Walk past a perfume counter and you will eventually find a bottle promising something close to magic. Spritz this on, the marketing suggests, and strangers will turn their heads. The pitch is built on a single word: pheromones. The chemistry of attraction, captured in a fragrance, ready to deploy.

It is a beautiful story. It is also, on the most generous reading of the evidence, almost entirely unsupported by research.

That does not mean smell is irrelevant to attraction, or that humans are the dull-nosed creatures we have been told we are. The real picture, drawn from decades of olfactory science, is more nuanced and far more interesting than any perfume ad. Humans are surprisingly capable smellers. Body odor genuinely shapes who we find pleasant. And the search for a true human pheromone — one specific molecule that reliably triggers a behavioral response — has produced, after roughly half a century of effort, exactly zero confirmed candidates.

This article walks through what we actually know: why the "poor human nose" idea was a 19th-century mistake, what your olfactory system is doing under the hood, why the marketed "pheromone perfumes" rest on disputed and probably extinct claims, and what the science of body odor really tells us about attraction, immunity, and disease.

The myth of the bad human nose

For more than a century, biology textbooks treated humans as microsmatic — animals with a poor sense of smell. The label felt obvious. We do not sniff lampposts. Dogs find drugs we cannot detect. Compared to a wolf, our nose seems almost decorative.

The story has a specific origin: the 19th-century French anatomist Paul Broca, who noticed that humans have proportionally smaller olfactory bulbs than many other mammals and concluded, somewhat creatively, that this was the price of our enlarged frontal lobes. Sigmund Freud later picked up the same theme and ran with it, arguing that the suppression of smell was central to becoming a rational, civilized animal. Neither claim was based on actual experiments measuring how well humans detect odors.

The modern reassessment came in a 2017 review by neuroscientist John P. McGann, Poor human olfaction is a 19th-century myth, published in Science. McGann pointed out that olfactory bulb size is a poor predictor of olfactory ability. What seems to matter is the number of neurons in the bulb — and on that measure, the human olfactory bulb sits comfortably alongside other mammals', with totals in the same ballpark whether you weigh 15 grams or 70 kilograms.

In behavioral tests, humans turn out to be more sensitive than rodents and even some primates to certain odors — and less sensitive to others. Smell, like vision, is tuned. Different species are good at detecting different things. The notion that all of human olfaction is degraded is simply wrong.

The most striking demonstration of this came in a 2014 Science paper by Bushdid and colleagues claiming that humans can discriminate more than one trillion olfactory stimuli. The number itself is now contested — a later analysis showed the math could yield wildly different estimates — but the headline that survived methodological scrutiny is the more important one: humans are clearly capable of distinguishing an enormous number of smells, far more than the textbook figure of 10,000 anyone learned in school.

What your olfactory system is built from

The machinery underneath all of this is genuinely impressive. In the early 1990s, Linda Buck and Richard Axel — work that later won them the Nobel Prize — discovered the family of genes that code for olfactory receptors. The current count, after the human genome was sequenced and refined, is roughly 400 functional olfactory receptor genes in humans, with several hundred more existing as inactive pseudogenes. Mice, by comparison, have around 1,400 functional ORs.

Before you decide that mice obviously win, remember the lesson from McGann's review: gene count is not the same as perceptual capacity. The 400 receptors humans express interact combinatorially. A given odor molecule typically activates several receptors at once, and the brain reads the pattern of activation rather than a single key-and-lock match. That combinatorial code is what allows a relatively small library of receptors to distinguish vast numbers of compounds.

There are a couple of features of the human olfactory system that genuinely differ from many other mammals. One is the absence of meaningful olfactory bulb neurogenesis in adults — humans do not seem to add new neurons to this brain region the way rodents do. Another, more relevant to the pheromone debate, is the vomeronasal organ.

The vomeronasal organ: a working part in mice, a vestige in us

Many mammals have an accessory olfactory system, anatomically distinct from the main one. Its sensory entry point is the vomeronasal organ (VNO), a tiny structure inside the nasal cavity. In rodents and many other mammals, the VNO is essential for detecting low-volatility chemical signals — including most known mammalian pheromones — and it sends its output to a separate accessory olfactory bulb in the brain.

