Plate I — Topography of the involuntary yawn
Figs 1–3 · lateral aspect · ×0.55tap to hold the peel
The eight leaders 1–8
Each number over the plate names a structure the reflex cannot do without. Hover to peel the layers back to it; tap to hold the peel while you read.
Ink on linen · 34 × 27 cm
Specimen No. 042 · lot xii.4
§ 1The mechanics, in three acts
Traces generated in real time from the same curve that drives Plate II. Fig. 8 below.
Physiologically the yawn is not one movement but a small, well-rehearsed orchestra, and it runs in three acts.2 What follows is pinned to a single specimen on the left; click any act heading to drive that plate to the corresponding beat.
A sensation of tightness under the chin, and the jaw begins to descend slowly — this is the part subjects describe as the yawn "coming on", and it is the part over which they retain the least. The lateral pterygoids pull the condyles forward off their eminences; the temporomandibular joint, which is a hinge for the first fifth of an opening and a sliding joint for the rest, performs both jobs at once. The eyes begin to narrow. The tongue presses the floor of the mouth.
The mandible reaches 25 to 35 millimetres of interincisal opening in the adult, the orbicularis oculi contracts and closes the lids — squeezing the lacrimal apparatus and producing the tear that so many readers will have noticed in Plate II — and the diaphragm descends in what is functionally a spasmodic flattening, dragging the lung bases with it. Between one and two litres of air are inspired against a glottis that has partly closed, which is why the yawn has a sound. The soft palate lifts; the tensor veli palatini opens the auditory tubes and the ears pop.
At the summit there is typically a brief extension spasm of the trunk and limbs — the arms go up, the fingers spread, the neck lengthens — followed by passive recoil: the jaw closes, the diaphragm rises, and the breath leaves audibly. Total duration in the laboratory averages five to six and a half seconds.3 Nothing about it has been demonstrated to be useful. The subject then apologises.
Plate II — The act itself, in four beats
Figs 4–9 · same specimen, successive · loop 7 sthe act from the top
Subject no. 042, unprepared
Yawned at 5.9 s. Reported, afterwards, that "nothing happened for a moment and then it was simply finished". No hypoxia, no stimulation, no towel. Lachimation left, lower punctum, graded 1 of 3.
Trace sampled 30 Hz
Plate engraved after the life
§ 2Contagion, and its discontents
That a yawn should be catching is the single most studied and least explained fact about it. Roughly a third to two-thirds of healthy adults will yawn within a minute of watching one, and the effect survives being told about it, surviving being paid not to, and — in our own theatre — surviving the presence of a second experimenter counting on his fingers.4 The susceptibility is not distributed evenly and is not, apparently, distributed on purpose.
Two things the reflex is not are worth stating plainly, because they are still repeated. It is not a response to low oxygen: raising inspired oxygen, lowering it, and loading the blood with carbon dioxide all leave the spontaneous yawn rate of the resting adult substantially unchanged.5 And it is not a sign of boredom in any strict sense, though it is faithfully reported by audiences during lectures — including, we are told, by the two experimenters during the control film of the towel.
What remains is a developmental and comparative profile that looks, uncomfortably, like a social capacity. Contagious yawning cannot be reliably demonstrated in human children before the second year and is established by four to six.6 It is present in chimpanzees watching a conspecific on video. It is present in dogs watching a human — and, in the study that settled the argument for us, not present in the same dogs watching another dog yawn, which remains one of the neatest negative results in the literature.7 It is present in budgerigars that have been housed together long enough to know one another. And in macaques, so far, nobody has managed to make it happen.
Fig. 10 · Susceptibility, tested on the reader
Three silent yawns, five seconds each, from our confederate (female, 24, no pathology). Press nothing, or press the truth. The log is session-only; nothing leaves this page.
Plate III — Comparative gaping, and the impostor
Figs 11–12 · counterphase · aquarium, 2nd flooropened against each other
Fig. 12 The human, run on the same clock at half a cycle's offset, so the two jaws open against each other like a canon. We keep this figure in the plate because it is the honest version of the comparison: the resemblance between a yawn and a threat gape is the reason the reflex has spent a century accumulating functions it cannot all have.
§ 3Method, briefly, and with the usual embarrassments
Two hundred and four adults (mean age 31.4 years; 96 women, 106 men, two who declined the category and one who declined the study after it began) were seated in the lower gallery of the physiology lecture theatre and exposed, in counterbalanced order, to three four-minute conditions: a yawning confederate seated two rows ahead; a silent film of the same confederate yawning; and a control film of a man folding a towel with great concentration.9 Yawns were counted by hand from the front, by a second observer who was not told which condition the room had received.
Of the 204, 138 yawned in at least one condition, and 61 in none — a susceptible fraction of 67.6 per cent, at the optimistic end of the published range, and probably a compliment to the confederate. Forty-six of the 138 who yawned at all did so within the first two seconds of the stimulus; the distribution of latencies is not normal and we have stopped pretending it will be. Nineteen participants asked afterwards whether they had been given anything. They had not. The towel was folded for four minutes by a man named Roel who is, by all accounts, extremely good at folding towels, and produced eleven yawns — a rate we attribute to the room, the hour, and the lecture preceding it.
