Yawn contagion occurs when individuals yawn
in response to the yawn of others (triggers).
This is the first account of yawn contagion in
wild geladas (Theropithecus gelada), a monkey
species that shows yawn contagion in captivity
and is organized in core units
(one-male/bachelor groups) forming multilevel
associations. In a population of geladas from
the Kundi plateau (Ethiopia) the authors found
that the yawning response was highest when
geladas could perceive a triggering yawn, which
confirms that yawn contagion is present in the
wild. Yawn duration, mouth-opening degree and
presence/absence of vocalisation (possibly
modulating yawn detectability) did not affect
the likelihood of contagion. Males and females,
known to be both implicated in movement
initiation within groups, were similarly
powerful as yawn triggers. Instead, group
membership and responder sex had a significant
role in shaping the phenomenon. Yawn contagion
was highest between individuals belonging to
different core units and males were most likely
to respond to others' yawns. Because males have
a non-negligible role in inter-group
coordination, these results suggest that yawn
contagion may have a communicative function that
goes beyond the basic unit level.
La contagion du bâillement se produit
lorsque des individus bâillent en
réponse au bâillement des autres
(déclencheurs). Il s'agit du premier
constat de la contagion du bâillement chez
les Geladas sauvages (Theropithecus gelada), une
espèce de singe qui présente une
contagion du bâillement en
captivité et est organisée en
groupes hiérarchisés (groupes d'un
seul mâle/célibataire) formant des
associations à plusieurs niveaux. Dans
une population de géladas du plateau de
Kundi (Éthiopie), les auteurs ont
constaté que la réponse du
bâillement était la plus
élevée lorsque les géladas
pouvaient percevoir un bâillement
déclencheur, ce qui confirme que la
contagion du bâillement est
présente dans la nature. La durée
du bâillement, le degré d'ouverture
de la bouche et la présence/absence de
vocalisation (éventuellement en modulant
la détectabilité du
bâillement) n'ont pas affecté la
probabilité de contagion. Les mâles
et les femelles, connus pour être tous
deux impliqués dans l'initiation du
mouvement au sein des groupes, avaient la
même faculté pour déclencher
des bâillements. Au lieu de cela,
l'appartenance à un groupe et le sexe des
répondants jouent un rôle important
dans la réplication. La contagion du
bâillement était la plus
fréquente entre les individus appartenant
à différents groupes de base et
les mâles étaient les plus
susceptibles de réagir aux
bâillements des autres. Parce que les
mâles ont un rôle non
négligeable dans la coordination
intergroupe, ces résultats
suggèrent que la contagion du
bâillement peut avoir une fonction de
communication qui dépasse le niveau du
groupe de base.
Introduction
Contagious yawning differs from spontaneous
yawning in that one individual (responder) yawns
in response to a yawn emitted by another
individual (trigger)1. In this respect, the yawn
of an individual acts as a triggering stimulus
(sensu Tinbergen2). While spontaneous yawning is
probably an ancient phenomenon widespread in
vertebrates3,4, intraspecific contagious yawning
is an evolutionary more recent phenomenon that
has been found mostly in highly social species.
Such species include mammals (domestic pig, Sus
scrofa5; sheep, Ovis aries6, lions, Panthera
leo7, elephant seals, Mirounga leonina8; Tonkean
macaques, Macaca tonkeana9; bonobos, Pan
paniscus10; chimpanzees, Pan troglodytes11,12;
humans, Homo sapiens13; but see: Pongo
pygmaeus14) and one bird species (budgerigars,
Melopsittacus undulates15). Moreover, there are
some social species in which yawn contagion is
virtually absent (e.g. lemurs, Lemur catta and
Varecia variegata16; lowland gorillas, Gorilla
gorilla gorilla17). In different species, yawn
contagion can be triggered both by vocalized and
non-vocalized yawns18,19, regardless of the uni-
or multi-modal nature of the stimulus13,18,20.
