Practical considerations regarding the implementation of a temperament test into horse performance tests: Results of a large-scale test run

Considering the ever-growing demand of various breeding organizations for an objective, inexpensive, reliable, and easily conducted assessment of the behavior of horses, the aim of our study was to implement a novel-object test and a startling test into any kind of breeding performance testing to assess horses' temperament. Additionally, the influence of testing areas (familiar or unfamiliar), riders, and horse factors such as levels of training, breed, and age were of interest. Furthermore, recommendations for the practical implementation concerning the parameters should be given. Therefore, 1,028 horses over a period of 3 years participated in a temperament test consisting of 5 different stimuli. The horses were either ridden (61.8 %) or led by hand (38.2 %) by an unfamiliar professional rider (N = 43) or a familiar rider (N = 20). Live behavioral observations were taken by a trained observer. Overall, horses' scores for reactivity in the present temperament test were distributed over the whole scale, with lower means and higher standard deviations (6.7 ± 2.2-7.6 ± 2.1) than corresponding scores from the conventional personality evaluation in performance tests (7.7 ± 0.8-8.2 ± 0.5; P < 0.01). High correlations (r = 0.3-0.9; P < 0.001) between the scores for reactivity and the other behavioral parameters (emotional expression, activity, time to calm down, rider's aids) show a large influence of these parameters in assessing the horses' temperament. Factors like breed type, sex, and age had significant influences (P < 0.001) on different scores of the temperament test. In most cases, the rider or handler had no influence on the different scores assessed during the temperament test. The training level and the testing modus never had a significant influence on different scores. Only the testing station or location had a small influence on the scores for the stimulus “bridge” in some horses. Based on the results, it could be concluded that an implementation of a temperament tests into performance testing is possible during various types of testing procedure. Especially the assessment of reactivity, emotional expression, interest in the stimulus and rider's aids during and after passing the stimulus, as well as the time to calm down are important parameters for analyzing the horses' personality.     

Analysis of Behaviors Observed During Mechanical Nociceptive Threshold Testing in Donkeys and Horses

The aims of the study were to analyze and compare behaviors in horses and donkeys observed during nociceptive threshold tests with a mechanical stimulus applied to the limb. The purpose was to identify end point behaviors suggesting the animals had perceived the stimulus to be noxious. Six male castrated horses (aged 3–4 years, weighing 415–503 kg) and eight castrated male donkeys (aged 4–9 years, weighing 152.5–170.5 kg) were studied. Video data recorded during mechanical nociceptive threshold test were analyzed by a single observer. Behaviors were classified into short-duration event behaviors and longer duration activity/state behaviors. Frequency of behaviors within a test (event behaviors) and percentage time spent during the test (activity/state behaviors) were calculated. Data were compared between horses and donkeys using Mann–Whitney tests (nonparametric data) or t-test (parametric data). Significance was taken as P < .05. Behaviors during the tests were observed which could indicate the animals perceived the stimulus as noxious. These included flattening ears back against the head, and turning the head (horses) and chewing (donkeys) although these were not consistent across both species. Foot lifts were often preceded by other behaviors which suggests that the foot lift was not purely a reflex withdrawal response. A shift in weight toward the contralateral limb was a consistent prodromal sign for an end point foot lift.     

Effects of 2 training methods on stress-related behaviors of the dog (Canis familiaris) and on the dog–owner relationship

Instrumental learning plays an important role in dog–human interactions. The recent demand for pet dog training has resulted in the development of various training methods. The present exploratory study aims to compare the effects of 2 training methods on both the behavioral welfare of the dog and the dog–owner relationship: the first method is based on positive reinforcement (appearance of an appetitive stimulus), whereas the second method is based on negative reinforcement (disappearance of an aversive stimulus). The study compared behaviors linked to signs of stress and attentive behaviors toward the owner in 2 dog training schools, which used different methods. Walking on-leash activity and obeying the “sit” command were studied. The results show that dogs from the school using a negative reinforcement–based method demonstrated lowered body postures and signals of stress, whereas dogs from the school using a positive reinforcement–based method showed increased attentiveness toward their owner. However, neither method affected avoidance behaviors. This exploratory study reveals the differential effects of the 2 training methods on dogs' behaviors; it suggests that training methods based on positive reinforcement are less stressful and potentially better for their welfare.     

