Joanna Newbolt


The behaviour of zoo animals can come to anticipate temporally predictable feeding times. However, there is a lack of consensus over the effects of such routines on behaviour and welfare. Few studies have been published in this area, perhaps in part due to methodological challenges. The current programme of work therefore aimed to extend knowledge on the effect of predictable feeding routines, and to develop a suitable methodology to overcome challenges. Prior to predictable feeding, tigers showed patterns of increased ‘active’ behaviours such as locomotion and pacing, and decreased inactivity. Geckos also showed increased ‘active’ behaviours such as locomotion and ‘stationary but moving head’ prior to predictable, compared to unpredictable feeding times, with significant differences between conditions (RSS = 0.059 and 0.047, p < .05). However, anticipatory patterns were not demonstrated for meerkats, which highlighted that other environmental factors may affect an animal’s response to temporal predictability (such as obtaining other food during the day, or signalled predictability). Studying patterns of behavioural change over time requires long periods of observation which is often not possible for researchers. The current programme of work argues that a measure which can be reliably used by many, relatively untrained observers is necessary to study predictability. The measure of ‘busyness’, a subjective rating of animal behaviour, was tested for reliability and validity. Busyness ratings showed good inter-observer reliability (ICC > .72) and correlated with traditional measures of behaviour. Busyness ratings demonstrated clear patterns related to feeding time and gave a useful compound measure of behavioural change. The use of multiple observers was extended to a citizen science approach, where useful data on anticipation in fish were obtained from aquarium visitors using a touch screen. The current programme of work successfully investigated the effects of predictable feeding routines on patterns of animal behaviour, alongside the development of suitable methods. The qualitative techniques developed here offer potential to increase the data obtained in future research into predictability and many other topics.

Document Type


Publication Date