Animal camouflage is a textbook example of natural selection. Despite substantial progress, one historical theory remains controversial: that conspicuous “distractive” markings draw predator attention away from the prey outline, preventing detection. Here, we present evidence from 4 experiments to resolve this controversy. In field experiments, we measured bird predation on artificial cryptic prey that were either unmarked or had distractive markings of various attributes (number, color, and location). Prey with 3 high-contrast distractive markings, and with markings located away from the body outline, suffered reduced survival compared with unmarked controls or prey with low-contrast markings. There was no effect of small single markings with different colors on the survival of targets. In 2 computer-based experiments with human subjects searching for hidden targets, distractive markings of various types (number, size, and location) reduced detection times compared with controls. This effect was greatest for targets that had large or 3 markings. In addition, small and centrally placed markings facilitated faster learning. Therefore, these 2 experimental approaches show that distractive markings are detrimental to camouflage, both facilitating initial detection and increasing the speed of predator learning. Our experiments also suggest that learning of camouflaged prey is dependent on the type of camouflage present. Contrary to current and historical discussion, conspicuous markings are more likely to impair camouflage than enhance it, presenting important implications for the optimization of prey coloration in general.