In humans, the situation is very different. Anatomical studies in adults find no functional sensory neurons in the VNO and no accessory olfactory bulb. The molecular evidence is even more decisive: the gene TRPC2, which encodes a critical ion channel for VNO signaling in mice, is a pseudogene in humans, riddled with stop codons that prevent it from making a working protein. Most vomeronasal receptor genes have suffered similar fates. Whatever role the VNO plays in catarrhine primates, it has been effectively dismantled in our lineage.

This matters because the marketing story behind "pheromone perfumes" implicitly relies on a working accessory system. The idea is that there is some hidden chemical channel — separate from conscious smell — by which subliminal signals trigger behavior. The anatomy and genetics say no such channel exists in adult humans. Whatever signaling we do via odor goes through the regular olfactory system, the same one that lets you tell coffee from cinnamon.

What is a pheromone, exactly?

The term was coined in 1959 by Peter Karlson and Martin Lüscher in a short Nature paper titled "'Pheromones': a New Term for a Class of Biologically Active Substances." Their definition was strict and is still the one biologists use: a pheromone is a chemical, or a defined mixture of chemicals, released by one member of a species and triggering a specific behavioral or physiological response in another member of the same species.

Two features of that definition matter. First, a pheromone is a particular molecule (or fixed blend), not a vague "chemistry." Second, the response has to be reproducible. Spray it, see the behavior. Without the molecule and the bioassay, you do not have a pheromone — you have an interesting hypothesis.

By this standard, plenty of animal pheromones are well documented. The textbook example is bombykol, identified in the female silk moth: a single molecule that, when released into the air, reliably draws males from astonishing distances. In mammals, androstenone — a steroid in boar saliva — pushes sows in heat into the lordosis mating posture so reliably that it is sold commercially under the name Boarmate to assist artificial insemination. (As an aside, the same compound is one of the components that gives truffles their distinctive smell, which is part of why pigs are good at finding them.)

By this same standard, no human pheromone has ever been confirmed.

The decades-long dead end of "human pheromones"

The most thorough takedown of the human-pheromone literature is a 2015 review by Oxford zoologist Tristram Wyatt, The search for human pheromones: the lost decades and the necessity of returning to first principles, published in Proceedings of the Royal Society B. Wyatt's argument is straightforward and uncomfortable: the four molecules most frequently cited in the popular and scientific press as "human pheromones" — androstenone, androstenol, androstadienone, and estratetraenol — were never actually identified as such by the methods used for every other species.

Two of those four trace back to a 1991 conference presentation by researchers associated with the EROX Corporation, a company with a direct commercial interest in selling pheromone-branded products. The presentation supplied no information about how the molecules were extracted from human secretions, no rigorous bioassay establishing a specific behavioral effect, and a single footnote acknowledging that the compounds had been provided by EROX itself. Yet hundreds of subsequent papers cited those compounds as established human pheromones, and a whole industry of "pheromone perfumes" grew up on the back of the citation chain.

Individual studies on, say, androstadienone activating sexually dimorphic brain regions have been published, but Wyatt and others note that almost none meet the standard required to call something a pheromone: small samples, inconsistent effect sizes, no replication of a specific behavioral response, and no clean chemical identification from a real human secretion. The most promising current candidate is something quite different — a compound or compounds in the secretion around the areolae of breastfeeding mothers that triggers a sucking and head-orientation response in newborns. Even there, the specific molecule has not yet been isolated and synthesized.

The honest conclusion is that pheromone perfumes, as currently sold, are perfume — sometimes pleasant, often expensive, and lacking any plausible mechanism for the effects their marketing promises.

What body odor does signal

Dismissing pheromone perfume is not the same as dismissing the role of smell in human social life. Quite the opposite — odor is doing real work, just not the kind that fits in a spray bottle.

The most famous study in this space is the "sweaty T-shirt" experiment by Claus Wedekind and colleagues, published in Proceedings of the Royal Society B in 1995. The setup was simple: men wore the same T-shirt for two nights using only unscented soap and detergent. Women, blind to the wearers' identity, then sniffed the shirts and rated their pleasantness. Wedekind had typed both groups for genes in the major histocompatibility complex (MHC), a cluster of immune-system genes whose enormous diversity helps the body recognize a wide range of pathogens.

