The characteristic striped appearance of zebras has provoked much speculation about its function and why the pattern has evolved, but experimental evidence is scarce. Here, we demonstrate that a zebra-striped horse model attracts far fewer horseflies (tabanids) than either homogeneous black, brown, grey or white equivalents. Such biting flies are prevalent across Africa and have considerable fitness impact on potential mammalian hosts. Besides brightness, one of the likely mechanisms underlying this protection is the polarization of reflected light from the host animal. We show that the attractiveness of striped patterns to tabanids is also reduced if only polarization modulations (parallel stripes with alternating orthogonal directions of polarization) occur in horizontal or vertical homogeneous grey surfaces. Tabanids have been shown to respond strongly to linearly polarized light, and we demonstrate here that the light and dark stripes of a zebra's coat reflect very different polarizations of light in a way that disrupts the attractiveness to tabanids. We show that the attractiveness to tabanids decreases with decreasing stripe width, and that stripes below a certain size are effective in not attracting tabanids. Further, we demonstrate that the stripe widths of zebra coats fall in a range where the striped pattern is most disruptive to tabanids. The striped coat patterns of several other large mammals may also function in reducing exposure to tabanids by similar mechanisms of differential brightness and polarization of reflected light. This work provides an experimentally supported explanation for the underlying mechanism leading to the selective advantage of a black-and-white striped coat pattern.