Yawning is rare in herbivores which
therefore may be an interesting group to
disentangle the potential function(s) of yawning
behaviour. Horses provide the opportunity to
compare not only animals living in different
conditions but also wild versus domestic
species. Here, the authors tested three
hypotheses by observing both domestic and
Przewalski horses living in semi-natural
conditions: (i) that domestic horses may show an
elevated rate of yawning as a result of the
domestication process (or as a result of life
conditions), (ii) that individuals experiencing
a higher level of social stress would yawn more
than individuals with lower social stress and
(iii) that males would yawn more often than
females.
The study involved 19 Przewalski horses
(PHs) and 16 domestic horses (DHs) of different
breeds living in large outdoor enclosures. The
results showed that there was no difference
between the PH and DH in yawning frequency (YF).
PHs exhibited much higher levels of social
interactions than DHs. There was a positive
correlation between yawning frequency and
aggressive behaviours in PHs, especially males,
supporting the idea that yawning may be
associated with more excitatory/stressful social
situations. A correlation was found between
yawning frequency and affiliative behaviours in
DHs, which supports the potential relationship
between yawning and social context. Finally, the
entire males, but not castrated males, showed
much higher levels of yawning than females in
both species. The intensity (rather than the
valence) of the interaction may be important in
triggering yawning, which could therefore be a
displacement activity that helps reduce
tension.
Résumé
Le bâillement est rare chez les
herbivores qui peuvent donc être un groupe
intéressant pour distinguer les fonctions
potentiellement différentes de ce
comportement. Les chevaux offrent la
possibilité de comparer non seulement des
animaux vivants dans des conditions
différentes, mais aussi des
espèces sauvages et des espèces
domestiquées.
Les auteurs testent trois hypothèses
en observant les chevaux domestiqués et
des Przewalski vivants dans des conditions
semi-naturelles:
-que les chevaux domestiques peuvent
présenter un taux élevé de
bâillements à la suite du processus
de domestication (ou en raison des conditions de
vie)
- que les individus connaissant un niveau
plus élevé de stress social
bâillerairent plus que les individus moins
de stress
-que les mâles bâilleraient plus
souvent que les femmelles.
L'étude a impliqué 19 chevaux
de Przewalski (PHS) et 16 chevaux domestiques
(EDS) de différentes races vivant dans de
grands enclos en plein air. Les résultats
montrent qu'il n'y a pas de différence
entre le PH et DH en fréquence des
bâillements (YF). PHS présentent
des niveaux beaucoup plus élevés
d'interactions sociales que EDS. Il existe une
corrélation positive entre
bâillement et fréquence des
comportements agressifs chez PHS, en particulier
les mâles, ce qui soutient l'idée
que le bâillement peut être
associé à plusieurs excitants /
situations sociales stressantes.
Une corrélation a été
trouvée entre la fréquence des
bâillements et des comportements
affiliatifs chez DHS, ce indique une relation
potentielle entre bâillement et contexte
social.
Enfin, l'ensemble des mâles, mais pas
ceux castrés, ont montré des
niveaux beaucoup plus élevés de
bâillements que les femelles dans les deux
espèces. L'intensité (plutôt
que la valence) de l'interaction peut être
important dans le déclenchement de
bâillements, ce qui pourrait donc
être « une activité de
substitution » permettant de réduire
les tensions.
Introduction
Yawning is a behaviour that is performed by
different vertebrate species (Siamese fighting
fish, lions and primates, Baenninger 1987; dogs,
Dreschel and Granger 2005) including humans
(Baenninger 1997). This physical event has been
described in detail in humans as being composed
of three distinct phases: a long inspiratory
phase, a brief acme and a rapid expiration
(Baenninger 1997). Yawning was studied in
rodents (e.g. Moyaho et al. 1995; Fundaro 1996),
carnivores (e.g. Bekoff 1974; Leyhausen 1979;
Joly-Mascheroni et al. 2008), birds (Gallup et
al. 2009, 2015) and, most extensively, non-human
(e.g. Deputte 1994; Hadidian 1980; Troisi et al.
1990) and human (e.g. Baenninger et al. 1996)
primates.