Importance of personality traits in horses to breeders and riders

Especially in horses, personality traits play an important role because horses' behavior influences their quality as a riding partner. In contrast to that, no objective assessment of horses' personality traits is available at present. Although initial efforts are made in this field, a successful implementation of behavior tests into horse performance tests depends on the acceptance of the riders and breeders. Therefore, the aim of the present study was to assess the importance of personality traits to breeders and riders as well as the degree of riders' acceptance of a temperament test as a more objective means of assessing equine personality traits. Using a web-based survey consisting of a 41-item questionnaire, a total of 1087 competition riders (49.3%), breeders (39.0%), leisure riders (37.9%), and professional riders (8.6%) of 13 countries were recruited to participate in the survey. When asked to split 1000 Euro among the different traits listed in the breeding goal, respondents clearly assigned more weight to the personality-related character and temperament traits (least squares mean ± standard error; P < 0.005: €228.7 ± 17.6) and willingness to work (€123.0 ± 9.6) compared with performance traits, such as the quality of trot (€77.7 ± 6.9) or show jumping (€68.0 ± 12.3). Nevertheless, expected differences in relative weighting of traits between the different groups of riders were confirmed (e.g., character and temperament: €209.3 ± 6.1 [leisure riders] vs. €149.7 ± 5.4 [competition riders], P < 0.0001). When asking why personality traits are so important, the simplification of daily work with the horses (47.9%) and relationship between horse and human (44.9%) as well as a more comfortable and safer handling (31.5%) were most commonly listed. As much as 45.6% of all participants see quality problems with the current assessment and suggested the evaluation of all breeding animals (30.1%), followed by a better standardization of assessment procedures (25.5%) and a move to more objective criteria such as the introduction of a temperament test (20.3%) for solving the problems. The present survey revealed that behavior traits are very important to all groups of riders and breeders, although there are diverse opinions about it. According to the participants, there is a need for and a high potential in the move toward more objective assessment methods of horses' personality traits, and participants would support a restructuring of the current assessment.     

Horses Could Perceive Riding Differently Depending on the Way They Express Poor Welfare in the Stable

This study investigated the relationships between four behavioral and postural indicators of a compromised welfare state in loose boxes (stereotypies, aggressive behaviors toward humans, withdrawn posture reflecting unresponsiveness to the environment, and alert posture indicating hypervigilance) and the way horses perceived riding. This perception was inferred using a survey completed by the usual riding instructor and during a standardized riding session (assessment of behaviors and postures, qualitative behavior assessment (QBA) and characterization of the horses’ locomotion using an inertial measurement unit). In accordance with ear and tail positions and the QBA, stereotypic and the most hypervigilant horses in loose boxes seemed to experience a more negative affective state during the riding session compared with nonstereotypic and less hypervigilant animals (P < .02 in all cases). Horses which were aggressive toward humans in loose boxes had higher scores regarding the occurrence of discomfort and defensive behaviors on the survey than nonaggressive horses (P = .03). They also presented higher dorsoventral accelerations at a canter during the riding session (P = .03), requiring the rider to increase his spinal movement (P = .005). These results suggest that aggressive horses may be harder to ride than nonaggressive animals. The expression of unresponsiveness to the environment in loose boxes was related to more reluctance to move forward, as assessed in the survey (P = .006). This study suggests that a compromised welfare state in the stable is related to horses having a more negative perception of riding. This perception could vary depending on the expression of poor welfare.     

