The degree of domestication of an animal is difficult to estimate because the animal's phenotype depends not only on its genetic make-up but also on its experiences during ontogeny. In addition, comparisons between wild and domestic animals suffer from the lack of a proper reference population. In this study, we compared the offspring of silver foxes (Vulpes vulpes) that had been selected specifically for tameness for more than 30 generations (NOVO) with normal Finnish farmed foxes (FIN) and with reciprocal hybrids between these two strains. All animals were housed under standard farm conditions without any extra handling. The NOVO foxes had higher domestication indexes and lower fearfulness scores than the FIN foxes, with their hybrids showing values in between. Almost all NOVO foxes started eating in the presence of a human and took a titbit from him, whereas only a few FIN foxes did so. Open field and urinary cortisol tests failed to discriminate between the genotypes. The NOVO cubs had lower levels of serum cortisol both before and after a stressful situation or ACTH injection and showed lower stress-induced hyperthermia than the FIN cubs, with the hybrids showing results in between. NOVO foxes can be used as a reference when developing and validating behavioural tests for use in the selection of less fearful silver foxes. The observed low fear levels and low stress hormone levels suggest better welfare in the NOVO foxes under standard farm conditions.

This study compares various morphometric features of two strains of broilers, selected and ‘relaxed’ (ie random-bred), raised under two feeding regimes, ad-libitum-fed and restricted-fed. We consider the possible consequences of the different body shapes on the musculoskeletal system. The ad-libitum-fed selected birds reached heavier bodyweights at younger ages, had wider girths, and developed large amounts of breast muscle which probably displaced their centre of gravity cranially. At cull weight, they had shorter legs than birds in the other groups and greater thigh-muscle masses; therefore, greater forces would have to be exerted by shorter lever arms in order to move the body. The tarsometatarsi were broader, providing increased resistance to greater loads, but the bones had a lower calcium and phosphorus content, which would theoretically make them weaker. Many of these morphological changes are likely to have detrimental effects on the musculoskeletal system and therefore compromise the walking ability and welfare of the birds.

Rat and mouse control methods potentially affect the welfare of many millions of animals every year. Here, the humaneness of the methods used in the UK and the USA is assessed in terms of their speed and mode of action, the appearance and behaviour of affected animals, experiences of human victims, long-term effects on animals that survive exposure, and welfare risks to non-target animals. Several methods emerge as relatively humane: cyanide, alpha-chloralose, electrocution traps and well-designed snap traps all usually kill swiftly and with little distress. Preventative methods such as rodent-proofing are also humane, as well as an essential — and probably under-used — component of effective control. However, anticoagulant poisons, the most common means of controlling rodents, generally take several days to kill, during which time they cause distress, disability and/or pain. Sub-lethally affected animals are also likely to experience haemorrhages and their sequelae, and carnivores feeding on affected rodents may be secondarily poisoned. The acute rodenticides zinc phosphide and calciferol are also generally inhumane, the former typically causing severe pain for several hours, and the latter, pain and illness for several days. Sticky boards, to which rodents become adhered by the feet and fur until they are killed or simply eventually die, also raise very serious welfare concerns. This evidence highlights remarkable paradoxes in the way society treats different classes of animal, and argues for more education, legislation and research targeted at reducing the vast numbers of rodents currently killed inhumanely.

Twelve Isa Brown hens (Gallus gallus domesticus) were trained to open a locked door for access to a pen containing an enclosed nest box (‘nest test’) and to return to a home pen containing food, water, litter and a perch (‘home test’). The door was connected to a computer-controlled load cell, which recorded work exerted on the door and unlocked the door when the hen had exceeded a predetermined workload. Following training, the workload was set at 10 Ns, and hens received one nest test per day at 80, 60, 40 or 20 min prior to oviposition, and then one home test per day after 1, 2, 3 or 4 h confinement in the nest pen. As oviposition approached, hens showed a higher work-rate for access to the nest pen, showed a shorter latency to use the nest box and spent a greater proportion of their visit time in the nest box. Hens also worked harder for the home pen, showed a lower latency to feed and spent more time feeding after their return as period of confinement increased. The hens’ work-rate for the nest pen at 40 min prior to oviposition was comparable with their work-rate for the home pen after 4 h confinement, while their work-rate was at its highest in nest tests at 20 min prior to oviposition. The technique appears to be a valid means of assessing the importance of environmental resources, the values of which vary with time. The results suggest that hens place a higher value on gaining access to a discrete nest-site prior to oviposition than they do on gaining access to food following 4 h food deprivation.

