A positional coding mechanism in pigeons after learning multiple three-item lists.

Pigeons were trained on three three-item lists (List 1: A(1) --> B(1) --> C(1;) List 2: A(2) --> B(2) --> C(2;) List 3: A(3) --> B(3) --> C(3)). After sessions in which any one of the three lists could be presented on a trial, derived-maintained list and derived-changed list probe trials were introduced. The derived list probe trials were composed of three items, one drawn from each of the training lists. On derived-maintained probe trials, each item was in the same ordinal position it occupied during training (e.g., A(3) --> B(1) --> C(2)). On derived-changed probe trials, items that occupied the second and third position during training were exchanged (e.g., A(2) --> C(3) --> B(1)). The performance of subjects on derived-maintained probe trials was significantly above chance and no different to that observed on the training lists. In contrast, subjects' performance on derived-changed probe trials was significantly below chance. To our knowledge, this is the first study that has demonstrated pigeons are able to learn and retain multiple three-item lists. In addition, subjects' performance on the derived-maintained probe trials suggests that they acquire knowledge of each list item's ordinal position when learning multiple lists.