Some of the main characteristics of second-order schedules were investigated.
In Section I, the component schedules were (drl 10-sec) and (drl 20-sec). These were held constant and the parameters of the overall schedule were varied. In
several important respects, the component schedule functioned in the same way as a simple operant is known to function with regard to the overall schedule. However, varying the value of the overall schedule also resulted in systematic changes in the response patterning within the component. Three studies were presented in Section II in which the effects of the parameters of the component
schedule on overall schedule control were investigated.
In one experiment , the lever presses which did not count as responses were found to have an important function on a DRL 20-sec schedule. The second
experiment examined the direct effect on fixed-interval control of changing the component schedule from (drl 10-sec) to (drl 20-sec) and vice versa. The final
report is of the effect of changing from DRL 5-sec or DRL 10-sec to a DRL 30-sec (drl x -sec) schedule. In all these cases, the nature of the operant was an important factor in determining the overall schedule control.
In Section III, five experiments into the effects of brief paired (S^p) and non - paired (S^N) stimuli were presented. Typical DRL patterning was evident under
the tandem schedules, and this was improved on adding S^N. The temporal distribution was sharpened even further on adding S^P. More detailed analysis revealed that S^N and S^P acted primarily on different aspects of the
behaviour. The main effect of adding S^N was to sharpen the temporal distribution of those responses following the completion of a non-reinforced component, whereas the effect of changing to S^p was to increase the postreinforcement pause. These effects are accounted for by a view which holds that the controlling effect of a stimulus event is increased by increasing the number of stimulus
elements present.