Resume : the
present research found no support for the
popular hypothesis that yawning is a response to
elevated CO2 or depressed O2 levels in the
blood.
Yawning is a common and probably
species-typical human behavior (Provine, 1986)
that is performed throughout the life span
(Gesell, l928). It is characterized by a gaping
of the mouth accompanied by a long inspiration
followed by a shorter expiration. Yawning is a
stereotyped behavior that is evoked by unknown
physiological stimuli or released by observed
yawns or y awn-related stimuli (Provine, 1986).
Yawns are easily triggered; even reading about
or thinking about yawns stimulates them. Yawning
is most frequent during the 'hour before
'bedtime and afterwaking (Provine,
Hamernik,& Curchack). People also yawn when
watching.uneventful stimuli or participating in
repetitious, uninteresting tasks (Provine &
Hamernik, 1986). The temporal proximity of
yawning to bed and waking times and other
occasions of low arousal is probably the basis
for yawning as a paralinguistic signal for
drowsiness.
Yawning is of medical importance because it
is symptomatic of pathology such as brain
lesions and tumors, hemorrhage, motion sickness,
chorea, and encephalitis (Barbizet, 1958,
Graybiel & Knepton, 1976; Heusner, 1946,
Jurko & Andy, 1975). Yawning is also
therapeutic in preventing postoperative
respiratory complications (Cahill, 1978) and in
adjusting the air pressure in the middle ear
(Laskiewicz, 1953).
There is much popular opinion but little
empirical evidence about the physiological cause
and function of yawning. The unsupported
hypothesis that yawning is a respiratory
maneuver that somehow increases the oxygenation
or decreases the CO, level of blood is often
repeated in the popular and scientific
literature (i.e., Abarbanel, 1964; Brody, 1984;
Montagu, 1962; Redican, 1982). The present
studies evaluated the effect of elevated levels
of CO2, 02, and exercise on human yawning. These
interventions alter blood levels of these gases
and breathing rates (Comme, 1965). If yawning
either increases or decreases in response to O2,
CO2 or exercise, some form of the CO2/O2
hypothesis is supported and a respiratory role
for yawning is suggested.
EFFECTS OF CO, AND O2 ON YAWNING
The prescent study tested the hypothesis
that yawning is facilitated by higher than
normal blood levels of CO, or lower than normal
blood levels of O2 by observing the effect on
yawning of breathing compressed air, 100% O2,
and gas mixtures with higher than normal levels
of CO2(3 or 5% CO2). (The normal composition of
air is 20,95% O2, 79.02% N2 and inert gases, and
0.03% CO2). If yawning is a response to
heightened blood CO2 levels, breathing gases
high in CO2 should increase yawning rate and/or
duration. If low blood O2 levels produce
yawning, breathing 100% O2 should inhibit
yawning. The effects of breathing gases with
subnormal levels of O2 were not investigated
because of concern about stressing the
subjects.
Method : Subjects. The 18 subjects (3
males, 15 females) were 17 to 21 year old
college freshmen who volunteered to earn
experiment participation credit for an
introductory psychology class. All subjects who
began the study completed all four experimental
conditions.
