The adaptability of horses and donkeys to different types of activity has seen the European equine industry become an important economic sector, giving rise to increasing concern regarding equine welfare. As part of the AWIN (Animal Welfare Indicators) project, this review focuses on scientific literature to find potential animal-based welfare indicators — the initial step in developing a valid, reliable and feasible on-farm welfare assessment protocol for equines. Forty-nine indicators were considered and classified in accordance with the four Principles and twelve Criteria developed by Welfare Quality®. Only practical indicators specifically for on-farm use were included, those requiring the use of specific instruments or laboratory analysis were excluded. Academic scientists, partners and collaborators of the AWIN project, discussed and agreed on validity, reliability, on-farm feasibility and acceptance by farmers for each indicator. Some aspects of equine welfare have been thoroughly investigated and appear to have indicators ready for on-farm use (eg ‘absence of prolonged hunger’, ‘absence of injuries and diseases’). On the other hand, a lack of animal-based measures were identified for other Criteria such as ‘absence of pain’ and ‘positive emotional state’. Ongoing research within the AWIN project has begun exploring some of the aforementioned Criteria — these preliminary results of promising indicators have been included (eg Horse Grimace Scale and Qualitative Behaviour Assessment). Further research should address the validity and reliability of indicators, such as human-animal relationship tests and signs of cold stress. As well as for working equines, the development and application of a welfare assessment protocol could be the first step for enhancing on-farm equine welfare.

Conservation biology and animal welfare science are multidisciplinary fields of research that address social concerns about animals. Conservation biology focuses on wild animals, works at the level of populations, ecological systems and genetic types, and deals with threats to biodiversity and ecological integrity. Animal welfare science typically focuses on captive (often domestic) animals, works at the level of individuals and groups, and deals with threats to the animals’ health and quality of life. However, there are many areas of existing or potential overlap: (i) many real-life problems, such as environmental contamination, urban development and transportation, create problems for animals that involve both welfare and conservation; (ii) research methods from each field are needed to address some of the scientific problems of the other; and (iii) policies and practices targeting either conservation or animal welfare may prove unproductive if they do not take account of both areas of concern. Moreover, scientists in both fields face the common challenge of applying science to guide policy and practice, often to issues that are both empirical and ethical, and often under conditions of uncertainty. There are many cases where communication and co-operation between the fields should lead to better science and better practical outcomes.

All systems of scoring animal units (groups, farms, slaughter plants, etc) according to the level of the animals’ welfare are based inevitably on normative decisions. Similarly, all methods of labelling, in terms of acceptability, are based on choices reflecting ethical values. The evaluative dimension of scoring and labelling does not mean that we should reject them, but it does mean that we need to make the normative and ethical background explicit. The Welfare Quality® scoring system is used as a case study in order to highlight the role of underlying value-based decisions. In this scoring system, which was designed in accordance with assessments and judgments from experts in animal and social sciences and stakeholders, we identify value-based decisions at the following five levels. First, there are several definitions of animal welfare (eg hedonist, perfectionist, and preferentialist), and any welfare scoring system will reflect a focus upon one or other definition. In Welfare Quality®, 12 welfare criteria were defined, and the entire list of criteria was intended to cover relevant definitions of animal welfare. Second, two dimensions can structure an overall evaluation of animal welfare: the individual animals and the welfare criteria (here 12). Hence, a choice needs to be made between the aggregation of information at the individual level (which results in a proportion of animals from the unit in a good vs bad state) and the aggregation at criterion level (which results in a proportion of criteria to which the unit complies vs does not comply). Welfare Quality® opted for the second alternative to facilitate the provision of advice to farmers on solving the welfare problems associated with their farms. Third, one has to decide whether the overall welfare assessment should reflect the average state of the animals or give priority to worse-off animals. In the Welfare Quality® scoring system the worse-off animals are treated as much more important than the others, but all welfare problems, major or minor, count. Fourth, one has to decide whether good scores on certain criteria can compensate for bad scores on others. In the opinion of most people, welfare scores do not compensate each other. This was taken into account in the Welfare Quality® scoring system by using a specific operator instead of mere weighted sums. Finally, a scoring system may either reflect societal demands for high levels of welfare or be based on what can be achieved in practice — in other words, an absolute assessment or a relative one may be proposed. Welfare Quality® adopted an intermediate strategy: absolute limits between welfare categories (Not classified, Acceptable, Enhanced, or Excellent level of welfare) were set, but the rules governing the assignment of an animal unit to a category take into account what had been observed on European farms. The scientists behind Welfare Quality® are keen to make the value-based choices underlying assessments of animal welfare transparent. This is essential to allow stakeholder groups to understand the extent to which their views are acknowledged and acted upon.

