Antagonistic color effects in spatial vision of honeybees

Summary

• 1. Combined spatial and color vision is studied by training honey bees to vertically oriented circular patterns consisting of two half circles of different brightness or hue. Spatial discrimination is quantified by determining the discrimination between the trained pattern and a test pattern with a different inclination of the contrast line between the two half circles.

• 2. Bees discriminate patterns of pure hue contrast. The spatial discrimination of a pattern of optimal hue contrast is better than the discrimination of the equivalent black and white pattern.

• 3. After training to a contrast line of α= −45 ° (Fig. 1), discrimination depends not only on the change of inclination of the contrast line (Δα, Fig. 1), but also on the direction of change (−Δα: clockwise rotated test pattern, +Δα: anticlockwise rotated test pattern). This is called the asymmetry effect. The asymmetry effect is a sensitive measure of the color specific action on spatial discrimination.

• 4. In monochromatic UV bright/dark patterns discrimination is better, when the tested pattern displays an increased area of bright UV in the upper visual field. The effect is reversed in monochromatic orange bright/dark patterns; discrimination is better when the tested pattern displays a decreased bright area in the upper visual field (Figs. 3, 4).

• 5. Achromatic (including UV) bright/dark patterns induce no asymmetry effect. Bright/dark patterns without UV show the asymmetry of the orange pattern. This confirms Wehner's (1972) findings.

• 6. The color-specific effects are enhanced in heterochromatic patterns with optimal hue contrast (UV/bluegreen, blue/orange; Fig. 5a, b).

• 7. The color specific effects depend on the inclination of the contrast line of the trained pattern. Asymmetry is reversed for both color pairs when α=90 ° (a vertical contrast line) is trained instead of α=45 °. There is no asymmetry effect if a horizontal contrast line (α=0 °) is trained.

• 8. We interpret our results on three different levels. First, we use Wehner's (1972) model which assumes higher weight of the lower part of the visual field in pattern vision. We conclude, however, that this applies only for long wavelength colors. In UV light the upper part of the visual field should have higher weight. Second, we assume that bees prefer spontaneously patterns with a higher proportion of UV in the upper visual field. This explains the higher choice performance if the trained pattern has a higher proportion of UV in the upper part of the visual field, and the lower choice performance if the alternative pattern has a larger UV area in the upper visual field. Third, we compare our results with recordings from visual interneurons and show that our psychophysical data are useful for interpreting interneuron recordings.