The assessment of animal welfare requires the collection of multiple indicators of welfare but quantification of their associations in different contexts is lacking. Previous studies have examined correlations between a few indicators, but not relationships between many different indicators, or between indicators taken from the same individuals in more than one environment. We housed 60 hens for six sequential 35-day phases in different pen environments. During each phase, a series of behavioural and physiological measures was taken for every bird: body and plumage condition, surface body temperature, behaviours observed in the home pens and during test periods, tonic immobility, physiological blood profiles, and faecal sample composition. Most variation in nearly all measures was not explained by either individual bird or grouping effects but varied across phases within the birds. Acknowledging this, we examined correlations between all parameters at the phase within-bird level, selecting a conservative P-value. A consistent set of correlations showed that a slow approach response and alert behaviour in the novel object test was associated with higher bodyweight, lower body temperature and lower acute phase protein, heterophil: lymphocyte ratio and blood glucose level. A cluster analysis confirmed these correlations. Other important parameters known to be linked to the hens’ environmental preference (eg comfort behaviour) were independent of the set described above. We conclude that statistical techniques can reveal patterns of independence and redundancy in the collection of behavioural and physiological measures of welfare.

The objective of this study was to identify farm characteristics that were associated with cow (Bos taurus) welfare outcomes, including mortality rate, culling by 60 days of lactation, survival to ≥ 6 years of age and ≥ 5th parity (aged cows), somatic cell score, milk yield, and milk composition. Data were collected on housing systems, feeding systems, pasture strategies, bedding type, labour management practices and other farm characteristics in face-to-face interviews with 314 Pennsylvania dairy herd owners where performance records were available. Five herd management systems were identified in the sample: free-stalls with complete confinement (n = 37); free-stalls that allowed outdoor access (n = 76); tie-stalls with complete confinement (n = 52); tie-stalls with outdoor access and that fed a total mixed ration (n = 72); and tie-stalls with outdoor access and that did not feed a total mixed ration (n = 77). Welfare outcomes were evaluated with multivariable linear regression models and marginal means were estimated for herd management system. Tie-stalls that allowed outdoor access and that did not feed total mixed rations had the lowest mortality rate (2.0%), culling in the first 60 days of lactation (5.1%), and the highest proportion of aged cows (13.8%). Those herds also had high lifetime-to-date milk yield, a low proportion of fat-protein inversions, and low somatic cell scores. Free-stalls with complete confinement had significantly higher levels of mortality (8.3%), culling in the first 60 days of lactation (9.7%), and fewer aged cows (6.4%). It was concluded that shifts toward more efficient herd management systems have not benefited cow health and welfare. This suggests that cow welfare has been compromised to facilitate the economic survival of dairy farms.

Atlantic bottlenose dolphins (Tursiops trucatus) are found in zoos and aquaria throughout the world. As the number of facilities with dolphin shows and interaction programmes increases, it becomes more important to understand the effects of such programmes on dolphin behaviour. The present study examined the short-term effects of dolphin shows and interaction programmes on the behaviour of Atlantic bottlenose dolphins at six facilities. Rates of affiliative behaviour, aggressive behaviour, repetitive behaviour and percentage of time spent socialising were found to be unrelated to dolphin shows or interaction programmes. Additionally, dolphins exhibited higher rates of behavioural diversity, diversity of swimming style, and play behaviour following shows and interaction programmes. These results suggest that dolphin shows and interaction programmes can be an important part of an enrichment programme for dolphins in zoological institutions. However, individual differences should be considered when animals participate in these types of programmes.

This study compared the effects of concrete slats (CS), synthetic rubber slats on aluminium profiles (RS) and slotted rubber mats on concrete slats (RM) in fully slatted floor pens on behaviour, claw and leg disorders, claw horn growth, cleanliness and production parameters of growing dairy bulls from 225 to 650 kg average liveweight. Each pen housed five bulls up to 400 kg average liveweight and four bulls thereafter. On CS, lying bouts were less frequent and longer than on RM and RS at 250 kg. Lying down phase 1 was longest on CS and shortest on RM. Interrupted attempts at lying down occurred twice as often on CS as on the rubber floors. Severity scores for white line haemorrhage and sole haemorrhage were higher in bulls on CS than on RM. Swelling on legs had highest scores on CS, whereas the severity score for heel horn erosion was lowest on CS. Floor type had no effect on dermatitis, leg hairlessness and skin damage. Both claw horn growth and wear were greater on CS than on RS and RM. Bulls on RS and CS were cleanest. Slaughter age tended to be higher and carcase conformation score tended to be lower on CS than on rubber, whereas feed intake, feed efficiency and other carcase traits were unaffected. The results indicate that rubber flooring improves animal welfare compared with concrete.

