In the expanding salmon industry, many farmers use production methods that could result in poor welfare of the fish at various points of their lifecycle. We have reviewed methods used for producing salmon for food with the aim of identifying and drawing attention to factors likely to affect farmed Atlantic salmon (Salmo salar) welfare. In addition to water conditions and high stocking density at sea, other issues are important for fish welfare. Handling and transport of salmon between fresh- and seawater phases and before slaughter can have severe negative effects and research should continue to seek improved methods. Stocking densities in fresh- or seawater have substantial effects on the welfare of salmon and a reduction in densities should be considered in order to reduce fin damage in particular. Currently used feeding systems result in starvation for some fish and fin damage for others, hence new systems should be developed. Some on-demand feeding systems improve welfare. All farmed fish should be stunned prior to slaughter, not left to die of asphyxia. Carbon dioxide and electrical stunning methods do not always stun salmon humanely. The widely used methods of percussive stunning, manual or automatic, must be precise to effectively stun large numbers of fish. Welfare outcome indicators, such as fin damage, morbidity and mortality rate, should be used in standards and laws relating to salmon welfare.

Recently, several farm assurance schemes in the United Kingdom have been adopting innovative approaches, such as welfare outcome assessment, into their routine procedures. In this paper, we present the findings of four consultation exercises, undertaken as part of a review process that examined farmer perspectives on planned or implemented changes to their current certification visits as members of UK-based dairy and laying hen schemes. The changes included the introduction of welfare outcome assessment by assessors, joint-scoring of welfare-outcome measures by farmers and assessors and self-assessment of welfare outcome measures by farmers between assessor visits. This study also explores the challenges that arise when schemes are aiming to adopt a scheme-level continuous improvement approach to promote welfare improvement on participating farms. The key challenges fall under three themes: the purpose and value for the farmer of the assessment of welfare outcomes as part of a farm assurance assessment process; the potential conflict rather than concordance with the role of the farmer in caring for their animals; and finally the technicalities of the assessment process, such as sample sizes for assessment being calibrated for gauging welfare prevalence at a scheme rather than farm level and the role of the farm assurance assessors both to assess impartially compliance against the standards and to provide welfare advice to support improvement. This study highlights that the involvement of farmers at all stages in the development and in the evaluation of outcome assessment initiatives is likely to be beneficial for welfare improvement on-farm.

Capturing wild birds with mist nets is very common for studying many aspects of avian biology. However, except for adrenocortical reactivity, little is known about other physiological responses to this potentially stressful procedure. Here, we focused on body temperature (Tb) in migratory songbirds, as large numbers are caught for population monitoring and studies of migratory biology. Tb is sensitive to acute stress in other vertebrates and contexts, usually showing an increase (stress-induced hyperthermia). We sampled garden warblers (Sylvia borin), whinchats (Saxicola rubetra) and barn swallows (Hirundo rustica) captured with mist nests using standard protocols at an island stopover site in the central Mediterranean during spring migration. Tb was measured within 3 min (T0), 30 min (T30) and 180 min (T180) from the time the bird hit the net, using an analogue probe inserted into the throat, and a body condition score was calculated for each bird. In the garden warbler, but not the other two species, a slight but significant reduction in Tb occurred after 180 min. In all species, the change in Tb after 3 h of confinement was positively correlated to the change in ambient temperature (Ta) but not to body condition. The mean (± SEM) change in Tb over 3 h was −1.68 (± 0.16)°C for garden warblers, therefore within the expected range of normothermy in small birds. Such reduction in Tb is contrary to the expectation of stress-induced hyperthermia; these results suggest that garden warblers are able to modulate their body temperature and that these small changes are influenced by fluctuations in ambient temperature. Therefore, Tb might not be a good indicator of capture stress in small passerine migrants. Our results also indicate that restraint for ringing procedures is unlikely to have adverse effects on Tb regulation of migratory birds, if held within the time-frame typical of ringing operations.

