John P. McGann of the Department of Psychology at Rutgers University has spent fourteen years looking at the olfactory (smelling) system of mammals, and has published a new paper in Science suggesting that what we think we know about our own sniffing ability, compared to the reputed Super Sniffers of dogs and rodents, is wrong. McGann suggests, to a fanfare of publicity in various venues, that we’re not that bad: in fact, he notes that humans may be just as good as dogs and mice at detecting odors.
His paper is referenced below, and access is free if you have the free and legal Unpaywall extension (get it!).
McGann suggests that the presumed olfactory inferiority of humans is based on observations of the relatively small size of human olfactory lobes in the brain—observations made by Broca and other early neuroanatomists, and promulgated by Freud, who said that the absence of an acute sense of smell in humans led to sexual repression, since sex was partly based on odors. After reviewing the historical evidence, McGann gives his biological evidence. Here are his main points:
- Although the relative size of olfactory lobes compared to brain size as a whole is smaller in humans than in rodents, the absolute size if the lobes is much larger.
This shows the relative sizes of human and mouse olfactory lobes; the mouse lobe is much larger relative to its whole brain:
But the absolute size of the human bulb, which McGann sees as one key to good olfaction, is much larger than that of the mouse:
- Further, the absolute number of neurons in mammalian olfactory bulbs is relatively constant (McGann sees that as an indicator of smelling ability as well).
This graph shows that the number of neurons in the olfactory regions varies among mammal species by less than a factor of ten, regardless of the much larger variation in body size—and humans aren’t particularly low:
- Data showing a larger number of inactivated olfactory genes in humans compared to dogs and mice is questionable.
In Why Evolution is True, I noted that many human “olfactory receptor genes”, each binding to a separate molecule and allowing us to smell it, have been inactivated by mutations: they’re dead “pseudogenes”. As McGann reports, humans have 1000 such genes, but “only” (his quotes) 390 of them code for receptor proteins. In contrast, mice have 1300 such genes, of which 1000 code for proteins. This has been taken as evidence that mice can smell a lot more acutely than can humans. But McGann points out that 60% of the human pseudogenes are transcribed into messenger RNA in the nose.
But that’s weak evidence, as we already know that many pseudogenes are transcribed into RNA but not translated into proteins, so this says nothing about the number of OR proteins made in humans versus mice. Besides, even if every transcribed pseudogene in humans was made into an active, odor-receptive protein, there would still be 1000 active genes in mice versus about 750 in humans. Further, McGann didn’t do the comparison for pseudogenes in dogs or mice.
Here’s a table from a 2007 PLoS paper by Nimura and Nei showing those three species. If McGann wants to make the pseudiogene argument for human sniffing, he has to take into account the data below, and the possibility that many dog, mouse, and rat pseudogenes may also produce receptor proteins. Thus, I’m not convinced by his pseudogene evidence. (Note that every olfactory gene in cetaceans like whales is a pseudogene, suggesting they really aren’t used to smell, as those genes aren’t functional underwater.)
- The structure of the olfactory system differs between humans and other mammals.
These difference include a larger number of glomeruli (cluster of nerve receptors) in the human olfactory bulb compared to rodents, and the observation that the bits of human cortex used for processing olfaction is more “elaborate” (i.e., has more neural connections) than in other species. To me, this says little about the relative abilities of humans vs. other mammals to detect odors.
The most crucial question, however, involves not just neurons or pseudogenes but is this: “How many different odors can humans detect compared to, say, dogs and mice, and do we detect them at different thresholds?” It’s classically assumed that dogs and mice are better sniffers than are humans, but McGann said the data are not convincing, are based on anecdotes, and also rely on tests of molecules that humans wouldn’t have been selected during their evolution to smell anyway. Here are a few quotes (my emphasis if you want just the high spots):
Human olfaction is excellent and impactful
Historical and anatomical expectations aside, is the human olfactory sense actually impoverished? No, the human olfactory system is excellent, although it depends on the criteria employed. For instance, dogs may be better than humans at discriminating the urines on a fire hydrant and humans may be better than dogs at discriminating the odors of fine wine, but few such comparisons have actual experimental support. When properly tested, the primate olfactory system is highly sensitive to many odors and can exert strong influences on behavior, physiology, and emotions.
