Abstract
Responses in the first component of a chained schedule produced a change to the terminal component according to a fixed-interval schedule. The number of responses emitted in the fixed interval determined whether a variable-interval schedule of food presentation or extinction prevailed in the terminal component. In one condition, the variable-interval schedule was in effect only if the number of responses during the fixed interval was less than that specified; in another condition, the number of responses had to exceed that specified. The number of responses emitted in the fixed interval did not shift markedly in the direction required for food presentation. Instead, responding often tended to change in the opposite direction. Such an effect indicated that differential food presentation did not modify the reference behavior in accord with the requirement, but it was consistent with other data on fixed-interval schedule performance. Behavior in the terminal component, however, did reveal sensitivity to the relation between total responses emitted in the fixed interval and the availability of food. Response rate in the terminal component was a function of the proximity of the response number emitted in the fixed interval to that required for food presentation. Thus, response number served as a discriminative stimulus controlling subsequent performance.