The design is unblinding; the hypothesis is not. We report it anyway, because a plate that admits its own limitations is still a plate.
| Condition | Yawned | Median latency | Mean yawns | Notes |
|---|---|---|---|---|
| Live confederate | 121 / 204 | 1.9 s | 2.4 | strongest — eye-line available |
| Film of confederate | 104 / 204 | 2.6 s | 1.8 | silent, 14 in., no sound cue |
| Control film (towel) | 11 / 204 | n/a | 0.1 | room heated to 23 °C |
| Reading this plate | — / — | — | — | you know who you are |
§ 4The remaining difficulty
There are three candidate explanations for why we yawn, and the honest state of the art is that two of them are unfashionable and the third is unproven. The oxygen hypothesis is dead and refuses to lie down.10 The thermoregulatory hypothesis — that a yawn is a small, undignified radiator — has attractive experiments attached to it and no consensus. The developmental hypothesis, that the antenatal yawn is maintenance work for the jaw, the ear and the lung, is the one our own plates do most to support and the least one anyone funds.
And then there is the contagion, which behaves exactly like a mild pathogen — with a dose, a latency, a susceptible fraction and a species barrier — and which has never been isolated. We have drawn it as carefully as we could. It is still a drawing.
Correspondence I. Marchetti-Vane, yawn desk, University of Groningen-sur-Meuse. The authors thank the confederate, the pufferfish, and Mr Roel.
—Notes, references & the small print
Notes
1 The comparison with the pulse is not rhetorical: subjects can suppress a yawn for a while, but suppression lengthens and distorts the act rather than preventing it, and the rebound is real.
2 Our three-act division follows the classical one; Provine's laboratory used two and never regretted it.
3 Duration measured from first descent of the mandible to the end of the audible exhalation. Means of 5.0–6.5 s are typical for the adult in the rest state.
4 Our own theatre, this volume, §3. Susceptible subjects yawn faster when instructed not to, which we regard as the most human result in the paper.
5 Goetz, Provine & Abling (1996) found no change in yawn rate under hyperoxia, hypoxia and CO2-enriched air. The towel condition was designed by the same instinct.
6 Contagious yawning is essentially absent before the second year, unreliable at two to four, and established by four to six — a developmental curve that overlaps, awkwardly, with the appearance of a theory of other minds.
7 Cordoni and colleagues recorded contagious yawning in 32 domestic dogs exposed to human yawning, and no such effect against dog models — the direction of the barrier, not its existence, is the surprise.
8 On "fake yawning" in vertebrates see the review of Hinde & Cameron (2003), who catalogue the mouth-opening displays of crocodilians, lizards, frogs and fishes and caution that the resemblance to our own yawn is precisely the difficulty.
9 The confederate was a second-year student paid in coffee. She requested that the record show she found the whole exercise "very tiring".
10 The antenatal record is the awkward one: yawning is observed in the fetus from about the eighteenth week, is common in the second trimester and diminishes towards term — which is why nobody accepts fetal yawning as evidence of fetal wakefulness, and why the oxygen theory cannot survive contact with it either.
References
Provine, R. R. (1996). Yawning. Scientific American 274(5), 108–111.
Anderson, J. R., Myowa-Yamakoshi, M. & Matsuzawa, T. (2004). Contagious yawning in chimpanzees. Proceedings of the Royal Society B 271(Suppl.), S468–S470.
Cordoni, A. et al. (2016). Dogs (Canis familiaris) yawn contagiously when exposed to human yawning. Biology Letters 12(11), 20160516.
Guggisberg, A. G., Mathis, J., Schnider, A. & Morishima, Y. (2010). Yawning: an update. Open Neurology Journal 4, 130–134.
Hinde, R. A. & Cameron, J. L. M. (2003). Yawning and fake yawning: a comparison between birds and mammals. Behaviour 140, 1219–1234.
Marchetti-Vane, I. & Nyström, O. K. (2025). The towel control: on boredom as an experimental confound. J. Cur. Physiol. 11(3), 88–104.
Verdoes, H. L. (1902). On the contagiousness of illustrations. Proc. Soc. Inquiries Too Small for Elsewhere xvii, 4–19.
Cited again in
- J. Cur. Physiol. 9:2 · "Stretching without a yawn: a survey of 300 gym-goers"
- Comparative Sleep Letters 44:1 · "The pufferfish defence"
- Rev. Unintended Gestures 3:2 · "Coughing, sneezing, and the polite reflex"
- Annals of the Lecture Theatre 12:4 · the towel paper
Acknowledgements
Dr H. L. Verdoes for the plate frames and for insisting, twice, that the diaphragm be drawn from the life. The aquarium staff of the second floor, who fed the fish on schedule. Mr Roel for four minutes of extraordinary towel.
Competing interests
None declared. One author yawns.
Ethics
Approved by the local committee, with the committee's own yawning unrecorded.
Colophon
Set in the journal's plate hand and composed for the press of the Society for Inquiries Too Small for Elsewhere, Groningen-sur-Meuse. Engravings by the authors from the life, on copper, re-inked twice; the animation you are watching is a modern affectation and was drawn with the same three paths. Print run 400. Of these, 380 were distributed and 20 were yawned on.