The function of nonvocal sounds in avian communication has rarely been investigated. I used playback of natural and digitally altered wing sounds from both sexes of Calliope Hummingbird (Stellula calliope) and Rufous Hummingbird (Selasphorus rufus) to test whether these stimuli altered their behavior. Stimuli played from behind target birds at artificial feeders were used to study responses that characterized the level of threat associated with each wing-sound stimulus. I classified the behaviors into one of five response categories and analyzed the data using log-linear (Poisson) models. I found that the wing-trill signal of males strongly altered the behavior of both conspecific and heterospecific recipients. Females responded more strongly than males to male stimuli, and stronger responses were elicited by male Rufous Hummingbird than by male Calliope Hummingbird stimuli. I used digitally altered wing-sound stimuli to investigate further which sound components are used in communication. The results, in conjunction with previous research that has shown fitness benefits to the sender of the signal, lead me to conclude that male wing trill is an important component of hummingbird communication. Female wing sounds elicited stronger responses than control sounds, but the base wing-hum sound itself appears to be a simple by-product of flight rather than a communication signal. These results of the first playback study using nonvocal avian sounds suggest that further research to identify which nonvocal sounds are communication signals is needed.