Asiatic black bears (Ursus thibetanus) face chronic stress in bile farms. In this study, we investigated whether bile-farmed bears show significantly high levels of stress at rescue and whether stress levels reduce over time in a bear sanctuary where the bears are supported with environmental enrichment and veterinary care to improve animal welfare. We measured stress hormone levels using faecal cortisol metabolites (FCM) in 16 Asiatic black bears freshly rescued from bile farms in Vietnam. Fresh faeces were collected from each bear on the rescue truck and on a weekly basis for a 22-week study period at a bear sanctuary in Vietnam. Results showed that for all 16 rescued bears (with one exception) individual FCM levels from truck samples were above mean baseline FCMs of bears previously rehabilitated to a bear sanctuary. This suggested the majority of the rescued bears were still capable of showing a stress endocrine response during the rescue operation despite being exposed to conditions causing chronic stress in bears on bile farms. Results showed that mean FCM levels of the rescued bears differed significantly between time-periods (higher at the rescue [on truck samples] compared to week 22 samples) and mean FCM levels showed an overall decline over the first 22 weeks after they arrived at the bear sanctuary. The bears also demonstrated acute FCM stress responses to management interventions at the sanctuary, such as veterinary health checks and transportation. In conclusion, rescued bears tend to modulate their stress endocrine response after rehoming at the bear sanctuary. This is an important result, indicating that the rescue effort and rehabilitation of bile-farm bears is effective. Whether this also coincides with behavioural adjustments in rehabilitating bears (eg lessening of stereotypic behaviour) warrants further investigation.

Free-ranging pet cats (Felis catus) frequently kill wildlife but also return live prey to their owners. This raises welfare concerns if live animals released by cat owners subsequently die, since this is preventable, eg through prompt euthanasia. To estimate the mortality rate of birds released alive by cat owners, we examined the fates of 3,597 cat-attacked individuals submitted to four RSPCA wildlife centres in the UK. Individuals from 64 species were received but most cases (77%) involved just seven species commonly found in urban areas. The overall mortality rate (based on all individuals received at centres, ie including those which perished in transport, those which were euthanased on arrival and those which were admitted for care after having been triaged) was 78%; the post-admittance mortality rate (n = 2,070 birds admitted for care) was 62%. On average, individuals that perished (n = 2,798) survived for 3.0 days before dying or being euthanased. Juveniles were more likely to survive to release than adults, possibly because their small size means they are less likely to receive injuries that are ultimately fatal. Extrapolating from the limited data currently available, and applying conservative estimates at each stage, we estimate that a minimum of 0.3 million birds are released annually by cat owners but subsequently die. Substantial welfare improvements could be achieved if owners were more prepared to adopt strategies to limit hunting behaviour (eg fitting cats with collars and bells) and if owners and rehabilitators were able to effectively identify individuals with fatal injuries. The latter will require studies that quantify the effects of identifiable physical injuries on the likelihood of survival to release, in order to establish effective triage criteria.

Small laboratory animals are commonly euthanased via intraperitoneal (IP) injection of sodium pentobarbital. However, there is concern that animals may experience pain prior to loss of consciousness with this delivery route. The present study investigated electroencephalographic (EEG) nociceptive responses of anaesthetised pigs to IP sodium pentobarbital injection using an established minimal anaesthesia model. Thirty commercial white line entire male pigs aged 10-15 days were minimally anaesthetised with halothane in oxygen. Following 10 min of baseline EEG data collection, pigs had their tails docked using side-cutters and, after a further 5-min interval, were euthanased via IP injection of sodium pentobarbital (250 mg kg-1). The summary variables median frequency (F50), 95% spectral edge frequency (F95) and total power (PTOT) were derived from the EEG data. For each variable in each pig, means were calculated for the following 60-s periods: immediately prior to tail-docking (baseline 1); immediately prior to pentobarbital injection (at least 4 min after docking; baseline 2); and for two consecutive 60-s periods immediately following pentobarbital injection (P1 and P2). Statistical analyses revealed no differences between the two baseline periods, indicating that transient EEG changes induced by tail-docking had resolved prior to pentobarbital injection. IP pentobarbital injection induced a significant increase in F50 and decrease in PTOT of the EEG during P1. This response is characteristic of acute nociception, indicating that conscious pigs likely perceive IP sodium pentobarbital as painful in the period prior to loss of consciousness.

