Almost every pheromone perfume on the market cites one of seven papers. The brands quote the title, paste the journal name, and hope you stop reading there. Most people do.
Here is the version where you keep reading. For each paper: full citation, what the researchers actually measured, what they found, the methodology caveats that get left off the marketing page, and what the result honestly lets you conclude about a $40 spray bottle.
These are all real papers. You can search any of them in Google Scholar in under thirty seconds. The exercise is not whether the studies exist. It is whether they support the claim a brand is hanging on them.
1. McClintock 1971 — the menstrual synchrony paper
Citation: McClintock, M. K. (1971). Menstrual synchrony and suppression. Nature 229: 244-245.
What it measured: a 23-year-old graduate student at Wellesley College tracked self-reported menstrual cycle onset dates among 135 women living in the same dormitory across an academic year, looking for convergence between roommates and close friends.
Result: cycle start dates of close friends and roommates appeared to drift closer together over the course of the year. McClintock argued this implied an airborne chemical signal between women. The paper became the foundational document for the entire 'human pheromones' field.
Caveats: self-reported dates with unclear precision, a small and demographically narrow sample, and a statistical design that later analysts showed will produce apparent 'synchrony' from random data simply because cycles vary in length. Replication attempts by Wilson (1992) and Strassmann (1997) failed to find the effect, and Strassmann in particular argued the original finding was a statistical artifact of the analysis method, not a biological signal.
What it lets you conclude: that the idea of human chemical communication is at least worth investigating. What it does not let you conclude: that menstrual synchrony is a real biological phenomenon, or that any compound in any cologne can drive it.
2. Cutler, Friedmann & McCoy 1998 — the Athena Institute paper
Citation: Cutler, W. B., Friedmann, E., & McCoy, N. L. (1998). Pheromonal influences on sociosexual behavior in men. Archives of Sexual Behavior 27(1): 1-13.
What it measured: 38 heterosexual men aged 26-42 were randomized to apply either an axillary extract (a proprietary blend of compounds collected from male underarm secretions) or an ethanol placebo to their aftershave each morning for six weeks. They kept daily logs of 'sociosexual' behaviors with women: petting, kissing, sleeping next to a partner, formal dates, sexual intercourse, and masturbation.
Result: men in the treatment group reported significant increases in three of the six measured behaviors (petting/kissing/affection, sleeping next to a partner, and sexual intercourse) compared to the placebo group. They did not show increased rates of masturbation or 'informal dates' that did not involve a partner, which the authors framed as evidence that the effect was relational rather than libido-driven.
Caveats: the sample was tiny (n=38), the study was conducted and funded by the founder of the company (Athena Institute) that sells the resulting product, the outcome measure was self-reported behavior with all the demand-characteristic problems that implies, and the result has never been independently replicated at scale by a research group with no commercial stake. Every pheromone brand citing 'a peer-reviewed study showed a significant increase in intimate behavior' is, almost without exception, citing this one.
What it lets you conclude: that one small commercial study found a behavioral effect from a specific axillary extract worn daily for six weeks. What it does not let you conclude: that the effect generalizes to any other product, formulation, or wear pattern, or that it survives independent replication. Full breakdown in our Athena Pheromones review .
3. Wyart et al. 2007 — androstadienone and cortisol
Citation: Wyart, C., Webster, W. W., Chen, J. H., Wilson, S. R., McClary, A., Khan, R. M., & Sobel, N. (2007). Smelling a single component of male sweat alters levels of cortisol in women. Journal of Neuroscience 27(6): 1261-1265.
What it measured: 21 heterosexual women were exposed to androstadienone (a steroid present in male sweat and saliva) or a control odor, with saliva samples collected at intervals to track cortisol, and a subset scanned by fMRI to measure brain response. This is a real physiology experiment with hard endpoints, not a behavioral self-report.
Result: exposure to androstadienone produced measurable, sustained elevations in cortisol relative to control, alongside altered activity in brain regions associated with social-emotional processing. It is the cleanest single piece of mechanism evidence in the human pheromone literature: a specific molecule producing a specific, measurable physiological change in a specific brain network.
Caveats: the androstadienone concentrations used were substantially higher than what is present in natural sweat, and dramatically higher than the trace amounts in a commercial cologne spray. The study also measured a physiological reaction, not attraction, not behavior, not 'effect on the person standing next to you'.
What it lets you conclude: androstadienone at the right concentration can produce measurable changes in women's hormone levels and brain activity. What it does not let you conclude: that a cologne containing trace androstadienone reproduces the effect on a date.
4. Saxton et al. 2008 — androstadienone at the speed-date
Citation: Saxton, T. K., Lyndon, A., Little, A. C., & Roberts, S. C. (2008). Evidence that androstadienone, a putative human chemosignal, modulates women's attributions of men's attractiveness. Hormones and Behavior 54(5): 597-601.
What it measured: women attending real speed-dating events were assigned to wear an upper-lip swab of androstadienone in clove oil, clove oil alone, or pure water. After each three-minute conversation with a male partner, they rated him for attractiveness. The study moved the research out of the fMRI scanner and into something close to a real social setting.
Result: women in the androstadienone condition rated the men they spoke with as significantly more attractive than women in either control condition. It is the strongest single behavioral data point linking androstadienone to actual social judgement in something approximating real life.
Caveats: the effect size was modest, the design was a single evening rather than a sustained social context, the sample size was on the smaller side, and the manipulation was applied to the rater rather than the rated person — meaning the result is about how androstadienone changes a woman's perception, not about a man wearing it being judged more attractive.
