Abstract: Recent studies raise serious welfare concerns regarding the estimated 93.6 million horses, donkeys and mules in developing countries. Most equids are used for work in poor communities, and are commonly afflicted with wounds, poor body condition, respiratory diseases, parasites, dental problems, and lameness. Non-physical welfare problems, such as fear of humans, are also of concern. Interventions to improve working equine welfare aim to prioritise the conditions that cause the most severe impositions on the animals' subjectively experienced welfare, but data identifying which conditions these may be, are lacking. Here we describe a stage in the validation of behavioural welfare indicators that form part of a working equine welfare assessment protocol. Over 4 years, behavioural and physical data were collected from 5481 donkeys, 4504 horses, and 858 mules across nine developing countries. Behaviours included the animals' general alertness, and their responses to four human-interaction tests, using the unfamiliar observer as the human stimulus. Avoidance behaviours correlated significantly with each other across the human-interaction tests, with 21% of animals avoiding the observer, but they showed no associations with likely anthropogenic injuries. Over 13% of equids appeared [`]apathetic': lethargic rather than alert. Measures of unresponsiveness correlated with each other across the five tests, and were associated with poor body condition, abnormal mucous membrane colour, faecal soiling, eye abnormalities, more severe wounds, and older age, depending on the equine species. This suggests that working equids in poor physical health show an unresponsive behavioural profile, consistent with sickness behaviour, exhaustion, chronic pain, or depression-like states.

Abstract: Unwanted oral investigative in horses has been anecdotally attributed to the practice of hand-feeding. Fears over such behaviour developing as a consequence of using food rewards, for example in clicker training, have been implicated as a common reason for not employing food-based positive reinforcement training techniques. This study used data generated as part of a larger research project, and explored associations between five common oral investigative behaviours and the practices of hand-feeding and clicker training. Data were from a convenience sample of UK leisure horse owners using two self-administered Internet surveys. Ninety-one percent of respondents reported giving their horse food by hand and this practice was significantly associated with three of the five oral investigative behaviours, licking hands (P = 0.006), gently searching clothing (P < 0.001) and roughly searching clothing (P = 0.003). Nipping hands and biting clothes were not associated with hand-feeding, suggesting that risk factors for these behaviours originate outside of this practice. Clicker training techniques were employed by 14% of respondents and their use was not associated with the incidence of any of the five oral investigative behaviours. These findings suggest that horse owners should not be deterred from using food-based positive reinforcement techniques with their horses, as fears that this practice will result in unwanted oral investigative behaviours from their horses appear unfounded.

Abstract: Using heart rate (HR) measurements we investigated whether potential calming effects of social licking were evident for both active (performers) and passive (receivers) licking partners. A HR decline was assumed to indicate relaxation and thus the experience of positive emotions. Effects of the licking category (spontaneous, solicited), the animals' basic activity (standing, lying) and the licked body region (head, neck, rest) were also considered. Two studies (A, B) were carried out in the same loose housed Austrian Simmental dairy herd. HR was recorded in up to 20 focal animals on 16 and 18 days, respectively. Using either direct observations (A) or video recordings (B), social licking interactions were continuously observed. The cow's basic activity was recorded using scan sampling at 5 min intervals. Linear mixed effects models were applied separately for Study A and B in order to compare the mean HR of the licking bouts with the mean of the respective 5 min pre- and post-licking periods. In receivers we found a significant calming effect in terms of a HR decline during allogrooming in both studies (A: -1.3 beats per minute, B: -1.1 bpm). This effect was more pronounced when animals were standing (A/B: -2.4 bpm/-3.8 bpm). However, it was not affected by the licked body region. In dairy cows performing social licking, we did not find an overall calming effect. On the contrary, in Study B, HR significantly increased during licking in lying performers (+2.5 bpm). This reaction was even stronger, when licking was directed to the receivers' head (+3.5 bpm) or neck (+3.0 bpm) as compared to the rest of the body (+1.4 bpm). The licking category had no effect on HR changes during the licking events. Our findings suggest that relaxation effects induced by social licking differ between performers and receivers and are affected by the cows' basic activity. In receivers, there were clear indications of a calming effect implying the experience of positive affective states. In performers, such calming effects during social licking were not identified.