Effect of owner experience,living environment,and dog characteristics on owner reports of behavior of Argentine Dogos in Italy

Previous research reported contrasting results on the effect of owner experience on dog behaviors. This may be because of dog breed genetic variability, differences in expectations, management choices, and attitude toward their dogs by owners of dogs of different breeds. A study of a single breed may help provide a better insight into this issue. A convenience sample of 94 participants, owners of 181 Argentine Dogos, was recruited to provide information on their dogs' behaviors through a questionnaire. Participants were classified according to their previous general experience or inexperience with dogs and whether they had previously owned an Argentine Dogo (expert vs. naive owners). Reasons for getting a dog (companionship, guarding, hunting, breeding) and attendance at dog-training classes were also explored. Pearson chi-square tests and binary logistic regressions were used to analyze the data. Previous inexperience with dogs was associated with a higher prevalence of dogs that were fearful of children and unknown dogs. Naive Dogo owners perceived that their dogs were more obedient and friendly toward strangers but aggressive toward children, than those belonging to Argentine Dogo experts. It also appeared that living environment and the dog's age are critical predictors of many outcome variables. Kennel living was found to be predictive of car and food protection, owner-directed aggression and aggression toward children, whereas house living was associated with fear of loud noises. Changes related to the dogs' age were probably related to behavioral development in the dogs. Destructiveness, excessive body licking, and protective aggression were more common in younger dogs, whereas aggression to unknown dogs was more common in older ones. Protective aggression was more common in males, and fear of startling noises was more common in females. Intact dogs showed a higher probability of having attention-getting behavior, and neutered dogs were more fearful of traffic. Our findings are correlative. Even if causative links cannot be made, our findings certainly provide direction for further investigation.     

Comparison of a 2-step insulin-response test to conventional insulin-sensitivity testing in horses

Equine insulin resistance is important because of its association with laminitis. The insulin-response test is described to diagnose insulin resistance in clinical settings. Practitioners may be reluctant to perform this test because of the time needed for the test and the fear of inducing hypoglycemia. The objective of the study was to compare a 2-step insulin-response test with a complete insulin-response test. A complete insulin-response test was performed on 6 insulin-resistant horses and 6 controls. A 2-step insulin-response test consisting of an intravenous injection of 0.1 IU/kg human insulin and blood glucose determination at 0 and 30 min after injection was performed on the same horses. Times to reach a 50% reduction of glucose baseline were compared between tests and horses. All the horses tolerated both tests well. No significant difference was observed between baseline glucose concentrations of insulin-resistant horses and controls (P = 0.09). Time to reach 50% reduction of glucose baseline for controls was not significantly different with the use of the complete insulin-response test or the 2-step test (P = 0.98). For insulin-resistant horses, the time to reach 50% reduction of glucose baseline with the use of the 2-step test was significantly longer than for controls (P = 0.004). With a cut-off time of 30 min, the 2-step test had the same characteristics as the complete test. The 2-step test provided a safe, rapid, and low-cost method to diagnose insulin resistance in horses in a clinical setting.     

Molecular detection of Coxiella burnetii in horse sera in Iran

Coxiella burnetii is a zoonotic bacterium that can infect a wide range of animals including horses. However, its circulation dynamics in and through horses are still unclear. The aim of this study was to evaluate prevalence of C. burnetii and its genomic characteristics in horse sera samples in the North of Iran (Golestan Province). The samples were collected in 2018 and the age, sex, and breed of each animal were recorded. Nested-PCR was used to detect C. burnetii based on the presence of the transposable gene IS1111. The results showed that 7.50 % (P < 0.05; 95 % CI: 0.5 %–0.12 %) of the examined sera samples were positive for C. burnetii. Based on the resuls, prevalence of C. burnetii in the age groupof < Years 1–5 (p-value <0.05, 95 % CI: 1 %–8 %) was less than the age group of >6 years old (p-value <0.05, 95 %, CI: 7 %–19.8 %). In previous studies, it was concluded that the horses' population in Golestan Province should be considered as an important factor in the epidemiology of Q fever and consequently in public health. Further studies should be implemented to evaluate if horses may be relevant indicators of zoonotic risk in urban and suburban endemic areas.     