Women with natural cycles tended to prefer the smell of men whose MHC profile was most different from their own. Women using oral contraceptives showed the opposite preference, leaning toward MHC similarity. A larger 1997 follow-up by Wedekind and Füri extended the result, with both men and women showing a tilt toward MHC dissimilarity in their pleasantness ratings.

The evolutionary logic is appealing. Children of MHC-dissimilar parents inherit a more diverse immune-gene repertoire, which broadens the range of pathogens they can recognize. MHC-dissimilarity is also crudely correlated with low genetic relatedness, so a preference for it could double as inbreeding avoidance. The reversal under hormonal contraception fits a story in which the pill mimics pregnancy and shifts preferences toward kin-like, "safe" social signals.

The MHC story is real, but messier than the headline

In the years since Wedekind's papers, MHC-and-body-odor results have become a textbook case of the replication challenge in human behavioral biology. A large 2018 analysis found that several attempts to reproduce the effect in male raters of female odors did not work, and a broader review of real-world couples found that partners are often more MHC-similar than chance, the opposite of what the dissimilarity hypothesis would predict at face value.

Why the inconsistency? Some of it is statistical noise around small effects in small samples. Some is methodological — the original studies used a handful of T-shirts, varied across studies whose nose was doing the rating, and could not control for diet, soap, perfume, smoking, and the rest of the modern olfactory landscape. And some is population structure: in ethnically heterogeneous societies, partners often share ancestry, which inflates apparent MHC similarity for reasons that have nothing to do with smell preference.

The related claim that women smell more attractive around ovulation has had a similarly bumpy ride. Early work suggested male raters preferred the body odor of women in their fertile window, but recent careful studies have struggled to replicate this. A 2024 study combining perceptual ratings with chemical analysis found no compelling evidence for ovulatory cycle shifts in women's axillary odor, although a follow-up the next year did identify some compounds that varied with cycle phase.

The takeaway is not "it's all fake." The takeaway is that smell-based signals in humans are real but subtle, easily swamped by hygiene products, diet, and context, and not strong enough to override the rest of how people choose partners. Body odor is a whisper. Marketing wants to sell you a shout.

What your nose can tell you about your health

If attraction is the romantic application of olfaction, disease detection is its most practical one. The most arresting case is Joy Milne, a retired Scottish nurse who noticed a distinctive musky odor on her husband years before he was diagnosed with Parkinson's disease. When researchers later tested her by having her smell shirts worn by people with and without the diagnosis, she identified the patients with near-perfect accuracy — including one apparent false positive who was diagnosed within the year.

Milne's hyperosmia is unusual, but the underlying chemistry is real. Researchers using mass spectrometry have since identified volatile compounds in skin sebum that differ between people with and without Parkinson's disease, opening the door to a possible early diagnostic test.

More broadly, smell is one of the earliest casualties of several neurodegenerative diseases. Multiple long-running cohort studies show that olfactory dysfunction precedes the cognitive symptoms of Alzheimer's disease by years and predicts later cognitive decline and dementia. This is part of why smell tests are increasingly used in research settings as one signal — not a diagnostic on their own — that something may be changing in the brain before more obvious symptoms appear.

COVID-19 dragged smell loss into the public conversation in a way it had never been before. At the peak of the pandemic, roughly half to two-thirds of infected patients reported some loss of smell or taste, with formal psychophysical testing finding even higher rates than self-report. Most people recover within weeks, but a meaningful minority do not. A two-year follow-up study of patients with persistent post-COVID smell loss found 38 percent reported full recovery, 54 percent partial, and 7.5 percent no recovery at all.

Smell training: rehab for the nose

The most evidence-supported intervention for post-viral smell loss is, somewhat counterintuitively, an exercise program. Olfactory training involves deliberately sniffing a small set of distinctive odors — classically rose, eucalyptus, lemon, and clove — twice daily for several months while consciously trying to recall what they used to smell like. The protocol was developed before COVID, but the pandemic produced a wave of new evidence.

A 2024 systematic review of olfactory training for chronic post-COVID smell loss concluded that the intervention provides meaningful benefit for many patients, particularly when sustained for at least 12 weeks and ideally longer. Effect sizes are modest and recovery is partial for many people, but compared to no treatment — which is essentially the alternative on offer — olfactory training is the standard of care.