Procedure. All subjects began with a
30-min baseline condition in which they breathed
normal room air. Three other 30-min experimental
sessions were spaced at least 1 day apart. The
order of presentation of all but the initial
baseline condition was counterbalanced. The
baseline session acclimated subjects to the
experimental procedures without the additional
novelty of breathing through a mask as in later
conditions. During each of the next three 30-min
sessions subjects breathed one of four gas
mixtures: (a) 1 00% 02; (6) 3% C02, 21% 02,
balance N2; (C) 5% CO2, 21% O2, balance N2; and
(d) normal compressed air (20-95% O2 0,03%c CO2,
79.02% N, and inert gases). After observing no
apparent effect of the 3% CO, mixture on the
yawning of the first six subjects, the next 12
subjects were presented the 5% CO2 mixture. The
compressed gases were fed into a 60 liter
Douglas bag that acted as a reservoir from which
subjects breathed using a small, hand held
(Vital Signs) mask that covered the nose and
mouth. The small mask was hand held by subjects
to avoid their possible concern about having the
mask strapped on. The reservoir (Douglas) bag
technique for gas presentation was chosen
instead of a demand regulator because the bag
was more likely to produce a normal breathing
pattern. Breathing was monitored with a chest
bellows transducer and recorded with a Grass
oscillograph. All gases were obtained from the
Puritan-Bennett Corp. The gas mixtures with
enhanced CO2 were guaranteed to be accurate
within +-0.5% but were usually accurate within
+- 0,1%. All conditions except baseline were
counterbalanced in order and involved breathing
gases through a hand held mask. The baseline
condition during which subjects breathed room
air was always first and served as a training
session. For these reasons, the counterbalanced
compressed air (air) condition that involved
mask breathing is the best control condition
with which to contrast the results for O2 and
CO2.
Each subject was seated at a table in a
well-lighted, fan-ventilated, sound-attenuating
chamber with floor dimensions of 2.5 m X 3.1 m.
After being seated, the participant was joined
by the experimenter who took a seat across the
table and explained the objectives and
procedures of the study. All subjects were read
the following information: "In this first part
of a four-part study, we will be examining the
naturally occurring frequency and duration of
yawns. Your breathing rate will be monitored by
a belt-like device that will be wrapped around
your chest. You will be recording your own
yawning behavior. Yawning is easy. It.is
important hat you relax. and do net force yawns.
It.will be unnecessary; most people can yawn by
simply thinking of it. When you start to yawn,
push down the button of the event recorder and
keep it depressed until the yawn is complete.
Consider the yawn to end when you finish
exhaling. You will record the duration of each
yawn until the end of the half-hour session. It
is important to record your yawns as accurately
as possible because other than the responses
that you record, you will not be observed in any
way. You will have complete privacy in the sound
proofed booth. I will demonstrate how to respond
to a yawn." (The experimenter demonstrated a
typical yawn and the appropiate reonse to it).
"Think about yawning and you can produce a real
yawn. Now you try a practice yawn. Do you have
any questions? After I leave the room, relax and
think about yawning throughout the session and
record your yawns, when they occur. You will be
notified when the time is up".
After the first baseline session, each
session was preceded with instructions about the
use of the face mask used to supply the gas
mixtures. Subjects were told, "In this part of
the study, we will bc examining the effects of
breathing one of threc gas mixtures on the
naturally occurring frequency and duration of
yawns. You will probably be unable to
distinguish any difference between the gases.
You will breathe the gas through a small mask
that you will hold over vour mouth and nose. Do
not alter your normal pattern of breathing while
using the mask. Breathe naturally. The gases are
safe to breathe and should produce no
distressing sensations, but if you feel
uncomfortable in any way, you may remove the
mask for a few seconds. You may also quit the
session at any time and.still receive credit for
participating. The rest of the procedures are
the same as in the first session. As before,
your breathing rate will be measured by a
belt-like device that will be wrapped around
your chest. Your task is to think about yawning
throughout the half-bour session and to record
your yawns when they occur. When you start to
yawn, push down the button of the event recorder
and keep it depressed until the yawn is
complete. Consider the yawn to end when you
finish exhaling. If you think that the mask is
inhibiting your yawning movements, you may
remove the mask until the yawn is complete. Do
von have any questions? After I leave the room,
relax and think about yawning throughout the
session and record your yawns when they occur.
You will be notified when the time is up.
Subjects were fully informed of the
experimental procedures and objectives, although
they were not told which gas they were breathing
during a given session. Informed subjects were
considered to be less likely than uninformed
subjects to speculate about their expected
behavior and, thus, become more sensitive to
possible experimental demand characteristics.