Anti-insect defensive behaviors in equines post-West Nile virus infection

West Nile virus (WNV), a zoonotic mosquito transmitted Flavivirus, has had significant health effects on horses in the United States, with over 23,000 United States equine cases since the disease was first recognized in 1999. Previous research has focused on how this disease progresses and affects equids days to weeks post infection. The purpose of this study was to evaluate if permanent equine behavioral changes had occurred in horses that had recovered from acute West Nile fever or encephalitis. Specifically, we examined if surviving this disease caused changes in the defensive behaviors of the animal against biting and stinging insects, presumably because of neurological sequelae that can result from the infection. Results from behavioral observations and neurologic reflex testing suggest that long-term survivors of WNV do not show a change in the frequency or types of behaviors used compared to uninfected horses, supporting the concept that lasting deficits from WNV usually resolve within the following 1–3 years post-infection. However, microhabitat and grouping behavior did have a significant impact on the frequency of defensive behaviors, with indoor locales and larger groups of horses showing less insect avoidance behaviors. These principles may play a more pivotal role in protecting equines from biting insects and disease than thought previously.     

Sex differences in the response of yearling horses to handling by unfamiliar humans

In adult horses, pronounced sex differences in behavior exist, and many riders prefer male horses to mares. In the present study, we tested the hypothesis that sex differences in the response of horses to handling emerge before puberty, that is, in sexually immature animals. Warm-blooded yearling horses (6 males and 9 females) were exposed to a stationary human test (exposure to an unfamiliar person for 5 minutes) and a tolerance test (haltering and direct contact with the horse) on 5 consecutive days. The horses' behavior and heart rate were recorded. Male yearling horses displayed significantly less exploratory behavior than females during the stationary human test on days 1 and 2 (P < 0.05). Latency to halter and ease with which horses could be haltered were significantly affected by day (P < 0.05), and a significant interaction of day and sex was detected (P < 0.05; e.g., time for haltering: females, day 1: 23.8 ± 7.8 seconds [±standard error of the mean] and day 5: 21.1 ± 1.7 seconds; males, day 1: 53.6 ± 10.4 seconds and day 5: 19.4 ± 2.2 seconds). Heart rate of the horses in response to the test situations was significantly affected by day (P < 0.01) but not by sex. In conclusion, sex differences in behavior do not only exist in adult horses but are already present in young horses before puberty. When exposed to unfamiliar humans for the first time, male yearling horses display more caution than females.     

Dog attachment to man: A comparison between pet and working dogs

Shared activities are known to create a good relationship between dog and man. Therefore, it can be expected that animals working in harness with their handlers, such as search and rescue dogs, show a higher attachment to man.The aim of the current research was to assess whether there are differences between pet and working dogs in the attachment to owners/handlers.Forty dogs were involved, 26 pets and 14 search and rescue dogs (taken home when not working). Their attachment to owners/handlers was analyzed by using a modified version of the Ainsworth Strange Situation Test and a recall test, while owners and handlers filled in a questionnaire and were scored on the Lexington Attachment to Pets Scale.Statistical analysis revealed no statistically significant differences between pet and working dogs for behaviors indicating an attachment bond: behavior oriented to door/chair/shoe during the absence of owner/handler (which represents “search of absent person”; U = 132.000, P = 0.156), contact with/proximity to owner/handler (i.e., “contact maintenance effect”; U = 157.000, P = 0.478), whining during owner's/handler's absence (“protest at the separation”; U = 147.000, P = 0.288), and individual play (U = 146.000, P = 0.288) and exploration (U = 138.500, P = 0.215) in presence of owner/handler, representing the “secure base effect.”Despite the lack of statistical differences, the formulated hypothesis may be not completely rejected, as a trend of a higher attachment in search and rescue dogs existed. It can be concluded that in carefully owned dogs, the level of attachment is increased by working as a search and rescue dog, but the difference does not lead to statistically relevant results.     