The aim of this paper was to develop a protocol to study the interpretation of ambiguous stimuli in cats as a measure of welfare. Ten cats were trained to discriminate between a rewarded position (R) and an unrewarded one (U), as measured by the approach latency for each position. After discrimination, they were exposed to three ambiguous unrewarded positions (R-near, R/U-equidistant, U-near) distributed at intermediate points between R and U. Approach latency increased as increasing the distance from the rewarded position: latencies to approach R and R-near were significantly shorter than for R/U-equidistant, U-near and U. This protocol should be further studied to assess its effectiveness in highlighting differences according to the welfare level of individual cats.

The aim of this paper was to develop a protocol to study the anticipatory response in cats as a measure of welfare. Seven experimental cats were trained in a classical conditioning paradigm to associate a sound with food arrival, while sound and food were presented without contingency in four control cats. Increasing the interval between sound and food up to 60 s allowed a detailed description of cats’ anticipatory response. Compared to control animals, experimental cats showed significantly shorter latencies to orient towards (average 2.96 s) and approach the source of the sound (12.98 s) as well as longer durations of exploring and standing by the source of the sound (namely 69.97 and 52.32%, respectively of the interval sound-food). Experimental cats also exhibited behaviours that may derive from predation patterns, eg short pauses and predatory crouch while approaching the source of the sound (namely in 28.93 and 29.64% of trials), rapid head movements while watching it (55.36% of trials) and pouncing on the food (9.29% of trials). This protocol should be further studied to assess its effectiveness in highlighting differences according to the welfare of individual cats.

Lameness is one of the most serious health and welfare problems faced by dairy cattle. The aim of this study was to assess how aware Czech farm managers were of this problem. The project was carried out on 14 Czech dairy farms. The proportion of lame cows observed (ie prevalence of moderate and severe lameness) on a farm varied between 9 and 64% while the farm managers’ estimation ranged between 0 and 20%, showing that lameness prevalence was under-perceived by the Czech farm managers. There were no correlations between the farm managers’ estimations and the observed total or severe lameness on each farm. Also, the observed prevalence of lameness did not differ between farms with managers who considered lameness to be a major problem in their herds and managers who did not, although their estimations did differ. The large variation in observed lameness prevalence between farms indicates that there is a large potential for reduction, which must start with increasing the farm managers’ awareness of dairy cow lameness.

Traditional husbandry practices for the public display of Callitrichidae involve strict separation of animals and public. An important consideration for the evaluation of such management is the occurrence of health problems and potential zoonotic risks. This study compared animal data and veterinary records from a captive, free-roaming population of golden-headed lion tamarins (Leontopithecus chrysomelas) with a population housed under indoor management. The captive, free-roaming group grew significantly faster than the indoor-housed group, as less animals died and offspring were more successfully weaned. No differences in the causes of death were detected between the two husbandry practices. However, bacterial diseases were most common and diagnosed significantly more often under indoor management. This study suggests that the captive, free-roaming management of golden-headed lion tamarins can be conducive to increased reproductive success, improved health and improved welfare. Therefore, current husbandry recommendations for captive, free-roaming conditions were supported by the findings of the present study as a valuable housing practice where health regulations and proper husbandry parameters permit.

