Yawning is contagious in humans and some
non-human primates. If there are social
functions to contagious behaviors, such as
yawning, they might occur in other highly social
vertebrates. To investigate this possibility, we
conducted an observational study of yawning and
an associated behavior, stretching, in
budgerigars (Melopsittacus undulatus), a social,
flock-living parrot. Flock-housed budgerigars
were videotaped for 1.5h at three time-blocks
during the day (early morning, afternoon and
early evening), and the times of all yawns and
stretches for each bird were recorded.
Both yawning and stretching were temporally
clumped within sessions, but were uniformly
distributed across the trials of a particular
time-block. This suggests that clumping was not
a result of circadian patterning and that both
behaviors could be contagious. There was
additional evidence of contagion in stretching,
which occurred in two forms - a posterior-dorsal
extension of either one foot or both feet. Birds
that could have observed a conspecific stretch,
and that then stretched themselves within 20s,
replicated the form of the earlier stretch
significantly more often than expected by
chance.
This study provides the first detailed
description of temporal patterns of yawning
under social conditions in a flock-living
species as well as the first support for
contagious yawning and stretching in a
non-primate species in a natural context.
Experimental evidence will be necessary to
confirm the extent of contagion in either
behavior
-Miller
ML, Gallup AC et al. Handling stress
initially inhibits, but then potentiates yawning
in budgerigars (Melopsittacus undulatus). Animal
Behaviour. 2010;80(4):615-619
Although yawning has been observed across
vertebrate classes (Baenninger, 1987; Craemer,
1924; Gallup et al., 2009; Luttenberger, 1975),
its function is still poorly understood
(Provine, 2005). It is characterized by an
involuntary opening of the mouth, with a deep
inspiration and shorter expiration, that is
stereotyped within and across individuals, and
is morphologically similar across species
(Provine, 1986a). Yawning is contextually
associated with transitions between activity and
inactivity, and for this reason it has been
suggested that yawning stimulates brain arousal
(Baenninger, 1997). Although physiological
evidence in support of this view is sparse
(Guggisberg et al., 2010), contextual evidence
is accumulating (Greco et al., 1993). For
instance, yawning is associated with fatigue in
humans (Zilli et al., 2008) and birds (Sauer and
Sauer, 1967), movement in humans (Baenninger et
al., 1996) and primates (Vick and Paukner,
2010), stress in rodents (Moyaho and Valencia,
2002) and birds (Miller et al., 2010), and
boredom in humans (Provine, 1986b). Recent
comparative research also supports a role of
yawning in brain thermoregulation (e.g., Gallup
and Gallup, 2007; Gallup, 2008; Gallup and Hack,
2011; Gallup and Eldakar, 2011; Shoup-Knox et
al., 2010), and it has been suggested that the
cooling component of yawning may facilitate
arousal by reinstating optimal brain
temperature. Thus, building evidence from
numerous laboratories suggests that yawning is
multifunctional (Vick and Paukner, 2010; Gallup,
2011), which may explain its ubiquity across
vertebrates (Baenninger, 1987).
In contrast with spontaneous yawning,
contagious yawning has been convincingly
documented only in humans and a few non-human
primates. Contagion is defined as the matching
of reflexive or involuntary behaviors (Zentall,
2003), of which yawning provides a classic
example. For instance, just observing or even
reading about yawns stimulates yawning in humans
(Baenninger and Greco, 1991), and attempts to
shield a yawn do not stop its contagion
(Provine, 2005). Under laboratory conditions,
watching videotaped yawns produces contagious
yawning for roughly 50% of human participants
(Gallup and Gallup, 2007; Platek et al., 2003).
Similar methods have been used to document
contagious yawning in chimpanzees (Pan
troglodytes) (Anderson et al., 2004), and
recently this result has been replicated using
threedimensional computer animations as a
stimulus (Campbell et al., 2009). Video-induced
yawning has also been reported in stumptail
macaques (Macaca arctoides) (Paukner and
Anderson, 2006), but since the same stimulus
also induced significantly more selfdirected
scratching responses, the degree to which the
increased yawning represents social contagion,
rather than social tension or stress, remains
unclear.
A more recent study tested for a contagious
yawning in red-footed tortoises (Geochelone
carbonaria) by either displaying video clips of
a yawning conspecific, or using a live model
trained to yawn in the presence of other
tortoises (Wilkinson et al., 2011). In
eithercase, however, there was no evidence
forcontagious yawning in this species. Further
research using a live demonstrator as a stimulus
has involved the testing of domesticated dogs
(Canis familiaris). The first report of this
topic provided evidence to suggest that dogs
yawn in response to human yawns (Joly-Mascheroni
et al., 2008). However, more recent research
attempting to replicate this finding, using both
live demonstrators as well as video clips, has
failed to demonstrate this cross-species
contagion effect (Harr et al., 2009; O'Hara and
Reeve, 2010). In addition, one report using
video clips of dog yawns also failed to provide
evidence for conspecific contagious yawning
(Harr et al., 2009), casting doubt on whether
dogs yawn contagiously at all. Using a different
approach, a recent observational study reported
evidence of contagious yawning in gelada baboons
(Theropithecus gelada) (Palagi et al., 2009).
