The current study sought to evaluate the efficacy of a stimulus equivalence training procedure in establishing auditory-tactile-visual stimulus classes with 2 children with autism and developmental delays. Participants were exposed to vocal-tactile (A-B) and tactile-picture (B-C) conditional discrimination training and were tested for the emergence of vocal-picture (A-C) and picture-vocal (C-A) responses. The results demonstrated that, following training, both participants responded successfully on both the training stimulus arrangements and the test probes that were never trained, illustrating emergence of cross-modal transitive and equivalence relations.