The high-amplitude, all-or-none, behavior of
yawning makes the yawner more reliable and
accurate at discriminating his yawns than an
outsider who must respond to sometimes ambiguous
cues. The between session reliability of
self-reponed yawns is documented in Provine
(1986). Also, yawns were self-reported because
attempts too bserve subjects directly may reduce
yawning frequency because of the social
sanctions against public yawning; several
subject participating in a pilot study
volunteered the comment that they did not like
to y awn while being watched. Subjects pressed a
button in response to a yawn. Their response was
recorded with an Esterline-Angus chart recorder
(A 620X) driven at 2.5 mm/s. Yawn duration was
the time the subject kept the response button
depressed. Yawn duration was measured manually
from the event recorder records with a
millimeter rule. The above procedures were
modeled after those of a previous study.
(Provine, 1986).
Results
The effects of breathing room air
(baseline), compressed air (air), 100% 02, and
C02 (3 or 5% C02, 21% 02, balance N2)on
breathingrate, yawing rate, and yawning duration
were examined. Breathing rate. The study
is a factorial design with groups (3% CO, and
5%. C02), conditions (baseline, air, 02, and
C02), and time periods (5, 10, 15, 20, 25, and
30. min).- The -.time periods were produced by
averaging data from 1 to. 5 min, 6 to 10 min,
and so on. The main effect of conditions was
first tested in terms of three orthogonal
contrasts baseline versus the average of all
other conditions, air versus the average of 02
and the C02 mixtures, and 02 versus the CO,
mixtures. The main effect of time period
(period) was analyzed in terms of polynomial
trend across the six 5-min periods. Both 100% 02
and the two CO2 mixtures increased breathing
rate. The analysis of breathing rate reveals a
highly sighificant (p < .0005) main effect of
conditions using both multivariate (F(3,14)
=14.94) and univariate (F3,48) = 14.46) tests.
The contrast of baseline (room air, no mask)
with other conditions is highly significant (F(l
,16) 13.67, p.002), as is the contrast of the
air (compressed air) with the O2 and CO2
conditions (F(l, 16) = 31.77, p <.0005). The
comparison of O2 with CO2 is marginally
significant (F(l, 16) = 4.21, p =.057). The air
versus O2 contrast is also significant as is the
air versus CO2 contrast. There is no interaction
with conditions and groups, either overall or
for any of the individual interaction contrasts
(all F ratios < 1). The main effect of period
and the interaction of period with groups is
nonsignificant. The interaction of Condition x
Period and the three way interaction of Group X
Condition X Period are also nonsignificant.
It should be noted that the above analyses
yield a very conservative estimate of CO2
effects. Although CO2 produced small but
significant increases in breathing rate,
previous work from other laboratories has
established that respiratory tidal volume almost
triples in response to breathing 5% CO2 (Dripp's
&-Comroe, 1947). The heavy breathing
produced by 5% C02 in the present study was not
experienced in response to the other gases.
Yawn frequency. Neither 100% 02 nor
either of the C02 mixtures influenced the rate
of yawning. A multivariate analysis of yawn
fiequency revealed no main effects of group
condition or period. There was, however, an
effect of condition by the univariate test (F(3,
48)= 3.17, p = .033) that was interpretable as a
difference between the haseline and the average
of the other three conditions (F(l, 16) =4.17, p
= .045). The detection of such a difference is
not surprising because, as noted above, the
baseline training condition was always first and
did not involve breathing through a mask. The
compressed air (air) condition is, thus, a
better control. The interaction of condition and
period was also significant (multivariate
F(15,2)=42,98, p = ,023) and suggests that the
fonctions for the four conditions are diverging
over time. However, the 30-min session length is
probably adequate to reveal experimental effects
if they were present. As noted above, breathing
5% C02 produces dramatic increases in breathing
rate and tidal volume: This respcinse is obvious
within a few minutes.
Yawn duration neither 100% O2 nor
either of the CO2 mixtures influenced the
duration of yawns. In the analysis of yawn
duration, the main effect of period is highly
significant by both multivariate (F(5, 12)=7.76,
p = .002) and univariate (F(5, 80) 3.34, p =
.009) tests. A trend analysis revealed a linear
increase in yawn duration over the six 5-min
time periods. No other main effects or
interactions concerning yawn duration were
significant.
