Reinforcement is considered the most important and most basic tool for motivating and teaching children with autism. However, the identification of potent reinforcers is problematic in autism, and creates a barrier to implementing behavioural intervention approaches. The Premack Principle of reinforcement has been investigated as a model for identifying and arranging reinforcers in children with autism. An assumption of the principle is that highly preferred behaviours can reinforce less preferred behaviours. Empirical research has investigated the repetitive, obsessional and stereotypic behaviour of children with autism as reinforcers. These behaviours are also associated with a number of clinical and educational problems that warrant intervention. The first paper, a literature review, highlights the difficulty of identifying reinforcers in children with autism and reviews the application of the Premack Principle as an alternative approach to determining reinforcers. In particular, the paper discusses the use of obsessional and repetitive behaviour as reinforcers. The second paper seeks to explore a new method of harnessing and using obsessional behaviour as reinforcement in children. The key development of this study is that it uses the Premack Principle to determine reinforcers and contingencies for increasing instrumental activity. The results indicated that instrumental behaviour increased during the contingent condition across all children; however, this increase was not due to a contingent effect. A decrease in contingent behaviour was also observed across all children. Overall, the results predicted by the Premack Principle were not observed.