Qualitative Behaviour Assessment (QBA) is part of the Welfare Quality® protocol for dairy cattle, although its inter- and intra-observer reliability have not been reported. This study evaluated inter- and intra-observer reliability of the QBA for dairy cattle in experienced and inexperienced observers using videos. Eight experienced observers performed the QBA (20 descriptors) twice for 16 video clips (60 s per clip; series 1) showing 4-17 animals. They assessed another 11 video clips showing herds (4 shots of 30 s per clip; series 2). Ten inexperienced observers performed the QBA on both video series one time. Inter-observer reliability of experienced observers ranged from slight to moderate (both assessments of series 1), and from low to high (series 2) for descriptors, and from slight to moderate for the QBA score. Inter-observer reliability of inexperienced observers ranged from low to moderate (series 1), and from low to high (series 2) for descriptors, and was moderate (both series) for the QBA score. Intra-observer correlations varied largely per descriptor and observer. They were both negative and positive, and ranged from low to very high. High correlations, however, were not necessarily associated with low paired differences. Values of half of the descriptors and the QBA score differed amongst experienced and inexperienced observers. The QBA appears insufficiently reliable as a tool for welfare assessment in dairy cattle.

Controlled atmosphere (gas) stunning (CAS) has the potential to improve the welfare of poultry at slaughter but there is a lack of consensus about which gas mixtures are most humane. The aim of this study was to evaluate the welfare consequences of different gas stunning approaches. Individual broilers were exposed to gas mixtures capable of inducing unconsciousness and euthanasia while their behavioural, cardiac, respiratory and neurophysiological responses were measured simultaneously. The approaches investigated included anoxia (N2 or Ar with < 2% residual O2), hypercapnic anoxia (30% CO2 in Ar, 40% CO2 in N2) and a biphasic method (40% CO2, 30% O2, 30% N2 for 60 s followed by 80% CO2 in air). Evaluation of the welfare implications of each approach centred on the likelihood of them inducing negative states or experiences during the conscious phase. Hypercapnic mixtures were associated with strong respiratory responses, while anoxic mixtures induced vigorous wing flapping. Electroencephalogram analysis using the correlation dimension (a non-linear measure of complexity) suggested that anoxic wing flapping occurred during periods in which a form of consciousness could not be excluded. Hypercapnic hyperoxygenation (biphasic approach) exacerbated respiratory responses but eliminated the possibility of vigorous behavioural responses occurring during a conscious phase. The relative importance of respiratory discomfort versus the potential to induce significant distress due to convulsive wing flapping and associated trauma is a matter for debate. We argue that respiratory discomfort is unpleasant but may be preferable to the risk of vigorous wing flapping and associated injury while conscious in poultry during CAS.

Since animal minds are private, so their perception of their own quality of life (QoL) must be also. Anthropocentrism, the interpretation of reality exclusively in terms of human values and experience, has to be guarded against in any assessment of animal welfare; for domestic pets, misapprehensions about their olfactory and cognitive abilities appear to present the greatest challenge to their welfare. Anthropomorphism, the attribution of human qualities to animals, presents a particular problem when considering companion animals, since most bonds between owners and their pets appear to be based upon a perception of the pet as almost human. Many owners report that their dogs, cats and horses are capable of feeling complex emotions, such as pride and guilt, that require a level of self-awareness that has been difficult to demonstrate even in chimpanzees. Such beliefs appear to contribute to the development of behavioural disorders in pets; for example, clinical experience suggests that the application of punishment by owners who attribute ‘guilt’ to their animals may unwittingly lead to compromised welfare. Anthropomorphic owners are also likely to be poor proxies for reporting their pets' QoL.

