The effects of brief stimuli intermittently paired with primary reinforcement were compared to the effects of brief stimuli never paired with reinforcement. Water-deprived rats were given water reinforcement under schedules consisting of two successive fixed-interval (FI) components, A lever-press following completion of the first-component time interval began the second-component time interval, after which a press on the same lever produced 5-sec. access to water. Component intervals were units of 1-min., 2-min., or 3-min., presented in combinations of two which totalled 4-min.: (FI1,FI3), (FI2,FI2), (FI3,FI1). These three schedules were each presented under three further procedures: no brief stimuli, paired brief stimuli, and non-paired brief stimuli. In the paired conditions, a brief tone-and-light combination occurred contingent on the response completing each component, thus accompanying water after the second component. In the non-paired conditions, this stimulus followed only the first component. In addition, in paired and non-paired conditions, after a measure of stable behavior had been obtained, the brief stimulus following the first component was occasionally omitted.
Under the non-paired procedure, two subjects showed a clear post-stimulus pause and within-component patterning in the (FI2,FI2) condition. There was less indication of these effects in other non-paired conditions and in other subjects. In at least two of the paired conditions, all subjects showed an occasional surge in responding immediately following the stimulus after the first component. There was little indication of enhancement of either first-component or overall response rate in the paired conditions. In some cases under each brief stimulus procedure, responding was strongest and patterning most marked under the (FI2,FI2) schedule. The association of the non-paired stimulus vrith a period of non-reinforcement is thought to be responsible for the non-paired stimulus effects. The paired stimulus effects are somewhat inconsistent with some earlier studies of conditioned reinforcement, but demonstrate the ability of the pairing operation to modify the properties of a stimulus. When component schedules are homogeneous, as they were here in the (FI2,FI2) conditions, the increased resemblance of the two components may increase within-component patterning under both paired and nonpaired procedures.