The aim of this study was to assess hair cortisol concentrations in New Zealand white rabbits (Oryctolagus cuniculus) that were subjected to relocation and surgery to evaluate HPA-axis activity; in addition, we used this marker of cortisol secretion to evaluate the allostatic load of animals undergoing surgery. After a period of acclimatisation, which lasted 40 days from their arrival at the enclosure, 19 rabbits were subjected to T1-T12 dorsal arthrodesis (RS), 19 were sham-operated (SS), and 19 were non-operated (CON). Hair samples were collected at the time of arrival (ST1) at the animal facility, and seven other sets of hair samples were collected at 40-day intervals from the same area of skin for a period of 240 days as re-shaved hair (anagen phase): immediately before surgery (ST2) and after the surgery (ST3, ST4, ST5, ST6, ST7, and ST8). The transition from the rabbitry to the animal breeding facility led to a significant increase in cortisol concentration (ST2) in all of the groups. At ST3, the RS group presented higher cortisol concentrations than those of the SS group and the CON group. At ST4, the experimental groups showed similar values that remained constant until ST8. The results show that the management of rabbits undergoing surgery should be evaluated very carefully, and hair cortisol concentrations may provide a means of avoiding the dangerous cumulative effects of additional stressors close to surgery.

Accelerometers are used to remotely monitor activity in various species in studies that quantify pain, document behavioural patterns, and measure individual activity differences. Studies validating accelerometers typically quantify various active states; however, targeting states specific to periods of inactivity, such as sitting, sleeping, and standing, has the potential to more accurately quantify inactive behaviours commonly associated with behavioural changes related to pain, sickness, or injury. Our objectives were two-fold: first, validate a commercially available accelerometer (Actical®) for quantifying inactivity in laying hens and, second, compare inactivity levels between hens with severely fractured keel bones and hens with minimal to no keel damage. Correlation between the inactivity level as measured by the accelerometer compared to live, focal observation of stationary, inactive behaviours was high; therefore, the Actical® accurately quantifies inactive states in laying hens. Following validation, the Actical® accelerometer was used to quantify inactivity level differences between hens with or without keel-bone damage. Severely fractured hens spent less time motionless, than hens with minimal to no keel damage. Further investigation into inactivity differences related to keel status before and after acquisition of keel fractures is warranted. Use of the accelerometer has the potential to improve animal welfare research by quantifying the effect of pain or sickness on activity level, mapping daily activity patterns, and measuring individual differences in general activity.

Exposure to high concentrations of CO2 is a common means of stunning and killing laboratory rodents. However, there is concern regarding the potential for animals to have aversive experiences, such as pain or breathlessness, prior to loss of awareness. This preliminary study evaluated the electroencephalographic (EEG) responses of rats (Rattus norvegicus) to CO2 inhalation, using a method based on a minimal anaesthesia model previously used to assess nociception in mammals. Fifteen adult female Sprague-Dawley rats were anaesthetised with halothane in oxygen and maintained at a minimal plane of anaesthesia. EEG was continuously recorded throughout a 10-min baseline period followed by sequential exposure to 5, 15, 30 and 50% CO2. The EEG summary variables median frequency (F50), 95% spectral edge frequency (F95) and total power (PTOT) were derived from the raw EEG. The F50 of the EEG, a sensitive indicator of nociception, increased significantly above baseline during exposure to 15% CO2, suggesting this concentration was noxious to rats. This is consistent with behavioural aversion in rats at around the same CO2 concentration. Stimulation of the rat mucosal nociceptors only occurs at CO2 concentrations of 37% or greater; therefore, it is hypothesised that the observed response was a result of what would have been CO2-induced aversive respiratory sensation in conscious animals, rather than pain. This study provides some evidence that an anaesthesia model may be utilised to study the nocuous effects of low-moderate CO2 exposure in rodents.