Behavioural management in zoos is often practiced within the confines of environmental enrichment, a well-known method which attempts to increase the welfare of captive animals. For the successful conservation and reintroduction of threatened or endangered species, however, it is also important to manage behaviour in such a way as to maintain behavioural diversity. The development of natural behaviour management (NBM) programs is advocated in this paper. These programs will act to maintain behavioural diversity in captivity and will encourage behaviour to be displayed in appropriate contexts through exposure of captive animals to naturalistic stimuli. The importance of developing appropriate antipredator and predation behaviours will be discussed in order to demonstrate how NBM strategies differ from, and can even conflict with, environmental enrichment strategies undertaken for welfare reasons.

There is an increasing requirement for tools to assess and improve animal welfare in an objective and scientifically based manner. In our research a concept of welfare is applied which states that welfare is determined by the balance between positive and negative experiences. This concept implies that an interaction exists between stress systems and reward systems in the brain and, as a consequence: (I) negative experiences induce an increased sensitivity (ie need) for positive experiences; and (II) negative experiences can be compensated for by positive experiences. On this basis, two uses of reward-related behaviour can be hypothesised: (I) reward sensitivity may be used as a tool to assess the state of an animal in terms of welfare because it can indicate the current state of the balance that is dependent on previous (stressful) experiences; and (II) regular presentation of rewards may serve as a tool to counteract stress by shifting the balance to the positive side and, thus, to improve welfare. In order to investigate this, we used the rat as a model. Reward sensitivity was determined by the spontaneous behavioural response shown during expectation of a reward (ie anticipatory behaviour). A third (III) use of reward-related behaviour derives from the fact that anticipatory behaviour is influenced by the (rewarding) properties of the forthcoming reward (or other event) and, thus, may serve as a tool to assess the animal's perception of this reward/event. This paper presents a descriptive overview of the evidence obtained thus far for the three proposed uses of reward-related behaviour. The biological background of our concept of welfare can be generalised to all (vertebrate) species, and anticipatory behaviour can be evoked in a wide range of other species. Therefore, this tool for measuring and improving the welfare of captive animals has great potential and will contribute to a good quality of life for captive animals.

The effect of individual differences on the behaviour and reproduction of zoo animals has long been recognised by zoo biologists, yet only recently has the quantitative assessment of personality been used to investigate some of the challenges faced in zoo animal management. We review the findings of animal personality studies carried out in zoos since 1995. Our results reveal that zoo animal personality is most commonly assessed using observer ratings, where people who are familiar with the animals are asked to rate them on various personality traits. The reviewed studies indicate that zoo keepers are able to reliably rate animal personality traits, and these ratings are valid and related to behaviour. We identify promising areas of development in zoo animal personality research and suggest applications of personality profiling to zoo animal welfare and management. We argue that a validated personality questionnaire is a valuable tool for zoo professionals and advocate the implementation of personality assessment into existing zoo management practices to inform decisions on welfare and captive breeding.

Equipping birds with tags (defined as any item externally attached to birds, including transmitters, loggers and flipper bands, or implanted devices, such as transponders) gives particular insights into animal biology, although researchers may not give systematic consideration of tag impact. We examined 357 papers published between 1986 and 2009 where tags (excluding rings attached to birds’ legs) were used on seabirds, to examine the extent to which researchers considered deleterious effects. Fifty-one (14.3%) papers considered instrumentation effects in their abstract, 31 (60.8%) of which showed statistically significant effects on seabird biology. Of the total data set, 302 (84.6%) articles were classified as ‘indirect’ (with no stated aim to assess the influence of the equipment used) and although most of these (237; 76.5%) did discuss instrumentation effects, this accounted for less than a mean of 2% of the total length of the text. Despite a clear increase in the number of papers based on tagging technology for seabird study over the previous 24 years, there has been no corresponding increase in documentation of the effects of devices on their bearers. We suggest mechanisms by which this issue might be addressed.

