Prey may use a variety of strategies to avoid predators, including behavioral modification and inducible defenses. For an inducible defense to evolve, one necessary component is that the defense must be costly; if it is not, then the defense should evolve into one that is always expressed (i.e., a constitutive defense). However, costs of inducible defenses have not been well-documented. Using predator exposure experiments in aquatic mesocosms, we demonstrate that dark tail coloration (or lack thereof) in Blanchard's Cricket Frog (Acris blanchardi) tadpoles is an inducible defense. Specifically, Acris tadpole tail spots in treatments with dragonfly predators (Anax sp.) were significantly larger when compared to treatments with fish predators (Lepomis macrochirus). However, tadpoles in control tanks (with no predators) had tail spots that were not significantly different in area from the Anax treatment. Therefore, this defense is unique among those known because the presence of fish induces the loss (not the appearance) of this morphology. Because Acris tadpoles express the tail spot in the absence of predation risk, this phenotype also does not appear to have any substantial allocation cost. We also document predation of Acris tadpoles by a fish predator and demonstrate reduction in movement and differential habitat use in the presence of fish predator cues. Under predation risk, tadpoles became less active and occupied the shallowest regions of their habitats. This combination of plastic morphological and behavioral defenses likely allows these frogs to successfully breed in a broad range of aquatic habitats with different assemblages of predators. While likely effective in reducing vulnerability to Anax attacks, the tail spot may increase vulnerability to fish. Our results suggest that the antagonistic effects of predator-specific inducible defenses may represent another type of cost relevant to the conditions under which inducible defenses are expected to evolve.