Why Cats Purr

FĒLIS Editorial Feb 2026 18 min read

The frequency of a cat's purr concentrates between 25 and 50 hertz, a number that has a completely different use in sports medicine and orthopedics: the optimal mechanical vibration window for activating osteoblasts. The working frequency of clinical vibration platforms used to treat osteoporosis falls right in this range.

Start with the sound-production mechanism. Neural oscillators in the brainstem send rhythmic electrical signals to the intrinsic laryngeal muscles at a rate of 25 to 150 times per second, the glottis opens and closes repeatedly, air passes through this narrow slit during both the inhalation and exhalation phases, and the purr does not interrupt with breathing. In recent years there is a new hypothesis, still unresolved, that a type of connective tissue pad is embedded in the vocal folds, possibly allowing passive airflow-induced vibration, similar to snoring, not entirely dependent on active neural drive. If confirmed, energy expenditure would be even lower than previously estimated.

Purring is not sound. Vocalization sends information into the air; purring is mechanical vibration, a physical force applied to tissue. Being audible to the ear is a side effect. 25 to 50 hertz falls in the optimal frequency band for bone repair; if the original function were communication, landing on this band would make no sense. It most likely was self-repair behavior from the very beginning.

Bone

Author's Note

The section below will be much longer than any other part of the article, not for structural balance, but because there is genuinely an order of magnitude more to talk about in this direction than in any other.

Felines spend 16 to 20 hours a day lying down. A basic rule of bone biology is that mechanoreceptors on the surface of bone cells need to continuously receive loading signals to keep the osteogenic pathway active; prolonged absence of loading input downregulates osteogenesis, upregulates osteoclast activity, and bone mass declines. Astronauts lose about 1.5% of bone density per month in microgravity. Cats lie around that long and by this rule should be in trouble. They are not.

The vibration of purring propagates through bone conduction to the entire skeleton; to the mechanoreceptors, this vibration signal is sufficiently similar to the loading signal produced by weight-bearing. The mechanical environment a cat's bones receive while lying down is of the same kind as that of a human patient standing on a clinical vibration platform.

There is something related to Wolff's law worth saying more about here. Wolff's law states that bone remodels its own structure and morphology along the direction of mechanical loading. Traditional discussion of this law focuses on macroscopic loading, the forces exerted by walking, running, and jumping, on the order of hundreds to thousands of microstrain. The vibrational load applied by purring is far lower in magnitude and far higher in frequency. If the vibration of purring is sufficient to maintain the osteogenic pathway in an active state, it means the response threshold of bone cell mechanoreceptors is much lower than what is discussed under the macroscopic mechanical framework. People doing research on the effects of low-magnitude vibration stimulation on bone density (Clinton Rubin's lab at Stony Brook has done considerable work in this direction) have directly discussed this threshold question, though their animal models were mainly turkeys and mice, not cats. Cats in this context are more like an uncontrolled experiment that nature has been running for tens of millions of years, the results of which appear consistent with the predictions of vibration platform experiments, except that no one has ever done a prospective study in this specific direction using cats.

When NASA researched anti-bone-loss measures for astronauts, it invested in low-frequency vibration therapy, and the final working frequency of the vibration platform overlaps heavily with the dominant frequency band of cat purring. Did the early literature explicitly write "referenced cats"? In some review-type articles one can indeed find openings that use feline resting time and skeletal health as a lead-in; whether that counts as "inspiration" or merely "a hook for the introduction" is a judgment left to the reader.

On the level of veterinary orthopedic experience, cats heal fractures at a speed and quality that is abnormally good among mammals of comparable size. The source of this claim is not systematic comparative data from any single paper; it is the same impression independently and repeatedly reported by clinical veterinarians across different regions over decades. The credibility of experiential consensus and the credibility of controlled experiments are not on the same level; this must be acknowledged. Why no one has done a controlled experiment will be discussed below; it is directly related to ethics review.

The cheetah situation. The cheetah is the only heavy-pursuit predator among felids, sprinting at 112 kilometers per hour, with bending stress on the spine and limbs during high-speed running far exceeding that of ambush-type felids. If purring only compensated for bone mass loss from "lying down too long," the cheetah's need should be the lowest. The cheetah's purr is among the loudest in the cat family. So the skeletal function of purring is not limited to "compensating for resting"; it may also include post-exercise microdamage repair. These two functions can coexist.

