The tiger moth Bertholdia trigona is the only animal in nature known to defend itself by jamming the sonar of its predators - bats. In this study we analyzed the three-dimensional flight paths and echolocation behavior of big brown bats (Eptesicus fuscus) attacking B. trigona in a flight room over seven consecutive nights to determine the acoustic mechanism of the sonar-jamming defense. Three mechanisms have been proposed: (1) the phantom echo hypothesis, which states that bats misinterpret moth clicks as echoes; (2) the ranging interference hypothesis, which states that moth clicks degrade the bats' precision in determining target distance; and (3) the masking hypothesis, which states that moth clicks mask the moth echoes entirely, making the moth temporarily invisible. On nights one and two of the experiment, the bats appeared startled by the clicks; however, on nights three through seven, the bats frequently missed their prey by a distance predicted by the ranging interference hypothesis (∼15-20 cm). Three-dimensional simulations show that bats did not avoid phantom targets, and the bats' ability to track clicking prey contradicts the predictions of the masking hypothesis. The moth clicks also forced the bats to reverse their stereotyped pattern of echolocation emissions during attack, even while bats continued pursuit of the moths. This likely further hinders the bats' ability to track prey. These results have implications for the evolution of sonar jamming in tiger moths, and we suggest evolutionary pathways by which sonar jamming may have evolved from other tiger moth defense mechanisms.