Employing a minimum 20 min intertrial interval (ITI) and extended acquisition training, it was found that a single alternation schedule of partial reward produced less resistance to extinction than an irregular schedule. These results support the sequential hypothesis and appear incompatible with the frustration hypothesis.

The 2X2 factorial design employed 48 rats in the runway. Manipulated were number of nonrewarded (N) trials preceding the rewarded trial (N-length, 3 vs. 1) and order of occurrence of the larger reward magnitude (.09 vs. .675 gm.), either before or after nonreward. The intertrial interval was IS min. Consistent with previous results under short intertrial intervals (ITI) (e.g., 30 sec.), but inconsistent with a previous finding under 15-min. ITI, resistance to extinction increased as N-length increased. Also found was that resistance to extinction was greater when large reward followed, rather than preceded, nonreward. The results were examined relative to several current hypotheses of the effects of partial reward at different ITIs.

Change in the stimulus produced by nonreward as a function of time! F or all rats half the daily acquisition intertrial intervals (III) were 8 min., half 30 sec. Reward followed nonreward at the 8 min. III in one group and at the 30 sec. III in the other. In extinction, trials occurred at a constant 8 min. III for half of each group, and at a constant 30 sec. III for the remaining half. Only the interaction between the acquisition nonreward-reward III and the extinction ITI was significant. These results suggest that the nonrewarded stimulus changes over time but that its response evoking power is neither increased nor decreased by the change.