Possible Relationship Between Performance and Oxidative Stress in Endurance Horses

The purpose of this study was to evaluate oxidative stress, antioxidant biomarkers, and performance during a multiday 210-km endurance race. Nine endurance athlete horses participated in this study. Samples were always taken at the same times of day, before the beginning of the race and after every day of competition. Analytic measurements included glutathione reductase (GR) and catalase activity, thiobarbituric acid–reactive substances (TBARs), and reactive carbonylated derivatives. Competition intensity was low, with an average speed of 12.56 ± 0.9 km/h. Four horses were unable to finish the race because of metabolic problems or fatigue. GR activity increased progressively (P < .001) throughout the competition, and TBARs showed a significant rise compared with baseline values (P < .01) but remained at the same levels throughout the 3 days of competition. Catalase and reactive carbonylated derivatives did not show any significant alterations in any time period. The best performance was obtained from horses who demonstrated higher GR capacity and/or lower TBAR concentration. In conclusion, redox status seems to modulate horses' performance in endurance races, but further studies are needed to better determine the adequate oxidant/antioxidant ratio to acquire optimal performance.     

Prevalence of aggression and fear-related behavioral problems in a sample of Argentine Dogos in Italy

The Argentine Dogo breed is identified frequently as potentially dangerous in worldwide legislation. To investigate the prevalence of problem behaviors in this breed, 94 Argentine Dogo owners provided information on aggressive and fearful behaviors shown by their dogs (83 males, 98 females) in a questionnaire. The results were consistent with the origin of the Argentine Dogo (i.e., descending from fighting dogs) and subsequent selection exerted to improve hunting qualities. Common aggressive behaviors included predation on small animals (92.0%), intra-specific conflicts (69.5%), and territoriality (i.e., vigilance of territory perimeter, 61.3%; aggression toward people visiting the dog's house, 45.3%; vigilance while inside the owners' car, 33.1%). Aggression directed to the owner was marginally represented. Reactions to leashing, bathing, punishment, sleep disturbance, toy removal, and in other conflict situations occurred in less than 10% of the sample. Reactions to food removal (13.3%) and handling (19.8%) showed higher prevalence. Chasing cars/running people (17.6%) and aggression toward strangers, either those who were friendly (19.1%) or those who seemed threatening on approach (29.4%), were noted. Low anxiety levels reported in social situations confirmed that these reactions were not common (e.g., fear of strangers, children, and unfamiliar dogs ≤10%). Common anxious behaviors were fear of traffic (10.3%), loud noises (36.2%), startling stimuli (30.5%), destructiveness (42.9%), excessive body licking (13.6%), and shyness in novel situations (32.4%). The context and targets of aggression when exhibited by these dogs (e.g., aggression toward animals when outside the dogs' territory, aggression towards people when inside territory) suggest that the Italian law on potentially dangerous dog breeds may be ineffective in preventing Argentine Dogos' aggression to humans. Social hazards could be reduced by promoting proper social exposure and reactions toward people.     

The interrelations of good welfare indicators assessed in working horses and their relationships with the type of work

This paper presents a different approach to the welfare of working horses through the indicators of good welfare, their interrelations and their relationships with the type of work performed by the horses. 697 working horses were assessed by using observation, behavioural tests, clinical exams and questionnaires. The correlated results highlight the complex and multiple interrelations between the good welfare indicators, showing once again that animal welfare is under the influence of all its three dimensions (physical and emotional wellbeing and naturalness). In working horses, the heavier the work they perform is, the more important it is for horse owners/breeders/farmers to fulfil their inherent needs so as to help them cope with the multiple specific risks imposed by their work.     