On the topic of bone there is one more thing that has always seemed strange, and in articles discussing purring almost no one has ever mentioned it: the impact loading on bones from a cat's jumping and landing is enormous. A four-kilogram housecat jumping down from the top of a refrigerator can produce peak ground reaction forces at the moment of landing several times its body weight. Cats perform large numbers of such jumps every day. If you stack the bone density maintenance demand from prolonged resting together with the microdamage repair demand from frequent high-impact landings, the skeletal maintenance function of purring is no longer the single narrative of "compensating for sleeping too much"; it may be simultaneously addressing problems from two directions: one is disuse-related bone mass decline from "too quiet," the other is stress microdamage from "too intense." These two demands alternate within the same cat on the same day and are both covered by the same 25-hertz vibration. Whether anyone has written about it from this angle is unknown; it has not been seen in the existing literature.

Analgesia

Purring vibration activates Aβ fibers, thickly myelinated, fast-conducting. The C fibers and Aδ fibers that transmit pain are slow. Melzack and Wall's gate control theory, 1965: fast non-nociceptive signals inhibit slow nociceptive signals from ascending at the dorsal horn of the spinal cord. The vibration of purring competes with pain signals for transmission channels at the neural circuit level. Rhythmic somatic stimulation also promotes the release of endogenous opioid peptides; gate-control analgesia and endorphin-mediated analgesia most likely run in parallel.

Multiple clinical veterinarians have independently reported that cats begin purring during the euthanasia injection process, and purring continues after the drug suppresses cortical function. General anesthesia data are the same: when the threshold of consciousness loss has just been passed, some cats are still purring; it disappears only at deeper levels. The control center is in the brainstem, not the cortex.

48 Hours and Genetic Encoding

At 48 hours after birth, kittens begin to purr. Hearing and vision are not yet fully developed. The mother cat purrs continuously during nursing and transmits this to the kittens through body-contact bone conduction. Orphan kittens without a mother cat to demonstrate still spontaneously initiate at 48 hours. Innate behavioral pattern, hardcoded in the genome. Meowing is highly dependent on postnatal social learning, with enormous variation across different environments.

Vibrational communication attenuates extremely fast, with an effective range limited to body-contact distance, leaking no information into the environment. For kittens in a nest, not being detected at a distance by predators means staying alive.

Vagal Nerve Stimulation Under Stress

The cat purring in the veterinary clinic, the cat purring in a cardboard box after moving house.

Rhythmic contraction of the laryngeal muscles coupled with controlled breathing stimulates the vagus nerve, raises parasympathetic tone, lowers heart rate, blood pressure, and cortisol. Same pathway as a human doing slow deep breathing. Humans must consciously initiate it; cats automatically initiate once stress crosses a threshold.

High-Frequency Embedding in Solicitation Purring

In 2009, Karen McComb's team at the University of Sussex discovered that when cats solicit food, a high-frequency component between 220 and 520 hertz is embedded in the purr's frequency spectrum, very weak or absent in the normal contentment purr. It overlaps with the fundamental frequency range of a human infant's distress cry. Test subjects who did not own cats also rated it as more urgent and harder to ignore.

The two signals have unrelated production mechanisms. The innate high sensitivity of the human brain to this frequency band was shaped by infant survival pressure. Through generational selection pressure during domesticated cohabitation, cats shifted the spectrum of the solicitation purr toward this range. Cats that could produce a purr more uncomfortable to humans reproduced more successfully, and the trait became fixed in the population. A parasitic signal strategy in signal theory.

After McComb 2009, in-depth follow-up work in this direction has been scarce. Tracing the citation chain on Google Scholar turns up some review citations and a small number of replication experiments, but little systematic follow-up.

Multi-Cat Households

In multi-cat households, the intensity of high-frequency embedding in different cats' solicitation purrs varies widely; cats with stronger embedding have an advantage in capturing human attention.