In summary, both 100% O2 and the combined
CO2 groups (3 and 5%) produced significantly
greater breathing rates than the baseline or the
compressed air control and the higher breathing
rate of the CO2 compared to the O2 group
approached significance. In contrast, breathing
CO2 and O2 had no significant impact on yawning
rate or duration when compared to either
baseline or compressed air controls. The
hypotheses that yawning is a response to a high
blood level of CO2 or a low level of O2 were not
confirmed.
EFFECT OF EXERCISE ON YAWNING
If yawning serves a respiratory function,
yawning and breathing may share common
regulatory mechanisms or be influenced by
similar stimuli. The present study examined the
effect of exercise on yawning rate and duration.
Exercise increases breathing rate. If yawning
either increases or decreases during exercise, a
relationship between yawning and respiration is
indicated. If no relation is found between the
response to exercise of yawning and breathing
separate and independent mechanisms control the
rates of the two behaviors.
Method Subjects. The 19 subjects (7
males, 12 females) were 17 to 21 year old
college freshmen who volunteered to earn
experiment participation credit in an
introductory psychology class.
Procedure. Subjects were seated alone
in the small, sound-attenuated experimental
chamber described previously, and were
instructed to think about yawning. Subjects
reported their yawns by pressing the button on a
small box held in their right hand that
activated an event recorder. This procedure was
the same as that used in the first study.
Breathing rate was detected with a small
thermister probe that was clipped to the
subject's left nostril and recorded using a
Grass oscillograph. After 10 min of baseline
data had been collected from seated subjects in
a resting condition, the experimenter entered
the chamber and instructed the subject te
exercise by stepping up and down a 25.4 cm step
at a rate of approximately on step cycle every 2
s. Subjects were also instructed to continue
thinking about yawning and to record their
responses. After 10 min.in the exercise
condition, the experimenter again entered the
chamber and instructed the subjects to be
seated, but to continue thinking about and
recording yawns. The experiment was terminated
after this second 10 mn baseline period. Average
breathing and yawning rates were tabulated for
every 1min interval. Yawn duration was not
analyzed because of potential problems that
subjects may have had in recording yawn duration
in a comparable manner while sitting in the
control condition and stepping in the exercise
condition. Such potential problems would be
unlikely to influence the all-or-none recording
of yawn occurrence, the variable of primary
interest.
Results The effect of moderate
exercise on breathing and yawning rates is
shown.in Fig 4. During the 10-min exercise
period, breathing rate rapidly increased until
it was almost twice that of the preceding 10-min
baseline period. Breathing rate declined rapidly
toward preexercise levels during the recovery
period. Significant differences in breathing
rate were detected among all three conditions
(baseline A; exercise, and baseline 13) using a
multivariate test (F(2, 17) = 42.711, p <
.0005) and a univariate repeated measures test
(F(2, 36) = 69.73, p < .0005). Both the
exercise versus baseline contrast and the
contrast between the two baseline conditions
were significant (FI, 18).- 90.02 and 18.55,
respectively, p < .0005). Yawning rate
declined across the three conditions of the
experiment and no effect of exercise was
detected. A multivarlate test of the effect of
conditions was significant (F(2, 17) = 5.36, p =
.016). The corresponding F ratio from a repeated
measures ANOVA was also significant (F(2, 36) =
5.67, p = .007). An evaluation of univariate
contrasts indicated that the number of yawns
produced in the exercise condition was not
different from the average number in the two
baseline conditions. The two baseline conditions
were, however, significantly different from each
other (F(I, 18) = 11.35, P = .003). These two
contrasts are equivalent to quadratic and linear
trends, respectively. Thus, yawn frequency
declined in a systematic linear manner across
the three conditions (time periods) of the
experiment. The fallure to detect a significant
effect of exercise on yawning rate was not due
to the inadequacy of exercise to produce a
substantial physiological effect on subjects;
exercise produced a significant effect on one
simultaneously recorded behavioral variable,
breathingy rate.
