Is discriminative responding in reward outcome serial learning mediated by item memories or by position cues?

In two experiments, rats in a runway were trained under fixed predictable series of 6 plain food pellets, 6 sucrose food pellets (both Noyes .045 g), and nonreward. Whether discriminative responding, faster running to reward than to nonreward, was controlled by cues associated with memories of earlier food items in the series or by cues associated with the position of the item in the series was investigated here. In Experiment 1 a series easy to learn from an item standpoint (item group), but not a position standpoint, was learned while a series having the opposite characteristics (position group) was not learned. Experiment 2 findings suggested that when item cues and position cues were equally relevant, item cues gained the greater control over responding and may have overshadowed position cues. Taken together, the acquisition findings of Experiments 1 and 2 cannot be explained by the specific spatial position hypothesis recommended by D'Amato (1991) or by position hypotheses in general. Following acquisition in Experiment 1, rats were shifted to series in which items in novel positions continued to be validly signaled by food items. The item groups responded discriminatively in shift, a finding consistent with the item view but not the position view. Cases of serial learning in which discriminative responding might be more strongly regulated by item cues than by position cues or vice-versa were considered in the General Discussion. (C) 1997 Academic Press.