A case control study with a questionnaire was carried out to compare feeding practices, diet composition, housing and management in 78 herds with or without a history of tail-biting in undocked pigs (Sus scrofa) in Finland. Tail-biting was measured as the mean annual prevalence score of tail-biting damage (TBD) for a farm. Logistic regression parameters were calculated separately for risk factors present in piglet (lactation), weaner, and finishing units. Risk factors found in piglet units for TBD were slatted floors and area of slats. In the weaner units, slatted floors, area of slats, use of whey or wheat in the diet, and use of purchased compound feeds were associated with a risk of TBD. In the finishing units, slatted floors, area of slats, increasing number of finisher pigs at the farm, absence of bedding, liquid feeding, several meals per day, specialised production type and a group size greater than nine pigs were found as risk factors for TBD. Increased farm size was connected to risk for TBD in the overall dataset. The nutritional risk factors seem to operate together with other risk factors, but with relatively lower odds. The risk factors of undocked herds in this study seem to be similar to the risk factors from earlier studies of docked pigs. This study provides information which can be used to refine decision-support tools for management of the potentially higher risk for tail-biting among long-tailed pigs, thus aiding compliance with EU law and enhancing pig welfare.

An understanding of the factors influencing the human-animal relationship is relevant in the development of welfare interventions for working equids. However, the link between the psychological attributes of animal owners, such as empathy level and pain perception, and the welfare state of working horses is yet to be studied. Here, we assessed working horse owners’ empathy, their perception of equine pain and how these relate to the welfare state of their horses. The relationship between empathy and perception of equine pain was also explored. One hundred owners of working horses were studied, along with their working horses (n = 127). Self-reported rating scales were used for measuring animal-oriented empathy, human-oriented empathy and owners’ perception of equine pain; higher scores in the three scales indicated higher levels of empathy and higher perception of equine pain. An animal welfare index based on direct measures was applied where higher scores indicate a poorer welfare state. Most of the owners had high levels of empathy towards both animals and humans and a high perception of equine pain. According to the animal welfare index, 15.7% of horses had a poor welfare state. Higher levels of empathy and equine pain perception in owners were correlated with a better animal welfare condition, with owner's empathy level towards animals explaining over 60% of horses’ welfare. Moreover, empathy towards animals was positively correlated with human-oriented empathy and perception of equine pain. We therefore suggest that strategies oriented to improve welfare of working horses should consider education programmes aimed at promoting the development of empathy in caretakers.

A previous study demonstrated the effectiveness of the Bock Industries Zephyr EXL non-penetrating captive bolt, using the abolition of visual-evoked potentials as a determination of brain death, in piglets in a laboratory. A second trial, reported here, involved the field-testing of this device, on-farm, in a commercial setting. Two hundred and seven piglets (mean [± SEM] dead weight: 1.86 [± 0.74] kg) requiring dispatch under the farm's protocols were euthanased with the device and demonstrated immediate loss of consciousness, subjectively assessed by behavioural signs and no recovery. Post mortem examination of the heads was undertaken confirming massive traumatic damage to the cerebrum with associated haemorrhage and bone plate shards forced down to the level of the corpus callosum in the majority of cases. A further trial of 106 piglets demonstrated that under commercial production conditions it took less than 7 s to select, place and euthanase a piglet using the device. One hundred percent of animals in the study were immediately killed. Given this complete kill rate and the sample size of the study, a statistical 95% confidence interval provides a maximum percentage of animals that would not immediately be stunned/killed, by this mechanical non-penetrating captive-bolt system, to be at most 1.2% and at least 0%. The results here, combined with the previous study, allow recommendation of the Bock Industries Zephyr EXL as being suitable as a single application euthanasia device for piglets up to 10.9 kg liveweight.

Humane euthanasia of cattle under field conditions presents special challenges for veterinarians and producers. The purpose of this study was to evaluate the effectiveness of the CASH Dispatch Kit captive-bolt system combined with improved shot placement landmarks as a single-step euthanasia method for cattle. Cattle destined for euthanasia for reasons unrelated to the study were utilised. Adult (> 2 years), young (6-24 months) and neonatal (< 1 month) cattle each received a single shot from the CASH penetrating captive-bolt pistol. An additional group of neonatal animals was shot with a non-penetrating muzzle attachment. The shot was placed on midline halfway between the top of the poll and an imaginary line connecting the lateral canthus of each eye. Following the shot, the animals were immediately assessed for loss of consciousness based upon: i) immediate collapse (if standing); ii) loss of eye reflexes with a centered, dilated pupil; iii) lack of co-ordinated respiration; iv) lack of vocalisation; and v) lack of a righting reflex. Lack of consciousness and heartbeat were assessed at 1-min intervals until cardiac arrest. All animals were adequately stunned by a single shot. Euthanasia via a single shot was successful in 28/31, 17/19, 8/10, and 9/10 adult, young, neonate (penetrating) and neonate (non-penetrating) animals, respectively. Reasons for failure included return of co-ordinated respiration and prolonged time until cardiac arrest. A single shot from the CASH Dispatch captive-bolt system will humanely euthanase most animals. However, the results of this study indicate that application of a follow-up step to ensure death is still needed in certain instances.

