Many young children with autism have complex communication needs, and many do not develop functional speech. Deficits in reciprocal social interactions and restricted or repetitive play are also at the core of this developmental disorder.Communicative interactions involve reciprocity or give-and-take between each communication partner; thus, the success of using AAC is highly reliant on the skills of each partner (Kent- Walsh and McNaughton, 2005). AAC systems allow children the ability to communicate independently; however, to date, the majority of studies reporting AAC effectiveness involves adult partners and rarely includes peers without disabilities. The purpose of this project is to examine a social communication intervention for nonverbal or minimally verbal preschool children with autism that integrates peer-mediated (PM) approaches with alternative and augmentative communication (AAC) instruction, called PM-AAC. Given the empirical support for training peers to be responsive communication partners, and the reported effectiveness of using AAC to teach functional communication, integrating PM and AAC approaches is a logical and innovative next step to advance intervention research for this population. Approaches will target and improve the course of development in core features of autism - social reciprocity, communication, and play skills. We will examine the effects of PM-AAC on communication and social interactions between preschoolers with and without autism in inclusive settings, and how skills generalize and maintain in non-treatment settings. Effectiveness will be evaluated using a multiple-baseline (MB) comparative AB design using block randomization based on cognitive abilities to randomly assign 12 children with autism each year to a PM-AAC treatment group (n=6) or an AAC comparison group (n=6). The MB design will be replicated across a total of four cohorts of 12 participants with autism over four years (n=48 total), with repeated measures on functional communication collected for both treatment and comparison groups allowing for analysis of between group differences at the end of year 4. Each child with autism will have three trained peer partners, for a total of n=144 peers without disabilities. Trained peers will learn responsive social strategies from a published preschool program, and how to use a selected AAC system. Fidelity of teacher and peer implementation, quality of peer behaviors, and skill generalization also will be measured. A structured communication sample and a structured play sample will be used to evaluate secondary increases in communication complexity and developmental play behaviors with peers. Outcomes will be a peer-mediated AAC intervention developed to enhance outcomes in core developmental areas for nonverbal or minimally verbal children with autism, and that attends to peer characteristics. A manual for treatment implementation will be produced, as will a compilation of videos for parent and teacher training. This work will provide a strong basis for a subsequent RCT study comparing PM- AAC intervention approach to alternative communication interventions for children with autism.