This paper summarises recent findings on the causation of stereotypic behaviours and other abnormal repetitive behaviours (ARBs) in captive animals: primarily motivational frustration and/or brain dysfunction, with possible contributory roles also being played by habit-formation and 'coping' effects. We then review the extent to which ARBs occur in zoos and similar, estimating that at least 10 000 captive wild animals are affected worldwide. We argue for 'zero tolerance' of such ARBs, because stress and poor welfare raise ethical issues, while abnormal behavioural phenotypes and possibilities of impaired brain development challenge both the indirect (e.g. educational) and the direct, intrinsic conservation value of affected animals. We then consider five potential means by which ARBs may be tackled: genetic selection; pharmacological treatment; the reinforcement of alternative behaviours; punishment; and environmental enrichment. All except punishment have potentially useful roles to play, but enrichment is the preferred approach: it is most likely to tackle the problems underlying stereotypic behaviours, and thence to improve both welfare and behaviour with few unwanted side-effects. Nevertheless, in zoos, environmental enrichment to date has only had partial success, with no study managing to abolish ARBs in all its subjects-suggesting either that the enrichments currently being used are never quite optimal, or that by the time they are tackled, ARBs have become resistant to change. We suggest some ways in which the effectiveness of enrichments may be enhanced; propose that certain properties of ARBs may usefully help evaluate their likely 'treatability'; and emphasise that if improving welfare is more important than just reducing ARB, then additional measures are needed in order to first, reliably identify those individuals most at risk from poor welfare, and then, to fully evaluate the welfare impact of enrichments. This paper also emphasises, with examples, the enormous potential value of zoo-derived data for helping understand

No abstract

In captivity, some species often seem to thrive, while others are often prone to breeding problems, poor health, and repetitive stereotypic behaviour. Within carnivores, for instance, the brown bear, American mink and snow leopard typically adapt well to captivity and show few signs of poor welfare, while the clouded leopard and polar bear are generally hard to breed successfully and/or to prevent from performing abnormal behaviour. Understanding the fundamental source of such differences could enable reproductive success and behavioural normalcy to be improved in zoos and breeding centres, by increasing the appropriateness of the enclosure designs and environmental enrichments offered particular species, and by allowing these to be offered pre-emptively instead of reactively. Here, we demonstrate that a significant proportion of the variation in apparent welfare between captive carnivore species stems from specific aspects of natural behaviour. We tested pre-existing hypotheses that species-typical welfare is predicted by natural hunting behaviour, general activity levels, ranging, or territorial patrolling (all activities that are constrained in captivity), by collating data on median stereotypy levels and infant mortality for multiple captive species, and then regressing these against median values for the relevant aspects of natural behavioural biology (e.g. hunts per day, proportion of flesh in the diet, home-range size, etc.). Our results revealed that instead of relating to foraging (e.g. hunting), as often assumed, carnivore stereotypy levels are significantly predicted by natural ranging behaviour (e.g. home-range size and typical daily travel distances). Furthermore, naturally wide-ranging lifestyles also predicted relatively high captive infant mortality rates. These results suggest that enclosure designs and enrichments focussing on carnivores' ranging tendencies (e.g. providing more space, multiple den sites, greater day-to-day environmental variability/novelty, and/or more control over exposure to aversive or rewarding stimuli) could be particularly effective means of improving welfare; and also, that targeting such enrichment programmes on wide-ranging species, before problems even emerge, might effectively pre-empt their development. Alternatively, species with relatively small ranges could instead be made the focus of future collections and breeding programmes, zoos phasing out wideranging carnivores in favour of those species inherently more suited to current or readily achievable enclosure sizes and enrichment regimes. #

14Keeping elephants in zoos is extremely costly, yet does not yield self-sustaining 15 populations. In Europe, which holds c. half the global zoo elephant population, a long-16 term decline of c.10% per year is expected in both species, if reliant on zoo-bred animals 17 under historically prevailing conditions. Fitness in zoos is compromised in several ways. 29The welfare implications of captivity, and the relative cost and effectiveness of ex versus 30 in situ conservation, determine the value of captive breeding for any given species. 31Several species apparently thrive in zoo conditions (1), and captive breeding has saved 32 some from extinction (2, 3). However, ex situ conservation is typically costlier than in situ 46$53 million per year in maintenance and, in the last decade, over $540 million in facility 47 upgrades worldwide (see Table S1