One million pet rabbits are kept in The Netherlands, but there are no data available on their behaviour and welfare. This study seeks to assess the welfare of pet rabbits in Dutch households and is a first step in the development of a welfare assessment system. In an internet survey, housing systems, general up-keep and behaviour of pet rabbits were reported by their owners. The answers of 912 respondents were analysed with behavioural observations carried out on 66 rabbits in as many households. The rabbits were observed in their home cage and during three fear-related tests: a contact test, a handling test and an open-field test. The survey revealed that the average lifespan of the rabbits is approximately 4.2 years (the maximum potential lifespan is 13 years) and solitary housing appears to reduce lifespan. Close to half of respondents subjected their rabbit(s) to solitary housing and the majority housed them in relatively small cages (< 5,000 cm2). Health risks may arise from a failure to inoculate rabbits and via inappropriate diet. During the contact test, solitary-housed rabbits made more contacts with a human than group-housed rabbits and rabbits in a small housing system made more contacts than those in a large system. Observations in the home cage differed greatly compared with the natural time budget of rabbits, ie displaying increased stereotypic behaviour and decreased foraging and, in solitary-housed rabbits, a complete lack of social behaviour. Nearly 25% of rabbits displayed strong resistance to being picked up, indicating socialisation problems. During the open-field test, solitary-housed rabbits sat up more than social-housed rabbits suggesting increased fearfulness. These findings indicate that the conditions in which pet rabbits are kept often have a negative impact on their welfare, further underlining the need to study this in greater detail.

People affect animals through four broad types of activity: (1) people keep companion, farm, laboratory and captive wild animals, often while using them for some purpose; (2) people cause deliberate harm to animals through activities such as slaughter, pest control, hunting, and toxicology testing; (3) people cause direct but unintended harm to animals through crop production, transportation, night-time lighting, and many other human activities; and (4) people harm animals indirectly by disturbing ecological systems and the processes of nature, for example by destroying habitat, introducing foreign species, and causing pollution and climate change. Each type of activity affects vast numbers of animals and raises different scientific and ethical challenges. In Type 1 activities (keeping animals), the challenge is to improve care, sometimes by finding options that benefit both people and animals. In Type 2 activities (deliberate harm), the challenge is to avoid compounding intentional harms with additional, unintended harms, such as animal suffering. For Type 3 and 4 activities, the challenges are to understand the unintended and indirect harms that people cause, to motivate people to recognise and avoid such harms, and to find less harmful ways of achieving human goals. With Type 4 activities, this may involve recognising commonalities between animal welfare, conservation and human well-being. Animal welfare science and animal ethics philosophy have traditionally focused on Type 1 and 2 activities. These fields need to include Type 3 and 4 activities, especially as they increase with human population growth.

Animal rescue shelters provide temporary housing for thousands of stray and abandoned dogs every year. Many of these animals fail to find new homes and are forced to spend long periods of time in kennels. This study examined the influence of the length of time spent in a rescue shelter (<1 month, 2-12 months, 1-5 years, >5 years) on the behaviour of 97 dogs. The dogs ‘ position in their kennels (front, back), their activity (moving, standing, sitting, resting, sleeping), and their vocalisation (barking, quiet, other) were recorded over a 4 h period at 10 min intervals. The dogs’ behaviour was significantly related to the length of time the animals had spent in the rescue shelter. Dogs housed in the shelter for over five years spent more of their time at the back of their kennels, more time resting, and less time barking than dogs housed in the shelter for shorter periods of time. The age of the dog could not account for the significant results found, suggesting that environmental factors were responsible for the change in the dogs’ behaviour. The findings suggest that lengthy periods of time spent in a captive environment may encourage dogs to behave in a manner that is generally considered unattractive by potential buyers. This may decrease the chances of such dogs being adopted, resulting in longer periods of time spent in the kennel environment and the possible development of further undesirable behaviours.

