Visual search using realistic camouflage: countershading is highly effective at deterring search.

Abstract

One of the most widespread patterns of colouration in the animal kingdom is countershading, a gradation of colour in which body parts that face a higher light intensity are darker. Countershading may help counterbalance the shadowing created by directional light, and, hence, reduce 3D object recognition via shape-from-shading. There is evidence that other animals, as well as humans, derive information on shape from shading. Here, we assessed experimentally the effect of optimising countershading camouflage on detection speed and accuracy, to explore whether countershading needs to be fine-tuned to achieve crypsis. We used a computational 3D world that included ecologically realistic lighting patterns. We defined 3D scenes with elliptical ‘distractor’ leaves and an ellipsoid target object. The scenes were rendered with different types of illumination and the target objects were endowed with different levels of camouflage: none at all, a countershading pattern optimized for the light distribution of the scene and target orientation in space, or optimized for a different illuminant. Participants (N=12) were asked to detect the target 3D object in the scene as fast as possible. The results showed a very significant effect of countershading camouflage on detection rate and accuracy. The extent to which the countershading pattern departed from the optimal pattern for the actual lighting condition and orientation of the target object had a strong effect on detection performance. This experiment showed that appropriate countershading camouflage strongly interferes with visual search by decreasing detection rate and accuracy. A field predation experiment using birds, based on similar stimuli, showed similar effects. Taken together, this suggests that countershading obstructs efficient visual search across species and reduces visibility, thus enhancing survival in prey animals that adopt it.

Meeting abstract presented at VSS 2015