Yawn contagion can be influenced by
environmental/social contexts and individual
factors, and it can follow a diurnal rhythm with
a peak in the morning21,22. Moreover, the
phenomenon has been found to be highest between
familiar or in-group subjects (wolves23;
bonobos24; chimpanzees12; humans18,25). Finally,
depending on the social role played by each sex
in social groups, the yawning response can
increase when triggers are males (dominant in
chimpanzees11) or females (dominant in
bonobos26). In humans, women may respond more to
others' yawns27,28, although this does not apply
to all cohorts18,29. Overall, yawn contagion
seems to promote inter-individual physiological
resonance30,31, synchronisation within social
groups7 and coordinated daily
activities7,32.
Geladas (Theropithecus gelada) are a diurnal
monkey species, endemic to Ethiopia, that lives
in core units (One-Male-Unit, OMU;
All-Male-Unit, AMU) forming discrete multilevel
associations (teams and bands33). The core units
are characterized by high levels of grooming,
low levels of aggression, and no coercive
control by males over females, with females
mediating group coordination with the alpha
male34. While grooming is observed only within
groups and never between groups33, high levels
of tolerance characterize the groups forming the
band, which show low aggressive levels when they
occupy the same foraging areas33.
Geladas have been the first monkey species
in which yawn contagion was demonstrated35. In
captive geladas yawn contagion was elicited by
all types of yawns (covered teeth, uncovered
teeth and uncovered gums) emitted with or
without vocalizations. Moreover, it was highest
between subjects sharing strong bonds (measured
via grooming exchange rates)35. So far, no study
has described the phenomenon of yawn contagion
in geladas, and more generally in primates,
under wild conditions.
Here, we investigated yawn contagion in a
population of wild geladas to verify whether the
phenomenon was present in the natural setting.
Moreover, we investigated the possible influence
of the factors that are only assessable under
wild conditions to understand if the phenomenon
can act on a multilayer scale reflecting the
multilevel social organization of the species.
Based on the previous framework, we formulated
the following predictions.
Prediction 1&emdash;Presence of yawn
contagion. As previously reported in
captivity35, we expected to find that
individuals would be more likely to yawn when
they perceived the yawn of a conspecific
compared to when they did not (presence of yawn
contagion).
Prediction 2&emdash;Perceptual factors. In
line with the findings obtained in captivity35,
we expected that also in the wild yawn contagion
would be induced at comparable levels by
different types of triggering yawns and
regardless of the presence of vocalization.
Prediction 3&emdash;Contextual and
individual factors. As it occurs in humans
(characterized by a diurnal lifestyle) that show
a peak of yawn contagion in the morning22, we
expected that in geladas yawn contagion could be
highest in the first part of the day (Prediction
3a). Because in primates, including captive
geladas, yawn contagion has been found to be
most frequent between familiar
subjects12,18,23,35, in geladas we expected to
find a higher rate of contagion within rather
than between groups (Prediction 3b). Finally,
the different role of each sex in managing
social groups can lead to sex differences in
yawn contagion25&endash;27. Because in geladas
both males and females play an important role in
managing the social groups34, we expected to
find no difference in yawn contagion rates
between sexes (Prediction 3c).
Discussion
The present study provides the first
evidence of yawn contagion in a wild primate,
the gelada (in line with Prediction 1). In the
considered time window, the study subjects were
more likely to yawn when they could perceive a
yawn from conspecifics compared to when they
could not (Fig. 2b; Table _Table1).1). Most of
the yawning responses occurred in the second
minute, thus confirming captive data35. The
physical distance between trigger and responder,
yawn duration, and trigger yawn type
(accompanied or not by vocalization) had no
effect on the occurrence of the phenomenon
(Table _(Table1).1). Hence, the level of yawn
detectability (possibly enhanced in case of
long, vocalized, and wide-open yawns) appeared
not to be strictly linked to the yawning
response probability (in agreement with
Prediction 2). This is in line with the previous
report showing that all yawn types can be
contagious in geladas35. Our data are also in
agreement with results on humans that are
infected by others' yawns by having access to
both total or partial view of the trigger's
face18,20,41. Palagi et al.35 found a matching
of the yawn type especially in the contagion
between females. We could not check for this
aspect due to data limitation, but it is
certainly an issue that would be worth exploring
in future investigations.