There is a question here not directly related to purring but directly related to the domestication identity of cats. Canine domestication was human-directed selective breeding; cat domestication is closer to self-domestication, with cats approaching granaries on their own, and individuals that could tolerate human presence having higher survival rates. The spectral shift of the solicitation purr is not selection imposed by humans on cats; it is exploitation of the human perceptual system that cats evolved on their own in a cohabitation environment. When the word domestication is used for cats, who domesticated whom is itself a matter of debate. The solicitation purr is a very small but very persuasive piece of evidence in that debate.

Hyoid Bone

Felids that can purr cannot roar; those that can roar cannot purr. The dividing line between Felinae and Pantherinae. In the Pantherinae hyoid chain there is an unossified elastic ligament, providing the larynx with greater range of motion, enabling roaring but not purring. In Felinae the hyoid is fully ossified, high rigidity, enabling purring but not roaring. The divergence was approximately 10.8 million years ago. The snow leopard belongs to Pantherinae but cannot produce a typical roar; there are reports that it can produce a vibration close to purring, and its degree of hyoid ossification may be intermediate between the two subfamilies.

Purring's Effect on Humans

People in prolonged close body contact with a purring cat are continuously receiving low-dose whole-body vibration at a frequency highly overlapping with that of clinical vibration platforms. Clinical data are lacking. The promoting effects of low-frequency mechanical vibration on fibroblast migration and angiogenesis have a fairly solid body of data in tissue engineering in vitro experiments. Whether prolonged body contact with a purring cat has a measurable effect on post-surgical recovery in humans has not been studied by anyone.

Why No One Has Done It

Canine research has sustained funding from military, law enforcement, guide dog, and search-and-rescue sectors. Cats have no corresponding application scenario. Virtually the only sources willing to fund non-clinical basic research on cats are pet food company sponsorships and a small number of veterinary school internal funds, disproportionate to the global population of over 600 million domestic cats.

Ethical barrier: to verify the efficacy of purring on fracture healing, the cleanest experiment would be creating fractures under controlled conditions and comparing a purring group to a purr-blocked group, which would not pass ethics review, nor should it. What remains is retrospective analysis of naturally occurring cases, with poor variable control.

Technical barrier: purring amplitude is low, the frequency falls in the low-frequency range where acoustic equipment sensitivity is poor, and cats frequently stop purring after sensors are attached. Some researchers have tried long-duration recording with non-contact microphone arrays in home environments; data can be collected, but the noise-separation workload is enormous.

There is another layer: purring does not kill, is not contagious, and causes no economic loss. Research funding review tends to allocate money to "problems." "Why do cats purr" is not a problem. People who write grant applications know how to package a phenomenon as a problem; purring can be packaged toward osteoporosis treatment and aerospace medicine, which makes logical sense, but compared to "this parasite causes X billion dollars in losses annually," the persuasive power is far weaker.

The near absence of major progress in the solicitation purr direction after McComb 2009 is the same reason. One beautiful paper was done, published, and afterward there was no money to support systematic follow-up. That paper is like a single marker post standing in an empty field.

On the bone repair direction, Clinton Rubin's lab at Stony Brook has done a large body of work on the effects of low-magnitude vibration on bone density, mainly using turkey tibia and mouse models, with results published in Journal of Bone and Mineral Research and similar venues. The connection between this line of research and cat purring is extremely obvious in logic and extremely rare in cross-citation in the literature. People working on vibration platforms do not study cats; people studying cats do not follow the vibration platform literature. There is a disciplinary wall between the two fields.

The Disciplinary Wall

How thick is this wall. A simple literature search experiment will tell you. Search "cat purr bone" on PubMed and the number of relevant results can be counted on one hand. Search "whole body vibration bone density" and there are hundreds of results. The two searches point to the same underlying physical mechanism; the literature volume differs by two orders of magnitude. This is not because the evidence for the relationship between cat purring and bone is insufficient to support research (the experiential evidence has always been there); it is because no one has ever built a stable funding channel between these two keywords.

If someday someone gets the money and does a comparison of fracture healing in purring versus non-purring cats (perhaps by utilizing naturally non-purring individuals, though such cats are extremely rare, or by utilizing post-surgical fracture healing data from cats that have temporarily lost the ability to purr after laryngeal surgery as a natural experiment), and the result is positive, it would have direct value for parameter optimization of vibration platform treatment protocols. Cats may have already completed tens of millions of years of parameter searching for human orthopedics, and humans have not yet gone to look at the results.