while for bees in the Different Group, choice of the different pair was rewarded. The bees learned to choose correctly on the basis of the rewarded relationship, same or different. Experiment 6 was a replication of Experiment 5 but with a divider added between the pairs in the expectation that the spatial separation would increase the discriminability of the pairs. The results were very similar to those of Experiment 5; the bees learned to choose the correct pair. The results of Experiments 1, 5 and 6 provide evidence that honeybees can learn about “sameness” and “differentness.” In the vertebrate literature, learning about relationships often is interpreted as concept learning. These results provide strong evidence for relational or concept learning in honeybees. Further studies should explore this capability in other kinds of same-different tasks.