The authors recorded all instances of yawning
from a colony of captive baboons, revealing that
the frequency of this behavior increased among
individuals when in the presence of both visual
and acoustic yawning signals from conspecifics.
Similar to other research on primates (Vick and
Paukner, 2010), several distinct types of yawns
were identified. Socially close baboons,
especially females, were more likely to yawn
contagiously, and these females matched the
observed yawn-type when they yawned immediately
after.
If contagious behaviors serve important
functions, e.g., group coordination, in social
mammals, it seems reasonable that yawning may be
contagious in social, non-mammalian species as
well. Furthermore, different behaviors could
also be contagious and serve the same function,
depending on the activity or social changes
signaled. Here we present an observational
study, in which we documented patterns of
yawning and an associated behavior, stretching,
in a flock of budgerigars (Melopsittacus
undulatus) housed in an indoor aviary.
Budgerigars are highly social, small parrots
indigenous to Australia. They move in highly
coordinated flocks throughout the year, even
breeding as pairs within a larger flock
(Wyndham, 1980), and signals of intention to
move could certainly play a role in coordinating
group activity. Stretching is a stereotyped
behavior that is associated with yawning in
humans and rodents (Baenninger, 1997), but there
is little evidence that stretching is contagious
in humans or other animals (for evidence of
synchronized group displays, see Stevens, 1991).
Nonetheless, stretching and yawning may predict
changes in activity and/or an individual's level
of stress, and therefore, the spread of either,
or perhaps both behaviors, may coordinate group
activity. Similar to yawning, stretching is
believed to be a reflexive, automatic action in
these birds, so unlike the copying of voluntary,
learnt behaviors, known as imitation or response
facilitation (e.g., Hoppitt et al., 2007), in
this study the temporal coupling of either
behavior refers to contagion.
Our previous research has recently explored
the contagious nature of these behaviors in
budgerigars through video stimuli, finding mixed
support for a social influence in yawning
(unpublished data). In particular, the latency
to yawn was significantly reduced following
clips of conspecific yawns compared with control
clips, but the frequency of yawning and
stretching did not increase following clips of
the respective behavior (unpublished data).
There were, however, limitations in the quality
of the stimulus (recorded from freely behaving
birds) and the degree to which the experimental
birds were attending to the video screen.
Therefore, in this study we tried to lay a
stronger foundation for future experimental work
by taking a naturalistic approach similar to the
study performed on gelada baboons (Palagi et
al., 2009). To explore how individual birds
responded to the actions of nearby group
members, we video recorded an undisturbed,
established flock of captive budgerigars, and
measured the time and occurrence of each yawn
and stretch. For yawning and stretching
separately, we analyzed the distribution of
successive behaviors.
We also looked for any diel patterns, and
associations between stretching and yawning at
three different times of the day (early morning,
afternoon and early evening). It was
hypothesized that, if contagious, each behavior
would be non-randomly clumped into closely
spaced bouts within recording sessions, as birds
were stimulated by their neighbors' behavior,
and separated by longer periods without these
behaviors. Even if clumped within a particular
testing session, we further predicted that these
behaviors would be evenly spaced across multiple
sessions, when comparing sessions that were
recorded at the same time of day. This would
suggest that it is not a specific time of day
that causes the clumping pattern. Lastly, a
strong circadian or other temporal pattern,
previously established for humans and rats
(Baenninger, 1997; Anias et al., 1984; Zilli et
al., 2007), would potentially illuminate the
context and function of these behaviors, whether
contagious or not. In summary, although these
behaviors may have a general circadian pattern
over distinct periods of the day (i.e., they may
occur more frequently in the morning or
evening), we predict that within a particular
session, behaviors will be clumped due to social
factors.
Occurrences of both yawning and stretching
were temporally clumped in an unmanipulated,
captive flock of budgerigars, as would be
expected if these behaviors are contagious.