Abstract: A previously described (Windschnurer et al., 2009) avoidance distance test was used to assess animals’ fear of humans in order to quantify the human–animal relationship (HAR). This study investigated the influence of the dominant and flightiest animals within a group on the responsiveness of animals during the avoidance distance test. Eighty-eight pregnant heifers comprised of four different genotypes were used (22 animals per genotype): Limousin × Holstein-Friesian, Limousin × Simmental, Charolais × Limousin, and Charolais × Simmental. Sixty of the 88 heifers were group housed (n = 5) into 12 pens with 3 pens per breed, while 28 heifers were singly housed (seven heifers per breed). A reactivity test was performed on days 10, 18, 25 and 30 post-housing on the singly housed heifers, and then on the group housed heifers, on the same days, to calculate a reactivity score. On days 33 and 37 flight and dominance tests, respectively, were performed to identify the flightiest and the dominant animal within each group. On day 41, an avoidance test, measuring both the avoidance distance towards a familiar and an unfamiliar human, was performed on all heifers. No difference (P > 0.05) in reactivity scores was found between the genotypes, between pens for the group housed heifers or between singly housed and group housed heifers (P = 0.28). The avoidance distance (AD) of singly (S) housed heifers towards a familiar (F) (ADSF) human was shorter (P < 0.001) than the avoidance distance of group (G) housed heifers towards an unfamiliar human (ADSU). The ADSF and ADGF were correlated with the ADSU and ADGU (R = 0.87 for singly housed heifers; R = 0.61 for group housed heifers, P < 0.001). For the singly housed heifers, no correlation was observed between reactivity score and ADSF (R = 0.36, P = 0.18), whereas the reactivity score and ADSU were correlated (R = 0.68, P = 0.004). For the group housed heifers no significant correlation was detected between the reactivity score and ADGF (R = 0.18, P = 0.22) or ADGU (R = &#8722;0.11, P = 0.39). No influence of the most dominant animal and the flightiest animals was found on the behaviour of the group in term of avoidance distance and reactivity (P > 0.05). It is concluded that the assessment of the fear of the animals towards humans using the avoidance test at the feed bunk may be useful for singly and group housed heifers and that the leaders of a group such as the flightiest animal or the dominant animal did not influence the avoidance distance test.

Abstract: The effects of riding style and various management factors on the prevalence of stereotypies and other behavioural problems among 346 mixed-breed saddle horses (phase 1) and 101 Arabian horses (phase 2) were analysed through a questionnaire answered by owners. In phase 1, the questionnaire data were partially validated through 20-min observations of 81 (23.3%) of the cases. Results indicate that horses primarily ridden in the English style were reported to be significantly more likely to display stereotypies (p < 0.001), problems when transported (p = 0.001), multiple behavioural problems (p < 0.001), and to have more restrictive stabling (p < 0.001) than horses ridden with other styles. When only Arabian horses were assessed in phase 2, however, there was no significant difference in behavioural problems between the Arabian horses ridden English style versus other riding styles. However Arabian horses were housed less restrictively than horses in phase 1 and English riding style and restrictive stabling tended to exacerbate each other's association with stereotypies. Management-related effects were found when, e.g., horses housed in restrictive stabling were more frequently reported to show locomotion stereotypies (p = 0.02) and those denied ad libitum hay displayed stereotypic wood-chewing behaviour (p = 0.02). To aid diagnosing and prioritizing interventions and care, the most predictive subsets of factors were computed for the various problem behaviours. E.g., among saddle horses, a statistical model comprised of the main riding style, duration of access to a paddock, and horse's age predicted whether a horse was reported to display any behavioural problem 62% of the time. This study supports the effects of management and handling on the prevalence of behavioural problems, and helps prioritize the relative importance of broad management categories on equine welfare. In particular, it underscores the importance of riding style on the well-being of saddle horses.