The function of zebra stripes has long puzzled biologists: contrasted and conspicuous colours are unusual in mammals. The puzzle appears solved: two lines of evidence indicate that they evolved as a protection against biting flies, the geographical coincidence of stripes and exposure to trypanosomiasis in Africa and field experiments showing flies struggling to navigate near zebras. A logical mechanistic explanation would be that stripes interfere with flies' analysis of the optic flow; however, both spatio-temporal aliasing and the aperture effect seem ruled out following recent experiments showing that randomly checked patterns also interfere with flies' capacity to navigate near zebras. No clear mechanistic hypothesis remains. Here, I model from first principles how flies assess their motion relative to stripes, from image forming to motion analysis. I show that, at short distances, flies would consistently misjudge the motion of a striped object and frequently and saliently misjudge the direction of movement of a randomised check pattern. The range of distances at which stripes the model predicts flies should be impaired is consistent with observations. The model shows that image formation is subject to spatial aliasing, preventing any form of motion analysis against a striped pattern at medium distances, while the motion computation of flies is subject to a second form of aliasing, which, although independent of the temporal resolution of flies, bears conceptual similarities to spatio-temporal aliasing. The findings highlight the necessity of accounting not only for processing and psychology but also for the optics of image formation when taking a perceptual perspective of animal colours and contrasts.
Read full abstract