Blog
Last week the Wall Street Journal published an article on a topic I’m quite familiar with – pheromones!
I thought I’d write a quick blog to give more context to the article for anyone curious about what the research actually says in this space.
The WSJ article has a paywall, so in case you can’t see the whole thing or would like a summary, here’s the gist.
Young people are pushing back against digital dating through apps. Part of that involves investing in how they smell to maximize the quality of in-person encounters. In particular, interest in fragrances involving pheromones has been growing.
Over the past two years, the number of people talking about pheromone perfume on social media surged nearly 300%, according to social media analytics company Sprout Social.
The article mentions a handful of products like this. One example, Pure instinct, is a “pheromone-infused” roll-on fragrance you can get for under $15 on Amazon. The author cites testimony from a young woman who says it works amazingly for her on dates.
Another product mentioned in the article is Kakou, a fragrance marketed as “pheromone perfume”, also available for about $15. A different woman interviewed for the article says that it causes men to be attracted to her to the point of it making her uncomfortable. Quite an endorsement.
On the other hand, a scientist interviewed for the article, British evolutionary biologist Tristam Wyatt, says that despite these testimonials there’s “no credible study” identifying chemicals that act as human sex pheromones.
So what gives? Is it all nonsense, or are consumers ahead of the curve on something that researchers don’t yet fully understand?
Fortunately, the totality of research on human pheromones is pretty clear: there aren’t any.
A pheromone is a molecule that an organism produces to trigger a development process (e.g. puberty in mice) or a predictable behavior (i.e. ants swarming in response to a threat) in another organism of the same species.
The basic idea is that pheromones are “external” hormones, or hormones that leave the body.
All plants and animals have hormones that communicate messages between tissues inside the body. For example, your adrenal glands attached to your kidneys make adrenaline that goes up to your brain and causes you not to fall asleep as you read this blog (hopefully).
When a molecule is used to communicate between organisms rather than within them, that’s a pheromone.
Biologists have known about pheromones since 1959, when a couple of European scientists found that Silkworms triggered mating behavior using a chemical called Bombykol. Since then, thousands more have been discovered across a wide variety of species.
The big question then is whether humans have pheromones of our own. The short answer is: possibly, but nobody has found any yet despite a decades-long search.
In 1971, University of Chicago psychologist Martha McClintock published a study showing that the menstrual cycles of women living together tended to synchronize over time, suggesting a possible pheromone-mediated effect. That study is often credited with starting the human pheromone phenomenon. Unfortunately, later studies found conflicting results.
Today, it’s understood that menstrual synchrony, the strongest and most long-standing candidate for a human pheromone-based effect, doesn’t really happen in humans. It seems to be just a statistical and psychological artifact. Check out this 2014 paper by UPenn psychology professor Richard Doty for a deeper analysis. Tristam Wyatt, the same researcher mentioned in the WSJ article, comes to the same conclusion in his 2020 review.
Since the early 1970s, a handful of other molecules have taken their turns in the research spotlight. For this blog, I’m just going to look at the ones related to human mating behavior since those have obvious implications for fragrance products.
That shortlist of human sex pheromone candidates includes: androstenol, androstenone, androstadienone, estratetraenol, and copulins.
I’ll walk through each of them in more detail below, but I can say one thing off the bat – none technically qualify as pheromones.
A few general reasons:
See the same 2014 review by Richard Doty for an elaboration on the problems here.
That said, it’s possible for a molecule to still have interesting psychological or behavioral effects without formally qualifying as a pheromone.
Even looking at it that way, there’s not much exciting news to report.
These two volatile steroid molecules were first identified in boars in the 1940s, and since then found use as a breeding tool in pig farming (still today). Their presence in human sweat wasn't confirmed until the late 1960s.
As it turns out, they’re basically just pig-pheromones that don’t do anything meaningful for humans.
There are dozens of studies going back to the 1970s on their psychological and behavioral effects in humans. Unfortunately most of them find only small, context-sensitive effects. Few have been replicated.
Here are some examples with notes on date, study size, and key outcome.
On androstenone:
On androstenol:
Note that all the problems of the replication crisis are potentially in play here: small sample sizes, p-value hacking, publication bias, HARKing, etc. I’m not accusing any specific authors or journals, just highlighting that methodological issues have plagued the field of behavioral psychology and that these results should all be viewed somewhat skeptically.
With that in mind, it’s tough to look at this body of research and conclude that there’s any reliable, useful effect from either of these molecules. For every exciting result, e.g. Filsinger et al. (1985), there’s a contradicting one, e.g. Black and Biron (1982), or just an obvious lack of studies replicating the exciting results. Even if you take the individual studies at face value, you have to consider context dependence to decide if the effects are worth caring about. For example, Gustavson et al. (1987) finds a gender-specific behavioral effect on bathroom stall choice mediated by androstenol presence. Assuming we don’t really care about bathroom stall choice, we’re left to wonder what this tells us about androsterone in other contexts. Perhaps the study’s outcome is actually bathroom specific.
On a pragmatic note, these molecules generally smell pretty bad. A large fraction of the population can’t smell androstenone at all, possibly up to 37%, although research on this has turned up a variety of estimates ranging from <5% to >50%. Qualitatively, people tend to describe the smell of androstenone and androstenol as “strongly urinous”, “musky,” and “unpleasant” – not super promising for artificial fragrance use.
If you come across products that use these androstenone or androstenol, you can safely assume they aren’t doing anything useful.
Androstadienone and estratetraenol, two more steroid molecules, rose to research prominence in the early 90s, after androstenone and androstenol,.