Modestly Improved Compliance and Apparent Comfort of Horses With Aversions to Mildly Aversive Routine Health Care Procedures Following Short-Term Alpha-Casozepine Supplementation

Alpha-casozepine, a decapeptide derived from bovine milk α-_S1 casein, has well-documented anxiolytic properties in several species. To evaluate potential benefit of alpha-casozepine to horses' compliance and comfort with routine management and health care procedures, we blindly compared behavior of alpha-casozepine–supplemented and control-supplemented horses with known specific aversions to specific health care procedures. Twenty-six light horse mares were first screened for aversions based on compliance and apparent comfort during a standard battery of 12 health care examinations and treatment procedures. Based on quantitative behavioral analysis of video-recorded sessions, baseline compliance and/or comfort with each procedure was scored from 0 (unable to progress) to 10 (excellent compliance and relaxed comfort). Based on those results, 10 of the 26 horses were selected as five pairs that were matched for breed and the same two specific aversions. One of each pair was randomly assigned to alpha-casozepine supplementation (2,000 mg PO daily) and the other to control supplement for 5 days. On day 5 of supplementation, the standard battery of 12 procedures was repeated. For the five alpha-casozepine–supplemented subjects, compliance and/or comfort scores improved for seven of their 10 aversions compared with one of 10 for the five matched controls (P < .01, Fisher Exact). Average score increased 1.5 (standard error [SE] = 0.87) points for alpha-casozepine aversions compared with an average decrease of 0.92 (SE = 0.61) points for matched control aversions (P < .05, Wilcoxon signed ranks). These results indicate a modest benefit of alpha-casozepine supplementation to horses for improvement of compliance and apparent comfort with mildly aversive routine health care procedures.     

Effects of different head–neck positions on the larynges of ridden horses

Hyperflexion, that is the strong deflection of the horse's head, poll and neck, is a prevalent training technique in equitation. Hyperflexion has come under criticism in recent years for being suspected of affecting the horses' well‐being contrary to animal welfare. The goal of the present study is a comparison between the impacts of different poll‐neck positions on findings in the upper respiratory tract of ridden horses. For this purpose, video recordings of the larynges of 14 horses were taken using an overground endoscope. The videos were recorded at rest and during three different riding phases: firstly, in a stretching posture, secondly, in a working position and, thirdly, in hyperflexion. A comparison between the analyses of the working position and hyperflexion phases revealed a significant reduction in the laryngeal opening area (p = 0.001) with a value of 8.2 ± 5.0%. Furthermore, other parameters of the larynx evaluated also showed a significant diminishment. These changes did not correlate with the age of the horses or their level of education, and they were independent of the individual anatomical conditions of the poll‐neck region. In summary, it can be stated that hyperflexion causes a considerable compression of the larynx.     

Investigating duration of nocturnal ingestive and sleep behaviors of horses bedded on straw versus shavings

Horses are stabled overnight for a number of practical reasons; however, there is little research quantifying nocturnal equine behavioral patterns or the extent to which different environments influence nocturnal behavior. The aim of this study was to establish whether differences in duration of sleep and ingestive behaviors were apparent for horses bedded on straw (group 1) or shavings (group 2). Ten geldings of mixed breed (mean age: 7.3 ± 3.53 years) bedded on either shavings (n = 5) or straw (n = 5) were observed between 7 pm and 7 am. Duration of behaviors according to a predefined ethogram was recorded in minutes using a video recorder and continuous focal sampling. Mann–Whitney U tests were used to identify whether any significant differences in duration of ingestion and sleep behaviors occurred for horses bedded on straw compared with shavings. Of the total observation period, group 1 spent, on average, 29.3% of their time budget engaged in recumbent behaviors, compared with 12.2% for group 2. However, no significant differences in duration were established between horses bedded on straw or shavings for standing sleep, sternal recumbency, and lateral recumbency behaviors (P > 0.05). Ingestive behaviors occupied approximately one-third of the time budget, with no significant difference (P > 0.05) observed between groups. On average, group 1 spent a longer proportion of the observation period ingesting bedding (8.1%) compared with group 2 (1%). Duration of bedding ingestion appeared to peak between 1 am and 7 am for both groups. Although not quantified, general observations revealed horses were motivated to alternate between eating hay and bedding in both groups, owing to the prevalence of bedding ingestion. The results indicate that straw bedding facilitates the display of ingestive and sleep behaviors, whereas horses bedded on shavings spent a greater proportion of their nocturnal time budget engaged in “other” behaviors. Further research is required to investigate the extent to which different types of bedding material enrich the environment of horses that are stabled overnight.     