In an effort to enhance welfare, behavioural management continually refines methods of non-human primate (NHP) care. Chimpanzees (Pan troglodytes) are one of the most cognitively complex captive NHPs and they have been observed to self-medicate in the wild. The population of captive chimpanzees in the US is aged (due to a breeding moratorium instituted in 1998) and will progressively require more medical care as they get older. To functionally simulate natural self-medication behaviour, provide chimpanzees with the opportunity to voluntarily participate in their own healthcare, and open new avenues of communication between caregivers and chimpanzees, we used a medication choice paradigm that allowed chimpanzees to choose their daily arthritis medication. We provided four arthritic, mobility-impaired chimpanzees with meloxicam or ibuprofen in blue or green Gatorade® to establish associations between the coloured drinks and the effects of the medications. We subsequently gave each chimpanzee a choice between the two medications. Behaviour was recorded using 15-min focal animal observations. Mobility was assessed using interactive mobility tests and a caregiver-rating system. One chimpanzee showed a medication preference (ibuprofen over meloxicam). The chimpanzees exhibited no significant behavioural or mobility differences over time, suggesting that ibuprofen and meloxicam may not differ significantly in their ability to alleviate arthritic symptoms. Whether or not the chimpanzees show a medication preference, the opportunity to make meaningful choices and the functional simulation of a complex behaviour, self-medication, is present when using this medication choice technique. Furthermore, the paradigm itself could have potential applications for additional medication options and treatment regimens.

In response to an increased awareness concerning the welfare of captive animals, several studies have investigated the effect of provisions on stress levels in model species, such as small mammals, birds and fish. In contrast, reptiles have received less attention. Although many reptilian species are becoming increasingly popular in the pet trade and are frequently used as model species in various branches of biology and a number of studies have explored how they react to stress in different contexts (eg social, predatory), little is known about how they react to stress induced by housing conditions or experimental treatments. In this study, we quantified the effect of provision of perches and leaves as refuges (provisioned) on the behaviour, morphology and physiology of the green anole (Anolis carolinensis). Our results showed that increased or decreased structural complexity of the cage had no effect on body mass, tail-base width, heterophil to lymphocyte ratios (H/L ratios), brightness, body colour, behaviour and faecal corticosterone metabolite (FCM) levels for both males and females in the experimental treatments (provisioned or deprived situation). Our study animals did score very highly for several stress-indicating variables in the three weeks preceding the experiments — suggesting that they had experienced considerable stress during capture, transport and temporary housing in the pet store.

This study was designed to determine whether feedback from welfare assessments, using the Horse Welfare Assessment Protocol, affected actual horse welfare in 21 stables. After the first assessment, stable managers in the high feedback (HF; n = 10 stables) group were supplied with extensive information and support regarding the welfare measures and relevance of the results. The low feedback (LF; n = 11 stables) group only received the results without additional information. Upon re-assessment, six months later, no significant changes were seen in the stable overall (SO) score in either group. Significant changes occurred in individual measures; in the HF group more fresh-air inlets were open but water drinker function and ocular discharge deteriorated. In the LF group, the feeding troughs were cleaner but mane and tail condition deteriorated. Both groups had cleaner water troughs and less equipment chafing but the sum of relative air humidity (RH) and temperature (T) deteriorated. Significant decreases occurred in the stable welfare issues (SWI) score; the HF group decreased from 93.3 to 72.0 and the LF group from 113.3 to 91.3. There were also nonsignificant changes; in the HF group, 71 measures and five stables improved while 63 measures and five stables (50%) deteriorated. In the LF group, 65 measures and seven stables improved while 62 measures and four stables deteriorated. The observed improvements in both groups suggest that assessment alone (with no detailed feedback) might raise awareness but we cannot yet conclude whether or not the type of feedback affects overall horse welfare.