Humans with intact olfactory systems can detect virtually all volatile chemicals larger than an atom or two, to the point that it has been a matter of scientific interest to document the few odorants that some people cannot smell (i.e., specific anosmias). A prominent recent study calculated that we could also tell virtually all odors apart, with an estimated ability to discriminate more than 1 trillion potential compounds . Although this exact number is highly sensitive to the assumptions made, it is clear that the human olfactory system is excellent at odor discrimination, far better even than the putative 10,000 odors claimed by folk wisdom and poorly sourced introductory psychology textbooks.
One key insight in comparing the olfactory system of primates and other animals has been that different species have different sensitivities to different odorants. . . . A recent experiment tested olfactory thresholds for six sulfur-containing odors in mice, spider monkeys, and humans. Relative olfactory sensitivity varied with odorant : Humans were three orders of magnitude more sensitive than mice or monkeys to 3-mercapto-3-methylbuytl-formate, with all 12 human subjects outperforming all of the individual animals, yet all 12 humans were worse than all of the mice (and comparable to the spider monkeys) on 3-mercapto-3-methylbutan-3-ol. Overall, the humans were most sensitive to two of the six odorants, whereas the mice were most sensitive to four of the odorants. This finding complements older literature. . .
Human behavior is strongly influenced by olfaction. Environmental odors can prime specific memories and emotions, influence autonomic nervous system activation, shape perceptions of stress and affect, and prompt approach and avoidance behavior . Humans can follow outdoor scent trails and even exhibit dog-like casting behavior when trails change direction . The human olfactory system also plays a major, sometimes unconscious, role in communication between individuals. Each person produces a distinct odor that reflects not only dietary and environmental factors but also interacts with the immune system’s “self/non-self” histocompatibility markers to incorporate genetic information that permits the discrimination of kin from non-kin . The contents of this “body odor cocktail” are interpreted in parallel with environmental odors in the brain and can drive mate and food choice, as well as communicating information about anxiety and aggression in other people. We even appear to unconsciously smell our hands after shaking hands with strangers [JAC: Not me!], suggesting an unexpected olfactory component to this common social interaction. Although many of these olfactory experiences do not recruit attentional resources, they can be exceptionally salient in traumatic circumstances . When such circumstances result in posttraumatic stress disorder, olfactory hallucinations frequently become part of the symptomology.
Well. it’s clear from this (assuming McGann is right) that we need good comparative tests involving a variety of different molecules—not just molecules that would have been important in species’ evolutionary past—before we can say that the bloodhound is a better tracker by odor than is a human. But statements like “human behavior is strongly influenced by olfaction” says nothing about our relative sniffability compared to other species. This kind of writing is, I think, a bit tendentious.
The Guardian has a summary article about this paper, and shows that our own Matthew Cobb, who studies olfaction in flies for a living, thought the paper was great:
Matthew Cobb, professor of zoology at Manchester University, said the review had altered his own perspective on a study that he has focused on for much of his career. “We have this myth that humans can’t smell very much,” he said. “McGann’s exploring the actual evidence for that, which it turns out is fairly poor. It’s going to change my teaching next year.”
But others disagree:
Alexandra Horowitz, a scientist at Barnard College in New York, whose work focusses on canine olfaction, notes that while dogs track scents, find drugs and detect ovarian cancer in plasma samples, humans merely “notice if there is a bad smell on the train or someone has been cooking when we come home.”
“That there are olfactory specialists, such as perfumers or animal trackers indicates that with attention, we can get much better,” she added. “But not dog-level.”
