Oddity Learning In Honeybees with Pattern Stimuli

Abstract

The aim of this series of experiments was to explore the traditional oddity problem in honeybees using black and white geometric patterns. The oddity problem is one way to study same-different concept learning. In the basic task, a set of three stimuli is presented, two non-odd and one “odd,” with choice of the odd stimulus rewarded. The results of previous work using solid colors as the stimuli in a traditional oddity problem suggest that honeybees may focus their attention on the specific colors rather than the relationship between them. Oddity learning may be easier for honeybees with geometric patterns, which are less salient than colors. Freely-flying foraging honeybees were trained to visit a laboratory window for sucrose reward. Once they were returning reliably to the window, the oddity training began. Choice of the correct pattern was rewarded with sucrose and choice of an incorrect pattern was punished with stevia solution. In Experiment 1, bees were trained with two patterns, black and white stripes and black and white concentric circles, with the concentric circle pattern target larger than the stripe pattern target. For half of the bees, choice of the stripe pattern was correct and rewarded, and for the other half, choice of the concentric circle pattern was correct and rewarded. The bees learned to choose correctly, either on the basis of size or pattern. In Experiment 2, honeybees were trained with identically-sized targets, again, the stripe pattern and the concentric circle pattern. The bees showed good discrimination of the patterns. In Experiment 3, honeybees were trained in a traditional oddity task using the stripe and concentric circle patterns of Experiment 2. Bees were rewarded for choosing the odd stimulus from a set of three. In training, bees had trials with two concentric circle patterns and one stripe pattern intermixed with trials with two stripe patterns and one concentric circle pattern. The bees’ performance was variable, although overall it was better than chance, suggesting that bees can learn an oddity problem with geometric patterns. In iv Experiment 4, honeybees were again trained to discriminate two geometric patterns, a white concentric circle pattern on a black background, used in Experiments 1-3, in addition to a black concentric circle pattern on a white background. The bees were able to discriminate the two concentric circle patterns. In Experiment 5, honeybees were trained in another oddity task using the two circle patterns of Experiment 4. The overall performance was better than chance, despite its variability, providing further evidence of oddity learning with geometric patterns. Overall, these results for honeybees in traditional oddity problems with geometric patterns indicate possible difficulty for the bees to attend to the relationship between patterns rather than to the specific patterns. The geometric patterns used in this series of experiments can be used in other same-different problems with honeybees, such as matching-to-sample and simultaneous discrimination of same and different pairs of stimuli. These problems have been used to study relational learning in vertebrate species, and they may be useful in exploring the conditions that promote relational learning in honeybees.