This works in part because the olfactory system retains real plasticity. Even though humans do not generate new neurons in the olfactory bulb the way rodents do, the higher cortical regions that interpret smell signals are remarkably trainable. Wine professionals, perfumers, and tea tasters are not anatomically different from the rest of us; they have spent thousands of hours pairing odors with labels, and their brains have built denser representational maps as a result.

If you have lost smell after a viral infection, head injury, or sinus inflammation, the evidence supports doing structured smell training rather than waiting passively. And if you notice a sudden change in your sense of smell that doesn't recover, it is worth flagging to a clinician — particularly if it accompanies other neurological symptoms — given olfaction's status as an early signal in several neurodegenerative diseases.

Why the pheromone perfume industry won't die

Given the state of the evidence, why does "pheromone" still sell? Several reasons converge.

The word itself is irresistible. It sounds like a real biological lever — secret, biochemical, beneath conscious awareness. That is exactly the kind of mechanism people want to believe in, especially around dating. A perfume that might tip the balance in your favor is a hard product to argue against, and the placebo loop is generous: spritz, feel more confident, behave more confidently, get better outcomes. The chemistry need not have done anything.

The regulatory environment is also permissive. Cosmetic claims are weakly policed compared to drug claims. As long as marketers do not explicitly promise medical effects, they can lean on suggestive language — "attraction-enhancing," "chemistry," "primal" — without ever proving a mechanism. And because there is no agreed standard for what a "pheromone product" must contain, the label has no scientific meaning.

Finally, the original literature gave the industry decades of citations to point to. Even after Wyatt's 2015 review, the cycle of underpowered studies on androstadienone or estratetraenol continues, and each one supplies another link in the marketing chain. The honest scientific position remains the one Wyatt argued: if a real human pheromone exists, it has not been identified, and the molecules currently sold under that name are not it.

What this means in practice

For everyday life, the practical implications are surprisingly clear.

If you are deciding whether to buy a pheromone perfume for romantic effect, the answer is: don't, unless you simply enjoy how it smells. There is no scientifically defensible reason to expect the labeled molecule to do anything beyond what any pleasant fragrance does.

If you are interested in how your own body odor influences how others perceive you, the boring news is that hygiene, diet, sleep, stress, and overall health probably matter more than any specific molecule. Diet has measurable effects on body odor, including the well-documented effect of eating large amounts of meat shifting axillary odor in ways some raters find less pleasant. Sleep deprivation and high stress shift the volatile compounds your skin emits. None of this is a substitute for showering, but it does explain why the same person smells different on different days.

If you have noticed a change in your sense of smell, take it seriously. Smell loss after viral infection often recovers but benefits from active training. Sudden or gradual smell loss without an obvious cause — particularly if it appears alongside other neurological symptoms — is worth raising with a clinician.

This is one of the small areas where general health tracking can do useful work. In WatchMyHealth, the wellbeing tracker lets you log day-to-day symptoms over weeks and months, including changes in senses such as smell or taste alongside sleep, mood, and energy. That kind of running record is more useful than memory when you are trying to figure out whether something started two weeks ago or two months ago — exactly the question a clinician will ask first if you bring up persistent anosmia.

The bottom line

Humans are not bad smellers. We have around 400 functional olfactory receptors and a brain that uses their combined output to distinguish enormous numbers of odors. Smell shapes memory, mood, and social perception in ways that are now well documented.

What we do not have is evidence for human pheromones in the strict biological sense — specific molecules that reliably trigger specific behaviors in other people. The vomeronasal organ that detects most mammalian pheromones is non-functional in adult humans. The molecules sold as human pheromones in commercial products were never properly identified by the standards applied to every other species. After half a century of looking, the search has produced a few promising leads — most prominently in mother-infant signaling — but no confirmed human pheromone.

Where odor genuinely seems to influence human social life is at the level of body odor as a complex cue. MHC-related preferences are real but subtle and inconsistently replicated. Hormonal cycles probably nudge body odor in detectable but small ways. Disease states — including Parkinson's, certain cancers, and the early stages of Alzheimer's — leave chemical fingerprints that the right nose, or the right mass spectrometer, can sometimes detect.

The difference between the science and the marketing is the difference between a faint, context-dependent signal and a guaranteed effect in a bottle. The marketing is selling something the science says is not there. The science, in turn, is selling something quieter but more interesting: a sensory system more capable than we were taught, doing work we are only now beginning to understand.