While I’m not an expert like Matthew or Alexandra, my own take is that this paper is provocative and convinces me that more experiments need to be done, especially involving the ability of different species to detect different odors. Perhaps we’ve been too cavalier in our claims that humans are an auditory and visual species while dogs and mice depend more on olfaction. Old assumptions may be wrong, and McGann’s paper is useful for re-examining a famous set of such assumptions. He may be right, but I’m not convinced that his data show that. At best they show that we need to provisionally withdraw the common claim that we can’t smell as well as mice and dogs. (One thing I can say, though, is that humans smell better than dogs, especially wet dogs! I’ll be here all week, folks.)
__________
McGann, J. P. 2017. Poor human olfaction is a 19th-century myth. Science Vol. 356, Issue 6338,DOI: 10.1126/science.aam7263
“Alexandra Horowitz, a scientist at Barnard College in New York, whose work focusses on canine olfaction, notes that while dogs track scents, find drugs and detect ovarian cancer in plasma samples, humans merely “notice if there is a bad smell on the train or someone has been cooking when we come home.”
Ok but wouldn’t that be partly explained by the relative sizes of the olfactory parts of the brain? For dogs, smelling things makes up a larger part of their brain relative to their brain size and therefore it is a more important part of their behaviour, while humans are usually busy using their brains for many other things that a dog isn’t. The dogs that find drugs and such things are also specifically trained for this purpose.
There are people with extraordinary noses – & I do not mean Cyrano. Some of us however are less than acute when it comes to distinguishing scents/smells/perfumes. One item I read about this story was that we lack the descriptive language for the smells – which is why wine tasters come up with strange comparisons.
Why some people are connoisseurs, and others drink wine from a box.
Connoisseurs are over rated. I’d sooner trust my dog. If I had a dog. 😉
Yeah, connoisseurship is overrated. But there’re probably super-smellers, just as there are “supertasters.”
Or is that a bit too on la nez?
There must be significant differences in capability between humans as well. How many times I have heard the question – can’t you smell that? Ah, no. Sensitivity must vary and if that is due to simply awareness, at least that provides some idea without getting into detailed examinations. My father had no sense of smell which must be genetic, but I know of no others in the family.
The ability to detect different odours can surely be improved by training; they do it in sommelier schools or in the fragrance industry. On the other hand it seems that some people possess sniffing abilities that go far beyond what others can learn. There is possibly a genetic component in it.
Do sommeliers really achieve a finely tuned nose, or do they simply learn what descriptors are traditionally associated with various wines, and to speak with confidence?
Double-blind studies have shown that expertly trained musicians can’t actually discriminate between the sounds of multi-million dollar Cremonese violins and cheaper ones.
It seems to me that even in individuals there can be wide variation caused by . . ., hell, I don’t know. How tired I am, dehydrated, sick, other? But I know that sometimes I can’t smell hardly anything and other times I can accurately detect an ingredient that no one else in the vicinity can.
It must be the same for other animals. I’d guess that even bloodhounds have bad days.
Ditto.
I think people are missing the point that the claims are about AVERAGES, and yes, of course there’s variation within a species. But that’s totally irrelevant to whether there are palpable differences on average between the species.
I would swear that my sense of smell has become REALLY acute as I’ve aged – I’m 72 now! Weird, I know.
Yes, I noted I could smell the rabbits when jogging in the early morning, but only then.
However, as long as we use dogs and rats to smell out mines -and not humans, I hope I’ll be forgiven to be somewhat skeptical of this thesis.
Human olfactory sensitivity to some chemicals have been identified as genetically related. (For instance, whether you can smell the scent effect eating asparagus has one one’s urine (yes, I can).)
In my experience (mainly as a wine taster and food taster) there are huge differences in taste/smell ability in people. I think the literature supports this. (E.g. some people are “super-tasters”.)
