Experimental and comparative evidence suggests that the striped coats of zebras deter biting fly attack, but the mechanisms by which flies fail to target black-and-white mammals are still opaque. Two hypotheses have been proposed: stripes might serve either to defeat polarotaxis or to obscure the form of the animal. To test these hypotheses, we systematically photographed free-living plains zebras in Africa. We found that black and white stripes both have moderate polarization signatures with a similar angle, though the degree (magnitude) of polarization in white stripes is lower. When we modeled the visibility of these signals from different distances, we found that polarization differences between stripes are invisible to flies more than 10 m away because they are averaged out by the flies' low visual resolution. At any distance, however, a positively polarotactic insect would have a distinct signal to guide its visual approach to a zebra because we found that polarization of light reflecting from zebras is higher than from surrounding dry grasses. We also found that the stripes themselves are visible to flies at somewhat greater distances (up to 20 m) than the polarization contrast between stripes. Together, these observations support hypotheses in which zebra stripes defeat visually guided orienting behavior in flies by a mechanism independent of polarotaxis.