Responding on random interval schedules with high rates of reinforcement

Four rats leverpressing for food were exposed to a series of random interval schedules that provided high rates of reinforcement. In three of the four rats, response rates decreased with increases in obtained reinforcement rates. In two subjects, response rates maintained by the two schedules with the highest reinforcement rates differed even though the obtained reinforcement rates were approximately equivalent. The results generally support theories which predict a bitonic relation between response rate and reinforcement rate. They also suggest that, in the development and testing of models, more attention needs to be paid to schedule feedback functions and the role of the scheduled reinforcement rate.