It would be interesting to see side-by-side box plots of the abilities of humans and dogs, with large sample sizes. (You’d need to choose certain chemicals and have a box plot for each I would think. Or maybe you could somehow average or sum these chemical smell abilities? A simple “volume” assessment wouldn’t tell you much I think.
RE asparagus pee, that is a great example. Being one who can also smell it, it is nearly incomprehensible to me that there are people who can’t.
Is that true? Are there really a substantial amount of people that cannot smell that? (of course there are ‘anosmics’, but I guess you are not talking about them)
It is true. For example, this article relates a study involving Europeans in which about 60% of the subjects could not smell asparagus pee. Per the article, in studying the genomes of the subjects the researchers found hundreds of genetic variations that seem to be linked to the inability to detect the odor.
Anecdotally, I have run into more than one person who had no idea what I was talking about when I mentioned the asparagus pee odor in some context.
Now I feel like I’ve been missing out on something. I’d never heard of asparagus pee until today. How does one discover this?
Ahem… throughout the above item, Professor Ceiling Cat [Emeritus] has mis-typed the canine – it should read ‘d*g’ throughout! 😉
Have started sniffing lamp-posts……
If dogs olfactory senses were so great they wouldn’t sniff their own butts.
Matter of taste. 🙂
What is ‘good’ and ‘interesting’ will of course differ among species.
Seems to be an easy experiment to me, conceptually at least. Train humans to find drugs, cancer and explosives by smell and see what the success rates are compared to dogs.
Or perhaps train wine tasting experts to detect wine in sealed containers in luggage at busy airports and compare with dogs trained to detect drugs and explosives.
I’m perfectly willing to believe that the common tropes about how poor humans’ sense of smell is are not accurate*, but I am skeptical that our sense of smell is comparable to dogs.
More generally I’ve always been skeptical about tropes like more intelligent animals (for example humans) have relatively poor senses because so much of their brain capacity is necessary to produce that intelligence that there isn’t enough left for good senses too. Or that animals that have an exceptional sense, for example vision in raptors, means that some other sense must be substandard because of limited brain capacity. Cats are one example that contradicts that bit of folk wisdom, but then they are gods.
My experience with cats is that they have much better hearing than I do and mine is pretty good. And of course, hearing is something we tend to get tested for although not as much as eyesight. The cat can hear things before I have a clue.
If humans actually can, on average, detect odors as well as dogs, why are dogs required to discover concealed drugs that humans don’t detect?
The theory smells suspicious to me, too.
Seems obvious. Dogs evolved to be junkies. Probably why cats look down on them too.
Maybe it’s just not cost-effective to train humans for a specialty that narrow.
Can you imagine a human on his hands and knees crawling through the customs line sniffing bags?
After some your descriptions of TSA employees in action, I can imagine anything :-).
I can imagine a human being trained to focus more intensely on odors that are strong enough to detect, and therefore gaining the ability to differentiate more finely between similar odors. But trained to detect odors they simply couldn’t detect before?
I mean, I suppose I can’t say it’s impossible.
I can imagine gaining the ability to, say, smell cut grass from 50 yards and and identify the species. I can’t imagine gaining the ability to smell cut grass from several miles away.
Depends on what you mean by “couldn’t detect before”. If, as Jerry suggests, human sniffers crawled around on hands and knees and stuck their noses in every bag or every corner of a vehicle, they could probably detect quite a few things that they can’t detect standing up.
So the kind of training I’m envisioning is less about honing their discriminatory powers than it is about motivating them to actually get down and do what needs to be done.
Could such people perform on par with drug-sniffing dogs? I don’t know, but I expect they’d cost a lot more and complain a lot more about their working conditions.
Another possibility is that dogs are more reliable not because they’re better sniffers but because they don’t overthink it or second-guess themselves. (“I’m smelling something, but am I really sure it’s cocaine?”) They smell drugs, they give the signal, end of story.
I gave the example of smelling a mown lawn from several miles away because that’s the distance from which some dogs can detect a scent. Ok. I might detect a drug if got near it; I don’t know: is cocaine or heroin volatile, smell-wise? I certainly won’t detect a scent from miles away.
