Is preference a predictor of enrichment efficacy in Galapagos tortoises (Chelonoidis nigra)?

It is widely acknowledged that environmental enrichment plays an important role in promoting the welfare of captive animals. However, a topic of debate is whether an animal's preference for an enrichment strategy is any indicator of its efficacy. In addition, relatively few studies have evaluated environmental enrichment strategies for non-mammalian species. In the present study, we compared the results of an observational evaluation of enrichment efficacy with the results of a paired-stimulus preference assessment for three Galapagos tortoises. In the observational study, object enrichment (boomer balls and a free-flowing sprinkler) and keeper interactions (shell scrubbing and neck rubbing) were evaluated systematically for their effects on locomotion, species-typical behavior, aggressive and non-aggressive conspecific interactions, and enclosure. Preference assessments were subsequently conducted in which subjects could choose the enrichment strategy to be implemented. All subjects preferred keeper interactions consistently over object enrichment. Our results suggest that enrichment preference was a variable predictor of efficacy across enrichment species-typical behavior, activity levels, enclosure use, and aggressive and non-aggressive conspecific interactions strategies. Preference predicted efficacy for promoting species-typical behavior (1/3 subjects), activity levels (2/3 subjects), and enclosure use (2/3 subjects), but not conspecific interactions (0/3 subjects). The results suggest that preference may be an efficient predictor of enrichment efficacy when daily observational evaluations are not practical; however, the predictive utility of preference assessments may depend on the behavioral goal of the enrichment strategy. We discuss the need for future research examining the relationship between preference and enrichment efficacy-as well as other potential indicators of enrichment effectiveness-in captive animals.