This study tests the hypothesis that growth rate and bodyweight affect walking ability in broilers by comparing objective measurements of the spatial and temporal gait parameters of several groups of birds. Two strains of birds were used (relaxed and selected), raised on two feeding regimes (ad-libitum and restricted), and culled at the same final bodyweight (commercial cull weight of 2.4 kg). The ad-libitum-fed selected birds walked more slowly, with lower cadences, and took shorter steps. The steps were wider, and the toes were pointed outwards, resulting in a wider walking base. They kept their feet in contact with the ground for longer periods, having longer percentage stance times, shorter percentage swing times and increased double-contact times compared to the relaxed birds. These changes serve to increase stability during walking and are a likely consequence of the morphological changes in the selected broiler — in particular, the rapid growth of breast muscle moving the centre of gravity forward, and the relatively short legs compared to their bodyweight (see Corr et al, pp 145-157, this issue). This altered gait would be very inefficient and would rapidly tire the birds, and could help to explain the low level of activity seen in the modern broiler.

When thwarted in a behaviour, laying hens show an increase in stereotyped pacing, displacement preening and a specific vocalisation known as the ‘gakel-call’. How these behaviours, which might serve as indicators of welfare, are influenced by social factors is not yet known. In this study, we investigated the effect of an audience (another bird or a human) on the expression of the gakel-call and other behaviours indicating frustration. Twenty-four Lohman Brown hens were trained to gain free access to food in a test cage. Sixteen hens were used as test birds and eight as non-test audience birds. The food-deprived test hens were tested for 15 min in a non-thwarting situation (food freely available) and for 15 min in a thwarting situation (food covered but visible). For both situations we investigated four different treatments: no audience in the adjacent cage; a non-thwarted audience bird in the adjacent cage; a thwarted audience bird in the adjacent cage; and finally a human audience. The durations of stereotyped pacing and displacement preening were significantly higher in test birds during thwarting than during non-thwarting; thwarted birds also gave significantly more gakel-calls compared to non-thwarted birds. The test birds, and also the audience birds, gave more gakel-calls when thwarted in the presence of a thwarted conspecific than when in the presence of a non-thwarted bird, but there were no significant differences in stereotyped pacing or displacement preening, which are usually associated with frustration. In conclusion, this study supports the view that the gakel-call signals frustration in laying hens. Furthermore, the state of the audience influences the occurrence of gakel-calls in thwarted hens. Thus, when using the gakel-call as a welfare-indicator, the social aspects of the vocal expression of frustration in laying hens should not be overlooked.

In the social debate about animal welfare we can identify three different views about how animals should be raised and how their welfare should be judged: (1) the view that animals should be raised under conditions that promote good biological functioning in the sense of health, growth and reproduction, (2) the view that animals should be raised in ways that minimise suffering and promote contentment, and (3) the view that animals should be allowed to lead relatively natural lives. When attempting to assess animal welfare, different scientists select different criteria, reflecting one or more of these value-dependent views. Even when ostensibly covering all three views, scientists may differ in what they treat as inherently important versus only instrumentally important, and their selection of variables may be further influenced by a desire to use measures that are scientifically respected and can be scored objectively. Value assumptions may also enter animal welfare assessment at the farm and group level (1) when empirical data provide insufficient guidance on important issues, (2) when we need to weigh conflicting interests of different animals, and (3) when we need to weigh conflicting evidence from different variables. Although value assumptions cannot be eliminated from animal welfare assessment, they can be made more explicit as the first step in creating animal welfare assessment tools. Different value assumptions could lead to different welfare assessment tools, each claiming validity within a given set of assumptions.