Things like forest fires, low tide, and landfills are pretty easy to smell from miles away. But of course the molecules we’re smelling aren’t miles away; they’re inside our nostrils.
So I think we need to clarify what we mean by detecting a scent “from miles away”. If the source is strong enough and the wind is in the right direction, you don’t need to be a dog to smell it. Or if some animal now miles away left a scent trail in the grass right here, you just have to know what you’re looking for and be willing to get down and put your nose in it.
Now maybe dogs are adapted to detect some molecules that we can’t detect (and vice versa). That’s a question of specificity, not sensitivity, and you’re right that no amount of training can overcome that.
Fair point. Some male dogs can smell a female in heat from 3+ miles away. That’s the kind of olfactory accuracy I’m skeptical humans have.
Well, smell can have a remarkable effect on humans. I think of all those times smell has triggered a memory for me. And, following Speaker to Animals above, we clearly experience smells differently than dogs. At the same time we’ve worked hard for the last century or so in making our environments scent-neutral. I’d like to see more studies.
This is pretty interesting, and it does make some good points that our sense of smell may be better than supposed. But it seems marred by some over-selling.
If true, it also seems true that we are not as keenly aware of our smelly world as much as some animals are. So perhaps our olfactory system does detect a lot more, but much is not passed on to our consciousness?
Mere anecdote, but when my wife was pregnant she would report a much greater awareness of smell. Since her last pregnancy, now many years ago, her sense of smell has remained much higher than it used to be. We both accept this, and it is now routine where she will react to something (usually not pleasant) that I cannot detect at all.
My wife also reported a greatly amplified sense of smell when pregnant. Both times. I have a terrible sense of smell. She will often complain about some terrible smell of which I am completely oblivious. I always attributed it to my having broken my nose a few times (one nostril barely functions at all), but maybe it’s genetic.
Margie Prophet made a compelling case that morning sickness -mainly effected through smell- in pregnant women was an evolutionary trait that prevented ingesting toxins that might harm the fetus.
BTW what has become or Margie Prophet?
I smell a rat.
AH, you have visited the white house lately?
Smell seems like our most primitive sense. Sometimes I’ll be on an elevator and a woman will get aboard wearing the same perfume as an old girlfriend. Even if (or maybe especially if) it’s a scent I haven’t smelled in years, the effect is powerful, like a smack up side the head.
Yes!
Some years ago I was host dad to an out-of-town student, 17 years old, living away from home for the first time. She had her own apartment, but it was my job to show her around town, take her shopping, feed her home-cooked meals, and generally look after her. We became very close, like a real father and daughter, for almost two years. Then she finished her training and moved away, leaving some of her stuff in storage at my place. It was one of the saddest days of my life.
A few months later, after she got settled in a new city, she asked me to ship her the clothes she’d left behind. So I packed it all into a big carton, and before sealing it up, pressed down on the stacks of clothes to squeeze out the excess air.
The smell that came out literally knocked me off my feet. Not a locker-room smell, the clothes were clean, but just that it smelled exactly like her, like she was suddenly right there in the room. It took me completely off guard and I had to sit down and cry for a few minutes before I could continue.
A moving story.
Yes. (Not to mention, much more platonic than the one told by Humbert Humbert, too. 🙂 )
There’s been some research about the close connection between the sense of smell and memory, but I haven’t read it in years and don’t remember any details. It is apparently fairly well established that there is a real phenomenon, due to brain architecture, that smells stimulate long buried memories more powerfully than other senses.
One at work, a coworker wore the men’s cologne Aramis, and I thought I was going to lose my mind. My first real boyfriend after college wore that cologne, and I simply cannot help but think of him (and his skills in bed) when I smell the scent.
