Sign language is an effective form of alternative communication for persons with autism and other developmental disabilities. Only a few studies have systematically measured the effects of sign language on the vocal responses of its users. This study employed a multiple baseline design to evaluate the effects of sign language on the vocal responses of a four-year-old boy with autism. Results indicate that a reinforcement contingency placed only on sign responses is inadequate for maintaining vocal responses. When a reinforcement contingency is placed on sign responses as well as vocal responses that the user is capable of emitting in verbal imitation, both sign and vocal responses are maintained. Results are discussed in terms of the need for a reinforcement contingency placed on vocal and sign responses, the effects of teaching procedures on response variability, and the need for future research to examine procedures utilized to teach sign language to persons within the developmental disabilities population.