DISCUSSION
The present
research found no support for the popular
hypothesis that yawning is a response to
elevated CO2 or depressed O2 levels in the
blood. Neither breathing pure O2 nor
gases high in CO2 had a significant effect on
yawning although both increased simultaneously
recorded breathing rate. Exercise that doubled
breathing rate also had no effect on yawning.
The increase in breathing rate to elevated CO2,
O2, and exercise established that all three
interventions had measurable physiological
effects. Breathing and yawning share some
similar motor processes that produce inspiration
and expiration, but the independent variation of
yawning and respiration suggested that breathing
and yawning are responses to different internal
states and are controlled by différent
mechanisms. Although it can be argued that other
gas concentrations, exposure times, or
instructions may have produced different
findings, there are no contrary data in the
hurnan yawning literature and the prescrit
results are a part of a broader pattern of
converging evidence.
Further evidence that yawning is not
primarily a respiratory act comes from an
analysis of the routes of inhalation and
exhalation during yawning and the relation
between yawn duration and inter yawn interval
(Provine, 1986; Provine, Hamernik, Tate, &
Curchack, 1986). Inhalation and exhalation occur
primarily through the mouth and this basic
pattern is highly inflexible. It is very
difficult, if not impossible, for most people to
perform a satisfying yawn with their mouth taped
shut even though they are free to respire
through their nose and rnove their jaw to a
limited extent within their closed mouth.
Yawning does not have the degree of behavioral
freedom characteristic, of normal breathing.
However, oral inhalation by itself is
insufficient to produce a satisfactory yawn if
the jaws are not free to move. Subjects
attempting to yawn with clenched teeth often
reported the unpleasant sensation of being stuck
in mid yawn and being unable to perform a
satisfying yawn although they could inhale and
exhale through their teeth. Thus, the yawn is
not simply a deep breath. The finding that
performers of short yawns do not compensate by
yawning more frequently and vice versa is also
inconsistent with the yawn as a respiratory
event (Provine, 1986).
The primary function of the highly
stereotyped action pattern of yawning may be
some nonrespiratory component of the behavior
such as sretching. A precedent exists for the
association of yawning and stretching. Subjects
who cannot stretch their jaws because their
teeth are clenched are unable to perform
satisfactory yawns (Provine, 1986), and
generalized body stretches often accompany yawns
in humans (Provine et al) and other animals
(Dourish & Cooper). Neurological case
reports of associated movements in human
hemiplegics provide further evidence for a
correlation between yawning and stretching.
Herniplegics often stretch otherwise paralyzed
body parts during yawns (Walshe, 1923).
Pharmacological evidence for a yawn-stretch
relation is provided by the observation that
drugs that produce yawning also trigger
stretching in a variety of animals (Dourish
& Cooper, in press; Gessa, Pisano, Vargiu,
Crabai, & Ferrari, 1967; Yamada &
Furukawa, 1980). There is also evidence of at
least partial autonomy of yawning and
stretching. Yawns may be performed without
concurrent stretches and vice versa. The
circadian patterning of yawns differs from that
of stretches; yawning is most frequent shortly
before sleeping and after waking, whereas
stretching is frequent only immediately after
waking (Provine et al).
The difference in the temporal distribution
of yawning and stretching and the ability to
perform elther behavior autonomously suggests
that yawning is not simply the facial component
of a generalized stretch response. During
evolution, yawning may have become emancipated
from a generalized stretch response or may have
evolved independently but served a furiction
homologous to stretching. Stretching has
consequences ranging from increases in blood
pressure and heart rate to improved flexibility
of muscles and joints. If the yawn is a
localized stretch of the face and upper body,
its consequences may be as diverse and plentiful
as those of other types of stretches. However,
yawning is unique in having important
conseqpences for both groups and individuals. If
yawning is assumed to,have a physiological
effect on the yawner, the tendency to yawning
response to observed yawns produces a behavioral
chain reaction among witnesses that synchronizes
the physiological state of a group (Provine et
al.). Both yawning and its resulting physiology
are contagious.
