Eight cat larynges sat on a laboratory bench at the University of Vienna. No brain attached, no nerves, no muscles firing. Air was pushed through them, and all eight began to oscillate between 25 and 30 Hz, which is the cat purr frequency you feel through your ribs when a cat decides your chest is furniture. The result was published in Current Biology on 3 October 2023, and it quietly demolished fifty years of textbook explanation.
The purr has a long history of making serious researchers look silly. It has been blamed on turbulent blood, on a bone in the throat, on a neural metronome in the brainstem. Along the way one number, 25 Hz, escaped the laboratory and became a wellness claim that cats heal broken bones. That claim has a real scientific ancestor, just not the one people think. We have covered the basic mechanics of purring before. This is the paper trail: who measured the sound, when, with what instruments, and where the numbers got bent.
Table of Contents
- A Sound Nobody Could Locate
- The Two Studies That Pinned Down Cat Purr Frequency
- Where the Cat Purr Frequency Healing Claim Was Born
- What Clinton Rubin Actually Did
- The 2023 Experiment That Broke the Model
- Which Cats Purr, Which Cats Roar, and Why the Hyoid Story Is Shaky
- The Purr Is Also a Negotiation
- Frequently Asked Questions
A Sound Nobody Could Locate
The strange part is not that the purr took long to explain. It is that for most of the twentieth century nobody could agree on which organ was making it. A cat purrs with its mouth shut, through inhalation and exhalation, at a rate an order of magnitude below human speech. So the search went sideways.
The Blood Flow Theory
One durable idea held that purring was hemodynamic, a fremitus produced by blood moving through the great vessels of the chest. In the version that circulated in veterinary writing, a cat arching its back to be petted kinks the caudal vena cava or the aorta, the flow goes turbulent, and the whole thorax buzzes like a badly seated pipe. It fails on anatomy, because those vessels keep their lumen when the spine arches, and on behaviour, because cats purr while lying flat, standing, walking and giving birth.
The other early framework came from taxonomy rather than physiology. In 1916 the British zoologist Reginald Innes Pocock formalised a split in the cat family: purring cats on one side, roaring cats on the other, with the difference pinned to whether the hyoid apparatus in the throat is fully ossified. That classification outlived the blood theory by a century and is still repeated in nature documentaries. It is also, as we will get to, wrong in an instructive way.
The Two Studies That Pinned Down Cat Purr Frequency
Two papers, nineteen years apart, built the standard model. Both used instruments rather than intuition, which is what finally moved the argument.
Remmers and Gautier, 1972
J. E. Remmers and H. Gautier published “Neural and mechanical mechanisms of feline purring” in Respiration Physiology in December 1972, volume 16, pages 351 to 361. They put electromyographic electrodes into the laryngeal muscles and the diaphragm of chronically instrumented cats and recorded electrical activity that came in bursts locked to the purr. The muscles were not just present during purring, they were firing in time with it, roughly thirty times per second.
Their interpretation set the agenda for half a century. Something in the central nervous system, a free running oscillator, drove the larynx open and shut at a fixed rate. This became the active muscle contraction hypothesis, and every popular explanation of purring written between 1975 and 2023 is downstream of it, usually without naming it.
Sissom, Rice and Peters, 1991
The follow up came from Dawn Sissom, D. A. Rice and Gustav Peters, who published “How cats purr” in the Journal of Zoology in 1991. They reviewed the competing explanations, noted there was no consensus among anatomists, then went and recorded animals: shelter cats, plus a cheetah and a cougar. Two channel acoustic measurement showed that the purr runs continuously across the whole respiratory cycle, changing character but never stopping as the airflow reverses, and that the source is laryngeal modulation of that airflow.
That combination, EMG plus acoustics, is what let textbooks state a number with confidence: domestic cats purr at roughly 20 to 30 cycles per second. Measuring something precisely enough to name it is usually where a field stops arguing, the way gravitational waves stopped being a debate once an instrument registered one. With the purr, precision arrived first and the explanation turned out to be wrong anyway.
Where the Cat Purr Frequency Healing Claim Was Born
In 2001 the bioacoustician Elizabeth von Muggenthaler, of the Fauna Communications Research Institute in North Carolina, presented work under a title designed to travel: “The felid purr: A healing mechanism?” It appeared in the Journal of the Acoustical Society of America, volume 110, issue 5 supplement, page 2666. That supplement is a meeting abstract volume, which matters. It is a conference presentation, not a full peer reviewed paper, and it has been cited for twenty five years as though it were the latter.
