Interactions between complex multisensory signal components result in unexpected mate choice responses

Multimodal (multisensory) signalling is common in many species and often facilitates communication. How receivers integrate individual signal components of multisensory displays, especially with regard to variance in signal complexity, has received relatively little attention. In nature, male túngara frogs, Physalaemus pustulosus, produce multisensory courtship signals by vocalizing and presenting their inflating and deflating vocal sac as a visual cue. Males can produce a simple call (whine only) or a complex call (whine þ one or more chucks). In a series of two-choice experiments, we tested female preferences for variation in acoustic call complexity and amplitude (unimodal signals). We then tested preferences for the same calls when a dynamic robotic frog was added to one call, generating a multi- modal stimulus. Females preferred a complex call to a simple call; when both calls contained at least one chuck, additional numbers of chucks did not further increase attractiveness. When calls contained zero or one chuck, the visual stimulus of the robofrog increased call attractiveness. When calls contained multiple chucks, however, the visual component failed to enhance call attractiveness. Females also preferred higher amplitude calls and the addition of the visual component to a lower amplitude call did not alter this preference. At relatively small amplitude differences, however, the visual signal increased overall discrimination between the calls. These results indicate that the visual signal component does not provide simple enhancement of call attractiveness. Instead, females integrate multisensory components in a nonlinear fashion. The resulting perception and behavioural response to complex signals probably evolved in response to animals that communicate in noisy environments.