The inflammatory response evokes changes in behaviour including increased thermoregulatory activities and sleep, reduced social exploration and appetite, and altered food preferences. This sickness response also includes feelings of lethargy, depression, and pain, collectively referred to as ‘malaise’. Recent experiments involving laboratory rodents reveal information about proximate mechanisms of sickness behaviour, but scant information exists about how sickness behaviour is expressed by farmed species or within social environments. The behavioural needs of ill individuals differ from those of conspecifics, and failure to accommodate the needs of ill individuals may exacerbate suffering. Policy makers, industry and animal welfare certification programs recommend hospital pens to address the housing and handling needs of ill livestock and to reduce risks of disease transmission. However, a survey of swine farms in Ontario, Canada revealed deficiencies in the use of hospital pens and gaps in knowledge about best management practices for this vulnerable population. There is considerable scope to improve the welfare and husbandry of ill and at risk animals through effective use of hospital pens and supportive therapies.

Academics working on animal welfare typically consider the animal's affective state (eg the experience of pain), biological functioning (eg the presence of injuries), and sometimes naturalness (eg access to pasture), but it is unclear how these different factors are weighed in different cases. We argue that progress can be informed by systematically observing how ordinary people respond to scenarios designed to elicit varying, and potentially conflicting, types of concern. The evidence we review illustrates that people vary in how much weight they place on each of these three factors in their assessments of welfare in different cases; in some cases, concerns about the animal's affective state are predominant, and in other cases other concerns are more important. This evidence also suggests that people's assessments can also include factors (like the animal's relationship with its caregiver) that do not fit neatly within the dominant three-circles framework of affect, functioning and naturalness. We conclude that a more complete understanding of the multiple conceptions of animal welfare can be advanced by systematically exploring the views of non-specialists, including their responses to scenarios designed to elicit conflicting concerns.

Tail and ear biting lesions have a negative effect both on the animal welfare status of pigs (Sus scrofus) and the economical revenue of the pig farm. Tail biting behaviour is an unpredictable, abnormal behaviour that is thought to have a multifactorial origin. On-farm factors influencing tail biting have been described, but the real triggers are poorly understood. Much of the research into tail biting has been done on a small scale within a well-controlled environment and small sample sizes. This well-controlled environment is not always representative of the contemporary commercial conditions. Therefore, an observational epidemiological approach at farm level was adopted to gain a better insight into the factors influencing the occurrence of tail and ear biting lesions. Tail and ear biting lesions were observed at pen level three times a year on sixty farms across Belgium. A questionnaire was conducted to build a multifactorial model indicating different risk factors concerning the lesions scored. The temperature and the number of feeding places per animal in the nursery, the percentage of floor space covered with slats in the farrowing unit, the feed type in the growing unit and the overall hygiene policy were the most important indicators for the appearance of tail and ear biting lesions during fattening. The leave-one-out cross validation of the model demonstrated an intraclass correlation coefficient of 0.55 between the predicted model outcomes and the observed data. This epidemiological study provides important potential risk factors in relation to the incidence of tail and ear biting lesions. However, experimental and/or longitudinal studies have to confirm that the correlations found in this work are causal factors.

