Ž .
Applied Animal Behaviour Science 68 2000 319–337 www.elsevier.comrlocaterapplanim
Aggression and social spacing in light horse ž
Equus caballus mares and foals
J.W. Weeks
a,
, S.L. Crowell-Davis
a
, A.B. Caudle
a
, G.L. Heusner
b
a
College of Veterinary Medicine, The UniÕersity of Georgia, Athens, GA 30602, USA
b
Department of Animal and Dairy Science, The UniÕersity of Georgia, Athens, GA 30602, USA Accepted 8 December 1999
Abstract
Aggression and social spacing were studied in 14 light horse mares and their foals living at pasture. Focal samples were collected on each mare–foal dyad for 6 to 10.5 h from 2 months of
foal age until weaning at approximately 4 months of age. Observations on foals continued until approximately 6 months of age for 7.5 to 10.5 h per foal. Every 2 min the identities of all
individuals within 5 m were recorded. All occurrences of agonistic behavior, and the participants, were recorded during the focal samples. In addition, during feeding of supplemental grain, all
occurrences of agonistic behavior by all subjects were recorded. Significant correlations were found between mare rank and the rank of foals both prior to and after weaning. Before weaning,
the rank of the foal was significantly correlated with birth order. No significant correlation between birth order and foal rank was found for the post-weaning hierarchy. An animal’s gender
had no significant effect on foal rank or the choice of preferred associate. Both prior to and after weaning, foals associated preferentially with the foal of their dam’s most preferred associate. In
addition, significant positive correlations were found between rank of mares and foals and the rate at which they directed aggression to other herd members. q 2000 Elsevier Science B.V. All rights
reserved.
Keywords: Horse; Social ontogeny; Sex differences; Dominance relationships
1. Introduction
Since the introduction of the concept of social dominance or ‘‘peck-order’’ Ž
. Schjelderup-Ebbe, 1922 , a great deal of debate has ensued revolving around the
Corresponding author. 0168-1591r00r - see front matter q 2000 Elsevier Science B.V. All rights reserved.
Ž .
PII: S 0 1 6 8 - 1 5 9 1 9 9 0 0 1 2 6 - 4
Ž .
concept Bernstein, 1981; Drews, 1993; Dewsbury, 1994 . The debate involves several key issues of dominance, the first being the actual definition of dominance. Dominance
Ž .
Ž has been defined as aggressiveness Wilson, 1975 ; a trait that conveys rank Baen-
. Ž
. ninger, 1981 ; priority of access to resources Wilson, 1975 ; and peck order, in which
Ž there is a consistent unidirectional pattern of agonistic interactions Schjelderup-Ebbe,
. 1922 . The concept of peck order has been melded into a modified peck order that takes
Ž into account more types of agonistic interactions and is less species-specific Drews,
. 1993 .
It is usually assumed that dominant animals accrue many benefits from their status. Dominant animals usually have a priority access to preferred resources such as food and
mates. However, there are now some questions as to the costs and benefits of being a high ranking animal. Recent studies have shown that dominant animals have higher
Ž .
levels of stress hormones than lower ranking animals Creel et al., 1996 . These elevated cortisol levels have been noted in female mongooses and wild dogs. The chronic
elevation of stress hormones could lead to a shorter life span for high ranking animals, as well as decreased reproductive ability. Therefore, the benefits of being a dominant
animal may not come without a cost. Further research will determine what these costs may be.
Even though there exists a great deal of disparity in reference to the concept of dominance, there is no debate about the existence of dominance relationships in many
socially living species. The study of the many facets of dominance in numerous species has created a wealth of data on the subject.
It is well established that feral horses, when in groups of two or more, develop and Ž
maintain dominance relationships Tyler, 1972; Clutton-Brock et al., 1976; Berger, 1977; Wells and vonGoldschmidt-Rothschild, 1979; Rubenstein, 1981; Waring, 1983;
Ellard and Crowell-Davis, 1989; Rutberg and Greenberg, 1990; Araba and Crowell- .
Davis, 1994 . These dominance relationships can take many forms. Dominant stallions Ž
may gain exclusive breeding capabilities over younger or subordinate stallions Feist and .
McCullough, 1976; Miller, 1981; Waring, 1983 . Particular bands of horses can be dominant over other bands. These dominant bands have priority of access to resources
Ž .
such as water, food and resting places Tyler, 1972; Feist and McCullough, 1976 . Ž
. Within each band there exists a dominance hierarchy Waring, 1983 . Feral horses are
seen most often in harem bands in which the stallion may be the dominant individual Ž
. Ž
Feist and McCullough, 1976; Miller, 1981 , though this is not always the case Houpt .
and Keiper, 1982 . Among the mares of the harem, there is normally a dominance Ž
. hierarchy Tyler, 1972; Houpt and Keiper, 1982 . This hierarchy is maintained through
Ž agonistic interactions and is relatively stable Wells and vonGoldschmidt-Rothschild,
. 1979 . Most changes in the dominance hierarchy among feral populations are due to
Ž .
changes in the younger age classes Wells and vonGoldschmidt-Rothschild, 1979 . Colts, upon reaching reproductive maturity, will often leave the natal herd and either
Ž .
join a bachelor herd or remain solitary Waring, 1983 . Fillies also will disperse and take Ž
. up membership in other harem bands Waring, 1983 .
