Consideration of the welfare of farm animals has been focused on physical well-being for most of the last century. In recent years, the need to address social needs as well has received increased attention [1
]. By nature, horses are highly social animals, depending on the group for survival. Social bonds are likely to play an important role for social cohesion in group living domestic horses [3
]. Management practices that minimizes aggression and give the horses ample opportunities to take part in affiliative interactions are to be recommended. In this study, data from several independent studies of social interactions among pastured horses in Iceland are used in order to identify factors that are associated with low levels of agonistic behaviour but enhance opportunities for behaviour that strengthens bonds among horses, such as mutual grooming.
Horses are highly social and as recent research has shown [4
] their cognition level and learning abilities are well advanced. Unusually long-lasting bonds between individuals in a family band [6
] may have contributed to the evolution of complex cognitive skills. Affiliation bonds between individuals are inferred from the choice of preferred allogrooming partners, play partners, a tendency to stay with certain individuals when resting, grazing or traveling or from affiliative approaches [7
]. It has been argued that close bonds such as these are of the same nature as between friends among humans and should, therefore, be called friendships [11
]. Horses have many ways of recognizing their group members [13
], for instance by employing long-term memories of previous interactions [15
] and using their ability to discriminate between individuals [16
] and recognize social status of familiar horses relative to their own [18
]. It is fair to conclude that horses can predict the outcome of encounters with familiar individuals and find ways of reducing tensions [19
]. Studying their social interactions and comparing the nature and frequency of such interactions in different social environments can improve understanding of the cognitive abilities of horses.
Group stability with respect to membership is high in most semi-feral groups [22
]. This allows both stable dominance relationships and stable networks of friendships, the latter most often established between two horses of similar age and the same sex [6
]. With frequent changes in group composition, levels of interactions are higher, especially agonistic interactions between resident horses and newcomers [24
]. Such a lack of group stability is considered to be one of the risk factors for potential injuries [25
Feral horses show low levels of aggressive interactions [6
]. The same is true for groups of high stability kept under semi-natural conditions [23
]. Fureix et al. proposed that the reasons for low aggression frequencies in such groups are the following: evolved ritualized communication signals, stable group composition, stable dominance hierarchies and opportunities for young animals to learn social skills [25
Much less is known about frequencies of allogrooming than agonistic behaviours. The frequency of mutual grooming is very variable and may depend on a variety of factors, such as weather, parasites and social factors [21
]. Different methodologies and limited information in published papers often make direct comparison of data impossible [9
]. In addition to serving the role of forming social bonds [11
], allogrooming following agonistic events is likely to reduce tension [28
]. Feh and Maziéres [7
] and Keiper [29
] suggest that allogrooming lowers both heart rate and blood pressure and reduces social stress in horses.
In contrast to the feral/natural situations where one stallion (occasionally two or three) keeps breeding mares with foals and their sexually immature offspring in a band (on average four to six horses [3
]), groups of domestic horses vary greatly with respect to composition and environment. Typically, groups of horses are kept in pastures or in open stables or are let outdoors into paddocks for part of the day. These groups can be composed of peers only or mixed with respect to sex and age [24
]. The composition of a group is known to influence the type and frequencies of social interactions [24
]. Other factors are also known to be influential, such as the stability of the group, access to resources, density and the presence of a stallion [13
]. However, as Hartmann et al. have pointed out, research on the effects of group size, group density and group composition on the social behaviour of horses is very limited [13
]. In the present study, our main aim is to investigate how group composition and other variables (see below) influence the frequencies of both agonistic and affiliative behaviour in horses. The data are based on independent studies of 20 groups of Icelandic horses, carried out over a period spanning 15 years and involving a total of 426 horses, all kept on pastures of similar vegetation. The same methodology was applied in all studies. All the members of the studied groups belong to the Icelandic breed, which is the only breed found in the country, making the question about possible breed characteristics irrelevant. Here we use the individual data to show the overall frequency distributions of different types of behaviour, compare the sexes and estimate the effect of age. The group factors considered are group size, density of horses in the pasture, sex ratio, proportion of adults, numbers of young foals present, median number of friends, group stability and the presence of stallions. The environmental factors considered are season and whether or not hay was provided
Results from the majority of the studies have been published previously [9
]. The first published study in 2003, which featured a mixed, stable group without a stallion, indicated that the horses were more socially active compared to studies conducted in other countries of groups with a stallion [9
]. Later, a comparison between four mixed groups in Iceland without a stallion and six groups with a stallion confirmed this; in groups where a stallion was present, the horses showed less agonistic behaviour [23
]. Stallions play an important role by herding their mares and interacting with youngsters. Their behaviour might, therefore, suppress the mares’ social activities. However, the same study showed that the stallions were not dominant over the mares and they did not intervene actively during ongoing interactions between the mares [23
By combining the data from the 20 groups, we intend to strengthen previous analyses of relations between social behaviour and group features. Furthermore, the combined data may reveal unknown influences of group features on social behaviour. It is hoped that the results will encourage horse owners to take up management practices that keep levels of aggression at a minimum level and increase opportunities for affiliative interactions, thus improving horses’ welfare.
This study emphasizes the need for horse caretakers to consider group stability and group composition in domestic horse management. This study is based on observations of horses held in spacious pastures but the findings also apply to the organization of group housing and the use of paddocks. Group composition which is similar to the natural social system with both sexes (sub-adults and foals) and stable membership of adult horses is likely to provide the best social environment for all horses because of very low aggression levels. Also, it is likely to offer the best conditions for young horses to learn social skills, because of the presence of older and experienced horses and the possibilities to associate with peers Such groups, or groups where stallions are replaced by adult geldings, could easily be taken up as a management practice. Peer groups should be avoided, especially if composed of young unfamiliar horses, because of high levels of aggression.
The level of allogrooming was more similar in the different group types than the frequency of agonistic behaviours. The two small unstable yearling groups had the highest allogrooming levels. The horses allogroomed less in large groups where they had more friends, but interestingly the horses that had few friends, allogroomed most.
Although only descriptive statistics were applied because of the nature of the data, the results of this paper are very interesting and should, first, provide a strong message to horse owners and caretakers and, second, encourage further research to test which factors are of the greatest significance. The importance of the welfare of horses can never be overemphasized. Recent research on the cognitive abilities of horses and their emotions [47
] has shown that horses are more complex in these areas than previously thought. It is our belief that awareness of such findings will affect both personal views and legislation on horse welfare in the near future.