Sperry (1961) and others working on split brain learning reported that the two halves function in a nearly identical manner, giving excellent controlled replication of a learning problem. The present experiments were carried out with cats trained in a tactile discrimination apparatus, sometimes being forced to use the left paw, and sometimes being forced to use the right paw. The data were analyzed in terms of parameters of the Audley-Jonckheere (1956) learning model and also in terms of the conventional measures of total trials, total correct responses and total errors to criterion. Varying results have been reported in the literature. Possibly there are some conditions under which first and second brain learning results agree (replications), others under which they are related randomly and still others under which there is transfer from one half brain to the other. If limited conditions can be clearly defined under which replication will occur, then an investigation of learning curve theories through replication by brain experiments would be possible. (Author/CK)