Previous trials of secretin for the treatment of autism have utilized a single or double dose administered intravenously. This is a report of a double-blind, randomized, controlled crossover trial of transdermally applied secretin in 15 children diagnosed with autism or pervasive developmental delay. Secretin or placebo was applied daily, in ointment form, to the backs of the children in randomized, successive 4 week periods with an intermediate 6 week washout period. Behavioral outcomes were measured by parents and teachers using the Autism Treatment Evaluation Checklist. Overall, there were no statistically significant differences in speech, sociability, sensory, and health scores for treatment versus placebo periods. In addition, there were no differences in such scores for children with a history of diarrhea. Severity of autism was significantly greater at baseline in children receiving concomitant medications. Improvement in speech was found during the treatment phase of the trial (p = 0.0479 for secretin versus placebo) only in children not using other medications.