Domestic herds under human management also form stable dominance hierarchies ŽHoupt and Wolski, 1980; Ellard and Crowell-Davis, 1989; Araba and Crowell-Davis,
. 1994 . Dominant individuals in domestic herds may have priority of access to resources
as well, though these resources are often not limited. Dominant individuals often Ž
maintain control over supplemental feed longer than subordinate horses Houpt and .
Wolski, 1977; Ellard and Crowell-Davis, 1989 and will often displace lower ranking individuals for access to food, water, or resting sites.
There are many factors that may influence the dominance rank an individual attains within its herd. Dominance rank has been found to correlate with age in feral horses
ŽTyler, 1972; Clutton-Brock et al., 1976; Wells and vonGoldschmidt-Rothschild, 1979; .
Rubenstein, 1981; Keiper and Sambraus, 1986; Rutberg and Greenberg, 1990 , in Ž
. Przewalski horses Feh, 1988; Keiper, 1988; Keiper and Receveur, 1992 , and in horse
Ž populations managed by humans Ellard and Crowell-Davis, 1989; Araba and Crowell-
. Davis, 1994 . In contrast, several studies have not found a significant correlation
Ž .
Ž between age and rank in feral Berger, 1977 and managed populations Houpt et al.,
. 1978; Haag et al., 1980 .
A horse’s body size is also thought to influence its dominance rank. Either height, Ž
weight or both have been found to correlate significantly with dominance rank Tyler, 1972; Berger, 1977; Houpt and Wolski, 1977; Ellard and Crowell-Davis, 1989; Rutberg
. Ž
and Greenberg, 1990 . As with age, these correlations are not found universally Houpt .
and Keiper, 1982; Houpt et al., 1978; Haag et al., 1980 . An integral part of the creation and maintenance of dominance relationships is
aggression. Significant positive correlations have been found between aggression and Ž
. dominance rank Houpt and Wolski, 1980; Houpt et al., 1978; Haag et al., 1980 .
Inverse correlations have been found between aggression rates and dominance rank in Ž
. foals Araba and Crowell-Davis, 1994 .
The rank of a foal’s dam may also influence the dominance rank a foal achieves. Two studies have reported a significant correlation between the dominance rank of mares
Ž within a mare herd and that of the foals born into the herd Houpt and Wolski, 1980;
. Araba and Crowell-Davis, 1994 . It is not know whether this influence is genetic,
environmental or a combination of both of these possibilities. It is not surprising that there are so many contrasting reports of the correlations
between individuals and their dominance status. There is no established norm for the measurement of dominance hierarchies in horses. Some hierarchies may be based solely
Ž on the occurrence of threats Wells and vonGoldschmidt-Rothschild, 1979; Keiper and
. Receveur, 1992 , while others use more complex requirements of interactions, using
Ž submissive responses as well Ellard and Crowell-Davis, 1989; Araba and Crowell-Davis,
. 1994 . Additionally, all of the studies were conducted on different types of horses
ranging from feral to human managed. Many different breeds of domestic horses were used as well. As there is little to no data reporting the similarities or differences in
behavior of different breeds of horses, it is assumed that their behavior is similar.
The dominance rank of an individual can influence the animals with whom it associates. A horse tends to associate with other individuals that are similar to it in age
Ž .
and rank Clutton-Brock et al., 1976; Ellard and Crowell-Davis, 1989 . One or more Ž
. preferred associates can usually be identified for herd members Estep et al., 1993 .
Preferred associates also appear to be chosen by gender. Foals will tend to preferentially Ž
associate with other foals of the same gender Crowell-Davis et al., 1986; Araba and .
Crowell-Davis, 1994 .
This study was designed to assess the determinants of dominance and preferred association in light breed mares and foals. Similar studies have previously been
Ž conducted to assess the determinants of dominance in draft horses Ellard and Crowell-
. Ž
. Davis, 1989; Araba and Crowell-Davis, 1994 and ponies Asa et al., 1979 . Studies
involving light breed horses have been of Thoroughbreds or groups of various breeds Ž
. Houpt and Wolski, 1980; Houpt et al., 1978 and were not conducted in the field.
Instead, these studies determined dominance relationships from the results of paired feeding tests. The following study used observations of spontaneous behavior in the field
to assess dominance and preferred associate relationships among light breed mares and foals. The following hypotheses were tested:
I. The dominance rank of the foal in the foal herd correlates with that of its dam in the mare herd, but not with its order of birth into the herd.
Ž .
II. The dominance rank of the foal correlates with its size height or weight , but not gender.
III. Foals associate most often with the foal of their dam’s most preferred associate. IV. The foal’s most preferred associate is more likely to be of the same gender than of
the opposite gender. V. Foals are more likely to aggress toward other foals than toward their most
preferred associate per unit time spent near them. VI. The rank of the foal positively correlates with its rate of giving aggression.
VII. The rank of the foal is negatively correlated with the rate at which its aggression is ignored.
VIII. Males and females differ in the type of aggression they most often display toward other foals.
2. Subjects and study site
