Keas rely on social information in a tool use task but abandon it in favour of overt exploration.

Academic Journal

Interaction Studies; 2011, Vol. 12 Issue 2, p304-323, 20p

To what extent do keas, Nestor notabilis, learn from each other? We tested eighteen captive keas, New Zealand parrots, in a tool use task involving visual feature discrimination and social learning. The keas were presented with two adjacent tubes, each containing a physically distinct baited platform. One platform could be collapsed by insertion of a block into the tube to release the bait; the other platform could not be collapsed. In contrast to birds that acted on their own (“individual learners”), birds that could observe a demonstrator bird operated the collapsible platform first. However, they soon changed their behaviour to inserting blocks indiscriminately in either tube. When we reversed the collapsibility of the platforms, only adult observers but neither their demonstrators that had individually learnt nor the juveniles immediately altered their former preference. Observers, however, did not simply reverse their initial preference but rather moved to and then stayed at a chance performance as to where to insert a block first. In conclusion, the keas' overt exploration soon overrode the effect of social learning. We argue that such behaviour might help keas to find more efficient extractive foraging techniques in their native variable, low-risk environment. Keywords: tool use behaviour; social learning; reversal learning; physical cognition [ABSTRACT FROM AUTHOR]