Captive jays exhibit reduced problem-solving performance compared to wild conspecifics.

Animal cognitive abilities are frequently quantified in strictly controlled settings, with laboratory-reared subjects. Results from these studies have merit for clarifying proximate mechanisms of performance and the potential upper limits of certain cognitive abilities. Researchers often assume that performance on laboratory-based assessments accurately represents the abilities of wild conspecifics, but this is infrequently tested. In this experiment, we quantified the performance of wild and captive corvid subjects on an extractive foraging task. We found that performance was not equivalent, and wild subjects were faster at problem-solving to extract the food reward. By contrast, there was no difference in the time it took for captive and wild solvers to repeat the behaviour to get additional food rewards (learning speed). Our findings differ from the few other studies that have statistically compared wild and captive performance on assessments of problem-solving and learning. This indicates that without explicitly testing it, we cannot assume that captive animal performance on experimental tasks can be generalized to the species as a whole. To better understand the causes and consequences of a variety of animal cognitive abilities, we should measure performance in the social and physical environment in which the ability in question evolved.