In non-human primates and rats, this
behaviour is more frequent in males (Berendsen
and Nickolson 1981; Troisi et al. 1990).
According to Smith (1999), yawning is related to
changes in the arousal state (i.e.
alertness/drowsiness, Baenninger and Greco 1991;
Greco et al. 1993), pandiculation (Gessa et al.
1966), thermoregulation and brain cooling
(Gallup and Gallup 2007, 2008) and/or can
correspond to the expression of social status
(Deputte 1978) or stress (Maestripieri et al.
1992; Beerda et al. 2000). It has also been
proposed that yawning could, in some contexts,
be a type of displacement behaviour (Tinbergen
1952; Troisi 2002).
In primates, increased yawning frequencies
were observed in groups with higher numbers of
conflicts, especially in cases of changes in
group composition (hence establishing hierarchy
and increased social stress), primarily in
males. This observation has led to the social
stress-related hypothesis of yawning (Troisi
2002; Deputte 1994). Deputte (1978) classifies
the yawning in non-human primates into three
types: (1) physiological yawn, related to
sleep/awakeness activity; (2) stress yawn,
occurring around conflict/frustration
situations; and (3) threat yawn. As compared to
primates and carnivores, yawning occurs at a
lower rate and has been described in a limited
number in herbivore species (Baenninger 1997;
Gallup 2011) for which the context and potential
functions are therefore still less known. Recent
research on domestic horses has, however,
evidenced a co-occurrence of yawning and
stereotypic behaviour in restricted domestic
situations, indicating frustration as a
potential indirect causal factor for those
behavioural patterns (Fureix et al. 2011).
Interestingly, the frequency of occurrence of
this behaviour was rather high (around 2 per h)
and did not differ between sexes which opposes
studies on other mammalian species (e.g.
Holmgren et al. 1980; Troisi et al. 1990).
However, with regard to sex differences, only
geldings (castrated males) and mares were
involved in the study of Fureix et al. (2011).
The authors argued that the restricted housing
conditions may be responsible for this rather
high frequency of yawning, but since no research
had been performed on horses living in natural
conditions, this was difficult to assess.
Additionally, it has been proposed that
domestication, due to selection of animals, may
change the frequency of occurrence of particular
behaviours, like, for instance, the decrease in
territorial competition or exaggerated
reproductive behaviour in domesticated species
(Price 1999). Thus, the idea that domestication
could have promoted more frequent yawning may be
further investigated by comparing the wild
species (Equus przewalskii) to the domesticated
species, in both cases within socially stable
groups of animals living in semi-natural
conditions.
The aim of the present study, therefore, was
triple: (i) testing the potential effects of
domestication by examining the frequency of
yawning in semi-natural conditions in wild
(Equus ferus przewalskii) and domesticated
(Equus caballus) species (and, secondarily, of
life conditions by comparing with published data
in restricted conditions of life); (ii) to some
extent, testing the social stress hypothesis
proposed by Deputte (1994) by exploring the
possible association between yawning and social
interactions, in particular agonistic behaviour
frequencies; and (iii) assessing sex differences
in yawning frequency in both species by
comparing males and females. Since testosterone
has been shown to be involved in yawning
frequency in other species and males being more
often involved in social rank competition, we
hypothesised that individuals (and potentially
males) in groups experiencing a higher level of
social stress would yawn more. The horse, having
different social structures from harems to
bacheloronly groups, is an interesting
non-primate mammal species to investigate
further potential determinants of yawning. Also,
yawning is considered to be an extremely rare
behaviour in ungulates (Gallup 2011). The
latter, as contradicted by the study of Fureix
et al. (2011), was then studied in two species
of Equidae: Przewalski and domestic horses.
Discussion
The present study, where social groups of
bothwild and domestic horses were observed,
reveals for the first time that in these
species, as in other species studied, adult
males yawn more than females and immature males.
It also shows that in natural conditions,
yawning is a rather rare behaviour in mares and
geldings with 0.07 occurrences per h, whether
wild or domestic horse species being considered.