Although they had been identified in human sweat and urine previously, both molecules were first presented at a corporate-sponsored conference (Erox corp, now defunct) in 1991 as “putative sex hormones.” Erox also patented and sold their own pheromone-based fragrances (Realm by Erox, launched 1933)
Scientific interest intensified several years later when McClintock's lab – the same McClintock responsible for the menstrual synchrony research – published several influential papers suggesting that androstadienone could affect women's mood and hormones. McClintock’s lab was not affiliated with Erox, granting more credibility to her work than that of other researchers in this domain who had clear commercial interests.
For example, her paper from 2000 found various subtle mood effects from exposure to androstadienone. McClintock and her colleagues were careful at the time to hedge their findings, explaining that the discovery of psychological effects from these molecules did not amount to confirmation of their status as human pheromones, and that the effects they did find needed further study.
Since then, many studies have been conducted, further muddying the picture. Here’s the set of relatively credible work on the psychological and behavioral effects of these molecules I was able to dig up (I filtered for minimum 20 participants, established research groups, reputable journals, etc.).
On androstadienone:
On estratetraenol:
Fortunately there’s more recent research into these molecules than androstenone and androstenol.
Unfortunately, we’re still looking at a bit of a mixed-bag of conclusions.
Androstadienone seems to have some effects on aggression and social behavior, reducing aggression in men while increasing reactive aggression in women (Yin Wu, et al. 2023). It also seems to influence cooperation and individualistic behavior, increasing cooperation in some contexts (Huoviala et al. 2013) but promoting individualistic responses in others (Banner et al. 2018). It may have an effect on mood and emotional processing too. In women, androstadienone can reduce negative mood states such as nervousness and tension (Grosser et al. 2000) and enhance attention to emotional information, particularly angry faces, in both men and women (Hornung et al 2017 and Hummer et al 2009).
Estratetraenol has an effect of increasing men’s social cognition, particularly in contexts related to intimacy and romantic relationships, as well as increasing general cooperative tendencies (Oren et al. 2019a and Oren et al. 2019b). It also affects gender perception by biasing men to perceive figures as more feminine. This effect is not observed in women, indicating a sex-specific influence of estratetraenol on gender perception (Zhou et al 2014). Estratetraenol also seems to subconsciously bias heterosexual men to perceive female figures as happier and more relaxed, indicating its role in priming emotional receptivity towards potential mates (Ye et al. 2019).
Like with androstenone and androstenol, I don’t see any smoking gun result here that would suggest androstadienone or estratetraenol would have a big desirable effect in artificial fragrances. But who am I to tell people something doesn’t work? There’s enough evidence here for some effects that it doesn’t bug me that people want to research these molecules more or even try putting them into fragrances. I’m just a bit skeptical.
So what do androstadienone and estratetraenol actually smell like?
Androstadienone smells putrid in general, although some people do bizarrely find it pleasant. It’s thought that two odor receptors account for the human ability to detect androstadienone. One higher-affinity receptor is responsible for the putrid quality and another low-affinity receptor is responsible for other vegetable-like, woody, and fruity qualities people report. Upon repeated exposure, the lower-affinity receptor tends to activate less, causing even the people who found the odor pleasant initially to come to dislike it.
Interestingly, estratetraenol does not have a consciously detectable odor. That’s right – all the studies above reporting behavior or mood effects from estratetraenol exposure are claiming to do so for a molecule with no noticeable odor.
Copulin(s) isn’t actually a single molecule; it’s a blend. All the constituents are naturally occuring aliphatic acids.
In 1971, psychiatry and anatomy researcher Richard P. Michael discovered this combination of molecules, which is produced in nature by rhesus monkeys to signal sexual receptivity.
In 1972 he patented the use of that blend in a human fragrance, which includes: acetic acid (yes, the same thing as white vinegar), propionic acid, butyric acid, isobutyric acid, 2-methylbutyric acid, isocaproic acid, and isovaleric acid.
Most of these molecules have since been found to occur naturally in human body odor, although they fall under the umbrella of general body odor more than anything seemingly specific to individual genetics, attractiveness, status, etc.
In all my searching I only turned up two papers on the psychological effects of copulin exposure in humans, and they contradict each other!
Meanwhile, a clear problem with copulins from a fragrance formulation standpoint is all these molecules smell bad. Some people do find some of their smells pleasant in very low concentration, but the overall picture is that the blend will just make you smell unhygienic, which makes sense given the high overlap between copulin and the set of molecules known to contribute to general human body malodor (underarm, foot, etc.).
Here are some scent notes on each copulin molecule from OlfactionBase:
Scientists still haven’t discovered any definitive human pheromone molecules. Surprises may still happen with additional research, however today there’s nothing.
Although they aren’t technically pheromones, these molecules can probably influence human mood and behavior in some subtle ways. But then so do regular old plant-based essential oils.
If I had to guess, the positive anecdotes reported for currently available “pheromone” products exist for two main reasons:
Today’s pheromone products are also a bit of a minefield for other reasons. Browsing on Amazon through a dozen or so pheromone products, only about half name an ingredient like one of the ones I talked about above. None of them mention the quantity or concentration of that ingredient in the final blend. You’re paying for a product that says it works using pheromones. Whether or not it really works, those pheromones at least better be in there.
These inexpensive pheromone products also tend to be secretive about where they’re manufactured – many are drop-shipped, like Kakou. They also don’t disclose the other ingredients they might contain, like parabens, phthalates, or other preservatives. It’s not a great sign if the fragrance’s preservatives are the only ingredients demonstrated to affect human hormone levels.
Given the lack of scientific rationale and shady marketing with these “pheromone” fragrances, I would suggest you just go out and find a traditional perfume you like. But if you’re feeling adventurous, then at least now you have all the info!