Fig. 2. Bees used spatial contextual cues to differentiate between two feeding sites. Top shows the arrangement of the apparatus with training and test stimuli. In the left flight box, the bees' task is to discriminate a rewarded (+) blue-black checkerboard from a non-rewarded (-) yellow-black checkerboard, and in the right flight box to discriminate a uniform rewarded yellow panel (+) from a blue one (-). In the absence of colour, yellow is always light grey and blue is dark grey. Bottom shows the bees' performance in tests with a choice between the reward stimuli from the two boxes. In this, and the remaining figures, choices are plotted cumulatively (ordinate) against the sequence of tests or trials for the indicated condition (abscissa), with the choices pooled over the group of bees tested (three in this case). We plot the total score, the number of correct choices and the number of wrong choices. The top edge of the grey wedge shows the null hypothesis of random choice. Bees prefer the blue-black checkerboard over the yellow panel in the left box and reverse their preference in the right box.