However, when looking only at adult males of
both species, yawning frequencies were
comparable to the frequencies observed for the
domestic mares and geldings kept in restricted
living conditions (Fureix et al. 2011). The
frequency of social interactions was higher in
Przewalski horses, especially in bachelors, as
mentioned in earlier studies (Feh 1988;
Christensen et al. 2002) which may explain that
the unequivocal relationship between agonistic
behaviour and yawning frequency was found only
in male Przewalski horses. However, the
correlation found between yawning frequency and
affiliative social interactions in the domestic
horses indicates that yawning may, to some
extent, relate to social (probably excitatory)
contexts. Further studies are needed to look
more precisely at which aspects of the social
interactions may be involved (for example,
sniffing may precede different types of
interactions of affiliative or negative valence)
and whether the structure of yawning may change
accordingly.
Domesticated versus wild species:
domestication and impact of life conditions The
absence of species differences in yawning
frequency clearly indicates that domestication
did not influence the prevalence of this
behaviour. However, the study of Fureix et al.
(2011) indicated very high frequencies of
yawning as compared to our findings here: 0.02
per min in one site where only geldings were
present and 0.05 per min in another site where
both geldings and mares were observed. Compared
to the frequencies observed for the domestic
mares and geldings in the present study, these
frequencies are about 60 times higher. Since in
the current study the comparison of Przewalski
and domestic horses living in stable groups in
semi-natural conditions did not indicate species
differences, it is likely that the increased
frequencies of this behaviour in the domestic
situation rather reflects an impact of the
housing conditions (see also findings on other
ambiguous behaviours in Hausberger et al. 2012;
Blois-Heulin et al. 2015) than of domestication
itself. Indeed, the conditions of life between
both studies were drastically different as
Fureix et al. (2011) observed the horses in
riding centres where they were housed in single
stalls, thus in restricted social, spatial and
feeding conditions known to elicit chronic
stress (McGreevy et al. 1995).
These horses were working in riding lessons,
which may also be a source of discomfort (e.g.
Lesimple et al. 2010). The finding that yawning
frequency co-occurred with stereotypic
behaviours, an admitted indicator of present
and/or past exposure to stressors reinforces the
idea that this behaviour was potentially
triggered by chronic stress (Fureix et al.
2011). Interestingly, the frequency of yawning
and stereotypic behaviours increased in the
pre-feeding situation, which likely corresponds
to anticipation and potential frustration, and,
therefore, an excitatory situation. Anticipatory
behaviours likely correspond to high-intensity
emotions (whether positive or negative), and the
results obtained in the present study may
indicate that yawning is associated with such
excitatory states whatever their valence (Mendl
et al. 2010; Peters et al. 2012). It could be
argued that the high level of excitability could
be attributed to mental characteristics of
horses (e.g. due to breed, Hausberger et al.
2004; Lloyd et al. 2008) from the cited study,
but almost a half of the domestic horses
observed were also warmbloods which did not show
higher yawning frequencies than the more native
Konik polski horses. Excitation may thus also
trigger this behaviour. Stress yawn
(yawning-stretching syndrome, Gessa et al. 1966)
was also observed out of social context but in
different anticipatory situations such as highly
aroused dogs before a walk
(Górecka-Bruzda, personal observations)
and before feeding in lions and mandrills
(Baenninger 1997).
Yawning and social stress
Prevalent occurrence of yawning in primates
was observed primarily in social conflict
situations (Hall 1962; Hinde and Rowell 1962;
Hadidian 1980; Troisi et al. 1990; Maestripieri
et al. 1992); therefore, social tension was
hypothesised to be the cause of stress yawns in
primates (Deputte 1978). Other arousal-provoking
situations like feeding competition were also
observed to be a context where subordinate
horses yawn more often (Hausberger, personal
observations). Here, we used agonistic behaviour
frequency as an indicator of social conflicts
and looked at the associated yawning frequency.
We found that male Przewalski horses (bachelors)
were more engaged in all measured social
interactions than younger individuals,
corroborating previous reports of increased
activity in adult male horses in natural
conditions (Duncan 1980; Berger 1986). According
to previous results in this species (Bourjade et
al. 2009), PH bachelors performed overall more
affiliative interactions than agonistic ones.