Experimental evaluation of attachment behaviors in owned cats

Attachment, a normal behavior among social animals, is quite significant since owners worry about their pets and take care of them because of this affective connection. There are not enough research studies that focus on attachment between owners and their cats. The general objective of this study was to identify attachment behaviors, directed toward their owners, in cats of different body types, age groups, and sexes in an experimental situation.Twenty-eight cats, ranging from 1 to 7 years of age and having different body types, were used in the study without taking into account sex or reproductive status. These cats underwent an Ainsworth’s Adapted Strange Situation Test. Event frequencies and behavioral state durations in individual type behaviors such as exploration/locomotion, alertness, and inactivity were registered using direct focal sampling. For data analysis, cats were divided by body type, sex, and reproductive status. Analysis of variance (ANOVA) of locomotion/exploration revealed a statistically significant difference (N = 28, F = 13.55, P < 0.001) between the episodes with the owner, alone, and with a stranger with cats spending more time engaged in locomotion/exploration while accompanied by their owner. On the alert behavior event frequency, difference (ANOVA, F = 7.44, P < 0.05) was found, which showed a higher frequency while in the company of a stranger. Last, in the inactivity time ratio, a significant difference was found (ANOVA, F = 18.55, P < 0.001), where the time spent on this behavior was considerably higher when the animal was alone.These results are consistent with the ones obtained by Ainsworth in children attached to their mothers; therefore, it can be said that cats can manifest attachment behaviors toward their owners. Further studies are indicated to see whether cats can develop separation anxiety.     

To jump or not to jump? Strategies employed by leisure and sport horses

Show jumping is one of the most popular equestrian disciplines although strategic jumping per se is seldom used by free-living horses when negotiating obstacles that they might otherwise avoid. To establish if horses are naturally motivated to jump, we examined horses in a free-choice situation when negotiating various obstacles under test conditions. Eighteen leisure horses (LHs) and 16 sport horses (SHs) participated in a number of 2-choice tests. First, horses were presented with 2 optional routes during a “free-choice test” to reach a food stimulus. The choices were a shorter route (over an obstacle) or a longer route (around the obstacle). During 8 consecutive trials, the horses encountered an increase in obstacle height on every second trial as follows: 0, 20, 35, and 50 cm. In the “ridden test” after an initial conditioning period, the horses were tested with a Y-maze formation during 3 consecutive trials, where one element of the maze contained an obstacle arm (OA). The horses walked or trotted over the obstacle most frequently (59.9%) and jumped the obstacle in only 10.7% of the cases. For both the LH and SH groups, the horses' motivation to traverse the obstacle decreased as the obstacle height increased and only 44.1% of horses actually negotiated the 50-cm obstacle as compared with the 20-cm obstacle (85.3%). The LH group preferred to go around the obstacle significantly more often than the SH group (24 trials, 16.7% vs. 9 trials, 7.0%, for SH and LH, respectively; χ² = 6.81, N = 33, P = 0.0090). The LH group used a jumping strategy to clear the obstacle far less frequently (6 trials, 4.2%) than the SH group (23 trials, 18.0%; χ² = 9.96, N = 29, P = 0.0023). In the Y-maze trials, the LH group exhibited preference (40 per 54 trials, 74.1%; χ² = 12.5, P < 0.0001), and the SH group failed to exhibit any preference (20 per 48 trials; χ² = 1.33, P = 0.3123). There was no evidence of any correlation between the motivation to clear the obstacle and total number of OA choices in either the LH group (r_s = 0.13, P = 0.9594) or the SH group (r_s = ?0.25, P = 0.3492). The findings from the present study indicate that SHs are motivated and willing to jump obstacles more often than are LHs under similar conditions. However, the apparent reluctance of the horses overall to continue jumping as the obstacle height increases suggests that, in general, many horses could easily encounter excessive demands (overfacing) in sport. This issue should be carefully monitored in terms of equine training, competition, and welfare.