Fig. 1.

Schematic overview of the experiments conducted. A Experiment 1: Visual learning with free-flying bees. The experimental set-up was a rotating screen apparatus where hangers could be attached to at various locations. The hangers displayed the stimuli during conditioning and testing. For reversal learning we used yellow and greenish-yellow cardboard squares as visual stimuli. For negative patterning we used checkerboard squares cut from pink or blue cardboards to create the single stimuli C/D and blue and pink cardboards to create the compound stimulus CD. Test performances were assessed by a 45 s. choice tests during which all contacts of the bees with the respective stimuli were counted and a percentage of correct choices was then calculated for each test. B Experiment 2: Olfactory learning with free-flying bees. The experimental set-up was a Y-maze apparatus where the bees could fly through a hole to enter and get access to the inside where the olfactory stimuli were applied to filter papers on the backwalls of the maze. In reversal learning we used the odors linalool and 2-hexanone as A and B. For negative patterning we used limonene as stimulus C, 2-octanol as stimulus D and a mixture of these odors as CD. Test performances were assessed by a 20 choice tests during which all contacts of the bees with the respective stimuli were counted and a percentage of correct choices was then calculated for each test. C Experiment 3: Visual learning with restrained bees. The experimental set-up was a box with five chambers covered by movable red Plexiglas^® ceilings preventing light stimulation. In reversal learning we used 400 nm and 600 nm monochromatic light discs as stimulus A and B. For negative patterning we used a blue or green checkerboard as single stimuli C/D and a blue-green checkerboard as compound stimulus CD. Test performances were assessed by a single presentation of each stimulus of the respective learning task. D Experiment 4: Olfactory learning experiments with restrained bees. The experimental set-up was also a movable box with compartments for ten bees in front of an exhaust fan. Odors were delivered through an airstream. In reversal learning we used the odors linalool and 2-hexanone as A and B. For negative patterning we used limonene as stimulus C, 2-octanol as stimulus D and a mixture of these odors as CD. Test performances were assessed by a single presentation of each stimulus of the respective learning task (color figure online)