However, bachelors involved in social conflicts,
exhibiting a more aggressive behaviour than
others, also showed increased yawning
frequencies. This result would support the
social stress hypothesis proposed for primates'
yawning behaviour. It is puzzling yet that
domestic horses did not behave the same way; we
found no correlation between yawning frequency
and aggressive or defensive interaction
frequencies in domestic horses, which may simply
be due to the very low frequency of agonistic
behaviours observed in these conditions. Since
there were less adult males in domestic horse
groups than in Przewalski horse groups, it might
be possible that the small number of conflicts
observed have hampered the association between
yawning frequency and agonistic behaviour in
domestic horses being revealed.
There was, however, a positive correlation
between yawning frequency and affiliative
behaviour in domestic horses. It would be
interesting to examine more precisely which
positive interactions were most involved with
yawning and what followed in terms of social
issue. In this respect, it is worth noting that
several contexts may be responsible for the
change of the arousal state, including its
increase or decrease, depending on trigger's
valence (positive or negative). The yawning, as
primarily connected with relaxation and
increased drowsiness (Walusinski and Deputte
2004; Guggisberg et al. 2010) may be, as
suggested by Deputte (1994), associated with
neural mechanisms of lowering arousal
level.
Sex differences
Here, it was remarkable that despite the
species difference, but also the differences in
group composition, the same overall frequency of
yawning appeared to be higher for adult intact
males than for females or immature horses in all
cases. In rats (Berendsen and Nickolson 1981;
Serra et al. 1984; Urbá-Holmgren et al.
1990) and primates (Troisi et al. 1990;
Walusinski and Deputte 2004), yawning was found
to be more frequent in males than in females.
Sex difference in yawning frequency was observed
mostly in primate males of Cercocebus albigena
and Macaca fascicularis (Deputte 1978; Troisi et
al. 1990), Macaca mulatta (Chambers and Phoenix
1981; Deputte et al. 1994) Macaca fuscata
(Troisi et al. 1990) and Macaca nigra (Hadidian
1980), usually after reaching sexual maturity.
It can be admitted that the testosterone level
was involved in the regulation of yawning in the
species observed. In rats, Berendsen and
Nickolson (1981) found that castration reduced
the frequency of yawning in males, but not in
females.
In the latter study, testosterone treatments
increased the number of yawns in castrated males
and in both intact and ovariectomised females,
suggesting that yawning is under androgenic
control. Experiments on apomorhine's
(non-selective dopamine agonist) effects in rats
have shown a higher increase of yawning
frequency in male than in female (Serra et al.
1984). The administration of 17_-oestradiol
reduced the effect of apomorhine-induced yawning
in male rats (Serra et al. 1984) which suggests
that sex hormones, mutually interdependent, are
involved in the regulation of yawning in
animals. Interestingly, the study of Fureix et
al. (2011), performed on females and castrated
males, mentioned no sex difference in yawning
frequency. This could be due to the low
testosterone levels in geldings since stallions
were not available for the latter study. The
geldings observed in the semi- natural
conditions of the present study also tended to
show lower frequencies of yawning than stallions
(or even mares, although not significant),
suggesting a relation between hormonal levels
and this behaviour also in the two herbivore
species studied here. The observation of higher
frequencies of yawning in mature stallions as
compared to immature males further promotes this
idea.
Conclusions
According to existing works, yawning may be
triggered by different factors like higher level
of testosterone and, simultaneously or
independently, by social stress caused by
agonistic interactions, mainly between males. In
the present study on horses living in favourable
environment close to natural settings, it seems,
however, that testosterone may be a major
factor, while social excitation (rather than
social stress) appears as a secondary potential
factor causing a higher frequency of yawning in
males, because they yawned also when not
involved in social conflicts. Since a high
frequency of yawning was related to increased
frustration in horses kept in a restricted
stabling environment (Fureix et al. 2011), it
may also be supposed that the lower frequency of
yawning in horses observed in undisturbed social
groups may reflect increased welfare in equine
groups living in favourable conditions
satisfying their behavioural needs. Increased
occurrence of yawning in domestic situations
could thus attract the attention of caretakers
to make the alterations to improve the welfare
of their horses. While the precise relationship
between the causes, contexts and functionality
of yawning remains to be explained, this study
contributes to the general knowledge on this
behavioural pattern in herbivores.
