Harp seals (Pagophilus groenlandicus) solve two-choice visual discriminations using learned object valence and a conditional cue

Abstract

Five harp seals (Pagophilus groenlandicus) experienced behavioural training on several learning tasks. In Experiment 1, the seals were trained on six 2-choice visual discrimination tasks. The ability of the seals to transfer learned object valence (i.e., S+ and S-) to new tasks involving 1 or 2 previously experienced stimuli was investigated. All seals learned to solve 2-choice discriminations and also performed significantly better on tasks involving 1 or 2 objects that had attained positive or negative valence from previous discriminative training than tasks involving 2 novel stimuli. This finding supports the hypothesis that harp seals can transfer learned object valence. Experiment 2 explored the capacity of harp seals to use the tank they were in as a conditional cue to solve a 2-choice visual discrimination reversal task. Seals that experienced a switch to a different tank that coincided with a reversal in reward contingencies showed significantly more improvement across five reversals than seals that did not experience a change in context. The results suggest that harp seals may be sensitive to spatial cues, which supports field observations of their ability to orient and navigate despite a lack of other cues (Kovacs, 1995). The findings are discussed in terms of harp seals' adaptations to the pack-ice environment, the constraints of the learning tasks, the nature of the subjects who were raised in captivity, and the number of subjects involved in the experiment.