Cats living long-term (over one month) in shelters were assessed for behavioural indicators of stress, using a stress scoring method in combination with behavioural observation. It is hypothesised that because of the inappropriate social grouping of unrelated adult cats and group instability, communal housing creates more stress than discrete-unit housing. Seventy-two cats were observed: 36 were housed communally with unfamiliar conspecifics, and 36 were housed in discrete units, either alone or with other previously familiar conspecifics. The mean stress score was greater in communal housing than in discrete-unit housing. Stress scores range from 1 to 7, with 1 indicating no stress experienced, and 7 indicating extreme stress. Individual scores showed that cats in discrete units, in comparison to those in communal housing, gained a significantly higher percentage of observations in the score 2 category, indicating that no stress was being experienced. Cats in communal housing gained a significantly higher percentage in the score 4 category (stressed). Score 5 was found exclusively in communal housing, but only in 2% of instances. Extreme stress was not found in cats housed under either condition. Cats in the different types of housing differed in their frequencies of hiding, play, sleeping/resting in close contact with one another, and agonistic behaviour. There was no difference between housing types in frequency of eating, drinking, grooming, and toilet use. In this study, cats housed communally experienced moderately higher levels of stress than cats housed in discrete units. Further research is recommended to determine the effect on stress levels of longer shelter residence time and of changes in group size and/or density.

Animal welfare is of increasing significance for European consumers and citizens. Previously, agricultural production focused mainly on supply, price and competition but consumers now expect their food to be produced and processed with greater respect for the welfare of the animals. Food quality is therefore determined by the welfare status of the animals from which it was produced as well as the nature and safety of the end product. Thus, practical welfare improvement strategies and reliable on-farm monitoring systems for assessing the animals’ welfare status and evaluating potential risks are urgently required to accommodate societal concerns and market demands. It is also of paramount importance to define the kind of information that consumers want about the final products and to develop effective strategies for communicating welfare standards to the public. Generating an intensified dialogue with all factions of society on welfare issues as well as appropriate labelling of animal products and farming systems that offer guarantees about welfare issues and production conditions will, in turn, promote transparency and the societal sustainability of European agriculture. Welfare is multidimensional. It cannot be measured directly but only inferred from external parameters. Therefore, the integration of the most appropriate specialist expertise in Europe is essential to develop, refine, standardise and intercalibrate welfare monitoring systems and to identify and validate remedial measures. We must establish a European standard for welfare assessment systems in order to facilitate intra-European trade and marketing. Only then can we harmonise labelling that is informative and relevant to all European consumers.

Laboratory primates are often housed in same-sex pairs to avoid single-housing and when breeding is to be prevented. However, pair formation is not without risks, as fights and injuries may occur. No data are available on pair formation in female common marmosets (Callithrix jacchus), a species used extensively in laboratories. Therefore, this study focuses on the pairing of unfamiliar common marmoset females, aiming to assess its success rate and whether age can predict the result. Data on the study animals and success of the pairings were extracted from laboratory back-records: a total of 28 pairings was obtained. In addition, behavioural data were collected on six of the 28 pairs. Almost 80% of pairs were compatible beyond one week, and most of the fights occurred well within the first week after pair formation. Pairs in which one of the females was sexually immature (ie < 15 months) were significantly more compatible than pairs in which both females were post-pubertal. First encounters were characterised by sniffing of the unfamiliar monkey. Aggressive behaviours occurred frequently following pair formation but they were unidirectional, and in only two pairs was veterinary treatment required. This study shows that pairing of unfamiliar common marmoset females is a safe practice if one monkey is sexually immature, a result that supports observations of both group and pair formation in other primate species. However, given the potentially detrimental effects of removing young females from their natal groups, we argue that it is preferable to remove two sisters from their natal group when female pairs are required. However, when a single sexually mature female requires a pair mate so as to avoid single-housing and no mature sibling is available, an older, but still sexually immature, unfamiliar female that has had a normal development within the family should be considered as a pair mate.

Various tools have been developed over previous years to study the welfare of laboratory animals. These include preference tests, which are commonly used to evaluate housing environments. Preference tests, however, have some pitfalls: they supply information only on the animals’ present preferences, and they allow the animal the choice only between the options offered. Other methods based upon the collection of clinico-chemical data require handling of the animals, which can be stressful in itself. An alternative may be to use telemetry to measure the changes in physiological parameters caused by different environmental conditions. The aim of this study was to use telemetry to evaluate the short-term impact of housing conditions on rodents. We monitored heart rate, blood pressure and body temperature in rats kept on three different types of flooring — bedding, grid floors and plastic floors. The study revealed significant differences in systolic and diastolic blood pressure, heart rate and body temperature between rats housed in the three conditions, indicating that both grid floors and plastic floors are more stressful for the animals than bedding. The observed differences did not diminish over the two-week observation period. The grid-floor housing induced elevations in blood pressure and heart rate. Blood pressure remained elevated even when the animals were returned to standard bedding, whereas the heart rate declined back to its original value immediately in response to this shift. This study shows that telemetry is a very effective tool but that it needs integrating with other methods: in addition, a greater understanding of the biological significance of the changes in cardiovascular parameters is required before the hypothesis that these changes represent an indication of distress can be accepted.