Despite the low frequency of yawning (1.28-2.96
yawns per bird per hour, depending on time of
day), a bird was more likely to yawn within 40s
or less of another bird's yawn. There were also
a substantial number of yawns separated by at
least 300s from the previous yawn, but few
spaced at intermediate intervals. Taken
together, the inter-yawn spacing distribution
(Fig. 2a) suggests that yawns were socially
influenced (Le., contagious). In other words,
long periods of no yawns were broken by a
budgerigar's yawn that was then followed by a
cascade of yawns among the others. A similar,
although less strongly bimodal temporal
distribution of stretching was observed. In
part, fewer stretches were separated by very
long intervals because there were a
substantially greater number of stretches than
yawns per session (566 yawns versus 1752
stretches) and stretching continued for longer
bouts among flock members. Stronger evidence to
support the social influence of this behavior
comes from stretch-type matching, illustrating
that birds were more likely to replicate the
specific stretch-type of a previous bird than
would be expected by chance. This result is
similar to the observational research on gelada
baboons (Palagi et al., 2009), showing yawn-type
matching. Although no functional distinctions
between mono- and bi-lateral stretches were
identified, this temporal pairing of identical
behaviors suggests that stretches of
conspecifics were closely observed and that the
form influenced the subsequently stretching
bird. It has been suggested that different
yawn-types may produce distinct physiological
outcomes among chimpanzees (Vick and Paukner,
2010), and thus the matching of different
behavioral-types may help coordinate group
activity.
An alternative interpretation of the
temporal patterns we observed is that flocking
birds tend to simultaneously reach the same
physiological state (i.e., well-rested, hungry,
alerted by outside event, etc.) (Sauer and
Sauer, 1967), and this tendency produced more,
or less, yawning or stretching among group
members. Differences in frequencies across the
day would be a reflection of these shared
changes, since there are clear diel patterns of
yawning in humans and rodents (Anias et al.,
1984; Baenninger et al., 1996). If yawning
patterns are related to daily rhythms of body
temperature, metabolism, and resultant arousal,
it is possible that flock members both share a
physiological rhythm and respond with some low
degree of contagion to another's behavior, thus
strengthening the diel pattern and producing
higher degrees of clumping. Although plausible,
this interpretation seems insufficient. Analyses
show that yawning was significantly clustered
within sessions, as would be expected by
contagion, but when looking across sessions
recorded at the same time of day, we notice that
both yawns and stretches occurred evenly
throughout the videos, and were not repeatedly
clustered at a particular time of day. This
suggests that a related circadian physiological
rhythm experienced by the birds does not explain
our results.
The combined temporal patterning and
significant "matching" of adjacent stretches
suggest that this behavior is contagious and
thus a potential social signal in this species.
Although the function of stretching is largely
unstudied, it is another stereotyped, unlearned
behavior that is ubiquitous among tetrapods.
Both yawning and stretching are homeostatic
behaviors believed to serve a purpose in the
maintenance of bodily functions through enhanced
circulation (Sauer and Sauer, 1967). Stretching
in humans is confined to specific, yet intense
state-change, occurring most frequently after
waking, but not prior to sleep (Provine et al.,
1987). Research on both humans and animals show
that yawning also typically occurs during
broader state-changes (Provine et al., 1987),
and is followed by modified activity or
increased locomotion (Baenninger et al., 1996;
Giganti et al., 2002; Vick and Paukner, 2010).
Taken together with the current results, we
suggest that these behaviors may coordinate
collective flock behavior, in addition to
serving as important preparatory responses to
flight. As evidence for this, flight and other
movement among perches is often preceded by
stretching or yawning in budgerigars
(unpublished data). Responding to another's
intention-movements has clear adaptive value for
any group-moving species. Based on these
observations, the study of yawning, stretching
and transitions in activity may provide a novel
approach to studying collective behavior. When
perched, budgerigars sit in close proximity to
one another and remain oriented towards adjacent
group members, providing a setting where
behaviors can spread across a line of birds,
coordinating flock movement. Recently we have
shown that auditory disturbances enhance both
stretching and yawning contagion among
budgerigars in small groups (Miller et al., in
press), suggesting that the close coupling of
these behaviors may be involved in collective
response to environmental stimuli. Future
research could investigate the role of yawning
and stretching contagion in group vigilance, and
more specifically how spatial position within a
group reflects another bird's information
processing, and how birds use local behaviors of
nearby conspecifics to infer collective-state
(Couzin, 2009).
4. Conclusions
Signals may frequently originate from
physiologically relevant behaviors adapted for
social purposes. Spontaneous yawning is
associated with stress, arousal and
thermoregulation in a variety of species,
including budgerigars. While the physiological
function of stretching is less clear,
vertebrates frequently stretch before beginning
to move. Stretching also co-occurs with yawning
in a variety of species and may therefore be
associated with arousal. The observational
results presented here suggest that yawning and
stretching are at least mildly contagious in
budgerigars under semi-natural flock-living
conditions. In line with each behavior's
presumed physiological function, contagious
yawning and stretching may ultimately coordinate
mental state and a group's collective movements,
but future research needs to test these
predictions. While experimental studies are
needed to confirm and clarify the degree and
precision of contagion, we propose that
experiments be designed using live birds as the
target stimulus. Nonetheless, the current
results provide a strong basis for understanding
the functional context of, and inferring an
adaptive role for, contagion in coordinated
flock-living species.