Methods for culling wild red deer (Cervus elaphus) were compared by observing behaviour and collecting post mortem samples from wild deer shot: (i) by a single stalker during daytime; (ii) by more than one stalker during daytime; (iii) by using a helicopter for the deployment of stalkers and carcase extraction; or (iv) by a single stalker at night, and compared with farmed red deer shot in a field or killed at a slaughterhouse. Culling by a single stalker during the day and shooting in a field were the most accurate in achieving placement of a shot in a target area, but when compared across all methods, there were no significant differences in the percentages of deer that were either wounded or appeared to have died immediately after the first shot. Plasma cortisol concentrations in deer shot using helicopter assistance were similar to those in deer at the slaughterhouse, but higher than deer shot at night or during the day by a single stalker, or in a field. Deer shot using helicopter assistance and also deer culled by a collaborative and single stalking during the day had lower muscle glycogen concentrations than those culled by a single stalker at night. There was no evidence that a particular culling method was associated with an increased risk of accidental or pre-culling injury. If a helicopter is used to assist culling, the deer are more likely to be disturbed before they are shot and therefore, measures should be taken to minimise the disturbance to the deer.

This study aimed to identify behaviours that could be used to assess post-operative pain and analgesic efficacy in male rabbits. In consideration of the ‘Three Rs’, behavioural data were collected on seven male New Zealand White rabbits in an ethically approved experiment requiring abdominal implantation of a telemetric device for purposes other than behavioural assessment. Prior to surgery, rabbits were anaesthetised using an isoflurane/oxygen mix and given Carprofen (2 mg kg−1) as a peri-operative analgesic. Rabbits were housed individually in standard laboratory cages throughout. Data were collected at three time periods: 24-21 h prior to surgery (T1) and, post-surgery, 0-3 h (T2) and 3-6 h (T3). Behavioural changes were identified using Observer XT, significance of which was assessed using a Friedman's test for several related samples. The frequency or duration of numerous pre-operative behaviours was significantly reduced in T2 and T3, as compared to T1. Conversely, novel or rare behaviours had either first occurrence or significant increase in T2 into T3 as compared to T1, these include ‘full-body-flexing’, ‘tight-huddling’, ‘hind-leg-shuffling’. We conclude that reduced expression of common pre-operative behaviours and the appearance of certain novel post-operative behaviours may be indicative of pain in rabbits. Behaviours identified as increased in T2 as compared to T1 but not consistently elevated into T3 were considered separately due to the potentially confounding effect of anaesthesia recovery. These included lateral lying, ‘drawing-back’, ‘staggering’ and ‘closed eyes’. We postulate that for effective application of best-practice post-operative care, informed behavioural observation can provide routes by which carers may identify requirements for additional post-operative analgesia. Additionally, improvement of the peri-operative pain management regimen may be required to ameliorate the immediate effects of abdominal surgery. Comparisons with other studies into post-operative pain expression in rabbits suggest behavioural indicators of pain may differ, depending on housing and surgical procedure.

Farm animal welfare is a societal concern, and the need exists for scientific protocols to assess welfare. This paper describes the development of a protocol to assess the welfare of sheep (Ovis aries) and its application in 36 farms in Norway. There were two parts to the protocol; the animal- and resource-based measurements obtained during farm visits, and the analysis of production data. Data collection took place during visits to 36 farms in the lambing season (April-May) in 2007 (n = 11) and 2008 (n = 25). A fear test was conducted, and ewes were scored on a scale from 0 to 3. Forty-one percent of the ewes tested had a fear score of 3, indicating the lowest level of fear. Mean (± SD) fear score across farms were 1.9 (± 0.5). Higher fearfulness was found to be associated with lower ewe body condition scores (BCS). Mean (± SD) BCS across farms was 2.6 (± 0.6). A relatively large proportion of the ewes had a BCS of 2 (41%), which may be associated with an increased risk of nutritional stress, disease and low productivity. Eight farms had more than 5% (range 5.4-24.4%) of lamb carcases categorised in the lowest conformation class, which may be an indication of a welfare problem. This study is the first step in the development and validation of a welfare assessment protocol for sheep, and further research is needed to assess the overall reliability of the protocol.