In this study, we analysed the effects of social and human contact on calves ‘ behaviour and stress responses. We also measured the effect of this contact on calves ‘ reactions to novel conspecifics and novel humans. Sixty-four calves were housed either alone or in pairs and received either minimal human contact or ‘additional’ human contact (stroking and talking). At six, 10 and 14 weeks of age, the behaviour of the calves was recorded in their home pens. Calves were then tested in an unfamiliar arena either alone, with an unfamiliar calf, or with an unfamiliar man, and in a Y-maze with one arm leading to a calf and the other to a man. An adrenocorticotrophic hormone (ACTH) challenge was performed in order to assess chronic stress responses. Compared with individually housed calves, pair-housed calves were more active and made fewer contacts with their neighbours when in their home pens; they were also less active in the arena, spent more time near the calf in the Y-maze, and had lower Cortisol responses to ACTH. Calves that had received additional human contact interacted more with the man in the arena and had lower mean heart rates than those that had received minimal contact. This study confirmed that calves feel a need for social contacts and that pair-housing can lower the stress felt by calves separated from their conspecifics. Additional contact from stockpersons increases calves ‘ likelihood of approaching humans but cannot compensate for their lack of social partners. Hence, when calves are separated, the duration of the separation should be limited, and visual and physical contact with other calves should be provided.

Catching of broilers is the first stage in the transfer of birds to the slaughterhouse. The catching process entails a high risk not only of stress but also of injury and death to the birds. Associated injury and mortality rates have important implications not only for animal welfare but also for the economics of the procedure. Catching machines are advantageous with regard to labour costs and standards, and they may also reduce damage to the birds. In the present investigation the use of a sweeper-type catching machine was compared with manual catching under commercial conditions, data being collected during 43 mechanical and 40 manual catching events evenly distributed over one year. Dead-on-arrival rates were recorded, and 108 068 mechanically caught and 87 916 manually caught birds were examined for injuries on the shackles at the processing plant. Injury rates of all types were significantly reduced after mechanical catching. This improvement was highest with respect to leg injuries. There was no significant difference in the number of dead-on-arrivals except during the spring period, when there were higher losses of birds caught mechanically; this was thought to be attributable to climatic conditions. The loading of the transport containers with equal numbers of birds and the initial familiarisation period of the catching team with the machine are potentially problematic factors with potential for improvement. The catching machine investigated here, with its lower risk of injury to broilers than commercial manual catching, has the potential to limit impairment of bird welfare during catching.

Most dairy producers believe that early separation of the cow and calf is necessary for reproductive efficiency, to minimise health problems and to improve the ease of milking. In addition, many consider that it is less distressing for both animals if separation occurs earlier rather than later. In this paper we review the welfare and production effects of early separation on the cow and calf. Research has shown that the cow's immediate behavioural response to separation from the calf increases with increased contact with the calf, but other work has shown that cow health and productivity are positively affected by the presence of the calf. Calf response to separation also increases when the calf spends more time with the cow, but there are long-term benefits of prolonged contact in terms of sociality, fearfulness and future maternal behaviour. Health, weight gain and future productivity are also improved when the calf is allowed to spend more time with the cow.

Welfare is a multidimensional construct and its quantification is of major scientific, societal and economic importance in veterinary medicine. The construction of indices that measure welfare validly and reliably remains a considerable challenge. A general methodology for constructing welfare indices can be adapted from human medicine (in particular, from methodologies to assess Quality of Life [QoL]) and modified to reflect the fact that all assessments of animal welfare must be observer-based. The methodology is based on the creation of individual, composite indices for distinct dimensions/domains of welfare such as pain, disease, or stress. The domains include behavioural, physiological and biochemical markers. We have established QoL methodologies in the assessment of acute and chronic pain in dogs and generalised this approach to farm animal welfare. We describe the development of a questionnaire with seven behavioural categories which are used to create a single pain score to assess acute pain in dogs. For chronic pain in dogs, a structured questionnaire with over 100 items has been devised, which the owner completes by indicating degree of agreement with each item using a seven-point Likert scale. The welfare measure includes pain as an integral component as well as husbandry, behavioural and physiological/biological measures. In each case, a profile of the individual indices can be studied and compared over time or among observers. These indices may also be combined to form a single composite welfare measure, should this be appropriate, using scaling models. In the welfare setting, we have both causal and indicator variables — and indeed, for farm animals, the causal variables may be sufficient cause for poor welfare (eg the presence of disease or inadequate husbandry).