As in captivity35, also in the wild both
vocalised and non-vocalised yawns could elicit
contagion at a comparable level (Table
_(Table1).1). This result is also consistent
with findings on humans, where contagion is
triggered by both the visual and acoustic
components of the stimulus19,25,47. The acoustic
component of the yawn might function when
geladas are located far from the trigger and
cannot visually perceive yawning. Further
investigation including larger observational
ranges and datasets per individual is necessary
to clarify the role of multimodality on this
phenomenon.
Contrary to our prediction (3b), in the
study population yawn contagion was highest
between individuals belonging to different
groups (Fig. 2c, Table _Table1).1). Although the
adult members of different groups do not engage
in affiliative behaviours, they show high
tolerance levels, spatial cohesion, and
collective moving33. Considering the complexity
of the gelada multilevel social system48, yawn
contagion could be a phenomenon acting on a
multilayer scale. Our limited dataset is silent
on whether yawn contagion is affected by the
affiliative relationships (e.g., grooming)
shared by group members, thus not allowing the
verification of this possibility, already tested
in captivity35.
Contrary to the expectation (Prediction 3a),
in the Kundi population the time of the day had
no effect on contagion (Table _(Table1).1).
Because geladas arrived on the plateau
relatively late in the morning (9&endash;10 am),
it was not possible to sample their behaviour
soon after they started their activities.
Expanding data collection to different contexts,
including sleeping sites, would be crucial to
determine how yawn contagion is distributed
within groups according to the social
affiliation shared by subjects over the 24-h
cycle. However, reaching out our geladas to
sample their behaviour at such sites on the
cliffs may be challenging.
Finally, we found that yawn contagion: (i)
was triggered to a similar extent by males and
females, and (ii) males most frequently
responded to others' yawns (Fig. 2d, Table
_Table1,1, Prediction 3c not supported). In the
wild controlling for all the potential arousal
stimuli may not be completely possible. However,
our conservative exclusion of possible
perturbing events (aggression, human
disturbance, and copulations) and the fact that
males did not preferentially respond with
uncovered teeth yawns (often associated with
social tension and agonistic contexts in
males38) makes it unlikely that the response by
males could be stress driven.
The trend of yawn contagion between sexes
and groups suggests that this phenomenon may
have a communicative function, related to
activity coordination. Males and females, which
were similarly powerful as yawn triggers (Table
_(Table1), 1), can be both implicated in
movement initiation within groups34. However,
males frequently initiate the group movement,
followed by the dominant female and the rest of
the groups34. Hence, the fact that males
automatically responded more frequently to
others' yawns and that yawn contagion was
highest between groups, suggests that in the
wild yawn contagion might also favour
inter-group synchronisation during daily
activities. The Social Communicative Hypothesis
predicts that in different
contexts&emdash;through
yawning&emdash;individuals can involuntarily
convey information about their internal state
and possibly about their subsequent
behaviour32,38,39,49. The communicative function
of yawning can have an adaptive role at
inter-group level (i.e., defence from predators,
vigilance50) because it promotes the
synchronization of groups' activity through
individuals that do not necessarily share strong
bonds31,51&endash;54. Consistently, in wild
lions Casetta et al.7 found that yawn contagion
increased coordinated movements. Hence, also in
wild geladas yawn contagion may be a phenomenon
that promotes the coordination within larger
ecological units, such as teams and, possibly,
bands33.
The low effect size of the significant
variables affecting the yawning response (Table
_(Table1)1) cautions against the generalisation
of the results and points toward the need of
further data collection on other sites and
populations. Despite this limitation, our
findings open new scenarios on the role of yawn
contagion in multilevel social systems. We
suggest that this phenomenon operates at a
multilayer scale reflecting the complex nested
organization of geladas. At the basic level
(core units), yawn contagion might improve
synchronisation between group members sharing
good relationship quality, as suggested for
captive geladas35. In the wild, yawn contagion
may also function at an upper level (team) by
promoting the coordination of daily movements
and activities between groups. Further studies
may verify whether yawn contagion indeed elicits
subsequent movements jointly started by
individuals involved in the contagion not only
between but also within groups (e.g.9).
On a broader perspective, our study
underlines the importance of combining wild and
captive approaches to thoroughly describe
behavioural phenomena that are sensitive to
different social and environmental conditions.