In the event of a notifiable disease outbreak, poultry may need to be culled in situ. This should be performed swiftly and humanely to prevent further spread of the pathogen while preserving the welfare of the animals prior to death. Here, we examined the aversion of broiler chicks (Gallus domesticus) to three lethal gas mixtures at various concentrations to determine the least aversive mix that could be used in whole-house gassing. For 1 h, individual chicks (n = 36) were allowed to place their heads inside three feeding and drinking stations (FDS) in order to access food and water. Each FDS was filled with a different gas mixture, and birds could access each FDS as much as they liked. Twelve chicks each were tested at low (50% carbon dioxide [CO2] in air, 70% argon [Ar] in CO2, 70% nitrogen [N2] in CO2), medium (55% CO2 in air, 80% Ar in CO2, 80% N2 in CO2) or high (60% CO2 in air, 90% Ar in CO2, 90% N2 in CO2) concentrations of gas mixtures. Aversion was assessed based on the time birds spent with head in each FDS (with more time indicating less aversive), and the frequency of head shakes relative to time spent with head in the FDS (with a lower proportion indicating less aversive). Data were analysed by ANOVA. On average, birds spent < 3 min with their head in any FDS. Mixtures containing 90% Ar or N2 in CO2 and 80% argon in CO2 were least aversive and mixtures containing 70% N2 in CO2 and 60% CO2 in air were most aversive, based on time spent with head in. Head shakes s−1 were more frequent with low concentration gas mixtures compared to high concentrations, and with all CO2 in air mixtures, which suggests that the intensity of head shaking is related to the concentrations of CO2. From these results, one concentration of each of the three gas mixtures (90% N2 in CO2, 80% Ar in CO2, and 50% CO2 in air) were chosen for assessment on a further 12 birds and the results showed that both inert gas mixtures were less aversive than 50% CO2 in air based on time spent with head in. Frequency of head shakes s−1 did not differ between the three mixtures. Birds found all gases aversive, however it is concluded that inert gas in CO2 mixtures were least aversive compared to CO2 in air and these gases also caused less signs of respiratory discomfort.

This study examined the sources of glare and range in luminance levels in eleven UK cattle abattoirs, and the effect of reflected glare from a footbath on balking behaviour of cows in a milking parlour. At the abattoirs, the mean luminance levels decreased from 240,000 cd m−2 outdoors to 100 cd m−2 in the stunning pen. In five of the abattoirs, the luminance of the glare from wet floor surfaces was three times higher than the luminance from the surrounding darker areas, and the luminance of the glare from shiny metal surfaces was ten times higher than the luminance from the adjacent darker areas. In the glare study, frequency of balking increased significantly from 10 to 23% when reflected glare increased from 0 (no lightbulb) to 873 cd m−2 (100 W bulb), but significantly fewer animals balked during the afternoon milking than at the morning milking times.

The aim of this study was to investigate how difficult it was to halter and separate a horse (Equus caballus) from a group for riding or training purposes and to describe how human-horse interactions could affect this common procedure. A total of 20 different horse groups, with a minimum of five horses in each, were studied on 14 farms in Norway, from June to October 2007. On each farm, information about the facilities, the horse and the handler was recorded. In each group, the horse owner or keeper (handler) was asked to enter the group, approach his/her horse, halter it and lead it out through the gate, then keep the horse standing out of sight from other horses for two minutes. The entire procedure was video recorded and the behaviour of the target horse, handler and other horses was subsequently scored. Only one of the 100 horses moved away from the handler when approached. Ninety-six percent of the target horses followed their handler without showing any resistance. In 75% of the tests, the other horses did not interact with the target horse and/or handler. At the gate, 26% of the handlers did not turn the target horse around before closing the gate, thereby placing themselves in a vulnerable position, with their body in very close proximity to the horse's hind legs. In conclusion, separating a horse from its group can be considered relatively safe and unproblematic, provided good management practices and trained handlers.

Severity classification systems (ie pain scales, categories of invasiveness, degrees of severity etc) are used to classify the adverse effects experienced by animals used for scientific purposes. Currently, eleven countries use severity classification systems. These systems have developed in various ways, depending on each country's process for overseeing the use of animals in science, as well as the particular aspects emphasised by those individuals who have championed their implementation. Severity classification serves four main purposes: as a tool to assist animal ethics committees in ethical review; education of animal users about concepts for humane animal experimentation; provision of data to inform the public about scientific animal use; and provision of data to inform national policies. At a time when the newly accepted European Union Directive will make the reporting of severity data mandatory, we review the characteristics of international severity classification systems and how they have evolved; analyse the effectiveness of some systems; and identify emerging challenges for severity classification.