Assessment of mass-produced animal-housing equipment can serve as a basis for improving animal welfare. A number of European countries have adopted various legal approaches to such assessment. In Germany, welfare assessment of housing equipment is voluntary, but minimum standards can be set by regulation for the assessment procedure and for the qualifications of the persons involved. From a scientific perspective, the time and resource constraints pose some problems, particularly as they apply to a voluntary procedure. For reasons of practicability, certain compromises will be required. Nevertheless, it is important to ensure that each assessment procedure is based upon scientific principles and considers animal welfare aspects to a sufficient extent. A proposal for the minimum standards of an assessment procedure has been elaborated by the Animal Welfare Committee of the German Agricultural Society (Deutsche Landwirtschafts-Gesellschaft, DLG), a shortened version of which is presented here. The animal welfare impact of such a regulated but voluntary procedure will be less than that of an obligatory assessment; however, the relatively flexible approach may still significantly contribute to the improvement of welfare aspects of livestock housing.

Farm animal welfare has now been studied, within a scientific framework, for several decades. The framework does not include ethical issues, but unless measurements of animal welfare at farm level are embedded in an ethical context, there is a danger that these measurements will not be properly utilised. This paper considers the relationship between ethical questions and animal welfare assessment. In it, the following four key ethical questions are identified. What is the baseline standard for morally acceptable animal welfare? What is a good animal life? What farming purposes are legitimate? What kinds of compromise are acceptable in a less-than-perfect world? The authors suggest that animal welfare scientists need to reflect carefully on these questions if welfare assessments are to be properly interpreted and put to practical use. Such reflection will lead to a more transparent appreciation of the values underlying welfare assessment. In this way, it will both offer welfare scientists a greater awareness of their own value-assumptions and enable the same scientists to communicate effectively with the wider audience — farmers, consumers, pressure groups, policy-makers and so on — for which the results of animal welfare assessments are of concern.

In the test described here, sheep are exposed to a situation of conflict between the motivation to approach other sheep and the motivation to avoid a human handler. The distance that the test sheep keep from the handler is a reflection of the relative aversiveness of this handler to the sheep. The test itself requires only a minimum amount of handling and gives the test animals the opportunity to choose their distance from the aversive stimulus, thereby reducing stress during the test itself. The two aversive stimuli chosen for comparison were a human handler facing toward the test arena (more aversive) or the same handler turning his back to the arena (less aversive). Ten Scottish Blackface sheep were tested individually a total of ten times, five times with each of the two stimuli in alternate tests. During the first two tests, nine of the sheep stayed further away when the human was facing toward the arena, compared to when he was facing away; this shows that the test is able to discriminate differences in aversiveness between two stimuli as perceived by individual sheep. This difference was not apparent in the following eight tests, probably because of the fact that the stimuli were not reinforced during the tests. Because the test is concerned with sheep's reaction to a stimulus (eg handler), the procedure associated with the stimulus itself (eg shearing, castration) does not have to be repeated in the test, which means that this method is ideal for studying procedures which cause distress to the animals or which are difficult to repeat.

A Delphi technique was used to gather the opinions of animal welfare experts on the most appropriate measures for welfare assessment of farm animals. The experts were asked to consider measures that were directed towards the animal (animal-based), rather than measurement of their environment. This systematic approach was designed to achieve a degree of consensus of opinion between a large number of experts. Two rounds of postal questionnaires were targeted at people with expertise in one or more of the species of interest. The respondents suggested measures based upon observations of health status, behaviour, and examination of records. These measures reflect the animal's welfare state — in other words, how the animal is coping within the environment and husbandry system in which it lives. The measures for cattle, pigs and laying hens were categorised into 22, 23 and 28 aspects, respectively, with the highest ranking of importance being given to observation of lameness in dairy cattle and pigs and to observation of feather condition in laying hens. This Delphi study was the basis for the development of a series of protocols to assess the welfare state of dairy cattle, pigs and laying hens.