The work itself was real. Forty four felids were recorded, including domestic cats, cheetahs, ocelots, pumas and servals, on a Sony TCD-D8 digital audio recorder with Statham microphones and an accelerometer for the domestic cats. Every animal produced strong frequency content between 25 and 150 Hz, with dominant frequencies clustered near 25 Hz.
Then came the sentence that did all the work. Those frequencies, the abstract noted, correspond to the vibrational ranges used therapeutically for bone growth, fracture repair, pain, edema, muscle strain, joint mobility and wound healing. That is a statement about overlap between two numbers. It is not a claim that a cat on your lap delivers a therapeutic dose. In the retelling, the hedge evaporated. By the 2010s the internet had converted it into the flat assertion that a purring cat heals human bones, a claim now filed alongside the older folklore we traced in our piece on black cat superstitions, except with a citation attached to make it look empirical.
What Clinton Rubin Actually Did
The therapeutic frequency range that von Muggenthaler was pointing at is not folklore. It comes from a specific line of biomedical engineering, most of it associated with Clinton Rubin at Stony Brook University. The landmark result was published in Nature in 2001, volume 412, pages 603 to 604, under the title “Low mechanical signals strengthen long bones.”
The subjects were adult sheep. They stood on a vibrating plate for twenty minutes a day, every day, for a full year, at 0.3 g of acceleration and 30 Hz. After twelve months, trabecular bone density in the proximal femur was 34.2 percent higher than in controls, at strain levels three orders of magnitude below what damages bone tissue. It is a genuinely odd piece of skeletal biology, and our roundup of strange human body facts has more like it.
Note what the protocol required. Direct coupling through the feet, a controlled acceleration, a fixed daily dose, a year of it. A cat lying on your sternum meets none of those conditions. Whatever vibration reaches your skeleton through fur, muscle and connective tissue is unmeasured, undosed, and not aimed at a load bearing bone.
What Happened When They Tried It on People
Rubin and colleagues ran a one year randomised double blind placebo controlled trial in 70 postmenopausal women, published in the Journal of Bone and Mineral Research in March 2004. Two ten minute sessions per day at 0.2 g and 30 Hz. Among the most compliant participants, femoral neck density held roughly flat while the placebo group lost 2.13 percent, a relative benefit of about 2.17 percent that landed at p equals 0.06. The clearest signal was in a subgroup: lighter women, under 65 kg, in the top compliance quartile, who gained 3.35 percent relative spinal density.
The larger test went the other way. The VIBES trial, led by Douglas Kiel and published in the same journal in 2015, put 174 elderly adults with a mean age of 82 through two years of low magnitude vibration at 0.3 g and 37 Hz for ten minutes daily. Adherence was around 70 percent. At 24 months there was no significant difference between active and placebo groups in volumetric bone density at the femur or the spine, and no difference in bone turnover markers.
So the honest summary is this. Mechanical vibration in the 20 to 50 Hz band does something to bone in animal models, produced modest and compliance dependent effects in one small human trial, and produced nothing measurable in a larger and older human population. That is the actual state of the evidence the cat claim is borrowing from, and it is much thinner than the meme.
The 2023 Experiment That Broke the Model
Christian T. Herbst, a voice scientist at the University of Vienna, studies how larynges make sound across species. With colleagues from Vienna, Palacký University Olomouc and the Research Institute of Organic Agriculture in Switzerland, he mounted excised domestic cat larynges in a rig that pushes controlled airflow through tissue. No brain, no nerves, no possibility of thirty contractions per second, because nothing was left to contract on command.
All eight larynges produced self sustaining oscillation at 25 to 30 Hz. The paper, “Domestic cat larynges can produce purring frequencies without neural input,” ran in Current Biology on 3 October 2023. Anatomical work turned up something the earlier studies had missed: a pad of connective tissue embedded in the vocal folds, which increases their mass and lets a four kilogram animal generate frequencies you would normally expect from something much larger.
This does not prove that living cats purr passively. It proves the active muscle contraction hypothesis is not required to explain the frequency, which is a different and more awkward result. The current working picture: the brain sends an initiating signal and sets laryngeal posture, then aerodynamics does the repetitive part, the way a reed keeps buzzing once you blow through it. Fifty years of confident diagrams turned into an open question in one afternoon of controlled airflow.