My coworker had liberally doused himself with Aramis that day, and stood about 3 feet away from me and talked on for a good 15 minutes. I didn’t hear a word he said. But suddenly he seemed a lot better looking than what I had remembered. It was the most wonderful and bizarre 15 minutes I’ve ever had at work
And the coworker never wore Aramis again. I don’t know what happened, and I never had the nerve to ask him about it.
One thought I had has to do with the fact that dogs, rodents and other such animals are low to the ground while we are upright. Maybe humans can smell as well as other animals but we don’t receive the same types and levels of scents that animals low to the ground do. Dogs and others are probably more specialized in scents close to the ground like the scents of other animals.
Humans walk upright to put the maximum distance between our noses and our feet.
Clearly dogs have evolved to detect explosives whereas humans have evolved to detect expensive wine.
Yes, it seemed like such a silly suggestion. 🙂 If we’re better at discriminating between expensive wines, then that seems to imply that it (or similar smells) contributed significantly to the reproductive success of our ancestors, or perhaps that we just got “lucky” with genetic drift.
I’m quite certain (though just guessing) that a dog could be trained to discriminate between wine vintages better than a human could. It’s testable…
Another flaw that seems obvious to me in the comparison that claims humans are better than mice is that humans can be easily told to communicate when they smell something, but how are you going to tell that to mice? Perhaps the mice only reacted when the smell of sulfur became obnoxious, and even if they could be trained to react, there’s no way to know whether they learned to react at the minimum olfactory threshold or somewhere far above it. Even dogs may learn “bark when it’s pretty smelly (for a dog)” and not “bark when you detect the slightest possible smell”.
Nobody has yet mentioned Feynman’s adventures in smell, recounted in “Surely you’re joking, Mr. Feynman!”
I notice Feynman doesn’t disclose whether she was wearing perfume or not. Or whether she used a scented lotion. Or etc. Yes. That would be easy.
Dogs would detect your own un-perfumed body’s scent on the book, which I don’t think many humans can do.
Here is some more on Feynman’s olfaction.
http://www.salon.com/1999/07/28/smell/
I think it was in “Making of the Fittest” (I can’t find my copy just now to confirm) that the author correlated diminished sense of smell in primates with development of red/green cones in color vision. The explanation went that if you can spot tender, young, often reddish leaves at a distance then you have no need to sniff them out.
It should therefore be interesting to correlate size of olfactory lobes with two-color vs three-color vision.
That explanation sounds a bit like a “Just So Story.”
That’s Sean B Carroll’s book, isn’t it? I can’t remember – do all Old World apes (not just ourselves) have poor sense of smell? From what I recall trichromatic vision arose after the separation of the continents and, hence, ape species.
The discussion of pseudogenes leaves me somewhat confused. What does it mean to say that we have 1000 “olfactory receptor genes” but only 390 of them code for “receptor proteins”? What else would an olfactory receptor gene code for? I get that many of them are deactivated, but if that means they no longer count as protein-coding genes, why would they still count as receptor genes?
If “protein-coding” requires that a gene be transcribed and translated, wouldn’t that put us in the peculiar position of having genes that are “protein-coding” in some cells but not others?
Sorry, what I mean is that we have 1000 genes that at one time coded for olfactory receptor proteins (in our ancestors), but only 390 of those ancestral genes remain active, known to produce a protein. Does that clear things up?
Yes, thanks!
I’ve read something that looks supporting Freud: that there is a cranial nerve located before the olfactory nerve that is so tiny in humans that got overlooked; later found in other species, it was called “zero” or “terminal nerve”; it starts from olfactory epithelium but instead of the olfactory bulbs, reaches a region controlling sexual behavior.
My thought is that while we do not have the sheer ability to smell small quantities of odors that are important to dogs, we actually have a remarkable ability to smell the forest by-products that signal the presence of water in extremely tiny concentrations.
Also, it wouldn’t surprise me if we don’t have a much greater ability to synthesize and understand odors – evolution has given us what WE need, not what dogs, or some other animals, need.
Sub