The relationship between the socio-economic status of working equine owners and the welfare status of their animals is yet to be documented. The aim of this study was to provide an approach to understanding socio-economic status, quality of life and working aspects of working horse owners, in order to establish their social vulnerability index and to determine how these measures correlate with the welfare state of their horses. Seventy-two owners and their urban working horses (n = 122) were studied. Owners’ socioeconomic and educational status was established together with their quality of life perception and multi-dimensional poverty index. The animal welfare index was constructed using animal-based measures. Whilst over 90% of owners were considered vulnerable, only 28.3% of horses were classified as being in a poor welfare state (eg presence of lesions and morphology not adequate for draught type). There were no significant correlations between owners’ factors and the animal welfare index. We conclude, therefore, that social vulnerability of owners does not necessarily imply that their animals will be in a poor welfare state.

Concerns over elephant welfare in UK zoos have implications for their future in captivity. To monitor improvements made to elephant welfare in UK zoos, non-invasive, valid and reliable indicators of welfare are needed. Using a rapid review strategy and critical appraisal tool, we aimed to appraise evidence from peer-reviewed literature on potential welfare indicators for captive elephants. Scopus, Web of Knowledge and Ovid were searched in January 2014 using terms relevant to captive elephants and welfare assessment. Inclusion and exclusion criteria were applied and remaining articles were critically appraised against a specially designed welfare indicator appraisal tool. Thirty-seven unique indicators of welfare were extracted from 30 peer-reviewed papers which met the inclusion criteria. Behavioural measures of welfare (n = 21) were more common than either physical (n = 11) or physiological (n = 5) measures. Stereotypies were the most frequently used behavioural measure, glucocorticoids were the most frequently used physiological measure and body condition scores were the most frequently used physical measure. There was most support for the following indicators of improved welfare state: reduced stereotypies, reduced glucocorticoids and improved body condition scores. Additional measures which require further validation but had strong associations with the most supported measures, and thus have potential use in welfare assessment, were: increased lying rest and positive social interactions. Further validation of the described measures is needed, but this information forms a crucial part of the knowledge required to efficiently monitor and improve the welfare of elephants in captivity.

Head-to-body stunning is regarded as ‘best practice’ stunning for sheep. The benefits are loss of consciousness followed by cardiac arrest, death, prevention of animal movements post stun/kill and improved meat quality. Commercial equipment places electrodes on the head and back, which is known to cause pelt burning, thus reducing the value of the skins. The aim was to demonstrate that passing current at 1.5 A and 50 Hz from the top of the head to the chest in lambs for 3.1 s would result in epilepsy. Electroencephalographic (EEG) and electrocardiographic (ECG) activity was recorded in sheep using non-invasive electrodes. Measurements in this trial were successfully performed on three lambs (live weight 25 to 39 kg) which were anaesthetised and given neuromuscular blockers to inhibit muscle activity. EEG information showed that the head-to-chest stunning produced an epileptic-like episode, which was followed by an isoelectric output. ECG recordings showed that ventricular fibrillation (VF) was induced and coincided with the epileptic brain activity observed. No animals regained brain activity or sinus heart rhythm after applying the stated stunning conditions. As a conclusion, it is postulated that modified stunning equipment passing an electrical current from the top of the head to the sternum in lambs (1.5 A, 50 Hz; 3.1 s) may induce an epileptic seizure and VF.

In this study, we sought to explain why dairy farmers give importance to various use and non-use values of animal welfare. In particular, we examined how the farmers could be segmented according to the relative importance they give to various use and non-use values in animal welfare and whether this segmentation could be explained by farmers’ personal values and personality traits. Based on a latent class analysis using best-worst scaling data on 123 Swedish dairy farmers, three segments of farmers were found: animalcentred, human-centred and business-orientated. These groups were related to measures of farmers’ personal values and personality traits in a point-biserial correlation and a hierarchal multinomial logistic regression analysis. The results suggest that the segmentation is related to personal values, but not to personality traits. This finding is important from a policy perspective, because the existence of different segments of farmers who are motivated by different values in animal welfare indicates a need to approach different farmers in different ways if policy is to succeed in improving animal welfare. It also indicates a possibility to influence the segments to improve animal welfare by measures that are sensitive to value dynamics of the farmers.