Which Cats Purr, Which Cats Roar, and Why the Hyoid Story Is Shaky
Pocock’s 1916 division is the version everyone learned. Cats with a fully ossified hyoid purr and cannot roar. Cats with an incompletely ossified hyoid roar and cannot purr. On the roaring side sit lion, tiger, jaguar and leopard. On the purring side sits everything else, including some large animals: cheetahs purr, and pumas purr, chirp and whistle.
The problem is the snow leopard. It is the fifth species with an incompletely ossified hyoid, it belongs to Panthera, and it does not roar. It chuffs, meows and makes purr type vocalisations. In 2002 Gerald Weissengruber and colleagues argued that hyoid ossification is not what decides the question at all, and pointed instead at the vocal fold tissue: the roaring cats carry a thick fibro elastic pad on the folds that would damp the fast small oscillation a purr requires. Given that Herbst’s team found a specialised pad inside the domestic cat’s folds doing the opposite job, the tissue explanation now looks considerably stronger than the bone one.
Much of the standard cat sound story rests on anatomy described once, a century ago, then repeated. The same goes for the stranger items in the repertoire, like the stuttering chatter cats aim at birds through a window, which we picked apart in our piece on chirping.
The Purr Is Also a Negotiation
Mechanism aside, there is the question of what the purr is for, and here the research is unusually unflattering to cats. In 2009 Karen McComb and colleagues at the University of Sussex published “The cry embedded within the purr” in Current Biology. They found that purrs recorded while cats were soliciting food contain a high frequency component sitting inside the low rumble, acoustically close to a cry or a meow. Played back at matched volume, listeners rated those solicitation purrs as more urgent and less pleasant than ordinary purrs, and people with no cat experience rated them the same way. The cat has found a frequency that piggybacks on the human response to an infant.
There is also a genetic angle, and it is recent. On 28 May 2025, Yume Okamoto and colleagues at the Kyoto University Wildlife Research Center published a study in PLOS One on 280 neutered mixed breed house cats, comparing owner reported behaviour against variants of the androgen receptor gene. Cats carrying the short type variant were reported to purr more. Males with the short type were more vocal towards people. Comparing 11 felid species, the team found the long type variant only in domestic cats, not in the fishing cat or the leopard cat, which suggests it appeared during domestication. The cats that talk at you least, in other words, may be the most domesticated ones, which is not the direction anyone expected. It pairs with the other habit cats reserve for humans, kneading.
Frequently Asked Questions
What is the actual frequency of a cat’s purr?
Roughly 20 to 30 Hz for domestic cats. The 2023 Vienna experiment measured excised larynges oscillating at 25 to 30 Hz, matching the range reported acoustically since the 1991 Journal of Zoology work. The wider 25 to 150 Hz range comes from von Muggenthaler’s 2001 survey of 44 felids and includes harmonic content, not the fundamental rate.
Can a cat’s purr heal broken bones?
There is no study showing that. The claim borrows evidence from vibration therapy research, where sheep standing on a calibrated plate for a year gained bone density. The largest human trial of that therapy, published in 2015, found no significant effect. Nobody has measured the vibration a purring cat actually delivers to a human skeleton, let alone its clinical effect.
Do cats purr only when they are happy?
No. Cats purr while injured, while being examined by a vet, while giving birth and while dying. The 2009 Sussex work also identified a distinct solicitation purr used to extract food from humans. Purring is better read as a signal about the cat’s state and intent than as a simple happiness meter.
Why can’t lions and tigers purr?
The traditional answer is the incompletely ossified hyoid bone, from Pocock’s 1916 classification. That answer has a hole in it, because the snow leopard has the same hyoid and does not roar. The 2002 Weissengruber analysis points instead to a thick fibro elastic pad on the vocal folds of roaring cats, which would damp the rapid low amplitude oscillation a purr needs.
The Short Version
The purr has been misexplained by professionals for a century, and the current best model is only three years old and still incomplete. The healing claim is not an invention, it is a misreading: one careful hedge in a conference abstract, welded to a real vibration therapy literature that has since produced a null result in its biggest trial. None of which makes a cat on your chest less pleasant. It just means the thing vibrating at 25 Hz is doing it for its own reasons, and has not agreed to fix your femur.
🐾 Visit the Pudgy Cat Shop for prints and cat-approved goodies, or find our illustrated books on Amazon.
Leave a Reply