This study is part of an Autism Centers of Excellence (ACE) Network. With the development of better diagnostic tools and a greater level of professional education, autism is being identified in two year olds and in even younger children, with the rationale that the earlier intervention begins, the better the outcomes may be. However, there are no published outcome data on intervention models or effectiveness for children who begin treatment by or before 24 months. This grant will support a multi-site randomized controlled trial of early intensive behavioral intervention (Early Start Denver Model - ESDM) compared to standard community treatment involving 108 toddlers with autism (18-24 months of age), 36 at each of three sites, to evaluate the efficacy of very early intervention for improving child outcomes after 12 and 24 months of treatment. The study will examine predictors, mediators, and moderators of outcomes for children in both groups, with variables involving social environmental risks and protective features, individual child risk or protective features, and biological risks and protective features.