What it lets you conclude: in one real-world social setting, a putative pheromone shifted women's attractiveness ratings of conversational partners. What it does not let you conclude: that the same effect generalizes to a man wearing a commercial cologne.
5. Hare et al. 2017 — the replication that broke the spell
Citation: Hare, R. M., Schlatter, S., Rhodes, G., & Simmons, L. W. (2017). Putative sex-specific human pheromones do not affect gender perception, attractiveness ratings or unfaithfulness judgements of opposite sex faces. Royal Society Open Science 4: 160831.
What it measured: a direct attempt to replicate the kind of attractiveness-rating effect Saxton 2008 had reported. The authors pre-registered their hypotheses, recruited 94 heterosexual participants, exposed them to androstadienone, estratetraenol (the putative female counterpart), or a control, and asked them to rate opposite-sex faces on gender, attractiveness, and likely unfaithfulness across two test sessions.
Result: nothing. No significant effect of either compound on any of the three outcome measures. The authors used larger and better-controlled samples than the original positive studies, pre-registered their analyses to prevent post-hoc cherry-picking, and reported a clean null.
Caveats: this is a single replication, and faces on a screen are not the same stimulus as a person across the table at a speed date. But pre-registered, well-powered nulls are exactly the kind of evidence the field most needed, and almost never gets cited by brands selling product.
What it lets you conclude: the headline attractiveness-modulation results for androstadienone and estratetraenol may not be as robust as a decade of citation made them look. What it does not let you conclude: that no human chemosignals do anything, or that the cortisol-level mechanism in Wyart 2007 is invalidated.
6. Wyatt 2015 — the honest reset
Citation: Wyatt, T. D. (2015). The search for human pheromones: the lost decades and the necessity of returning to first principles. Proceedings of the Royal Society B 282: 20142994.
What it is: not an experiment, a review and critique. Tristram Wyatt, an Oxford chemical-ecology researcher and the author of the standard academic textbook on pheromones, argues that the entire human pheromone field has spent forty years chasing a handful of molecules (androstadienone and estratetraenol most notably) that were never rigorously identified as pheromones to begin with — they were guessed, popularised, and then re-cited until they sounded established.
His argument: a real pheromone, by the Karlson-Lüscher 1959 definition the rest of biology uses, has to be a chemical signal that triggers a specific, repeatable response in a member of the same species. The bar is high on purpose. For humans, only one candidate (a nipple-area secretion that elicits suckling in newborns) currently clears it. Everything else, including the molecules every pheromone cologne is built on, is provisional at best.
Caveats: a review paper is an argument, not a measurement. Other researchers in the field disagree, particularly on whether the standard is too strict for a primate as behaviourally complex as humans.
What it lets you conclude: that the most respected skeptic in the field thinks the evidence base is much thinner than the marketing implies, and that the field needs to start over with proper bioassay-guided identification. It is the single most important paper for understanding where the science actually stands.
7. Liman & Innan 2003 — the missing receptor
Citation: Liman, E. R., & Innan, H. (2003). Relaxed selective pressure on an essential component of pheromone transduction in primate evolution. PNAS 100(6): 3328-3332.
What it measured: a comparative genetics analysis of the TRPC2 gene across primate species. In mice and most other mammals, TRPC2 codes for an ion channel that is essential for signal transduction in the vomeronasal organ — the dedicated 'pheromone-sensing' organ in the nasal cavity. The researchers asked whether the gene is intact and functional in humans and our close evolutionary relatives.
Result: TRPC2 is pseudogenized — broken by accumulated mutations and no longer functional — in humans and all other catarrhine primates (Old World monkeys, apes). The breakage looks ancient and the gene has been under relaxed selective pressure for tens of millions of years, meaning evolution has not been maintaining it because it has not been doing anything.
Caveats: a broken VNO transduction gene does not prove humans have no chemical communication. Plenty of mammalian chemical signalling routes through the main olfactory epithelium rather than the VNO. But it does strongly suggest that whatever humans are doing chemically, we are not doing it with the dedicated pheromone hardware other mammals use.
What it lets you conclude: humans almost certainly lack the dedicated pheromone-detection system the rest of the mammal lineage uses. What it does not let you conclude: that chemical communication between humans is impossible. Longer treatment in our piece on the VNO debate .
The honest summary
Read all seven papers together and the picture is not 'pheromones are fake' and not 'pheromones obviously work'. It is something more interesting and more frustrating: a small handful of molecules, including androstadienone , androstenone , and the vaginal-secretion copulins , produce measurable physiological and brain-imaging effects in tightly controlled laboratory exposures.
The translation from lab to real social settings is contested. The flagship attractiveness study (Saxton 2008) has been directly challenged by a pre-registered null (Hare 2017). The original menstrual-synchrony finding (McClintock 1971) has repeatedly failed to replicate. The most-cited behavioral effect (Cutler 1998) was run by the company selling the product and has never been independently reproduced at scale. Humans do not have the dedicated VNO hardware (Liman 2003) that other mammals use for pheromone detection.
So the pheromone-cologne category is built on a contested but not zero evidence base. Real lab signal. Soft and mixed translation to real-world social settings. Wyatt 2015 is right that the field needs better methodology, and the brands selling product are wrong to imply the science is settled. If you are still deciding whether one of these bottles is worth your $40, the longer answer lives in our do pheromone perfumes work breakdown, with placebo and confidence effects treated honestly alongside the chemistry.
No comments yet