Previous research has shown that pre-parturient primiparous pigs (gilts) housed in behaviourally restrictive farrowing crates without straw redirect their nest-building behaviour to non-manipulable substrates such as the bars of the crate. These gilts also show elevated plasma adrenocorticotropic hormone (ACTH) and Cortisol levels, particularly around the peak of nest-building activity, when compared to gilts in larger pens that have been provided with a manipulable substrate (straw). It remains unclear whether these behavioural and physiological responses to crating result from the lack of a suitable nesting substrate or from the restricted space. This study investigated the effects of space (crate [C] versus pen [P] and straw (straw [S] versus no straw [NS]) using a 2 × 2 factorial design. Thirty-four gilts were implanted with an indwelling jugular catheter at around 12 days before parturition. They were moved to one of the four environments five days before parturition, and blood sampling and recording of behaviour were carried out during the pre-parturient period. Penned gilts (P), irrespective of straw availability, spent more time standing and walking and performed more total substrate-directed behaviour than crated (C) gilts. When straw was not available to penned gilts, a large proportion of their substrate-directed behaviour was redirected to the floor. Space also had an effect on ACTH and Cortisol levels across the entire pre-parturient phase, with C gilts having higher levels than P gilts irrespective of straw availability, but particularly so at the peak of nest-building activity. There was no effect of straw on ACTH or Cortisol levels. Overall, it appears that increased space, perhaps through allowing locomotion, increases substrate-directed behaviour of pre-parturient gilts. When space is available but straw is absent, pre-parturient gilts redirect their nest-building behaviour to the floor. The ability to express substrate-directed behaviour as a result of increased space is reflected in lower levels of indicators of physiological stress.

The behaviour of broilers reared at pasture from 4 to 12 weeks of age on a low density diet supplied either indoors or outdoors, was compared with that of birds reared inside on deep litter. This single study found few differences in behaviour. Activity levels of birds outside were initially greater, but from six weeks of age lying increased to comparable levels in all groups.

Surprisingly little use was made of the extra space and facilities such as perches at pasture. It is proposed that the main reason for this was leg weakness as 80 per cent of the birds had a detectable gait abnormality at seven weeks of age. There was no evidence of reduced motivation to extend the behavioural repertoire, as, for example, ground pecking remained at significantly higher levels in the outdoor groups because it could also be performed from a lying posture.

Appropriate space allowances for animals are yet to be specifically determined for lairage. Space allowances that may be suitable for animals in lairage are suggested, based on reviewed studies of animals in transport, lairage and on farm. The longer animals are in lairage the more space they require, in order to be able to get up and lie down and lie undisturbed by congeners. Little work has been done on air quality and air flow characteristics in lairages. The range of ventilation must be sufficient to control levels of toxic or irritant gases such as carbon dioxide and ammonia and to remove excess heat and humidity; the latter being particularly relevant for pig lairages in hot weather. Intensities of sound measured in lairages often exceed 85 dB and there is evidence to suggest that such levels can be stressful especially for pigs; and human shouting appears particularly aversive to animals. Cattle vocalise in response to painful stimuli and to convey information to conspecifics that may be related to fear and distress. There is limited evidence that sheep adapt to continuous sound, provided it is not too loud, but respond to intermittent sounds such as gates banging and human shouting. Vocal communication between sheep may be less important than that between cattle and pigs. Levels of vocalisation are potential indices of animal welfare. Animals' prior experiences and factors such as sex, group size and constitution, pen design, and climatic or environmental conditions affect their welfare and responses to conditions in lairage.

This study examined the effects of environmental enrichment on aggressive behaviour and dominance relationships in growing pigs. Three hundred and twenty pigs were reared from birth to 15 weeks of age in either barren or enriched environments. The barren environments were defined by common intensive housing conditions (ie with slatted floors and in terms of recommended space allowances), while the enriched environments incorporated extra space and substrates for manipulation. Aggressive behaviour was observed in a social confrontation test during the suckling period and dominance relationships were assessed from a food competition test at 12 weeks of age. Animals were weighed at regular intervals throughout the experiment. Environmental enrichment reduced the expression of aggressive behaviour. Pigs from enriched rearing environments fought significantly less with unfamiliar animals than those from barren environments when tested under standard conditions (mean of 1.46 vs 2.75 fights per 30min test for enriched vs barren environments; SEM 0.20, P < 0.001). The nature of dominance relationships also appeared to differ between barren and enriched environments. In barren environments, dominance among pen mates was correlated with aggression (r = 0.33, P < 0.01), whereas in enriched environments it was correlated with body weight (r = 0.24, P < 0.01). Correlations between behaviour in the social confrontation and food competition tests suggested that dominance characteristics were established early in life and remained stable through the growing period.